Introduction/Objective: To compare the diagnostic effectiveness of bronchoalveolar lavage (BAL) and gastric lavage (GL) in isolating Mycobacterium Tuberculosis (Mtb) in children with pulmonary Tuberculosis (PTB).

Methods: Studies addressing to the problem of bacteriological isolation of Mtb from GL or BAL in PTB in children were accessed and systematically reviewed. Total of 3 published studies in children and one individual patient data base constituted by authors were meta-analysed for the diagnostic outcome measure of number of patients with a positive culture of Mtb. Typical Odds Ratios (Peto) and Dersimonian and Laird Method were used for pooling the data.

Results: The individual odds ratios and rate differences for individual studies and typical odds ratio and pooled rate difference for the pooled effect are given in the table.

Study Year
GL/Total
(95% CI) Bal/Total Odds ratio Rate Difference
(95% CI)
1992 2/20:10/20 0.16 (0.04-.59) -0.4(-0.6 to -0.1 4)
1994 6/50:16/50 0.32 (.12-.81) -0.2(-0.3 to -0.04)
1995 2/16:3/.16 0.63 (.1-4.3) -0.06(-0.3 to -0.18)
1997 9/33:5/33 2.04 (0.63-6.59) 0.12(-0.07 to 0.3)
Pooled effect
19/119:34/119 0.49(.26-.89) -0.13(-0.3 to 0.07)

Discussion: The typical Odds Ratio is in favour of gastric lavage being more effective in isolating Mtb. However, according to Dersimonian and Laird pooled rate difference, which includes a test of heterogeniety the pooled effect is still inconclusive and will need more number of cases to be worked up to provide a conclusive evidence.