Many epidemiologic studies have explored the association between dairy product consumption and the risk of non-Hodgkin lymphoma (NHL), but the results remain controversial. A literature search was performed in PubMed, Web of Science and Embase for relevant articles published up to October 2015. Pooled relative risks (RRs) with 95% confidence intervals (CIs) were calculated with a random-effects model. The dose-response relationship was assessed by restricted cubic spline. A total of 16 articles were eligible for this meta-analysis. The pooled RRs (95% CIs) of NHL for the highest vs. lowest category of the consumption of total dairy product, milk, butter, cheese, ice cream and yogurt were 1.20 (1.02, 1.42), 1.41 (1.08, 1.84), 1.31 (1.04, 1.65), 1.14 (0.96, 1.34), 1.57 (1.11, 2.20) and 0.78 (0.54, 1.12), respectively. In subgroup analyses, the positive association between total dairy product consumption and the risk of NHL was found among case-control studies (RR = 1.41, 95% CI: 1.17-1.70) but not among cohort studies (RR = 1.02, 95% CI: 0.88-1.17). The pooled RRs (95% CIs) of NHL were 1.21 (1.01, 1.46) for milk consumption in studies conducted in North America, and 1.24 (1.09, 1.40) for cheese consumption in studies that adopted validated food frequency questionnaires. In further analysis of NHL subtypes, we found statistically significant associations between the consumption of total dairy product (RR = 1.73, 95% CI: 1.22-2.45) and milk (RR = 1.49, 95% CI: 1.08-2.06) and the risk of diffuse large B-cell lymphoma. The dose-response analysis suggested that the risk of NHL increased by 5% (1.05 (1.00-1.10)) and 6% (1.06 (0.99-1.13)) for each 200 g/day increment of total dairy product and milk consumption, respectively. This meta-analysis suggested that dairy product consumption, but not yogurt, may increase the risk of NHL. More prospective cohort studies that investigate specific types of dairy product consumption are needed to confirm this conclusion.
Keywords: dairy product; meta-analysis; milk; non-Hodgkin lymphoma.