Jann-Yuan Wang, MD, Yih-Leong Chang, MD, Li-Na Lee, MD, PhD, Jen-Hau Chen, MD, MPH, Jih-Luh Tang, MD, Pan-Chyr Yang, MD, PhD, and Yung-Chie Lee, MD, PhD Departments of Internal Medicine, Pathology, Laboratory Medicine, and Surgery, National Taiwan University Hospital and Department of Internal Medicine, Far Eastern Memorial Hospital, Taipei, Taiwan

Background. Diffuse pulmonary infiltrates is the major complication and cause of mortality after bone marrow transplantation. We analyzed the etiologies and prognostic factors in bone marrow recipients with diffuse pulmonary infiltrates and assessed the role of open lung biopsy in managing this complication. Methods. Medical records of patients with diffuse pulmonary infiltrates after bone marrow transplantation were reviewed. Possible prognostic factors were analyzed by multivariate logistic regression. Results. Sixty-eight (20%) of 341 bone marrow recipients had diffuse pulmonary infiltrates and 34 died. Thirty-five underwent open lung biopsy, resulting in therapeutic changes in 22 (63%) and clinical improvement in 16 (46%). The leading diagnoses were idiopathic interstitial pneumonitis (40%) and cytomegalovirus pneumonitis (20%). Cytomegalovirus pneumonitis caused radiographically observable interstitial infiltrates exclusively and was frequently associated with hepatitis. Idiopathic

interstitial pneumonitis resulted in either diffuse ground-glass opacity or interstitial infiltrates. Three (9%) patients had miliary tuberculosis. Respiratory failure (p < 0.001) and acute graft-versus-host disease (p ⴝ 0.016) were the poor prognostic factors. Conclusions. Among bone marrow recipients, we found diffuse pulmonary infiltrates in 20% and a mortality rate of 50%. Idiopathic interstitial pneumonitis and cytomegalovirus pneumonitis were the most common causes and should be suspected in patients with diffuse interstitial infiltrates. In endemic areas, miliary tuberculosis should be suspected in bone marrow recipients with diffuse reticulonodular lesions. Respiratory failure and acute graft-versus-host disease were poor prognostic factors. By establishing a correct diagnosis, open lung biopsy led to treatment changes in about two-thirds of these patients. (Ann Thorac Surg 2004;78:267–72) © 2004 by The Society of Thoracic Surgeons

B

Patients and Methods

one marrow transplantation (BMT), including peripheral blood stem cell transplantation, has become a successful therapy for patients with various hematologic disorders, some solid tumors, and immunodeficient states [1]. Although long-term disease-free survival after BMT could exceed 60%, pulmonary infiltrates, due to either inflammatory or infectious pneumonitis, occur in 40% to 60% of bone marrow recipients. These infiltrates are the major causes of morbidity and mortality in as many as 80% of transplant-related deaths [2, 3]. Because the possible etiologies of pulmonary infiltrates in immunocompromised patients are numerous and may coexist, open lung biopsy was previously recommended to make a definitive diagnosis in these seriously ill patients and thereby to select appropriate treatment [4 –7]. However, studies exploring the prognostic factors in BMT patients with diffuse pulmonary infiltrates (DPI) are few [8 –10]. In addition, open lung biopsy does not appear to improve overall survival [5, 6, 11]. Thus we conducted this study to investigate the etiologies, clinical characteristics, and prognostic factors of DPI after BMT. The effect of open lung biopsy was also evaluated. Accepted for publication Jan 7, 2004. Address reprint requests to Dr Yung-Chie Lee, Department of Surgery, National Taiwan University Hospital, No 7, Chung-Shan South Rd, Taipei, Taiwan; e-mail: [email protected].

© 2004 by The Society of Thoracic Surgeons Published by Elsevier Inc

This study was conducted in the National Taiwan University Hospital, a tertiary-care referral center with 2,000 beds, by review of the medical records and chest images of 341 patients with hematologic diseases who underwent BMT between July 1995 and June 2001. Patients in whom DPI developed were included, with DPI defined as the presence of infiltrates over all four quadrants of the lung fields as seen on chest radiograph. The date of onset was defined as the earliest date when image study demonstrated DPI. All chest images were reviewed by 1 radiologist and 1 chest specialist. If discrepancies were noted between their interpretations, the image was further reviewed by 1 chest specialist blinded to the results. Patients in whom DPI developed after septic shock or at the terminal stage of other diseases were excluded from this study. Prophylactic ganciclovir against cytomegalovirus (CMV) for allogeneic recipients and trimethoprimsulfamethoxazole against Pneumocystis carinii for all recipients were prescribed during transplantation if a patient’s leukocyte count was higher than 2,000 k/␮L and platelet count above 50 k/␮L. All allogeneic bone marrow recipients received cyclosporin-containing regimens for preventing graft-versus-host disease (GVHD). Once DPI developed, computed tomography of the chest, bacteriology studies, serology tests for Mycoplasma 0003-4975/04/$30.00 doi:10.1016/j.athoracsur.2004.03.002

GENERAL THORACIC

Diffuse Pulmonary Infiltrates After Bone Marrow Transplantation: The Role of Open Lung Biopsy

268

WANG ET AL DIFFUSE PULMONARY INFILTRATES AFTER BMT

GENERAL THORACIC

Abbreviations and Acronyms BMT ⫽ bone marrow transplantation CMV ⫽ cytomegalovirus DPI ⫽ diffuse pulmonary infiltrates Fio2 ⫽ fractional concentration of inspiratory oxygen GVHD ⫽ graft-versus-host disease IIP ⫽ idiopathic interstitial pneumonitis MODS ⫽ multiple organ dysfunction score Pao2 ⫽ partial pressure of arterial oxygen Paco2 ⫽ partial pressure of arterial carbon dioxide PCP ⫽ Pneumocystis carinii pneumonia PTLD ⫽ posttransplant lymphoproliferative disorder

and Chlamydia, antigen assays for Legionella and Cryptococcus, microscopic examination of sputum for Pneumocystis carinii and assays for CMV antigenemia or viremia were performed. Patients underwent open lung biopsy when diagnosis was not reached by the above-mentioned studies (early biopsy). Alternatively, if the patient’s condition did not improve within 3 days with empirical therapy, open lung biopsy was then arranged (late biopsy). Bronchoscopy, including transbronchial biopsy and bronchoalveolar lavage, was performed if the patient was too weak to undergo general anesthesia or refused open lung biopsy. In truly diffuse processes, the lingula was biopsied. In cases with nonuniform involvement, the biopsy was taken from the most involved area. The biopsy was performed by video-assisted thoracoscopy if the patient could tolerate single-lung ventilation, otherwise by minithoracotomy. All biopsy specimens were reviewed by 2 pulmonary pathologists. Therapeutic changes included use, discontinuation, or change of antimicrobial or immunosuppressive agents. The response of the treatment specific to the pathologic diagnosis was assessed according to the patient’s symptoms and laboratory and radiographic findings as assessed by primary care physicians, including a specialist in BMT and a specialist in critical care. Cytomegalovirus pneumonitis was diagnosed if the histologic examination showed interstitial pneumonitis characterized by intranuclear or cytoplasmic inclusion bodies with confirmed immunohistochemical CMV antibody stain, or a positive CMV culture from the lung biopsy specimen [12]. Active CMV infection was diagnosed by positive CMV antigenemia [13] or viremia [14], or CMV pneumonitis. Idiopathic interstitial pneumonitis (IIP) was diagnosed if the histologic examination revealed diffuse interstitial pneumonitis with some cases resembling diffuse alveolar damage or nonspecific interstitial pneumonia in the absence of an identifiable infectious agent [15].

Statistical Analysis Variables with possible survival impact were analyzed. The variables included age at transplantation, sex, un-

Ann Thorac Surg 2004;78:267–72

derlying hematologic disease and remission status while undergoing BMT, smoking, pulmonary function before BMT, type of BMT (autologous or allogeneic), conditioning regimens [16], timing of the onset of DPI (more than 100 days or 100 or fewer days after BMT), initial multiple organ dysfunction score (MODS) [17], respiratory failure, acute GVHD, and open lung biopsy. Respiratory failure was defined as partial pressure of arterial oxygen (Pao2) of less than 60 mm Hg or partial pressure of arterial carbon dioxide (Paco2) of more than 50 mm Hg while breathing room air, or requiring assisted ventilation [18]. The diagnosis, grading, and treatment of GVHD were determined as reported previously [19]. Logistical regression was applied to identify factors independently associated with mortality. Variables were entered into the multivariate analysis if a significant difference (p ⬍ 0.05) was obtained in the univariate analysis. Intergroup differences were analyzed using either the independent-samples t test (continuous variables) or ␹2 test (categorical variables).

Results From July 1995 through June 2001, pulmonary infiltrates after BMT occurred in 188 (55.1%) of 341 patients. Diffuse pulmonary infiltrates developed in 68 (19.9%) of the 341 patients, including 63 (23.7%) of 266 allogeneic recipients and 5 (6.67%) of 75 autologous recipients. The clinical characteristics of the 68 patients are listed in Table 1. Relapse of underlying disease before the onset of DPI was proved in 5 (7.4%) of the 68 patients. Sputum collected during the course of DPI was cytologically free of malignant cells. Nineteen patients underwent early biopsy and 16 underwent late biopsy. The median interval between BMT and open lung biopsy was 137 days (range 17 to 925 days). At the time of lung biopsy, 10 patients required ventilator support and 7 patients had ratios of Pao2/Fio2 (fractional concentration of inspiratory oxygen) of more than 300. The pathologic diagnoses of the 35 patients who underwent biopsy and their radiographic findings are listed in Table 2. Major discordance between clinical and pathologic diagnoses was noted in 22 (62.9%) patients, including 12 early and 10 late biopsy patients. The results of open lung biopsy led to therapeutic change in 22 (62.9%) patients. Ten early biopsy patients and 6 late biopsy patients responded to the altered treatment in their clinical symptoms and pulmonary infiltrates (␹2 test, p ⫽ 0.500). One other late biopsy patient had transient improvement. Thirty-four (50.0%) of the 68 patients with DPI died, 33 before undergoing biopsy. The cause of death was respiratory failure in 30 (88.2%), brain abscess resulting in fatal brain herniation in 2 (5.88%), and septic shock in 1 (2.94%). Of the 35 patients undergoing biopsy, 1 (2.9%) patient died of massive hemothorax after minithoracotomy refractory to reoperation. The surgical morbidity among the 35 patients undergoing biopsy (n ⫽ 2; 5.7%) included transient wound dehiscence in 1 patient and bleeding that required blood transfusion in another patient. In the 35 biopsy patients, we noted 11 episodes of adverse drug reaction (31.4%), including four cases of

WANG ET AL DIFFUSE PULMONARY INFILTRATES AFTER BMT

Table 1. Characteristics of the 68 Patients With DPI After BMT Characteristic

Survivors Nonsurvivors (n ⫽ 34) (n ⫽ 34)

Age at transplantation, median 31 (1–57) 24 (2–58) (range), years Sex, n (male/female) 19/15 18/16 Onset of DPI after BMT, median 102.5 (16–1,328) 87.5 (7– 884) (range), days Smoking, n 6/28 5/29 (ex-smoker/nonsmoker) Diagnosis (n) Acute myelocytic leukemia 11 8 Chronic myelocytic leukemia 7 7 Acute lymphocytic leukemia 4 9 Non-Hodgkin lymphoma 3 5 Thalassemia major 6 2 Others 3a 3b Cytomegalovirus serostatus (n) Donor ⫹, Recipient ⫹ 24 20 Donor ⫺, Recipient ⫹ 0 4 Donor ⫹, Recipient ⫺ 2 5 Donor ⫺, Recipient ⫺ 7 6 Conditioning drugs Cyclophosphamide ⫹ busulphan 17 14 Cyclophosphamide 12 14 Other 5 6 Total body irradiation (n) 13 20 Graft-versus-host disease (n) Acute 14 25 Chronic 11 6 No 12 6 Timing of the onset of DPI (n) ⱕ100 days 16 18 ⬎100 days 19 15 MODSc at the onset of DPI (n) 0 2 1 1– 4 17 11 5– 8 11 16 9 –12 4 6 Respiratory failured (n) 11 33 a

The diagnoses of the 3 patients were severe aplastic anemia in 2 and b myelodysplastic syndrome in 1. The diagnoses of the 3 patients were c myelodysplastic syndrome in 2 and myelofibrosis in 1. MODS was d noted according to Marshall and colleagues [17]. Respiratory failure was defined as partial pressure of arterial oxygen (Pao2) ⬍ 60 mm Hg or partial pressure of arterial carbon dioxide (Pao2) ⬎ 50 mm Hg while breathing room air, or requiring use of ventilator [18]. BMT ⫽ bone marrow transplantation; DPI ⫽ diffuse pulmonary infiltrates; MODS ⫽ multiple organ dysfunction score.

neutropenia (neutrophil counts ⬍ 1,000 k/␮L on 2 consecutive days), five of renal dysfunction, and two of hepatic dysfunction. However, among the 33 patients without open lung biopsy we noted 18 episodes (54.5%) of adverse drug reaction, including nine cases of neutropenia, seven of renal dysfunction, and two of hepatic dysfunction (␹2 test, p ⫽ 0.054). Four (57.1%) of the 7 CMV pneumonitis patients had abnormal liver function on admission, compared with

269

none of the IIP patients. The onset of DPI in CMV pneumonitis patients was earlier than that among IIP patients (64 ⫾ 48 days compared with 115 ⫾ 93 days). Respiratory failure developed in 5 of the 7 CMV pneumonitis patients and all 5 died, whereas 2 of the 14 patients with IIP died (mortality 71.4% compared with 14.3%). Of the 35 patients who received open lung biopsy, tuberculosis caused DPI in 3 (8.6%) patients with the onset at 124, 584, and 605 days after BMT, respectively. Tuberculosis was also the most common etiology of radiographically diffuse reticulonodular infiltrates (Table 2). Mycobacterium tuberculosis was isolated from the biopsy specimens of all 3 patients. The last patient also had tuberculous hepatitis proved by liver biopsy. Additionally, all 3 of these patients had GVHD. Bronchoscopy was performed in 13 patients, 7 of whom had the procedure before open lung biopsy. Among these 7, bronchoscopy was nondiagnostic in 4. Of the 4 patients, the open lung biopsy revealed IIP in 2, pulmonary tuberculosis in 1, and CMV pneumonitis in the other. Among the 3 for whom bronchoscopy was diagnostic, the bronchoscopic diagnosis was compatible with the result of open lung biopsy in 2 patients with IIP. The bronchoscopy revealed pulmonary candidiasis in the other patient, whose open lung biopsy showed lymphoma relapse. Six patients who did not receive open lung biopsy underwent bronchoscopy, in which 5 were nondiagnostic and the other had a diagnosis of pulmonary aspergillosis. Five (38.5%) of the 13 patients received therapeutic changes after bronchoscopy. Table 3 shows the intubation rate and the number of days on a ventilator in the 6 patients who received only bronchoscopy and the 35 patients who underwent open lung biopsy. The clinical diagnostic categories and outcome of the 33 patients who did not receive open lung biopsy were listed in Table 4. In the univariate analysis, factors significantly associated with outcome (survival or death at discharge) were respiratory failure, acute GVHD, and open lung biopsy. Respiratory failure and acute GVHD were independently associated with mortality in the multivariate analysis (Table 5). Respiratory failure developed in 44 patients with a mean time of 14 days after the onset of respiratory symptoms (range 0 to 107 days). Of them, 19 underwent open lung biopsy. The most common pathologic diagnoses were CMV pneumonitis (5 patients, 26.3%) and IIP (5 patients, 26.3%), with mortality rates of 100% and 0%, respectively. In patients with acute GVHD, the mean interval between BMT and the onset of DPI was 111 days, whereas the mean interval in patients without acute GVHD was 255 days. Of the 39 patients with acute GVHD, active CMV infection was diagnosed in 16 (41%), of whom 13 (81.3%) died. However, only 1 of the 29 patients without acute GVHD had active CMV infection.

Comment Consistent with previous reports [2, 3], our incidence of pulmonary infiltrates after BMT was 55.1% (188 of 341 patients). About one-third of these patients had DPI (68

GENERAL THORACIC

Ann Thorac Surg 2004;78:267–72

270

WANG ET AL DIFFUSE PULMONARY INFILTRATES AFTER BMT

Ann Thorac Surg 2004;78:267–72

Table 2. Disease Mortality and Correlation Between Pathologic Diagnoses and Radiographic Findings of the 35 Patients Receiving Open Lung Biopsy

GENERAL THORACIC

Infectious etiology CMV pneumonitis Tuberculosis PCP Aspergillosis Noninfectious etiology IIP Organizing pneumonia Alveolar hemorrhage Capillaritis Lymphoma relapse PTLD

Number of Cases

Mortality

12/35 (34.3%) 7 3 1 1 23/35 (65.7%) 14 5 1 1 1 1

6 (50.0%) 5 (71.4%) 1 (33.39%) 0 0 7 (30.4%) 2 (14.3%) 1 (20.0%) 1 (100%) 1 (100%) 1 (100%) 1 (100%)

Interstitial Infiltrates With or Without Focal Consolidation (n ⫽ 15)

Diffuse Ground-Glass Opacity (n ⫽ 12)

Multifocal Consolidation (n ⫽ 3)

Diffuse Reticulonodular Pattern (n ⫽ 5)

7 0 0 0

0 0 1 0

0 0 0 1

0 3 0 0

6 1 0 1 0 0

8 2 1 0 0 0

0 2 0 0 0 0

0 0 0 0 1 1

CMV ⫽ cytomegalovirus; IIP ⫽ idiopathic interstitial pneumonitis; lymphoproliferative disorder.

of 188 patients; 36.2%), with a high mortality rate (50%). The results of open lung biopsy caused major therapeutic change in two-thirds and were associated with clinical improvement in nearly half of the biopsy patients. Patients who underwent open lung biopsy had a much shorter length of mechanical ventilation than patients who underwent only bronchoscopy. The most common pathologic diagnoses were IIP and CMV pneumonitis. Idiopathic interstitial pneumonitis had a much lower mortality rate. All of the CMV pneumonitis patients with respiratory failure died. Cytomegalovirus pneumonitis presented with radiographically confirmed interstitial pulmonary infiltrates exclusively, whereas IIP caused either interstitial infiltrates or diffuse ground-glass opacity. Tuberculosis was the most common etiology of diffuse reticulonodular infiltrates. Only respiratory failure and acute GVHD were significant prognostic factors in multivariate analysis. Bronchoscopy has historically been the primary diagnostic test to evaluate pulmonary diseases in immunocompromised patients, with a diagnostic rate ranging from 40% to 60% and a therapeutic change rate around 30% [9, 20 –24]. For infections, the diagnostic rate of bronchoscopy may approach 80% [25, 26]. However, bronchoscopy has not been shown to improve survival [20 –22, 24, 27]. Although the therapeutic change rate (38.5%) associated with bronchoscopy in our study was similar to that in previous reports, the diagnostic rate (30.7%) was slightly lower, thus limiting the usefulness of bronchoscopy in evaluating DPI in our BMT patients. Table 3. Intubation Rates and Ventilator Days Diagnostic Group

Number of Cases

Intubation, n (%)

Ventilator Days (n)

Bronchoscopy only Open lung biopsy

6 35

3 (50.0) 20 (57.1)

36.0 13.7

PCP ⫽ Pneumocystis carinii pneumonia;

PTLD ⫽ post-transplant

These findings may have two reasons. First, our percentage of infection-related DPI (36.2%) was lower than previous studies (50% to 73%) [15, 28], probably because we included only patients with DPI, not localized infiltrates, and we excluded patients in whom DPI developed after septic shock. These criteria may have excluded most patients with bacterial or fungal pneumonia. In addition, the incidence of infectious complication after transplantation may have been largely reduced by preemptive therapy for CMV disease and trimethoprim-sulfamethoxazole prophylaxis [29 –31]. Thus the overall diagnostic rate of bronchoscopy among our BMT patients with pulmonary infiltrates was reduced. Second, fiber optic bronchoscopy is performed under local anesthesia rather than general anesthesia in our hospital, thus greatly increasing the technical difficulty in sampling. Open lung biopsy is the gold standard for the diagnosis of pulmonary disease in BMT patients [3–7]. As in previous reports [11, 32], open lung biopsy failed to demonstrate a survival benefit in our study (Table 3). However, nearly two-thirds of the clinical diagnoses were wrong and uncertain, leading to the administration of many unnecessary drugs to cover all possible etiologies of DPI and subsequently increasing drug-related toxicities and costs of treatment. Although the intubation rates were similar, the duration of mechanical ventilation was much Table 4. Clinical Diagnosis and Outcome of 33 Patients Without Open Lung Biopsy

Infectious pneumonitis Cytomegalovirus Pneumocystis carinii Others Inflammatory pneumonitis

Number of Cases

Expired (n)

Mortality Rate (%)

22 11 7 4 11

16 9 5 2 5

72.7 81.8 71.4 50.0 45.5

Ann Thorac Surg 2004;78:267–72

WANG ET AL DIFFUSE PULMONARY INFILTRATES AFTER BMT

271

Table 5. Univariate and Multivariate Analysis of Prognostic Factors of Diffuse Pulmonary Infiltrates After BMT in 68 Patients

Respiratory failure Present Not present Acute GVHD Yes No Open lung biopsy Received Not received

Mortailty, n (%)

Univariate p

OR (95% CI)

p

⬍ 0.001

60.35 (6.62 to 550.21)

⬍ 0.001

44 24

33 (75) 1 (4.2)

39 29

25 (64.1) 9 (31)

0.006

6.26 (1.28 to 30.75)

0.016

35 33

13 (37.1) 21 (63.6

0.028

0.23 (0.05 to 1.11)

0.053

BMT ⫽ bone marrow transplantation;

CI ⫽ confidence interval;

GVHD ⫽ graft-versus-host disease;

shorter for patients who underwent open lung biopsy than that for patients who underwent only bronchoscopy. Although bleeding tendency and poor pulmonary reserve might be concerns, open lung biopsy, either performed by mini-thoracotomy or video-assisted thoracoscopy, is safe with a surgical mortality rate of less than 3%. The most common pathologic diagnoses were IIP and CMV pneumonitis, which were also the leading causes of radiographically confirmed interstitial infiltrates. However, the mortality rate of infectious pneumonitis, especially CMV pneumonitis, was much higher than that of inflammatory pneumonitis (Tables 2 and 4). Most BMT patients with diffuse pulmonary ground-glass opacity also had IIP. Our results suggest that CMV pneumonitis is more common than IIP in patients with diffuse interstitial pulmonary infiltrates and concomitant hepatitis [33] shortly after BMT. Miliary tuberculosis, which was rarely reported as the cause of DPI after BMT, was the fourth leading cause in our study, probably reflecting the high incidence of tuberculosis among the general population in Taiwan [34]. It was also the most common disease causing diffuse reticulonodular infiltrates on chest radiographs. GVHD seemed to be the precipitating factor. Thus, in endemic areas, there should be a strong index of suspicion for miliary tuberculosis in BMT patients with GVHD and diffuse reticulonodular pulmonary infiltrates. In our study, the variables that were independently associated with outcome were respiratory failure and acute GVHD. Severe lung damage leading to respiratory failure has been noted as the most common cause of death in BMT patients [9, 10, 20, 21, 35]. The mortality is even higher in patients with CMV pneumonitis and respiratory failure (100% in our study) [36]. Although the response rates in our study were not statistically different between the early biopsy group and the late biopsy group, these dire findings suggest that if the patient’s pulmonary condition continues to deteriorate and a pathologic diagnosis is desired, open lung biopsy should be performed before respiratory failure develops in BMT patients with DPI. Acute GVHD has been shown to increase mortality, especially infection-related mortality, in BMT patients

OR ⫽ odds ratio.

[37]. In our study, the mortality rate doubled and the onset of DPI was significantly earlier if acute GVHD developed. Acute GVHD and active CMV infections were also highly correlated, probably because the immunosuppression associated with GVHD precipitates opportunistic infections [37]. Conversely, CMV may also initiate and enhance the development of GVHD [38, 39]. Our study has some limitations. First, the criteria for open lung biopsy were not clearly defined. Although the MODS at the onset of DPI were not statistically different between the biopsy and nonbiopsy groups, open lung biopsy might still be inappropriate as a prognostic factor because of the possible selection bias that the decision to proceed to open lung biopsy was related specifically to the severity and extent of disease. Second, the treatment protocols were devised by the primary care physicians and were not standardized. Third, the causes of death in the 34 patients were not known because no autopsy was performed. In addition, our patient group was too small to draw powerful conclusions. Further, large controlled studies are needed to evaluate the benefits, safety, and costs of open lung biopsy in patients with DPI after BMT. In conclusion, 20% of BMT patients developed DPI with a mortality rate of 50%. The mortality rate was even higher among patients with respiratory failure or acute GVHD. By providing a definite diagnosis by open lung biopsy early before the onset of respiratory failure, treatment could become more specific, thus reducing the drug-related complications and probably shortening the length of mechanical ventilation and improving the outcome. Idiopathic interstitial pneumonitis and CMV pneumonitis accounted for 60% of the etiologies of DPI and were also the most common cause of radiographically confirmed interstitial pulmonary infiltrates. But CMV pneumonitis is frequently associated with hepatitis and tends to occur shortly after BMT. Idiopathic interstitial pneumonitis should also be considered in those with diffuse ground-glass opacity. In areas endemic for tuberculosis, miliary tuberculosis should be highly suspected in BMT patients with GVHD and diffuse pulmonary reticulonodular lesions.

GENERAL THORACIC

Multivariate N

272

WANG ET AL DIFFUSE PULMONARY INFILTRATES AFTER BMT

Ann Thorac Surg 2004;78:267–72

References

GENERAL THORACIC

1. O’Reilly RJ. Allogeneic bone marrow transplantation: current status and future directions. Blood 1983;62:941–4. 2. Soubani AO, Miller KB, Hassoun PM. Pulmonary complications of bone morrow transplantation. Chest 1996;109:1066–77. 3. Gentile G, Micozzi A, Girmenia C, et al. Pneumonia in allogenic and autologous bone marrow recipients. A retrospective study. Chest 1993;104:371–5. 4. Prober CG, Whyte H, Smith CR. Open lung biopsy in immunocompromised children with pulmonary infiltrates. Am J Dis Child 1984;138:60 –3. 5. McKenna RJ Jr, Mountain CF, McMurtrey MJ. Open lung biopsy in immunocompromised patients. Chest 1984;86:671–4. 6. Haverkos HW, Dowling JN, Pasculle AW, Myerowitz RL, Lerberg DB, Hakala TR. Diagnosis of pneumonitis in immunocompromised patients by open lung biopsy. Cancer 1983; 52:1093–7. 7. Crawford SW, Hackman RC, Clark JG. Open lung biopsy diagnosis of diffuse pulmonary infiltrates after marrow transplantation. Chest 1988;94:949 –53. 8. Poe RH, Wahl GW, Qazi R, Kallay MC, Utell MJ, Morrow GR. Predictors of mortality in the immunocompromised patient with pulmonary infiltrates. Arch Intern Med 1986; 146:1304 –8. 9. Rano A, Agusti C, Benito N, et al. Prognostic factors of non-HIV immunocompromised patients with pulmonary infiltrates. Chest 2002;122:253–61. 10. Agusti C, Rano A, Filella X, et al. Pulmonary infiltrates in patients receiving long-term glucocorticoid treatment. Chest 2003;123:488 –98. 11. Hayes-Jordan A, Benaim E, Richardson S, et al. Open lung biopsy in pediatric bone marrow transplant patients. J Pediatr Surg 2002;37:446 –52. 12. Mori T, Okamoto S, Matsuoka S, et al. Risk-adapted preemptive therapy for cytomegalovirus disease in patients undergoing allogeneic bone marrow transplantation. Bone Marrow Transplant 2000;25:765–9. 13. Machado CM, Menezes RX, Macedo MC, et al. Extended antigenemia surveillance and late cytomegalovirus infection after allogeneic BMT. Bone Marrow Transplant 2001;28:1053–9. 14. Kurihara T, Hayashi J, Ito A, Asai T. CMV antigenemia assay using indirect ALP-immunostaining in bone marrow transplant recipients. Transplant Proc 1996;28:1750 –3. 15. Wingard JR, Mellits ED, Sostrin MB, et al. Interstitial pneumonitis after allogeneic bone marrow transplantation. Medicine 1988;67:175–86. 16. Atkinson K, Nivison-Smith I, Dodds A, Concannon A, Milliken S, Downs K. A comparison of the pattern of interstitial pneumonitis following allogeneic bone marrow transplantation before and after the introduction of prophylactic ganciclovir therapy in 1989. Bone Marrow Transplant 1998;21: 691–5. 17. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: a reliable predictor of a complex clinical outcome. Crit Care Med 1995;23:1638 –52. 18. Ewig S, Schafer H, Torres A. Severity assessment in community-acquired pneumonia. Eur Respir J 2000;16:1193–201. 19. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus conference on acute GVHD grading. Bone Marrow Transplant 1995;15:825–8. 20. Dunagan DP, Baker AM, Hurd DD, Haponik EF. Bronchoscopic evaluation of pulmonary infiltrates following bone marrow transplantation. Chest 1997;111:135–41. 21. Gruson D, Hilbert G, Valentino R, et al. Utility of fiberoptic bronchoscopy in neutropenic patients admitted to the intensive care unit with pulmonary infiltrates. Crit Care Med 2000;28:2224 –30. 22. Rano A, Agusti C, Jimenez P, et al. Pulmonary infiltrates in non-HIV immunocompromised patients: a diagnostic ap-

23.

24.

25.

26.

27.

28.

29.

30.

31.

32. 33. 34. 35.

36. 37.

38.

39.

proach using non-invasive and bronchoscopic procedures. Thorax 2001;56:379 –87. Feinstein MB, Mokhtari M, Ferreiro R, Stover DE, Jakubowski A. Fiberoptic bronchoscopy in allogeneic bone marrow transplantation: findings in the era of serum cytomegalovirus antigen surveillance. Chest 2001;120:1094 –100. Murray PV, O’Brien ME, Padhani AR, et al. Infection post transplant: use of first line bronchoalveolar lavage in the immunosuppressed oncology patient. Bone Marrow Transplant 2001;27:967–71. Stover DE, Zaman MB, Hajdu SI, Lange M, Gold J, Armstrong D. Bronchoalveolar lavage in the diagnosis of diffuse pulmonary infiltrates in the immunosuppressed host. Ann Intern Med 1984;101:1–7. Von Eiff M, Zuhlsdorf M, Roos N, Hesse M, Schulten R, van de Loo J. Pulmonary fungal infections in patients with haematological malignancies— diagnostic approaches. Ann Haematol 1995;70:135–41. Campbell JH, Blessing N, Burnett AK, Stevenson RD. Investigation and management of pulmonary infiltrates following bone marrow transplantation: an eight year review. Thorax 1993;48:1248 –51. Chandrasekar PH, Weinmann A, Shearer C. Autopsyidentified infections among bone marrow transplant recipients: a clinico-pathologic study of 56 patients. Bone Marrow Transplantation Team. Bone Marrow Transplant 1995;16:675–81. Conti DJ, Freed BM, Singh TP, Gallichio M, Gruber SA, Lempert N. Preemptive ganciclovir therapy in cytomegalovirus-seropositive renal transplants recipients. Arch Surg 1995;130:1217–21. Hibberd PL, Tolkoff Rubin NE, Conti D, et al. Preemptive ganciclovir therapy to prevent cytomegalovirus disease in cytomegalovirus antibody-positive renal transplant recipients: a randomized controlled trial. Ann Intern Med 1995; 123:18 –26. Fox BC, Sollinger HW, Belzer FO, Maki DG. A prospective, randomized, double-blind study of trimethoprimsulfamethoxazole for prophylaxis of infection in renal transplantation: clinical efficacy, absorption of trimethoprimsulfamethoxazole, effects on the microflora, and the costbenefit of prophylaxis. Am J Med 1990;89:255–74. White P. Evaluation of pulmonary infiltrates in critically ill patients with cancer and marrow transplant. Crit Care Clin 2001;17:647–70. Jordan MC. Cytomegalovirus infections. In: Hoeprich PD, Jordan MC, Ronald AR, eds. Infectious disease: a treatise of infectious processes. Philadelphia: Lippincott, 1994:835–40. Bureau for Chronic Disease Control. TB statistics (Taiwan)— 1999 [in Chinese]. Taipei, Taiwan: Bureau for Chronic Disease Control, 2001:8 –17. Weisdorf DJ, McGlave PB, Ramsay NK, et al. Allogeneic bone marrow transplantation for acute leukemia: comparative outcome for adults and children. Br J Haematol 1988;69: 351–8. Winston DJ, Ho WG, Ghamplin RE. Cytomegalovirus infections after allogeneic bone marrow transplantation. Rev Infect Dis 1990;12(Suppl 7):S776 –92. Broers AE, van Der Holt R, van Esser JW, et al. Increased transplant-related morbidity and mortality in CMVseropositive patients despite highly effective prevention of CMV disease after allogeneic T-cell-depleted stem cell transplantation. Blood 2000;95:2240 –5. Bostrom L, Ringden O, Gratama JW, Jacobsen N, Zwaan F, Nilsson B. The impact of pre-transplant herpes virus serology on acute and chronic graft-versus-host disease. Leukemia Working Party of the European Group for Bone Marrow Transplantation. Transplant Proc 1990;22:206 –7. Grefte A, van der Giessen M, van Son W, The TH. Circulating cytomegalovirus (CMV)-infected endothelial cells in patients with an active CMV infection. J Infect Dis 1993;167: 270 –7.