PERINATAL BRAIN DAMAGE IN VERY PRETERM INFANTS Prenatal inflammation and neurologic outcome in children born term and preterm

TUULA KAUKO LA Faculty of Medicine, Department of Paediatrics, Biocenter Oulu, University of Oulu

OULU 2005 Tiivistelmä suomeksi

TUULA KAUKOLA

PERINATAL BRAIN DAMAGE IN VERY PRETERM INFANTS Prenatal inflammation and neurologic outcome in children born term and preterm

Academic Dissertation to be presented with the assent of the Faculty of Medicine, University of Oulu, for public discussion in the Auditorium 12 of the Department of Paediatrics, Oulu University Hospital, on October 21st, 2005, at 12 noon

O U L U N Y L I O P I S TO, O U L U 2 0 0 5

Copyright © 2005 University of Oulu, 2005

Supervised by Professor Mikko Hallman Docent Leena Vainionpää

Reviewed by Professor Leena Haataja Docent Anna-Liisa Järvenpää

ISBN 951-42-7839-9 (nid.) ISBN 951-42-7840-2 (PDF) http://herkules.oulu.fi/isbn9514278402/ ISSN 0355-3221

OULU UNIVERSITY PRESS OULU 2005

http://herkules.oulu.fi/issn03553221/

Kaukola, Tuula, Perinatal brain damage in very preterm infants. Prenatal inflammation and neurologic outcome in children born term and preterm Faculty of Medicine, Department of Paediatrics, Biocenter Oulu, University of Oulu, P.O.Box 5000, FIN-90014 University of Oulu, Finland 2005 Oulu, Finland

Abstract Despite improvements in peri- and neonatal care and an increase in the overall survival of very preterm infants, the incidence of neurologic sequelae has remained high. The pathogenesis of many brain imaging findings, such as white matter damage, WMD, is poorly understood. The factors predisposing to brain damage differ between term and preterm infants. More detailed information is needed of how brain imaging correlates with neurodevelopmental impairment after the neonatal period. The present study investigated the pre- and perinatal factors leading to brain damage and their effects on neurologic and neurodevelopmental outcome in very preterm children. We also analyzed the differences in umbilical cord serum cytokines in term and preterm children with cerebral palsy, CP. Furthermore, the correlations between the findings on diffusion-weighted imaging, DWI, measurements in brainstem auditory evoked potentials, and neurodevelopmental outcome were assessed. We demonstrated that pregnancies complicated by combined histologic chorioamnionitis and placental insufficiency independently predicted abnormal neurologic outcome at 2 years of corrected age. WMD additively predicted poor outcome. Isolated fetal inflammatory response, umbilical cord serum acute phase cytokines (IL-1α, IL-1β, IL-6, IL-8, TNF-α), did not associate with neurologic outcome in either term or preterm children. Instead, a cluster of cytokines different from acute phase cytokines were related to CP, and the protein profile differed between term and preterm children. Disturbed hemodynamics during the pre- and perinatal period affected outcome in very preterm infants. In severe placental insufficiency, fetal cardiac compromise associated with suboptimal neurodevelopmental outcome at 1 year of corrected age. In addition, several clinical factors characterising cardiorespiratory status after birth associated with abnormal neurologic outcome at 2 years of corrected age. We found the apparent diffusion coefficient, ADC, a quantitative measurement of water diffusion, in pons to correlate with the conduction rate of impulses travelling through the auditory tract. We also demonstrated a high value of ADC in corona radiata to associate with poor outcome in gross motor and eye-hand coordination skills at 2 years of corrected age. Both pre- and perinatal factors associate with later outcome in very preterm infants. An isolated fetal inflammatory response does not predict neurologic outcome. Findings on DWI in specific brain regions predict abnormal neurodevelopmental outcome.

Keywords: brain, cytokines, diffusion magnetic resonance imaging, infant, inflammation, neurologic outcome, placenta, preschool child, preterm birth

Kaukola, Tuula, Syntymänaikainen aivovaurio hyvin ennenaikaisilla vastasyntyneillä. Syntymää edeltävän tulehduksen vaikutus neurologiseen kehitykseen täysiaikaisena ja ennenaikaisena syntyneillä lapsilla Lääketieteellinen tiedekunta, Lastentautien klinikka, Biocenter Oulu, Oulun yliopisto, PL 5000, 90014 Oulun yliopisto 2005 Oulu

Tiivistelmä Huolimatta vastasyntyneisyyskauden parantuneista hoitotuloksista ja että yhä useampi hyvin ennenaikaisena syntynyt lapsi jää eloon, heidän neurologisen vammautuneisuuden ilmaantuvuus on edelleen korkea. Monien aivojen kuvantamislöydösten, kuten valkean aineen vaurion, syntymekanismit tunnetaan huonosti. Aivojen vaurioitumiselle altistavat tekijät eroavat täysiaikaisena ja ennenaikaisena syntyneillä lapsilla. Tarvitaan myös aiempaa yksityiskohtaisempaa tietoa aivojen kuvantamislöydösten merkityksestä lasten vastasyntyneisyyskauden jälkeiseen kehitykseen. Tässä tutkimuksessa selvitettiin raskauden- ja syntymänaikaisia tekijöitä, jotka vaikuttavat aivojen vaurioitumiseen hyvin ennenaikaisena syntyneillä lapsilla sekä näiden tekijöiden merkitystä lasten neurologiseen kehitykseen. Tarkastelimme myös napaveren seerumin välittäjäaineiden, sytokiinien, eroavuuksia täysiaikaisena ja ennenaikaisena syntyneillä CP-lapsilla. Lisäksi selvitimme diffuusiomagneettitutkimus- ja aivorunkoherätevastelöydösten sekä neurologisen kehityksen välisiä yhteyksiä. Tämän tutkimuksen mukaan kohdunsisäinen tulehdus ja istukan vajaatoiminta yhtä aikaa esiintyessään ovat poikkeavan neurologisen kehityksen itsenäisiä riskitekijöitä lapsilla 2 vuoden korjatussa iässä tutkittuna. Valkoisen aivoaineen vaurio edelleen lisäsi näiden lasten huonon neurologisen kehityksen ennustetta. Raskauden kestosta riippumatta, sikiön tulehdusvastetta kuvaavat napaveren akuutin vaiheen tulehdusvälittäjäaineet (IL-1α, IL-1β, IL-6, IL-8, TNF- α) eivät vaikuttaneet lapsen neurologiseen kehitykseen. Sen sijaan, CP-lasten napaverestä löytyi erityinen joukko ei-akuutin vaiheen välittäjäaineita. Nämä valkuaisaineet erosivat toisistaan täysiaikaisena ja ennenaikaisena syntyneillä CP-lapsilla. Raskauden- ja syntymänaikaiset verenkierron häiriöt vaikuttivat hyvin ennenaikaisena syntyneiden lasten myöhempään kehitykseen. Vaikeassa istukan vajaatoiminassa sikiön sydämen toiminnan heikkeneminen liittyi lapsen suboptimaaliin neurologiseen kehitykseen 1 vuoden korjatussa iässä tutkittuna. Lisäksi useat syntymänjälkeiset keuhkojen ja verenkierron tilaa kuvaavat kliiniset tekijät liittyivät lapsen poikkeavaan neurologiseen kehitykseen 2 vuoden korjatussa iässä tutkittuna. Tutkimuksemme mukaan, veden diffuusiota määrällisesti kuvaava diffuusiokerroin, ADC, aivosillasta mitattuna, liittyi impulssien johtumisnopeutueen kuuloradastossa. Lisäksi korkea ADCarvo aivojen sepelviuhkassa liittyi karkean motoriikan ja silmä-käsi-yhteistyötaitojen huonoon kehitykseen 2 vuoden korjatussa iässä tutkittuna. Sekä raskauden- että syntymänaikaiset tekijät vaikuttavat hyvin ennenaikaisena syntyneiden lasten myöhempään kehitykseen. Yksittäinen sikiön tulehdusvaste ei ennakoi lapsen neurologista kehitystä. Tiettyjen aivoalueiden diffuusiokuvantamislöydökset ennustavat lapsen poikkeavaa neurologista kehitystä.

Asiasanat: aivot, diffuusiomagneettikuvantaminen, ennenaikainen synnytys, istukka, lapsi, neurologinen kehitys, sytokiinit, tulehdus

To children born premature

Acknowledgements This work was carried out at the Department of Paediatrics, University of Oulu, during the years 1998–2004. I wish to express my deepest gratitude to my supervisor Professor Mikko Hallman, M.D., Ph.D. Head of the Department of Paediatrics. It has been a privilege to work in his research group. His everlasting optimism and encouragement throughout these years have been essential for me. I admire his enthusiastic attitude to science and his immeasurable knowledge about neonatology. I also wish to acknowledge Professor Matti Uhari, M.D., Ph.D., for teaching me critical scientific thinking during these years. I owe my most sincere gratitude to my second supervisor, Docent Leena Vainionpää, M.D., Ph.D. She has a special skill of practical guidance not only in child neurology but also in the field of science. Besides science, we have shared many joyful moments by a cup of coffee in between the busy moments of the workdays. I express my sincere thanks to Professor Marjatta Lanning, M.D., Ph.D., Chief Physician of the Department of Paediatrics, for her supportive attitude towards scientific research. I sincerely thank my official referees, Professor Leena Haataja, M.D., Ph.D. and Docent Anna-Liisa Järvenpää, M.D., Ph.D., for their valuable and constructive criticism during the preparation of the final version of my thesis. I wish to warmly express my gratitude to my co-authors, Docent Juha Räsänen, M.D., Ph.D., for introducing me to the field of perinatology and initiating me into the care of a fetus as a patient, Docent Riitta Herva, M.D., Ph.D., for her excellent knowledge about human pathology, Docent Uolevi Tolonen, M.D., Ph.D., for his professionalism in clinical neurophysiology, and Docent Eija Pääkkö, M.D., Ph.D., and Dr. Marja Perhomaa, M.D., for their skilful evaluation of brain ultrasound and MRI scans. I appreciate your flexible co-operation and the most supportive attitude towards my research work during these years. It has been a pleasure to work with you all. My special thanks are due to Docent Pentti Koskela, M.D., Ph.D., Docent Helena Pihko, M.D., Ph.D., Docent Outi Tammela, M.D., Ph.D., and Docent Tuula Äärimaa, M.D., Ph.D., for their valuable assistance. I also want to acknowledge Professor Aimo Ruokonen, M.D., Ph.D., and Jukka Jauhiainen, Ph.D., for their co-operation. I further wish to express my gratitude to

the collaborators in laboratory in New Haven, Connecticut, CT, and in Duke University Medical Center, Durham, North Carolina, NC. I thank Mrs. Sirkka-Liisa Leinonen, Lic.Phil., for her careful revision of the English language of my thesis. I owe my thanks to Risto Bloigu, M.Sc., for his skilful advice on statistical analyses, to Juha Turtinen, M.Sc., for his friendly help with technical problems in computers, and Mrs. Marjatta Paloheimo and Mrs. Maija Veikkola for their kind assistance throughout these years. I am most deeply grateful to all the babies and their families for attending this followup programme all the way from Sodankylä to Helsinki and to the wonderful staff in the neonatal intensive care unit. Without your sympathetic attitude, this work would not have been possible. I also want to express my thanks to my colleagues in the central hospitals in Rovaniemi, Kemi, Kajaani and especially Kokkola for their flexible co-operation during these years. The research nurses Liisa Siermala-Hiironen and Nana Jaakola have been irreplaceable in organizing innumerable practical matters concerning this research. You are always so positive-minded and ready to assist. I wish to express my most sincere thanks to my colleagues in the neonatal intensive care unit, especially to Marja-Leena Pokela, M.D., Ph.D., for her kind and supportive attitude and for teching me so much of practical neonatology during my fellow years in NICU, and to Timo Saarela, M.D., Ph.D., and Marita Valkama, M.D., Ph.D., for their kind encouragement during these years. I also want to thank Docent Maila Koivisto, M.D., Ph.D., for the valuable support and inspiring conversations we had when sharing the same research room together. I thank all my friends and colleagues in the Department of Peadiatrics. We have shared some inspiring and refreshing conversations in Pannuhuone. My very special thanks go to my friend and colleague Satu Lehtinen, M.D., Ph.D. We have shared both laughs and tears during the past years. I am ever so grateful for having a friend like you. I thank my dear friends Anne Ollikainen, M.D., and Mr. Esa Haapa-aho and Mr. and Mrs. Juha and Pirjo Nurmi for their warm and valuable friendship and also my little godchildren Saini Haapa-aho and Elias Nurmi for just being there. I am deeply grateful for your friendship. I wish to express my deepest thanks to my dear uncle Veikko and her wife Maire Juanto. I feel I come home when visiting you. You are always so kind and ready to help me. Your hearts are pure gold. Finally, I want to honor my dear late parents, Eino and Nina Kaukola. They always supported and encouraged me to study and to get on with my life. I am so proud of having been their daughter. This research was financially supported by Alma and K. A. Snellman Foundation, Oulu, Finland, Foundation for Paediatric Research, Finland, the Academy of Finland, Sigrid Jusèlius Foundation, and the Department of Paediatrics and Adolescence, Oulu University Hospital, and a research subsidy from the Ministry of Social Affairs and Health. Oulu, August 2005

Tuula Kaukola

Abbreviations ADC AF AMPA AoI AoV BAEP BBB BDNF BE BPD BV BW C CBF CCA CCO CI CINC CNS CNTF CP cPVL CRIB CSF DAo DEHSI DQ DV DWI EGF ELBW

apparent diffusion coefficient amniotic fluid alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate aortic isthmus aortic valve brainstem auditory evoked potentials blood brain barrier brain-derived neurotrophic factor base excess bronchopulmonary dysplasia bacterial vaginosis birth weight cysteine cerebral blood flow clinical chorioamnionitis combined cardiac output confidence interval cytokine-induced neutrophil chemoattractant central nervous system ciliary neurotrophic factor cerebral palsy cystic periventricular leukomalacia critical risk index for babies cerebrospinal fluid descending aorta diffuse extensive high signal intensity developmental quotient ductus venosus diffusion-weighted imaging epidermal growth factor extremely low birth weight

FHR FO FOV GA GABA GBS G-CSF GFAP GLT GM GM-CSF HCA HI ICAM IFN-γ IL IPL IQ IUGR IVC IVH kPa LHV LIF LPS LVCO MAP MCA MCP MDI MIP MRI mRNA MS NEC NF-κB NK NMDA NO NT O-2A OL OR PaO2 PDI PET

fetal heart rate foramen ovale field of view gestational age gamma-amino butyric acid group B streptococci granulocyte colony-stimulating factor glial fibrillary acid protein glutamate transporter germinal matrix granulocyte-macrophage colony-stimulating factor histologic chorioamnionitis hypoxia-ischemia intercellular adhesion molecule interferon gamma interleukin inter-peak latency intellectual quotient intrauterine growth retardation inferior vena cava intraventricular hemorrhage kilopascal left hepatic vein leukemia inhibitory factor lipopolysaccharide left ventricular cardiac output mean arterial pressure middle cerebral artery monocyte chemoattractant protein mental developmental index macrophage inflammatory protein magnetic resonance imaging messenger ribonucleic acid multiple sclerosis necrotizing enterocolitis nuclear factor- kappa B natural killer N-methyl-D-aspartate nitric oxide neurotrophin oligodendrocyte-2 astrocyte oligodendrocyte odds ratio partial pressure of oxygen, arterial psychomotor developmental index positron emission tomography

PG PHI PI PIV PLIC PMN PPROM PV PV-IVH PVE PVL Q RANTES RDS ROC ROI RR RVCO SD SEM SGA STV Th TLR TNF-α UA US VEGF VLBW WMD

prostaglandin periventricular hemorrhagic infarction pulsatility index pulsatility index for vein posterior limb of internal capsule polymorphonuclear preterm premature rupture of membranes pulmonary valve periventricular-intraventricular hemorrhage periventricular echodensity periventricular leukomalacia volume blood flow (ml/min) regulated upon activation, normal T-cell expressed and presumably secreted respiratory distress syndrome receiver operator curve region of interest relative risk right ventricular cardiac output standard deviation standard error of mean small for gestational age short-term variation T helper toll-like receptor tumor necrosis factor-alpha umbilical artery ultrasound vascular endothelial growth factor very low birth weight white matter damage

For abbreviations of the cytokines, readers are referred to page 55.

List of original papers The thesis is based on the following articles, which are referred to in the text by their Roman numerals: I

Kaukola T, Satyaraj E, Patel DD, Tchernev VT, Grimwade BG, Kingsmore SF, Koskela P, Tammela O, Vainionpää L, Pihko H, Äärimaa T & Hallman M (2004) Cerebral palsy is characterized by protein mediators in cord serum. Ann Neurol 55:186–194.

II

Kaukola T, Herva R, Perhomaa M, Pääkkö E, Kingsmore S, Vainionpää L & Hallman M (2005) Chorioamnionitis and cord serum proinflammatory cytokines: lack of association with brain damage and neurologic outcome in very preterm infants. Pediatr Res, in press.

III Kaukola T, Räsänen J, Herva R, Patel DD & Hallman M (2005) Suboptimal neurodevelopment in very preterm infants is related to fetal cardiovascular compromise in placental insufficiency. Am J Obstet Gynecol 193:414–420. IV Kaukola T, Perhomaa M, Vainionpää L, Tolonen U, Jauhiainen J, Pääkkö E & Hallman M (2005) Correlation of apparent diffusion coefficient in diffusionweighted imaging with brainstem auditory evoked potentials and neurodevelopment in very preterm infants. Submitted for publication. The original articles are reprinted with permission from John Wiley & Sons, Inc., Copyright © (2004) American Neurological Association (I), Lippincott Williams & Wilkins (II), and American Journal of Obstetrics and Gynecology, Copyright © (2005) Elsevier Inc., (III).

Contents Abstract Acknowledgements Abbreviations List of original papers Contents 1 Introduction ................................................................................................................... 19 2 Review of literature ....................................................................................................... 21 2.1 Definitions and epidemiology of prematurity and low birth weight.......................21 2.2 Brain development and maturation.........................................................................22 2.2.1 Neuronal migration and organization ..............................................................22 2.2.2 Synaptogenesis ................................................................................................23 2.2.3 Vascularization.................................................................................................23 2.2.4 Brain folding....................................................................................................24 2.2.5 Myelination .....................................................................................................24 2.3 Brain damage..........................................................................................................25 2.3.1 Periventricular leukomalacia ...........................................................................25 2.3.1.1 Incidence...................................................................................................26 2.3.1.2 Pathogenesis .............................................................................................26 2.3.2 Gray matter damage ........................................................................................29 2.3.3 Periventricular-intraventricular hemorrhage and infarction.............................29 2.3.3.1 Incidence...................................................................................................29 2.3.3.2 Pathogenesis .............................................................................................30 2.4 Placental insufficiency............................................................................................30 2.4.1 Placental hemodynamics .................................................................................31 2.4.2 Fetal hemodynamics ........................................................................................32 2.5 Cytokines................................................................................................................33 2.5.1 Classification ...................................................................................................33 2.5.2 Function...........................................................................................................33 2.5.3 Cytokines and brain .........................................................................................35

2.6 Chorioamnionitis ....................................................................................................38 2.6.1 Brain damage...................................................................................................39 2.6.2 Cerebral palsy..................................................................................................41 2.7 Brain imaging .........................................................................................................44 2.8 Neurologic and neurocognitive outcome of ELBW infants....................................47 3 Purpose of the study ...................................................................................................... 50 4 Subjects and methods .................................................................................................... 51 4.1 Subjects ..................................................................................................................52 4.2 Methods ..................................................................................................................53 4.2.1 Analysis of cord serum proteins ......................................................................53 4.2.2 Data quality control .........................................................................................54 4.2.3 Pathology of placenta ......................................................................................55 4.2.4 Brain imaging ..................................................................................................56 4.2.4.1 Brain ultrasound .......................................................................................56 4.2.4.2 Magnetic resonance imaging ....................................................................56 4.2.5 Placental and fetal hemodynamics...................................................................57 4.2.6 Brainstem auditory evoked potentials..............................................................58 4.2.7 Neurologic and neurodevelopmental outcome ................................................58 4.2.8 Pre- and postnatal clinical data ........................................................................59 4.3 Statistics..................................................................................................................59 5 Results ........................................................................................................................... 61 5.1 Prenatal inflammation (I, II, III) .............................................................................61 5.1.1 Umbilical cord serum proteins and gestational age (I, II)................................61 5.1.2 Placental pathology, cytokines and active preterm labor (II, III).....................62 5.1.3 IVH grade II to III, WMD and inflammation (II) ............................................63 5.1.4 Inflammation and neurologic outome in preterm children (I–III) ...................64 5.1.5 Cytokines and cerebral palsy in term-born children (I) ...................................65 5.2 Pre- and neonatal risk factors and neurologic outcome (II)....................................66 5.3 Placental insufficiency and neurodevelopment (III)...............................................68 5.4 Functional neuroanatomy (IV) ...............................................................................69 5.4.1 Association between ADC and BAEP .............................................................69 5.4.2 Neuroanatomy and functional neurodevelopment ...........................................69 6 Discussion ..................................................................................................................... 71 6.1 Prenatal inflammation and brain.............................................................................71 6.2 Cytokine profiles in term and preterm CP children ................................................74 6.3 Placental insufficiency and neurodevelopmental outcome .....................................76 6.4 DWI and correlation with BAEP and neurodevelopmental outcome .....................77 7 Conclusions ................................................................................................................... 80 References

1 Introduction Prematurity continues to be a leading cause of peri- and neonatal mortality and morbidity. Despite significant improvements in perinatal care and an increase in the overall survival rate of very preterm infants during the last decade, the incidence of later neurologic sequelae has remained high (Hagberg et al. 2001). Intraventricular hemorrhage and white matter injury, referred to as cystic periventricular leukomalacia, are the well-known brain lesions leading to long-term neurologic and neurodevelopmental deficits. Conventional magnetic resonance imaging (MRI) as well as brain ultrasound (US) detect these lesions accurately. A specialized MRI technique, diffusion-weighted imaging (DWI), allows visualization of the more subtle and diffuse changes that predominantly affect the most premature infants. Apparent diffusion coefficient (ADC) is a quantitative indicator of molecular diffusion of water, and it relates to the organization of fiber tracts and the myelination process in white matter (Morriss et al. 1999). The pathogenesis of white matter injury has remained poorly understood. It is suggested to involve a complexity of predisposing factors (Saliba & Marret 2001) acting additively or synergistically during the pre- and perinatal period. The factors predisposing to brain damage differ between term and preterm infants (Hagberg et al. 2001). In addition, more detailed information is needed of how MRI findings correlate with the neurodevelopmental impairments seen after the neonatal period. Cytokines are small soluble proteins that are released by different cell types in the body. They have important functions in both innate and adaptive immunity to induce a response against an activating stimulus during infection and inflammation. Moreover, some cytokines act as growth factors and enhance maturation, while others promote the survival and regeneration of a target organ. Intrauterine infection associates with spontaneous preterm birth (Hillier et al. 1988). Elevated cytokine levels of fetal origin and histologic chorioamnionitis have previously been related with brain damage in both term and in preterm children (Yoon et al. 1996, Yoon et al. 1997a, Nelson et al. 1998). Recent evidence suggests that prenatal factors are the main determinants of cerebral palsy (CP) among children born at term, but that perinatal/neonatal etiology accounts for some 80% of CP in children born before 32 weeks of gestation (Hagberg et al. 2001). Apart from intrauterine inflammation, reduction in uteroplacental blood flow has an effect on fetal well-being. Placental insufficiency relates to increased perinatal mortality

20 (Kurkinen-Räty et al. 1997). Longitudinal studies have demonstrated that changes in placental and fetal hemodynamics appear progressively, and that the late changes relate to short-term peri- and neonatal outcome (Hecher et al. 2001, Ferrazzi et al. 2002). The objectives of the present study were to investigate the pre- and perinatal factors leading to brain damage and their effects on later neurologic and neurodevelopmental outcome. We also aimed to íncrease the current knowledge of DWI and to assess the correlates between neurodevelopmental outcome and findings on DWI. The study included both children born at term and very preterm children born before 32 weeks of gestation.

2 Review of literature 2.1 Definitions and epidemiology of prematurity and low birth weight Preterm birth is defined as birth before 37 completed weeks of gestation. There are significant racial differences in the rates of preterm birth and low birth weight. The percentage of preterm birth rose from 9.4% in 1981 to 11.6% in 2000 in the United States (MacDorman et al. 2002). In Finland, the incidence of preterm birth has shown an upward trend, increasing from 5.6% in 1987 to 6.3% by the year 2000. In 2003, the incidence of preterm birth again declined to 5.7% (Vuori & Gissler 2004). The rate of very preterm birth (< 32 completed weeks of gestation) in the United States rose from 1.81% in 1981 to 1.93% in 2000 (MacDorman et al. 2002). In Finland, the proportion of very preterm birth out of all births was 0.86% in 2003 (Vuori & Gissler 2004). In the United States, the rate of very low birth weight (VLBW, birth weight < 1500 grams) infants was 1.43% in 2001 (MacDorman et al. 2004). In Finland, there was a moderate decline in the incidence of VLBW infants from 0.9% in 2001 to 0.7% in 2003, and a similar trend was seen in the incidence of extremely low birth weight (ELBW, birth weight < 1000 grams) infants. In 2001 ELBW infants constituted 0.4% of all births, while in 2003 they accounted for 0.3% of all births (Vuori & Gissler 2004). During the past decades, the overall survival of preterm infants with birth weight between 500 to 1000 grams has increased up to 80% (Goldenberg et al. 2000). Survival has been reported to rise with each completed week in utero by 6 to 9% for newborns delivered after 23 completed weeks. Ninety percent of infants born at 27 to 28 weeks of gestation now survive (Ward & Beachy 2003). Reported survival at 22 weeks of gestation has ranged from 0% (Allen et al. 1993) to 21% (Lemons et al. 2001). In infants born at 23 completed weeks, survival rate has varied from 30% (Lemons et al. 2001) to near 50% (Ward & Beachy 2003). In Finland, the reported survival rates after the perinatal period in 2003 were 4.2%, 38.9%, 50.0%, 48.3%, and 82.4% among infants born at 22, 23, 24, 25, and 26 completed weeks, respectively (Vuori & Gissler 2004). In 2003, 54.7% of all

22 ELBW infants and 77.0% of all live-born ELBW infants survived the perinatal period (Vuori & Gissler 2004). Twenty to twenty-five percent of all preterm births are induced due to medical or obstetric conditions that deteriorate the health of mother, fetus, or both (Lockwood & Kuczynski 1999, Iams 2003). Seventy to eighty percent of preterm births are spontaneous due to preterm labor with intact membranes or to preterm premature rupture of membranes (PPROM) (Lockwood & Kuczynski 1999, Iams 2003). The most common reasons for induced preterm birth are preeclampsia, fetal distress, intrauterine growth restriction, placental abruption, and fetal demise (Meis et al. 1998). Uterine abnormality, proteinuria before 24 weeks of pregnancy, history of chronic hypertension, previous induced preterm birth, maternal lung disease, previous spontaneous preterm birth, maternal age over 30 years, black ethnicity, and working during pregnancy increase the risk of induced preterm birth (Meis et al. 1998). Risk factors associated with spontaneous preterm birth include young maternal age (OR 2.0, 95%CI 1.43–2.81), low maternal weight (OR 1.83, 95%CI 1.33–2.53), low (OR 1.32, 95%CI 1.06–1.65) or high parity (OR 1.48, 95%CI 1.13–1.94), previous abortion (OR 1.24, 95%CI 1.03–1.49), vaginal bleeding (OR 1.96, 95%CI 1.61–2.39), cigarette smoking (OR 1.33, 95%CI 1.12–1.59) (Meis et al. 1995), bacterial vaginosis (OR 2.19, 95%CI 1.54–3.12) (Leitich et al. 2003), severe maternal periodontal disease (OR 4.45, 95%CI 2.16–9.18) (Jeffcoat et al. 2001), black race (OR 1.4, 95%CI 1.1–1.7), and short cervical length (OR 7.7, 95%CI 4.5–13.4) (Iams 2003).

2.2 Brain development and maturation The neural plate becomes recognizable during the third postconceptional week. It forms the neural tube at the end of the fourth postconceptional week. (Evrard et al. 1993). The anterior end of the neural tube closes at approximately 24 days and the posterior end at approximately 26 days (Volpe 1995). During the following three weeks, the neural tube enlarges rostrally and forms the primary brain vesicles: prosencephalon, mesencephalon, and rhombencephalon. During prosencephalic cleavage, five secondary brain vesicles: two paired vesicles of telencephalon, (cerebral hemispheres, lateral ventricles, basal ganglia), diencephalon, (thalamus, hypothalamus), mesencephalon (midbrain), and metencephalon (pons, cerebellum), are formed. Midline prosencephalic development occurs from the latter half of the second month through the third month, and it is crucial for the formation of corpus callosum, septum pellucidum, optic nerve-chiasm, and hypothalamus. Forebrain is formed of prechordial mesoderm (Evrard et al. 1992, Volpe 1995).

2.2.1 Neuronal migration and organization The neural tube is composed of a single germinal epithelial layer of highly proliferative cells that give rise to all major classes of cell types in the central nervous system (CNS), i.e. both neurons and glial cells (McKay 1997). After proliferation and differentiation,

23 newly formed neurons (neuroblasts) migrate mostly radially (Volpe, 1995, Gressens 2002) along the glial processes out from the germinative zone (ventricular zone) to establish a six-layer neocortex, following an inside-out pattern: each neuron bypasses the earlier generated neurons and comes to lie between the subplate layer and the marginal layer (Rakic 1971). Each proliferative unit in the ventricular zone has a distinct corresponding area in the neocortex (Rakic 1988). In humans, neuronal migration occurs mostly between the 12th and the 24th weeks of gestation (Volpe 1995, Gressens 2002) (Table 1). Some neurotransmitters, e.g. glutamate and gamma-aminobutyric acid (GABA), as well as some growth factors play a role in the control or modulation of neuronal migration (Gressens 2002). The germinative zone involutes during the third trimester of gestation (Volpe 1995). The subplate zone both reaches its maximal width at the beginning of the third trimester and begins to dissolve during the third trimester (Kostovic & Rakic 1990). Subplate neurons are largely absent after 6 months of postnatal age (Volpe 1995) and undergo programmed cell death to a much greater extent than other cortical neurons (Price et al. 1997). The transient subplate zone is an important site for the development of connections between the thalamus and the cerebral cortex (Kostovic & Rakic 1990). The germinative zone (germinal matrix) can be visualized by conventional MRI (Battin et al. 1997, Battin et al. 1998). MRI and diffusion tensor imaging can reveal cortical maturation of the human brain (Girard et al. 1995, McKinstry et al. 2002a).

2.2.2 Synaptogenesis Synaptic formation for intercellular communication begins during the period of neuronal migration (Kostovic et al. 1989), but the most rapid phase of synaptogenesis occurs after the neurons have completed their migration, at 20 to 24 weeks of gestation, and proceeds until the first months after birth (Volpe 1995, Bourgeios 2002) (Table 1). The rapid phase of synaptogenesis is important for the establishment of different cortical functions, including motor, sensory, and cognitive functions. The kinetics of synaptic formation and elimination differ between regions in the human brain, although within a certain time window (Bourgeois 2002).

2.2.3 Vascularization Angiogenesis starts during the first half of gestation (Table 1). In telencephalon, the first vessels originate from the primitive leptomeningeal plexus during the second month of gestation. Around the 12th week of gestation these vessels reach the subventricular germinative zone, where they form a subventricular plexus (Duckett 1971). During the second half of gestation these radial vessels originating from the leptomeningeal vascular bed form horizontal branches and recurrent collaterals, which extend into the cortical layers in an inside-out direction (Kuban & Gilles 1985, Norman & O'Kusky 1986). The total number of cerebral arteries reaches its maximum at birth, after which rarefaction begins to occur (Pearce 2002). The dense network of vessels in the germinative zone

24 lacks a smooth muscular layer. They are lined by a mere single layer of endothelial cells and are thus vulnerable to rupture in response to changes in cerebral hemodynamics (Kuban & Gilles 1985, Anstrom et al. 2004). The blood-brain barrier (BBB) is formed by brain endothelial cells lining the cerebral vasculature.

2.2.4 Brain folding Anatomo-pathological studies describe fetal brains as lissencephalic up to week 18 of pregnancy (Chi et al. 1977). Sulcal and gyral development starts in the central areas and continues to involve the parieto-occipital cortex. The frontal cortex develops last (Battin et al. 1998, Ruoss et al. 2001). The interhemispheric fissure is the earliest primary fissure to appear. The sylvian fissure and central sulcus appear at approximately 14 and 20 weeks of gestation, respectively (Chi et al. 1977). Between the weeks 28 and 30, there is an increase in the number of sulci and gyri as well as maturational and developmental evolution (Dorovini-Zis & Dolman 1977) (Table 1). In MRI studies, cortical surface area has been reported to increase by 12% per week at the gestational age between 25 and 42 weeks (Ajayi-Obe et al. 2000). At 34 to 36 weeks of gestation, the normal pattern of gyral maturation on MRI should include the presence of the marginal sulcus and paracentral gyrus. By 36 to 38 weeks, the anterior and posterior orbital gyri should be established. The inferior temporal and inferior occipital gyri and sulci should be established at the gestational age of 40 weeks (Inder et al. 2003a). Anatomical studies (Chi et al. 1977) detect the beginning of cortical folding earlier than in vivo imaging techniques (Battin et al. 1998, Levine & Barnes 1999, Ruoss et al. 2001). Disturbance in cortical convolution has previously been reported in preterm infants without parenchymal lesions compared to infants born at term (Ajayi-Obe et al. 2000).

2.2.5 Myelination From 20 weeks on, glial cells proliferate and differentiate to form astroglia and oligodendroglia, called myelination gliosis (Girard et al. 1995) (Table 1). Myelin, which is produced by mature oligodendrocytes, is composed of a bilayer of lipids with several large proteins. The outer layer is composed mainly of cholesterol and glycolipids, and the inner layer is composed of phospholipids. The protein molecules, including myelin basic protein and proteolipid protein, span the bilayer (Barkovich et al. 1988). The premyelination changes taking place prior to the development of the myelin sheath, have been attributed to the thickening of the axolemmal membrane and the development of the transmembrane pumps, which restrict water motion across the axon (Morriss et al. 1999). Histologically, myelination has been demonstrated to occur in a predictable fashion, beginning at the second trimester and continuing at least until the second year of life (Yakolev 1967, Kinney et al. 1988). Complete myelination is reached in early adulthood (van der Knaap & Valk 1990). The progression of myelination takes place in a distinct caudo-cranial order and proceeds from central white matter towards the poles, as shown both in autopsy specimens (Kinney et al. 1988) and on MRI (McArdle et al. 1987, Girard

25 et al. 1995, Sie et al. 1997, Battin et al. 1998, Hüppi et al. 1998a, Counsell et al. 2002). On MRI, myelination reaches the brainstem by 29 weeks of gestation and centrum semiovale by 42 weeks of gestation (Hüppi 2002). The visualization of myelination on MRI lags some weeks behind compared to its histologic appearance, histology being able to show small amounts of myelin indiscernible by MRI (Sie et al. 1997). The absolute volume of myelinated white matter has been reported to increase fivefold between 29 and 41 weeks of gestation (Hüppi et al. 1998a). The maturity of myelination at term differs between term and preterm infants (Hüppi et al. 1996). Table 1. Timetable of brain maturation. Gestational age

Migration

6th week–8th week 12th week–17th week

Neuronal

Synaptogenesis

Vascularization

Synaptogenesis

Vascularization

Brain folding

Synaptogenesis

Vascularization

Brain folding

Synaptogenesis

Vascularization

Brain folding

Synaptogenesis

Vascularization

Brain folding

Myelination

migration 20th week–24th week

Neuronal

Myelination

migration Birth 2–4–8–12 months---

Synaptogenesis

Myelination Myelination

2.3 Brain damage 2.3.1 Periventricular leukomalacia Periventricular leukomalacia (PVL) refers to white matter necrosis of a characteristic distribution, i.e. dorsal and lateral to the external angles of the lateral ventricles, followed after hours and days by microglial (macrophage) infiltration, axonal swelling, and astrocytic proliferation, and ending with cyst formation in the lesion some weeks later. Cystic PVL (cPVL) represents a focal necrotic lesion with loss of all cellular elements (Banker & Larroche 1962). Diffuse PVL undergoes less commonly cystic changes, usually in less mature infants, and the histologic characteristics include diffuse loss of oligodendrocytes, prominent microglial activation throughout white matter, and an increase of hypertrophic astrocytes (Paneth et al. 1990, Haynes et al. 2003). Myelin loss, manifested as diminished volume of cerebral white matter and ventricular dilatation, is a later correlate (Volpe 1995, Haynes et al. 2003). Figure 1 shows the anatomic location of both cPVL and diffuse PVL. Prenatal clinical factors that have been associated with PVL include gestational age at birth (OR 0.8, 95%CI 0.7–0.9) (Vergani et al. 2004), placental vascular anastomoses in monochorionic multiple pregnancies (Bejar et al. 1988, Zupan et al. 1996), intrauterine growth retardation (IUGR) (OR 2.7, 95%CI 1.3–5.6), oligohydramnion (OR 2.9, 95%CI 1.2–7.0) (Gaffney et al. 1994), and intrauterine inflammation (Wu & Colford, Jr. 2000). Perinatal events that have been associated with PVL include hypocarbia (OR 5.43, 95%CI 1.33–22.2) (Wiswell et al. 1996), intraventricular hemorrhage (IVH) grade ≥ 2

26 (OR 5.3, 95%CI 1.2–23.0) (Yoon et al. 1997a), mechanical ventilation (Graziani et al. 1992), patent ductus arteriosus (Shortland et al. 1990), episodes of apnea with bradycardia (Perlman & Volpe 1985), and outborn delivery (de Vries et al. 1988).

Fig. 1. Anatomic location of periventricular hemorrhagic infarction and focal and diffuse periventricular leukomalacia (Olsén & Vainionpää 2000. Reprinted with permission from Duodecim).

2.3.1.1 Incidence The incidence of cPVL in infants with birth weight (BW) less than 1500 g or gestational age under 33 weeks at birth ranges from 3% to 15% (Zupan et al. 1996, Perlman et al. 1996, Baud et al. 1999a, Leviton et al. 1999, Inder & Volpe 2000, Yanowitz et al. 2002, Inder et al. 2003a, Blumenthal 2004). It declines with increasing gestational age at birth (Zupan et al. 1996, Baud et al. 1999b). The peak incidence occurs among infants born between 27 and 30 weeks of gestation (Perlman et al. 1996, Zupan et al. 1996). The incidence of diffuse PVL has been reported to range between 16% to 75%, depending on population size and the definition used for the diffuse brain injury detected on conventional MRI (Maalouf et al. 1999, Counsell et al. 2003, Inder et al. 2003a).

2.3.1.2 Pathogenesis The pathogenesis of PVL consists of at least three major interacting factors: vascular immaturity (Takashima & Tanaka 1978) and impaired cerebrovascular autoregulation (Lou et al. 1979), which increase the risk of cerebral hypoperfusion, and the intrinsic vulnerability of differentiating oligodendrocytes within white matter (Back et al. 1998). The penetrating cerebral vessels, which include long and short branches of the anterior, middle, and posterior cerebral arteries, form the border zones in deep periventricular white matter and subcortical white matter, respectively (De Reuck 1971, Takashima & Tanaka 1978, De Reuck 1984). cPVL occurs principally within the end zones of the long penetrating arteries. Diffuse PVL occurs principally in the border zones

27 between the long penetrating arteries and in the end zones of the short penetrating arteries (Volpe 1997). From 32 weeks on, there is an increase in vascular supply as a result of an increase in the length and anastomosis of the blood vessels (Volpe 1997, Shalak & Perlman 2002). The minimum cerebral blood flow (CBF) needed to sustain neuronal viability in preterm infants is unknown. Positron emission tomography (PET) studies have reported the mean CBF to be as low as 4.9 ml/100g/min in preterm infants who survived with normal neurologic outcome later on (Altman et al. 1988). A persistent decrease in CBF has been reported in white matter during reperfusion followed by hemorrhagic hypotension in premature lamb despite recovery of CBF in all other brain regions (Szymonowicz et al. 1990). In a study of 12 normotensive and normoxic preterm infants at a mean gestational age of 27.7 wk at birth, cortical blood flow exceeded five times global CBF, suggesting that blood flow into white matter in preterm infants is extremely low (Borch & Greisen 1998). Cerebrovascular autoregulation has been shown to be intact both in guinea pigs and in human infants born preterm (Papile et al. 1985, Tsuji et al. 2000). These infants have an ability to maintain constant CBF despite changes in mean arterial blood pressure. In premature lamb, the pressure range over which autoregulation can maintain cerebral perfusion is narrow (Papile et al. 1985). In mechanically ventilated preterm infants and in newborns with birth asphyxia, CBF has shown to be pressure-passive and to correlate linearly with changes in mean arterial blood pressure (Lou et al. 1979, Tsuji et al. 2000). In premature infants, pressure-passive cerebral circulation has been related to an increased risk of PVL (Tsuji et al. 2000). Elevated carbon dioxide tension is one of the best-known vasodilatators in cerebral circulation. The relationship between hypocarbia and cPVL in human premature infants is contradictory (Wiswell et al. 1996, Dammann et al. 2001a). Presumably, hypocarbia contributes to PVL by inducing cerebral vasoconstriction (Saliba & Marret 2001). Oxidative damage to vital cellular structures contributes to the pathogenesis of brain injury in both immature and mature nervous systems (Palmer 1995). Before 30 weeks of gestation, the majority of human oligodendrocytes (OL) are in an early stage of development known as pre-OL, after which a progressive increase in a number of more mature OLs has been observed (Back et al. 2001). Vulnerability of immature white matter to hypoxia-ischemia has been suggested based on studies of hypoxic-ischemic injury caused by hemorrhagic hypotension (Matsuda et al. 1999) or by placental insufficiency (Mallard et al. 1998, Duncan et al. 2000) in fetal lamb. Having been subjected to 25-minute umbilical cord occlusion, the fetal lamb showed either PVL or diffuse injury in white matter and subcortical gray matter (Mallard et al. 2003). Impaired placental circulation together with hypoxemia has been shown to result in neuronal damage, reduced myelination, and focal white matter lesions in fetal lamb (Mallard et al. 1998). Reactive oxygen radicals are formed during ischemia-reperfusion (Fellman & Raivio 1997) and demonstrated in fetal sheep afterwards (Bagenholm et al. 1998). Cultured OL precursors, but not mature OLs, are shown to be vulnerable to free oxygen radicals (Back et al. 1998) and reactive nitrogen radicals (Baud et al. 2004). Activated microglia has been shown to release reactive oxygen radicals (Colton & Gilbert 1993) and nitric oxide (NO) in response to lipopolysaccharide (LPS) or cytokine (IFN-γ, IL-1β, or TNF-α) stimulation (Possel et al. 2000). Direct comparison between OL precursors and mature

28 OLs under conditions of free radical attack has shown that precursor cells accumulate free radicals, whereas mature OLs do not (Volpe 2001). Antioxidant administration in cultured rat pre-OLs suppressed free radical production and prevented OL death caused by free radical attack (Back et al. 1998). Elevated products of protein oxidation and lipid peroxidation in cerebrospinal fluid have been detected in preterm infants with white matter injury defined on MRI at term (Inder et al. 2002a). Markers of protein nitration (nitrotyrosine) and lipid peroxidation have been demonstrated in premyelinating OLs in autopsy series of premature infants with diffuse periventricular leukomalacia (Haynes et al. 2003). Markers for lipid peroxidation in gliotic white matter have been found to correlate with the expression of interferon-γ (IFN-γ) immunopositive glial cells in children with PVL (Folkerth et al. 2004a). The mechanism of the vulnerability of the immature OL is not known (Volpe 2001) but mitochondrial dysfunction (Baud et al. 2004), excessive local iron in immature OLs (Volpe 2001), and deficiency of enzymatic antioxidants, glutathione peroxidase, catalase (Vexler & Ferriero 2001), and superoxide dismutase (Folkerth et al. 2004b), in immature OLs have been suggested. Excessive free radical production may occur via a variety of mechanisms, many of which are commonly implicated in the pathogenesis of PVL. These mechanisms include ischemia-reperfusion (Bagenholm et al. 1998), infection, and inflammation (Inder et al. 2002b). Glutamate is an excitatory neurotransmitter in synapses, and it is also needed to control e.g. neuronal migration (Gressens 2002). The potential role of glutamate in the pathogenesis of PVL is based on the neuronal and axonal damage seen in PVL (Banker & Larroche 1962). Immature OLs (Yoshioka et al. 1996, Follett et al. 2000) and neuronal cells (Qiu et al. 1998, Conroy et al. 2004) express glutamate receptors. Glutamate release leads to receptor activation and calcium influx into the cell. An increase in intracellular calcium may trigger the production of toxic substances, e.g. NO, and free oxygen radicals and finally lead to neuronal cell death (Saliba & Marret 2001). Elevation in extracellular glutamate is suggested to lead to the death of OL precursors either by activation of a specific glutamate transport system (Oka et al. 1993) or by activation of alpha-amino-3hydroxy-5-methyl-4-isoxazole propionate (AMPA)/kainate glutamate receptors (Yoshioka et al. 1996, Follett et al. 2000). In cultured OL precursors, an increase in glutamate uptake leads to depletion of intracellular glutathione and to free radicalmediated cell death, which can be prevented by free radical scavengers, i.e. vitamin E (Oka et al. 1993) or N-acetyl-L-cysteine (Yoshioka et al. 1996). Receptor-mediated glutamate-induced OL death has also been established. In seven-day-old rats, white matter injury caused by hypoxia-ischemia was attenuated by a non-NMDA glutamate receptor antagonist (Follett et al. 2000). In mouse pups, the cytokines tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, IL-6, and IL-9 have been shown to potentiate PVL damage caused by a glutamateric analogue, ibotenate (Dommergues et al. 2000).

29

2.3.2 Gray matter damage Cortical and subcortical gray matter volumes at term have been shown to be reduced in preterm infants with white matter injury (Inder et al. 1999a, Inder et al. 2003a, Inder et al. 2005). The cause of this alteration of gray matter development in association with white matter injury is unknown, but it may relate to disturbance in cortical neuronal development (Inder et al. 1999a). In autopsy studies of infants with white matter lesions, alterations in the morphology and organization of neurons in the cortex overlying neonatal PVL have been demonstrated (Marin-Padilla 1997). In preterm infants, later cognitive outcome has been related to reduction of cortical gray matter volumes (Isaacs et al. 2001, Peterson et al. 2003).

2.3.3 Periventricular-intraventricular hemorrhage and infarction Periventricular-intraventricular hemorrhage (PV-IVH) is graded according to the extent of bleeding into the adjacent ventricles and to the surrounding structures detected by brain ultrasound (US) (Papile et al. 1978). According to Papile et al., grade I hemorrhage refers to subependymal hemorrhage, grade II hemorrhage refers to IVH without ventricular dilatation, grade III hemorrhage refers to IVH with ventricular dilatation, and grade IV hemorrhage refers to IVH with parenchymal hemorrhage (Papile et al. 1978). According to Volpe, grade I hemorrhage refers to germinal matrix hemorrhage with no or minimal IVH. Grade II and grade III hemorrhages refer to IVH involving 10% to 50% and over 50% of the ventricular area on a parasagittal US view, respectively (Volpe 1995). Periventricular hemorrhagic infarction (PHI) represents an area of hemorrhagic necrosis that is usually large and asymmetric and lies within periventricular white matter dorsal and lateral to the external angle of the lateral ventricle (Guzzetta et al. 1986, Gould et al. 1987). Intraparenchymal echodensity (IPE) observed in brain ultrasound (US) reflects PHI (Guzzetta et al. 1986). Figure 1 presents the anatomic location of PHI. The clinical factors associated with PV-IVH include low gestational age at birth (Heuchan et al. 2002, Heep et al. 2003, Vergani et al. 2004), respiratory distress syndrome (RDS) (Vergani et al. 2000), hypercarbia (Wallin et al. 1990), intravascular volume expansion (Salafia et al. 1995), pneumothorax (Hill et al. 1982), hypotension (Bada et al. 1990), vaginal delivery (Leviton et al. 1991), coagulation disorders (Whitelaw 2001), and intrauterine inflammation (Salafia et al. 1995, Vergani et al. 2000).

2.3.3.1 Incidence The incidence of PV-IVH ranges between 15% and 30% among infants born preterm (Salafia et al. 1995, Verma et al. 1997, Vergani et al. 2000, Lemons et al. 2001, Heuchan et al. 2002). A declining rate of the overall incidence of PV-IVH has been reported (Heuchan et al. 2002). Both PV-IVH as a whole (Lemons et al. 2001) and the most severe forms of PV-IVH (Salafia et al. 1995, Lemons et al. 2001) occur most frequently in the

30 group of preterm infants with the lowest birth weight (Lemons et al. 2001) and gestational age (Salafia et al. 1995). Approximately 15% of all infants with IVH exhibit PHI (Volpe 1995). The incidence of PHI has been reported to be 13% in infants with birth weight of 501 to 750 g, 6% in infants with birth weight of 751 to 1000 g, 3% in infants with birth weight of 1001 to 1250 g, and 1% in infants with birth weight of 1251 to 1500 g. Altogether, the incidence of PHI in infants with birth weight of 501 to 1500 g has been reported to be 5% (Lemons et al. 2001).

2.3.3.2 Pathogenesis The pathogenesis of PV-IVH implies structural immaturity of blood vessels in the germinal matrix (GM) (Kuban & Gilles 1985, Nakamura et al. 1991, Anstrom et al. 2004) and deficient autoregulation of CBF (Lou et al. 1979, Pryds et al. 1989). Changes in CBF were measured in 57 preterm infants during 48 hours after birth. Severe intracranial hemorrhage was related to pressure-passive cerebral blood circulation, suggesting that either cerebral hypoperfusion secondary to a decrease in systemic blood pressure or systemic hypertension and an increase in intravascular pressure leads to a rupture of GM (Pryds et al. 1989). PHI is suggested to be closely linked to PV-IVH. Both GM and periventricular white matter are border zone regions. This increases the risk for PV-IVH and PHI as a phenomenon secondary to reperfusion injury (Perlman 1998, Shalak & Perlman 2002). Elevations in intracerebral venous pressure may also be considered a mechanism leading to PHI (Perlman 1998). Obstructed venous drainage in the periventricular area as a result of PV-IVH has been shown to associate with PHI (Gould et al. 1987, Taylor 1995).

2.4 Placental insufficiency Villous vascularization during the first and second trimesters leads to the formation of immature villous trees with a continuously growing network of fetal capillaries. In normal pregnancies during the third trimester, the fetoplacental blood supply constitutes terminal villi with a capillary network responsible for nutrient and gas exchange between the mother and the fetus (Benirschke & Kaufmann 2000). Placental insufficiency relates to, for instance, maternal hypertension, preeclampsia (Ferrazzi et al. 1999), and maternal diabetes (Huisman 2001). Uteroplacental vascular lesions have also been reported in normotensive pregnancies (Ferrazzi et al. 1999) and in pregnancies complicated by maternal autoimmune disease (Labarrere et al. 1986). In pregnancies complicated by impaired uteroplacental circulation, the predominant feature on the maternal side is an abnormal conversion of branches of the uterine arteries (spiral arteries) into uteroplacental arteries due to poor extravillous trophoblastic invasion (Pijnenborg et al. 1991). On the fetal side, the intervillous space is wide, and terminal villi are poorly vascularized with slender, unbranched capillary loops. Cytotrophoblast cells in villi are reduced. This leads to a decrease in gas exchange and increases the risk of fetal hypoxia and acidosis (Macara et al. 1996). Villous infarcts are generally due to

31 disturbances of intervillous circulation (Benirschke & Kaufmann 2000). Endothelial cell activation with up-regulation of various biochemical substances has been recently related to placental insufficiency in preeclampsia (Dietl 2000).

2.4.1 Placental hemodynamics Placental circulatory insufficiency is associated with an increase in placental vascular resistance and downstream vascular impedance (Thompson & Trudinger 1990). This leads to a decrease in umbilical artery (UA) end-diastolic blood flow and secondary fetal hypoxia (Wladimiroff et al. 1986). In Doppler ultrasonography, this is manifested as an increase in systolic to diastolic velocity ratio (S/D ratio) (Trudinger et al. 1985) and expressed as an increase in pulsatility index (PI) in UA or descending aorta (DAo) velocity waveform (Wladimiroff et al. 1986). Loss of small muscular arteries in stem villi and poorly capillarized terminal villi correlate with an increase in placental resistance (Giles et al. 1985, Thompson & Trudinger 1990, Todros et al. 1999). In the blood flow profile, reduction in diastolic blood flow velocity in UA precedes the total absence of end-diastolic blood flow (Bekedam et al. 1990) and reversed end-diastolic blood flow (Brodszki et al. 2002). In animal experiments, placental insufficiency has been related to white matter damage. Fetal sheep predisposed to placental isufficiency and hypoxia demonstrated gliosis in cerebral cortex and reduced myelination in subcortical white matter (Mallard et al. 1998, Duncan et al. 2000). In a prospective study of preterm infants, Doppler measurements of UA blood flow were performed in 54 cases before delivery. Nine cases had absent or reversed end-diastolic blood flow in UA. Four of the 8 surviving infants had CP at 2 years of corrected age compared to none of the 38 surviving infants with detectable end-diastolic blood flow in UA (Spinillo et al. 1997). Recently, premature infants with IUGR born from pregnancies complicated by placental insufficiency showed significant reduction of brain tissue and cerebral cortical gray matter volumes during the first 2 weeks after birth and at term, when measured with a volumetric three-dimensional MRI technique, compared to control preterm infants with birth weight appropriate for gestational age and normal blood flow in UA. Furthermore, the cerebral cortical gray matter volume at term correlated directly with the attention-interaction capacity measured at term (Tolsa et al. 2004). Placental insufficiency relates to premature birth and fetal growth failure in utero (Trudinger et al. 1985, Wladimiroff et al. 1987, Trudinger et al. 1991). IUGR complicates 5% to 10% of all pregnancies (Galan et al. 2002). The risk of CP has been associated with a prenatal diagnosis of IUGR, OR 6.6, 95%CI 1.8-25.2 (Gray et al. 2001). IUGR is diagnosed when the estimated fetal weight falls below the 10th percentile for gestational age. This can vary, depending on the population studied (Goldenberg et al. 1989).

32

2.4.2 Fetal hemodynamics During the second half of normal pregnancy, increase in the diastolic blood flow velocity relative to the peak systolic blood flow velocity in the uteroplacental artery has been reported (Trudinger et al. 1985). Normally, PI in UA decreases in the course of the third trimester (Wladimiroff et al. 1986). In pregnancies complicated by IUGR, reduced PI in fetal cerebral arteries has been demonstrated (Wladimiroff et al. 1986, Wladimiroff et al. 1987) with a concomitant rise in PI in UA and in DAo. This suggests a reduction of vascular resistance in the fetal cerebral circulation, i.e. a ´brain-sparing effect´ (Wladimiroff et al. 1986, Wladimiroff et al. 1987), and a decrease in blood flow to the major parts of the subdiaphragmatic circulation (Block et al. 1990) mediated by alphaadrenergic innervation (Reuss et al. 1982). During the second half of uncomplicated pregnancies, there is a higher cardiac output from the right than the left fetal ventricle (al Ghazali et al. 1989, Räsänen et al. 1996). Combined cardiac output (CCO) and placental blood flow increase with advancing gestation (Sutton et al. 1991). In fetuses with IUGR, redistribution of cardiac output has been shown to favor the left ventricle (al Ghazali et al. 1989, Mäkikallio et al. 2002). Evidence to support the relationship between redistribution of fetal blood flow and fetal hypoxemia has been presented (Hecher et al. 1995a). The inferior vena cava (IVC), hepatic veins (HV), and ductus venosus (DV) play a major role in venous return flow to the fetal heart (Hecher et al. 1995b). Well-oxygenated blood from the placental circulation flows through DV and is preferentially directed toward the foramen ovale (FO) and the left atrium (Edelstone & Rudolph 1979). A gestational age dependent increase in DV flow has been demonstrated (Van Splunder et al. 1996). In pregnancies complicated by placental insufficiency, DV PI has been shown to increase (Hecher et al. 2001). This indicates a decrease in forward diastolic blood flow velocity in DV during atrial contraction (Hecher et al. 1995b). The rise in the right ventricular afterload coincides with the increase in systemic venous pressure (Mäkikallio et al. 2002). A temporal sequence of abnormal Doppler changes in pregnancies complicated by placental insufficiency has been detected. An increase in PI in UA and DAo and a decrease in PI in the middle cerebral artery (MCA) precede changes in the blood flow profiles of venous circulation and decreases in fetal cardiac outputs (Hecher et al. 2001, Ferrazzi et al. 2002). Hypoxemia and acidemia correlate with changes in both arterial and venous blood flow profiles in the fetus (Ferrazzi et al. 1995, Hecher et al. 1995a). In pregnancies complicated by placental insufficiency, abnormal velocity waveforms in UA (Trudinger et al. 1991, Kurkinen-Räty et al. 1997, Vossbeck et al. 2001, Kutschera et al. 2002) and in DAo (Ley et al. 1996a, Ley et al. 1996b), aortic isthmus (Fouron et al. 2001), cerebral arteries (Scherjon et al. 2000, Kutschera et al. 2002), and umbilical vein (Hecher et al. 1995b, Baschat et al. 2000, Hecher et al. 2001, Ferrazzi et al. 2002) have been associated with poor short-term or long-term outcomes in infants. Abnormal fetal venous velocity waveform has a strong correlation with fetal death and associates with perinatal mortality and morbidity (Baschat et al. 2000).

33

2.5 Cytokines Cytokines are small soluble polypeptides that are released in response to an activating stimulus by a variety of cells, e.g. monocyte-macrophages, T lymphocytes (Janeway et al. 2001), microglia, astrocytes (Lee et al. 1993a), neurons (Jonakait 1997), endothelial cells (Rollins 1997), and decidual cells (Romero et al. 1989a, Dudley et al. 1992). Cytokines induce responses through signal transduction by binding to specific cell-surface receptors. They act via three different mechanisms: 1) in an autocrine manner, by affecting the behavior of the cell that releases the cytokine, 2) in a paracrine manner, by affecting the behavior of adjacent cells, and 3) in an endocrine manner, by affecting the behavior of distant cells (Janeway et al. 2001). Cytokines participate in a wide variety of biologic responses, including the growth and development and modulation of inflammation and the immune response (Nesin & Cunningham-Rundles 2000).

2.5.1 Classification Cytokines and their receptors are grouped according to their structure. There are three major cytokine families: the hematopoietin family, which includes growth factors, e.g. ciliary neurotrophic factor (CNTF) and granulocyte-macrophage colony-stimulating factor (GM-CSF), and many interleukins; the TNF family; and the chemokines. Chemokines are divided into four groups, depending on the position of the first two cysteine residues, which can be either adjacent (CC) or separated by one amino acid (CXC) or three amino acids (CX3C). Chemokine C contains only one cysteine residue (Janeway et al. 2001).

2.5.2 Function Cytokines can be produced by a human fetus at the end of the first trimester of gestation (Adinolfi 1993). They have an important role in innate immunity, as they induce an inflammatory response against foreign pathogens and connect innate immunity with adaptive immunity (Janeway et al. 2001). The fetus depends mostly on innate immunity. Development of the components of adaptive immunity starts with the appearance of the first lymph nodes and lymphocytes in the second trimester (Nesin & CunninghamRundles 2000). Chemokines, e.g. IL-8, and cell adhesion molecules participate in the inflammatory response by controling leukocyte migration. This leads to recruitment of neutrophils from the circulation, adherence to the vascular endothelium, and migration through the vessel wall to the site of microbial invasion. The effects of chemokines on leukocytes are mediated by their receptors. CXC chemokines promote the migration of neutrophils and lymphocytes. CC chemokines promote the migration of monocytes (Janeway et al. 2001). CXC and CC chemokine receptors recognize chemokines only of the corresponding subfamily (Baggiolini 1998). Chemokines have been detected in cord blood in both term and preterm infants without infection (Sullivan et al. 2002).

34 As a response to a foreign antigen, the proinflammatory cytokines TNF-α, IL-1, and IL-6 initiate an acute phase response, which helps to coordinate the body`s response to infection while the adaptive immune response is being developed. This process involves the production of acute-phase proteins, e.g. C-reactive protein, which binds to the pathogen, activation of bone marrow to release neutrophils into circulation, and induction of fever (Janeway et al. 2001). Monocytes/macrophages produce cytokines in response to antigen stimulation. Toll-like transmembrane receptors (TLR) include at least 10 signaling receptors. They bind components from microbes and initiate specific intracellular signaling pathways resulting in the induction of proinflammatory cytokines. This process has been best described between lipopolysaccharide (LPS) from gramnegative bacteria and TLR4. LPS-induced activation of TLR4 is mediated via the signal transduction pathway through activation of the transcription factor, nuclear factor (NF)κB (Hallman et al. 2001). Anti-inflammatory cytokines, e.g. IL-10 and IL-4, produced by T helper lymphocytes (Th2 cells), are needed to downregulate the inflammatory process (Ng et al. 2003). Immaturity of the immune system, including the reduced capacity of cytokine production, has been proposed to be responsible for increased susceptibility to bacterial infections during the neonatal period. However, LPS-stimulated cord blood monocytes have been shown to express a comparable level of IL-1 (Weatherstone & Rich 1989) and higher levels of IL-6 and IL-8 (Schultz et al. 2002) both in term and preterm infants compared to adult monocytes. On the other hand, lower levels of TNF-α in term (Pillay et al. 1994) and preterm (Weatherstone & Rich 1989) infants and lower levels of IL-6 in term infants (Pillay et al. 1994) have been detected compared to adult monocytes. The production of some cytokines seems to be developmentally regulated. Maximum levels of TNF-α, IL-1β, IL-6, and granylocyte colony-stimulating factor (G-CSF) in LPSstimulated cord blood have been shown to be lower in very preterm infants (≤ 32 weeks) compared to infants with more advanced gestational age (Dembinski et al. 2003). Also, GM-CSF measured from amniotic fluid (AF) has been demonstrated to increase as a function of gestational age in uncomplicated pregnancies (Bry et al. 1997). In physiological conditions, cytokines interact by signal transmission from one cell to another. This complex system may be disrupted if some cytokines are produced in excess or at an inappropriate site (Adinolfi 1993). Cytokines and chemokines have been related to neonatal mortality and morbidity. A genetic polymorphism in TNF-α has been found to associate with increased mortality from sepsis in preterm infants (Hedberg et al. 2004). On the other hand, it has been suggested that preterm infants with compromised ability to produce IL-6 are susceptible to sepsis (Harding et al. 2003). High levels of IL-1ß have been associated with sepsis in preterm infants (Berner et al. 1998). High levels of G-CSF, IL-6, IL-8, and TNF-α have been associated with sepsis in term and preterm infants (Berner et al. 1998, Kashlan et al. 2000, Nupponen et al. 2001, Silveira & Procianoy 2003). High levels of IL-6 have also been associated with RDS (Gomez et al. 1998), bronchopulmonary dysplasia (BPD) (Yoon et al. 1999), and necrotizing enterocolitis (NEC) (Goepfert et al. 2004) in preterm infants and with asphyxia in term infants (Shalak et al. 2002, Silveira & Procianoy 2003). High levels of IL-1β in cord blood have been detected in infants with severe perinatal complications, including marked fetal distress during labor, and severe maternal preeclampsia (Miller et al. 1990).

35

2.5.3 Cytokines and brain Cytokines have several neurotrophic activities in developing brain. They participate in cell proliferation and differentiation and interact with each other (Jonakait 1997). Neurotrophins also participate in neuronal migration. Intraventricular injection of neurotrophin-4 (NT-4) (Brunstrom et al. 1997) or overexpression of brain-derived neurotrophic factor (BDNF) (Ringstedt et al. 1998) have been shown to produce heterotopic accumulation of neurons in the marginal zone in developing brain. The IL-6 family also affects the developing nervous system. IL-6, ciliary neurotrophic factor (CNTF), and leukemia inhibitory factor (LIF) have been shown to regulate glial cell differentiation by causing a transient increase in glial fibrillary acid protein (GFAP) in OL-2 astrocyte (O-2A) progenitor cells in a rat model. (Kahn & De Vellis 1994). CNTF promotes survival and maturation of cultured OL and protects OL from death induced by TNF (Louis et al. 1993). CNTF also enhances regeneration and neuroprotection by activating glial cells and by upregulating another neurotrophic factor, i.e. fibroblast growth factor-2 (FGF-2), in cultured spinal cord astrocytes and neurons in a rat model (Albrecht et al. 2002). Epidermal growth factor (EGF) and transforming growth factor-α (TGF-α) stimulate astrocyte differentiation and establish protection against glutamate by inducing expression of glial glutamate transporter-1 (GLT-1) in cultured astrocytes (Zelenaia et al. 2000). A broad variety of growth factors influence the maturation of cerebrovascular endothelial cells and angiogenesis in developing brain (Pearce 2002). Composite action of different cytokines and different cell types is needed to produce neuronal survival and recovery in developing brain (Jonakait 1997). Cytokines are expressed constitutively or upon activation in human brain. TNF-α messenger ribonucleic acid (mRNA) and TNF-α have been detected in normally developing brain in two autopsy series of human neonates (Lee et al. 1993a, Deguchi et al. 1996), and TNF-α mRNA has also been found in cultured nonstimulated human fetal astrocytes (Lee et al. 1993a). In purified human fetal cultures of microglia and astrocytes, microglia was shown to produce TNF-α, IL-1β, and IL-6 in response to LPS. Furthermore, in stimulated microglia, IL-1β was found to be produced by IL-1β and TNF-α. IL-1β was also found to induce a strong stimulus to astrocytes to produce TNF-α and IL-6. (Lee et al. 1993a). TNF-α and IL-6 have been shown to induce proliferation of astrocytes in cultured mature bovine brain (Selmaj et al. 1990). Apparently, cytokines may act differently in different species. LPS has been shown to induce TNF-α synthesis in neonatal rat astrocytes (Chung & Benveniste 1990) but not in human astrocytes (Lee et al. 1993a). Recombinant IL-2 infusion has been shown to induce systemic production of TNF-α in rats (Ellison & Merchant 1991). Cytokines can also initiate dual effects resulting in either cell growth and survival or cell death (Conroy et al. 2004, Buntinx et al. 2004). It has been suggested that the effect induced by cytokines depends on the duration and magnitude of the cytokine challenge (Louis et al. 1993, Qiu et al. 1998, Cammer & Zhang 1999, Conroy et al. 2004, Buntinx et al. 2004), genotype variation (Harding et al. 2004), and immaturity of the target cell (Baerwald & Popko 1998, Conroy et al. 2004). Cytokines have a toxic effect on immature OLs. TNF-α (Louis et al. 1993, Buntinx et al. 2004) and IFN-γ (Baerwald & Popko 1998, Buntinx et al. 2004) have been shown to induce OL apoptosis in vitro.

36 TNF-α has been shown to inhibit the expression of myelin basic protein and also to influence the maturation of cultured OLs in rat pups (Cammer & Zhang 1999). On the other hand, TNF-α also induces genes that regulate OL cell survival (Buntinx et al. 2004), and IL-6 has growth-promoting effects on CNS cells (Conroy et al. 2004). In vitro cytokines have also been demonstrated to participate in reactions involved in ischemia/reperfusion injury. IL-6 has enhanced the neurotoxic activation of NMDA glutamate receptor in cerebellar granulae neuron cultures (Qiu et al. 1998, Conroy et al. 2004). IL-1β-stimulated human fetal astrocytes have been shown to produce NO (Lee et al. 1993b), which has been found to be toxic to developing OLs (Baud et al. 2004). It has been proposed that, during gestation, cytokines gain access into fetal brain through an impaired blood-brain barrier (BBB) (Dammann & Leviton 1997). BBB is formed by endothelial cells lining the cerebral vasculature interconnected with tight junctions that develop early in fetal life and continue to differentiate throughout fetal and early postnatal life (Pearce 2002). BBB forms an important mechanism that protects the brain from fluctuations in plasma composition (Abbott 2002). The present findings suggest that fetal BBB is functionally specialized, utilizes mechanisms not found in adult cerebral endothelium, and is well adapted to fetal life. Permeability to small lipidinsoluble molecules is greater in developing brain (Saunders et al. 1999). Numerous cytokines, e.g. IL-1α, IL-1β, IL-2, IL-6, and TNF-α (Abbott 2002), but also chemokines (Anthony et al. 1998) and other humoral agents, e.g. glutamate, NO, and free radicals (Abbott 2002), have been shown to increase the permeability of BBB. In a rat model, 1 hour of hypoxia (6% O2) followed by a 10-minute period of reoxygenation produced an increase in BBB permeability not attributable to an increase in CBF or alterations in the expression of tight-junctional proteins (Witt et al. 2003). Moreover, chemokines may potentiate barrier breakdown. Juvenile rats treated with IL-1β and anti-cytokine-induced neutrophil chemoattractant-1 (CINC-1), a chemokine-neutralizing antibody, showed a significant reduction in neutrophil recruitment in brain and almost complete absence of BBB breakdown. Importantly, neutrophil depletion with anti-neutrophil antiserum eliminated the cytokine breakdown effect of BBB in both adult and juvenile rats (Anthony et al. 1998). It has been proposed that, once activated inflammatory cells, i.e. macrophage and Tcells, have entered the brain, they activate parenchymal astrocytes and microglia, thereby recruiting them into the inflammatory activity to up-regulate cytokine production (Merrill & Benveniste 1996, Kadhim & Sebire 2002, Duncan et al. 2002). High concentrations of proinflammatory cytokines (TNF-α, IL-1β, IL-6) and IL-10 were recently detected together with upregulation of activated T-cells in umbilical cord blood in preterm infants with various cerebral lesions demonstrated on conventional MRI soon after birth (Duggan et al. 2001). On the other hand, no expression of adhesion molecules, which are usually associated with transendothelial migration of white blood cells, were observed in PVL lesions in an autopsy series of neonatal brains (Kadhim et al. 2001). Various neuropathological studies using animal models have pointed out the association between white matter damage, induced either by endotoxin or by a hypoxicischemic episode, and cytokine expression. Yet, it is not clear whether neuronal expression of cytokines in PVL is a cause or a consequence of the injury (Kadhim et al. 2003).

37 In the classical study of Gilles et al., newborn kittens were administered daily LPS intraperitoneally. Three morphologically distinct abnormalities were detected in neonatal brains from endotoxin-exposed animals: astrogliosis in telencephalic white matter, focal areas of cystic necrosis with macrophage infiltration, and deposits of eosinophilic or mineralized debris adjacent to necrotic lesions. The appearance of astrogliosis and cystic necrosis appeared to be dose-dependent. None of the control animals demonstrated white matter lesions. There was no evidence of vascular occlusions in any case. (Gilles et al. 1976). In another study of newborn rabbits delivered from pregnancies complicated by intrauterine inoculation of Escherichia coli, 12 rabbits out of 190 were diagnosed to have white matter lesions compared to none in the control group. The histopathologic changes in brain included nuclear fragmentation and apoptosis in glial cells and focal rarefaction and disorganization of white matter resembling changes observed in human neonatal brains with coagulation necrosis in PVL (Yoon et al. 1997b). In a recently published study using six preterm ovine fetuses, corresponding to gestational age of 24 to 25 weeks in a human fetus, series of intravenous injections of LPS were given in utero. Neural injury was observed in all LPS-exposed fetuses, most prominently in cerebral white matter. Injuries ranged from diffuse subcortical damage with gliosis and axonal damage to focal patches of necrosis, gliosis, and axonal damage in white matter adjacent to the lateral ventricles. Resident microglia or infiltrating macrophages were present. In the brainstem, the cross-sectional area of the corticospinal tract was reduced by 30%. (Duncan et al. 2002). The effect of hypoxia-ischemia (HI) was studied in 7-day-old rat pups. After unilateral carotid artery ligation and hypoxia for 70 to 100 minutes, transient increases of cerebral IL-1 and IL-6 bioactivity and IL-1β and IL-6 mRNA were seen. Pretreatment with IL-1 receptor antagonist reduced HI brain damage and posttreatment increased the portion of animals devoid of brain injury (40%) compared with vehicle-treated controls (13%) (Hagberg et al. 1996). A few studies of autopsy series in human neonates have also been reported. Cytokine expression in brain was studied in 17 term and preterm infants, most of whom died during the neonatal period. Immunohistochemical staining demonstrated expression of TNF-α, IL-1β, or IL-6 in 88% of the cases with PVL compared to 24% of the cases without PVL. Cytokines were mainly expressed in hypertrophic astrocytes and microglial cells (Yoon et al. 1997c). In another autopsy series of 19 neonatal brains with findings of coagulative necrosis and cystic lesions and altogether four non-PVL brains with anoxia, the highest expressions of TNF-α and, to a lesser extent, IL-1β in microglia, foamy macrophages, and reactive astrocytes were detected in brains with coagulative necrosis and defined perinatal infection. Some immunoreactivity for TNF-α and IL-1β was also detected in brains with anoxia but to a lesser extent compared to brains with PVL (Kadhim et al. 2001). Recently, TNF-α immunoreactivity has been demonstrated in neurons in both cortical gray matter and deep gray matter structures of human infants with PVL (Kadhim et al. 2003). Moreover, IFN-γ-immunopositive macrophages and reactive astrocytes have been demonstrated in PVL lesions in human brain (Folkerth et al. 2004a).

38

2.6 Chorioamnionitis Histological chorioamnionitis (HCA) refers to intrauterine infection that affects tissues of either feto-maternal origin (choriodecidual space and placenta) or fetal origin (chorioamniotic membrane, amniotic fluid, and umbilical cord) (Goldenberg et al. 2000, Hagberg et al. 2002). Infection of the fetal membranes is called chorioamnionitis, infection of the umbilical cord is called funisitis, and infection of AF is called amnionitis (Goldenberg et al. 2000). HCA is diagnosed and graded according to the appearance and density of polymorphonuclear (PMN) leukocytes in the fetal membranes, umbilical cord vessels, and chorionic plate (Salafia et al. 1989). HCA associates with spontaneous preterm labor (Hillier et al. 1988, Mueller-Heubach et al. 1990, Verma et al. 1997). PPROM is present in 30% to 40% of spontaneous preterm deliveries (Parry & Strauss, III 1998). PPROM is also associated with an increased incidence of HCA (Verma et al. 1997, Gomez et al. 1998), which, on the other hand, is the more common, the longer is the duration between membrane rupture and delivery (Leviton et al. 1999). HCA is rare in late preterm deliveries (at 34 to 36 weeks), but it is present in most cases where birth occurs at less than 30 weeks of gestation (Goldenberg et al. 2000). The frequency of HCA is inversely related to gestational age at birth (Hillier et al. 1988, Mueller-Heubach et al. 1990, Tauscher et al. 2003, Lahra & Jeffery 2004). In a cohort of preterm infants, the incidence of HCA ranged from 66% at 20 to 24 weeks of gestation to 16% at 34 weeks of gestation (Lahra & Jeffery 2004), and it has been reported to be 18% in pregnancies at term gestation (Mueller-Heubach et al. 1990). Signs of bacterial etiology in AF have been found in 23% to 58% of cases with PPROM (Gomez et al. 1998, Shim et al. 2004) and in 10% of cases with preterm labor with intact membranes (Gomez et al. 1998, Yoon et al. 2001). Preterm labor and HCA have recently also been associated with a genetic polymorphism of TNF-α in women with preterm labor before 34 weeks of gestation (Amory et al. 2004). Preterm labor with HCA with or without PPROM has also been associated with increased expression of TLR2 and TLR4 in the chorioamniotic membranes (Kim et al. 2004). No uniform criteria for the diagnosis of clinical chorioamnionits (CCA) have been established, and the criteria differ widely between the published studies (Wu 2002). CCA is usually defined as the presence of maternal fever (≥ 37.8 C) during labor together with two or more of the following symptoms or signs: maternal or fetal tachycardia, uterine tenderness, malodorous AF, and maternal leukocytosis (Gibbs et al. 1982), the additional criteria including duration of ruptured membranes > 24h and an elevated level of maternal C-reactive protein (Baud et al. 1999a). In preterm pregnancies, the incidence of CCA ranges between 4% to 24%, depending on the study population and the CCA criteria used (Yoon et al. 1995, Perlman et al. 1996, Alexander et al. 1998, Dexter et al. 1999), and CCA correlates with the severity of HCA (Redline et al. 2000). In term infants, a CCA incidence of 7% has been reported (Shalak et al. 2002). Microorganisms may gain access to the choriodecidual layer and the amniotic cavity via a transcervical route, hematogenously through the placenta, iatrogenically at the time of amniocentesis, or by seeding from the peritoneal cavity through the fallopian tubes. The most common pathway of intrauterine infection is the ascending route (Asrat 2001, Goldenberg & Culhane 2003). Upon adherence of the membranes to the decidua at about

39 20 weeks of gestation, the inflammatory process accelerates (Goldenberg & Culhane 2003), as bacteria can penetrate the intact fetal membranes and contaminate the amniotic fluid (Bobitt & Ledger 1977, Gravett et al. 1986). Bacterial vaginosis (BV) has been shown to increase the risk of preterm labor (Gravett et al. 1986, Hillier et al. 1995, Leitich et al. 2003), and BV-associated organisms have been observed in the uterus in otherwise healthy nonpregnant women (Korn et al. 1995). Of non-genital tract infections, severe or generalized periodontal disease has recently been found to associate with preterm delivery before 35 weeks of pregnancy (OR 5.28, CI95% 2.05–13.60) and before 32 weeks of pregnancy (OR 7.07, CI95% 1.70–27.4) (Jeffcoat et al. 2001). Trophoblast cells are able to synthesize IL-6 (Kameda et al. 1990). Decidual cells are able to synthesize IL-1 (Romero et al. 1989a), IL-6 (Dudley et al. 1992), and TNF-α (Casey et al. 1989) in response to either cytokine (Dudley et al. 1992) or bacterial endotoxin (Casey et al. 1989, Romero et al. 1989a) stimulation. Bacterial endotoxin or TNF-α has been shown to cause an increase in the production of prostaglandin (PG) F2α in decidua (Casey et al. 1989, Romero et al. 1989b) and an increase in PGE2 in amnion cells (Casey et al. 1989) and in decidua (Romero et al. 1989b). In an experimental model of intraamniotic infection and preterm labor in rhesus monkeys, an increase in AF cytokines TNF-α, IL-1β, and IL-6 and prostaglandins PGE2 and PGF2α preceded the increase of uterine contractility after group B streptococci (GBS) inoculation (Gravett et al. 1994). Cytokines in AF, cervical secretions, umbilical cord blood, or tracheal aspirates have been shown to associate with spontaneous preterm labor (Silver et al. 1993, Rizzo et al. 1997, Goepfert et al. 2004) and with HCA in preterm (Yoon et al. 1995, Salafia et al. 1997, Kashlan et al. 2000, Naccasha et al. 2001, Dollner et al. 2002, Rogers et al. 2002, Yanowitz et al. 2002, De Dooy et al. 2003) and term pregnancies (Rogers et al. 2002). Elevated levels of IL-6 in umbilical cord blood (Kashlan et al. 2000, Rogers et al. 2002) or in AF (Yoon et al. 1995) have been found to correlate with the severity of HCA. Funisitis represents the most severe form of HCA (Dollner et al. 2002). Elevated levels of umbilical cord IL-6 associate with funisitis (Kashlan et al. 2000, Naccasha et al. 2001). Fetal vasculitis defined as PMN leukocyte infiltration in the vessel walls in the chorionic plate or in the umbilical cord (Leviton et al. 1999) or elevated levels of IL-6 in umbilical cord blood have been defined as systemic responses to inflammation in fetus (Gomez et al. 1998).

2.6.1 Brain damage Some 10 years ago, it was suggested that cytokines, particularly TNF-α, may be the connecting factor between preterm birth and perinatal brain damage (Leviton 1993). Both HCA and CCA have been related to cPVL and PV-IVH in preterm infants with spontaneous preterm labor with or without PPROM. Increasing severity of PV-IVH correlates with increasing severity of HCA (Salafia et al. 1995, Vergani et al. 2000). Table 2 presents the studies addressing the association between brain injury and HCA or CCA. Two meta-analyses have recently been published underlining chorioamnionitis as a

40 risk factor for cPVL in preterm infants. The relative risk (RR) for cPVL in CCA was 3.0, 95%CI 2.2–4.0 (Wu & Colford, Jr. 2000) and 2.6, 95%CI 1.7–3.9 (Wu 2002). RR for cPVL in HCA was 2.1 95%CI 1.5–2.9 (Wu & Colford, Jr. 2000) and 1.6, 95%CI 1.0–2.5 (Wu 2002). Table 2. Summary of publications investigating the association between chorioamnionitis and brain lesion in preterm infants. Reference

Type of infection

Association between brain pathology and intrauterine inflammation

Bejar et al. 1988

Purulent AF, funisitis Echolucency by day 3, significant association in logistic regression analysis (numbers not shown in the article)

Salafia et al. 1995

HCA

Early PV-IVH, OR 1.407, 95%CI 1.129 to 1.753

Perlman et al. 1996

CCA

cPVL, OR 5.14, 95%CI 1.27 to 20.79

Verma et al. 1997

CCA

IVH grade 3 to 4, PV-IVH with PVL, echolucent PVL, OR

Yoon et al. 1997a

HCA

IVH grade ≥ 2; OR 5.3, 95%CI 1.2 to 23.0

Disalvo et al. 1998

HCA

PV-IVH, OR 2.4, 95%CI 1.0 to 6.0

Alexander et al. 1998

CCA

IVH grade 3 to 4, OR 2.8, 95%CI 1.6 to 4.8; cPVL, OR 3.4,

Leviton et al. 1999

HCA

Late echolucency, OR 10.8, 95%CI 1.03 to 114

Vergani et al. 2000

HCA

PV-IVH, p = 0.02

2.23, 95%CI 1.23 to 3.94

95%CI 1.6 to 7.3

Dexter et al. 2000

HCA

PV-IVH, RR 1.6, 95%CI 1.1 to 2.4

Tauscher et al. 2003

HCA

PV-IVH, p = 0.006

Yoon et al. 1995

HCA

No association

Wiswell et al. 1996

CCA

No association

Dexter et al. 1999

CCA

No association

Baud et al. 1999a

HCA

No association

Kumazaki et al. 2002

HCA

No association

AF, amniotic fluid; CCA, clinical chorioamnionitis; CI, confidence interval; cPVL, cystic periventricular leukomalacia; HCA, histologic chorioamnionitis; OR, odds ratio; PV-IVH, periventricular-intraventricular hemorrhage; RR, relative risk.

The systemic fetal response to infection/inflammation during pregnancy, defined as elevated levels of cord blood cytokines, has been reported to associate with cPVL and PV-IVH in preterm infants, indicating that the injury responsible for the disease may begin before birth. In a prospective study of 172 preterm infants with gestational age < 36 weeks at birth, IL-6 levels in umbilical cord blood were significantly higher in preterm infants with PVL compared to infants without PVL. An IL-6 value ≥ 400 pg/ml had a sensitivity of 72% and a specificity of 74% to identify PVL lesions. PVL was defined at autopsy or by ultrasonography as persistent echogenic or cystic lesions in periventricular white matter. Multiple logistic regression analysis revealed IL-6 as an independent risk factor for PVL after adjustment for gestational age at birth. However, the levels of TNFα, IL-1β, and IL-1receptor antagonist in umbilical cord blood did not differ between newborns with or without PVL, nor were there significant associations between PVL and HCA or CCA (Yoon et al. 1996). In a series of 50 infants, born between 23 to 29 weeks

41 of gestation, conventional MRI was performed at a median age of 2 days after birth. The mean values of cord blood cytokines TNF-α, IL-1β, IL-6, and IL-10 were higher in infants with a detected brain lesion, i.e. PV-IVH, discrete periventricular lesion, and/or cystic lesion in caudate nucleus, compared to infants without a brain lesion. The investigators also found an increase of cell surface antigen CD45RO+, indicating a fetal T lymphocyte response to an antigen stimulus in situ (Duggan et al. 2001). More recently, umbilical cord plasma levels of IL-6 and neonatal outcomes were assessed among 309 infants born between 24 to 31 weeks of pregnancy. In a multivariate analysis, an IL-6 value ≥ 107.7 pg/ml remained an independent risk factor for PVL (Goepfert et al. 2004). In a retrospective study of 88 infants with gestational age < 28 weeks at birth, PV-IVH was found more often in the infants with serum IL-6 levels > 100 pg/ml recorded within 12 hours after birth compared to those with IL-6 levels ≤ 100 pg/ml. No association was found between PVL and IL-6. The multiple logistic regression model confirmed the independence of high IL-6 as a risk factor for IVH grade 3 to 4 after adjustment for confounding factors (Heep et al. 2003). In a prospective study of 106 preterm infants with gestational age < 32 weeks, infants with PV-IVH had higher median levels of cord blood IL-1β, IL-6, and IL-8 than infants without PV-IVH, P < 0.001 (Tauscher et al. 2003). Not all investigators, however, have reached a consistent conclusion about intrauterine inflammation and its effects on brain injury in preterm infants (Yoon et al. 1995, Wiswell et al. 1996, Gomez et al. 1998, Dexter et al. 1999). In some cases the significance has disappeared after adjustment for covariates, e.g. gestational age and birth weight (Yoon et al. 1997a, Baud et al. 1999b). In a study of 31 newborn infants with gestational age < 32 weeks at birth, TNF-α, IL-1β, and IL-6 were measured from AF on hospital admission. Both cystic and non-cystic PVLs were defined by ultrasonographic scans and by conventional MRI, performed at postconceptional age of 32 to 36 weeks. No correlation was found between AF cytokine levels, HCA, and the occurrence of white matter lesions in brain, although a positive correlation between AF cytokine levels and extension of HCA was found. Furthermore, all of the cytokines were detected in higher concentrations in AF when there was proven sepsis during the first 48 hours after birth. Cytokine levels from umbilical cord blood were not measured (Baud et al. 1999a). In a recently published analysis of infants born at < 34 weeks of pregnancy, frequencies of HCA (46.9%) and funisitis (37.5%) in infants with cPVL were slightly higher than in control infants without cPVL (37.9% and 33.7%, respectively), although the differences were not significant. Instead, both the gross findings and the histologic placental findings indicating disturbance in uteroplacental blood flow were found to be significant in detecting cPVL (Kumazaki et al. 2002).

2.6.2 Cerebral palsy Cerebral palsy (CP) refers to a group of non-progressive, but often changing motor impairment syndromes secondary to lesions or anomalies in the developing brain, often accompanied by neurocognitive and sensory disabilities (Mutch et al. 1992). CP includes persistent abnormality of muscle tone and abnormal control of movement and posture

42 (Morton 2001). The live birth prevalence of CP fell from 2.3 per 1000 to 1.4 per 1000 between the periods 1954 to 1958 and 1967 to 1970 (Hagberg & Hagberg 1996). According to a recent report, CP prevalence is 2.12 per 1000 (Hagberg et al. 2001). Although infants born < 32 weeks of gestation represent only 2% of all births (MacDorman et al. 2002), they account for almost one third of all children with CP (Hagberg et al. 2001). The gestational age-specific CP prevalence shows clustering to the most preterm infants: 86 per 1000 for children born < 28 weeks of gestation (extremely preterm), 60 per 1000 for children born between 28 to 31 weeks of gestation (very preterm), 6 per 1000 for those born between 32 to 36 weeks of gestation (moderately preterm), and 1.3 for children born at term (Hagberg et al. 2001). This is consistent with the peak incidence of cPVL between 27 and 30 weeks of gestation (Zupan et al. 1996). Sixty to 100% of infants with cPVL defined by US later develop CP (Leviton & Paneth 1990, Spinillo et al. 1998). As a result of increasing survival among those born most preterm (Goldenberg et al. 2000), the gestational age-specific CP prevalence has increased marginally among the two most immature subgroups of infants but at the same time decreased among full-term infants, although the latter represent over half of the diagnosed CP cases (Hagberg et al. 2001). Recent evidence suggests that prenatal factors are the major determinants of CP especially among children born at term (Croen et al. 2001, Hagberg et al. 2001), and that perinatal asphyxia leading to hypoxic-ischemic encephalopathy accounts for approximately 10% to 20% of all CP cases (Perlman 1997). Peri- and neonatal etiologies account for the majority of CP cases among the most immature newborns (Hagberg et al. 2001). Out of 227 CP children born between 1991 and 1994, a prenatal etiology was recorded in 51% of term births, in 28% of moderately preterm births and only in 3% of very preterm and extremely preterm births. Perinatal and neonatal etiology was defined in 36% of term births, in 25% of moderately preterm births and in 79% of very and extremely preterm births. Postnatal etiology was recorded in 4.5% of term births and only 3% of moderately preterm births (Hagberg et al. 2001). The prenatal risk factors of CP include prematurity, low birth weight, IUGR in term or near-term infants, fetal coagulation disorders, multiple pregnancy, especially with death of the co-twin, antepartum hemorrhage, and chromosomal or congenital abnormalities (Gibson et al. 2003). Maldevelopment of the central nervous system, including neuronal migration disorders or vascular disorders, such as intracranial hemorrhage and/or infarction, account for most of the prenatal factors leading to CP in term infants (Sugimoto et al. 1995, Hagberg et al. 2001). Instead, maternal factors, e.g. socioeconomic status or smoking history, have not been found to associate with CP (Nelson & Ellenberg 1986). Parity has been established as a risk factor for CP in VLBW infants (Grether et al. 1996) born < 32 weeks of gestation (Murphy et al. 1995). Observations made since 1955 (Eastman & DeLeon 1955) have suggested that intrauterine exposure to infection, defined as elevated levels of cytokines in AF or in umbilical cord blood or as HCA or CCA, can influence the development of CP especially in term or near-term infants. According to the previous meta-analysis, HCA (RR 1.6, 95%CI 0.9–2.7) and CCA (RR 1.9, 95%CI 1.4–2.5) were found to predict CP in preterm infants, and CCA (RR 4.7 95%CI 1.3–16.2) was found to predict CP in term infants (Wu & Colford, Jr. 2000). The results are consistent with the recently published meta-analysis

43 (Wu 2002). Table 3 represents a summary of studies describing the association between intrauterine inflammation and CP in both term and preterm infants. Table 3. Intrauterine inflammation and CP in children born either term or preterm. Reference

Subjects

Association between prenatal inflammation and CP

Nelson et al. 1986

Both term and

CCA, RR 0.7, 95%CI 0.4 to 1.2

Cooke 1990

Preterm infants

CCA, p ≤ 0.002

Grether et al. 1997

Term infants

HCA, OR 8.9, 95%CI 1.9 to 40; CCA, OR 9.3, 95%CI 2.7 to 31

Spinillo et al. 1997

Preterm infants

preterm infants

CCA, OR 2.48, 95%CI 0.64 to 9.61; PPROM > 48h, OR 2.98, 95%CI 1.12 to 7.96; stained AF, OR 3.96, 95%CI 1.42 to 11.1

Allan et al. 1997

Preterm infants

CCA, p = 0.02 in univariate analysis; no association in multivariate analysis

Yoon et al. 1997a

Preterm infants

CCA, no association; HCA, p < 0.05, univariate analysis

Redline et al. 1998

Preterm infants

HCA, no association in univariate analysis

Wilson-Costello et

Preterm infants

CCA, no association in univariate analysis

al. 1998 Yoon et al. 2000

Preterm infants

HCA, OR 3.3, 95%CI 0.6 to 18.0; funisits, OR 5.5, 95%CI 1.2 to 24.5

Gray et al. 2001

Preterm infants

Jacobsson et al.

Preterm infants

HCA, OR 1.0, 95%CI 0.4 to 2.4; funisitis, OR 1.0, 95%CI 0 to 2.3; CCA, OR 1.7, 95%CI 0.8 to 3.9

2002 Nelson et al. 2003

CCA, OR 1.77, 95%CI 0.88 to 3.55; HCA, OR 3.61, 95%CI 1.16 to 12.1

Preterm infants

CCA, HCA, no association

AF, amniotic fluid; CCA, clinical chorioamnionitis; CI, confidence interval; CP, cerebral palsy; HCA, histologic chorioamnionitis; OR, odds ratio; PPROM, preterm premature rupture of membranes; RR, relative risk.

In a retrospective study of 31 children with spastic CP and 65 control children, a peripheral blood sample was taken at the median age of 2 days after birth in order to analyze altogether 51 markers of inflammation, autoimmune and coagulation disorders. Eighty-one percent of the children were born at term. CP was diagnosed by the age of 3 years. Concentrations of the cytokines TNF-α, IL-1, IL-8, and IL-9 and the chemokine RANTES were found to be higher in CP children compared to control children with sensitivity and specificity of 100%. The concentrations of IL-6, IL-11, and IL-13, and the chemokines macrophage inflammatory protein (MIP) -1α, -1β, -2, and monocyte chemoattractant protein (MCP) -1, and -2 differentiated CP children from control children with a sensitivity and a specificity that exceeded 88%. Only 4 children with CP were born to women with clinically recognized infection in the admission for delivery. The mean concentration of inflammatory cytokines tended to be higher in the CP children whose Apgar score at 5 minutes was below 6. Furthermore, CP children were also found to have more often evidence of blood coagulation and autoimmune disorders than their controls (Nelson et al. 1998). In an analysis of the same two groups of CP children and controls, 45% of the CP children were found to have higher peripheral blood concentrations of IFN-α, IFN-β, and IFN-γ together with higher concentrations of inflammatory mediators

44 compared to the control children, P < 0.0001 (Grether et al. 1999). In a previous report, IFN-α, given as treatment for hemangiomas to infants, was associated with the development of spastic diplegia (Barlow et al. 1998). The reports on the association between cytokines and CP in preterm infants are somewhat conflicting. In a previous retrospective study of children born between 26 and 35 weeks of pregnancy, 83 infants had neurologic follow-up for at least up to 6 months of age. Seventeen infants were diagnosed to have white matter lesions by US after birth. Eight of these children had CP. None of the cytokines detected in AF, TNF-α, IL-1-β, IL6, or IL-1ra, predicted CP. A higher proportion of neonates who had CP had elevated concentrations of TNF-α, IL-1β, and IL-6 in AF than non-CP children, but the difference between the two groups disappeared after adjustment for birth weight and gestational age (Yoon et al. 1997a). When 14 children (mean gestational age 30.6 wk at birth) with CP diagnosed by the age of 3 years were compared with 109 non-CP children, the median AF concentrations of IL-6 and IL-8 were higher in the pregnancies with children having subsequent development of CP compared to those without CP, P < 0.01. An AF value of IL-6 ≥ 2.95 ng/ml had a sensitivity of 85% and a specificity of 68% and a value of IL8 ≥ 3.0 ng/ml had a sensitivity of 85% and a specificity of 66% to predict CP. After adjustment for gestational age at birth, elevated AF concentrations of IL-6 and IL-8 had OR 6.4, 95%CI 1.3–33.0 and OR 5.9, 95%CI 1.1–30.7, respectively, to predict the development of CP (Yoon et al. 2000). In a recently published study of 64 CP children and 107 controls born at < 32 weeks of gestation, a peripheral blood sample was taken an average of 2.4 days after birth. CP was diagnosed by 2 years of age. None of the measured cytokines, including TNF-α, IL-1, IL-6, and IL-8, were higher in the children with CP than in the control children. Moreover, the cytokine concentrations were not clinically predictive of lesions found after birth in brain ultrasonography, PVL, ventricular enlargement, and/or moderate or severe germinal matrix hemorrhage (Nelson et al. 2003).

2.7 Brain imaging Brain US has been proposed as the first-line method to study brain damage in preterm infants (Ment et al. 2002). In the acute phase of injury, increased echogenity in the periventricular area appears within 24 to 48 hours after a hypoxic-ischemic incident. The affected periventricular white matter is usually equally bright as or brighter than the choroid plexus in contrast to a normal periventricular halo, which is less bright than the choroid plexus. Two to 4 weeks later, cysts may appear in the hyperechogenic areas. Finally, the cysts resolve with development of ventricular enlargement (de Vries et al. 1992). The Increased echogenity is thought to reflect circulatory disturbances in cerebral white matter (Paneth et al. 1990, de Vries et al. 1992, Dammann & Leviton 1997b). The extent, localization, and persistence of increased echogenity together with cyst formation correlate with the later neurologic outcome. In a study of infants born at < 33 weeks of gestation, mild periventricular echodensity (PVE) without cysts and moderate to severe PVE without cysts had specificities of 58% and 83% and negative predictive values of 69% and 76%, respectively for CP diagnosed at the age of 18 to 24 months. By contrast,

45 the presence of moderate to severe PVEs with large cyst formation predicted CP with a sensitivity of 69% and a positive predictive value of 90% and with a specificity of 98% and a negative predictive value of 93% (Pidcock et al. 1990). Transient PVE has been shown to associate with minor motor handicaps in preterm children at the corrected age of 2 years (Ringelberg & van de Bor 1993). Normal white matter echogenity on US has not been found to be a good predictor of normal white matter signal intensity on MRI. Paired US and MRI studies were performed on 32 infants with gestational age of 23 to 30 weeks at birth, beginning at the 4th day of life and proceeding until term age. Mild or no white matter echogenity on US had a sensitivity and specificity of some 50% to predict normal white matter signal intensity on MRI (Maalouf et al. 2001). In infants born at gestational age < 30 weeks, diffuse excessive high signal intensity (DEHSI) has been found to associate with the development of cerebral atrophy (ventricular squaring or dilatation and widened interhemispheric fissure) detected at term age on conventional MRI (Maalouf et al. 1999). The spectrum of leukomalacia diagnosed by US includes transient PVE persisting for ≥ 7 days (grade I), PVE evolving into small localized frontoparietal cystic lesions (grade II), PVE evolving into extensive periventricular cystic lesions (grade III), and PVE extending into deep white matter and evolving into extensive subcortical cystic lesions (grade IV) (de Vries et al. 1992). The size of the cysts and the extent of the lesions associate with the later outcome. In a study of 3451 infants born ≤ 32 weeks of gestation, 96 infants were diagnosed to have cPVL. After 24-month follow-up, 9 children out of the 38 with PVL grade II were free of motor sequelae at 2 years of age compared with only one of the 27 children who had PVL grade III (OR 8.07, 95%CI 0.92–181.7). Twenty-two of the 29 children with PVL grade II who developed CP achieved independent walking compared with only 3 of the 26 children with PVL grade III (OR 75.0, 95%CI 11.4–662) (Pierrat et al. 2001). US is equal to MRI in detecting extensive cystic lesions in brain (Inder et al. 2003b), and it has turned out to be a good predictor for PV-IVH detected on MRI with sensitivities ranging between 69% and 74% and specificities of 83% to 90% (Maalouf et al. 2001), but US has limited sensitivity and specificity to detect diffuse white matter injury (Maalouf et al. 2001, Inder et al. 2003b). In a retrospective study of 51 infants with a mean gestational age of 32 weeks and 6 days at birth, postmortem neuropathologic findings were compared with the premorbid diagnoses of brain US. At autopsy, PVL was diagnosed in altogether 21 infants. US had detected PVL in 7 of these 21 infants. There were 14 false negative and one false positive diagnosis of PVL by US. In 10 out of the 14 false negative US diagnoses, PVL could only be established histologically. The sensitivity and specificity of US to predict PVL were 33% and 97%, respectively. The corresponding figures for detecting PV-IVH and parenchymal hemorrhage by US were 60% and 86%, and 40% and 100%, respectively (Adcock et al. 1998). MR is comparable to autopsy findings in detecting maturation of the cerebral cortex and white matter, hemorrhagic lesions, and infarction (Felderhoff-Mueser et al. 1999). Apart from extensive cPVL diagnosed by US (Leviton & Paneth 1990, Spinillo et al. 1998), MRI predicts neurodevelopmental outcome better than US. Fifty-one children with a mean gestational age of 29.3 weeks at birth underwent neurodevelopmental follow-up until the corrected age of 18 months. Both US and MRI were performed at term. MRI had a sensitivity of 100% and a specificity of 79% to predict CP, whereas the

46 corresponding figures for US were 67% and 85% (Valkama et al. 2000). In a study of 61 infants with a gestational age < 30 weeks at birth, brain US was performed at least twice during the first 2 weeks of life and thereafter in casu. MRI was performed at term age. MRI had a sensitivity of 71% and a specificity of 91% for predicting CP at 20 months of age. The sensitivity and specificity of brain US for predicting CP were 29% and 86%, respectively (Mirmiran et al. 2004). Diffusion-weighted imaging (DWI) is a novel technique, which was first introduced into clinical use some twenty years ago (Le Bihan et al. 1986). It is based on onedimensional molecular displacement of water in tissues. DWI differs from conventional MRI by its ability to infer microstructural and physiological information (Basser 1995). In concordance with brain maturation, the axonal fiber architecture becomes organized, myelination proceeds, and extracellular water content in brain decreases (Morriss et al. 1999, Melhem 2002). This leads to restriction of the molecular movement of water, a phenomenon characterized by the apparent diffusion coefficient (ADC) (Le Bihan et al. 1986, Basser 1995). In developing brain with no evidence of injury, ADC values continue to decline with increasing age after birth (Neil et al. 1998, Morriss et al. 1999, Engelbrecht et al. 2002). This is compatible with a consistent order of myelination in brain (McArdle et al. 1987). For example, at term age, the ADC value in the posterior limb of the internal capsule (PLIC) is lower compared to the ADC value in the anterior limb of the internal capsule (Neil et al. 1998, Morriss et al. 1999), compatible with the progression of myelination in the internal capsule (Barkovich et al. 1988). DWI in brain has been used to evaluate the more diffuse white matter injury known to predominate in very premature infants (Counsell et al. 2003). Conventional MRI and DWI were performed at term on 50 infants born at the median gestational age of 29 weeks. Thirteen infants had normal white matter, 23 infants had DEHSI, and 11 infants had parenchymal white matter lesions. ADC values were measured in frontal, central, and posterior white matter at the level of centrum semiovale. ADC values were significantly higher in infants with DEHSI and with parenchymal white matter lesions than in infants with normal white matter. No difference was found between ADC values in infants with DEHSI and with parenchymal white matter lesions (Counsell et al. 2003). In a case report of an infant born at the gestational age of 30 weeks, both conventional MRI and DWI were performed at the age of 5 days. DWI showed symmetric and diffuse areas of restricted water diffusion in periventricular white matter, and the ADC value was lower compared to the reference value for normal brain. A repeat brain US scan and conventional MRI some weeks later demonstrated bilateral cystic lesions characteristic of cPVL. DWI, again performed at 10 weeks of age, showed an increase in the ADC value in the periventricular white matter surrounding the cystic lesions (Inder et al. 1999b). The latter report demonstrates changes in water diffusion in white matter similar to those observed after acute ischemic injury in term infants, in whom the maximal reduction of diffusion took place at around the 3rd day after injury, and transient normalization of ADC occurred between 7 and 8 days after injury (McKinstry et al. 2002b). In the case of acute stroke, changes in water diffusion are probably related to cytotoxic oedema, i.e. a change in the relative volume of water between the intra- and extracellular spaces (Le Bihan et al. 1992, van Gelderen et al. 1994). Recently, conventional MRI, DWI, and brain US were performed on 11 preterm infants soon after birth. Three infants were defined to have decreased ADC values in PLIC, corona radiata, frontal white matter, and

47 parietal white matter at the mean age of 3.4 days. During follow-up for up to 2 months, every child was confirmed to have severe white matter damage defined as PVL both on US and on conventional MRI (Bozzao et al. 2003). Since axonal and oligodendrocytic damage are the principal characteristics in the pathogenesis of PVL, leading afterward to neurologic sequelae (Volpe 2001), DWI may offer a possibility to investigate white matter maturation in preterm infants who have had sustained white matter injury documented during the neonatal period (Hüppi et al. 2001, Miller et al. 2002). In a longitudinal study of 23 infants with gestational age < 34 weeks, two consecutive MRI and DWI procedures were performed, one soon after birth and the other near term or just before discharge form hospital. On conventional MRI, white matter was defined as normal or minimally or moderately injured. ADC values were measured in gray and white matter. In normal or minimally injured white matter, each week’s increase in age resulted in a significant decrease in the ADC value. In infants with moderately injured white matter, ADC values did not decrease significantly in posterior white matter, while a significant increase in ADC was seen in frontal white matter and in the visual association areas. No neurologic outcome data was included in the study (Miller et al. 2002).

2.8 Neurologic and neurocognitive outcome of ELBW infants Since the overall survival of preterm infants with birth weight of 500 to 1000 g has now increased up to 80% (Goldenberg et al. 2000), interest has been focused on assessing the different aspects of subsequent neurologic, behavioral, and cognitive development among the survivors. Despite the increased survival of ELBW infants, there is little evidence to suggest that their long-term neurodevelopmental outcome has improved during the past two decades (Hack & Fanaroff 2000, Lorenz 2001). In general, the rates of neurologic and neurocognitive disability increase with decreasing gestational age and birth weight (Hack & Fanaroff 2000, Hack et al. 2000, Bhutta et al. 2002). As estimated, one fifth to one quarter of surviving extremely preterm infants (gestational age ≤ 26 weeks at birth) have at least one major disability; impaired mental development, cerebral palsy, blindness, or deafness, and approximately half of all ELBW (birth weight up to 1000 g) infants without major disabilities will have one or more subtle neurodevelopmental impairments at school age (Lorenz 2001). Neurologic evaluation at the mean corrected age of 12.6 months of infants born < 31 weeks of gestation has been shown to predict gross motor outcome at 2 years of age (Frisone et al. 2002). Boys are more likely to have disadvantageous neurologic and especially neurodevelopmental outcome compared to girls (Wood et al. 2000, Hack et al. 2000). Tables 4 and 5 present neurologic and neurodevelopmental outcome studies published between 2000 and 2003 concerning very preterm children born between the years 1986 to 1997 and assessed at the age between 12 to 36 months. According to these reports, the majority of children who survive are, however, free from major neurosensory and neurocognitive disabilities.

48 Table 4. Summary of outcome profiles in very preterm children. Study

Follow-up

Outcome

Hack et al. N = 221;

Population

Corrected age of

No neurodevelopmental disability, 52%;

2000

mean 20 months

CP, 15%;

Birth years 1992–1995 GA, mean 26.4 ± 1.8

Bayley Scales,

wk; BW, mean

MDI < 70, 42%

813 ± 125 g Wood et

N = 283;

Corrected age of

No disability, 49%;

al. 2000

Birth years 1998–1999

median 30 months

Bayley Scales,

GA 22 to 25 wk

MDI/PDI < -3SD, 19%, MDI/PDI < -2SD, 11%; CP, 18%; Other motor impairment, 6%; No recognizable speech, 6%

Rijken et

N = 236;

Corrected age of 2

Abnormal neurologic outcome

al. 2003

Birth years 1996–1997

years

35% < 27 wk,

GA < 32 wk

9% ≥ 27 wk, P < 0.001;

(N = 30, < 27 wk;

Abnormal outcome (neurologic outcome and/or

N = 206, ≥ 27 wk)

Bayley Scales), 36% < 27 wk, 16% ≥ 27 wk (OR 3.0,95%CI 1.3–7.3, P = 0.02)

Tommiska

N = 211;

Corrected age of

et al. 2003

Birth years 1996–1997

mean 18 ± 2 months CP, 11%;

Normal motor development, 76%;

GA, mean 27.3 wk

Other motor impairment, 11.5%;

(range, 22.3 to 34.9 wk);

Severe delay in speech development, 6%

BW, mean 807 g (range, 447 to 995 g) BW, birth weight; CI, confidence interval; CP, cerebral palsy; GA, gestational age; MDI, mental developmental index; OR, odds ratio; PDI, psychomotor developmental index; SD, standard deviation.

Despite the perceived importance of CP, the most common disability during the first two years of life is developmental or cognitive impairment (Hack et al. 2000, Wood et al. 2000), which also has great significance during the school years. The recently published meta-analysis reported cognitive and behavioral outcomes of school-aged children (age range 5 to 14 years) born preterm by combining the results of case-control studies published since 1988. Preterm children had significantly lower cognitive scores compared to their controls. The mean cognitive test scores correlated with birth weight (R2 = 0.51, P < 0.001) and with gestational age at birth (R2 = 0.49, P < 0.001). In addition, this meta-analysis further showed that children born preterm had a 2.64-fold risk for developing attention-deficit/hyperactivity disorder compared to their controls, and that they frequently manifested externalizing or internalizing behaviors at school age (Bhutta et al. 2002). These results are in concordance with the recently published report of extremely preterm children re-assessed at the median age of 6 years. Forty-one percent of the preterm children studied had moderate to severe impairment in their cognitive development (a score more than 2 SD below the mean score of the comparison group) compared to only 1% in their term-born classmates. Severe disability at the age of 30

49 months was also found to be highly predictive for the outcome at 6 years of age. Eightysix percent of children classified as having severe disability at the corrected age of 30 months had either severe or moderate disability at 6 years of age. Furthermore, 24% of children categorized as having no disability at the age of 30 months had moderate or severe disability when assessed at 6 years of age (Marlow et al. 2005). The apparent deterioration of neurodevelopmental outcome in preterm children during the school years (O'Brien et al. 2004) may predict future difficulties in educational and academic areas. In a prospective case-control study of 242 VLBW infants assessed at 20 years of age, 40% of the VLBW cases had repeated a grade at school compared with 27% of the normal-birth-weight participants. Furthermore, 74% of the cases had graduated from high school compared with 83% of the controls. The VLBW cases also had lower mean intellectual quotients (IQ) and lower academic achievement scores than the controls. Thirty percent of the VLBW men were enrolled in postsecondary studies compared with 53% of men in the control group (Hack et al. 2002). Table 5. Outcomes among children with extreme immaturity according to different follow-up studies published between the years 1986 and 1996 (Hack & Fanaroff 2000). Outcome

GA 24 wk

GA 25 wk

BW 500 to 800 g

Severe disability*

22% (95% CI 6–48) to

12% (95% CI 3–27) to

9% (95% CI 3–21) to

45% (95% CI 28–64)

35% (95% CI 15–59)

37% (95% CI 20–56)

Cerebral palsy

11% (95% CI 1–35) to

3% (95% CI 0–17) to

5% (95% CI 1–25) to

15% (95% CI 5–32)

20% (95% CI 8–39)

37% (95% CI 21–56)

Subnormal cognitive

14% (95% CI 2–43) to

10% (95% CI 2–26) to

13% (95% CI 7–19) to

function†

39% (95% CI 17–64)

30% (95% CI 23–37)

Blindness

0% to

2% (95% CI 0–11) to

9% (95% CI 2–24)

25% (95% CI 12–43)

47% (95% CI 29–65)

BW, birth weight; CI, confidence interval; GA, gestational age. * Severe disability includes cerebral palsy, other major neurologic impairment, blindness, deafness, or subnormal cognitive function. † Subnormal cognitive function includes developmental quotient (DQ) on Griffiths Scales < -2SD or mental developmental index (MDI) on Bayley Scales < 68 or < 70.

3 Purpose of the study Among very preterm children, abnormalities in neuromotor and neurocognitive outcome continue to be the major problems during infancy and later childhood. Pre- and perinatal events account for the brain damage and neurologic sequelae. New techniques of brain imaging offer an insight to view more diffuse injury not evident on conventional imaging. The principal purposes of the present cohort studies were: 1. To examine the relationships between several soluble protein mediators in umbilical cord blood and CP in full-term and preterm children. 2. To investigate whether prenatal inflammation serves as a risk factor of cPVL, IVH grade IV, or IVH without parenchymal involvement, and whether it relates to the neurologic and neurodevelopmental outcome in very preterm ELBW children. 3. To evaluate fetal cardiovascular hemodynamics in relation to neurodevelopmental outcome at 1 year of corrected age in children born from pregnancies complicated by placental insufficiency and delivery before 32 weeks of gestation. 4. To investigate whether the ADC value, a quantitative indicator of water diffusion on DWI in brain, correlates with measurements of brainstem auditory evoked potentials (BAEPs) and with neurodevelopmental outcome in very preterm ELBW children.

4 Subjects and methods The study protocols were approved by the ethics committee of Oulu University Hospital. A written informed consent was signed by the parents. The children in the studies II–IV were part of a larger prospective cohort of 163 infants live-born < 32 weeks of gestation between November 1998 and November 2002 in Oulu University Hospital. Table 6 represents a summary of the children enrolled in the studies I–IV. Detailed descriptions of the subjects and methods are also presented in the original articles (I–IV). Table 6. Children included in studies I–IV. Study number and type of

Gestational age and birth

Number of children

Characteristic of the final

weight of the study

included in the study

study population

population

population

Study I, retrospective,

Term (GA ≥ 37 wk at

41 cases, 41 gestation-

cross-sectional study

birth) and preterm

matched controls, 41

controls, 19 gestation-

(GA < 37 wk at birth)

term-born random

matched pairs, 30 random

children

controls

controls

Preterm children,

61

the study

Study II, prospective, longitudinal study

GA < 32 wk at birth,

27 cases, 25 gestational

N = 54 One child excluded from

BW < 1000 g

follow-up because of a brain tumor diagnosed at 1 year of age.

Study III, prospective,

Preterm children,

cross-sectional study

GA < 32 wk at birth

17

N = 17 Suboptimal outcome (group 1), N = 7 Normal outcome (group 2), N = 10

Study IV, prospective, longitudinal study

Preterm children,

53

N = 30

GA < 32 wk at birth,

Division into groups 1 to

BW < 1000 g

3 according to scores on Griffiths subscales

BW, birth weight; GA, gestational age.

52

4.1 Subjects Study I comprised children born in four university hospitals (Helsinki, Oulu, Tampere, and Turku) in Finland between September 1, 1992 and August 31, 1993. Patient data were collected from databases and medical records to identify all children with CP born during the period in question in the above-mentioned university hospitals. CP was defined as a syndrome that includes persistent abnormality of muscle tone, movement and posture, resulting in functional impairment due to a nonprogressive lesion of the immature brain. The diagnoses of CP were made by child neurologists by the age of 5 years on the basis of standard criteria (Hagberg & Hagberg 1993). Children with congenital brain malformation, prenatal viral or protozoal infection, brain damage after the neonatal period, or chromosomal abnormality were excluded. Sixty-one term and preterm children were tentatively identified as having CP. The following CP cases were excluded: in 11 cases no cord blood specimens were available, 5 parents refused to participate in the study, 1 parent could not be reached, and 3 children were excluded because the diagnostic criteria were not met. The final study group consisted of 41 CP children. Twenty-two had spastic diplegia, 12 had spastic hemiplegia, and the remaining 7 had quadriplegia (3 the spastic form, 4 the dystonic form). Four children had a twin pair. One paired control matched for gestational age was selected for each of the 41 infants with CP. To recruit the gestation-paired controls, the data of the next born infant of similar gestational age (± 1.5 weeks) and without CP and the exclusion criteria were extracted from the records of the neonatal intensive care unit of Oulu University Hospital. Because all gestation-matched control children required intensive care after birth, an additional reference group of presumably healthy term-born children was enrolled in the study by extracting every fifth consecutively numbered umbilical cord specimen after the case without individual identifier. A healthy sibling of each of the four pairs of twins served as an additional control to its sister or borther. The studies II and IV covered a regional cohort of 78 consecutive very preterm (gestational age < 32 wk) ELBW (birth weight < 1000 g) infants born between November 1998 and November 2002. Only 4 additional ELBW infants were born alive in Northern Finland during the period in question. Three infants were born in secondary hospitals and one was born in a regional health center. Two of the 4 infants died soon after birth. In study II, the following exclusion criteria were used: 1) refusal of consent, 2) unavailability of umbilical cord sample, 3) death in delivery room, 4) lethal disease due to extreme immaturity, 5) serious congenital disease, and 6) inability to participate in follow-up. Altogether 61 infants were eligible for the study protocol. The following 7 infants were additionally excluded: 5 infants died during the neonatal period, 1 parent refused to continue the study, and 1 infant was lost to follow-up. The final study population comprised 54 infants. Study III consisted of 17 singleton infants. The following inclusion criteria were used: 1) pregnancy complicated by placental insufficiency, defined as abnormal UA blood velocity waveform pattern for gestational age, 2) Doppler ultrasonography of placental and fetal cardiovascular hemodynamics performed within 24 hours before delivery, and 3) none of the infants had congenital malformations or chromosomal abnormalities.

53 Study IV consisted of the same prospectively collected regional cohort (N = 78) as study II, provided that both the conventional MRI and DWI assessments were performed during the same imaging procedure at term. Conventional MRI was successfully performed on 53 infants. However, 23 infants had to be excluded from the final study group because of inaccurate DWIs due to movement artefacts or technical problems. The final study group consisted of 30 very preterm ELBW infants. Sixteen infants made up an additional subgroup that attended prospectively a BAEP examination at term age. One infant with normal results on brain imaging at term was excluded from neurodevelopmental follow-up in the studies II and IV because of a brain tumor diagnosed at 1 year of corrected age. In the studies I–IV, gestational age was confirmed by ultrasonographic examination before 20 weeks of gestation. Infants with birth weight below 2 standard deviations from the mean of the gestation-adjusted birth weight according to the Finnish growth standards were classified as small for gestational age (SGA).

4.2 Methods 4.2.1 Analysis of cord serum proteins For the studies I–III, umbilical cord blood was collected into dry sterile tubes immediately after birth. After centrifugation at 3000 rpm for 15 minutes, the sera were separated and stored frozen until the analyses. In study I, the leftover sera for screening of congenital hypothyreoidism were used. The protein mediators were analyzed by using antibody-based protein microarrays with DNA amplification (Schweitzer et al. 2002, Kingsmore & Patel 2003). The protein mediators analyzed in the study I are listed in Table 7. In study II, altogether 4 proinflammatory cytokines, IL-1α, IL-1β, IL-6, and TNF-α, and one chemokine, IL-8, were analyzed. In study III, altogether 9 protein mediators, IL-1α, IL-1β, IL-6, IL-8, TNF-α, intercellular adhesion molecule (ICAM)-1, -3, vascular endothelial growth factor (VEGF), and VEGF-receptor2 (R2), were analyzed. Antibody microarrays were printed on cyanosilane-coated glass slides divided by Teflon boundaries into 16 circular wells 0.5 cm in diameter, using a Packard Biosciences (Downers Grove, IL) BCA-II piezoelectric dispenser. Monoclonal antibodies for the cytokines (R&D Systems, Minneapolis, MN; PharMingen, San Diego, CA) were dispensed onto arrays. A titration of biotinylated mouse IgG spots (Bio-mIgG) was printed and used as internal calibrators for the assay. The printed slides were blocked (Schweitzer et al. 2002) and stored at 4ºC until use. The printed spots were examined by subjecting batches of slides to a quality control procedure consisting of incubation with a fluorescently labeled anti-mouse antibody followed by washing, scanning, and quantitation. The assays were performed by a liquid-handling robot (Biomek 2000; Beckman Instruments, Fullerton, CA) in the laboratory of Molecular Staging Inc., New Haven, CT. After incubation of serum samples on microarrays, captured proteins were detected by

54 specific biotinylated second antibodies, after which a universal anti-biotin antibody was bound to the second antibodies. The anti-biotin antibody contained an oligonucleotide DNA primer to generate a fluorescent signal. In the process of rolling-circle signal amplification, a circular DNA hybridizes to the oligonucleotide DNA primer in the presence of DNA polymerase and fluorescent nucleotides (Schweitzer et al. 2002, Kingsmore & Patel 2003). For every slide, a set of negative controls were run, and the intensity values were used to correct for background signal. The slides were scanned for fluorescence (GenePix, Axon Instruments, Foster City, CA). The fluorescence intensity of each analyte in each sample was used to determine cytokine values. The coefficient of the variation and assay sensitivities were as described (Schweitzer et al. 2002, Kingsmore & Patel 2003). Untransformed fluorescent intensities were used as data values.

4.2.2 Data quality control Samples were excluded from statistical analyses if 1) the fluorescent intensities were generally weak {[sum (all cytokine signals)] /[sum(all bio-mIgG signals)] > 1SD below median}, indicating sample degradation during storage, 2) there were visible defects in the array, or 3) there was a high-intensity background signal (> 1000 intensity units). In study I, 22 specimens could not be analyzed because of insufficient sample volume, and 22 samples were excluded on the basis of excessive background signal. In study II, altogether 2 specimens were excluded. In study I, the accuracy of EGF analysis was confirmed using the immunoassay kit (R&D Systems, Minneapolis, MN). Because of 4 twin pairs, only 1 sample was successfully analyzed, and the results of twin pair analysis were not presented.

55 Table 7. List of the proteins analyzed in study I. Amphiregulin (AR)

Macrophage colony-stimulating factor (M-CSF)

Angiogenin (ANG)

Macrophage derived chemokine (MDC)

Apoptotic receptor molecule on lymphocytes (FAS)

Macrophage inflammatory protein (MIP)-1α,-1β,-1δ

Brain-derived neurotrophic factor (BDNF)

Macrophage migration inhibitory factor (MIF)

B-lymphocyte chemoattractant (BLC)

Macrophage-stimulating protein (MSP)

Ciliary neurotrophic factor (CNTF)

Monocyte chemoattractant protein (MCP)-1,-2,-3

Eotaxin (Eot)

Monokine induced by interferon γ (MIG)

Eotaxin 2 (Eot 2)

Myeloid progenitor inhibitory factor (MPIF)-1

Epidermal growth factor (EGF)

Neurotrophin (NT)-3,-4

Epithelial cell-derived neutrophil-activating peptide -

Neutrophil-activating peptide-2 (NAP)

78 (ENA-78) Fibroblast growth factor (FGF) -6,-7,-9

Oncostatin M (OSM)

Fms-like tyrosine kinase-3-ligand (Flt-3Lig)

Placental growth factor (PLGF)

Glial cell line-derived neurotrophic factor (GDNF)

Pulmonary and activation-regulated chemokine

Granulocyte colony-stimulating factor (G-CSF)

RANTES

Granulocyte-macrophage colony-stimulating factor

Soluble CD23 (sCD23)

(PARC)

(GM-CSF) Hemofiltrate CC-chemokine-4 (HCC4)

Soluble glycoprotein 130 (sGP130)

I-309

Stem cell factor (SCF)

Interferon (IFN)-α,-γ

Stromal cell-derived factor (SDF)-1α

Interferon-inducible protein (IP)-10

Thymus and activation-regulated chemokine (TARC)

Interleukin (IL)-1α,-1β,-2,-3,-4,-5,-6,-7,-8,-10,-11,-

Transforming growth factor (TGF)-β1,- β2

12,-13,-15,-16,-17,-18 IL-1 receptor antagonist (IL-1Ra)

Tumor necrosis factor (TNF)-α,-β

IL-1 soluble receptor 1 (IL-1sR1)

TNF-receptor (TNF-R)-I,-II

IL-2 soluble receptor antagonist (IL-2sRa)

TNF-related apoptosis inducing ligand (TRAIL)

IL-6soluble recetor (IL-6sR)

Urokinase-type plasminogen activator receptor (uPAR)

Leukemia inhibitory factor (LIF)

Vascular endothelial growth factor (VEGF) Vascular endothelial growth factor-receptor2 (VEGFR2)

4.2.3 Pathology of placenta All the placentas examined in the studies II and III were fixed in 10% neutral buffered formalin immediately after birth. The rim of membranes was taken from the site of membrane rupture. The umbilical cord specimens were taken from the fetal and placental sides of the umbilical cord and from midway between both sides of insertion. The fullthickness specimen of placental parenchyma was taken from midway between the umbilical cord insertion and the placental margin.

56 For histologic evaluation, paraffin blocks were made, cut into 5µm slices, and stained with hematoxylin-eosin. HCA was defined as the prescence of PMN leukocytes in at least one of the three compartments examined, extraplacental membranes, chorionic plate, and umbilical cord. Placental perfusion defect was defined as poorly vascularized villi, multiple capillary lumina in some villi, increased intervillous volume, and reduced total villous capillary bed. Both the HCA and the placental perfusion defect were assessed according to the standard criteria (Benirschke & Kauffmann 2000). In order to compare the placental histology and the levels of protein mediators in cord blood, the placentas were divided into groups. In study II, the placentas were divided into four groups: isolated HCA, isolated perfusion defect, HCA and perfusion defect, and placentas without HCA or perfusion defect. In study III, all of the index placentas examined (n = 17) were further compared with 25 placentas with isolated HCA (HCA placenta group) and with 10 placentas without either HCA or placental perfusion defect (normal placenta group). Eight out of the 25 placentas with isolated HCA and 7 out of the 10 placentas without either HCA or placental perfusion defect were the same placentas as in study II. The other 20 placentas were derived from the cohort of 163 infants mentioned above.

4.2.4 Brain imaging 4.2.4.1 Brain ultrasound Brain US scans in the studies II and III were obtained by using HDI 5000 (Advanced Technology Laboratories Ultrasound, Botwell, WA, USA) with a curved-array 5–8 MHz transducer. The infants underwent serial brain ultrasound assessments at the following ages: 1 to 3 days, 1, 2, and 4 weeks, and thereafter every 4th week until discharge or term. A pediatric radiologist performed the prospective evaluation and filled in a form after each examination. Afterward, the results were finally reviewed by a single radiologist. The final outcome was the consensus of these two evaluations. PVE (Pidcock et al. 1990), PV-IVH (Papile et al. 1978), and PVL (de Vries et al. 1992) were classified using the standard criteria.

4.2.4.2 Magnetic resonance imaging Conventional MRI was performed in the studies II and IV, and DWI was performed in study IV at term age. A 1.5-T system (Signa Horizon Echo Speed; General Electrics, Milwaukee, Wis) and a head coil were used. The study protocol consisted of T1-weighted sagittal images [TR/TE = 400–460 / 9, matrix 256 × 192, 4 mm slices, 1 mm slice gap and 20 × 20 cm field of view (FOV)], fast-spin echo T2-weighted axial images (TR/TE = 3000–4660 /83–168, matrix 256 × 192 or 512 × 320, 5 mm slices, 0.5 mm slice gap and 20 × 15 cm FOV), and T1-weighted axial images (TR/TE = 460–600/10–20, 256 × 192 or 256 × 256 matrix, 4–5 mm slices, 0.5 mm slice gap and 20 × 15 cm FOV).

57 Axial DWI was obtained by using a single-shot spin-echo echo planar sequence with the following imaging parameters: TR 10 000 ms, TE 69.1–100.7 ms, slice thickness 7 mm, matrix 128×128, and FOV 36×22 cm or 25×18 cm. Each slice was acquired with a b value (diffusion weighted factor) of 0 s/mm2 and a diffusion-sensitive gradient pulse with a b value of 1000 s/mm2 applied separately in three orthogonal directions (x, y, z). The analysis of DWI data was performed on a separate workstation (Advantage Windows 4.0, GE Medical Systems, Milwaukee, Wis.). Apparent diffusion coefficient (ADC) maps were obtained by using software created by the manufacturer. In addition to diffusionsensitized images in three orthogonal directions, a combined image of each section was encoded by the software. Signal intensity measurements were performed in circular or ovoid regions of interest (ROI) ranging from 21 mm2 to 202 mm2 in size, depending on the diameter of the target area, and placed on reference scans (b = 0 s/mm2). The system automatically transferred the ROI to the same areas in the corresponding diffusionweighted images. ADC value was expressed as a mean and standard deviation (SD) within the ROI. The anatomical regions of signal intensity measurements included pons, bilateral PLIC, corona radiata, frontal white matter, posterior white matter, and centrum semiovale. ADC in cerebrospinal fluid, measured from the lateral ventricles, where accurate setting of ROI was possible, served as a reference of free diffusion. Before and during the MRI examination, the infants were fed and swaddled to keep them warm, fitted with disposable ear plugs, and given a dose of chloral hydrate (25– 50 mg per kilogram). Oxygen saturation and cardio-respiratory status were monitored during the imaging procedure by using a MR-compatible pulse oximeter probe (MR-9500, Berner/Nissan, USA). The infant´s condition was evaluated throughout the examination by a pediatrician. White matter damage was defined as abnormal signal intensity on conventional T1and T2-weighted images, presenting with parenchymal hemorrhage or cystic lesions with or without petechial hemorrhage in the white matter area together with reduction of white matter thickness and compensatory ventricular enlargement. Ventricular size, white matter reduction, widening of sulci, abnormal signal intensity, and irregularity in the shape of the lateral ventricles were assessed visually.

4.2.5 Placental and fetal hemodynamics In study III, Doppler ultrasonographic examinations of placental and fetal hemodynamics were performed using image-directed pulsed and color Doppler equipment (Acuson Sequoia 512; Acuson Corporation, Mountain View, CA) with a 4 to 8 MHz convex or a 5 MHz sector probe. Values were measured from three consecutive cardiac cycles, and the mean value was used for further analysis. UA blood velocity waveforms were obtained from free loops of the umbilical cord. Placental vascular impedance was assessed by calculating the UA pulsatility index value (PI = [peak systolic velocity – end-diastolic velocity]/time-averaged maximum velocity over the cardiac cycle). The distribution of fetal arterial circulation was evaluated by calculating PI values of DAo and MCA. From AoI, the net antegrade and net retrograde blood velocity waveform components were measured and their ratio was calculated

58 (Fouron et al. 2001). Volumetric blood flows (Q) across the pulmonary (PV) and aortic valves (AoV) were calculated (Mäkikallio et al. 2002). Right ventricular cardiac output (RVCO) equals QPV, left ventricular cardiac output (LVCO) equals QAoV, and CCO is their sum. The proportions (%) of RVCO and LVCO of CCO and the weight-indexed RVCO, LVCO, and CCO were calculated. PI values for veins were calculated from the DV, LHV, and IVC blood velocity waveforms (Mäkikallio et al. 2002). Actual birth weight was used for indexing purposes because the interval between Doppler ultrasonography and delivery was less than 24 hours.

4.2.6 Brainstem auditory evoked potentials In study IV, BAEPs were performed by using the Nicolet Viking IV (Nicolet Viking Inc., Madison, Wis). Rarefraction clicks of 100 μs were delivered separately to both ears through earphones. The stimulation rate was 11.1 Hz, the intensity both 40 dB and 75 dB the normal hearing level with contralateral masking. The vertex responses to the stimulation of ears were recorded with cup electrodes and referenced to the ipsilateral and contralateral mastoids. The reproducible waveform responses to 2000 clicks or more to both ears were averaged at least twice. The following variables were determined from the screens: 1) absolute peak latencies of the waves I, III, and V, 2) interpeak latencies (IPL) of I–III, III–V, and I–V, and 3) amplitude ratio of wave V to wave I (V/I). Wave I reflects the compound action potential in the peripheral portion of the 8th nerve. Wave III is generated in caudal pons. Wave V originates in the midbrain. IPL I–III represents conduction through the proximal acoustic nerve to pons. IPL III–V represents conduction through pons to midbrain. IPL I–V represents the total central conduction through the brainstem (Markand 1994).

4.2.7 Neurologic and neurodevelopmental outcome In study I the children were examined until the age of 5 years. In the studies II and IV the children were evaluated at the corrected age of 2 years, and in study III the children were evaluated at the corrected age of 1 year. In study I, CP was diagnosed by a child neurologist using the criteria of Hagberg et al. (Hagberg & Hagberg 1993). In the studies II and III, the children were assessed for neurologic status with special attention to the pattern of movements and posture, muscular tone, and reflex status. An abnormal neurologic outcome was defined as neuromotor dysfunction based on the functional ability of the most affected limb. Neurodevelopmental evaluation was carried out on the children in studies II–IV using the Griffiths Developmental Scales (Grifftihs 1954). For each child, a total score of developmental quotient (DQ) was calculated (studies II–III). In study IV the 5 different subscales on the Griffiths Developmental Scales evaluating gross motor, personal-social, hearing and speech, eye-hand coordination, and performance abilities were assessed. In study II the neurodevelopmental outcome was also reported by using the Bayley Scales of Infant Development (Bayley 1993).

59 Two outcome groups based on neurologic and neurodevelopmental evaluation were presented in study III. Three outcome groups were defined in study IV on the basis of the scores on the Griffiths Developmental Subscales. The division into the study groups is presented in Table 8. Table 8. Criteria of division of children into groups in studies III–IV Study number

Study groups

Study III

The highest DQ on Griffiths Scales among the infants with abnormal neurologic outcome was 97. Group 1 consisted of infants with DQ ≤ 97 (= suboptimal outcome group); Group 2 consisted of infants with DQ > 97 (= normal outcome group).

Study IV

Group 1 consisted of children with subscale scores < - 1SD of the mean of the reference (Griffiths 1954); Group 2 consisted of children with subscale scores within ± 1SD of the mean of the reference; Group 3 consisted of children with subscale scores > + 1SD.

DQ, developmental quotient; SD, standard deviation.

4.2.8 Pre- and postnatal clinical data In study II, the prenatal clinical data included maternal age at birth, socioeconomic factors, parity, history of previous preterm births, pre-existing maternal medical conditions, administration of antenatal steroid, antibiotic use before delivery, the time point of PPROM, mode of delivery, and the presence of active labor defined as regular contractions, cervical shortening and dilatation (Hartmann et al. 1999). The criteria of clinical chorioamnionitis were fever ≥ 37.8° C and at least two of the following: leukocytosis ≥ 15,000 cells/mm3; maternal pulse rate > 100/min or fetal pulse rate ≥ 160/min; uterine tenderness; and foul-smelling amniotic fluid (Gibbs et al. 1982). In study III, during the prenatal period, computerized analysis of short-term variation (STV) of fetal heart rate (FHR) was performed daily (Sonicaid, Oxford Instruments, England), and the last recording prior to delivery was used for analysis. The intrapartum and postnatal clinical data in studies II and III included umbilical artery blood gas values, 5-minute Apgar scores, gestational age, birth weight, and evaluation of the infant´s cardiorespiratory status.

4.3 Statistics The Statistical Package for Social Sciences software (SPSS Inc., Chicago, Illinois, IL) was used to make all the statistical analyses. All the tests were two-tailed. Chi-squared test or Fisher´s exact test was used for the differences in the categorical variables between the two groups. The differences in the continuous variables were studied by using unpaired t-test if the variable between the two groups was normally distributed, otherwise the Mann-Whitney U-test was used. In study I the primary analysis was done with paired

60 t-test to compare the differences of the cytokines between pairs of CP children and their gestation-matched controls. When the differences of the continuous variables were used to study more than two groups, one-way analysis of variance (ANOVA) with a post hoc test of Bonferroni or Tukey (studies I and IV, respectively) or Kruskal-Wallis with a post hoc Mann-Whitney U-test (study III) were used. In study I ANOVA was used after logarithmic transformation. In study II, when comparing the umbilical cord serum cytokines between the 4 placental groups, gestational age was adjusted by using univariate analysis of variance, and unpaired t-test was used as a post hoc test. Cytokines were correlated with the length of gestation by using Spearman´s rank correlation coefficient (study I) or Pearson´s correlation coefficient after logarithmic transformation (study II). The ADC values were correlated with the neurophysiologic measurements (BAEP) and with the scores on the neurodevelopmental subscales by using Pearson´s correlation coefficient after logarithmic transformation. A receiver-operator characteristic (ROC) curve was used to evaluate the relationship between the sensitivity and specificity for a cut-off value of IL-6 to predict active preterm labor (study II) and for a cut-off value of ADC in corona radiata to predict neurodevelopmental outcome (study IV). Finally, multiple logistic regression analysis (study II) and stepwise linear regression analysis (study IV) were used to evaluate covariates as independent predictors for dependent variable.

5 Results 5.1 Prenatal inflammation (I, II, III) 5.1.1 Umbilical cord serum proteins and gestational age (I, II) We found gestational age at birth to associate with umbilical cord serum levels of certain protein mediators. First, IL-12p70 (r = 0.501, P = 0.008), IL-13 (r = 0.435, P = 0.02), MCP-3 (r = 0.433, P = 0.02), and MIG (r = 0.442, P = 0.02) correlated with gestational age in children with CP, but no correlation was found in gestation-matched controls (study I). On the other hand, the cord serum level of EGF correlated with gestational age in control children (r = 0.459, P = 0.02) but not in children with CP (study I). Finally, the umbilical cord serum levels of altogether 11 protein mediators correlated with the length of gestation among the CP cases and their gestation-matched controls. BDNF (r = 0.476, P < 0.001), IL-1sR1 (r = 0.400, P = 0.003), IL-11 (r = 0.444, P = 0.001), LIF (r = 0.493, P < 0.001), MDC (r = 0.580, P < 0.001), PARC (r = 0.471, P < 0.001), and TGF-β1 (r = 0.394, P = 0.004) increased with gestation (study I). Cord serum TNF-α level was found to increase with gestation in the population of 5-year-old CP children and their controls (r = 0.367, P = 0.007) (study I). However, among very preterm ELBW children, TNF-α declined with increasing gestational age (r = -0.424, P = 0.002) (study II). IL-6 (r = -0.353, P = 0.010) (study I), (r = -0.372, P = 0.007) (study II) and IL-8 (r = -0.405, P = 0.003) (study I), (r = -0.382, P = 0.006) (study II) also declined with increasing gestation. The levels of IL-1α (r = -0.416, P = 0.003) and IL-1β (r = -0.458, P = 0.001) declined with gestation in very preterm ELBW children (study II), and uPAR declined with gestation in 5-year-old CP children and their controls (r = -0.301, P = 0.030) (study I).

62

5.1.2 Placental pathology, cytokines and active preterm labor (II, III) Study II covered 25 infants born to 24 mothers with HCA. Two mothers had clinical chorioamnionitis, and they both had HCA. PPROM was evident in 16 (67%) mothers with HCA compared to 4 (14%) mothers without HCA, P < 0.001. Active labor was evident in altogether 27 (52%) mothers. HCA predicted active labor with a sensitivity of 79% and a specificity of 87% (OR 24.44, 95%CI 5.39–110.9, P < 0.001). Cord serum IL6 was higher in the isolated HCA group than in the group with no HCA or perfusion defect, P = 0.036, and cord serum IL-8 was higher in the isolated HCA group than in the perfusion defect group, P = 0.008. Furthermore, cord serum IL-1β tended to be higher in the isolated HCA group than in the group with no HCA or perfusion defect, P = 0.120. No differences were found between HCA and cord serum IL-1α or TNF-α. In addition, in study III, the levels of cord serum proinflammatory cytokines IL-1β, IL-6, and IL-8 were higher in the isolated HCA group than in the index group of placentas with perfusion defect, P ≤ 0.01. In study II, high IL-6 in cord serum (cut-off level > 800 using ROC curve analysis) predicted active preterm labor with a sensitivity of 85% and specificity of 57% (OR 7.67, 95%CI 1.95–30.1, P = 0.005). Infants born to mothers with isolated HCA had lower gestational age (26.1 wk ± 1.5) than other infants (27.6 wk ± 1.7), P = 0.002 (study II). Table 9 presents the cytokine levels according to the histological findings in placenta in study II.

63 Table 9. Differences between cord serum cytokines and placental histology (study II). Variable

Isolated HCA

No HCA, no

Isolated perfusion

HCA+perfusion

perfusion defect

defect

defect

N = 19

N = 13

N = 16

N=6

26.1 ± 1.5

26.5 ± 1.4

28.3 ± 1.7

28.1 ± 1.6

Mean

1730.3

658.2

510.4

742.6

SEM

527.0

92.1

59.2

272.9

Adjusted mean

1606.9

610.5

631.2

844.8

Mean

9737.6

733.2

529.4

1771.2

SEM

4096.9

137.3

90.1

1207.6

Adjusted mean

8788.8

366.3

1458.6

2557.6

Mean

26711.1

5990.0

847.0

9763.9

SEM

6514.4

4635.4

89.6

8970.1

26289.0*

5841.2

1314.0

10160.2

Mean

23214.5

12249.0

4488.5

11661.2

SEM

4721.5

5222.1

945.4

7786.6

22673.6†

12058.4

5087.0

12169.0

Mean

630.5

486.5

525.3

498.3

SEM

91.1

111.7

79.0

106.3

Adjusted mean

533.2

448.9

620.5

578.9

Gestational age (wk, mean ± SD) IL-1α

IL-1β

IL-6

Adjusted mean IL-8

Adjusted mean TNF-α

HCA, histologic chorioamnionitis; SEM, standard error of the mean. The results were adjusted for gestational age using univariate analysis of variance and t-test as post hoc.* Isolated HCA different from the placentas without HCA or perfusion defect, P = 0.036. † Isolated HCA different from the placentas with isolated perfusion defect, P = 0.008.

5.1.3 IVH grade II to III, WMD and inflammation (II) Seven infants out of 54 were diagnosed to have WMD. Three infants had IVH grade IV and 4 infants had cPVL. The diagnosis of WMD was made in consensus with the findings of brain US and conventional MRI. Altogether 22 infants had no IVH. IVH grade I was diagnosed in 18 infants. Eleven infants had IVH grade II to III, and this was evident within the first seven days after birth in each infant. In five of these infants IVH grade II to III was diagnosed during the first three days after birth, and each of them was born from a pregnancy complicated by HCA. HCA associated with IVH grade II to III. 81.8% of the infants with IVH grade II to III were born from pregnancies complicated by HCA compared to 40.0% of the infants with

64 no IVH or IVH grade I, P = 0.019. HCA independently predicted IVH grade II to III. Infants born from pregnancies complicated by HCA were found to have a 8.10-fold risk of having IVH grade II to III than those without HCA (OR 8.10, 95%CI 1.55–42.5, P = 0.013) after adjustment for confounders. No detectable association was found between the levels of umbilical cord serum cytokines and IVH grade II to III. HCA of any degree of severity or in any combination (fetal chorionic plate vasculitis and/or funisitis included) failed to associate with cPVL and IVH grade IV (WMD). Umbilical cord serum cytokine levels were equal in infants with and without cPVL and IVH grade IV.

5.1.4 Inflammation and neurologic outome in preterm children (I–III) In study I, the analysis of 5-year-old CP children (gestational age < 37 wk at birth) and their paired gestation-matched controls revealed that altogether 5 protein mediators in umbilical cord serum differed significantly between the preterm cases and their paired controls. EGF was higher among the cases. GM-CSF, IL-2, MDC, and PARC were higher among the controls (Table 10). Cord serum EGF was found to be higher in the preterm CP children with diplegia, P = 0.02 and the CP children with other type of CP, P = 0.03, compared with their paired controls. In non-paired analysis between the preterm cases (n = 16) and their preterm controls (n = 13), EGF remained higher in the preterm CP group compared with the preterm control group, P = 0.045. Cord serum levels of the proinflammatory cytokines IL-1α, IL-1β, IL-6, and TNF-α, and a chemokine IL-8 did not associate with CP diagnosed by the age of 5 years in children born before term (study I), nor with abnormal neurologic outcome in 2-year-old ELBW children born very preterm (gestational age < 32 wk) (study II). Also, no differences were found in the umbilical cord serum proinflammatory cytokines, ICAM-1, ICAM-3, VEGF, or VEGFR2 between the suboptimal outcome group and the normal outcome group of children born < 32 weeks of gestation (study III). Furthermore, HCA was not more prevalent among the very preterm ELBW children with abnormal neurologic outcome than in children with normal outcome. In fact, 50% of the children with abnormal neurologic outcome at 2 years of age and 44.2% of the children with normal outcome were born from pregnancies complicated by HCA, P = 0.739 (study II). Instead, study II showed that, when the pregnancy was complicated by both HCA and placental perfusion defect, the compound effect had an independent association with abnormal neurologic outcome among very preterm ELBW children (OR 15.2, 95%CI 1.3-181). Furthermore, compound placental defect and WMD additively predicted abnormal neurologic outcome (OR 14.9, 95%CI 1.2–192 and OR 20.7, 95%CI 2.63–163, respectively). Compound placental defect and/or WMD predicted abnormal neurologic outcome with a specificity of 88%, sensitivity of 70%, positive predictive value of 58%, and negative predictive value of 93%, P < 0.001. In study II, the effect of prenatal inflammation was further assessed by comparing the total scores on Griffiths Developmental Scales and the Mental Developmental Index on Bayley Scales between very preterm ELBW children with and without HCA. No differences were found in the developmental outcome scores of the children with and

65 without HCA (Griffiths Scales, 92.4 ± 22.9 and 95.2 ± 22.5, respectively, P = 0.680 and Bayley Scales, 92.3 ± 22.8 and 91.5 ± 23.2, respectively, P = 0.914).

5.1.5 Cytokines and cerebral palsy in term-born children (I) Altogether 15 proteins measured from cord serum were higher in full-term CP children compared with their gestation-matched paired controls (Table 10). Cord serum proteins were further compared between three groups of children born at term: term-born CP children, gestation-matched controls, and random controls. The cord serum proteins showed similar differences in term CP children and the term children in the random reference group as in the term CP children compared with their gestation-matched paired controls. IL-3, IL-13, IL-12p70, MCP-3, and MIG were higher in the term CP children than in the term controls in the random reference group, P ≤ 0.05. There was no association between the cord serum proinflammatory cytokines (IL-1α, IL-1β, IL-6, IL-8, TNF-α) and CP in the term children. It was noteworthy that the proinflammatory cytokine IL-6 was higher in the term controls than in the random reference children, P = 0.015.

66 Table 10. Cord serum proteins in preterm and term CP children and their paired controls. Cord blood protein

CP children

Control children

P value

Preterm children (n = 10) EGF*

50047 (27005)

34050 (29824)

0.002

GM-CSF

462 (877)

794 (1326)

0.010

IL-2

965 (698)

1568 (1864)

0.050

MDC

4673 (6627)

7191 (6610)

0.001

PARC

7781 (10183)

12178 (12846)

0.040

BDNF

54662 (6461)

51128 (16174)

0.054

BLC*

1876 (5046)

124 (1252)

0.028 0.021

Term children (n = 9)

CNTF*

2233 (5157)

389 (767)

EOT2

33942 (29937)

49433 (17200)

0.035

1539 (3508)

537 (1044)

0.021

IL-3

13790 (33779)

1041 (2778)

0.012

IL-4

2461 (6556)

1926 (1703)

0.040

IL-5*

1746 (4976)

150 (482)

0.034 0.009

GM-CSF

IL-12p40*

1633 (3611)

410 (1002)

IL-12p70*

11854 (36684)

3664 (3714)

0.029

IL-13*

12776 (31519)

781 (2652)

0.017

IL-15*

1963 (3601)

97 (1071)

0.016

MCP-3*

22637 (33039)

1034 (2780)

0.007

MIG*

21249 (30373)

5835 (6443)

0.012

1607 (5334)

99 (556)

0.024

TRAIL*

CP, cerebral palsy. Data are reported as median (interquartile range). The analysis was made with the paired ttest. There were 10 pairs of preterm children and 9 pairs of term children. The level of significance was set at P ≤ 0.05. *In the primary analysis of CP children (both preterm and term) and their gestation-matched pairs, the 11 analytes and MIF were higher in cases than in controls, P ≤ 0.05.

5.2 Pre- and neonatal risk factors and neurologic outcome (II) Factors describing infants´ cardiorespiratory stability after birth: high critical risk index for babies (CRIB) score, long duration of assisted ventilation, low mean intra-arterial blood pressure during the first 24 hours after birth, and postnatal dexamethasone treatment, were more common in the infants with brain damage compared to the infants without IVH grade II to III or WMD, P < 0.05. Five out of 10 children with abnormal neurologic outcome had WMD compared to only 2 of the 43 (4.6%) children with normal neurologic outcome at 2 years of corrected age, P = 0.002. Compared to the children with normal neurologic outcome, the children with abnormal neurologic outcome had higher CRIB score (6.63 ± 4.5 vs 10.3 ± 3.6), P = 0.015 and longer duration of assisted ventilation in days (18.1 ± 17.5 vs 49.8 ± 29.4), P = 0.003. Furthermore, 9 of the 10 children with abnormal neurologic outcome at 2

67 years of corrected age had received dexamethasone treatment during the neonatal period compared to 12 of the 43 (28%) children with normal neurologic outcome, P = 0.001. According to multiple logistic regression analysis, the duration of assisted ventilation in days (OR 1.06, 95%CI 1.01–1.12) together with dexamethasone treatment after birth (OR 15.07, 95%CI 1.28–177) independently and additively predicted abnormal outcome at 2 years of age. WMD (OR 12.57, 95%CI 1.31–121), and dexamethasone treatment after birth (OR 21.80, 95%CI 1.88–253) also additively predicted abnormal neurologic outcome. Table 11 presents the pre- and neonatal risk factors and their association with brain injury and neurologic outcome at the corrected age of 2 years in very preterm ELBW infants. Table 11. Pre- and neonatal factors and their association with brain injury and neurologic outcome. Factor

IVH* grade II–III

0–I

n = 11

n = 40

P value

WMD

P

Yes

No

n=7

n = 47

value

Neurologic outcome† Abnormal n = 10

P

Normal value n = 43

Prenatal Placenta HCA, n (%)

9 (82)

16 (40) 0.019

2 (29)

23 (49) 0.431

5 (50)

19 (44) 0.739

Isolated HCA,

7 (64)

12 (30) 0.075

1 (14)

18 (38) 0.400

2 (20)

16 (37) 0.464

0

14 (35) 0.023

2 (29)

14 (30)

1.0

2 (20)

14 (33) 0.487

2 (18)

4 (10)

0.598

1 (14)

5 (11)

1.0

3 (30)

2 (18)

10 (25) 0.716

3 (43)

10 (21) 0.340

n (%) Isolated perfusion defect, n (%) HCA + perfusion

3 (7)

0.073

defect, n (%) No HCA, no

3 (30)

10 (23) 0.692

perfusion defect, n (%) Gestational age at birth, wk

25.4

27.4

26.1

27.4

26.4

27.4

(27.4;

(28.3;

(28.9;

(28.1;

(28.9;

(28.1;

24.6)

26.3)

24.1)

25.9)

24.8)

26.1)

0.012

0.212

0.373

Neonatal CRIB score

10 ± 4

6±4

0.023

11 ± 2

7±5

0.008

10 ± 4

7±5

0.015

Lowest MAP on the

18 ± 6

24 ± 7

0.015

19 ± 4

23 ± 7

0.366

19 ± 5

23 ± 7

0.179

19 ± 22 0.032

22 ± 8

20 ± 3

0.001

50 ± 29

18 ± 18 0.003

11 (28) 0.037

6 (86)

15 (32) 0.011

9 (90)

12 (28) 0.001

1st day, kPa Assisted ventilation in 30 ± 19 days Steroid treatment after

7 (64)

birth, n (%) CRIB score, critical risk index for babies score; HCA, histologic chorioamnionitis; IVH, intraventricular hemorrhage; kPa, kilopascal; MAP, mean arterial pressure; WMD, white matter damage. Continuous variables are reported as median (upper quartile; lower quartile) or as mean ± SD. * Infants with IVH grade IV (n = 3) were not included in the analyses. † One child was excluded from the analyses because of a brain tumor diagnosed at 1 year of age.

68

5.3 Placental insufficiency and neurodevelopment (III) The severity of placental insufficiency was defined by Doppler ultrasonographic examination as an increase in the UA PI value. All placentas (n = 17) showed histologic evidence of impaired perfusion. No evidence of inflammation was detected in any case. Total DQ on Griffiths Scales in the suboptimal outcome group (group 1) was (mean ± SD) 91.6 ± 6.8 and in the normal outcome group (group 2) 108.3 ± 5.4, P < 0.001. There were no differences in peri- or neonatal clinical characteristics between the two groups. Preeclampsia was diagnosed in five (71%) mothers in group 1 and in seven (70%) mothers in group 2. One mother in group 1 had essential hypertension without preeclampsia. None of the four mothers without hypertensive disorder but with placental insufficiency had evidence of any other chronic disease. STV values were not different between group 1 and group 2 (7.3 ± 5.0 and 7.4 ± 2.9, respectively, P = 0.973). Also, no differences were found in gestational age at delivery, birth weight, umbilical artery blood gas values (pH, BE, PaO2,), or 5-minute Apgar scores between the two groups. Duration of mechanical ventilation and time spent in neonatal intensive care unit were identical in both groups. Increased PVE in brain US within 72 h after birth was diagnosed in 3 infants in group 1 but in none in group 2 (P = .05). Two infants in group 1 and one in group 2 had later PVL. Two infants in group 1 had convulsions after birth, whereas no neurologic symptoms were present in group 2 infants during hospitalization. The values reflecting placental and fetal hemodynamics differed between the suboptimal (group 1) and normal (group 2) neurodevelopmental outcome groups. In group 1, UA PI values were higher than in group 2 (P = 0.005). The weight-indexed RVCO, LVCO, and CCO values were significantly lower in group 1 than in group 2 (P < 0.05). Between the groups 1 and 2, no significant difference in the incidence of retrograde AoI net blood flow (29% vs 40%) was detected. Also, MCA PI values were similar in both groups. DV and IVC PIVs were higher in group 1 than in group 2 (P < 0.05) (Table 12).

69 Table 12. Parameters of fetal cardiovascular hemodynamics between the outcome groups. Factor

Group 1

Group 2

N=7

N = 10

P value

RVCO (ml/min/kg)

221.3 ± 68.4

367.0 ± 122.8

0.030

LVCO (ml/min/kg)

146.8 ± 23.6

212.0 ± 59.9

0.038

CCO (ml/min/kg)

0.022

368.1 ± 89.8

579.0 ± 166.7

RVCO%

59.4 ± 4.5

62.7 ± 6.2

0.307

LVCO%

40.6 ± 4.7

37.3 ± 6.2

0.307

UA PI

3.7 ± 0.8

2.0 ± 1.0

0.005

MCA PI

1.4 ± 0.3

1.4 ± 0.3

0.974

DAo PI

3.1 ± 0.9

2.5 ± 0.5

0.106

2 (29)

4 (40)

0.633

4.9 ± 2.49

3.6 ± 1.69

0.218

Retrograde AoI, n (%) LHV PIV DV PIV

1.3 ± 0.7

0.7 ± 0.3

0.029

IVC PIV

4.6 ± 1.6

3.1 ± 0.9

0.034

Values are mean ± SD. RVO, right ventricular output; LVO, left ventricular output; CCO, combined cardiac output; UA PI, umbilical artery pulsatile index; MCA, middle cerebral artery; DAo, descending aorta; AoI, aortic isthmus; LHV PIV, left hepatic vein pulsatility index for veins; DV, ductus venosus; IVC, inferior vena cava.

5.4 Functional neuroanatomy (IV) 5.4.1 Association between ADC and BAEP Of the specific brain regions on DWI, only pons represented a region in the auditory pathway. The correlations were therefore only calculated between ADC in pons and the different components in the BAEP measurements. There was a positive correlation between ADC measured in pons and the latency of wave III (r = 0.621, P = 0.018) and a trend toward a positive correlation between ADC measured in pons and IPL I–III (r = 0.506, P = 0.077). In stepwise linear regression analysis, 33.8% of the variability in the latency of wave III was explained by ADC in pons after adjustment for gender and gestational age at birth, P = 0.017.

5.4.2 Neuroanatomy and functional neurodevelopment The gross motor and eye-hand coordination subscales had the lowest mean values of the five subscales (Table 13). A negative correlation was found between the scores on the gross motor subscale and the ADC value in corona radiata (r = -0.401, P = 0.038). In stepwise linear regression analysis, 27.3% of the variability on the gross motor subscale

70 was explained by gestational age after adjustment for ADC in corona radiata, P = 0.003. Moreover, 14.5% of the variability in the results of the gross motor subscale was explained by ADC in corona radiata after adjustment for the gross findings on conventional MRI, including ventricular dilatation, white matter reduction, widening of sulci, irregular shape of lateral ventricle, and abnormal signal intensity, P = 0.029. The further analyses of the gross motor and eye-hand coordination subscales, on which the children had most difficulties, are presented on Table 14. The children in group 1 with scores below - 1 SD on the gross motor subscale had higher ADC values in corona radiata than the children in group 2, with scores between -1 SD and +1 SD, P = 0.008. Only one child belonged to group 3 and had gross motor scores above + 1 SD. A cut-off value of ADC 1.29 in corona radiata had a sensitivity of 75% and a specificity of 96% in finding differences on the gross motor subscale between the groups. Again, on the eyehand coordination subscale, the children in group 1 had higher ADC values in corona radiata than the children in group 2, P = 0.018. A cut-off value of ADC 1.29 in corona radiata had a sensitivity of 43% and specificity of 100% in finding differences on the eyehand coordination subscale between the groups. Table 13. Results of the subscales on Griffiths Developmental Scales in study population (N = 27). Subscale

Mean ± SD

Minimum

Maximum

Gross Motor

93.2 ± 20.6

30.0

112.5

Personal-Social

98.1 ± 15.9

42.0

124.0

Hearing and Speech

100.2 ± 16.7

56.0

131.3

Eye-hand coordination

93.6 ± 16.1

46.0

116.7

Performance

96.7 ± 26.6

26.0

143.8

SD, standard deviation. Two children out of 30 were unable to participate in the neurodevelopmental followup, and 1 child was excluded because of a brain tumor diagnosed at 1 year of corrected age. In the analysis of the eye-hand coordination subscale 1 child was excluded because of blindness.

Table 14. Differences in ADC values in corona radiata between the groups divided on the basis of the scores on Griffiths Developmental Subscales. Factor Gross Motor subscale ADC in corona radiata Eye-hand coordination subscale ADC in corona radiata

Group 1

Group 2

Group 3

N=4

N = 22

N=1

1.31 ± 0.07

1.19 ± 0.06

1.28

N=7

N = 16

N=3

1.27 ± 0.08

1.18 ± 0.06

1.22 ± 0.07

P value 0.008 0.018

ADC, apparent diffusion coefficient. ADC values are expressed as mean ± SD, unit 10-3mm2/s. To evaluate the differences in the ADC values between the three groups, one-way analysis of variance (ANOVA) with Tukey post hoc test was used.

6 Discussion 6.1 Prenatal inflammation and brain Of the homogenous population of very preterm ELBW infants, 82% survived the neonatal period. This agrees with the overall survival rate presented in the current literature (Goldenberg et al. 2000). The present study found prenatal inflammation, defined as histologic inflammation in any of the three placental compartments, to independently predict the severe forms of intracranial haemorrhage diagnosed during the first week of life. This was evident regardless of gestational age at birth, even though the infants with IVH grade II to III were more immature than the other infants in the study. This finding confirms some earlier findings (Salafia et al. 1995, Disalvo 1998, Dexter et al. 2000, Yanowitz et al. 2002) and emphasizes the significance of the intrauterine environment for brain envolvement. However, we failed to confirm the association between the acute-phase proinflammatory cytokines in cord blood or HCA and intraparenchymal WMD as previously proposed (Bejar et al. 1988, Leviton et al. 1999, Wu & Colford, Jr. 2000, Duggan et al. 2001, Wu 2002). This challenges the hypothesis on the direct relationship between the fetal inflammatory response and its detrimental effect on the immature brain (Dammann & Leviton 1997a). We used standard criteria to diagnose both PV-IVH (Papile et al. 1978) and cPVL (de Vries et al. 1992) on US but decided to analyze IVH grade IV together with cPVL since both are well defined white matter lesions that share the same contributing factor in their pathogenesis, ischemia-reperfusion, and account for the majority of severe WMD leading to long-term neurologic sequelae among preterm infants (Perlman 1998). The final assessment of WMD was based on conventional MRI performed at term. Consistent with previous reports, the present study found US to be comparable to conventional MRI, and it detected cPVL (Inder et al. 2003b) with a sensitivity of 100% and a specificity of 98% and IVH grade IV (Maalouf et al. 2001) with a sensitivity of 67% and a specificity of 100%. The overall incidence of IVH in our study was 54%. This exceeds the incidence of PV-IVH usually reported (Heuchan et al. 2002). The study protocol we obtained enabled us to assess the brain US prospectively from day 1 from birth until term age and thus to

72 detect every occurrence of even small cerebral hemorrhage. This may explain the difference between the figures in the present and previous studies. On the other hand, IVH grade IV was diagnosed in 5.5% and cPVL in 7.4% of very preterm ELBW infants which is in line with the incidence figures published recently (Lemons et al. 2001, Blumenthal 2004). Although the levels of proinflammatory cytokines in cord serum per se did not differ between the infants with IVH grade II to III and the other infants in the study, this does not rule out the possibility that inflammatory mediators at some point during pregnancy affected endothelial cells in the germinal matrix, leading to upregulation of chemokines and adhesion molecules and recruitment of white blood cells, resulting in damage of BBB (Anthony et al. 1998, Stanimirovic & Satoh 2000). Several factors are likely to influence the levels of acute phase cytokines in umbilical cord blood. The kinetic analysis of proinflammatory cytokines in vitro has revealed that, after LPS stimulation, TNF-α reached a peak level at 4 hours, whereas IL-1β and IL-6 production peaked at approximately 16 hours. The levels of all the 3 cytokines were considerably decreased by 48 hours after stimulation (Lee et al. 1993a). The fact that each individual cytokine is expressed within a certain time window emphasizes the timing of sample drawing. Consistent with the latter, the time point of stimulation of the immune system in vivo seems to associate with the cytokine levels detected. Infants with early-onset sepsis (≤ 96 hours after birth) had high plasma levels of the acute-phase proteins IL-6 and IL-8 in the first blood sample drawn within 24 hours after birth, with the values declining significantly during ≥ 48 hours after birth (Berner et al. 1998). This is consistent with the present finding, according to which the umbilical cord blood level of IL-6 was higher in the term control children who needed intensive care treatment after birth but did not develop CP afterwards compared to their healthy random controls. Carefully defined criteria of both CCA (Gibbs et al. 1982) and HCA (Benirschke & Kaufmann 2000) were used, and the systemic fetal response to inflammation was separately assessed. This enabled an accurate analysis of the inflammatory effects on brain lesions and the neurologic outcome, also carefully detected. The exact criteria for the diagnosis of chorioamnionitis are not always mentioned in the literature (Murphy et al. 1995, Grether et al. 1996). Consistent with the previous findings (Gravett et al. 1994, Yoon et al. 1995), we found a high level of cord serum IL-6 to be a good predictor of active preterm labor and high levels of IL-6 and IL-8 in cord serum to associate with isolated HCA (Shimoya et al. 1992, Yanowitz et al. 2002) after adjustment for gestational age at birth. In 46% of the mothers, pregnancy was complicated by HCA. Only two mothers had clinical signs of CA, but they were not adequate to predict surrogate outcomes or abnormal neurologic outcome. However, both mothers had HCA. The one mother with twin pregnancy had group B streptococci septicemia on admission into hospital, and her A fetus was stillborn. The other mother had severe chorionic plate inflammation and funisitis. Both of the infants were diagnosed for neurologic sequelae later on. One child had abnormal neuromotor function of both legs, and the other was diagnosed to have WMD and severe neurodevelopmental disability on Griffiths Scales at 2 years of age. Invasive pathogens are likely to affect the outcome of children born from pregnancies complicated by HCA (Berner et al. 2001). In a case-control study of placental and clinical risk factors associated with neurologic impairment at 20 months of corrected age, severe

73 fetal inflammation, i.e. abundant neutrophils in chorionic plate vessels, had a high level of concordance with CCA, although neither CCA or HCA independently predicted neurologic impairment (Redline et al. 2000). In the present study, some 19% of the very preterm ELBW infants were defined to have abnormal neurologic outcome at the corrected age of 2 years. High levels of cord serum proinflammatory cytokines and HCA were not found to be more prevalent among those with neuromotor impairment than those with normal outcome. Also, we found no association between prenatal inflammation and impaired neurodevelopmental outcome scores among very preterm infants, evaluated both by Griffiths Developmental Scales and by Bayley Scales at 2 years of corrected age. Furthermore, among the pregnancies complicated by placental insufficiency, no differences in cord serum proinflammatory cytokines (IL-1α, IL-1β, IL-6, IL-8, TNF-α) were found between the infants with suboptimal and normal outcome at 1 year of corrected age. These findings are consistent with the recent observations (Dexter et al. 2000, Redline et al. 2000, Gray et al. 2001, Grether et al. 2003, Nelson et al. 2003). However, our study demonstrated that preterm infants born from pregnancies complicated by both HCA and placental perfusion defect were at an increased risk of abnormal neurologic outcome at 2 years of corrected age. Also, compound placental defect together with WMD diagnosed during hospitalization after birth additively predicted poor outcome. Two recently published reports have underlined the importance of non-inflammatory placental pathology in predicting brain damage and later neurologic sequelae (Redline et al. 2000, Kumazaki et al. 2002). In a study of infants delivered before 34 weeks of gestation, 41.7% of those having PVL were born from pregnancies complicated by disturbed uteroplacental circulation manifesting as gross findings of the placenta compared to 13.7% of control children without PVL. Both the gross lesions and the histologic findings of ischemic changes in villi remained significant in predicting PVL after adjustment for confounding factors (OR 4.04, 95%CI 1.40–11.67 and OR 7.28, 95%CI 2.50–21.20, respectively). Fifty-eight percent of the placentas with HCA also had disturbed uteroplacental circulation (Kumazaki et al. 2002). In another study of VLBW infants, multiple placental lesions were independently related to neurologic impairment assessed at 20 months of corrected age (OR 13.2, 95%CI 1.3–137) (Redline et al. 2000). In the present study, 3 mothers with a combined placental defect, HCA and impaired placental perfusion, had preeclampsia. Animal experiments have produced evidence for infection as a possible cause of preeclampsia. In pregnant rats, a single dose of endotoxin triggered the kind of clinical symptoms and histopathologic renal findings commonly related to the predominant features of human preeclampsia (Faas et al. 1994). It is also noteworthy that intra-amniotic administration of endotoxin to pregnant mice has been shown to lead to the kind of hemodynamic changes seen in severe placental insufficiency, i.e. increase in UA PI and in DV PIV and a decrease in fetal cardiac output (Ferrazzi et al. 2002, Rounioja et al. 2003). Since both intrauterine inflammation and placental perfusion insufficiency seem to relate to fetal cardiac function, their synergistic effect might potentially affect cerebral blood flow, decreasing perfusion in the distal fields of the watershed area in white matter, thus resulting in WMD and subsequent disabilities in neurologic outcome. Only 1 out of 6 infants with compound placental defects had normal brain US and conventional MRI at term. Abnormal neurologic outcome was detected in 3 of the 6 children at 2 years of corrected age.

74

6.2 Cytokine profiles in term and preterm CP children We found a cluster of protein mediators, different from the conventional acute phase proteins, to be related with CP. This is a novel finding and differs from the results reported in the previous literature, where the role of the acute-phase response is emphasized (Nelson et al. 1998, Yoon et al. 2000). In term and preterm children with CP diagnosed by the chronological age of 5 years, the umbilical cord serum levels of 8 cytokines (CNTF, IL-5, IL-12p40, IL-12p70, IL-13, IL-15, MIF, TRAIL), growth factor EGF, and 3 chemokines (BLC, MCP-3, MIG) were higher compared to their paired gestation-matched controls. Furthermore, we found the profile of protein mediators in umbilical cord serum to differ between term and preterm CP children. Several factors may explain the differences between the present and previous studies. We used umbilical cord serum, unlike most studies that have evaluated the differences in protein mediators between CP and non-CP children by measuring cytokine levels either from amniotic fluid (Yoon et al. 1997a, Yoon et al. 2000) or by using peripheral blood samples withdrawn a few days after birth (Nelson et al. 1998). Furthermore, the clinical definitions used for studies must be precise. In a recently published report, the levels of cord blood cytokines (IL-1β, IL-6, TNF-α, IL-18) were compared between children without exact definitions of CP (Minagawa et al. 2002). This does not enable a reliable analysis of the study population. In addition to the source of samples, gestational age may influence the cytokine profile analyzed. We used matched and paired controls to take gestational age into account. In a recent study, altogether 16 analytes were measured from peripheral blood samples of infants born ≤ 28 weeks of pregnancy and compared to measurements from infants born at term. The medians of some protein mediators, e.g. IL-1, IL-8, and TNF-α, were reported to be higher in the preterm compared to the term infants, although no statistical analysis was performed (Dammann et al. 2001b). In the present study, the umbilical cord serum levels of 4 proteins (IL-12p70, IL-13, MCP-3, MIG) correlated with the length of gestation among the CP cases, and the cord serum levels of 11 proteins (BDNF, LIF, MDC, PARC, TGF-β1, TNF-α, IL-6, IL-8, IL-1sR1, IL-11, uPAR) correlated with the length of gestation among both the CP cases and their gestation-matched controls. Based on the present study, we propose that high levels of certain cytokines measured in CP children indicate prenatal immune activation, and that gestational age may influence the levels of some protein mediators, while others may be high regardless of gestational age. Some of the protein mediators analyzed in the present study may be directly involved in neural injury, some may be involved in the repair of such injury, while yet some others could be coincidental. Cytokines have several protective or trophic effects, and their adverse effects may depend on interactions with constitutional or environmental factors (Dammann & Leviton 1999, Harding et al. 2004). CNTF present in high concentrations in the cord blood of term infants developing CP could reflect a cytoprotective role or a compensatory mechanism as a result of cytokine-mediated injury in brain (Louis et al. 1993, Albrecht et al. 2002). The expression of BDNF relates to cerebral maturation in progress, and overexpression affects the distortion of the cortical architecture (Ringstedt et al. 1998). EGF was high in preterm infants who developed CP. EGF receptor has been detected from various tissues in human fetuses (Oliver 1988). It has a neurotrophic effect

75 on newborn brain. EGF stimulates astrocyte proliferation and differentiation and establishes cellular protection against glutamate by inducing the expression of glial glutamate transporter (GLT-1) in astrocytes (Zelenaia et al. 2000). EGF could thus participate in a protective cascade triggered by a process leading to brain damage in premature infants. Previously, EGF has been found to increase during gestation in umbilical cord serum (Scott et al. 1989). We found EGF to correlate with gestational age in our control children but not in the children with CP, which emphasizes its role in the process that eventually leads to brain damage. The elevated levels of the cytokines IL-5, IL-12, IL-13, IL-15, and the chemokines MCP3 and MIG in the cord serum of full-term infants developing CP demonstrate that inflammatory cells in newborns are capable of secreting protein mediators in response to inflammatory stimuli in vivo. IL-12 is secreted by mononuclear cells in response to especially invasive gram-positive pathogens (Berner et al. 2001). IL-12 links up innate and adaptive immune responses. It activates T cells and natural killer (NK) cells to secrete IFN-γ, which has been shown induce a toxic effect on immature OLs in vitro (Baerwald & Popko 1998, Buntinx et al. 2004) and to associate with high levels of other inflammatory mediators in term or near-term infants with CP (Grether et al. 1999). IL-15 secreted by Th1 cells and IL-5 and IL-13 secreted by Th2 cells also show involvement of the adaptive immune response during the prenatal period among term CP children. The latter is complementary to the previous report of protein mediators detected from peripheral blood after birth in term or near-term infants later developing CP (Nelson et al. 1998). The MCP family of CC chemokines has been found to be involved in the development of multiple sclerosis (MS) lesions in brain (McManus et al. 1998). Elevated levels of MCP-3 have also been found in CSF in humans with acute bacterial meningitis (Pashenkov et al. 2002). Nelson and colleagues found members of the MCP family to have a good sensitivity and specificity to identify children with CP from control children (Nelson et al. 1998). The CXC chemokine MIG is induced by IFN-γ and has been shown to control the chemotaxis of T lymphocytes in the central nervous system during viral infections in rodents (Liu et al. 2001), and it is also implicated in the pathogenesis of MS (Simpson et al. 2000). The present study related high levels of MIG in umbilical cord serum to CP in children for the first time. The present study found no specific increase in the several proinflammatory mediators previously described to associate with CP in either preterm or term infants. This is consistent with the recent reports (Grether et al. 2003, Nelson et al. 2003) but contradicts one previous report (Nelson et al. 1998). The umbilical cord serum levels of IL-6 and IL8 were generally high in preterm infants. IL-6 was high in term gestation controls, who differed from the healthy reference group in that they developed an acute neonatal disease and were treated in a neonatal intensive care unit. Various factors reflecting infants´ cardiorespiratory stability after birth were found to associate independently and additively with abnormal neurologic outcome at 2 years of corrected age. Our study supports the opinion that the process leading to CP is likely to involve a complex cascade of potential factors starting before birth (Blair & Stanley 2002), and we propose that infants predisposed to perinatal distress are susceptible to neural damage during compromised neonatal adaptation.

76

6.3 Placental insufficiency and neurodevelopmental outcome Our study is the first to demonstrate that fetal cardiac compromise, defined as decreased weight-indexed cardiac outputs and increased systemic venous pressure, associates with suboptimal neurodevelopmental outcome assessed at 1 year of corrected age in infants born from pregnancies complicated by placental insufficiency and delivery before 32 weeks of gestation. Previously, the effect of absent or reversed end-diastolic velocity waveforms in UA or in DAo on later neurologic or neurodevelopmental compromise have been focused on (Ley et al. 1996a, Ley et al. 1996b, Vossbeck et al. 2001, Kutschera et al. 2002). Abnormal end-diastolic blood velocity waveform on Doppler ultrasonography relates to a rise in UA PI and in DAo PI, which in turn relates to a reduction of vascular resistance in fetal cerebral circulation, called the ´brain-sparing effect´, in the prescence of fetal hypoxia (Wladimiroff et al. 1986). We found no differences in MCA PI between infants with suboptimal (group 1) and normal (group 2) neurodevelopmental outcome, although the infants in group 1 had higher UA PI, indicating higher downstream vascular impedance and higher placental resistance and thus more severe placental insufficiency, than the infants in group 2. This suggests that the maximal cerebrovascular compensation for hypoxemia had already been reached in both groups. Longitudinal studies of pregnancies complicated by placental insufficiency have shown that both the increase in UA PI and the decrease in PI in fetal cerebral circulation are early changes that can manifest weeks before delivery and do not necessarily predict the perinatal outcome (Hecher et al. 2001, Ferrazzi et al. 2002). In compromised fetuses, MCA PI has even shown a trend toward normalization close to delivery (Hecher et al. 2001). In severe placental insufficiency, hypoxemia induces increased shunting through FO, and cardiac output begins to favor the left ventricle (al Ghazali et al. 1989). Oxygen content in fetal blood has been shown to correlate directly with LVCO (Ferrazzi et al. 1995). We found diminished weight-indexed cardiac output of both ventricles within 24 hours before birth in infants with suboptimal neurodevelopmental outcome but no differences in umbilical artery pH or partial pressure of arterial oxygen (PaO2) between infants with suboptimal and normal outcome. This suggests that hypoxemia itself cannot explain the reduction of weight-indexed cardiac outputs. In normal pregnancies, the volume of blood flow through FO (QFO) correlates with increasing gestational age (Räsänen et al. 1996). In a study of singleton pregnancies complicated by IUGR and placental insufficiency, fetuses < 32 weeks of gestation had a smaller size of FO compared to normally grown fetuses at the same age of gestation, P = 0.003 (Kiserud et al. 2004). Furthermore, in the present study, infants with suboptimal outcome did not have higher levels of umbilical cord serum proinflammatory cytokines compared to infants with normal outcome, suggesting that the cardiac compromise was not related to the fetal inflammatory response. We propose that the fetuses with suboptimal neurodevelopmental outcome were unable to shift their cardiac output in favor of the left ventricle since there were no differences in the proportions (%) of RVCO or LVCO of CCO between the infants with suboptimal and normal outcomes. Our study is consistent with the recent report demonstrating that decreased peak systolic velocity in the

77 ascending aorta, reflecting diminished LVCO, was the late stage change in haemodynamics in IUGR fetuses before delivery (Ferrazzi et al. 2002). Fetuses with significant myocardial cell damage have been shown to present signs of increased systemic venous pressure defined as increased PI in venous circulation (Mäkikallio et al. 2002). In the present study, infants with suboptimal neurodevelopmental outcome had higher DV PIV and IVC PIV compared to infants with normal outcome. Our finding of fetal cardiac compromise, defined as a decrease of weight-indexed cardiac outputs and an increase of systemic venous pressure, associates with later outcome and supports the preliminary finding of a recent report (Ferrazzi et al. 2002). Very preterm infants with reversed flow in DV and reduced peak systolic velocities in both the pulmonary artery, reflecting low RVCO, and the ascending aorta, reflecting low LVCO, had evidence of brain injury after birth and neurologic deficit at 6 months of age (Ferrazzi et al. 2002). Increased PVE detected by US within 72 hours after birth was prevalent in the present infants with suboptimal outcome (43%), but it was not found in infants with normal outcome. Three infants later developed PVL, 2 of them in the suboptimal outcome group with PVE. The one infant with severe neuromotor dysfunction at 1 year of corrected age was born from a pregnancy complicated by IUGR and severe fetal cardiac compromise. After birth, he was defined to have cPVL on US. We propose that, in pregnancies complicated by placental insufficiency, diminished fetal cardiac output may affect cerebral perfusion and thus give rise to circulatory changes in poorly vascularized periventricular white matter in immature brain. A similar phenomenon has been proposed in fetuses with a single ventricular chamber restricting the compensatory increase of cardiac output for sufficient blood flow in cerebral circulation (Donofrio et al. 2003). Since endothelial cell dysfunction has been proposed to underline maternal preeclampsia with a reduction in placental perfusion, we analyzed the umbilical cord serum levels of two adhesion molecules, ICAM-1 and ICAM-3, and growth factor VEGF and its receptor VEGFR2, known to be induced by hypoxia and to be involved in endothelial cell activation (Wang et al. 2002, Bartha et al. 2003). We did not find the levels of ICAM or VEGF in the fetal compartment to reflect the severity of placental compromise. Also, the levels of umbilical cord serum proinflammatory cytokines (IL-1, IL-6, IL-8, TNF-α) did not differ between placentas with impaired perfusion and normal histology. These findings differ from the previous reports suggesting that placental insufficiency is accompanied by elevated levels of inflammatory mediators in cord blood (Trudinger et al. 2002, Wang et al. 2002). In both previous studies, however, infants with placental insufficiency and elevated levels of protein mediators had significantly lower gestational age than control infants born from normal pregnancies. In the present study, on the other hand, gestational age at delivery was strictly controlled.

6.4 DWI and correlation with BAEP and neurodevelopmental outcome We found the ADC value in pons to correlate with the measurement of a BAEP component, the latency of wave III, both measured at term. This was confirmed with a

78 regression model after adjustment for gender and gestational age at birth. Furthermore, a trend toward a positive correlation between ADC and IPL I-III was detected. This relation between molecular displacement of water on DWI and conduction of axonal impulses through the auditory tract is reported for the first time in very preterm ELBW infants, and it integrates microstructural information detected on DWI and neurophysiological information measured by BAEP. Myelination is crucial for the conduction of impulses through both motor and sensory nerve fibers. The absolute peak latencies and interpeak latencies of auditory waves represent conduction of speed along the auditory pathway in the central nervous system. Previously, prolonged latencies of the waves I and II on BAEP components have been related to a congenital hypomyelination disorder (Ono et al. 1994). Since the white matter tracts in the brainstem are known to be myelinated at term even in preterm infants and since, in concordance with brain maturation, as the diffusion of water becomes restricted, the ADC value declines (Hüppi et al. 1998, Morriss et al. 1999, Melhem 2002), we propose that the ADC value measured in pons relates to the myelination and organization of the fiber architecture in the auditory tract. On the other hand, ADC values have been shown to fail to decline in widespread areas of white matter damage detected originally on conventional MRI (Miller et al. 2002). Abnormalities in BAEP measurements may predict neurosensory disabilities among preterm infants (Valkama et al. 2001) and suggest involvement in the central nervous system (Markand 1994). In the present study, only one infant was diagnosed to have cPVL on conventional MRI at term. Ventricular dilatation was seen in 16 infants, white matter reduction in 21 infants, widening of sulci in 17, irregular shape of the lateral ventricles in 9, and abnormal signal intensity in 2 infants. No association was found between the gross findings on MRI and the different BAEP components measured. The latter is in concordance with a previous study showing no correlation between BAEP measurements and cerebral findings on conventional MRI at the age of 8 years in children born at a mean gestational age of 31 wk (Olsén et al. 2002). We found no correlation between ADC in pons and the other components of BAEP measurements. It is possible that the ROIs in pons studied for ADC value may not have matched neuroanatomically some other regions in the auditory pathway. As a rule, major neurologic abnormalities, e.g. CP, correlate with gross findings on brain US or conventional MRI (Leviton & Paneth 1990, Valkama et al. 2000, Mirmiran et al. 2004). Apart from the major disabilities, the current literature contains scant information about more subtle impairments, such as minor motor dysfunction (Lorenz 2001), and their relation with brain envolvement. Yet, 25% to 50% of the infants with birth weight < 1500 g continue to have less prominent neuromotor or neurodevelopmental disabilities later on (Volpe 1997). Corona radiata represents an area in periventricular white matter through which the corticospinal motor fibers traverse on their way to the cerebral cortex. We found a negative correlation between ADC in corona radiata and scores on the gross motor subscale on the Griffiths Scales assessed at 2 years of corrected age. This is a novel finding and suggests that a high ADC value in corona radiata reveals a more diffuse injury in central white matter not evident on conventional MRI, and that this could influence later functional neurodevelopmental abilities in very preterm ELBW infants. None of the gross findings on conventional MRI associated with any of the child´s

79 abilities on the 5 Griffiths subscales. Furthermore, after the children had been divided into 3 groups and compared according to the scores on the gross motor and eye-hand coordination subscales, the groups with the lowest scores had the highest ADC in corona radiata. Eye-hand coordination implies integration of motor functions in the upper limbs and visual activation. Preterm children without any major neurologic disability have shown significant weakness in their eye-hand coordination skills (Bowen et al. 1993). Our population was homogenous and comparable in size to the previously published series (Hüppi et al. 1998, Morriss et al. 1999, Miller et al. 2002). The ADC values measured correspond to those reported in the current literature (Engelbrecht et al. 2002, Bozzao et al. 2003). The associations between ADC value and neurodevelopment were constant. We propose that the quantitative measurement of water diffusion, ADC, could widen the indications of DWI in clinical use and serve as an early diagnostic tool to select the very high-risk infants who require more specific neurologic follow-up.

7 Conclusions 1. In a study population of ELBW infants born before 32 weeks of gestation, high levels of acute-phase proinflammatory cytokines, IL-6 and IL-8, measured in umbilical cord serum related to isolated histologic chorioamnionitis and high levels of IL-6 in umbilical cord serum predicted active preterm labor. Infants born to mothers with isolated histologic chorioamnionitis had lower gestational age at birth than infants born from other pregnancies. Several proinflammatory cytokines in cord serum associated negatively with the length of gestation. 2. Neither chorioamnionitis nor isolated fetal inflammatory response, defined as high levels of acute phase proinflammatory cytokines in umbilical cord serum, served as a risk factor of WMD or abnormal neurologic or neurodevelopmental outcome in very preterm children. Instead, clinical factors describing cardiorespiratory failure after birth associated with IVH grade II to III, WMD, and with abnormal neurologic outcome in very preterm ELBW infants. 3. Prenatal inflammation defined as histologic chorioamnionitis in any placental compartments independently predicted IVH grade II to III detected by the age of 1 week in very preterm ELBW infants. This suggests periventricular endothelial injury as a response to intrauterine inflammation. 4. Pregnancies complicated by compound placental defect, both histologic chorioamnionitis and placental perfusion defect, independently predicted abnormal neurologic outcome assessed at the corrected age of 2 years in very preterm ELBW infants. Focal WMD additively predicted poor neurologic outcome. 5. A cluster of protein mediators in umbilical cord serum, different from acute-phase proinflammatory cytokines, was related with CP. The protein profile in umbilical cord serum differed between term and preterm children with CP. 6. Fetal cardiac compromise defined as decreased weight-indexed cardiac outputs and increased systemic venous pressure associated with suboptimal neurodevelopmental

81 outcome at 1 year of corrected age in infants born from pregnancies complicated by placental insufficiency and delivery before 32 weeks of gestation. 7. Apparent diffusion coefficient, ADC, on DWI in pons had a positive correlation with a BAEP component, the latency of wave III, both measured at term. ADC in corona radiata had a negative correlation with scores on the gross motor subscale on the Griffiths Scales assessed at 2 years of corrected age. These findings integrate both neurofunctional and neurodevelopmental measurements with quantitative findings on DWI and extend the current knowledge of the importance of DWI in predicting the outcome of high-risk very preterm ELBW infants who would benefit from later neurologic follow-up.

References Abbott NJ (2002) Astrocyte-endothelial interactions and blood-brain barrier permeability. J Anat 200: 629-638. Adcock LM, Moore PJ, Schlesinger AE & Armstrong DL (1998) Correlation of ultrasound with postmortem neuropathologic studies in neonates. Pediatr Neurol 19: 263-271. Ajayi-Obe M, Saeed N, Cowan FM, Rutherford MA & Edwards AD (2000) Reduced development of cerebral cortex in extremely preterm infants. Lancet 356: 1162-1163. al Ghazali W, Chita SK, Chapman MG & Allan LD (1989) Evidence of redistribution of cardiac output in asymmetrical growth retardation. Br J Obstet Gynaecol 96: 697-704. Albrecht PJ, Dahl JP, Stoltzfus OK, Levenson R & Levison SW (2002) Ciliary neurotrophic factor activates spinal cord astrocytes, stimulating their production and release of fibroblast growth factor-2, to increase motor neuron survival. Exp Neurol 173: 46-62. Alexander JM, Gilstrap LC, Cox SM, McIntire DM & Leveno KJ (1998) Clinical chorioamnionitis and the prognosis for very low birth weight infants. Obstet Gynecol 91: 725-729. Allan WC, Vohr B, Makuch RW, Katz KH & Ment L (1997) Antecedents of cerebral palsy in a multicenter trial of indomethacin for intraventricular hemorrhage. Arch Pediatr Adolesc Med 151: 580-585. Allen MC, Donohue PK & Dusman AE (1993) The limit of viability--neonatal outcome of infants born at 22 to 25 weeks' gestation. N Engl J Med 329: 1597-1601. Altman DI, Powers WJ, Perlman JM, Herscovitch P, Volpe SL & Volpe JJ (1988) Cerebral blood flow requirement for brain viability in newborn infants is lower than in adults. Ann Neurol 24: 218-226. Amory JH, Adams KM, Lin MT, Hansen JA, Eschenbach DA & Hitti J (2004) Adverse outcomes after preterm labor are associated with tumor necrosis factor-alpha polymorphism -863, but not 308, in mother-infant pairs. Am J Obstet Gynecol 191: 1362-1367. Anstrom JA, Brown WR, Moody DM, Thore CR, Challa VR & Block SM (2004) Subependymal veins in premature neonates: implications for hemorrhage. Pediatr Neurol 30: 46-53. Anthony D, Dempster R, Fearn S, Clements J, Wells G, Perry VH & Walker K (1998) CXC chemokines generate age-related increases in neutrophil-mediated brain inflammation and blood-brain barrier breakdown. Curr Biol 8: 923-926. Asrat T (2001) Intra-amniotic infection in patients with preterm prelabor rupture of membranes. Pathophysiology, detection, and management. Clin Perinatol 28: 735-751. Back SA, Gan X, Li Y, Rosenberg PA & Volpe JJ (1998) Maturation-dependent vulnerability of oligodendrocytes to oxidative stress-induced death caused by glutathione depletion. J Neurosci 18: 6241-6253.

83 Back SA, Luo NL, Borenstein NS, Levine JM, Volpe JJ & Kinney HC (2001) Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury. J Neurosci 21: 1302-1312. Bada HS, Korones SB, Perry EH, Arheart KL, Ray JD, Pourcyrous M, Magill HL, Runyan W, III, Somes GW, Clark FC & . (1990) Mean arterial blood pressure changes in premature infants and those at risk for intraventricular hemorrhage. J Pediatr 117: 607-614. Baerwald KD & Popko B (1998) Developing and mature oligodendrocytes respond differently to the immune cytokine interferon-gamma. J Neurosci Res 52: 230-239. Bagenholm R, Nilsson UA, Gotborg CW & Kjellmer I (1998) Free radicals are formed in the brain of fetal sheep during reperfusion after cerebral ischemia. Pediatr Res 43: 271-275. Baggiolini M (1998) Chemokines and leukocyte traffic. Nature 392: 565-568. Banker BQ & Larroche JC (1962) Periventricular leukomalacia of infancy. A form of neonatal anoxic encephalopathy. Arch Neurol 7: 386-410. Barkovich AJ, Kjos BO, Jackson DE, Jr. & Norman D (1988) Normal maturation of the neonatal and infant brain: MR imaging at 1.5 T. Radiology 166: 173-180. Barlow CF, Priebe CJ, Mulliken JB, Barnes PD, Mac DD, Folkman J & Ezekowitz RA (1998) Spastic diplegia as a complication of interferon Alfa-2a treatment of hemangiomas of infancy. J Pediatr 132: 527-530. Bartha JL, Romero-Carmona R, Escobar-Llompart M, Paloma-Castro O & Comino-Delgado R (2003) Human chorionic gonadotropin and vascular endothelial growth factor in normal and complicated pregnancies. Obstet Gynecol 102: 995-999. Baschat AA, Gembruch U, Reiss I, Gortner L, Weiner CP & Harman CR (2000) Relationship between arterial and venous Doppler and perinatal outcome in fetal growth restriction. Ultrasound Obstet Gynecol 16: 407-413. Basser PJ (1995) Inferring microstructural features and the physiological state of tissues from diffusion-weighted images. NMR Biomed 8: 333-344. Battin M, Maalouf EF, Counsell S, Herilhy AH & Edwards AD (1997) Magnetic resonance imaging of the brain of premature infants. Lancet 349: 1741. Battin MR, Maalouf EF, Counsell SJ, Herlihy AH, Rutherford MA, Azzopardi D & Edwards AD (1998) Magnetic resonance imaging of the brain in very preterm infants: visualization of the germinal matrix, early myelination, and cortical folding. Pediatrics 101: 957-962. Baud O, Emilie D, Pelletier E, Lacaze-Masmonteil T, Zupan V, Fernandez H, Dehan M, Frydman R & Ville Y (1999a) Amniotic fluid concentrations of interleukin-1beta, interleukin-6 and TNFalpha in chorioamnionitis before 32 weeks of gestation: histological associations and neonatal outcome. Br J Obstet Gynaecol 106: 72-77. Baud O, Foix-L'Helias L, Kaminski M, Audibert F, Jarreau PH, Papiernik E, Huon C, Lepercq J, Dehan M & Lacaze-Masmonteil T (1999b) Antenatal glucocorticoid treatment and cystic periventricular leukomalacia in very premature infants. N Engl J Med 341: 1190-1196. Baud O, Li J, Zhang Y, Neve RL, Volpe JJ & Rosenberg PA (2004) Nitric oxide-induced cell death in developing oligodendrocytes is associated with mitochondrial dysfunction and apoptosisinducing factor translocation. Eur J Neurosci 20: 1713-1726. Bayley N (1993) Bayley scales of infant development. The psychological corporation, Harcourt Brace and Co, San Antonio. Bejar R, Wozniak P, Allard M, Benirschke K, Vaucher Y, Coen R, Berry C, Schragg P, Villegas I & Resnik R (1988) Antenatal origin of neurologic damage in newborn infants. I. Preterm infants. Am J Obstet Gynecol 159: 357-363. Bekedam DJ, Visser GH, van der Zee AG, Snijders RJ & Poelmann-Weesjes G (1990) Abnormal velocity waveforms of the umbilical artery in growth retarded fetuses: relationship to antepartum late heart rate decelerations and outcome. Early Hum Dev 24: 79-89.

84 Benirschke K, Kauffmann P (2000) Pathology of the human placenta. Springer-Verlag, New York, 16-28, 425, 453-456, 591-608. Berner R, Niemeyer CM, Leititis JU, Funke A, Schwab C, Rau U, Richter K, Tawfeek MS, Clad A & Brandis M (1998) Plasma levels and gene expression of granulocyte colony-stimulating factor, tumor necrosis factor-alpha, interleukin (IL)-1beta, IL-6, IL-8, and soluble intercellular adhesion molecule-1 in neonatal early onset sepsis. Pediatr Res 44: 469-477. Berner R, Csorba J & Brandis M (2001) Different cytokine expression in cord blood mononuclear cells after stimulation with neonatal sepsis or colonizing strains of Streptococcus agalactiae. Pediatr Res 49: 691-697. Bhutta AT, Cleves MA, Casey PH, Cradock MM & Anand KJ (2002) Cognitive and behavioral outcomes of school-aged children who were born preterm: a meta-analysis. JAMA 288: 728737. Block BS, Schlafer DH, Wentworth RA, Kreitzer LA & Nathanielsz PW (1990) Intrauterine asphyxia and the breakdown of physiologic circulatory compensation in fetal sheep. Am J Obstet Gynecol 162: 1325-1331. Blumenthal I (2004) Periventricular leucomalacia: a review. Eur J Pediatr 163: 435-442. Bobitt JR & Ledger WJ (1977) Unrecognized amnionitis and prematurity: a preliminary report. J Reprod Med 19: 8-12. Borch K & Greisen G (1998) Blood flow distribution in the normal human preterm brain. Pediatr Res 43: 28-33. Bourgeois JP (2002) Synaptogenesis in the neocortex of the newborn: the ultimate frontier for individuation. In: Lagercrantz H, Hanson M, Evrard P & Rodeck C (eds) The newborn brain. Cambridge University Press, Cambridge, 91-113. Bowen JR, Starte DR, Arnold JD, Simmons JL, Ma PJ & Leslie GI (1993) Extremely low birthweight infants at 3 years: a developmental profile. J Paediatr Child Health 29: 276-281. Bozzao A, Di Paolo A, Mazzoleni C, Fasoli F, Simonetti A, Fantozzi LM & Floris R (2003) Diffusion-weighted MR imaging in the early diagnosis of periventricular leukomalacia. Eur Radiol 13: 1571-1576. Brodszki J, Hernandez-Andrade E, Gudmundsson S, Dubiel M, Mandruzzato GP, Laurini R & Marsal K (2002) Can the degree of retrograde diastolic flow in abnormal umbilical artery flow velocity waveforms predict pregnancy outcome? Ultrasound Obstet Gynecol 19: 229-234. Brunstrom JE, Gray-Swain MR, Osborne PA & Pearlman AL (1997) Neuronal heterotopias in the developing cerebral cortex produced by neurotrophin-4. Neuron 18: 505-517. Bry K, Hallman M, Teramo K, Waffarn F & Lappalainen U (1997) Granulocyte-macrophage colony-stimulating factor in amniotic fluid and in airway specimens of newborn infants. Pediatr Res 41: 105-109. Buntinx M, Gielen E, Van Hummelen P, Raus J, Ameloot M, Steels P & Stinissen P (2004) Cytokine-induced cell death in human oligodendroglial cell lines. II: Alterations in gene expression induced by interferon-gamma and tumor necrosis factor-alpha. J Neurosci Res 76: 846-861. Cammer W & Zhang H (1999) Maturation of oligodendrocytes is more sensitive to TNF alpha than is survival of precursors and immature oligodendrocytes. J Neuroimmunol 97: 37-42. Casey ML, Cox SM, Beutler B, Milewich L & MacDonald PC (1989) Cachectin/tumor necrosis factor-alpha formation in human decidua. Potential role of cytokines in infection-induced preterm labor. J Clin Invest 83: 430-436. Chi JG, Dooling EC & Gilles FH (1977) Gyral development of the human brain. Ann Neurol 1: 8693. Chung IY & Benveniste EN (1990) Tumor necrosis factor-alpha production by astrocytes. Induction by lipopolysaccharide, IFN-gamma, and IL-1 beta. J Immunol 144: 2999-3007.

85 Colton CA & Gilbert DL (1993) Microglia, an in vivo source of reactive oxygen species in the brain. Adv Neurol 59: 321-326. Conroy SM, Nguyen V, Quina LA, Blakely-Gonzales P, Ur C, Netzeband JG, Prieto AL & Gruol DL (2004) Interleukin-6 produces neuronal loss in developing cerebellar granule neuron cultures. J Neuroimmunol 155: 43-54. Cooke RWI (1990) Cerebral palsy in very low birthweight infants. Arch Dis Child 65: 201-206. Counsell SJ, Maalouf EF, Fletcher AM, Duggan P, Battin M, Lewis HJ, Herlihy AH, Edwards AD, Bydder GM & Rutherford MA (2002) MR imaging assessment of myelination in the very preterm brain. AJNR Am J Neuroradiol 23: 872-881. Counsell SJ, Allsop JM, Harrison MC, Larkman DJ, Kennea NL, Kapellou O, Cowan FM, Hajnal JV, Edwards AD & Rutherford MA (2003) Diffusion-weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics 112: 1-7. Croen LA, Grether JK, Curry CJ & Nelson KB (2001) Congenital abnormalities among children with cerebral palsy: More evidence for prenatal antecedents. J Pediatr 138: 804-810. Dammann O & Leviton A (1997a) Maternal intrauterine infection, cytokines, and brain damage in the preterm newborn. Pediatr Res 42: 1-8. Dammann O & Leviton A (1997b) Duration of transient hyperechoic images of white matter in very-low-birthweight infants: a proposed classification. Dev Med Child Neurol 39: 2-5. Dammann O & Leviton A (1999) Brain damage in preterm newborns: might enhancement of developmentally regulated endogenous protection open a door for prevention? Pediatrics 104: 541-550. Dammann O, Allred EN, Kuban KC, Van Marter LJ, Stewart JE, Pagano M & Leviton A (2001a) Hypocarbia during the first 24 postnatal hours and white matter echolucencies in newborns < or = 28 weeks gestation. Pediatr Res 49: 388-393. Dammann O, Phillips TM, Allred EN, O'Shea TM, Paneth N, Van Marter LJ, Bose C, Ehrenkranz RA, Bednarek FJ, Naples M & Leviton A (2001b) Mediators of fetal inflammation in extremely low gestational age newborns. Cytokine 13: 234-239. De Dooy J, Colpaert C, Schuerwegh A, Bridts C, Van Der PM, Ieven M, De Clerck L, Stevens W & Mahieu L (2003) Relationship between histologic chorioamnionitis and early inflammatory variables in blood, tracheal aspirates, and endotracheal colonization in preterm infants. Pediatr Res 54: 113-119. De Reuck J (1971) The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 5: 321-334. De Reuck JL (1984) Cerebral angioarchitecture and perinatal brain lesions in premature and fullterm infants. Acta Neurol Scand 70: 391-395. de Vries LS, Regev R, Dubowitz LM, Whitelaw A & Aber VR (1988) Perinatal risk factors for the development of extensive cystic leukomalacia. Am J Dis Child 142: 732-735. de Vries LS, Eken P & Dubowitz LM (1992) The spectrum of leukomalacia using cranial ultrasound. Behav Brain Res 49: 1-6. Deguchi K, Mizuguchi M & Takashima S (1996) Immunohistochemical expression of tumor necrosis factor alpha in neonatal leukomalacia. Pediatr Neurol 14: 13-16. Dembinski J, Behrendt D, Martini R, Heep A & Bartmann P (2003) Modulation of pro- and antiinflammatory cytokine production in very preterm infants. Cytokine 21: 200-206. Dexter SC, Malee MP, Pinar H, Hogan JW, Carpenter MW & Vohr BR (1999) Influence of chorioamnionitis on developmental outcome in very low birth weight infants. Obstet Gynecol 94: 267-273. Dexter SC, Pinar H, Malee MP, Hogan J, Carpenter MW & Vohr BR (2000) Outcome of very low birth weight infants with histopathologic chorioamnionitis. Obstet Gynecol 96: 172-177. Dietl J (2000) The pathogenesis of pre-eclampsia: new aspects. J Perinat Med 28: 464-471.

86 Disalvo D (1998) The correlation between placental pathology and intraventricular hemorrhage in the preterm infant. The Developmental Epidemiology Network Investigators. Pediatr Res 43: 15-19. Dollner H, Vatten L, Halgunset J, Rahimipoor S & Austgulen R (2002) Histologic chorioamnionitis and umbilical serum levels of pro-inflammatory cytokines and cytokine inhibitors. BJOG 109: 534-539. Dommergues MA, Patkai J, Renauld JC, Evrard P & Gressens P (2000) Proinflammatory cytokines and interleukin-9 exacerbate excitotoxic lesions of the newborn murine neopallium. Ann Neurol 47: 54-63. Donofrio MT, Bremer YA, Schieken RM, Gennings C, Morton LD, Eidem BW, Cetta F, Falkensammer CB, Huhta JC & Kleinman CS (2003) Autoregulation of cerebral blood flow in fetuses with congenital heart disease: the brain sparing effect. Pediatr Cardiol 24: 436-443. Dorovini-Zis K & Dolman CL (1977) Gestational development of brain. Arch Pathol Lab Med 101: 192-195. Duckett S (1971) The establishment of internal vascularization in the human telencephalon. Acta Anat (Basel) 80: 107-113. Dudley DJ, Trautman MS, Araneo BA, Edwin SS & Mitchell MD (1992) Decidual cell biosynthesis of interleukin-6: regulation by inflammatory cytokines. J Clin Endocrinol Metab 74: 884-889. Duggan PJ, Maalouf EF, Watts TL, Sullivan MH, Counsell SJ, Allsop J, Al Nakib L, Rutherford MA, Battin M, Roberts I & Edwards AD (2001) Intrauterine T-cell activation and increased proinflammatory cytokine concentrations in preterm infants with cerebral lesions. Lancet 358: 1699-1700. Duncan JR, Cock ML, Harding R & Rees SM (2000) Relation between damage to the placenta and the fetal brain after late-gestation placental embolization and fetal growth restriction in sheep. Am J Obstet Gynecol 183: 1013-1022. Duncan JR, Cock ML, Scheerlinck JP, Westcott KT, McLean C, Harding R & Rees SM (2002) White matter injury after repeated endotoxin exposure in the preterm ovine fetus. Pediatr Res 52: 941-949. Eastman NJ & DeLeon M (1955) The etiology of cerebral palsy. Am J Obstet Gynecol 69: 950961. Edelstone DI & Rudolph AM (1979) Preferential streaming of ductus venosus blood to the brain and heart in fetal lambs. Am J Physiol 237: H724-H729. Ellison MD & Merchant RE (1991) Appearance of cytokine-associated central nervous system myelin damage coincides temporally with serum tumor necrosis factor induction after recombinant interleukin-2 infusion in rats. J Neuroimmunol 33: 245-251. Engelbrecht V, Scherer A, Rassek M, Witsack HJ & Modder U (2002) Diffusion-weighted MR imaging in the brain in children: findings in the normal brain and in the brain with white matter diseases. Radiology 222: 410-418. Evrard P, Miladi N, Bonnier C & Gressens P (1992) Normal and abnormal development of the brain. In: Rapin I & Segalowitz SJ (eds) Handbook of neuropsychology. Elsevier Science Publishers B. V., 11-44. Faas MM, Schuiling GA, Baller JF, Visscher CA & Bakker WW (1994) A new animal model for human preeclampsia: ultra-low-dose endotoxin infusion in pregnant rats. Am J Obstet Gynecol 171: 158-164. Felderhoff-Mueser U, Rutherford MA, Squier WV, Cox P, Maalouf EF, Counsell SJ, Bydder GM & Edwards AD (1999) Relationship between MR imaging and histopathologic findings of the brain in extremely sick preterm infants. AJNR Am J Neuroradiol 20: 1349-1357. Fellman V & Raivio KO (1997) Reperfusion injury as the mechanism of brain damage after perinatal asphyxia. Pediatr Res 41: 599-606.

87 Ferrazzi E, Bellotti M, Marconi A, Flisi L, Barbera A & Pardi G (1995) Peak velocity of the outflow tract of the aorta: correlations with acid base status and oxygenation of the growthretarded fetus. Obstet Gynecol 85: 663-668. Ferrazzi E, Bulfamante G, Mezzopane R, Barbera A, Ghidini A & Pardi G (1999) Uterine Doppler velocimetry and placental hypoxic-ischemic lesion in pregnancies with fetal intrauterine growth restriction. Placenta 20: 389-394. Ferrazzi E, Bozzo M, Rigano S, Bellotti M, Morabito A, Pardi G, Battaglia FC & Galan HL (2002) Temporal sequence of abnormal Doppler changes in the peripheral and central circulatory systems of the severely growth-restricted fetus. Ultrasound Obstet Gynecol 19: 140-146. Folkerth RD, Keefe RJ, Haynes RL, Trachtenberg FL, Volpe JJ & Kinney HC (2004a) Interferongamma expression in periventricular leukomalacia in the human brain. Brain Pathol 14: 265274. Folkerth RD, Haynes RL, Borenstein NS, Belliveau RA, Trachtenberg F, Rosenberg PA, Volpe JJ & Kinney HC (2004b) Developmental lag in superoxide dismutases relative to other antioxidant enzymes in premyelinated human telencephalic white matter. J Neuropathol Exp Neurol 63: 990-999. Follett PL, Rosenberg PA, Volpe JJ & Jensen FE (2000) NBQX attenuates excitotoxic injury in developing white matter. J Neurosci 20: 9235-9241. Fouron JC, Gosselin J, Amiel-Tison C, Infante-Rivard C, Fouron C, Skoll A & Veilleux A (2001) Correlation between prenatal velocity waveforms in the aortic isthmus and neurodevelopmental outcome between the ages of 2 and 4 years. Am J Obstet Gynecol 184: 630-636. Frisone MF, Mercuri E, Laroche S, Foglia C, Maalouf EF, Haataja L, Cowan F & Dubowitz L (2002) Prognostic value of the neurologic optimality score at 9 and 18 months in preterm infants born before 31 weeks' gestation. J Pediatr 140: 57-60. Gaffney G, Squier MV, Johnson A, Flavell V & Sellers S (1994) Clinical associations of prenatal ischaemic white matter injury. Arch Dis Child Fetal Neonatal Ed 70: F101-F106. Galan HL, Ferrazzi E & Hobbins JC (2002) Intrauterine growth restriction (IUGR): biometric and Doppler assessment. Prenat Diagn 22: 331-337. Gibbs RS, Blanco JD, St Clair PJ & Castaneda YS (1982) Quantitative bacteriology of amniotic fluid from women with clinical intraamniotic infection at term. J Infect Dis 145: 1-8. Gibson CS, MacLennan AH, Goldwater PN & Dekker GA (2003) Antenatal causes of cerebral palsy: associations between inherited thrombophilias, viral and bacterial infection, and inherited susceptibility to infection. Obstet Gynecol Surv 58: 209-220. Giles WB, Trudinger BJ & Baird PJ (1985) Fetal umbilical artery flow velocity waveforms and placental resistance: pathological correlation. Br J Obstet Gynaecol 92: 31-38. Gilles FH, Leviton A & Kerr CS (1976) Endotoxin leucoencephalopathy in the telencephalon of the newborn kitten. J Neurol Sci 27: 183-191. Girard N, Raybaud C & Poncet M (1995) In vivo MR study of brain maturation in normal fetuses. AJNR Am J Neuroradiol 16: 407-413. Goepfert AR, Andrews WW, Carlo W, Ramsey PS, Cliver SP, Goldenberg RL & Hauth JC (2004) Umbilical cord plasma interleukin-6 concentrations in preterm infants and risk of neonatal morbidity. Am J Obstet Gynecol 191: 1375-1381. Goldenberg RL, Cutter GR, Hoffman HJ, Foster JM, Nelson KG & Hauth JC (1989) Intrauterine growth retardation: standards for diagnosis. Am J Obstet Gynecol 161: 271-277. Goldenberg RL, Hauth JC & Andrews WW (2000) Intrauterine infection and preterm delivery. N Engl J Med 342: 1500-1507. Goldenberg RL & Culhane JF (2003) Infection as a cause of preterm birth. Clin Perinatol 30: 677700. Gomez R, Romero R, Ghezzi F, Yoon BH, Mazor M & Berry SM (1998) The fetal inflammatory response syndrome. Am J Obstet Gynecol 179: 194-202.

88 Gould SJ, Howard S, Hope PL & Reynolds EO (1987) Periventricular intraparenchymal cerebral haemorrhage in preterm infants: the role of venous infarction. J Pathol 151: 197-202. Gravett MG, Hummel D, Eschenbach DA & Holmes KK (1986) Preterm labor associated with subclinical amniotic fluid infection and with bacterial vaginosis. Obstet Gynecol 67: 229-237. Gravett MG, Witkin SS, Haluska GJ, Edwards JL, Cook MJ & Novy MJ (1994) An experimental model for intraamniotic infection and preterm labor in rhesus monkeys. Am J Obstet Gynecol 171: 1660-1667. Gray PH, Jones P & O'Callaghan MJ (2001) Maternal antecedents for cerebral palsy in extremely preterm babies: a case-control study. Dev Med Child Neurol 43: 580-585. Graziani LJ, Spitzer AR, Mitchell DG, Merton DA, Stanley C, Robinson N & McKee L (1992) Mechanical ventilation in preterm infants: neurosonographic and developmental studies. Pediatrics 90: 515-522. Gressens P (2002) Neuronal migration. In: Lagercrantz H, Hanson M, Evrard P & Rodeck C (eds) The newborn brain. Cambridge University Press, Cambridge, 69-90. Grether JK, Nelson KB, Emery ES, III & Cummins SK (1996) Prenatal and perinatal factors and cerebral palsy in very low birth weight infants. J Pediatr 128: 407-414. Grether JK & Nelson KB (1997) Maternal infection and cerebral palsy in infants of normal birth weight. JAMA 278: 207-211. Grether JK, Nelson KB, Dambrosia JM & Phillips TM (1999) Interferons and cerebral palsy. J Pediatr 134: 324-332. Grether JK, Nelson KB, Walsh E, Willoughby RE & Redline RW (2003) Intrauterine exposure to infection and risk of cerebral palsy in very preterm infants. Arch Pediatr Adolesc Med 157: 2632. Griffiths R (1954) The abilities of babies. A study in mental measurement. McGraw-Hill Book Company Inc., London. Guzzetta F, Shackelford GD, Volpe S, Perlman JM & Volpe JJ (1986) Periventricular intraparenchymal echodensities in the premature newborn: critical determinant of neurologic outcome. Pediatrics 78: 995-1006. Hack M & Fanaroff AA (2000) Outcomes of children of extremely low birthweight and gestational age in the 1990s. Semin Neonatol 5: 89-106. Hack M, Wilson-Costello D, Friedman H, Taylor GH, Schluchter M & Fanaroff AA (2000) Neurodevelopment and predictors of outcomes of children with birth weights of less than 1000 g: 1992-1995. Arch Pediatr Adolesc Med 154: 725-731. Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E & Klein N (2002) Outcomes in young adulthood for very-low-birth-weight infants. N Engl J Med 346: 149-157. Hagberg B & Hagberg G (1993) The origins of cerebral palsy. In: David TJ (ed) Recent advances in paediatrics. Churchill Livingstone Inc., New York, 67-83. Hagberg B & Hagberg G (1996) The changing panorama of cerebral palsy--bilateral spastic forms in particular. Acta Paediatr Suppl 416: 48-52. Hagberg B, Hagberg G, Beckung E & Uvebrant P (2001) Changing panorama of cerebral palsy in Sweden. VIII. Prevalence and origin in the birth year period 1991-94. Acta Paediatr 90: 271277. Hagberg H, Gilland E, Bona E, Hanson LA, Hahin-Zoric M, Blennow M, Holst M, McRae A & Soder O (1996) Enhanced expression of interleukin (IL)-1 and IL-6 messenger RNA and bioactive protein after hypoxia-ischemia in neonatal rats. Pediatr Res 40: 603-609. Hagberg H, Wennerholm UB & Savman K (2002) Sequelae of chorioamnionitis. Curr Opin Infect Dis 15: 301-306. Hallman M, Ramet M & Ezekowitz RA (2001) Toll-like receptors as sensors of pathogens. Pediatr Res 50: 315-321.

89 Harding D, Dhamrait S, Millar A, Humphries S, Marlow N, Whitelaw A & Montgomery H (2003) Is interleukin-6 -174 genotype associated with the development of septicemia in preterm infants? Pediatrics 112: 800-803. Harding DR, Dhamrait S, Whitelaw A, Humphries SE, Marlow N & Montgomery HE (2004) Does interleukin-6 genotype influence cerebral injury or developmental progress after preterm birth? Pediatrics 114: 941-947. Hartmann K, Thorp JM, Jr., McDonald TL, Savitz DA & Granados JL (1999) Cervical dimensions and risk of preterm birth: a prospective cohort study. Obstet Gynecol 93: 504-509. Haynes RL, Folkerth RD, Keefe RJ, Sung I, Swzeda LI, Rosenberg PA, Volpe JJ & Kinney HC (2003) Nitrosative and oxidative injury to premyelinating oligodendrocytes in periventricular leukomalacia. J Neuropathol Exp Neurol 62: 441-450. Hecher K, Snijders R, Campbell S & Nicolaides K (1995a) Fetal venous, intracardiac, and arterial blood flow measurements in intrauterine growth retardation: relationship with fetal blood gases. Am J Obstet Gynecol 173: 10-15. Hecher K, Campbell S, Doyle P, Harrington K & Nicolaides K (1995b) Assessment of fetal compromise by Doppler ultrasound investigation of the fetal circulation. Arterial, intracardiac, and venous blood flow velocity studies. Circulation 91: 129-138. Hecher K, Bilardo CM, Stigter RH, Ville Y, Hackeloer BJ, Kok HJ, Senat MV & Visser GH (2001) Monitoring of fetuses with intrauterine growth restriction: a longitudinal study. Ultrasound Obstet Gynecol 18: 564-570. Hedberg CL, Adcock K, Martin J, Loggins J, Kruger TE & Baier RJ (2004) Tumor necrosis factor alpha -- 308 polymorphism associated with increased sepsis mortality in ventilated very low birth weight infants. Pediatr Infect Dis J 23: 424-428. Heep A, Behrendt D, Nitsch P, Fimmers R, Bartmann P & Dembinski J (2003) Increased serum levels of interleukin 6 are associated with severe intraventricular haemorrhage in extremely premature infants. Arch Dis Child Fetal Neonatal Ed 88: F501-F504. Heuchan AM, Evans N, Henderson Smart DJ & Simpson JM (2002) Perinatal risk factors for major intraventricular haemorrhage in the Australian and New Zealand Neonatal Network, 1995-97. Arch Dis Child Fetal Neonatal Ed 86: F86-F90. Hill A, Perlman JM & Volpe JJ (1982) Relationship of pneumothorax to occurrence of intraventricular hemorrhage in the premature newborn. Pediatrics 69: 144-149. Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK & Eschenbach DA (1988) A case-control study of chorioamnionic infection and histologic chorioamnionitis in prematurity. N Engl J Med 319: 972-978. Hillier SL, Nugent RP, Eschenbach DA, Krohn MA, Gibbs RS, Martin DH, Cotch MF, Edelman R, Pastorek JG, Rao AV & . (1995) Association between bacterial vaginosis and preterm delivery of a low-birth-weight infant. The Vaginal Infections and Prematurity Study Group. N Engl J Med 333: 1737-1742. Huisman TW (2001) Doppler assessment of the fetal venous system. Semin Perinatol 25: 21-31. Hüppi PS, Schuknecht B, Boesch C, Bossi E, Felblinger J, Fusch C & Herschkowitz N (1996) Structural and neurobehavioral delay in postnatal brain development of preterm infants. Pediatr Res 39: 895-901. Hüppi PS, Maier SE, Peled S, Zientara GP, Barnes PD, Jolesz FA & Volpe JJ (1998) Microstructural development of human newborn cerebral white matter assessed in vivo by diffusion tensor imaging. Pediatr Res 44: 584-590. Hüppi PS, Murphy B, Maier SE, Zientara GP, Inder TE, Barnes PD, Kikinis R, Jolesz FA & Volpe JJ (2001) Microstructural brain development after perinatal cerebral white matter injury assessed by diffusion tensor imaging. Pediatrics 107: 455-460. Hüppi PS (2002) Advances in postnatal neuroimaging: relevance to pathogenesis and treatment of brain injury. Clin Perinatol 29: 827-856.

90 Iams JD (2003) The epidemiology of preterm birth. Clin Perinatol 30: 651-664. Inder TE, Hüppi PS, Warfield S, Kikinis R, Zientara GP, Barnes PD, Jolesz F & Volpe JJ (1999a) Periventricular white matter injury in the premature infant is followed by reduced cerebral cortical gray matter volume at term. Ann Neurol 46: 755-760. Inder TE, Hüppi PS, Zientara GP, Maier SE, Jolesz FA, di Salvo D, Robertson R, Barnes PD & Volpe JJ (1999b) Early detection of periventricular leukomalacia by diffusion- weighted magnetic resonance imaging techniques. J Pediatr 134: 631-634. Inder TE & Volpe JJ (2000) Mechanisms of perinatal brain injury. Semin Neonatol 5: 3-16. Inder T, Mocatta T, Darlow B, Spencer C, Volpe JJ & Winterbourn C (2002a) Elevated free radical products in the cerebrospinal fluid of VLBW infants with cerebral white matter injury. Pediatr Res 52: 213-218 Inder T, Mocatta T, Darlow B, Spencer C, Senthilmohan R, Winterbourn CC & Vople JJ (2002b) Markers of oxidative injury in the cerebrospinal fluid of a premature infant with meningitis and periventricular leukomalacia. J Pediatr 140: 617-621. Inder TE, Wells SJ, Mogridge NB, Spencer C & Volpe JJ (2003a) Defining nature of the cerebral abnormalities in the premature infant: a qualitative magnetic resonance imaging study. J Pediatr 143: 171-179. Inder TE, Anderson NJ, Spencer C, Wells S & Volpe JJ (2003b) White matter injury in the premature infant: a comparison between serial cranial sonographic and MR findings at term. Am J Neuroradiol 24: 805-809. Inder TE, Warfield SK, Wang H, Hüppi PS & Volpe JJ (2005) Abnormal cerebral structure is present at term in premature infants. Pediatrics 115: 286-294. Isaacs EB, Edmonds CJ, Lucas A & Gadian DG (2001) Calculation difficulties in children of very low birthweight: a neural correlate. Brain 124: 1701-1707. Jacobsson B, Hagberg G, Hagberg B, Ladfors L, Niklasson A & Hagberg H (2002) Cerebral palsy in preterm infants: a population-based case-control study of antenatal and intrapartum risk factors. Acta Paediatr 91: 946-951. Janeway CA, Travers P, Walport M & Shlomchik M (2001) Innate immunity. In: Immunobiology. Garland Publishing, New York, 35-91. Jeffcoat MK, Geurs NC, Reddy MS, Cliver SP, Goldenerg RL & Hauth JC (2001) Periodontal infection and preterm birth: results of a prospective study. J Am Dent Assoc 132: 875-880. Jonakait GM (1997) Cytokines in neuronal development. Adv Pharmacol 37: 35-67. Kadhim H, Tabarki B, Verellen G, De Prez C, Rona AM & Sebire G (2001) Inflammatory cytokines in the pathogenesis of periventricular leukomalacia. Neurology 56: 1278-1284. Kadhim H & Sebire G (2002) Immune mechanisms in the pathogenesis of cerebral palsy: implication of proinflammatory cytokines and T lymphocytes. Eur J Paediatr Neurol 6: 139-142. Kadhim H, Tabarki B, De Prez C & Sebire G (2003) Cytokine immunoreactivity in cortical and subcortical neurons in periventricular leukomalacia: are cytokines implicated in neuronal dysfunction in cerebral palsy? Acta Neuropathol (Berl) 105: 209-216. Kahn MA & De Vellis J (1994) Regulation of an oligodendrocyte progenitor cell line by the interleukin-6 family of cytokines. Glia 12: 87-98. Kameda T, Matsuzaki N, Sawai K, Okada T, Saji F, Matsuda T, Hirano T, Kishimoto T & Tanizawa O (1990) Production of interleukin-6 by normal human trophoblast. Placenta 11: 205213. Kashlan F, Smulian J, Shen-Schwarz S, Anwar M, Hiatt M & Hegyi T (2000) Umbilical vein interleukin 6 and tumor necrosis factor alpha plasma concentrations in the very preterm infant. Pediatr Infect Dis J 19: 238-243.

91 Kim YM, Romero R, Chaiworapongsa T, Kim GJ, Kim MR, Kuivaniemi H, Tromp G, Espinoza J, Bujold E, Abrahams VM & Mor G (2004) Toll-like receptor-2 and -4 in the chorioamniotic membranes in spontaneous labor at term and in preterm parturition that are associated with chorioamnionitis. Am J Obstet Gynecol 191: 1346-1355. Kingsmore SF & Patel DD (2003) Multiplexed protein profiling on antibody-based microarrays by rolling circle amplification. Curr Opin Biotechnol 14: 74-81. Kinney HC, Brody BA, Kloman AS & Gilles FH (1988) Sequence of central nervous system myelination in human infancy. II. Patterns of myelination in autopsied infants. J Neuropathol Exp Neurol 47: 217-234. Kiserud T, Chedid G & Rasmussen S (2004) Foramen ovale changes in growth-restricted fetuses. Ultrasound Obstet Gynecol 24: 141-146. Korn AP, Bolan G, Padian N, Ohm-Smith M, Schachter J & Landers DV (1995) Plasma cell endometritis in women with symptomatic bacterial vaginosis. Obstet Gynecol 85: 387-390. Kostovic I, Seress L, Mrzljak L & Judas M (1989) Early onset of synapse formation in the human hippocampus: a correlation with Nissl-Golgi architectonics in 15- and 16.5-week-old fetuses. Neuroscience 30: 105-116. Kostovic I & Rakic P (1990) Developmental history of the transient subplate zone in the visual and somatosensory cortex of the macaque monkey and human brain. J Comp Neurol 297: 441-470. Kuban KC & Gilles FH (1985) Human telencephalic angiogenesis. Ann Neurol 17: 539-548. Kumazaki K, Nakayama M, Sumida Y, Ozono K, Mushiake S, Suehara N, Wada Y & Fujimura M (2002) Placental features in preterm infants with periventricular leukomalacia. Pediatrics 109: 650-655. Kurkinen-Räty M, Kivelä A & Jouppila P (1997) The clinical significance of an absent enddiastolic velocity in the umbilical artery detected before the 34th week of pregnancy. Acta Obstet Gynecol Scand 76: 398-404. Kutschera J, Tomaselli J, Urlesberger B, Maurer U, Hausler M, Gradnitzer E, Burmucic K & Muller W (2002) Absent or reversed end-diastolic blood flow in the umbilical artery and abnormal Doppler cerebroplacental ratio--cognitive, neurological and somatic development at 3 to 6 years. Early Hum Dev 69: 47-56. Labarrere CA, Catoggio LJ, Mullen EG & Althabe OH (1986) Placental lesions in maternal autoimmune diseases. Am J Reprod Immunol Microbiol 12: 78-86. Lahra MM & Jeffery HE (2004) A fetal response to chorioamnionitis is associated with early survival after preterm birth. Am J Obstet Gynecol 190: 147-151. Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E & Laval-Jeantet M (1986) MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 161: 401-407. Le Bihan D, Turner R, Douek P & Patronas N (1992) Diffusion MR imaging: clinical applications. AJR Am J Roentgenol 159: 591-599. Lee SC, Liu W, Dickson DW, Brosnan CF & Berman JW (1993a) Cytokine production by human fetal microglia and astrocytes. Differential induction by lipopolysaccharide and IL-1 beta. J Immunol 150: 2659-2667. Lee SC, Dickson DW, Liu W & Brosnan CF (1993b) Induction of nitric oxide synthase activity in human astrocytes by interleukin-1 beta and interferon-gamma. J Neuroimmunol 46: 19-24. Leitich H, Bodner-Adler B, Brunbauer M, Kaider A, Egarter C & Husslein P (2003) Bacterial vaginosis as a risk factor for preterm delivery: a meta-analysis. Am J Obstet Gynecol 189: 139147.

92 Lemons JA, Bauer CR, Oh W, Korones SB, Papile LA, Stoll BJ, Verter J, Temprosa M, Wright LL, Ehrenkranz RA, Fanaroff AA, Stark A, Carlo W, Tyson JE, Donovan EF, Shankaran S & Stevenson DK (2001) Very low birth weight outcomes of the National Institute of Child health and human development neonatal research network, January 1995 through December 1996. NICHD Neonatal Research Network. Pediatrics 107: E1. Levine D & Barnes PD (1999) Cortical maturation in normal and abnormal fetuses as assessed with prenatal MR imaging. Radiology 210: 751-758. Leviton A & Paneth N (1990) White matter damage in preterm newborns--an epidemiologic perspective. Early Hum Dev 24: 1-22. Leviton A, Fenton T, Kuban KC & Pagano M (1991) Labor and delivery characteristics and the risk of germinal matrix hemorrhage in low birth weight infants. J Child Neurol 6: 35-40. Leviton A (1993) Preterm birth and cerebral palsy: is tumor necrosis factor the missing link? Dev Med Child Neurol 35: 553-558. Leviton A, Paneth N, Reuss ML, Susser M, Allred EN, Dammann O, Kuban K, Van Marter LJ, Pagano M, Hegyi T, Hiatt M, Sanocka U, Shahrivar F, Abiri M, Disalvo D, Doubilet P, Kairam R, Kazam E, Kirpekar M, Rosenfeld D, Schonfeld S, Share J, Collins M, Genest D, ShenSchwarz S & . (1999) Maternal infection, fetal inflammatory response, and brain damage in very low birth weight infants. Developmental Epidemiology Network Investigators. Pediatr Res 46: 566-575. Ley D, Laurin J, Bjerre I & Marsal K (1996a) Abnormal fetal aortic velocity waveform and minor neurological dysfunction at 7 years of age. Ultrasound Obstet Gynecol 8: 152-159. Ley D, Tideman E, Laurin J, Bjerre I & Marsal K (1996b) Abnormal fetal aortic velocity waveform and intellectual function at 7 years of age. Ultrasound Obstet Gynecol 8: 160-165. Liu MT, Armstrong D, Hamilton TA & Lane TE (2001) Expression of Mig (monokine induced by interferon-gamma) is important in T lymphocyte recruitment and host defense following viral infection of the central nervous system. J Immunol 166: 1790-1795. Lockwood CJ & Kuczynski E (1999) Markers of risk for preterm delivery. J Perinat Med 27: 5-20. Lorenz JM (2001) The outcome of extreme prematurity. Semin Perinatol 25: 348-359. Lou HC, Lassen NA & Friis-Hansen B (1979) Impaired autoregulation of cerebral blood flow in the distressed newborn infant. J Pediatr 94: 118-121. Louis JC, Magal E, Takayama S & Varon S (1993) CNTF protection of oligodendrocytes against natural and tumor necrosis factor-induced death. Science 259: 689-692. Maalouf EF, Duggan PJ, Rutherford MA, Counsell SJ, Fletcher AM, Battin M, Cowan F & Edwards AD (1999) Magnetic resonance imaging of the brain in a cohort of extremely preterm infants. J Pediatr 135: 351-357. Maalouf EF, Duggan PJ, Counsell SJ, Rutherford MA, Cowan F, Azzopardi D & Edwards AD (2001) Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 107: 719-727. Macara L, Kingdom JC, Kaufmann P, Kohnen G, Hair J, More IA, Lyall F & Greer IA (1996) Structural analysis of placental terminal villi from growth-restricted pregnancies with abnormal umbilical artery Doppler waveforms. Placenta 17: 37-48. MacDorman MF, Minino AM, Strobino DM & Guyer B (2002) Annual summary of vital statistics-2001. Pediatrics 110: 1037-1052. Mallard EC, Rees S, Stringer M, Cock ML & Harding R (1998) Effects of chronic placental insufficiency on brain development in fetal sheep. Pediatr Res 43: 262-270. Mallard C, Welin AK, Peebles D, Hagberg H & Kjellmer I (2003) White matter injury following systemic endotoxemia or asphyxia in the fetal sheep. Neurochem Res 28: 215-223. Marin-Padilla M (1997) Developmental neuropathology and impact of perinatal brain damage. II: white matter lesions of the neocortex. J Neuropathol Exp Neurol 56: 219-235. Markand ON (1994) Brainstem auditory evoked potentials. J Clin Neurophysiol 11: 319-342.

93 Marlow N, Wolke D, Bracewell MA & Samara M (2005) Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 352: 9-19. Matsuda T, Okuyama K, Cho K, Hoshi N, Matsumoto Y, Kobayashi Y & Fujimoto S (1999) Induction of antenatal periventricular leukomalacia by hemorrhagic hypotension in the chronically instrumented fetal sheep. Am J Obstet Gynecol 181: 725-730. McArdle CB, Richardson CJ, Nicholas DA, Mirfakhraee M, Hayden CK & Amparo EG (1987) Developmental features of the neonatal brain: MR imaging. Part I. Gray-white matter differentiation and myelination. Radiology 162: 223-229. McKay R (1997) Stem cells in the central nervous system. Science 276: 66-71. McKinstry RC, Mathur A, Miller JH, Ozcan A, Snyder AZ, Schefft GL, Almli CR, Shiran SI, Conturo TE & Neil JJ (2002a) Radial organization of developing preterm human cerebral cortex revealed by non-invasive water diffusion anisotropy MRI. Cereb Cortex 12: 1237-1243. McKinstry RC, Miller JH, Snyder AZ, Mathur A, Schefft GL, Almli CR, Shimony JS, Shiran SI & Neil JJ (2002b) A prospective, longitudinal diffusion tensor imaging study of brain injury in newborns. Neurology 59: 824-833. McManus C, Berman JW, Brett FM, Staunton H, Farrell M & Brosnan CF (1998) MCP-1, MCP-2 and MCP-3 expression in multiple sclerosis lesions: an immunohistochemical and in situ hybridization study. J Neuroimmunol 86: 20-29. Meis PJ, Michielutte R, Peters TJ, Wells HB, Sands RE, Coles EC & Johns KA (1995) Factors associated with preterm birth in Cardiff, Wales. II. Indicated and spontaneous preterm birth. Am J Obstet Gynecol 173: 597-602. Meis PJ, Goldenberg RL, Mercer BM, Iams JD, Moawad AH, Miodovnik M, Menard MK, Caritis SN, Thurnau GR, Bottoms SF, Das A, Roberts JM & McNellis D (1998) The preterm prediction study: risk factors for indicated preterm births. Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development. Am J Obstet Gynecol 178: 562567. Melhem ER (2002) Time-course of apparent diffusion coefficient in neonatal brain injury: the first piece of the puzzle. Neurology 59: 798-799. Ment LR, Bada HS, Barnes P, Grant PE, Hirtz D, Papile LA, Pinto-Martin J, Rivkin M & Slovis TL (2002) Practice parameter: neuroimaging of the neonate: report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 58: 1726-1738. Merrill JE & Benveniste EN (1996) Cytokines in inflammatory brain lesions: helpful and harmful. Trends Neurosci 19: 331-338. Miller LC, Isa S, LoPreste G, Schaller JG & Dinarello CA (1990) Neonatal interleukin-1 beta, interleukin-6, and tumor necrosis factor: cord blood levels and cellular production. J Pediatr 117: 961-965. Miller SP, Vigneron DB, Henry RG, Bohland MA, Ceppi-Cozzio C, Hoffman C, Newton N, Partridge JC, Ferriero DM & Barkovich AJ (2002) Serial quantitative diffusion tensor MRI of the premature brain: development in newborns with and without injury. J Magn Reson Imaging 16: 621-632. Minagawa K, Tsuji Y, Ueda H, Koyama K, Tanizawa K, Okamura H & Hashimoto-Tamaoki T (2002) Possible correlation between high levels of IL-18 in the cord blood of pre-term infants and neonatal development of periventricular leukomalacia and cerebral palsy. Cytokine 17: 164170. Mirmiran M, Barnes PD, Keller K, Constantinou JC, Fleisher BE, Hintz SR & Ariagno RL (2004) Neonatal brain magnetic resonance imaging before discharge is better than serial cranial ultrasound in predicting cerebral palsy in very low birth weight preterm infants. Pediatrics 114: 992-998.

94 Morriss MC, Zimmerman RA, Bilaniuk LT, Hunter JV & Haselgrove JC (1999) Changes in brain water diffusion during childhood. Neuroradiology 41: 929-934. Mueller-Heubach E, Rubinstein DN & Schwarz SS (1990) Histologic chorioamnionitis and preterm delivery in different patient populations. Obstet Gynecol 75: 622-626. Murphy DJ, Sellers S, MacKenzie IZ, Yudkin PL & Johnson AM (1995) Case-control study of antenatal and intrapartum risk factors for cerebral palsy in very preterm singleton babies. Lancet 346: 1449-1454. Mutch L, Alberman E, Hagberg B, Kodama K & Perat MV (1992) Cerebral palsy epidemiology: where are we now and where are we going? Dev Med Child Neurol 34: 547-551. Mäkikallio K, Vuolteenaho O, Jouppila P & Räsänen J (2002) Ultrasonographic and biochemical markers of human fetal cardiac dysfunction in placental insufficiency. Circulation 105: 20582063. Naccasha N, Hinson R, Montag A, Ismail M, Bentz L & Mittendorf R (2001) Association between funisitis and elevated interleukin-6 in cord blood. Obstet Gynecol 97: 220-224. Nakamura Y, Okudera T & Hashimoto T (1991) Microvasculature in germinal matrix layer: its relationship to germinal matrix hemorrhage. Mod Pathol 4: 475-480. Neil JJ, Shiran SI, McKinstry RC, Schefft GL, Snyder AZ, Almli CR, Akbudak E, Aronovitz JA, Miller JP, Lee BC & Conturo TE (1998) Normal brain in human newborns: apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology 209: 57-66. Nelson KB & Ellenberg JH (1986) Antecedents of cerebral palsy. Multivariate analysis of risk. N Engl J Med 315: 81-86. Nelson KB, Dambrosia JM, Grether JK & Phillips TM (1998) Neonatal cytokines and coagulation factors in children with cerebral palsy. Ann Neurol 44: 665-675. Nelson KB, Grether JK, Dambrosia JM, Walsh E, Kohler S, Satyanarayana G, Nelson PG, Dickens BF & Phillips TM (2003) Neonatal cytokines and cerebral palsy in very preterm infants. Pediatr Res 53: 600-607. Nesin M & Cunningham-Rundles S (2000) Cytokines and neonates. Am J Perinatol 17: 393-404. Ng PC, Li K, Wong RP, Chui K, Wong E, Li G & Fok TF (2003) Proinflammatory and antiinflammatory cytokine responses in preterm infants with systemic infections. Arch Dis Child Fetal Neonatal Ed 88: F209-F213. Norman MG & O'Kusky JR (1986) The growth and development of microvasculature in human cerebral cortex. J Neuropathol Exp Neurol 45: 222-232. Nupponen I, Andersson S, Jarvenpaa AL, Kautiainen H & Repo H (2001) Neutrophil CD11b expression and circulating interleukin-8 as diagnostic markers for early-onset neonatal sepsis. Pediatrics 108: E12. O'Brien F, Roth S, Stewart A, Rifkin L, Rushe T & Wyatt J (2004) The neurodevelopmental progress of infants less than 33 weeks into adolescence. Arch Dis Child 89: 207-211. Oka A, Belliveau MJ, Rosenberg PA & Volpe JJ (1993) Vulnerability of oligodendroglia to glutamate: pharmacology, mechanisms, and prevention. J Neurosci 13: 1441-1453. Oliver AM (1988) Epidermal growth factor receptor expression in human foetal tissues is agedependent. Br J Cancer 58: 461-463. Olsén P & Vainiompää L (2000) [Neurological prognosis of the premature infant]. Duodecim 116: 2032-2037. Olsén P, Yliherva A, Pääkkö E, Järvelin MR & Tolonen U (2002) Brainstem auditory- evoked potentials of 8-year-old preterm children in relation to their psycholinguistic abilities and MRI findings. Early Hum Dev 70: 25-34. Ono J, Harada K, Sakurai K, Kodaka R, Shimidzu N, Tanaka J, Nagai T & Okada S (1994) MR diffusion imaging in Pelizaeus-Merzbacher disease. Brain Dev 16: 219-223.

95 Palmer C (1995) Hypoxic-ischemic encephalopathy. Therapeutic approaches against microvascular injury, and role of neutrophils, PAF, and free radicals. Clin Perinatol 22: 481-517. Paneth N, Rudelli R, Monte W, Rodriguez E, Pinto J, Kairam R & Kazam E (1990) White matter necrosis in very low birth weight infants: neuropathologic and ultrasonographic findings in infants surviving six days or longer. J Pediatr 116: 975-984. Papile LA, Burstein J, Burstein R & Koffler H (1978) Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 92: 529-534. Papile LA, Rudolph AM & Heymann MA (1985) Autoregulation of cerebral blood flow in the preterm fetal lamb. Pediatr Res 19: 159-161. Parry S & Strauss JF, III (1998) Premature rupture of the fetal membranes. N Engl J Med 338: 663670. Pashenkov M, Teleshova N, Kouwenhoven M, Smirnova T, Jin YP, Kostulas V, Huang YM, Pinegin B, Boiko A & Link H (2002) Recruitment of dendritic cells to the cerebrospinal fluid in bacterial neuroinfections. J Neuroimmunol 122: 106-116. Pearce WJ (2002) Cerebrovascular regulation in the neonate. In: Lagercrantz H, Hanson M, Evrard P & Rodeck C (eds) The newborn brain. Cambridge University Press, Cambridge, 252-277. Perlman JM & Volpe JJ (1985) Episodes of apnea and bradycardia in the preterm newborn: impact on cerebral circulation. Pediatrics 76: 333-338. Perlman JM, Risser R & Broyles RS (1996) Bilateral cystic periventricular leukomalacia in the premature infant: associated risk factors. Pediatrics 97: 822-827. Perlman JM (1997) Intrapartum hypoxic-ischemic cerebral injury and subsequent cerebral palsy: medicolegal issues. Pediatrics 99: 851-859. Perlman JM (1998) White matter injury in the preterm infant: an important determination of abnormal neurodevelopment outcome. Early Hum Dev 53: 99-120. Peterson BS, Anderson AW, Ehrenkranz R, Staib LH, Tageldin M, Colson E, Gore JC, Duncan CC, Makuch R & Ment LR (2003) Regional brain volumes and their later neurodevelopmental correlates in term and preterm infants. Pediatrics 111: 939-948. Pidcock FS, Graziani LJ, Stanley C, Mitchell DG & Merton D (1990) Neurosonographic features of periventricular echodensities associated with cerebral palsy in preterm infants. J Pediatr 116: 417-422. Pierrat V, Duquennoy C, van Haastert IC, Ernst M, Guilley N & de Vries LS (2001) Ultrasound diagnosis and neurodevelopmental outcome of localised and extensive cystic periventricular leucomalacia. Arch Dis Child Fetal Neonatal Ed 84: F151-F156. Pijnenborg R, Anthony J, Davey DA, Rees A, Tiltman A, Vercruysse L & van Assche A (1991) Placental bed spiral arteries in the hypertensive disorders of pregnancy. Br J Obstet Gynaecol 98: 648-655. Pillay V, Savage N & Laburn H (1994) Circulating cytokine concentrations and cytokine production by monocytes from newborn babies and adults. Pflugers Arch 428: 197-201. Possel H, Noack H, Putzke J, Wolf G & Sies H (2000) Selective upregulation of inducible nitric oxide synthase (iNOS) by lipopolysaccharide (LPS) and cytokines in microglia: in vitro and in vivo studies. Glia 32: 51-59. Price DJ, Aslam S, Tasker L & Gillies K (1997) Fates of the earliest generated cells in the developing murine neocortex. J Comp Neurol 377: 414-422. Pryds O, Greisen G, Lou H & Friis-Hansen B (1989) Heterogeneity of cerebral vasoreactivity in preterm infants supported by mechanical ventilation. J Pediatr 115: 638-645. Qiu Z, Sweeney DD, Netzeband JG & Gruol DL (1998) Chronic interleukin-6 alters NMDA receptor-mediated membrane responses and enhances neurotoxicity in developing CNS neurons. J Neurosci 18: 10445-10456.

96 Rakic P (1971) Guidance of neurons migrating to the fetal monkey neocortex. Brain Res 33: 471476. Rakic P (1988) Specification of cerebral cortical areas. Science 241: 170-176. Redline RW, Wilson-Costello D, Borawski E, Fanaroff AA & Hack M (1998) Placental lesions associated with neurologic impairment and cerebral palsy in very low-birth-weight infants. Arch Pathol Lab Med 122: 1091-1098. Redline RW, Wilson-Costello D, Borawski E, Fanaroff AA & Hack M (2000) The relationship between placental and other perinatal risk factors for neurologic impairment in very low birth weight children. Pediatr Res 47: 721-726. Reuss ML, Parer JT, Harris JL & Krueger TR (1982) Hemodynamic effects of alpha-adrenergic blockade during hypoxia in fetal sheep. Am J Obstet Gynecol 142: 410-415. Ringelberg J & van de Bor M (1993) Outcome of transient periventricular echodensities in preterm infants. Neuropediatrics 24: 269-273. Ringstedt T, Linnarsson S, Wagner J, Lendahl U, Kokaia Z, Arenas E, Ernfors P & Ibanez CF (1998) BDNF regulates reelin expression and Cajal-Retzius cell development in the cerebral cortex. Neuron 21: 305-315. Rizzo G, Capponi A, Vlachopoulou A, Angelini E, Grassi C & Romanini C (1997) The diagnostic value of interleukin-8 and fetal fibronectin concentrations in cervical secretions in patients with preterm labor and intact membranes. J Perinat Med 25: 461-468. Rogers BB, Alexander JM, Head J, McIntire D & Leveno KJ (2002) Umbilical vein interleukin-6 levels correlate with the severity of placental inflammation and gestational age. Hum Pathol 33: 335-340. Rollins BJ (1997) Chemokines. Blood 90: 909-928. Romero R, Wu YK, Brody DT, Oyarzun E, Duff GW & Durum SK (1989a) Human decidua: a source of interleukin-1. Obstet Gynecol 73: 31-34. Romero R, Mazor M, Wu YK, Avila C, Oyarzun E & Mitchell MD (1989b) Bacterial endotoxin and tumor necrosis factor stimulate prostaglandin production by human decidua. Prostaglandins Leukot Essent Fatty Acids 37: 183-186. Rounioja S, Räsänen J, Glumoff V, Ojaniemi M, Mäkikallio K & Hallman M (2003) Intra-amniotic lipopolysaccharide leads to fetal cardiac dysfunction. A mouse model for fetal inflammatory response. Cardiovasc Res 60: 156-164. Ruoss K, Lovblad K, Schroth G, Moessinger AC & Fusch C (2001) Brain development (sulci and gyri) as assessed by early postnatal MR imaging in preterm and term newborn infants. Neuropediatrics 32: 69-74. Räsänen J, Wood DC, Weiner S, Ludomirski A & Huhta JC (1996) Role of the pulmonary circulation in the distribution of human fetal cardiac output during the second half of pregnancy. Circulation 94: 1068-1073. Salafia CM, Weigl C & Silberman L (1989) The prevalence and distribution of acute placental inflammation in uncomplicated term pregnancies. Obstet Gynecol 73: 383-389. Salafia CM, Minior VK, Rosenkrantz TS, Pezzullo JC, Popek EJ, Cusick W & Vintzileos AM (1995) Maternal, placental, and neonatal associations with early germinal matrix/intraventricular hemorrhage in infants born before 32 weeks' gestation. Am J Perinatol 12: 429-436. Salafia CM, Sherer DM, Spong CY, Lencki S, Eglinton GS, Parkash V, Marley E & Lage JM (1997) Fetal but not maternal serum cytokine levels correlate with histologic acute placental inflammation. Am J Perinatol 14: 419-422. Saliba E & Marret S (2001) Cerebral white matter damage in the preterm infant: pathophysiology and risk factors. Semin Neonatol 6: 121-133. Saunders NR, Habgood MD & Dziegielewska KM (1999) Barrier mechanisms in the brain, II. Immature brain. Clin Exp Pharmacol Physiol 26: 85-91.

97 Scherjon S, Briet J, Oosting H & Kok J (2000) The discrepancy between maturation of visualevoked potentials and cognitive outcome at five years in very preterm infants with and without hemodynamic signs of fetal brain-sparing. Pediatrics 105: 385-391. Schultz C, Rott C, Temming P, Schlenke P, Moller JC & Bucsky P (2002) Enhanced interleukin-6 and interleukin-8 synthesis in term and preterm infants. Pediatr Res 51: 317-322. Schweitzer B, Roberts S, Grimwade B, Shao W, Wang M, Fu Q, Shu Q, Laroche I, Zhou Z, Tchernev VT, Christiansen J, Velleca M & Kingsmore SF (2002) Multiplexed protein profiling on microarrays by rolling-circle amplification. Nat Biotechnol 20: 359-365. Scott SM, Buenaflor GG & Orth DN (1989) Immunoreactive human epidermal growth factor concentrations in amniotic fluid, umbilical artery and vein serum, and placenta in full-term and preterm infants. Biol Neonate 56: 246-251. Selmaj KW, Farooq M, Norton WT, Raine CS & Brosnan CF (1990) Proliferation of astrocytes in vitro in response to cytokines. A primary role for tumor necrosis factor. J Immunol 144: 129135. Shalak L & Perlman JM (2002) Hemorrhagic-ischemic cerebral injury in the preterm infant: current concepts. Clin Perinatol 29: 745-763. Shalak LF, Laptook AR, Jafri HS, Ramilo O & Perlman JM (2002) Clinical chorioamnionitis, elevated cytokines, and brain injury in term infants. Pediatrics 110: 673-680. Shim SS, Romero R, Hong JS, Park CW, Jun JK, Kim BI & Yoon BH (2004) Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 191: 1339-1345. Shimoya K, Matsuzaki N, Taniguchi T, Jo T, Saji F, Kitajima H, Fujimura M, Nakayama M & Tanizawa O (1992) Interleukin-8 in cord sera: a sensitive and specific marker for the detection of preterm chorioamnionitis. J Infect Dis 165: 957-960. Shortland DB, Gibson NA, Levene MI, Archer LN, Evans DH & Shaw DE (1990) Patent ductus arteriosus and cerebral circulation in preterm infants. Dev Med Child Neurol 32: 386-393. Sie LT, van der Knaap MS, Wezel-Meijler G & Valk J (1997) MRI assessment of myelination of motor and sensory pathways in the brain of preterm and term-born infants. Neuropediatrics 28: 97-105. Silveira RC & Procianoy RS (2003) Interleukin-6 and tumor necrosis factor-alpha levels in plasma and cerebrospinal fluid of term newborn infants with hypoxic-ischemic encephalopathy. J Pediatr 143: 625-629. Silver RM, Schwinzer B & McGregor JA (1993) Interleukin-6 levels in amniotic fluid in normal and abnormal pregnancies: preeclampsia, small-for-gestational-age fetus, and premature labor. Am J Obstet Gynecol 169: 1101-1105. Simpson JE, Newcombe J, Cuzner ML & Woodroofe MN (2000) Expression of the interferongamma-inducible chemokines IP-10 and Mig and their receptor, CXCR3, in multiple sclerosis lesions. Neuropathol Appl Neurobiol 26: 133-142. Spinillo A, Capuzzo E, Orcesi S, Stronati M, Di Mario M & Fazzi E (1997) Antenatal and delivery risk factors simultaneously associated with neonatal death and cerebral palsy in preterm infants. Early Hum Dev 48: 81-91. Spinillo A, Capuzzo E, Stronati M, Ometto A, De Santolo A & Acciano S (1998) Obstetric risk factors for periventricular leukomalacia among preterm infants. Br J Obstet Gynaecol 105: 865871. Stanimirovic D & Satoh K (2000) Inflammatory mediators of cerebral endothelium: a role in ischemic brain inflammation. Brain Pathol 10: 113-126. Sugimoto T, Woo M, Nishida N, Araki A, Hara T, Yasuhara A, Kobayashi Y & Yamanouchi Y (1995) When do brain abnormalities in cerebral palsy occur? An MRI study. Dev Med Child Neurol 37: 285-292.

98 Sullivan SE, Staba SL, Gersting JA, Hutson AD, Theriaque D, Christensen RD & Calhoun DA (2002) Circulating concentrations of chemokines in cord blood, neonates, and adults. Pediatr Res 51: 653-657. Sutton MG, Plappert T & Doubilet P (1991) Relationship between placental blood flow and combined ventricular output with gestational age in normal human fetus. Cardiovasc Res 25: 603-608. Szymonowicz W, Walker AM, Yu VY, Stewart ML, Cannata J & Cussen L (1990) Regional cerebral blood flow after hemorrhagic hypotension in the preterm, near-term, and newborn lamb. Pediatr Res 28: 361-366. Takashima S & Tanaka K (1978) Development of cerebrovascular architecture and its relationship to periventricular leukomalacia. Arch Neurol 35: 11-16. Tauscher MK, Berg D, Brockmann M, Seidenspinner S, Speer CP & Groneck P (2003) Association of histologic chorioamnionitis, increased levels of cord blood cytokines, and intracerebral hemorrhage in preterm neonates. Biol Neonate 83: 166-170. Taylor GA (1995) Effect of germinal matrix hemorrhage on terminal vein position and patency. Pediatr Radiol 25 Suppl 1: S37-S40. Thompson RS & Trudinger BJ (1990) Doppler waveform pulsatility index and resistance, pressure and flow in the umbilical placental circulation: an investigation using a mathematical model. Ultrasound Med Biol 16: 449-458. Todros T, Sciarrone A, Piccoli E, Guiot C, Kaufmann P & Kingdom J (1999) Umbilical Doppler waveforms and placental villous angiogenesis in pregnancies complicated by fetal growth restriction. Obstet Gynecol 93: 499-503. Tolsa CB, Zimine S, Warfield SK, Freschi M, Sancho RA, Lazeyras F, Hanquinet S, Pfizenmaier M & Hüppi PS (2004) Early alteration of structural and functional brain development in premature infants born with intrauterine growth restriction. Pediatr Res56: 132-138. Trudinger BJ, Giles WB & Cook CM (1985) Uteroplacental blood flow velocity-time waveforms in normal and complicated pregnancy. Br J Obstet Gynaecol 92: 39-45. Trudinger BJ, Cook CM, Giles WB, Ng S, Fong E, Connelly A & Wilcox W (1991) Fetal umbilical artery velocity waveforms and subsequent neonatal outcome. Br J Obstet Gynaecol 98: 378-384. Trudinger B, Wang J, Athayde N, Beutler L & Wang X (2002) Association of umbilical placental vascular disease with fetal acute inflammatory cytokine response. J Soc Gynecol Investig 9: 152-157. Tsuji M, Saul JP, du PA, Eichenwald E, Sobh J, Crocker R & Volpe JJ (2000) Cerebral intravascular oxygenation correlates with mean arterial pressure in critically ill premature infants. Pediatrics 106: 625-632. Valkama AM, Pääkkö EL, Vainionpää LK, Lanning PF, Ilkko EA & Koivisto ME (2000) Magnetic resonance imaging at term and neuromotor outcome in preterm infants. Acta Paediatr 89: 348355. Valkama AM, Tolonen EU, Kerttula LI, Pääkkö EL, Vainionpää LK & Koivisto ME (2001) Brainstem size and function at term age in relation to later neurosensory disability in high-risk, preterm infants. Acta Paediatr 90: 909-915. van der Knaap MS & Valk J (1990) MR imaging of the various stages of normal myelination during the first year of life. Neuroradiology 31: 459-470. van Gelderen P, de Vleeschouwer MH, DesPres D, Pekar J, van Zijl PC & Moonen CT (1994) Water diffusion and acute stroke. Magn Reson Med 31: 154-163. Van Splunder P, Huisman TW, DeRidder MA & Wladimiroff JW (1996) Fetal venous and arterial flow velocity wave forms between eight and twenty weeks of gestation. Pediatr Res 40: 158162.

99 Vergani P, Patane L, Doria P, Borroni C, Cappellini A, Pezzullo JC & Ghidini A (2000) Risk factors for neonatal intraventricular haemorrhage in spontaneous prematurity at 32 weeks gestation or less. Placenta 21: 402-407. Vergani P, Locatelli A, Doria V, Assi F, Paterlini G, Pezzullo JC & Ghidini A (2004) Intraventricular hemorrhage and periventricular leukomalacia in preterm infants. Obstet Gynecol 104: 225-231. Verma U, Tejani N, Klein S, Reale MR, Beneck D, Figueroa R & Visintainer P (1997) Obstetric antecedents of intraventricular hemorrhage and periventricular leukomalacia in the low-birthweight neonate. Am J Obstet Gynecol 176: 275-281. Vexler ZS & Ferriero DM (2001) Molecular and biochemical mechanisms of perinatal brain injury. Semin Neonatol 6: 99-108. Volpe JJ (1995) Neurology of the newborn. W.B. Saunders Company, Philadelphia, 3-92, 291-299, 403-463. Volpe JJ (1997) Brain injury in the premature infant. Neuropathology, clinical aspects, pathogenesis, and prevention. Clin Perinatol 24: 567-587. Volpe JJ (2001) Neurobiology of periventricular leukomalacia in the premature infant. Pediatr Res 50: 553-562. Vossbeck S, de Camargo OK, Grab D, Bode H & Pohlandt F (2001) Neonatal and neurodevelopmental outcome in infants born before 30 weeks of gestation with absent or reversed end-diastolic flow velocities in the umbilical artery. Eur J Pediatr 160: 128-134. Vuori E & Gissler M (2004) Statistical summary. National Research and Development Center for Welfare and Health 26. Wallin LA, Rosenfeld CR, Laptook AR, Maravilla AM, Strand C, Campbell N, Dowling S & Lasky RE (1990) Neonatal intracranial hemorrhage: II. Risk factor analysis in an inborn population. Early Hum Dev 23: 129-137. Wang X, Athayde N & Trudinger B (2002) Endothelial cell expression of adhesion molecules is induced by fetal plasma from pregnancies with umbilical placental vascular disease. BJOG 109: 770-777. Ward RM & Beachy JC (2003) Neonatal complications following preterm birth. BJOG 110 Suppl 20: 8-16. Weatherstone KB & Rich EA (1989) Tumor necrosis factor/cachectin and interleukin-1 secretion by cord blood monocytes from premature and term neonates. Pediatr Res 25: 342-346. Whitelaw A (2001) Intraventricular haemorrhage and posthaemorrhagic hydrocephalus: pathogenesis, prevention and future interventions. Semin Neonatol 6: 135-146. Wilson-Costello D, Borawski E, Friedman H, Redline R, Fanaroff AA & Hack M (1998) Perinatal correlates of cerebral palsy and other neurologic impairment among very low birth weight children. Pediatrics 102: 315-322. Wiswell TE, Graziani LJ, Kornhauser MS, Stanley C, Merton DA, McKee L & Spitzer AR (1996) Effects of hypocarbia on the development of cystic periventricular leukomalacia in premature infants treated with high-frequency jet ventilation. Pediatrics 98: 918-924. Witt KA, Mark KS, Hom S & Davis TP (2003) Effects of hypoxia-reoxygenation on rat bloodbrain barrier permeability and tight junctional protein expression. Am J Physiol Heart Circ Physiol 285: H2820-H2831. Wladimiroff JW, Tonge HM & Stewart PA (1986) Doppler ultrasound assessment of cerebral blood flow in the human fetus. Br J Obstet Gynaecol 93: 471-475. Wladimiroff JW, vd Wijngaard JA, Degani S, Noordam MJ, van Eyck J & Tonge HM (1987) Cerebral and umbilical arterial blood flow velocity waveforms in normal and growth-retarded pregnancies. Obstet Gynecol 69: 705-709.

100 Wood NS, Marlow N, Costeloe K, Gibson AT & Wilkinson AR (2000) Neurologic and developmental disability after extremely preterm birth. EPICure Study Group. N Engl J Med 343: 378-384. Wu YW & Colford JM, Jr (2000) Chorioamnionitis as a risk factor for cerebral palsy: A metaanalysis. JAMA 284: 1417-1424. Wu YW (2002) Systematic review of chorioamnionitis and cerebral palsy. Ment Retard Dev Disabil Res Rev 8: 25-29. Yakolev PI & Lecours AR (1967) The myelogenetic cycles of regional maturation of the brain. In: Minkowski A (ed) Regional development of the brain in early life. Blackwell Scientific Publications, Oxford and Edinburgh, 3-70. Yanowitz TD, Jordan JA, Gilmour CH, Towbin R, Bowen A, Roberts JM & Brozanski BS (2002) Hemodynamic disturbances in premature infants born after chorioamnionitis: association with cord blood cytokine concentrations. Pediatr Res 51: 310-316. Yoon BH, Romero R, Kim CJ, Jun JK, Gomez R, Choi JH & Syn HC (1995) Amniotic fluid interleukin-6: a sensitive test for antenatal diagnosis of acute inflammatory lesions of preterm placenta and prediction of perinatal morbidity. Am J Obstet Gynecol 172: 960-970. Yoon BH, Romero R, Yang SH, Jun JK, Kim IO, Choi JH & Syn HC (1996) Interleukin-6 concentrations in umbilical cord plasma are elevated in neonates with white matter lesions associated with periventricular leukomalacia. Am J Obstet Gynecol 174: 1433-1440. Yoon BH, Jun JK, Romero R, Park KH, Gomez R, Choi JH & Kim IO (1997a) Amniotic fluid inflammatory cytokines (interleukin-6, interleukin-1beta, and tumor necrosis factor-alpha), neonatal brain white matter lesions, and cerebral palsy. Am J Obstet Gynecol 177: 19-26. Yoon BH, Kim CJ, Romero R, Jun JK, Park KH, Choi ST & Chi JG (1997b) Experimentally induced intrauterine infection causes fetal brain white matter lesions in rabbits. Am J Obstet Gynecol 177: 797-802. Yoon BH, Romero R, Kim CJ, Koo JN, Choe G, Syn HC & Chi JG (1997c) High expression of tumor necrosis factor-alpha and interleukin-6 in periventricular leukomalacia. Am J Obstet Gynecol 177: 406-411. Yoon BH, Romero R, Kim KS, Park JS, Ki SH, Kim BI & Jun JK (1999) A systemic fetal inflammatory response and the development of bronchopulmonary dysplasia. Am J Obstet Gynecol 181: 773-779. Yoon BH, Romero R, Park JS, Kim CJ, Kim SH, Choi JH & Han TR (2000) Fetal exposure to an intra-amniotic inflammation and the development of cerebral palsy at the age of three years. Am J Obstet Gynecol 182: 675-681. Yoon BH, Romero R, Moon JB, Shim SS, Kim M, Kim G & Jun JK (2001) Clinical significance of intra-amniotic inflammation in patients with preterm labor and intact membranes. Am J Obstet Gynecol 185: 1130-1136. Yoshioka A, Bacskai B & Pleasure D (1996) Pathophysiology of oligodendroglial excitotoxicity. J Neurosci Res 46: 427-437. Zelenaia O, Schlag BD, Gochenauer GE, Ganel R, Song W, Beesley JS, Grinspan JB, Rothstein JD & Robinson MB (2000) Epidermal growth factor receptor agonists increase expression of glutamate transporter GLT-1 in astrocytes through pathways dependent on phosphatidylinositol 3-kinase and transcription factor NF-kappaB. Mol Pharmacol 57: 667-678. Zupan V, Gonzalez P, Lacaze-Masmonteil T, Boithias C, d'Allest AM, Dehan M & Gabilan JC (1996) Periventricular leukomalacia: risk factors revisited. Dev Med Child Neurol 38: 10611067.