Radiological Findings in Children with Acute Pneumonia: Age More Important Than Infectious Agent

ORIGINAL ARTICLE ACTA RADIOLOGICA Radiological Findings in Children with Acute Pneumonia: Age More Important Than Infectious Agent H. WAHLGREN, W. M...
31 downloads 0 Views 171KB Size
ORIGINAL ARTICLE

ACTA RADIOLOGICA

Radiological Findings in Children with Acute Pneumonia: Age More Important Than Infectious Agent H. WAHLGREN, W. MORTENSSON, M. ERIKSSON, Y. FINKEL, M. FORSGREN & M. LEINONEN Department of Pediatric Radiology and Department of Pediatrics, Astrid Lindgren’s Children’s Hospital, Karolinska Hospital, Stockholm, Sweden; Department of Clinical Virology, Karolinska Institute, Stockholm, Sweden; National Public Health Institute, Oulu, Finland Wahlgren H, Mortensson W, Eriksson M, Finkel Y, Forsgren M, Leinonen M. Radiological findings in children with acute pneumonia: age more important than infectious agent. Acta Radiol 2005;46:431–436. Purpose: To evaluate whether radiological findings and healing time in children with pneumonia are correlated to etiologic agent. Material and Methods: A total of 346 children with radiologically verified acute pneumonia, and with accomplished serological tests for bacteria and viruses, were included in the study. Five etiological groups were analysed: children with bacterial etiology only, with viral etiology only, with mixed bacterial and viral etiology, with Mycoplasma only, and children with no etiology. Results: The chest films of each etiological group were analysed and the findings were correlated to the children’s age. The radiological findings did not differ between the etiological groups. Radiological findings correlated significantly with the patient’s age. The radiological healing frequency at check-up X-ray was found to be significantly lower in children with mixed bacterial and viral etiology compared to children in each of the other groups and to the material as a whole. Conclusion: Conclusions about the etiology could not be drawn from the chest X-ray findings. Key words: Pediatrics; lung; infection; chest X-ray; pneumonia Hedwig Wahlgren, Astrid Lindgren’s Children’s Hospital, Department of Pediatric Radiology, Karolinska University Hospital, Solna, SE-171 76 Stockholm, Sweden (e-mail. [email protected]) Accepted for publication 11 February 2005

In 1995, we published (from S:t Go¨ran’s Children’s Hospital) the results of a study of 479 children with pneumonia; we compared radiological findings in viral and non-viral etiological groups (23). At the time, we did not have access to reliable methods for diagnosing bacterial infection – a problem we shared with most studies published on the subject. Since then, methods for serologic diagnosis of bacterial infections have become available, and we now have the possibility of analyzing the frozen sera of these children for antibodies against some of the common respiratory bacteria. This study aimed at examining whether chest X-ray can be used to predict the etiology of acute pneumonia in children and guide the choice of treament. In addition, we studied the influence of etiologic agent on radiological healing.

Material and Methods This presentation is part of a major study (in manuscript) with the aim of describing different aspects of childhood pneumonia. The study included 715 pneumonic episodes in children 0–15 years who came to the emergency department at S:t Go¨ran’s Children’s Hospital during a period of 4 years. The children had symptoms suggesting acute pneumonia; a chest X-ray was taken to confirm the diagnosis. At the initial visit, a blood sample was taken for serology. Blood cultures were taken from all hospitalized children who were prescribed parenteral antibiotics. Nasopharyngeal samples were taken for rapid viral detection in children less than 1 year of age during RSV (respiratory syncytial virus) season. Treatment was given by the physician DOI 10.1080/02841850510021238

# 2005 Taylor & Francis

432

H. Wahlgren et al.

in charge of the patient. A follow-up appointment was given after 4–6 weeks, at which time a second chest X-ray was taken, along with a new blood sample for serology. Serological studies were done on paired sera for virus and bacteria (14). The following viral etiologies were included: RSV, influenza A and B, parainfluenza 1, 2, and 3, and herpes virus. The following bacterial etiologies were included: Pneumococcus, Moxharella catharalis, Hemophilus influenzae, Mycoplasma and Chlamydia. Complete information and a positive chest X-ray were available for 346 children, who constitute the present study. The children were allotted to one of five groups according to the microbial etiology: Group B+ V+ (n567) Group B+ V2 (n572) Group B2 V+ (n5111) Group B2 V2 (n588) Group MP only (n58)

Both bacterial and viral tests positive Bacterial etiology positive, viral etiology not found Viral etiology positive, bacterial etiology not found Neither bacterial nor viral etiology found Mucoplasma with no other etiology found

Within each etiology group the children were divided into three age groups: 0–1 years, 2–6 years and >6 years. A chest X-ray was taken on the day of consultation, and findings consistent with acute lower respiratory tract infection were obligatory for being included in the study. A check-up chest Xray was taken after a mean of 6 weeks. The radiographs were re-evaluated by two experienced pediatric radiologists and the diagnosis was made in consensus. The findings were described as follows: presence of infiltrates, type of infiltrates, whether predominantly interstitial (int inf), alveolar (alv inf), or mixed (mix inf), and presence of hyperinflation (air trapping). In the re-evaluation,

Table 1. Etiology pattern related to age. Etiology is described as mixed bacterial and viral infection (B+ V+), bacterial infection but no viral infection (B+ V2), viral infection but no bacterial infection (B2 V+), and no signs of either bacterial or viral infection (B2 V2). Mucoplasma infection enters as a separate group (MP only)

95 children (95/715; 13%) initially diagnosed as suffering from pneumonia were found to have normal chest X-rays and were excluded from the study. Statistical analysis Chi-square test was used to analyse differences between proportions with a 0.05 level of significance. The prevalence of different agents was analysed as a function of radiographic findings and age by fitting a logistic regression model. Results The most common viral etiology was RSV (29%) followed by parainfluenza, adenovirus, and influenza (each approximately 10%). The most common bacterial etiology was pneumococci (60%), followed by Hemophilus influenzae. Mycoplasma was found in 6% of the patients and Chlamydia in 2%. When comparing age with etiology, we found a strong correlation (pv0.000) between low age and viral infection, and between higher age and bacterial infection (Table 1). Comparing the age groups, there was a significant difference in radiological findings, with more interstitial infiltrations and hyperinflation in the younger children, and more alveolar infiltrations in the older children (p50.024) (Table 2). When comparing the radiological findings in the different etiological groups, we analysed each age group separately. In each age group, the distribution of parenchymal changes and hyperinflation in relation to etiology was completely random (Table 3A–C). The radiological healing at check-up X-ray was found to be significantly lower in group B+ V+, 54%, while the other groups had healed in about the same

Table 2. Radiological findings related to age. The parenchymal infiltrates are described as interstitial (int inf), alveolar (alv inf), or mixed interstitial and alveolar (mix inf). The presence of general hyperinflation is noted separately, and is found with, as well as without, infiltrates Age (years) 0–1

Age (years) 0–1 +

+

B V B+ V2 B2 V+ B2 V2 MP only Total

40 20 62 44

(24%) (12%) (37%) (27%)

166 (100%)

Acta Radiol 2005 (4)

2–5 20 26 41 32

(17%) (22%) (34%) (27%)

119 (100%)

w6 7 26 8 12 8 61

(11%) ( 43%) (13%) (20%) (13%) (100%)

Hyperinflation with or without infiltrates Hyperinflation without infiltrates Int inf Alv inf Mix inf Total

87 (52%) 5 (3%) 65 68 28 166

2–5

w6

45 (38%) 10 (19%) 5 (4%)

(39%) 45 (41%) 48 (17%) 21 (100%) 119

(38%) (40%) (18%) (100%)

2 (3%) 15 40 4 61

(25%) (66%) (6%) (100%)

Radiological Findings in Children with Acute Pneumonia

433

Table 3. A–C. Etiological pattern related to X-ray findings, presented for each age group separately. Etiology is described as mixed bacterial and viral infection (B+ V+), bacterial infection but no viral infection (B+ V2), viral infection but no bacterial infection (B2 V+), and no signs of either bacterial or viral infection (B2 V2). Mucoplasma infection without other infectious agent is noted separately. The parenchymal infiltrates are described as interstitial (int inf), alveolar (alv inf), or mixed interstitial and alveolar (mix inf). The presence of general hyperinflation is noted separately, and is found together with, as well as without, infiltrates (hyp w/wo infiltrates) Int inf

Alv inf

A. 0–1 years (n5166) 14 (21%) B+ V+ 7 (11%) B+ V2 26 (40%) B 2 V+ 2 2 18 (28%) B V Total 65 (100%) B. 2–5 years (n5119) 9 (20%) B+ V+ 10 (22%) B+ V2 2 + 13 (29%) B V 13 (29%) B 2 V2 Total 45 (100%) C. 6 years and older (n553) (counting mucoplasma the total is 61) 2 (14%) B+ V+ 6 (43%) B+ V2 2 (14%) B 2 V+ 2 2 4 (29%) B V MP only 1 (7%) Total 15

Mix inf

Hyp w/wo infiltrates

21 10 24 13 68

(31%) (15%) (35%) (19%) (100%)

4 3 9 12 28

(14%) (11%) (32%) (43%) (100%)

18 10 36 23 87

(21%) (12%) (41%) (26%) (100%)

7 12 19 10 48

(15%) (25%) (39%) (21%) (100%)

3 2 8 8 21

(14%) (10%) (38%) (38%) (100%)

12 9 14 10 45

(27%) (20 %) (31%) (22%) (100%)

3 18 5 8 6 40

(8%) (45%) (12%) (20%) (15%)

2 4 1 3

(20%) (40%) (10%) (30%)

1 (25%) 1 (25%) 1 (25%) – 1 (25%) 4

10

A. Hyp without infiltrates 5 children. B. Hyp without infiltrates 5 children. C. Hyp without infiltrates 2 children.

proportion as the material as a whole, just above 70%. Children with Mucoplasma infection only, however, were found to have healed in 100%. The mean length of the check-up time did not differ significantly between the etiology groups (Table 4). Neither was there any significant difference with respect to healing time in the three age groups. The healing frequency in the different age groups was found to be around 70%, not differing in any significant way from the material as a whole. This holds also for the mean length of the check-up time, with the exception of the oldest children, above 6 years, who for some reason had a mean check-up time significantly shorter than the other groups (data not shown). Discussion Several studies and surveys of radiologic findings in childhood pneumonia published during the 1970s and 1980s propose that it is possible to differentiate

between bacterial and viral pneumonia with the help of chest X-ray (6, 16, 18, 19, 21, 24). A common problem in these articles and studies is the lack of reliable microbiological diagnosis, especially bacterial, and the radiological findings are mostly related to clinical signs and symptoms presumed to reflect etiology. Since the end of the 1980s, a number of studies have been published contradicting this thesis (1, 3, 4, 15, 17, 20, 22), most of them comparing the radiologic picture to the etiologic diagnosis based on viral serology and culture, and urinary antigen analyses for bacteria. These authors have found no correlation between radiological picture and etiology, and thus concluded that chest X-ray is of little useful help in the decision whether the infection is bacterial or viral. In the 1990s, more reliable methods for microbiological etiologic diagnosis were developed (14), and these have facilitated the comparison between X-ray appearance and etiology. In a study of 61 children in 1993, KORPPI et al. (10) compared the type of

Table 4. Radiological healing and mean check-up time related to etiology. Healing (healed) means complete radiological resolution. Etiology is described as mixed bacterial and viral infection (B+ V+), bacterial infection but no viral infection (B+ V2), viral infection but no bacterial infection (B2 V+), and no signs of either bacterial or viral infection (B2 V2). Mucoplasma infection enters as a separate group (MP only). Mean check-up time as number of days (check-up) and standard deviation (SD) is noted for each etiology group

n Healed Check-up SD

B2 V+

B+ V2

B+ V+

B 2 V2

MP only

Total

111 83 (75%) 43 d 16 d

72 52 (72%) 42 d 14 d

67 36 (54%) 44 d 33 d

88 67 (76%) 41 d 14 d

8 8 (100%) 41 d 4d

346 246 (71%)

Acta Radiol 2005 (4)

434

H. Wahlgren et al.

infiltrates, interstitial or alveolar, with microbiologically verified etiology, viral and/or bacterial. A correlation was found between alveolar infiltrates and bacterial etiology, as well as between interstitial infiltrates and viral etiology. It was concluded that the presence of alveolar infiltrates was a good indicator of bacterial etiology, and should suggest antibiotic treatment. In a later study (11) dealing only with the 28 children with pneumococcal pneumonia, further data were published suggesting a significant correlation between radiological pattern and inflammatory activity, as reflected by CRP and ESR. These inflammatory parameters in turn were associated with the serological results, being higher in pneumococcal pneumonias than in viral pneumonias. In 1996, NOHYNEK (14), comparing two groups of children with indication of bacterial (n548) versus viral (n527) etiology and chest X-ray, found a higher probability for bacterial infection with a dense, unisegmental infiltrate, and a higher probability for viral infection with central or peribronchial infiltrates, though the positive predictive value of the radiological interpretation of etiology was found to be poor. Our findings, however, seem to contradict these results. In a study published from our centre in 1995 of 479 children with pneumonia, we were unable to find any correlation between the radiological findings and the presence or absence of viral etiology, concluded from viral culture and/or serology (23). At the time, we did not have access to reliable methods for bacterial diagnosis. In the present study, investigating the same material, adding bacterial serology in 346 of the children, we still find no significant correlation between the radiological picture and the type of etiological agent, i.e. it was not possible from the chest X-ray to conclude whether the child had viral infection, bacterial infection, or mixed infection. A common problem in the studies cited above of chest X-ray appearance is the fact that relatively few children were included, i.e. less than 100 in all but one study. In a rather extensive study (13), NascimientoCarvalho and her group investigated 1456 children with radiologically verified pneumonia and blood culture performed, relating the chest X-ray findings (alveolar versus interstitial infiltrates) to blood culture results. They found no correlation between the radiological findings and the blood culture tests, results that seem to be in accordance with our findings. However, the culture yield was only 2–3% in both X-ray groups, so even if the difference was significant, the actual number of children with a verified bacteriological diagnosis was very small. Acta Radiol 2005 (4)

It is well known that the lower respiratory tract of the infant differs from that of the adult (1). Consequently, this difference could be expected to influence the X-ray picture. The effect of the child’s age on the radiological pattern, though, has seldom been discussed. A correlation between air-trapping and young age (4, 6, 9) has been discussed particularly for RSV virus infection. In the above-mentioned study of 61 children (10), where it was noted that alveolar infiltrates were more common with bacterial etiology, it was also noted that alveolar infiltrates were more common with increasing age. In our material, we found not only a highly significant correlation between young age and air-trapping, but also a higher frequency of interstitial infiltrates in younger children, and of alveolar infiltrates in the older ones. The fact that the relative frequencies of different etiologic agents varies between age groups is well known and documented (8, 12, 14). Accordingly, we found that the etiologic pattern varied with age, with a higher frequency of viral infections among the youngest children, and more bacterial infections in the older groups. The correlation between young age and interstitial infiltrates and hyperinflation, along with the higher frequency of virus infection in younger children and bacterial infection in older children, can perhaps explain the widespread feeling that interstitial infiltrates reflect viral infection while alveolar infiltrates reflect bacterial infection. To investigate the roles of both age and etiology, a larger group than previously investigated was needed to demonstrate age as the important determinant. There are few studies describing the resolution of radiological changes in acute pneumonia in the normal child. In a material of 70 children, GROSSMAN found that after 3–4 weeks the radiologic changes had resolved in 80%, and after 3 months all of the children had normal X-rays (7). GIBSON found that in 55 of 59 children with pneumonia the X-ray changes had resolved completely within 3–4 weeks, and in the rest at repeat check-up. He suggests that in uncomplicated childhood pneumonias follow-up radiographs are unnecessary (5). In an article in 1993 relating to adult pneumonias, CORLEY refers to a slower resolution of chest X-ray changes in infections resulting from Pneumococci and Legionella, while Mycoplasma is reported to heal more quickly (2). The relation between radiological healing and etiology is poorly investigated in children. In our study, we found that the proportion of children in whom the radiological changes had resolved at

Radiological Findings in Children with Acute Pneumonia

435

Fig. 1. AP and lateral views of 6-month-old baby with respiratory syncytial virus (RSV) infection. Bilateral perihilar interstitial infiltrates and generalized hyperinflation with pronounced flattening of diaphragma seen in lateral view.

Fig. 2. AP and lateral views of a 4-year-old boy with Hemophilus infection. Dense alveolar infiltrate in the right upper lobe. No hyperinflation.

check-up chest X-ray after 4–6 weeks did not differ significantly between the etiological groups B+ V2, B2 V+ and B2 V2, varying between 72% and 76% (Table 4). Group B+ V+, however, with mixed bacterial/viral infection, had a much lower frequency of healing, 54%, while all the children with Mucoplasma infection only had healed. The mean time between the initial chest X-ray and the check-up did not differ much between groups (Table 4); note, however, that in children with mixed infection, group B+ V+, even though the mean check-up time was within the same range as in the other groups, the variation was much wider. In this group, there was a greater number of children

that had a very early check-up, possibly beacause of a more serious clinical condition. In our material, the group of children with negative etiological diagnosis for both virus and bacteria, the B2 V2 group, is problematic. These children did not differ to any significant degree from the other groups in X-ray pattern or in radiological healing time. Although the children in this group had pneumonia, we could not find the cause of infection with the tests performed. One possible explanation is that the infection was caused by bacteria and virus not investigated in our study. Another source of uncertainty could be that serological methods in general are known to be Acta Radiol 2005 (4)

436

H. Wahlgren et al.

less sensitive in the youngest, immunologically immature children. However, the proportion of B2 V2 did not differ significantly between the age groups, amounting to 27% in the two youngest groups, and to 28% in children above 6 years of age (Table 3). In the present study, we excluded children with pertussis infection and no other proven agent. In conclusion, we found that the X-ray picture is a poor predictor of etiology in acute pneumonia in children. Hence, the decision whether or not to treat with antibiotics should be based primarily on clinical grounds. Persisting pathological findings were seen in check-up chest X-ray after 4–6 weeks in about half of the children with mixed bacterial/viral infections, and about one-fourth of the children in the other groups. No difference in healing time could be demonstrated between children with viral and children with bacterial infection. References 1. Bettenay FAL, de Campo JF, McCrossin DB. Differentiating bacterial from viral pneumonia in children. Pediatr Radiol 1988;18:453–4. 2. Corley DE, Winterbauer RH. Infectious diseases that result in slowly resolving and chronic pneumonia. Semin Resp Inf 1993;8:3–13. 3. Courtoy I, Lande AE, Turner RB. Accuracy of radiographic differentiation of bacterial from non-bacterial pneumonia. Clin Pediatr 1989;28:261–4. 4. Friis B, Eiken A, Hornsteth, Jensen A. Chest X-ray appearances in pneumonia and bronchiolitis. Acta Pediatr Scand 1990;79:219–25. 5. Gibson NA, Hollman AS, Paton JY. Value of radiological follow up of childhood pneumonia. Br Med J 1993;307:1117. 6. Griscom NT, Wohl ME, Kirkpatrick JA Jr. Lower respiratory/infection, how infants differ from adults. Radiol Clin North Am 1978;15:367–86. 7. Grossman L, Wald E, Prasanna N, Papiez J. Roentgenographic follow-up of acute pneumonia in children. Pediatrics 1979;63:30–1. 8. Khamapirad T, Glezen P. Clinical and radiographic assessment of acute lower respiratory tract disease in infants and children. Semin Resp Inf 1987;2:130–44.

Acta Radiol 2005 (4)

9. Kirkpatrick JA. Pneumonia in children as it differs from adult pneumonia. Semin Roentgenol 1980;15:96–103. 10. Korppi M, Kiekara O, Heiskanen-Kosma T, Soimakallio S. Comparison of radiological findings and microbial aetiology of childhood pneumonia. Acta Paediatr 1993;82:360–3. 11. Korppi M, Leinonen M. Pneumococcal pneumonia in children; new data from circulating immune complexes. Eur J Pediatr 1996;156:341–2. 12. Murphy TF, Henderson FW, Clyde WA, Collier AM, Denny FW. Pneumonia; an eleven-year study in pediatric practice. Am J Epidemiol 1981;113:12–21. 13. Nascimiento-Carvalho MC, Alves NN, Athayde LA, Caldas RM, Barberina MGMA, Duarte J, et al. Is there any association of a specific chest X-ray pattern and bacteremia in children with pneumonia? J Trop Pediatr 2002;48:253–4. 14. Nohynek H. Acute lower respiratory tract infection in children. National Public Health Institute, University of Helsinki, 1996: 22–23, 30, 39–40, 48. 15. Redd SC, Patrick E. Comparison of the clinical and radiologic diagnosis of pediatric pneumonia. Trans R Soc Trop Med Hyg 1994;88:307–10. 16. Scanlon GT, Unger JD. The radiology of viral pneumonias. Rad Clin North Am 1973;9:317–37. 17. Sunakorn P, Chunchit L, Niltawat S, Wangweerawong M, Jacobs R. Epidemiology of acute respiratory infections in young children from Thailand. Pediatr Infect Dis J 1990;9:873–7. 18. Swischuk LE, Hayden CK. Viral vs bacterial pulmonary infections in children (Is roentgenographic differentiation possible?). Pediatr Radiol 1986;16:278–84. 19. Wildin SR, Chonmaitree T, Swischuk LE. Roentgenographic features of common pediatric viral respiratory tract infection. Am J Dis Child 1988;142:43–6. 20. Teague WG. Approach to pneumonia in infants, children and adolescents. Immunol Allergy Clin North Am 1993;13:159–67. 21. Tew J, Calenoff L, Berlo¨in BS. Bacterial or nonbacterial pneumonia: accuracy of radiographic diagnosis. Diagn Radiol 1977;124:607–12. 22. Turner RB, Lande AE, Chase P, Hilton N, Weinberg D. Pneumonia in pediatric outpatients: cause and clinical manifestations. J Pediatr 1987;11:194–9. 23. Wahlgren H, Mortensson W, Eriksson M, Finkel Y, Forsgren M. Radiographic patterns and viral studies in childhood pneumonia at various ages. Pediatr Radiol 1995;25:627–30. 24. Zukin DD, Hoffman JR, Cleveland RH, Kushner DC, Herman TE. Correlation of pulmonary signs and symptoms with chest radiographs in the pediatric age group. Ann Emerg Med 1986;15:792–6.

Suggest Documents