Body mass index and risk of non-Hodgkin's and Hodgkin's lymphoma: a meta-analysis of prospective studies.

We conducted a meta-analysis of prospective studies to summarise the epidemiologic evidence regarding the association of body mass index (BMI) with non-Hodgkin's lymphoma (NHL) and Hodgkin's lymphoma (HL) incidence and NHL mortality. Pertinent studies were identified by searching PubMed (1966-May 2011) and the reference lists of retrieved articles. For each study, we estimated a relative risk (RR) for a 5 kg/m(2) increase in BMI. A random-effects model was used to combine the RR estimates from individual studies. The summary RRs for a 5 kg/m(2) increase in BMI were 1.07 (95% confidence intervals (CI), 1.04-1.10) for NHL incidence (16 studies, n=17,291 cases) and 1.14 (95% CI, 1.04-1.26) for NHL mortality (five studies, n=3407 cases). BMI was significantly positively associated with risk of diffuse large B-cell lymphoma (RR, 1.13; 95% CI, 1.02-1.26), but not other NHL subtypes. The difference in risk estimates for subtypes was not statistically significant (P=0.10). There was evidence of a nonlinear association between BMI and HL (P for nonlinearity=0.01) (five studies, n=1557 cases). The summary RRs of HL were 0.97 (95% CI, 0.85-1.12) for overweight and 1.41 (95% CI, 1.14-1.75) for obesity. These results indicate that BMI is positively associated with risk of NHL and HL as well as with NHL mortality.