Dairy foods (DFs) contain complex ingredients that could affect different diseases. The control of lactose digestion phenotypically divides populations into those who can [lactase persistent (LP)] and those who cannot [lactase nonpersistent (LNP)] assimilate lactose. LNP subjects, however, can adapt to lactose intolerance through intestinal bacteria. The DF/LNP status interactions may function as disease risk modifiers. We evaluated the relationship between DF and LNP with colorectal, breast, prostate, ovarian, lung, and stomach cancer and inflammatory bowel diseases (IBD; Crohn's disease and ulcerative colitis). Yearly per capita DF consumption, LNP national prevalence, cancer mortality, and incidence of IBD were obtained from several sources. A negative binomial regression model was used to derive incremental risks. There were statistically significant (P <or= 0.05) increases in risk for colorectal and prostate cancer and ulcerative colitis with DFs and a statistically significant decreased risk for stomach cancer. There were trends (P<0.1) for lung and ovarian cancers and Crohn's disease. As LNP prevalence increased, stomach cancer risk increased, whereas risks of all other conditions decreased (P<0.01). In 3 cancers (prostate, ovarian, and breast), meta-analyses of case-based studies support ecological data. In colorectal cancer, on the contrary, meta-analyses of case-based studies suggest protection. The possible importance of distinguishing LNP/LP status in studies is discussed.