ABSTRACT

Background: The association between intake of dairy products and the risk of type 2 diabetes has been investigated in several studies, but the evidence is not conclusive. Objective: We conducted an updated systematic review and dose-response meta-analysis of dairy product intake and the risk of type 2 diabetes. Design: We searched the PubMed database for prospective cohort and nested case-control studies of dairy product intake and risk of type 2 diabetes up to 5 June 2013. Summary RRs were estimated by use of a random-effects model. Results: Seventeen cohort studies were included in the meta-analysis. In the dose-response analysis, the summary RRs (95% CIs) were 0.93 (0.87, 0.99; I2 = 33%) per 400 g total dairy products/d (n = 12), 0.98 (0.94, 1.03; I2 = 8%) per 200 g high-fat dairy products/d (n = 9), 0.91 (0.86, 0.96; I2 = 40%) per 200 g low-fat dairy products/d (n = 9), 0.87 (0.72, 1.04; I2 = 94%) per 200 g milk/d (n = 7), 0.92 (0.86, 0.99; I2 = 0%) per 50 g cheese/d (n = 8), and 0.78 (0.60, 1.02; I2 = 70%) per 200 g yogurt/d (n = 7). Nonlinear inverse associations were observed for total dairy products (P-nonlinearity < 0.0001), low-fat dairy products (P-nonlinearity = 0.06), cheese (P-nonlinearity = 0.05), and yogurt (P-nonlinearity = 0.004), and there was a flattening of the curve at higher intakes. Conclusions: This meta-analysis suggests that there is a significant inverse association between intakes of dairy products, low-fat dairy products, and cheese and risk of type 2 diabetes. Any additional studies should assess the association between other specific types of dairy products and the risk of type 2 diabetes and adjust for more confounding factors.

INTRODUCTION

The prevalence of type 2 diabetes is increasing rapidly around the world, parallel to the increase in obesity, the reduction in physical activity, and dietary changes. It has been estimated that 366 million persons had diabetes (mostly type 2) in 2011, and the number has been projected to increase to 552 million by 2030 (1). Although diet is thought to be of major importance for the increased prevalence of type 2 diabetes, few dietary factors have been established as risk factors for type 2 diabetes (2–5).

Dairy products have been hypothesized to protect against type 2 diabetes because of their high content of calcium, magnesium, vitamin D, and whey proteins, which may reduce body fat and insulin resistance (6). However, some dairy products, such as cheese and cream, also have a high fat content that might offset any benefits of increased intake of calcium or other potentially beneficial dairy components. Epidemiologic studies have yielded mixed results: some have suggested a decreased risk associated with higher intake of dairy products (7–14), whereas other studies suggested no association (15–23). Studies of specific types of dairy products have also shown mixed results: some have reported inverse associations for low-fat dairy products (7, 10, 13, 15, 19), milk (12, 16), low-fat or skim milk (7, 10, 14), cheese (10, 14, 21), and yogurt (7, 10, 19, 21), whereas other studies suggested no association (8, 17, 18, 20, 22–24). In contrast, most studies of high-fat dairy products reported no association (7, 8, 10, 13, 15, 20, 24), and only one study reported a reduced risk (23). The dose-response relation between dairy products and type 2 diabetes needs more detailed examination to establish whether there could be potential threshold effects. In addition, it is important to establish whether the associations may differ according to the type of dairy product consumed and by study characteristics such as geographic location and adjustment for confounding factors. To clarify the association between dairy product intake and risk of type 2 diabetes, we conducted a systematic review and meta-analysis of the available prospective studies, with specific aims of analyzing different types of dairy products, to clarify whether the association differed by study characteristics and to clarify any dose-response relation between dairy product intake and the risk of type 2 diabetes.

METHODS

Search strategy

We searched the PubMed database (http://www.ncbi.nlm.nih.gov/pubmed) up to 5 June 2013 for cohort studies of dairy intake and type 2 diabetes risk. As part of a larger systematic review of dietary factors and type 2 diabetes risk we used broad search terms on a wide range of dietary factors and type 2 diabetes. The search terms included the following: (cereal OR grain OR grains OR rice OR bread OR roots OR tubers OR vegetable OR fruits OR pulse OR pulses OR bean OR beans OR lentil OR lentils OR legume OR legumes OR soy OR soya OR pea OR chickpeas OR chickpea OR nut OR seed OR peanut OR peanuts OR meat OR beef OR pork OR lamb OR poultry OR chicken OR fish OR egg OR eggs OR seafood OR shellfish OR dairy OR dairy products OR milk OR cheese OR yoghurt OR ice cream OR butter OR drink OR drinks OR beverage OR soda OR sodas OR juice OR juices OR punch OR foods) AND diabetes. We also searched the reference lists of previous reviews of the subject (25–28) and of the studies included in the analysis for any further studies.

Study selection

To be included, the study had to have a prospective cohort, a case-cohort, or a nested case-control design and to investigate the association between intake of dairy products and the risk of type 2 diabetes. Estimates of the RR (HR, risk ratio, OR) had to be available with the 95% CIs, and for the dose-response analysis a quantitative measure of intake and the total number of cases and person-years had to be available in the publication or on request from the authors. We identified 22 potentially relevant studies (7–24, 29–32). One study was excluded because of a cross-sectional design (31), 2 studies did not report any risk estimates for type 2 diabetes (29, 30), and one study was excluded because it reported on a combined outcome of impaired fasting blood glucose and type 2 diabetes (32). The European Investigation into Cancer and Nutrition (EPIC)–Potsdam Study (24) reported on different dairy food items than the EPIC-InterAct Study (21) and was therefore included despite the overlap between these studies.

Data extraction

We extracted the following data from each study: the first author’s last name, publication year, country in which the study was conducted, study name, follow-up period, sample size, sex, age, number of cases, dietary assessment method (type, number of food items, and whether the method had been validated), exposure, quantity of intake, RRs and 95% CIs for the highest compared with the lowest intake, and variables adjusted for in the analysis.

Statistical methods

We used random-effects models to calculate summary RRs and 95% CIs for the highest compared with the lowest amount of dairy product intake and for the dose-response analysis (33). The average of the natural logarithm of the RRs was estimated, and the RR from each study was weighted by the inverse of its variance. A 2-tailed P value <0.05 was considered significant. For one study that reported results separately for men and women (18), but not combined, we combined the results by using a fixed-effects model to obtain an estimate for both sexes combined, which was used for the overall analysis.

We used the method described by Greenland and Longnecker (34) for the dose-response analysis and computed study-specific slopes (linear trends) and 95% CIs from the natural logs of the RRs and CIs across categories of dairy product intake. The method requires that the distribution of cases and person-years or noncases and the RRs with the variance estimates for at least 3 quantitative exposure categories are known. We estimated the distribution of cases or person-years in studies that did not report these but reported the total number of cases/person-years (35). The median or mean amount of dairy product intake in each category of intake was assigned to the corresponding RR for each study. For studies that reported dairy product intake by ranges of intake we estimated the midpoint for each category by calculating the average of the lower and upper bound. When the highest or lowest category was open-ended, we assumed the open-ended interval length to be the same as the adjacent interval. In studies that reported the intakes by frequency, we used 43 g as a serving size for cheese, 177 g as a serving size for total dairy products, and 244 g as a serving size for milk and yogurt intake to recalculate the intakes to a common scale (g/d) (36). We examined a potential nonlinear dose-response relation between dairy intake and type 2 diabetes by using fractional polynomial models (37). We determined the best-fitting second-order fractional polynomial regression model, defined as the one with the lowest deviance. A likelihood ratio test was used to assess the difference between the nonlinear and linear models to test for nonlinearity (37). The intake in the reference category was subtracted from the intake in each category for the linear dose-response analysis but not for the nonlinear dose-response analysis. Supplementary information was requested from 3 studies (9, 14, 16) and was obtained from 2 of these studies (14, 16).

Heterogeneity between studies was assessed by the Q test and I2 (38). I2 is the amount of total variation that is explained by between-study variation. I2 values of ~25%, 50%, and 75% are considered to indicate low, moderate, and high heterogeneity, respectively. To investigate sources of heterogeneity, we conducted subgroup and meta-regression analyses stratified by study characteristics such as sex, duration of follow-up, number of cases, and adjustment for confounding factors.

Publication bias was assessed with Egger’s test (39), and the results were considered to indicate publication bias when P < 0.10. We conducted sensitivity analyses excluding one study at a time to ensure that the results were not simply due to one large study or to a study with an extreme result. Results from these sensitivity analyses are presented excluding the 2 studies with the largest negative and positive impact on the summary estimates. The statistical analyses were conducted by using Stata, version 10.1 (StataCorp).

RESULTS

We identified 17 cohort studies (18 publications) (7–24) that could be included in the analysis. All the studies were included in the high compared with low meta-analysis (7–23), and 15 cohort studies (16 publications) (7, 8, 10, 12–24) could be included in the dose-response meta-analysis (Table 1 and Figure 1). Seven studies were from the United States, 6 studies were from Europe, 2 were from Asia, and 2 were from Australia (Table 1).

TABLE 1 First author, publication year, country/ region (ref) Study name Follow-up period Study size, sex, age, number of cases Dietary assessment Exposure Quantity RR (95% CI) Adjustment for confounders Choi, 2005, USA (7) Health Professionals Follow-Up Study 1986–1998, 12-y follow-up 41,254 men, age 40–75 y, 1243 cases Validated FFQ, 131 food items Total dairy intake ≥2.9 vs <0.9 servings/d 0.77 (0.62, 0.95) Age, total energy intake, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, alcohol intake, cereal fiber, trans FAs, PUFA:SFA ratio, glycemic load Per 1 serving/d 0.91 (0.85, 0.97) Low-fat dairy foods ≥1.58 vs <0.14 servings/d 0.74 (0.60, 0.91) Per 1 serving/d 0.88 (0.81, 0.94) High-fat dairy foods ≥1.72 vs <0.38 servings/d 0.97 (0.78, 1.21) Per 1 serving/d 0.99 (0.91, 1.07) Skim, low-fat milk ≥2 servings/d vs <1 serving/mo 0.78 (0.63, 0.97) Whole milk ≥2 servings/wk vs <1 serving/mo 1.19 (1.00, 1.43) Yogurt ≥2 servings/wk vs <1 serving/mo 0.83 (0.66, 1.06) Sherbet ≥2 servings/wk vs <1 serving/mo 0.85 (0.66, 1.09) Cottage, ricotta cheese ≥2 servings/wk vs <1 serving/mo 0.96 (0.80, 1.17) Ice cream ≥2 servings/wk vs <1 serving/mo 0.78 (0.64, 0.95) Other cheese ≥2 servings/wk vs <1 serving/mo 0.88 (0.67, 1.16) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.06 (0.81, 1.39) Cream ≥2 servings/wk vs <1 serving/mo 0.89 (0.72, 1.09) Sour cream ≥2 servings/wk vs <1 serving/mo 1.04 (0.80, 1.36) Montonen, 2005, Finland (8) Finnish Mobile Clinic Health Examination Survey 1967–1990, 23-y follow-up 4304 men and women, age 40–69 y, 383 cases Dietary history interview, >100 food items Regular dairy products ≥305 vs <39 g/d 0.81 (0.62, 1.08) Age, sex, BMI, energy intake, smoking, FH of DM, geographic area Reduced-fat dairy products >0 vs 0 g/d 0.90 (0.60, 1.36) Whole milk ≥878 vs <326 g/d 1.06 (0.75, 1.50) Butter >59 vs <27 g/d 1.15 (0.80, 1.67) Pittas, 2006, USA (9) Nurses’ Health Study 1980–2000, 20-y follow-up 83,779 women, age 30–55 y, 4843 cases Validated FFQ, 61–116 food items Dairy foods ≥3 vs <1 serving/d 0.89 (0.81, 0.99) Age, BMI, hypertension, FH of DM, smoking, physical activity, caffeine, alcohol, residence, SFAs, PUFAs, trans FAs, cereal fiber, Mg, GL, retinol, energy Liu, 2006, USA (10) Women’s Health Study 1993–2003, 10-y follow-up 37,183 women, age ≥45 y, 1603 cases Validated FFQ, 131 food and beverage items Total dairy intake >2.9 vs <0.85 servings/d 0.80 (0.67, 0.95) Age, total energy intake, randomized treatment assignment, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, hormones, alcohol, dietary fiber, total fat, GL Per 1 serving/d 0.97 (0.93, 1.01) Low-fat dairy foods >2.00 vs 0.27 servings/d 0.82 (0.68, 0.98) Per 1 serving/d 0.95 (0.90, 1.01) High-fat dairy foods >1.329 vs <0.20 servings/d 0.97 (0.80, 1.17) Per 1 serving/d 1.00 (0.95, 1.04) Skim milk ≥2 servings/wk vs <1 serving/mo 0.92 (0.78, 1.09) Whole milk ≥2 servings/wk vs <1 serving/mo 1.04 (0.84, 1.30) Yogurt ≥2 servings/wk vs <1 serving/mo 0.82 (0.70, 0.97) Sherbet ≥2 servings/wk vs <1 serving/mo 0.92 (0.77, 1.05) Cottage cheese ≥2 servings/wk vs <1 serving/mo 0.86 (0.71, 1.05) Ice cream ≥2 servings/wk vs <1 serving/mo 0.88 (0.74, 1.05) Other cheese ≥2 servings/wk vs <1 serving/mo 0.80 (0.64, 1.01) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.19 (0.97, 1.47) Cream ≥2 servings/wk vs <1 serving/mo 1.03 (0.86, 1.25) Sour cream ≥2 servings/wk vs <1 serving/mo 0.93 (0.74, 1.18) van Dam, 2006, USA (15) Black Women’s Health Study 1995–2003, 8-y follow-up 41,186 women, age 21–69 y, 1964 cases Validated FFQ, 68 food items Total dairy products 2.53 vs 0.07 servings/d 0.93 (0.75, 1.15) Age, total energy intake, alcohol, BMI, smoking status, strenuous physical activity, parental history of DM, education, coffee, sugar-sweetened soft drinks, processed meat, red meat, whole grains Low-fat dairy products 1.33 vs 0.07 servings/d 0.87 (0.76, 1.00) High-fat dairy products 1.33 vs 0.07 servings/d 1.03 (0.88, 1.20) Lecomte, 2007, France (11) NA 1995/1997–2000/2002, 5-y follow-up 743 men with IFG, age 20–60 y, 127 cases FFQ, 18 food items Dairy products Not daily vs daily 1.86 (1.21, 2.86) FH of DM, BMI, TG, glucose at baseline Elwood, 2007, United Kingdom (16) Caerphilly Prospective Study 1979/1983–NA, 20-y follow-up 640 men, age 45–59 y, 41 cases 7-d weighed records Milk 396.8 vs 76.5 g/d2 0.57 (0.20, 1.63) Age, smoking, BMI, social class Dairy foods 430.4 vs 107.0 g/d2 0.59 (0.21, 1.69)2 Vang, 2008, USA (17) Adventist Health Study 1960–1976, 17-y follow-up 8401 men and women, age 45–88 y, 543 cases Validated FFQ Cheese ≥1 time/wk vs never 0.90 (0.71, 1.15) Age, sex Milk ≥4 times/d vs never 0.92 (0.65, 1.29) Villegas, 2009, China (12) Shanghai Women’s Health Study 2000–2006, 6.9-y follow-up 64,191 women, age 40–70 y, 2270 cases (1514 confirmed cases) Validated FFQ, 77 food items Fresh milk, all participants 250.0 vs 0 g/d 0.46 (0.32, 0.64) Age, energy intake, BMI, WHR, smoking status, alcohol, physical activity, income, education, occupation, hypertension Powdered milk Yes vs no 0.74 (0.67, 0.82) Fresh milk, confirmed cases 250.0 vs 0 g/d 0.60 (0.41, 0.88) Powdered milk Yes vs no 0.85 (0.75, 0.96) Kirii, 2009, Japan (18) Japan Public Health Center–based Prospective Study 1995/1998–2000/2003, 5-y follow-up 59,796 men and women, age 40–59 y, 1114 cases Validated FFQ, 147 food and beverage items Dairy products, men ≥300 vs <50 g/d 1.18 (0.90, 1.56) Age, area, BMI, FH of DM, smoking status, alcohol intake, history of hypertension, exercise frequency, coffee consumption, energy-adjusted Mg, total energy intake Milk ≥200 vs <50 g/d 1.02 (0.85, 1.24) Cheese ≥5 vs 0 g/d 0.88 (0.64, 1.21) Yogurt ≥60 vs 0 g/d 1.01 (0.75, 1.36) Dairy products, women ≥300 vs <50 g/d 0.71 (0.51, 0.98) Milk ≥200 vs <50 g/d 0.87 (0.70, 1.09) Cheese ≥5 vs 0 g/d 1.12 (0.80, 1.57) Yogurt ≥60 vs 0 g/d 0.77 (0.58, 1.01) Malik, 2011, USA (13) Nurses’ Health Study II 1997–2005, 7-y follow-up 37,038 women, age 24–42 y, 550 cases Validated FFQ, 133 food items Total dairy 2.14 vs 0.62 servings per 1000 kcal/d 0.75 (0.55, 1.02) Age, BMI, total energy, FH of DM, smoking status, physical activity, alcohol use, OC use, HRT, PUFA:SFA ratio, GL, cereal fiber, trans fat, processed meat, carbonated soft drinks, fruit drinks, coffee; mutual adjustment: high- and low-fat dairy products Low-fat dairy foods 1.44 vs 0.18 servings per 1000 kcal/d 0.74 (0.54, 1.01) High-fat dairy foods 1.14 vs 0.19 servings per 1000 kcal/d 0.72 (0.53, 0.99) Margolis, 2011, USA (19) Women’s Health Initiative 1994/1998–2005, 7.9-y follow-up 82,076 women, age 50–79 y, 3946 cases FFQ, >300 foods and beverages Low-fat dairy products 2.8 vs 0.05 servings/d 0.65 (0.44, 0.96) Age, race-ethnicity, total energy intake, income, education, BMI, smoking, alcohol intake, FH of DM, HRT, SBP, DBP, physical activity, interaction of low-fat dairy × BMI, interaction of yogurt × time Total dairy products 3.4 vs 0.5 servings/d 0.93 (0.83, 1.04) Yogurt ≥2 servings/wk vs <1 serving/mo 0.46 (0.31, 0.68) Sluijs, 2012, Europe (21) EPIC-Interact Study 1991–2007, 11.7-y follow-up 24,475 men and women, mean age 52 y, 10,694 cases Validated FFQ, 24-h recall Total dairy products 628.9 vs 79.7 g/d 1.01 (0.89, 1.13) Age; center; sex; BMI; education; smoking status; physical activity; intakes of alcohol, fruit and vegetables, red meat, processed meat, sugar-sweetened soft drinks, coffee, cereals, cereal products, energy Milk 486.1 vs 0.3 g/d 1.10 (0.92, 1.31) Yogurt and thick, fermented milk 190.4 vs 0 g/d 0.91 (0.81, 1.02) Cheese 73.7 vs 3.2 g/d 0.88 (0.76, 1.02) Combined fermented dairy intake 220.7 vs 11.6 g/d 0.88 (0.78, 0.99) Struijk, 2012, Denmark (22) The Inter99 Study 1999/2001–2006, 5-y follow-up 5953 men and women, age 30–60 y, 214 cases Validated FFQ, 198 food items Total dairy 578 vs 47 g/d 0.96 (0.58, 1.58) Age; sex; intervention group; FH of DM; education; physical activity; smoking status; intakes of alcohol, whole-grain cereal, meat, fish, coffee, tea, fruit, vegetables, energy; change in diet from baseline to 5-y follow-up; waist circumference Per 1 serving/d 0.95 (0.86, 1.06) Low-fat dairy 536 vs 6 g/d 0.85 (0.52, 1.40) Per 1 serving/d 0.95 (0.85, 1.06) Full-fat dairy 89 vs 4 g/d 0.94 (0.56, 1.58) Per 1 serving/d 1.03 (0.77, 1.36) Milk and milk products 546 vs 16 g/d 0.95 (0.58, 1.57) Per 1 serving/d 0.96 (0.86, 1.06) Cheese 49 vs 4 g/d 0.78 (0.47, 1.29) Per 1 serving/d 0.97 (0.82, 1.15) Fermented dairy 260 vs 13 g/d 0.86 (0.50, 1.47) Per 1 serving/d 0.88 (0.69, 1.11) Louie, 2012, Australia (23) The Blue Mountains Eye Study 1992/1994–2002/2004, 10-y follow-up 1824 men and women, age ≥49 y, 145 cases Validated FFQ, 145 food items Total dairy 3.1 vs 0.5 servings/d 0.96 (0.54, 1.17) Age, sex, smoking status, physical activity, GL, vegetable fiber, total energy, FH of DM2, SBP, BMI, HDL, total cholesterol, TG Reduced/low-fat dairy 2.1 vs 0 servings/d 1.00 (0.58, 1.71) Regular dairy 1.9 vs 0.1 servings/d 0.83 (0.47, 1.48) Grantham, 2013, Australia (14) The Australian Diabetes Obesity and Lifestyle Study 1999/2000–2004/2005, 5-y follow-up 5582 men and women, age ≥25 y, 209 cases FFQ, 121 food items Total dairy foods 477 vs 205 g/d2 0.71 (0.48, 1.05) Age, sex, energy intake, FH DM, education, physical activity, smoking status, TG, HDL cholesterol, SBP, waist circumference, hip circumference Low-fat milk 375 vs 0 g/d2 0.65 (0.44, 0.94) Full-fat milk 375 vs 0 g/d2 1.18 (0.78, 1.79) Yogurt 114 vs 3 g/d2 1.14 (0.78, 1.67) Cheese 29 vs 4 g/d2 0.78 (0.53, 1.15) von Ruesten, 2013, Germany (24) EPIC-Potsdam Study 1994/1998–NA, 8-y follow-up 23,531 men and women, age 35–65 y, 837 cases Validated FFQ, 148 food items Low-fat dairy products Per 100 g/d 1.02 (0.96, 1.09) Age, sex, smoking status, pack-years of smoking, alcohol, leisure-time physical activity, BMI, WHR, hypertension, high blood lipids, education, vitamin supplementation, total energy intake, other food groups High-fat dairy products Per 100 g/d 1.00 (0.92, 1.08) Low-fat cheese Per 30 g/d 0.98 (0.83, 1.15) High-fat cheese Per 30 g/d 0.96 (0.85, 1.08) Soedamah-Muthu, 2013, UK (20) The Whitehall II Prospective Study 1985/1988–2009, 10-y follow-up 4526 men and women, 35–55 y, 273 cases FFQ, 114 food items Total dairy 575 vs 246 g/d 1.30 (0.95, 1.77) Age; ethnicity; employment grade; smoking; BMI; physical activity; FH of CHD/hypertension; intakes of alcohol, fruit and vegetables, bread, meat, fish, coffee, tea, total energy High-fat dairy 182 vs 27 g/d 1.23 (0.91, 1.67) Low-fat dairy 458 vs 28 g/d 0.98 (0.73, 1.31) Total milk 441 vs 147 g/d 0.97 (0.71, 1.32) Fermented dairy 105 vs 17 g/d 1.17 (0.87, 1.58) Yogurt 117 vs 0 g/d 1.04 (0.77, 1.42) Cheese 31 vs 6 g/d 1.20 (0.88, 1.64) First author, publication year, country/ region (ref) Study name Follow-up period Study size, sex, age, number of cases Dietary assessment Exposure Quantity RR (95% CI) Adjustment for confounders Choi, 2005, USA (7) Health Professionals Follow-Up Study 1986–1998, 12-y follow-up 41,254 men, age 40–75 y, 1243 cases Validated FFQ, 131 food items Total dairy intake ≥2.9 vs <0.9 servings/d 0.77 (0.62, 0.95) Age, total energy intake, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, alcohol intake, cereal fiber, trans FAs, PUFA:SFA ratio, glycemic load Per 1 serving/d 0.91 (0.85, 0.97) Low-fat dairy foods ≥1.58 vs <0.14 servings/d 0.74 (0.60, 0.91) Per 1 serving/d 0.88 (0.81, 0.94) High-fat dairy foods ≥1.72 vs <0.38 servings/d 0.97 (0.78, 1.21) Per 1 serving/d 0.99 (0.91, 1.07) Skim, low-fat milk ≥2 servings/d vs <1 serving/mo 0.78 (0.63, 0.97) Whole milk ≥2 servings/wk vs <1 serving/mo 1.19 (1.00, 1.43) Yogurt ≥2 servings/wk vs <1 serving/mo 0.83 (0.66, 1.06) Sherbet ≥2 servings/wk vs <1 serving/mo 0.85 (0.66, 1.09) Cottage, ricotta cheese ≥2 servings/wk vs <1 serving/mo 0.96 (0.80, 1.17) Ice cream ≥2 servings/wk vs <1 serving/mo 0.78 (0.64, 0.95) Other cheese ≥2 servings/wk vs <1 serving/mo 0.88 (0.67, 1.16) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.06 (0.81, 1.39) Cream ≥2 servings/wk vs <1 serving/mo 0.89 (0.72, 1.09) Sour cream ≥2 servings/wk vs <1 serving/mo 1.04 (0.80, 1.36) Montonen, 2005, Finland (8) Finnish Mobile Clinic Health Examination Survey 1967–1990, 23-y follow-up 4304 men and women, age 40–69 y, 383 cases Dietary history interview, >100 food items Regular dairy products ≥305 vs <39 g/d 0.81 (0.62, 1.08) Age, sex, BMI, energy intake, smoking, FH of DM, geographic area Reduced-fat dairy products >0 vs 0 g/d 0.90 (0.60, 1.36) Whole milk ≥878 vs <326 g/d 1.06 (0.75, 1.50) Butter >59 vs <27 g/d 1.15 (0.80, 1.67) Pittas, 2006, USA (9) Nurses’ Health Study 1980–2000, 20-y follow-up 83,779 women, age 30–55 y, 4843 cases Validated FFQ, 61–116 food items Dairy foods ≥3 vs <1 serving/d 0.89 (0.81, 0.99) Age, BMI, hypertension, FH of DM, smoking, physical activity, caffeine, alcohol, residence, SFAs, PUFAs, trans FAs, cereal fiber, Mg, GL, retinol, energy Liu, 2006, USA (10) Women’s Health Study 1993–2003, 10-y follow-up 37,183 women, age ≥45 y, 1603 cases Validated FFQ, 131 food and beverage items Total dairy intake >2.9 vs <0.85 servings/d 0.80 (0.67, 0.95) Age, total energy intake, randomized treatment assignment, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, hormones, alcohol, dietary fiber, total fat, GL Per 1 serving/d 0.97 (0.93, 1.01) Low-fat dairy foods >2.00 vs 0.27 servings/d 0.82 (0.68, 0.98) Per 1 serving/d 0.95 (0.90, 1.01) High-fat dairy foods >1.329 vs <0.20 servings/d 0.97 (0.80, 1.17) Per 1 serving/d 1.00 (0.95, 1.04) Skim milk ≥2 servings/wk vs <1 serving/mo 0.92 (0.78, 1.09) Whole milk ≥2 servings/wk vs <1 serving/mo 1.04 (0.84, 1.30) Yogurt ≥2 servings/wk vs <1 serving/mo 0.82 (0.70, 0.97) Sherbet ≥2 servings/wk vs <1 serving/mo 0.92 (0.77, 1.05) Cottage cheese ≥2 servings/wk vs <1 serving/mo 0.86 (0.71, 1.05) Ice cream ≥2 servings/wk vs <1 serving/mo 0.88 (0.74, 1.05) Other cheese ≥2 servings/wk vs <1 serving/mo 0.80 (0.64, 1.01) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.19 (0.97, 1.47) Cream ≥2 servings/wk vs <1 serving/mo 1.03 (0.86, 1.25) Sour cream ≥2 servings/wk vs <1 serving/mo 0.93 (0.74, 1.18) van Dam, 2006, USA (15) Black Women’s Health Study 1995–2003, 8-y follow-up 41,186 women, age 21–69 y, 1964 cases Validated FFQ, 68 food items Total dairy products 2.53 vs 0.07 servings/d 0.93 (0.75, 1.15) Age, total energy intake, alcohol, BMI, smoking status, strenuous physical activity, parental history of DM, education, coffee, sugar-sweetened soft drinks, processed meat, red meat, whole grains Low-fat dairy products 1.33 vs 0.07 servings/d 0.87 (0.76, 1.00) High-fat dairy products 1.33 vs 0.07 servings/d 1.03 (0.88, 1.20) Lecomte, 2007, France (11) NA 1995/1997–2000/2002, 5-y follow-up 743 men with IFG, age 20–60 y, 127 cases FFQ, 18 food items Dairy products Not daily vs daily 1.86 (1.21, 2.86) FH of DM, BMI, TG, glucose at baseline Elwood, 2007, United Kingdom (16) Caerphilly Prospective Study 1979/1983–NA, 20-y follow-up 640 men, age 45–59 y, 41 cases 7-d weighed records Milk 396.8 vs 76.5 g/d2 0.57 (0.20, 1.63) Age, smoking, BMI, social class Dairy foods 430.4 vs 107.0 g/d2 0.59 (0.21, 1.69)2 Vang, 2008, USA (17) Adventist Health Study 1960–1976, 17-y follow-up 8401 men and women, age 45–88 y, 543 cases Validated FFQ Cheese ≥1 time/wk vs never 0.90 (0.71, 1.15) Age, sex Milk ≥4 times/d vs never 0.92 (0.65, 1.29) Villegas, 2009, China (12) Shanghai Women’s Health Study 2000–2006, 6.9-y follow-up 64,191 women, age 40–70 y, 2270 cases (1514 confirmed cases) Validated FFQ, 77 food items Fresh milk, all participants 250.0 vs 0 g/d 0.46 (0.32, 0.64) Age, energy intake, BMI, WHR, smoking status, alcohol, physical activity, income, education, occupation, hypertension Powdered milk Yes vs no 0.74 (0.67, 0.82) Fresh milk, confirmed cases 250.0 vs 0 g/d 0.60 (0.41, 0.88) Powdered milk Yes vs no 0.85 (0.75, 0.96) Kirii, 2009, Japan (18) Japan Public Health Center–based Prospective Study 1995/1998–2000/2003, 5-y follow-up 59,796 men and women, age 40–59 y, 1114 cases Validated FFQ, 147 food and beverage items Dairy products, men ≥300 vs <50 g/d 1.18 (0.90, 1.56) Age, area, BMI, FH of DM, smoking status, alcohol intake, history of hypertension, exercise frequency, coffee consumption, energy-adjusted Mg, total energy intake Milk ≥200 vs <50 g/d 1.02 (0.85, 1.24) Cheese ≥5 vs 0 g/d 0.88 (0.64, 1.21) Yogurt ≥60 vs 0 g/d 1.01 (0.75, 1.36) Dairy products, women ≥300 vs <50 g/d 0.71 (0.51, 0.98) Milk ≥200 vs <50 g/d 0.87 (0.70, 1.09) Cheese ≥5 vs 0 g/d 1.12 (0.80, 1.57) Yogurt ≥60 vs 0 g/d 0.77 (0.58, 1.01) Malik, 2011, USA (13) Nurses’ Health Study II 1997–2005, 7-y follow-up 37,038 women, age 24–42 y, 550 cases Validated FFQ, 133 food items Total dairy 2.14 vs 0.62 servings per 1000 kcal/d 0.75 (0.55, 1.02) Age, BMI, total energy, FH of DM, smoking status, physical activity, alcohol use, OC use, HRT, PUFA:SFA ratio, GL, cereal fiber, trans fat, processed meat, carbonated soft drinks, fruit drinks, coffee; mutual adjustment: high- and low-fat dairy products Low-fat dairy foods 1.44 vs 0.18 servings per 1000 kcal/d 0.74 (0.54, 1.01) High-fat dairy foods 1.14 vs 0.19 servings per 1000 kcal/d 0.72 (0.53, 0.99) Margolis, 2011, USA (19) Women’s Health Initiative 1994/1998–2005, 7.9-y follow-up 82,076 women, age 50–79 y, 3946 cases FFQ, >300 foods and beverages Low-fat dairy products 2.8 vs 0.05 servings/d 0.65 (0.44, 0.96) Age, race-ethnicity, total energy intake, income, education, BMI, smoking, alcohol intake, FH of DM, HRT, SBP, DBP, physical activity, interaction of low-fat dairy × BMI, interaction of yogurt × time Total dairy products 3.4 vs 0.5 servings/d 0.93 (0.83, 1.04) Yogurt ≥2 servings/wk vs <1 serving/mo 0.46 (0.31, 0.68) Sluijs, 2012, Europe (21) EPIC-Interact Study 1991–2007, 11.7-y follow-up 24,475 men and women, mean age 52 y, 10,694 cases Validated FFQ, 24-h recall Total dairy products 628.9 vs 79.7 g/d 1.01 (0.89, 1.13) Age; center; sex; BMI; education; smoking status; physical activity; intakes of alcohol, fruit and vegetables, red meat, processed meat, sugar-sweetened soft drinks, coffee, cereals, cereal products, energy Milk 486.1 vs 0.3 g/d 1.10 (0.92, 1.31) Yogurt and thick, fermented milk 190.4 vs 0 g/d 0.91 (0.81, 1.02) Cheese 73.7 vs 3.2 g/d 0.88 (0.76, 1.02) Combined fermented dairy intake 220.7 vs 11.6 g/d 0.88 (0.78, 0.99) Struijk, 2012, Denmark (22) The Inter99 Study 1999/2001–2006, 5-y follow-up 5953 men and women, age 30–60 y, 214 cases Validated FFQ, 198 food items Total dairy 578 vs 47 g/d 0.96 (0.58, 1.58) Age; sex; intervention group; FH of DM; education; physical activity; smoking status; intakes of alcohol, whole-grain cereal, meat, fish, coffee, tea, fruit, vegetables, energy; change in diet from baseline to 5-y follow-up; waist circumference Per 1 serving/d 0.95 (0.86, 1.06) Low-fat dairy 536 vs 6 g/d 0.85 (0.52, 1.40) Per 1 serving/d 0.95 (0.85, 1.06) Full-fat dairy 89 vs 4 g/d 0.94 (0.56, 1.58) Per 1 serving/d 1.03 (0.77, 1.36) Milk and milk products 546 vs 16 g/d 0.95 (0.58, 1.57) Per 1 serving/d 0.96 (0.86, 1.06) Cheese 49 vs 4 g/d 0.78 (0.47, 1.29) Per 1 serving/d 0.97 (0.82, 1.15) Fermented dairy 260 vs 13 g/d 0.86 (0.50, 1.47) Per 1 serving/d 0.88 (0.69, 1.11) Louie, 2012, Australia (23) The Blue Mountains Eye Study 1992/1994–2002/2004, 10-y follow-up 1824 men and women, age ≥49 y, 145 cases Validated FFQ, 145 food items Total dairy 3.1 vs 0.5 servings/d 0.96 (0.54, 1.17) Age, sex, smoking status, physical activity, GL, vegetable fiber, total energy, FH of DM2, SBP, BMI, HDL, total cholesterol, TG Reduced/low-fat dairy 2.1 vs 0 servings/d 1.00 (0.58, 1.71) Regular dairy 1.9 vs 0.1 servings/d 0.83 (0.47, 1.48) Grantham, 2013, Australia (14) The Australian Diabetes Obesity and Lifestyle Study 1999/2000–2004/2005, 5-y follow-up 5582 men and women, age ≥25 y, 209 cases FFQ, 121 food items Total dairy foods 477 vs 205 g/d2 0.71 (0.48, 1.05) Age, sex, energy intake, FH DM, education, physical activity, smoking status, TG, HDL cholesterol, SBP, waist circumference, hip circumference Low-fat milk 375 vs 0 g/d2 0.65 (0.44, 0.94) Full-fat milk 375 vs 0 g/d2 1.18 (0.78, 1.79) Yogurt 114 vs 3 g/d2 1.14 (0.78, 1.67) Cheese 29 vs 4 g/d2 0.78 (0.53, 1.15) von Ruesten, 2013, Germany (24) EPIC-Potsdam Study 1994/1998–NA, 8-y follow-up 23,531 men and women, age 35–65 y, 837 cases Validated FFQ, 148 food items Low-fat dairy products Per 100 g/d 1.02 (0.96, 1.09) Age, sex, smoking status, pack-years of smoking, alcohol, leisure-time physical activity, BMI, WHR, hypertension, high blood lipids, education, vitamin supplementation, total energy intake, other food groups High-fat dairy products Per 100 g/d 1.00 (0.92, 1.08) Low-fat cheese Per 30 g/d 0.98 (0.83, 1.15) High-fat cheese Per 30 g/d 0.96 (0.85, 1.08) Soedamah-Muthu, 2013, UK (20) The Whitehall II Prospective Study 1985/1988–2009, 10-y follow-up 4526 men and women, 35–55 y, 273 cases FFQ, 114 food items Total dairy 575 vs 246 g/d 1.30 (0.95, 1.77) Age; ethnicity; employment grade; smoking; BMI; physical activity; FH of CHD/hypertension; intakes of alcohol, fruit and vegetables, bread, meat, fish, coffee, tea, total energy High-fat dairy 182 vs 27 g/d 1.23 (0.91, 1.67) Low-fat dairy 458 vs 28 g/d 0.98 (0.73, 1.31) Total milk 441 vs 147 g/d 0.97 (0.71, 1.32) Fermented dairy 105 vs 17 g/d 1.17 (0.87, 1.58) Yogurt 117 vs 0 g/d 1.04 (0.77, 1.42) Cheese 31 vs 6 g/d 1.20 (0.88, 1.64) View Large

TABLE 1 First author, publication year, country/ region (ref) Study name Follow-up period Study size, sex, age, number of cases Dietary assessment Exposure Quantity RR (95% CI) Adjustment for confounders Choi, 2005, USA (7) Health Professionals Follow-Up Study 1986–1998, 12-y follow-up 41,254 men, age 40–75 y, 1243 cases Validated FFQ, 131 food items Total dairy intake ≥2.9 vs <0.9 servings/d 0.77 (0.62, 0.95) Age, total energy intake, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, alcohol intake, cereal fiber, trans FAs, PUFA:SFA ratio, glycemic load Per 1 serving/d 0.91 (0.85, 0.97) Low-fat dairy foods ≥1.58 vs <0.14 servings/d 0.74 (0.60, 0.91) Per 1 serving/d 0.88 (0.81, 0.94) High-fat dairy foods ≥1.72 vs <0.38 servings/d 0.97 (0.78, 1.21) Per 1 serving/d 0.99 (0.91, 1.07) Skim, low-fat milk ≥2 servings/d vs <1 serving/mo 0.78 (0.63, 0.97) Whole milk ≥2 servings/wk vs <1 serving/mo 1.19 (1.00, 1.43) Yogurt ≥2 servings/wk vs <1 serving/mo 0.83 (0.66, 1.06) Sherbet ≥2 servings/wk vs <1 serving/mo 0.85 (0.66, 1.09) Cottage, ricotta cheese ≥2 servings/wk vs <1 serving/mo 0.96 (0.80, 1.17) Ice cream ≥2 servings/wk vs <1 serving/mo 0.78 (0.64, 0.95) Other cheese ≥2 servings/wk vs <1 serving/mo 0.88 (0.67, 1.16) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.06 (0.81, 1.39) Cream ≥2 servings/wk vs <1 serving/mo 0.89 (0.72, 1.09) Sour cream ≥2 servings/wk vs <1 serving/mo 1.04 (0.80, 1.36) Montonen, 2005, Finland (8) Finnish Mobile Clinic Health Examination Survey 1967–1990, 23-y follow-up 4304 men and women, age 40–69 y, 383 cases Dietary history interview, >100 food items Regular dairy products ≥305 vs <39 g/d 0.81 (0.62, 1.08) Age, sex, BMI, energy intake, smoking, FH of DM, geographic area Reduced-fat dairy products >0 vs 0 g/d 0.90 (0.60, 1.36) Whole milk ≥878 vs <326 g/d 1.06 (0.75, 1.50) Butter >59 vs <27 g/d 1.15 (0.80, 1.67) Pittas, 2006, USA (9) Nurses’ Health Study 1980–2000, 20-y follow-up 83,779 women, age 30–55 y, 4843 cases Validated FFQ, 61–116 food items Dairy foods ≥3 vs <1 serving/d 0.89 (0.81, 0.99) Age, BMI, hypertension, FH of DM, smoking, physical activity, caffeine, alcohol, residence, SFAs, PUFAs, trans FAs, cereal fiber, Mg, GL, retinol, energy Liu, 2006, USA (10) Women’s Health Study 1993–2003, 10-y follow-up 37,183 women, age ≥45 y, 1603 cases Validated FFQ, 131 food and beverage items Total dairy intake >2.9 vs <0.85 servings/d 0.80 (0.67, 0.95) Age, total energy intake, randomized treatment assignment, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, hormones, alcohol, dietary fiber, total fat, GL Per 1 serving/d 0.97 (0.93, 1.01) Low-fat dairy foods >2.00 vs 0.27 servings/d 0.82 (0.68, 0.98) Per 1 serving/d 0.95 (0.90, 1.01) High-fat dairy foods >1.329 vs <0.20 servings/d 0.97 (0.80, 1.17) Per 1 serving/d 1.00 (0.95, 1.04) Skim milk ≥2 servings/wk vs <1 serving/mo 0.92 (0.78, 1.09) Whole milk ≥2 servings/wk vs <1 serving/mo 1.04 (0.84, 1.30) Yogurt ≥2 servings/wk vs <1 serving/mo 0.82 (0.70, 0.97) Sherbet ≥2 servings/wk vs <1 serving/mo 0.92 (0.77, 1.05) Cottage cheese ≥2 servings/wk vs <1 serving/mo 0.86 (0.71, 1.05) Ice cream ≥2 servings/wk vs <1 serving/mo 0.88 (0.74, 1.05) Other cheese ≥2 servings/wk vs <1 serving/mo 0.80 (0.64, 1.01) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.19 (0.97, 1.47) Cream ≥2 servings/wk vs <1 serving/mo 1.03 (0.86, 1.25) Sour cream ≥2 servings/wk vs <1 serving/mo 0.93 (0.74, 1.18) van Dam, 2006, USA (15) Black Women’s Health Study 1995–2003, 8-y follow-up 41,186 women, age 21–69 y, 1964 cases Validated FFQ, 68 food items Total dairy products 2.53 vs 0.07 servings/d 0.93 (0.75, 1.15) Age, total energy intake, alcohol, BMI, smoking status, strenuous physical activity, parental history of DM, education, coffee, sugar-sweetened soft drinks, processed meat, red meat, whole grains Low-fat dairy products 1.33 vs 0.07 servings/d 0.87 (0.76, 1.00) High-fat dairy products 1.33 vs 0.07 servings/d 1.03 (0.88, 1.20) Lecomte, 2007, France (11) NA 1995/1997–2000/2002, 5-y follow-up 743 men with IFG, age 20–60 y, 127 cases FFQ, 18 food items Dairy products Not daily vs daily 1.86 (1.21, 2.86) FH of DM, BMI, TG, glucose at baseline Elwood, 2007, United Kingdom (16) Caerphilly Prospective Study 1979/1983–NA, 20-y follow-up 640 men, age 45–59 y, 41 cases 7-d weighed records Milk 396.8 vs 76.5 g/d2 0.57 (0.20, 1.63) Age, smoking, BMI, social class Dairy foods 430.4 vs 107.0 g/d2 0.59 (0.21, 1.69)2 Vang, 2008, USA (17) Adventist Health Study 1960–1976, 17-y follow-up 8401 men and women, age 45–88 y, 543 cases Validated FFQ Cheese ≥1 time/wk vs never 0.90 (0.71, 1.15) Age, sex Milk ≥4 times/d vs never 0.92 (0.65, 1.29) Villegas, 2009, China (12) Shanghai Women’s Health Study 2000–2006, 6.9-y follow-up 64,191 women, age 40–70 y, 2270 cases (1514 confirmed cases) Validated FFQ, 77 food items Fresh milk, all participants 250.0 vs 0 g/d 0.46 (0.32, 0.64) Age, energy intake, BMI, WHR, smoking status, alcohol, physical activity, income, education, occupation, hypertension Powdered milk Yes vs no 0.74 (0.67, 0.82) Fresh milk, confirmed cases 250.0 vs 0 g/d 0.60 (0.41, 0.88) Powdered milk Yes vs no 0.85 (0.75, 0.96) Kirii, 2009, Japan (18) Japan Public Health Center–based Prospective Study 1995/1998–2000/2003, 5-y follow-up 59,796 men and women, age 40–59 y, 1114 cases Validated FFQ, 147 food and beverage items Dairy products, men ≥300 vs <50 g/d 1.18 (0.90, 1.56) Age, area, BMI, FH of DM, smoking status, alcohol intake, history of hypertension, exercise frequency, coffee consumption, energy-adjusted Mg, total energy intake Milk ≥200 vs <50 g/d 1.02 (0.85, 1.24) Cheese ≥5 vs 0 g/d 0.88 (0.64, 1.21) Yogurt ≥60 vs 0 g/d 1.01 (0.75, 1.36) Dairy products, women ≥300 vs <50 g/d 0.71 (0.51, 0.98) Milk ≥200 vs <50 g/d 0.87 (0.70, 1.09) Cheese ≥5 vs 0 g/d 1.12 (0.80, 1.57) Yogurt ≥60 vs 0 g/d 0.77 (0.58, 1.01) Malik, 2011, USA (13) Nurses’ Health Study II 1997–2005, 7-y follow-up 37,038 women, age 24–42 y, 550 cases Validated FFQ, 133 food items Total dairy 2.14 vs 0.62 servings per 1000 kcal/d 0.75 (0.55, 1.02) Age, BMI, total energy, FH of DM, smoking status, physical activity, alcohol use, OC use, HRT, PUFA:SFA ratio, GL, cereal fiber, trans fat, processed meat, carbonated soft drinks, fruit drinks, coffee; mutual adjustment: high- and low-fat dairy products Low-fat dairy foods 1.44 vs 0.18 servings per 1000 kcal/d 0.74 (0.54, 1.01) High-fat dairy foods 1.14 vs 0.19 servings per 1000 kcal/d 0.72 (0.53, 0.99) Margolis, 2011, USA (19) Women’s Health Initiative 1994/1998–2005, 7.9-y follow-up 82,076 women, age 50–79 y, 3946 cases FFQ, >300 foods and beverages Low-fat dairy products 2.8 vs 0.05 servings/d 0.65 (0.44, 0.96) Age, race-ethnicity, total energy intake, income, education, BMI, smoking, alcohol intake, FH of DM, HRT, SBP, DBP, physical activity, interaction of low-fat dairy × BMI, interaction of yogurt × time Total dairy products 3.4 vs 0.5 servings/d 0.93 (0.83, 1.04) Yogurt ≥2 servings/wk vs <1 serving/mo 0.46 (0.31, 0.68) Sluijs, 2012, Europe (21) EPIC-Interact Study 1991–2007, 11.7-y follow-up 24,475 men and women, mean age 52 y, 10,694 cases Validated FFQ, 24-h recall Total dairy products 628.9 vs 79.7 g/d 1.01 (0.89, 1.13) Age; center; sex; BMI; education; smoking status; physical activity; intakes of alcohol, fruit and vegetables, red meat, processed meat, sugar-sweetened soft drinks, coffee, cereals, cereal products, energy Milk 486.1 vs 0.3 g/d 1.10 (0.92, 1.31) Yogurt and thick, fermented milk 190.4 vs 0 g/d 0.91 (0.81, 1.02) Cheese 73.7 vs 3.2 g/d 0.88 (0.76, 1.02) Combined fermented dairy intake 220.7 vs 11.6 g/d 0.88 (0.78, 0.99) Struijk, 2012, Denmark (22) The Inter99 Study 1999/2001–2006, 5-y follow-up 5953 men and women, age 30–60 y, 214 cases Validated FFQ, 198 food items Total dairy 578 vs 47 g/d 0.96 (0.58, 1.58) Age; sex; intervention group; FH of DM; education; physical activity; smoking status; intakes of alcohol, whole-grain cereal, meat, fish, coffee, tea, fruit, vegetables, energy; change in diet from baseline to 5-y follow-up; waist circumference Per 1 serving/d 0.95 (0.86, 1.06) Low-fat dairy 536 vs 6 g/d 0.85 (0.52, 1.40) Per 1 serving/d 0.95 (0.85, 1.06) Full-fat dairy 89 vs 4 g/d 0.94 (0.56, 1.58) Per 1 serving/d 1.03 (0.77, 1.36) Milk and milk products 546 vs 16 g/d 0.95 (0.58, 1.57) Per 1 serving/d 0.96 (0.86, 1.06) Cheese 49 vs 4 g/d 0.78 (0.47, 1.29) Per 1 serving/d 0.97 (0.82, 1.15) Fermented dairy 260 vs 13 g/d 0.86 (0.50, 1.47) Per 1 serving/d 0.88 (0.69, 1.11) Louie, 2012, Australia (23) The Blue Mountains Eye Study 1992/1994–2002/2004, 10-y follow-up 1824 men and women, age ≥49 y, 145 cases Validated FFQ, 145 food items Total dairy 3.1 vs 0.5 servings/d 0.96 (0.54, 1.17) Age, sex, smoking status, physical activity, GL, vegetable fiber, total energy, FH of DM2, SBP, BMI, HDL, total cholesterol, TG Reduced/low-fat dairy 2.1 vs 0 servings/d 1.00 (0.58, 1.71) Regular dairy 1.9 vs 0.1 servings/d 0.83 (0.47, 1.48) Grantham, 2013, Australia (14) The Australian Diabetes Obesity and Lifestyle Study 1999/2000–2004/2005, 5-y follow-up 5582 men and women, age ≥25 y, 209 cases FFQ, 121 food items Total dairy foods 477 vs 205 g/d2 0.71 (0.48, 1.05) Age, sex, energy intake, FH DM, education, physical activity, smoking status, TG, HDL cholesterol, SBP, waist circumference, hip circumference Low-fat milk 375 vs 0 g/d2 0.65 (0.44, 0.94) Full-fat milk 375 vs 0 g/d2 1.18 (0.78, 1.79) Yogurt 114 vs 3 g/d2 1.14 (0.78, 1.67) Cheese 29 vs 4 g/d2 0.78 (0.53, 1.15) von Ruesten, 2013, Germany (24) EPIC-Potsdam Study 1994/1998–NA, 8-y follow-up 23,531 men and women, age 35–65 y, 837 cases Validated FFQ, 148 food items Low-fat dairy products Per 100 g/d 1.02 (0.96, 1.09) Age, sex, smoking status, pack-years of smoking, alcohol, leisure-time physical activity, BMI, WHR, hypertension, high blood lipids, education, vitamin supplementation, total energy intake, other food groups High-fat dairy products Per 100 g/d 1.00 (0.92, 1.08) Low-fat cheese Per 30 g/d 0.98 (0.83, 1.15) High-fat cheese Per 30 g/d 0.96 (0.85, 1.08) Soedamah-Muthu, 2013, UK (20) The Whitehall II Prospective Study 1985/1988–2009, 10-y follow-up 4526 men and women, 35–55 y, 273 cases FFQ, 114 food items Total dairy 575 vs 246 g/d 1.30 (0.95, 1.77) Age; ethnicity; employment grade; smoking; BMI; physical activity; FH of CHD/hypertension; intakes of alcohol, fruit and vegetables, bread, meat, fish, coffee, tea, total energy High-fat dairy 182 vs 27 g/d 1.23 (0.91, 1.67) Low-fat dairy 458 vs 28 g/d 0.98 (0.73, 1.31) Total milk 441 vs 147 g/d 0.97 (0.71, 1.32) Fermented dairy 105 vs 17 g/d 1.17 (0.87, 1.58) Yogurt 117 vs 0 g/d 1.04 (0.77, 1.42) Cheese 31 vs 6 g/d 1.20 (0.88, 1.64) First author, publication year, country/ region (ref) Study name Follow-up period Study size, sex, age, number of cases Dietary assessment Exposure Quantity RR (95% CI) Adjustment for confounders Choi, 2005, USA (7) Health Professionals Follow-Up Study 1986–1998, 12-y follow-up 41,254 men, age 40–75 y, 1243 cases Validated FFQ, 131 food items Total dairy intake ≥2.9 vs <0.9 servings/d 0.77 (0.62, 0.95) Age, total energy intake, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, alcohol intake, cereal fiber, trans FAs, PUFA:SFA ratio, glycemic load Per 1 serving/d 0.91 (0.85, 0.97) Low-fat dairy foods ≥1.58 vs <0.14 servings/d 0.74 (0.60, 0.91) Per 1 serving/d 0.88 (0.81, 0.94) High-fat dairy foods ≥1.72 vs <0.38 servings/d 0.97 (0.78, 1.21) Per 1 serving/d 0.99 (0.91, 1.07) Skim, low-fat milk ≥2 servings/d vs <1 serving/mo 0.78 (0.63, 0.97) Whole milk ≥2 servings/wk vs <1 serving/mo 1.19 (1.00, 1.43) Yogurt ≥2 servings/wk vs <1 serving/mo 0.83 (0.66, 1.06) Sherbet ≥2 servings/wk vs <1 serving/mo 0.85 (0.66, 1.09) Cottage, ricotta cheese ≥2 servings/wk vs <1 serving/mo 0.96 (0.80, 1.17) Ice cream ≥2 servings/wk vs <1 serving/mo 0.78 (0.64, 0.95) Other cheese ≥2 servings/wk vs <1 serving/mo 0.88 (0.67, 1.16) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.06 (0.81, 1.39) Cream ≥2 servings/wk vs <1 serving/mo 0.89 (0.72, 1.09) Sour cream ≥2 servings/wk vs <1 serving/mo 1.04 (0.80, 1.36) Montonen, 2005, Finland (8) Finnish Mobile Clinic Health Examination Survey 1967–1990, 23-y follow-up 4304 men and women, age 40–69 y, 383 cases Dietary history interview, >100 food items Regular dairy products ≥305 vs <39 g/d 0.81 (0.62, 1.08) Age, sex, BMI, energy intake, smoking, FH of DM, geographic area Reduced-fat dairy products >0 vs 0 g/d 0.90 (0.60, 1.36) Whole milk ≥878 vs <326 g/d 1.06 (0.75, 1.50) Butter >59 vs <27 g/d 1.15 (0.80, 1.67) Pittas, 2006, USA (9) Nurses’ Health Study 1980–2000, 20-y follow-up 83,779 women, age 30–55 y, 4843 cases Validated FFQ, 61–116 food items Dairy foods ≥3 vs <1 serving/d 0.89 (0.81, 0.99) Age, BMI, hypertension, FH of DM, smoking, physical activity, caffeine, alcohol, residence, SFAs, PUFAs, trans FAs, cereal fiber, Mg, GL, retinol, energy Liu, 2006, USA (10) Women’s Health Study 1993–2003, 10-y follow-up 37,183 women, age ≥45 y, 1603 cases Validated FFQ, 131 food and beverage items Total dairy intake >2.9 vs <0.85 servings/d 0.80 (0.67, 0.95) Age, total energy intake, randomized treatment assignment, FH of DM, smoking status, BMI, hypercholesterolemia, hypertension, physical activity, hormones, alcohol, dietary fiber, total fat, GL Per 1 serving/d 0.97 (0.93, 1.01) Low-fat dairy foods >2.00 vs 0.27 servings/d 0.82 (0.68, 0.98) Per 1 serving/d 0.95 (0.90, 1.01) High-fat dairy foods >1.329 vs <0.20 servings/d 0.97 (0.80, 1.17) Per 1 serving/d 1.00 (0.95, 1.04) Skim milk ≥2 servings/wk vs <1 serving/mo 0.92 (0.78, 1.09) Whole milk ≥2 servings/wk vs <1 serving/mo 1.04 (0.84, 1.30) Yogurt ≥2 servings/wk vs <1 serving/mo 0.82 (0.70, 0.97) Sherbet ≥2 servings/wk vs <1 serving/mo 0.92 (0.77, 1.05) Cottage cheese ≥2 servings/wk vs <1 serving/mo 0.86 (0.71, 1.05) Ice cream ≥2 servings/wk vs <1 serving/mo 0.88 (0.74, 1.05) Other cheese ≥2 servings/wk vs <1 serving/mo 0.80 (0.64, 1.01) Cream cheese ≥2 servings/wk vs <1 serving/mo 1.19 (0.97, 1.47) Cream ≥2 servings/wk vs <1 serving/mo 1.03 (0.86, 1.25) Sour cream ≥2 servings/wk vs <1 serving/mo 0.93 (0.74, 1.18) van Dam, 2006, USA (15) Black Women’s Health Study 1995–2003, 8-y follow-up 41,186 women, age 21–69 y, 1964 cases Validated FFQ, 68 food items Total dairy products 2.53 vs 0.07 servings/d 0.93 (0.75, 1.15) Age, total energy intake, alcohol, BMI, smoking status, strenuous physical activity, parental history of DM, education, coffee, sugar-sweetened soft drinks, processed meat, red meat, whole grains Low-fat dairy products 1.33 vs 0.07 servings/d 0.87 (0.76, 1.00) High-fat dairy products 1.33 vs 0.07 servings/d 1.03 (0.88, 1.20) Lecomte, 2007, France (11) NA 1995/1997–2000/2002, 5-y follow-up 743 men with IFG, age 20–60 y, 127 cases FFQ, 18 food items Dairy products Not daily vs daily 1.86 (1.21, 2.86) FH of DM, BMI, TG, glucose at baseline Elwood, 2007, United Kingdom (16) Caerphilly Prospective Study 1979/1983–NA, 20-y follow-up 640 men, age 45–59 y, 41 cases 7-d weighed records Milk 396.8 vs 76.5 g/d2 0.57 (0.20, 1.63) Age, smoking, BMI, social class Dairy foods 430.4 vs 107.0 g/d2 0.59 (0.21, 1.69)2 Vang, 2008, USA (17) Adventist Health Study 1960–1976, 17-y follow-up 8401 men and women, age 45–88 y, 543 cases Validated FFQ Cheese ≥1 time/wk vs never 0.90 (0.71, 1.15) Age, sex Milk ≥4 times/d vs never 0.92 (0.65, 1.29) Villegas, 2009, China (12) Shanghai Women’s Health Study 2000–2006, 6.9-y follow-up 64,191 women, age 40–70 y, 2270 cases (1514 confirmed cases) Validated FFQ, 77 food items Fresh milk, all participants 250.0 vs 0 g/d 0.46 (0.32, 0.64) Age, energy intake, BMI, WHR, smoking status, alcohol, physical activity, income, education, occupation, hypertension Powdered milk Yes vs no 0.74 (0.67, 0.82) Fresh milk, confirmed cases 250.0 vs 0 g/d 0.60 (0.41, 0.88) Powdered milk Yes vs no 0.85 (0.75, 0.96) Kirii, 2009, Japan (18) Japan Public Health Center–based Prospective Study 1995/1998–2000/2003, 5-y follow-up 59,796 men and women, age 40–59 y, 1114 cases Validated FFQ, 147 food and beverage items Dairy products, men ≥300 vs <50 g/d 1.18 (0.90, 1.56) Age, area, BMI, FH of DM, smoking status, alcohol intake, history of hypertension, exercise frequency, coffee consumption, energy-adjusted Mg, total energy intake Milk ≥200 vs <50 g/d 1.02 (0.85, 1.24) Cheese ≥5 vs 0 g/d 0.88 (0.64, 1.21) Yogurt ≥60 vs 0 g/d 1.01 (0.75, 1.36) Dairy products, women ≥300 vs <50 g/d 0.71 (0.51, 0.98) Milk ≥200 vs <50 g/d 0.87 (0.70, 1.09) Cheese ≥5 vs 0 g/d 1.12 (0.80, 1.57) Yogurt ≥60 vs 0 g/d 0.77 (0.58, 1.01) Malik, 2011, USA (13) Nurses’ Health Study II 1997–2005, 7-y follow-up 37,038 women, age 24–42 y, 550 cases Validated FFQ, 133 food items Total dairy 2.14 vs 0.62 servings per 1000 kcal/d 0.75 (0.55, 1.02) Age, BMI, total energy, FH of DM, smoking status, physical activity, alcohol use, OC use, HRT, PUFA:SFA ratio, GL, cereal fiber, trans fat, processed meat, carbonated soft drinks, fruit drinks, coffee; mutual adjustment: high- and low-fat dairy products Low-fat dairy foods 1.44 vs 0.18 servings per 1000 kcal/d 0.74 (0.54, 1.01) High-fat dairy foods 1.14 vs 0.19 servings per 1000 kcal/d 0.72 (0.53, 0.99) Margolis, 2011, USA (19) Women’s Health Initiative 1994/1998–2005, 7.9-y follow-up 82,076 women, age 50–79 y, 3946 cases FFQ, >300 foods and beverages Low-fat dairy products 2.8 vs 0.05 servings/d 0.65 (0.44, 0.96) Age, race-ethnicity, total energy intake, income, education, BMI, smoking, alcohol intake, FH of DM, HRT, SBP, DBP, physical activity, interaction of low-fat dairy × BMI, interaction of yogurt × time Total dairy products 3.4 vs 0.5 servings/d 0.93 (0.83, 1.04) Yogurt ≥2 servings/wk vs <1 serving/mo 0.46 (0.31, 0.68) Sluijs, 2012, Europe (21) EPIC-Interact Study 1991–2007, 11.7-y follow-up 24,475 men and women, mean age 52 y, 10,694 cases Validated FFQ, 24-h recall Total dairy products 628.9 vs 79.7 g/d 1.01 (0.89, 1.13) Age; center; sex; BMI; education; smoking status; physical activity; intakes of alcohol, fruit and vegetables, red meat, processed meat, sugar-sweetened soft drinks, coffee, cereals, cereal products, energy Milk 486.1 vs 0.3 g/d 1.10 (0.92, 1.31) Yogurt and thick, fermented milk 190.4 vs 0 g/d 0.91 (0.81, 1.02) Cheese 73.7 vs 3.2 g/d 0.88 (0.76, 1.02) Combined fermented dairy intake 220.7 vs 11.6 g/d 0.88 (0.78, 0.99) Struijk, 2012, Denmark (22) The Inter99 Study 1999/2001–2006, 5-y follow-up 5953 men and women, age 30–60 y, 214 cases Validated FFQ, 198 food items Total dairy 578 vs 47 g/d 0.96 (0.58, 1.58) Age; sex; intervention group; FH of DM; education; physical activity; smoking status; intakes of alcohol, whole-grain cereal, meat, fish, coffee, tea, fruit, vegetables, energy; change in diet from baseline to 5-y follow-up; waist circumference Per 1 serving/d 0.95 (0.86, 1.06) Low-fat dairy 536 vs 6 g/d 0.85 (0.52, 1.40) Per 1 serving/d 0.95 (0.85, 1.06) Full-fat dairy 89 vs 4 g/d 0.94 (0.56, 1.58) Per 1 serving/d 1.03 (0.77, 1.36) Milk and milk products 546 vs 16 g/d 0.95 (0.58, 1.57) Per 1 serving/d 0.96 (0.86, 1.06) Cheese 49 vs 4 g/d 0.78 (0.47, 1.29) Per 1 serving/d 0.97 (0.82, 1.15) Fermented dairy 260 vs 13 g/d 0.86 (0.50, 1.47) Per 1 serving/d 0.88 (0.69, 1.11) Louie, 2012, Australia (23) The Blue Mountains Eye Study 1992/1994–2002/2004, 10-y follow-up 1824 men and women, age ≥49 y, 145 cases Validated FFQ, 145 food items Total dairy 3.1 vs 0.5 servings/d 0.96 (0.54, 1.17) Age, sex, smoking status, physical activity, GL, vegetable fiber, total energy, FH of DM2, SBP, BMI, HDL, total cholesterol, TG Reduced/low-fat dairy 2.1 vs 0 servings/d 1.00 (0.58, 1.71) Regular dairy 1.9 vs 0.1 servings/d 0.83 (0.47, 1.48) Grantham, 2013, Australia (14) The Australian Diabetes Obesity and Lifestyle Study 1999/2000–2004/2005, 5-y follow-up 5582 men and women, age ≥25 y, 209 cases FFQ, 121 food items Total dairy foods 477 vs 205 g/d2 0.71 (0.48, 1.05) Age, sex, energy intake, FH DM, education, physical activity, smoking status, TG, HDL cholesterol, SBP, waist circumference, hip circumference Low-fat milk 375 vs 0 g/d2 0.65 (0.44, 0.94) Full-fat milk 375 vs 0 g/d2 1.18 (0.78, 1.79) Yogurt 114 vs 3 g/d2 1.14 (0.78, 1.67) Cheese 29 vs 4 g/d2 0.78 (0.53, 1.15) von Ruesten, 2013, Germany (24) EPIC-Potsdam Study 1994/1998–NA, 8-y follow-up 23,531 men and women, age 35–65 y, 837 cases Validated FFQ, 148 food items Low-fat dairy products Per 100 g/d 1.02 (0.96, 1.09) Age, sex, smoking status, pack-years of smoking, alcohol, leisure-time physical activity, BMI, WHR, hypertension, high blood lipids, education, vitamin supplementation, total energy intake, other food groups High-fat dairy products Per 100 g/d 1.00 (0.92, 1.08) Low-fat cheese Per 30 g/d 0.98 (0.83, 1.15) High-fat cheese Per 30 g/d 0.96 (0.85, 1.08) Soedamah-Muthu, 2013, UK (20) The Whitehall II Prospective Study 1985/1988–2009, 10-y follow-up 4526 men and women, 35–55 y, 273 cases FFQ, 114 food items Total dairy 575 vs 246 g/d 1.30 (0.95, 1.77) Age; ethnicity; employment grade; smoking; BMI; physical activity; FH of CHD/hypertension; intakes of alcohol, fruit and vegetables, bread, meat, fish, coffee, tea, total energy High-fat dairy 182 vs 27 g/d 1.23 (0.91, 1.67) Low-fat dairy 458 vs 28 g/d 0.98 (0.73, 1.31) Total milk 441 vs 147 g/d 0.97 (0.71, 1.32) Fermented dairy 105 vs 17 g/d 1.17 (0.87, 1.58) Yogurt 117 vs 0 g/d 1.04 (0.77, 1.42) Cheese 31 vs 6 g/d 1.20 (0.88, 1.64) View Large

Total dairy products

Fourteen cohort studies (7, 9–11, 13–16, 18–23) investigated the association between total dairy product intake and type 2 diabetes risk and included 26,976 cases among 426,055 participants. Twelve of these studies were included in the dose-response analysis (7, 9–11, 13–16, 18–23), and 2 studies were excluded because they only reported results for a high and low comparison (9, 11). The summary RR for high compared with low intake was 0.89 (95% CI: 0.82, 0.96), with moderate heterogeneity between studies (I2 = 42.1% and P-heterogeneity = 0.05) (see Supplemental Figure S1 under “Supplemental data” in the online issue). The summary RR per 400 g/d was 0.93 (95% CI: 0.87, 0.99; I2 = 33.1% and P-heterogeneity = 0.13) (Figure 2A). In sensitivity analyses excluding one study at a time from the analysis, the summary RRs for type 2 diabetes ranged from 0.91 (95% CI: 0.85, 0.97) when the EPIC-Interact Study (21) was excluded to 0.95 (95% CI: 0.90, 1.00) when the Health Professionals Follow-Up Study (7) was excluded. There was no indication of publication bias with Egger’s test (P = 0.20). There was evidence of a nonlinear association between dairy product intake and type 2 diabetes, and most of the benefit was observed when increasing the intake from low amounts (P-nonlinearity < 0.0001), and there was no further reduction in risk above an intake of 300–400 g/d (seeFigure 2B and Supplemental Table S1 under “Supplemental data” in the online issue).

FIGURE 2. View largeDownload slide A, B: Intake of total dairy products and risk of type 2 diabetes. The summary RR per 400 g/d by using random-effects models was 0.93 (95% CI: 0.87, 0.99; I2 = 33.1%; P-heterogeneity = 0.13; n = 341,533). There was evidence of a nonlinear association between total dairy products and type 2 diabetes, P-nonlinearity < 0.0001.

FIGURE 2. View largeDownload slide A, B: Intake of total dairy products and risk of type 2 diabetes. The summary RR per 400 g/d by using random-effects models was 0.93 (95% CI: 0.87, 0.99; I2 = 33.1%; P-heterogeneity = 0.13; n = 341,533). There was evidence of a nonlinear association between total dairy products and type 2 diabetes, P-nonlinearity < 0.0001.

High-fat dairy products

Nine cohort studies (7, 8, 10, 13, 15, 20, 22–24) investigated the association between intake of high-fat dairy products and type 2 diabetes risk and included 7222 cases among 196,799 participants. One of these studies was only included in the dose-response analysis because it reported only a continuous result (24). The summary RR for high compared with low intake was 0.96 (95% CI: 0.87, 1.06; I2 = 15.8% and P-heterogeneity = 0.31) (see Supplemental Figure S2 under “Supplemental data” in the online issue). The summary RR per 200 g/d was 0.98 (95% CI: 0.94, 1.03; I2 = 7.6% and P-heterogeneity = 0.37) (Figure 3A). The summary RRs ranged from 0.97 (95% CI: 0.91, 1.03) when the Women’s Health Study was excluded (10) to 0.99 (95% CI: 0.96, 1.03) when the Nurses’ Health Study II (13) was excluded. There was no evidence of publication bias (P = 0.77), and there was no evidence of a nonlinear association (P-nonlinearity = 0.57) (Figure 3B and Supplemental Table S2 under “Supplemental data” in the online issue).

FIGURE 3. View largeDownload slide A–D: Intake of high- and low-fat dairy products and risk of type 2 diabetes. The summary RRs per 200 g/d by using random-effects models were 0.98 (95% CI: 0.94, 1.03; I2 = 7.6%; P-heterogeneity = 0.37; n = 196,799) for high-fat dairy products (A) and 0.91 (95% CI: 0.86, 0.96; I2 = 40.2%; P-heterogeneity = 0.10; n = 274,571) for low-fat dairy products (C). There was an indication of a nonlinear association between low-fat dairy products and type 2 diabetes (P-nonlinearity = 0.06) (D) but not between high-fat dairy products and type 2 diabetes (P-nonlinearity = 0.57) (B).

FIGURE 3. View largeDownload slide A–D: Intake of high- and low-fat dairy products and risk of type 2 diabetes. The summary RRs per 200 g/d by using random-effects models were 0.98 (95% CI: 0.94, 1.03; I2 = 7.6%; P-heterogeneity = 0.37; n = 196,799) for high-fat dairy products (A) and 0.91 (95% CI: 0.86, 0.96; I2 = 40.2%; P-heterogeneity = 0.10; n = 274,571) for low-fat dairy products (C). There was an indication of a nonlinear association between low-fat dairy products and type 2 diabetes (P-nonlinearity = 0.06) (D) but not between high-fat dairy products and type 2 diabetes (P-nonlinearity = 0.57) (B).

Low-fat dairy products

Ten cohort studies (7, 8, 10, 13, 15, 19, 20, 22–24) were included in the analysis of low-fat dairy products and type 2 diabetes risk and included 11,168 cases among 278,875 participants. Nine of the studies were included in the dose-response analysis (7, 10, 13, 15, 19, 20, 22–24). One of the studies was excluded from the dose-response analysis because results were presented in only 2 categories (8), and one study was excluded from the high compared with low analysis because it presented only continuous results (24). The summary RR for high compared with low intake was 0.83 (95% CI: 0.76, 0.90; I2 = 0% and P-heterogeneity = 0.67) (see Supplemental Figure S3 under “Supplemental data” in the online issue). The summary RR per 200 g/d was 0.91 (95% CI: 0.86, 0.96; I2 = 40.2% and P-heterogeneity = 0.10) (Figure 3C). The summary RRs ranged from 0.90 (95% CI: 0.85, 0.94) when the EPIC-Potsdam Study was excluded (24) to 0.92 (95% CI: 0.87, 0.97) when the Women’s Health Initiative Observational Study (19) was excluded. There was no evidence of publication bias (P = 0.49). There was some indication of a nonlinear association between low-fat dairy products and type 2 diabetes (P-nonlinearity = 0.06), with no further reduction in risk above 300–400 g/d (seeFigure 3D and Supplemental Table S3 under “Supplemental data” in the online issue).

Milk

Seven cohort studies (12, 16–18, 20–22) were included in the analysis of high compared with low milk intake and type 2 diabetes risk and included a total of 15,149 cases among 167,982 participants. The summary RR was 0.87 (95% CI: 0.70, 1.07), and there was high heterogeneity (I2 = 70.5%; P-heterogeneity = 0.002) (see Supplemental Figure S4 under “Supplemental data” in the online issue). The summary RR per 200 g/d was 0.87 (95% CI: 0.72, 1.04), and there was very high heterogeneity (I2 = 93.6%; P-heterogeneity < 0.0001) (Figure 4A). In a sensitivity analysis, the summary RRs of type 2 diabetes ranged from 0.84 (95% CI: 0.66, 1.06) when the EPIC-Interact Study (21) was excluded to 0.99 (95% CI: 0.95, 1.04) when the Shanghai Women’s Health Study (12) was excluded. The heterogeneity was also reduced when the latter study was excluded (I2 = 0% and P-heterogeneity = 0.76). There was no indication of publication bias with Egger’s test (P = 0.41). There was evidence of a nonlinear inverse association between milk intake and type 2 diabetes (P-nonlinearity < 0.0001); however, the CIs were wide (seeFigure 4B and Supplemental Table S4 under “Supplemental data” in the online issue). When the Shanghai Women’s Health Study (which seemed to be an outlier) was excluded, the test for nonlinearity was no longer significant (P-nonlinearity = 0.62) and the association was null over the whole range of intake (results not shown).

FIGURE 4. View largeDownload slide A–D: Intake of milk and types of milk and risk of type 2 diabetes. The summary RR per 200 g/d by using random-effects models was 0.87 (95% CI: 0.72, 1.04; I2 = 93.6%; P-heterogeneity < 0.0001; n = 167,982) (A). There was evidence of a nonlinear inverse association between milk intake and type 2 diabetes (P-nonlinearity < 0.0001) (B) and evidence of a nonlinear positive association between whole-milk intake and type 2 diabetes (P-nonlinearity = 0.01) (C), whereas the association between low-fat or skim milk and type 2 diabetes appeared to be linear (P-nonlinearity = 0.44) (D).

FIGURE 4. View largeDownload slide A–D: Intake of milk and types of milk and risk of type 2 diabetes. The summary RR per 200 g/d by using random-effects models was 0.87 (95% CI: 0.72, 1.04; I2 = 93.6%; P-heterogeneity < 0.0001; n = 167,982) (A). There was evidence of a nonlinear inverse association between milk intake and type 2 diabetes (P-nonlinearity < 0.0001) (B) and evidence of a nonlinear positive association between whole-milk intake and type 2 diabetes (P-nonlinearity = 0.01) (C), whereas the association between low-fat or skim milk and type 2 diabetes appeared to be linear (P-nonlinearity = 0.44) (D).

Cheese

Eight cohort studies (7, 10, 14, 17, 18, 20–22) were included in the analysis of cheese intake and type 2 diabetes risk and included a total of 17,620 cases among 242,960 participants. The summary RR for high compared with low intake was 0.91 (95% CI: 0.84, 0.98), with no heterogeneity (I2 = 0%; P-heterogeneity = 0.57) (see Supplemental Figure S5 under “Supplemental data” in the online issue). The summary RR per 50 g/d was 0.92 (95% CI: 0.86, 0.99), with no heterogeneity (I2 = 0%; P-heterogeneity = 0.79) (Figure 5A). The summary RRs ranged from 0.91 (95% CI: 0.72, 1.14) when the EPIC-Interact Study (21) was excluded to 0.93 (95% CI: 0.86, 0.99) when the Australian Diabetes Obesity and Lifestyle Study (14) was excluded. There was no evidence of publication bias with Egger’s test (P = 0.74). There was some indication of a nonlinear association between cheese intake and type 2 diabetes (P-nonlinearity = 0.05), with a reduction in risk up to an intake of ~50 g/d (seeFigure 5B and Supplemental Table S5 under “Supplemental data” in the online issue); however, there were few data points above that value.

FIGURE 5. View largeDownload slide A–D: Intake of cheese and yogurt and risk of type 2 diabetes. The summary RRs by using random-effects models were 0.92 (95% CI: 0.86, 0.99; I2 = 0%; P-heterogeneity = 0.79; n = 242,960) per 50 g cheese/d (A, B) and 0.78 (95% CI: 0.60, 1.02; I2 = 69.9%; P-heterogeneity = 0.003; n = 254,892) per 200 g yogurt/d (C, D).

FIGURE 5. View largeDownload slide A–D: Intake of cheese and yogurt and risk of type 2 diabetes. The summary RRs by using random-effects models were 0.92 (95% CI: 0.86, 0.99; I2 = 0%; P-heterogeneity = 0.79; n = 242,960) per 50 g cheese/d (A, B) and 0.78 (95% CI: 0.60, 1.02; I2 = 69.9%; P-heterogeneity = 0.003; n = 254,892) per 200 g yogurt/d (C, D).

Yogurt

Seven cohort studies (7, 10, 14, 18–21) were included in the analysis of high compared with low yogurt intake and type 2 diabetes risk and included a total of 19,082 cases among 254,892 participants. The summary RR was 0.86 (95% CI: 0.75, 0.98), with moderate heterogeneity (I2 = 58.9%; P-heterogeneity = 0.02) (see Supplemental Figure S6 under “Supplemental data” in the online issue). The summary RR per 200 g/d was 0.78 (95% CI: 0.60, 1.02), with high heterogeneity (I2 = 69.9%; P-heterogeneity = 0.003) (Figure 5C). The summary RRs ranged from 0.73 (95% CI: 0.55, 0.96) when the Australian Diabetes Obesity and Lifestyle Study (14) was excluded to 0.86 (95% CI: 0.72, 1.02) when the Women’s Health Initiative Observational Study (19) was excluded. Most of the heterogeneity was explained by the latter study as well and when excluded (I2 = 34.3%; P-heterogeneity = 0.18). There was no evidence of publication bias with Egger’s test (P = 0.37). There was evidence of a nonlinear association between yogurt intake and type 2 diabetes (P-nonlinearity = 0.004), and there was no further reduction in risk with an intake >120–140 g/d (seeFigure 5D and Supplemental Table S6 under “Supplemental data” in the online issue).

Other dairy food items

Fewer studies reported results for other specific dairy food items. Inverse associations were observed for high compared with low intakes of low-fat or skim milk (summary RR = 0.82; 95% CI: 0.69, 0.97; I2 = 40.1%; P-heterogeneity = 0.19) (7, 10, 14), fermented dairy products (summary RR = 0.88; 95% CI: 0.79, 0.98; I2 = 0%; P-heterogeneity > 0.99) (21, 22), and ice cream (summary RR = 0.83; 95% CI: 0.73, 0.95; I2 = 0%; P-heterogeneity = 0.37) (7, 10), but there was no significant association with intake of whole milk (summary RR = 1.12; 95% CI: 0.99, 1.27; I2 = 0%; P-heterogeneity = 0.79) (7, 8, 10, 14), cottage cheese (summary RR = 0.91; 95% CI: 0.79, 1.04; I2 = 0%; P-heterogeneity = 0.43) (7, 10), cream (summary RR = 0.96; 95% CI: 0.84, 1.11; I2 = 4.9%; P-heterogeneity = 0.31) (7, 10), sour cream (summary RR = 0.98; 95% CI: 0.82, 1.16; I2 = 0%; P-heterogeneity = 0.54) (7, 10), or sherbet (summary RR = 0.90; 95% CI: 0.79, 1.03; I2 = 0%; P-heterogeneity = 0.60) (7, 10) (Table 2). The summary RRs were 1.06 (95% CI: 0.93, 1.20; I2 = 22.5%; P-heterogeneity = 0.28) per 200 g whole milk/d (see Supplemental Figure S7 under “Supplemental data” in the online issue) and 0.89 (95% CI: 0.84, 0.95; I2 = 0%; P-heterogeneity = 0.57) per 200 g low-fat or skim milk/d (see Supplemental Figure S8 under “Supplemental data” in the online issue). Because of the few studies we did not conduct a dose-response analysis for the remaining subtypes of dairy foods. There was a suggestion of a nonlinear positive association between whole milk (full-fat milk) intake and type 2 diabetes (P-nonlinearity = 0.01), with a rapid increase in risk when increasing intakes from low amounts to 40–50 g/d (seeFigure 4C and Supplemental Table S7 under “Supplemental data” in the online issue), whereas the association between low-fat or skim milk and type 2 diabetes appeared to be linear (P-nonlinearity = 0.44) (seeFigure 4D and Supplemental Table S8 under “Supplemental data” in the online issue).

TABLE 2 Type of dairy product n RR (95% CI) I2 P-heterogeneity Whole milk 4 1.12 (0.99, 1.27) 0 0.79 Low-fat/skim milk 3 0.82 (0.69, 0.97) 40.1 0.19 Fermented dairy products 2 0.88 (0.79, 0.98) 0 >0.99 Cottage cheese 2 0.91 (0.79, 1.04) 0 0.43 Cream 2 0.96 (0.84, 1.11) 0 0.31 Sour cream 2 0.98 (0.82, 1.16) 0 0.54 Ice cream 2 0.83 (0.73, 0.95) 0 0.37 Sherbet 2 0.90 (0.79, 1.03) 0 0.60 Type of dairy product n RR (95% CI) I2 P-heterogeneity Whole milk 4 1.12 (0.99, 1.27) 0 0.79 Low-fat/skim milk 3 0.82 (0.69, 0.97) 40.1 0.19 Fermented dairy products 2 0.88 (0.79, 0.98) 0 >0.99 Cottage cheese 2 0.91 (0.79, 1.04) 0 0.43 Cream 2 0.96 (0.84, 1.11) 0 0.31 Sour cream 2 0.98 (0.82, 1.16) 0 0.54 Ice cream 2 0.83 (0.73, 0.95) 0 0.37 Sherbet 2 0.90 (0.79, 1.03) 0 0.60 View Large

TABLE 2 Type of dairy product n RR (95% CI) I2 P-heterogeneity Whole milk 4 1.12 (0.99, 1.27) 0 0.79 Low-fat/skim milk 3 0.82 (0.69, 0.97) 40.1 0.19 Fermented dairy products 2 0.88 (0.79, 0.98) 0 >0.99 Cottage cheese 2 0.91 (0.79, 1.04) 0 0.43 Cream 2 0.96 (0.84, 1.11) 0 0.31 Sour cream 2 0.98 (0.82, 1.16) 0 0.54 Ice cream 2 0.83 (0.73, 0.95) 0 0.37 Sherbet 2 0.90 (0.79, 1.03) 0 0.60 Type of dairy product n RR (95% CI) I2 P-heterogeneity Whole milk 4 1.12 (0.99, 1.27) 0 0.79 Low-fat/skim milk 3 0.82 (0.69, 0.97) 40.1 0.19 Fermented dairy products 2 0.88 (0.79, 0.98) 0 >0.99 Cottage cheese 2 0.91 (0.79, 1.04) 0 0.43 Cream 2 0.96 (0.84, 1.11) 0 0.31 Sour cream 2 0.98 (0.82, 1.16) 0 0.54 Ice cream 2 0.83 (0.73, 0.95) 0 0.37 Sherbet 2 0.90 (0.79, 1.03) 0 0.60 View Large

Subgroup, sensitivity, and meta-regression analyses

In subgroup and meta-regression analyses, we found no significant heterogeneity between subgroups when studies were stratified by sex, duration of follow-up, geographic location, or study size (Tables 3 and 4). There was some indication that the inverse association between total dairy products and yogurt and type 2 diabetes was restricted to the American but not the European studies; however, the tests for heterogeneity were not significant (P = 0.10 and 0.75, respectively). Further subgroup analyses by whether studies had adjusted for confounding factors did not show significant heterogeneity between strata, although associations were not always significant. In a sensitivity analysis, we reconducted the analysis of total dairy, high-fat dairy products, and low-fat dairy products restricted to the 6 studies that were common for the 3 analyses (7, 10, 13, 15, 20, 22), and the summary RRs were 0.90 (95% CI: 0.80, 1.01; I2 = 49.7%; P-heterogeneity = 0.08) for total dairy products, 0.99 (95% CI: 0.93, 1.05; I2 = 30.9%; P-heterogeneity = 0.20) for high-fat dairy products, and 0.91 (95% CI: 0.86, 0.96; I2 = 24.2%; P-heterogeneity = 0.25) for low-fat dairy products, which was similar to the overall analysis.

TABLE 3 Dairy products, per 400 g/d High-fat dairy products, per 200 g/d Low-fat dairy products, per 200 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 12 0.93 (0.87, 0.99) 33.1 0.13 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Duration of follow-up <10 y 6 0.92 (0.84, 1.00) 15.3 0.32 0.63 4 0.96 (0.86, 1.07) 26.1 0.26 0.64 5 0.89 (0.80, 0.99) 58.7 0.05 0.65 ≥10 y 6 0.94 (0.85, 1.05) 50.7 0.07 5 0.99 (0.94, 1.04) 7.1 0.37 4 0.92 (0.87, 0.97) 12.6 0.33 Sex Men 4 0.85 (0.62, 1.18) 62.7 0.05 0.08/0.904 1 0.99 (0.90, 1.08) 0.87/0.854 1 0.87 (0.79, 0.93) 0.08/0.964 Women 6 0.90 (0.82, 0.98) 39.0 0.15 3 0.97 (0.88, 1.07) 55.1 0.11 4 0.87 (0.79, 0.96) 53.7 0.09 Men and women 4 1.01 (0.93, 1.09) 0 0.46 4 0.99 (0.84, 1.16) 25.4 0.26 4 0.98 (0.91, 1.06) 0 0.67 Geographic location Europe 4 1.01 (0.92, 1.11) 4.1 0.37 0.10 4 1.00 (0.88, 1.12) 22.2 0.28 0.83 3 0.99 (0.91, 1.07) 0 0.55 0.18 America 5 0.90 (0.83, 0.97) 41.4 0.15 4 0.98 (0.93, 1.04) 32.2 0.22 5 0.87 (0.82, 0.94) 44.2 0.13 Asia 1 0.96 (0.77, 1.19) 0 0 Australia 2 0.77 (0.53, 1.11) 0 0.33 1 0.89 (0.64, 1.24) 1 0.91 (0.71, 1.17) Number of cases <1000 6 0.86 (0.66, 1.12) 47.1 0.09 0.20 6 0.95 (0.85, 1.06) 29.6 0.21 0.28 5 0.95 (0.87, 1.04) 28.3 0.23 0.33 1000 to <1500 2 0.86 (0.73, 1.02) 38.9 0.20 1 0.99 (0.90, 1.08) 1 0.87 (0.79, 0.93) ≥1500 4 0.96 (0.91, 1.00) 0 0.60 2 1.00 (0.96, 1.05) 0 0.88 3 0.89 (0.81, 0.97) 53.3 0.12 Adjustment for confounding factors Alcohol Yes 9 0.93 (0.88, 1.00) 41.5 0.09 0.30 7 0.99 (0.94, 1.04) 17.2 0.30 0.29 8 0.91 (0.86, 0.96) 47.7 0.06 0.98 No 3 0.75 (0.53, 1.06) 0 0.57 2 0.92 (0.81, 1.04) 0 0.86 1 0.91 (0.71, 1.17) Smoking Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 BMI, waist circumference Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 Physical activity Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 8 0.99 (0.95, 1.04) 8.4 0.37 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 1 0.92 (0.80, 1.06) 0 Coffee, caffeine Yes 6 0.96 (0.86, 1.07) 40.9 0.13 0.31 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 6 0.91 (0.85, 0.97) 15.3 0.32 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Soft drinks Yes 3 0.91 (0.78, 1.07) 64.7 0.06 0.66 3 0.95 (0.84, 1.08) 48.8 0.14 0.58 3 0.90 (0.76, 1.06) 71.0 0.03 0.90 No 9 0.92 (0.86, 0.99) 18.5 0.28 6 0.99 (0.95, 1.04) 0 0.50 6 0.91 (0.86, 0.96) 21.8 0.27 Meat Yes 5 0.95 (0.83, 1.09) 52.6 0.08 0.33 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 7 0.92 (0.87, 0.97) 1.2 0.42 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Fat Yes 3 0.84 (0.71, 0.99) 63.0 0.07 0.14 3 0.97 (0.89, 1.05) 53.5 0.12 0.98 3 0.89 (0.81, 0.97) 58.5 0.09 0.53 No 9 0.97 (0.92, 1.02) 0 0.48 6 0.99 (0.92, 1.07) 0 0.50 6 0.92 (0.86, 1.00) 39.5 0.14 Grains, whole grains Yes 4 0.99 (0.90, 1.08) 17.6 0.30 0.15 4 1.03 (0.94, 1.13) 0 0.50 0.38 4 0.95 (0.88, 1.03) 27.8 0.25 0.19 No 8 0.90 (0.83, 0.97) 25.4 0.23 5 0.97 (0.91, 1.02) 26.6 0.24 5 0.88 (0.81, 0.95) 43.0 0.14 Fiber Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.90, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Glycemic load Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.89, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Magnesium Yes 1 0.96 (0.77, 1.19) 0.83 0 NC 0 NC No 11 0.92 (0.86, 0.99) 39.1 0.09 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Energy intake Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 0 0 Dairy products, per 400 g/d High-fat dairy products, per 200 g/d Low-fat dairy products, per 200 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 12 0.93 (0.87, 0.99) 33.1 0.13 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Duration of follow-up <10 y 6 0.92 (0.84, 1.00) 15.3 0.32 0.63 4 0.96 (0.86, 1.07) 26.1 0.26 0.64 5 0.89 (0.80, 0.99) 58.7 0.05 0.65 ≥10 y 6 0.94 (0.85, 1.05) 50.7 0.07 5 0.99 (0.94, 1.04) 7.1 0.37 4 0.92 (0.87, 0.97) 12.6 0.33 Sex Men 4 0.85 (0.62, 1.18) 62.7 0.05 0.08/0.904 1 0.99 (0.90, 1.08) 0.87/0.854 1 0.87 (0.79, 0.93) 0.08/0.964 Women 6 0.90 (0.82, 0.98) 39.0 0.15 3 0.97 (0.88, 1.07) 55.1 0.11 4 0.87 (0.79, 0.96) 53.7 0.09 Men and women 4 1.01 (0.93, 1.09) 0 0.46 4 0.99 (0.84, 1.16) 25.4 0.26 4 0.98 (0.91, 1.06) 0 0.67 Geographic location Europe 4 1.01 (0.92, 1.11) 4.1 0.37 0.10 4 1.00 (0.88, 1.12) 22.2 0.28 0.83 3 0.99 (0.91, 1.07) 0 0.55 0.18 America 5 0.90 (0.83, 0.97) 41.4 0.15 4 0.98 (0.93, 1.04) 32.2 0.22 5 0.87 (0.82, 0.94) 44.2 0.13 Asia 1 0.96 (0.77, 1.19) 0 0 Australia 2 0.77 (0.53, 1.11) 0 0.33 1 0.89 (0.64, 1.24) 1 0.91 (0.71, 1.17) Number of cases <1000 6 0.86 (0.66, 1.12) 47.1 0.09 0.20 6 0.95 (0.85, 1.06) 29.6 0.21 0.28 5 0.95 (0.87, 1.04) 28.3 0.23 0.33 1000 to <1500 2 0.86 (0.73, 1.02) 38.9 0.20 1 0.99 (0.90, 1.08) 1 0.87 (0.79, 0.93) ≥1500 4 0.96 (0.91, 1.00) 0 0.60 2 1.00 (0.96, 1.05) 0 0.88 3 0.89 (0.81, 0.97) 53.3 0.12 Adjustment for confounding factors Alcohol Yes 9 0.93 (0.88, 1.00) 41.5 0.09 0.30 7 0.99 (0.94, 1.04) 17.2 0.30 0.29 8 0.91 (0.86, 0.96) 47.7 0.06 0.98 No 3 0.75 (0.53, 1.06) 0 0.57 2 0.92 (0.81, 1.04) 0 0.86 1 0.91 (0.71, 1.17) Smoking Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 BMI, waist circumference Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 Physical activity Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 8 0.99 (0.95, 1.04) 8.4 0.37 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 1 0.92 (0.80, 1.06) 0 Coffee, caffeine Yes 6 0.96 (0.86, 1.07) 40.9 0.13 0.31 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 6 0.91 (0.85, 0.97) 15.3 0.32 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Soft drinks Yes 3 0.91 (0.78, 1.07) 64.7 0.06 0.66 3 0.95 (0.84, 1.08) 48.8 0.14 0.58 3 0.90 (0.76, 1.06) 71.0 0.03 0.90 No 9 0.92 (0.86, 0.99) 18.5 0.28 6 0.99 (0.95, 1.04) 0 0.50 6 0.91 (0.86, 0.96) 21.8 0.27 Meat Yes 5 0.95 (0.83, 1.09) 52.6 0.08 0.33 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 7 0.92 (0.87, 0.97) 1.2 0.42 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Fat Yes 3 0.84 (0.71, 0.99) 63.0 0.07 0.14 3 0.97 (0.89, 1.05) 53.5 0.12 0.98 3 0.89 (0.81, 0.97) 58.5 0.09 0.53 No 9 0.97 (0.92, 1.02) 0 0.48 6 0.99 (0.92, 1.07) 0 0.50 6 0.92 (0.86, 1.00) 39.5 0.14 Grains, whole grains Yes 4 0.99 (0.90, 1.08) 17.6 0.30 0.15 4 1.03 (0.94, 1.13) 0 0.50 0.38 4 0.95 (0.88, 1.03) 27.8 0.25 0.19 No 8 0.90 (0.83, 0.97) 25.4 0.23 5 0.97 (0.91, 1.02) 26.6 0.24 5 0.88 (0.81, 0.95) 43.0 0.14 Fiber Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.90, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Glycemic load Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.89, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Magnesium Yes 1 0.96 (0.77, 1.19) 0.83 0 NC 0 NC No 11 0.92 (0.86, 0.99) 39.1 0.09 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Energy intake Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 0 0 View Large

TABLE 3 Dairy products, per 400 g/d High-fat dairy products, per 200 g/d Low-fat dairy products, per 200 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 12 0.93 (0.87, 0.99) 33.1 0.13 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Duration of follow-up <10 y 6 0.92 (0.84, 1.00) 15.3 0.32 0.63 4 0.96 (0.86, 1.07) 26.1 0.26 0.64 5 0.89 (0.80, 0.99) 58.7 0.05 0.65 ≥10 y 6 0.94 (0.85, 1.05) 50.7 0.07 5 0.99 (0.94, 1.04) 7.1 0.37 4 0.92 (0.87, 0.97) 12.6 0.33 Sex Men 4 0.85 (0.62, 1.18) 62.7 0.05 0.08/0.904 1 0.99 (0.90, 1.08) 0.87/0.854 1 0.87 (0.79, 0.93) 0.08/0.964 Women 6 0.90 (0.82, 0.98) 39.0 0.15 3 0.97 (0.88, 1.07) 55.1 0.11 4 0.87 (0.79, 0.96) 53.7 0.09 Men and women 4 1.01 (0.93, 1.09) 0 0.46 4 0.99 (0.84, 1.16) 25.4 0.26 4 0.98 (0.91, 1.06) 0 0.67 Geographic location Europe 4 1.01 (0.92, 1.11) 4.1 0.37 0.10 4 1.00 (0.88, 1.12) 22.2 0.28 0.83 3 0.99 (0.91, 1.07) 0 0.55 0.18 America 5 0.90 (0.83, 0.97) 41.4 0.15 4 0.98 (0.93, 1.04) 32.2 0.22 5 0.87 (0.82, 0.94) 44.2 0.13 Asia 1 0.96 (0.77, 1.19) 0 0 Australia 2 0.77 (0.53, 1.11) 0 0.33 1 0.89 (0.64, 1.24) 1 0.91 (0.71, 1.17) Number of cases <1000 6 0.86 (0.66, 1.12) 47.1 0.09 0.20 6 0.95 (0.85, 1.06) 29.6 0.21 0.28 5 0.95 (0.87, 1.04) 28.3 0.23 0.33 1000 to <1500 2 0.86 (0.73, 1.02) 38.9 0.20 1 0.99 (0.90, 1.08) 1 0.87 (0.79, 0.93) ≥1500 4 0.96 (0.91, 1.00) 0 0.60 2 1.00 (0.96, 1.05) 0 0.88 3 0.89 (0.81, 0.97) 53.3 0.12 Adjustment for confounding factors Alcohol Yes 9 0.93 (0.88, 1.00) 41.5 0.09 0.30 7 0.99 (0.94, 1.04) 17.2 0.30 0.29 8 0.91 (0.86, 0.96) 47.7 0.06 0.98 No 3 0.75 (0.53, 1.06) 0 0.57 2 0.92 (0.81, 1.04) 0 0.86 1 0.91 (0.71, 1.17) Smoking Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 BMI, waist circumference Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 Physical activity Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 8 0.99 (0.95, 1.04) 8.4 0.37 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 1 0.92 (0.80, 1.06) 0 Coffee, caffeine Yes 6 0.96 (0.86, 1.07) 40.9 0.13 0.31 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 6 0.91 (0.85, 0.97) 15.3 0.32 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Soft drinks Yes 3 0.91 (0.78, 1.07) 64.7 0.06 0.66 3 0.95 (0.84, 1.08) 48.8 0.14 0.58 3 0.90 (0.76, 1.06) 71.0 0.03 0.90 No 9 0.92 (0.86, 0.99) 18.5 0.28 6 0.99 (0.95, 1.04) 0 0.50 6 0.91 (0.86, 0.96) 21.8 0.27 Meat Yes 5 0.95 (0.83, 1.09) 52.6 0.08 0.33 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 7 0.92 (0.87, 0.97) 1.2 0.42 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Fat Yes 3 0.84 (0.71, 0.99) 63.0 0.07 0.14 3 0.97 (0.89, 1.05) 53.5 0.12 0.98 3 0.89 (0.81, 0.97) 58.5 0.09 0.53 No 9 0.97 (0.92, 1.02) 0 0.48 6 0.99 (0.92, 1.07) 0 0.50 6 0.92 (0.86, 1.00) 39.5 0.14 Grains, whole grains Yes 4 0.99 (0.90, 1.08) 17.6 0.30 0.15 4 1.03 (0.94, 1.13) 0 0.50 0.38 4 0.95 (0.88, 1.03) 27.8 0.25 0.19 No 8 0.90 (0.83, 0.97) 25.4 0.23 5 0.97 (0.91, 1.02) 26.6 0.24 5 0.88 (0.81, 0.95) 43.0 0.14 Fiber Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.90, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Glycemic load Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.89, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Magnesium Yes 1 0.96 (0.77, 1.19) 0.83 0 NC 0 NC No 11 0.92 (0.86, 0.99) 39.1 0.09 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Energy intake Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 0 0 Dairy products, per 400 g/d High-fat dairy products, per 200 g/d Low-fat dairy products, per 200 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 12 0.93 (0.87, 0.99) 33.1 0.13 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Duration of follow-up <10 y 6 0.92 (0.84, 1.00) 15.3 0.32 0.63 4 0.96 (0.86, 1.07) 26.1 0.26 0.64 5 0.89 (0.80, 0.99) 58.7 0.05 0.65 ≥10 y 6 0.94 (0.85, 1.05) 50.7 0.07 5 0.99 (0.94, 1.04) 7.1 0.37 4 0.92 (0.87, 0.97) 12.6 0.33 Sex Men 4 0.85 (0.62, 1.18) 62.7 0.05 0.08/0.904 1 0.99 (0.90, 1.08) 0.87/0.854 1 0.87 (0.79, 0.93) 0.08/0.964 Women 6 0.90 (0.82, 0.98) 39.0 0.15 3 0.97 (0.88, 1.07) 55.1 0.11 4 0.87 (0.79, 0.96) 53.7 0.09 Men and women 4 1.01 (0.93, 1.09) 0 0.46 4 0.99 (0.84, 1.16) 25.4 0.26 4 0.98 (0.91, 1.06) 0 0.67 Geographic location Europe 4 1.01 (0.92, 1.11) 4.1 0.37 0.10 4 1.00 (0.88, 1.12) 22.2 0.28 0.83 3 0.99 (0.91, 1.07) 0 0.55 0.18 America 5 0.90 (0.83, 0.97) 41.4 0.15 4 0.98 (0.93, 1.04) 32.2 0.22 5 0.87 (0.82, 0.94) 44.2 0.13 Asia 1 0.96 (0.77, 1.19) 0 0 Australia 2 0.77 (0.53, 1.11) 0 0.33 1 0.89 (0.64, 1.24) 1 0.91 (0.71, 1.17) Number of cases <1000 6 0.86 (0.66, 1.12) 47.1 0.09 0.20 6 0.95 (0.85, 1.06) 29.6 0.21 0.28 5 0.95 (0.87, 1.04) 28.3 0.23 0.33 1000 to <1500 2 0.86 (0.73, 1.02) 38.9 0.20 1 0.99 (0.90, 1.08) 1 0.87 (0.79, 0.93) ≥1500 4 0.96 (0.91, 1.00) 0 0.60 2 1.00 (0.96, 1.05) 0 0.88 3 0.89 (0.81, 0.97) 53.3 0.12 Adjustment for confounding factors Alcohol Yes 9 0.93 (0.88, 1.00) 41.5 0.09 0.30 7 0.99 (0.94, 1.04) 17.2 0.30 0.29 8 0.91 (0.86, 0.96) 47.7 0.06 0.98 No 3 0.75 (0.53, 1.06) 0 0.57 2 0.92 (0.81, 1.04) 0 0.86 1 0.91 (0.71, 1.17) Smoking Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 BMI, waist circumference Yes 12 0.93 (0.87, 0.99) 33.1 0.13 NC 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 0 0 0 Physical activity Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 8 0.99 (0.95, 1.04) 8.4 0.37 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 1 0.92 (0.80, 1.06) 0 Coffee, caffeine Yes 6 0.96 (0.86, 1.07) 40.9 0.13 0.31 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 6 0.91 (0.85, 0.97) 15.3 0.32 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Soft drinks Yes 3 0.91 (0.78, 1.07) 64.7 0.06 0.66 3 0.95 (0.84, 1.08) 48.8 0.14 0.58 3 0.90 (0.76, 1.06) 71.0 0.03 0.90 No 9 0.92 (0.86, 0.99) 18.5 0.28 6 0.99 (0.95, 1.04) 0 0.50 6 0.91 (0.86, 0.96) 21.8 0.27 Meat Yes 5 0.95 (0.83, 1.09) 52.6 0.08 0.33 4 0.99 (0.83, 1.19) 57.5 0.07 0.93 5 0.93 (0.85, 1.02) 46.0 0.12 0.42 No 7 0.92 (0.87, 0.97) 1.2 0.42 5 0.99 (0.87, 1.12) 43.8 0.13 4 0.89 (0.83, 0.96) 40.1 0.17 Fat Yes 3 0.84 (0.71, 0.99) 63.0 0.07 0.14 3 0.97 (0.89, 1.05) 53.5 0.12 0.98 3 0.89 (0.81, 0.97) 58.5 0.09 0.53 No 9 0.97 (0.92, 1.02) 0 0.48 6 0.99 (0.92, 1.07) 0 0.50 6 0.92 (0.86, 1.00) 39.5 0.14 Grains, whole grains Yes 4 0.99 (0.90, 1.08) 17.6 0.30 0.15 4 1.03 (0.94, 1.13) 0 0.50 0.38 4 0.95 (0.88, 1.03) 27.8 0.25 0.19 No 8 0.90 (0.83, 0.97) 25.4 0.23 5 0.97 (0.91, 1.02) 26.6 0.24 5 0.88 (0.81, 0.95) 43.0 0.14 Fiber Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.90, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Glycemic load Yes 4 0.85 (0.74, 0.97) 44.5 0.14 0.13 4 0.97 (0.89, 1.04) 35.5 0.20 0.87 4 0.89 (0.83, 0.97) 37.8 0.19 0.55 No 8 0.97 (0.91, 1.02) 5.6 0.39 5 0.99 (0.92, 1.07) 0 0.41 5 0.92 (0.85, 1.01) 51.5 0.08 Magnesium Yes 1 0.96 (0.77, 1.19) 0.83 0 NC 0 NC No 11 0.92 (0.86, 0.99) 39.1 0.09 9 0.98 (0.94, 1.03) 7.6 0.37 9 0.91 (0.86, 0.96) 40.2 0.10 Energy intake Yes 11 0.93 (0.87, 0.99) 36.9 0.10 0.52 9 0.98 (0.94, 1.03) 7.6 0.37 NC 9 0.91 (0.86, 0.96) 40.2 0.10 NC No 1 0.58 (0.17, 1.97) 0 0 View Large

TABLE 4 Milk, per 200 g/d Yogurt, per 200 g/d Cheese, per 50 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 7 0.87 (0.72, 1.04) 93.6 <0.0001 7 0.78 (0.60, 1.02) 69.9 0.003 8 0.92 (0.86, 0.99) 0 0.79 Duration of follow-up <10 y 3 0.77 (0.52, 1.16) 96.5 <0.0001 0.31 3 0.65 (0.26, 1.62) 85.3 0.001 0.66 3 0.86 (0.60, 1.23) 0 0.56 0.71 ≥10 y 4 1.01 (0.96, 1.05) 0 0.71 4 0.83 (0.68, 1.02) 43.6 0.15 5 0.92 (0.86, 0.99) 0 0.62 Sex Men 2 0.98 (0.84, 1.15) 0 0.33 0.46/0.474 3 0.86 (0.59, 1.27) 11.6 0.32 0.87/0.534 3 0.74 (0.41, 1.31) 0 0.57 0.47/0.624 Women 2 0.66 (0.39, 1.14) 95.3 <0.0001 4 0.67 (0.38, 1.16) 76.9 0.005 3 0.89 (0.61, 1.29) 0 0.61 Men and women 4 1.00 (0.96, 1.05) 0 0.80 2 0.92 (0.83, 1.02) 0 0.40 4 0.93 (0.86, 1.00) 0 0.49 Geographic location Europe 4 1.00 (0.95, 1.06) 0 0.58 0.22 2 0.92 (0.83, 1.02) 0 0.40 0.75 3 0.93 (0.86, 1.00) 0 0.40 0.33 America 1 0.99 (0.92, 1.07) 3 0.52 (0.30, 0.88) 72.5 0.03 3 0.84 (0.60, 1.18) 0 0.79 Asia 2 0.69 (0.37, 1.28) 97.8 <0.0001 1 0.80 (0.51, 1.24) 1 1.23 (0.28, 5.42) Australia 0 1 1.42 (0.75, 2.68) 1 0.60 (0.28, 1.31) Number of cases <1000 4 0.98 (0.92, 1.04) 0 0.75 0.32 2 1.22 (0.83, 1.81) 0 0.58 0.16 3 0.96 (0.64, 1.44) 31.9 0.23 0.83 1000 to <1500 1 0.95 (0.84, 1.08) 2 0.74 (0.54, 1.01) 0 0.60 2 0.86 (0.44, 1.68) 0 0.60 ≥1500 2 0.72 (0.36, 1.43) 98.9 <0.0001 3 0.60 (0.36, 1.00) 87.0 <0.0001 3 0.92 (0.86, 0.99) 0 0.74 Adjustment for confounding factors Alcohol Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 6 0.73 (0.55, 0.96) 71.7 0.003 0.30 6 0.93 (0.86, 1.00) 0 0.82 0.24 No 2 0.99 (0.92, 1.07) 0 0.33 1 1.42 (0.75, 2.68) 2 0.59 (0.31, 1.16) 0 0.94 Smoking Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) BMI, waist circumference Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) Physical activity Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Coffee, caffeine Yes 4 1.00 (0.95, 1.05) 0 0.65 0.17 3 0.91 (0.82, 1.01) 0 0.59 0.39 4 0.93 (0.86, 1.00) 0 0.58 0.37 No 3 0.71 (0.41, 1.24) 97.4 <0.0001 4 0.64 (0.37, 1.10) 77.4 0.004 4 0.79 (0.58, 1.09) 0 0.79 Soft drinks Yes 1 1.02 (0.96, 1.09) 0.54 1 0.91 (0.82, 1.01) 0.72 1 0.92 (0.86, 1.00) 0.91 No 6 0.84 (0.66, 1.06) 93.8 <0.0001 6 0.74 (0.51, 1.06) 70.3 0.005 7 0.91 (0.72, 1.14) 0 0.69 Meat Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fat Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Grains, whole grains Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fiber Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Glycemic load Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Magnesium Yes 1 0.95 (0.84, 1.08) 0.74 1 0.80 (0.51, 1.24) 0.93 1 1.23 (0.28, 5.42) 0.72 No 6 0.85 (0.69, 1.06) 94.6 <0.0001 6 0.77 (0.57, 1.05) 74.8 0.001 7 0.92 (0.86, 0.99) 0 0.71 Energy intake Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Milk, per 200 g/d Yogurt, per 200 g/d Cheese, per 50 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 7 0.87 (0.72, 1.04) 93.6 <0.0001 7 0.78 (0.60, 1.02) 69.9 0.003 8 0.92 (0.86, 0.99) 0 0.79 Duration of follow-up <10 y 3 0.77 (0.52, 1.16) 96.5 <0.0001 0.31 3 0.65 (0.26, 1.62) 85.3 0.001 0.66 3 0.86 (0.60, 1.23) 0 0.56 0.71 ≥10 y 4 1.01 (0.96, 1.05) 0 0.71 4 0.83 (0.68, 1.02) 43.6 0.15 5 0.92 (0.86, 0.99) 0 0.62 Sex Men 2 0.98 (0.84, 1.15) 0 0.33 0.46/0.474 3 0.86 (0.59, 1.27) 11.6 0.32 0.87/0.534 3 0.74 (0.41, 1.31) 0 0.57 0.47/0.624 Women 2 0.66 (0.39, 1.14) 95.3 <0.0001 4 0.67 (0.38, 1.16) 76.9 0.005 3 0.89 (0.61, 1.29) 0 0.61 Men and women 4 1.00 (0.96, 1.05) 0 0.80 2 0.92 (0.83, 1.02) 0 0.40 4 0.93 (0.86, 1.00) 0 0.49 Geographic location Europe 4 1.00 (0.95, 1.06) 0 0.58 0.22 2 0.92 (0.83, 1.02) 0 0.40 0.75 3 0.93 (0.86, 1.00) 0 0.40 0.33 America 1 0.99 (0.92, 1.07) 3 0.52 (0.30, 0.88) 72.5 0.03 3 0.84 (0.60, 1.18) 0 0.79 Asia 2 0.69 (0.37, 1.28) 97.8 <0.0001 1 0.80 (0.51, 1.24) 1 1.23 (0.28, 5.42) Australia 0 1 1.42 (0.75, 2.68) 1 0.60 (0.28, 1.31) Number of cases <1000 4 0.98 (0.92, 1.04) 0 0.75 0.32 2 1.22 (0.83, 1.81) 0 0.58 0.16 3 0.96 (0.64, 1.44) 31.9 0.23 0.83 1000 to <1500 1 0.95 (0.84, 1.08) 2 0.74 (0.54, 1.01) 0 0.60 2 0.86 (0.44, 1.68) 0 0.60 ≥1500 2 0.72 (0.36, 1.43) 98.9 <0.0001 3 0.60 (0.36, 1.00) 87.0 <0.0001 3 0.92 (0.86, 0.99) 0 0.74 Adjustment for confounding factors Alcohol Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 6 0.73 (0.55, 0.96) 71.7 0.003 0.30 6 0.93 (0.86, 1.00) 0 0.82 0.24 No 2 0.99 (0.92, 1.07) 0 0.33 1 1.42 (0.75, 2.68) 2 0.59 (0.31, 1.16) 0 0.94 Smoking Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) BMI, waist circumference Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) Physical activity Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Coffee, caffeine Yes 4 1.00 (0.95, 1.05) 0 0.65 0.17 3 0.91 (0.82, 1.01) 0 0.59 0.39 4 0.93 (0.86, 1.00) 0 0.58 0.37 No 3 0.71 (0.41, 1.24) 97.4 <0.0001 4 0.64 (0.37, 1.10) 77.4 0.004 4 0.79 (0.58, 1.09) 0 0.79 Soft drinks Yes 1 1.02 (0.96, 1.09) 0.54 1 0.91 (0.82, 1.01) 0.72 1 0.92 (0.86, 1.00) 0.91 No 6 0.84 (0.66, 1.06) 93.8 <0.0001 6 0.74 (0.51, 1.06) 70.3 0.005 7 0.91 (0.72, 1.14) 0 0.69 Meat Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fat Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Grains, whole grains Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fiber Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Glycemic load Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Magnesium Yes 1 0.95 (0.84, 1.08) 0.74 1 0.80 (0.51, 1.24) 0.93 1 1.23 (0.28, 5.42) 0.72 No 6 0.85 (0.69, 1.06) 94.6 <0.0001 6 0.77 (0.57, 1.05) 74.8 0.001 7 0.92 (0.86, 0.99) 0 0.71 Energy intake Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) View Large

TABLE 4 Milk, per 200 g/d Yogurt, per 200 g/d Cheese, per 50 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 7 0.87 (0.72, 1.04) 93.6 <0.0001 7 0.78 (0.60, 1.02) 69.9 0.003 8 0.92 (0.86, 0.99) 0 0.79 Duration of follow-up <10 y 3 0.77 (0.52, 1.16) 96.5 <0.0001 0.31 3 0.65 (0.26, 1.62) 85.3 0.001 0.66 3 0.86 (0.60, 1.23) 0 0.56 0.71 ≥10 y 4 1.01 (0.96, 1.05) 0 0.71 4 0.83 (0.68, 1.02) 43.6 0.15 5 0.92 (0.86, 0.99) 0 0.62 Sex Men 2 0.98 (0.84, 1.15) 0 0.33 0.46/0.474 3 0.86 (0.59, 1.27) 11.6 0.32 0.87/0.534 3 0.74 (0.41, 1.31) 0 0.57 0.47/0.624 Women 2 0.66 (0.39, 1.14) 95.3 <0.0001 4 0.67 (0.38, 1.16) 76.9 0.005 3 0.89 (0.61, 1.29) 0 0.61 Men and women 4 1.00 (0.96, 1.05) 0 0.80 2 0.92 (0.83, 1.02) 0 0.40 4 0.93 (0.86, 1.00) 0 0.49 Geographic location Europe 4 1.00 (0.95, 1.06) 0 0.58 0.22 2 0.92 (0.83, 1.02) 0 0.40 0.75 3 0.93 (0.86, 1.00) 0 0.40 0.33 America 1 0.99 (0.92, 1.07) 3 0.52 (0.30, 0.88) 72.5 0.03 3 0.84 (0.60, 1.18) 0 0.79 Asia 2 0.69 (0.37, 1.28) 97.8 <0.0001 1 0.80 (0.51, 1.24) 1 1.23 (0.28, 5.42) Australia 0 1 1.42 (0.75, 2.68) 1 0.60 (0.28, 1.31) Number of cases <1000 4 0.98 (0.92, 1.04) 0 0.75 0.32 2 1.22 (0.83, 1.81) 0 0.58 0.16 3 0.96 (0.64, 1.44) 31.9 0.23 0.83 1000 to <1500 1 0.95 (0.84, 1.08) 2 0.74 (0.54, 1.01) 0 0.60 2 0.86 (0.44, 1.68) 0 0.60 ≥1500 2 0.72 (0.36, 1.43) 98.9 <0.0001 3 0.60 (0.36, 1.00) 87.0 <0.0001 3 0.92 (0.86, 0.99) 0 0.74 Adjustment for confounding factors Alcohol Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 6 0.73 (0.55, 0.96) 71.7 0.003 0.30 6 0.93 (0.86, 1.00) 0 0.82 0.24 No 2 0.99 (0.92, 1.07) 0 0.33 1 1.42 (0.75, 2.68) 2 0.59 (0.31, 1.16) 0 0.94 Smoking Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) BMI, waist circumference Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) Physical activity Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Coffee, caffeine Yes 4 1.00 (0.95, 1.05) 0 0.65 0.17 3 0.91 (0.82, 1.01) 0 0.59 0.39 4 0.93 (0.86, 1.00) 0 0.58 0.37 No 3 0.71 (0.41, 1.24) 97.4 <0.0001 4 0.64 (0.37, 1.10) 77.4 0.004 4 0.79 (0.58, 1.09) 0 0.79 Soft drinks Yes 1 1.02 (0.96, 1.09) 0.54 1 0.91 (0.82, 1.01) 0.72 1 0.92 (0.86, 1.00) 0.91 No 6 0.84 (0.66, 1.06) 93.8 <0.0001 6 0.74 (0.51, 1.06) 70.3 0.005 7 0.91 (0.72, 1.14) 0 0.69 Meat Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fat Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Grains, whole grains Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77 (0.54, 1.11) 96.2 <0.0001 5 0.68 (0.45, 1.02) 70.9 0.008 5 0.81 (0.60, 1.10) 0 0.85 Fiber Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Glycemic load Yes 0 NC 2 0.67 (0.52, 0.87) 0 1.00 0.71 2 0.86 (0.61, 1.23) 0 0.78 0.73 No 7 0.87 (0.72, 1.04) 93.6 <0.0001 5 0.82 (0.56, 1.20) 74.8 0.003 6 0.92 (0.86, 0.99) 0 0.59 Magnesium Yes 1 0.95 (0.84, 1.08) 0.74 1 0.80 (0.51, 1.24) 0.93 1 1.23 (0.28, 5.42) 0.72 No 6 0.85 (0.69, 1.06) 94.6 <0.0001 6 0.77 (0.57, 1.05) 74.8 0.001 7 0.92 (0.86, 0.99) 0 0.71 Energy intake Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Milk, per 200 g/d Yogurt, per 200 g/d Cheese, per 50 g/d n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 n RR (95% CI) I2 P2 P3 % % % All studies 7 0.87 (0.72, 1.04) 93.6 <0.0001 7 0.78 (0.60, 1.02) 69.9 0.003 8 0.92 (0.86, 0.99) 0 0.79 Duration of follow-up <10 y 3 0.77 (0.52, 1.16) 96.5 <0.0001 0.31 3 0.65 (0.26, 1.62) 85.3 0.001 0.66 3 0.86 (0.60, 1.23) 0 0.56 0.71 ≥10 y 4 1.01 (0.96, 1.05) 0 0.71 4 0.83 (0.68, 1.02) 43.6 0.15 5 0.92 (0.86, 0.99) 0 0.62 Sex Men 2 0.98 (0.84, 1.15) 0 0.33 0.46/0.474 3 0.86 (0.59, 1.27) 11.6 0.32 0.87/0.534 3 0.74 (0.41, 1.31) 0 0.57 0.47/0.624 Women 2 0.66 (0.39, 1.14) 95.3 <0.0001 4 0.67 (0.38, 1.16) 76.9 0.005 3 0.89 (0.61, 1.29) 0 0.61 Men and women 4 1.00 (0.96, 1.05) 0 0.80 2 0.92 (0.83, 1.02) 0 0.40 4 0.93 (0.86, 1.00) 0 0.49 Geographic location Europe 4 1.00 (0.95, 1.06) 0 0.58 0.22 2 0.92 (0.83, 1.02) 0 0.40 0.75 3 0.93 (0.86, 1.00) 0 0.40 0.33 America 1 0.99 (0.92, 1.07) 3 0.52 (0.30, 0.88) 72.5 0.03 3 0.84 (0.60, 1.18) 0 0.79 Asia 2 0.69 (0.37, 1.28) 97.8 <0.0001 1 0.80 (0.51, 1.24) 1 1.23 (0.28, 5.42) Australia 0 1 1.42 (0.75, 2.68) 1 0.60 (0.28, 1.31) Number of cases <1000 4 0.98 (0.92, 1.04) 0 0.75 0.32 2 1.22 (0.83, 1.81) 0 0.58 0.16 3 0.96 (0.64, 1.44) 31.9 0.23 0.83 1000 to <1500 1 0.95 (0.84, 1.08) 2 0.74 (0.54, 1.01) 0 0.60 2 0.86 (0.44, 1.68) 0 0.60 ≥1500 2 0.72 (0.36, 1.43) 98.9 <0.0001 3 0.60 (0.36, 1.00) 87.0 <0.0001 3 0.92 (0.86, 0.99) 0 0.74 Adjustment for confounding factors Alcohol Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 6 0.73 (0.55, 0.96) 71.7 0.003 0.30 6 0.93 (0.86, 1.00) 0 0.82 0.24 No 2 0.99 (0.92, 1.07) 0 0.33 1 1.42 (0.75, 2.68) 2 0.59 (0.31, 1.16) 0 0.94 Smoking Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) BMI, waist circumference Yes 6 0.84 (0.66, 1.07) 94.4 <0.0001 0.61 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 1 0.99 (0.92, 1.07) 0 1 0.57 (0.16, 1.99) Physical activity Yes 5 0.85 (0.66, 1.10) 95.5 <0.0001 0.84 7 0.78 (0.60, 1.02) 69.9 0.003 NC 7 0.92 (0.86, 0.99) 0 0.76 0.48 No 2 0.99 (0.92, 1.07) 0 0.33 0 1 0.57 (0.16, 1.99) Coffee, caffeine Yes 4 1.00 (0.95, 1.05) 0 0.65 0.17 3 0.91 (0.82, 1.01) 0 0.59 0.39 4 0.93 (0.86, 1.00) 0 0.58 0.37 No 3 0.71 (0.41, 1.24) 97.4 <0.0001 4 0.64 (0.37, 1.10) 77.4 0.004 4 0.79 (0.58, 1.09) 0 0.79 Soft drinks Yes 1 1.02 (0.96, 1.09) 0.54 1 0.91 (0.82, 1.01) 0.72 1 0.92 (0.86, 1.00) 0.91 No 6 0.84 (0.66, 1.06) 93.8 <0.0001 6 0.74 (0.51, 1.06) 70.3 0.005 7 0.91 (0.72, 1.14) 0 0.69 Meat Yes 3 1.01 (0.95, 1.06) 0 0.62 0.30 2 0.92 (0.83, 1.02) 0 0.40 0.40 3 0.93 (0.86, 1.00) 0 0.40 0.42 No 4 0.77