Broncho-Alveolar Lavage and Lung Biopsy in Patients with Hematological Malignancy and Hematopoietic Stem Cell Transplantation Recipients: A Systematic Review and Meta-Analysis

Background: Hematological malignancy patients receiving chemotherapy and hematopoietic stem cell transplant (HSCT) recipients experience considerable treatment related toxicities. Radiographic findings during febrile episodes present a therapeutic dilemma. Little is known about the trade-offs between diagnostic yield and risks for different diagnostic approaches in the investigation of lung lesions, namely broncho-alveolar lavage (BAL) and lung biopsy.

Objective: The primary objective of this review was to describe the diagnostic yield of BAL and lung biopsy in the evaluation of pulmonary lesions in patients with hematological malignancies and HSCT recipients. The secondary objectives were to describe the rate of complications and procedure-related mortality of BAL and lung biopsy.

Methods: Electronic searches of Ovid Medline from 1980 to March 14, 2014, EMBASE from 1980 to 2014 week 10, and Cochrane Central Register of Controlled Trials until January 2014 were conducted. Studies were included if pediatric and/or adult patients had hematological malignancy or were HSCT recipients and if patients underwent BAL or lung biopsy for the evaluation of a pulmonary lesion. We limited studies to full-text articles published in the English language after 1980. Studies with lung procedures conducted for initial diagnosis of cancer, surveillance and evaluation for drug toxicity, studies reporting only patients with positive diagnostic tests and studies done to validate a diagnostic test were excluded. Studies exclusively focusing on Pneumocystis jiroveci pneumonia (PCP) were excluded. Two reviewers independently identified articles and abstracted all data. Agreement of study inclusion between the two reviewers was evaluated using the kappa statistic. Synthesis of proportions was conducted using RevMan. All analyses were conducted using the natural logarithm of the proportion as the outcome. All estimates are presented as the proportion with the 95% confidence interval (CI). Heterogeneity was described using the I² value and heterogeneity between sub-groups was evaluated using the chi-square statistic.

Results: 14,148 studies identified by the search strategy and 266 were retrieved for full evaluation; 61 studies of BAL and 29 of lung biopsy were included in the final meta-analysis. Agreement of study inclusion between the two reviewers was almost perfect with kappa statistic = 87.9% (95% CI 82.0 to 93.9%). The proportion of procedures leading to an infectious diagnosis was 0.50 (95% confidence interval (CI) 0.46-0.55; n=44) for BAL and was 0.34 (95% CI 0.28-0.41; n=28) for lung biopsy. The proportion of procedures leading to a non-infectious diagnosis was 0.07 (95% CI 0.05-0.10; n=36) for BAL and 0.43 (95% CI 0.35-0.52; n=27) for lung biopsy. Change in management occurred more often with lung biopsy (0.47, 95% CI 0.39 to 0.57; n=15) compared with BAL (0.31, 95% CI 0.24 to 0.39; n=20). Transthoracic lung biopsies (0.38, 95% CI 0.32 to 0.47; n=23) were more likely to yield an infectious diagnosis compared to transbronchial procedures (0.12, 95% CI 0.06 to 0.24; n=5; P=0.002) and associated with more complications as compared to transbronchial procedures (P=0.02). The proportion of procedures with complications was 0.06 (95% CI 0.04-0.10; n=28) for BAL and 0.13 (95% CI 0.09-0.19; n=21) for lung biopsy. Procedure-related mortality was 0.20% (5/2,447) for BAL and 0.85% (5/589) for lung biopsy.

Conclusions: BAL may be the preferred diagnostic modality for the evaluation of potentially infectious pulmonary lesions among patients with hematological malignancies and HSCT recipients because of lower complication and mortality rates and similar yield. Guidelines to promote consistency in the approach to the evaluation of lung infiltrates may improve clinical care of patients.