The aims of our meta‐analysis were to examine the pattern and gender's influence on body mass index (BMI) – pneumonia relationship. Published studies were searched from PubMed, Web of Science, Cochrane Library databases using keywords of pneumonia, BMI and epidemiologic studies. Random‐effects analysis was applied to estimate pooled effect sizes from individual studies. The Cochrane Q‐test and index of heterogeneity (I²) were used to evaluate heterogeneity, and Egger's test was used to evaluate publication bias. Random‐effects meta‐regression was applied to examine the pattern and gender's influence on BMI–pneumonia relationship.

A total of 1,531 studies were initially identified, and 25 studies finally were included. The pooled relative risk (RR) and meta‐regression model revealed a J‐shaped relationship between BMI and risk of community‐acquired pneumonia (underweight, RR 1.8, 95% confidence interval [CI], 1.4–2.2, P < 0.01; overweight, 0.89, 95%CI, 0.8–1.03, P, 0.1; obesity, 1.03, 95% CI, 0.8–1.3, p. 8) and U‐shaped relationship between BMI and risk of influenza‐related pneumonia (underweight, RR 1.9, 95% CI, 1.2–3, P < 0.01; overweight, 0.89, 95% CI, 0.79–0.99, P, 0.03; obesity, 1.3, 95% CI, 1.05–1.63, p. 2; morbidity obesity, 4.6, 95% CI, 2.2–9.8, P < 0.01); whereas, no difference in risk of nosocomial pneumonia was found across the BMI groups. Gender difference did not make significant contribution in modifying BMI–pneumonia risk relationship.