We used Cox proportional hazards modeling to estimate the relative risk (RR) for incident type 2 diabetes in all multivariate analyses (SAS Institute Inc, Cary, NC). Responses to the individual dairy items were converted to average daily number of servings of each item. The average daily intakes of individual dairy items were combined to compute dairy intake: low-fat dairy products, including skim/lowfat milk, sherbet, yogurt, and cottage/ricotta cheese; high-fat dairy foods, including whole milk, cream, sour cream, ice cream, cream cheese, and other cheese; and all dairy products, including all of the above. The average daily dairy intake was categorized into quintiles of intake and each quintile was compared with the lowest quintile. Multivariate models were adjusted for age (continuous), total energy intake (continuous), family history of diabetes (yes or no), smoking status (never smoked; former smoker; current smoker, 1-14 cigarettes per day; current smoker, 15-24 cigarettes per day; or current smoker, ≥25 cigarettes per day), body mass index (<23.0, 23.0-23.9, 24.0-24.9, 2.5-26.9, 2.7-28.9, 2.9-30.9, 31.0-34.9, or ≥35.0), hypercholesterolemia at baseline (yes or no), hypertension at baseline (yes or no), physical activity (quintiles of metabolic equivalent tasks [METs]), cereal fiber intake (quintiles), trans-fat intake (quintiles), ratio of polyunsaturated fat to saturated fat (quintiles), glycemic load (quintiles), and alcohol consumption (0, 0.1-4.9, 5.0-14.9, 15-29.9, or ≥30.0 g/d). We evaluated potential confounding by other possible dietary risk factors for type 2 diabetes (ie, nuts,1 processed meat,22 fruits,23 vegetables,23 coffee intake,24 Western dietary pattern,25 and prudent dietary pattern25) by entering each term (5 categories for coffee intake and quintiles for the others) into the multivariate model. The RR for the continuous measures indicates the change in risk associated with an average increment of one serving per day of the standard portion size. To assess the trends, we used the median values of intake for each category to minimize the influence of outliers. We conducted analyses stratified by body mass index (<25 vs ≥25), family history, or physical activity (<12.2 METs/wk [ie, median value] vs ≥12.2 METs/wk) to assess possible effect modification. We tested the significance of the interaction using the likelihood ratio test by comparing a model that included the main effects of dairy intake along with the stratifying variable and the interaction terms with a reduced model that included only the main effects. We calculated 95% confidence intervals (CIs) for all RRs. All P values are two sided.