ABSTRACT
Background
Yogurt consumption and low-fat dairy consumption have been associated with reduced incidence of type 2 diabetes (T2D) in some studies.
Objective
We assessed the relation of yogurt and other dairy consumption to incidence of T2D in black women, a population group with a disproportionately high incidence of T2D.
Methods
The Black Women's Health Study has followed 59,000 US black women since 1995 through biennial questionnaires which update health information. Each questionnaire inquired about doctor-diagnosed diabetes in the previous 2 y. FFQs completed by participants in 1995 and 2001 provided information on yogurt and other dietary intake. HRs with 95% CIs for yogurt (nonfrozen or frozen) and other dairy consumption in relation to incident T2D (n = 8061 cases) were estimated with Cox proportional hazards regression, controlling for risk factors for T2D.
Results
The HR for consumption of ≥1 serving of yogurt/d relative to <1 serving/mo was 0.99 (95% CI: 0.87, 1.13, P trend = 0.65) after control for dietary and nondietary risk factors for T2D. The multivariable HR was 0.97 (95% CI: 0.75, 1.27; P trend = 0.74) for 2 or more servings/d of low-fat dairy other than yogurt relative to <1 serving/mo and 1.06 (95% CI: 0.91, 1.25, P trend = 0.36) for 2 or more servings/d of regular dairy relative to <1 serving/mo.
Conclusion
Results from this study do not support an inverse association of yogurt consumption or other dairy consumption with T2D risk in black women.
Keywords: yogurt, dairy, type 2 diabetes, black, African American, women
Introduction
Type 2 diabetes (T2D) affects >30 million Americans, with African Americans affected almost twice as often as whites (1). The prevalence has continued to rise (2). Yogurt has been found to improve lipid profiles and antioxidant status in T2D patients (3, 4) and to have beneficial effects on cholesterol concentrations (5), which may reflect effects of probiotic bacteria in yogurt. Several meta-analyses have found that yogurt consumption is associated with decreased T2D risk (6–9), with an 18% reduction estimated for a serving of yogurt in a 2014 meta-analysis (6) and a 14% reduction in a 2017 meta-analysis (8). The addition of yogurt to one's diet might be a behavioral change that could be successfully implemented, but convincing scientific evidence is needed to support such a recommendation.
Dairy products in general are high in calcium, magnesium, vitamin D, whey protein, and certain fatty acids, which may contribute to lower body fat and insulin resistance (7, 10, 11). For that reason, numerous observational studies have assessed low-fat and regular dairy intake in relation to T2D. Results have generally ranged from null to inverse associations with T2D, with low-fat consumption more often associated with lower risk than regular consumption (6, 7, 9, 12).
Dietary studies of yogurt and other dairy intake in relation to T2D to date have been conducted in populations consisting mostly of whites (8). The credibility of a causal inverse association of yogurt or other dairy products with T2D would be increased by documentation of similar associations in different populations having different eating habits. Black and white Americans differ in their eating patterns (13–16) and black women are of particular interest for T2D prevention because of their disproportionately high incidence of T2D (1). We assessed whether yogurt intake and other dairy intake are associated with a reduced incidence of T2D among black women, using data collected in a large follow-up study, the Black Women's Health Study (BWHS).
Methods
The BWHS
The BWHS began in 1995 when health questionnaires with an invitation to participate in a long-term health study of black women were mailed to over 400,000 subscribers to a magazine targeted to black women (Essence), members of the Black Nurses’ Association, and friends and relatives of early respondents (17, 18). A total of 59,000 black women aged 21–69 y from across the continental USA enrolled by filling out the questionnaires. The median age of participants was 38 y, and >80% were from California, Georgia, Illinois, Indiana, Louisiana, Maryland, Massachusetts, Michigan, New Jersey, New York, South Carolina, Virginia, and the District of Columbia. The questionnaires obtained data on demographics, health history, behaviors, dietary intake, and a range of other factors that might affect health. The participants have been followed biennially by mailed and web questionnaires on which they update health information (available at www.bu.edu/bwhs). Deaths are identified through yearly linkage to the National Death Index. After 11 completed biennial follow-up cycles, follow-up has been successful for 85% of potential person-years. The study was approved by the Institutional Review Board of Boston University Medical Center. Participants indicate their consent to participate by filling out questionnaires.
Cases of T2D
Every biennial questionnaire asked about physician-diagnosed diabetes during the previous 2 y. For the current analyses, self-reported cases of incident T2D were ascertained from all follow-up questionnaires from 1997 through to 2015. We excluded cases diagnosed before the age of 30 y to eliminate possible cases of type 1 diabetes. In a BWHS validation study of 293 women who reported having been diagnosed with diabetes and gave permission to contact their physicians (19), the physicians provided data on 229 women, of whom 217 (95%) were confirmed to have T2D.
Dietary information
Diet was assessed in 1995 and in 2001 with modified versions of the National Cancer Institute-Block short-form FFQ (20, 21). The 1995 FFQ asked about consumption of 68 food items in the previous year, with frequencies ranging from “never or <1/mo” to “≥2/d.” For each item, the questionnaire indicated the size of a “medium” serving [e.g., 8 ounce (226 g) glass], and respondents then indicated the serving size that they consumed as “small,” “medium,” or “large,” where small was ½ of medium size and large was 1.5 times medium size. The 2001 FFQ included an additional portion size, “super,” which was 2 times medium size. As an example, the medium serving size of a glass of skim milk was specified as 8 ounces. A woman who indicated that she had 2 large servings of skim milk each day would be considered to have consumed 3 medium servings (2 × 1.5 = 3) per day. DIETSYS software (version 4.01) was used to calculate nutrients in 1995 (20) and DIETCALC (version 1.4.1, National Cancer Institute) was used in 2001.
Yogurt (nonfrozen) and frozen yogurt were asked about separately on both the 1995 and 2001 FFQs. Our main analysis was of total yogurt consumption (sum of nonfrozen and frozen yogurt). In a sensitivity analysis we assessed nonfrozen yogurt consumption, which accounted for 94% of total yogurt consumption. The questionnaires also asked about whole milk; 2% milk; 1% milk; skim milk; buttermilk; milk/cream in coffee/tea; regular ice cream; low-fat ice cream; and cheese and cheese spreads (not cottage cheese). Total low-fat dairy intake was the sum of servings of skim/low-fat milk, low-fat ice cream, and yogurt. Evidence from food tables (21) and other studies of yogurt (22, 23) supports classification of yogurt as low-fat. Regular dairy intake was the sum of servings of whole milk, milk/cream in coffee/tea, regular ice cream, and cheese and cheese spreads. The 1995 FFQ asked about ice cream as a single item. We assigned ice cream reported in 1995 as regular or low-fat according to the participant's answer to the question “How often do you eat a low-fat or nonfat version of ice cream?” Ice cream intake was assigned as regular dairy for women who said they rarely ate low-fat/nonfat ice cream, and otherwise was assigned as low-fat. The 2001 FFQ asked about regular and low-fat ice cream separately.
In a validation study of the 1995 FFQ, each of ∼400 BWHS participants provided 3 nonconsecutive 24-h recall interviews by telephone and 1 3-d food record over a period of a year. Energy-adjusted and deattenuated Pearson correlations of FFQ nutrient estimates with averages from the combined recall/record data ranged from 0.5–0.8 (21); the correlation coefficient for calcium, a prominant constituent of dairy products, was 0.58 (95% CI: 0.23, 0.89) In a comparison of food group servings estimated by the FFQ and the average of the 3 dietary recalls, mean servings per day for the total dairy group was 1.0 ± 1.3 (SD) from the FFQ and 1.1 ± 1.3 (SD) from the dietary recalls (24).
Covariates
The baseline questionnaire in 1995 collected information on height, weight, alcohol consumption, cigarette smoking, vigorous exercise, years of education, geographic region, and other factors. Biennial follow-up questionnaires updated variables and ascertained incident disease. A neighborhood socioeconomic status (SES) score was derived for each address at which a participant lived during follow-up, based on census block data on median household income; median housing value; percent households receiving interest, dividends, or net rental income; percent adults ≥25 y who had completed college; percent employed persons aged ≥16 in white-collar occupations; and percent of families with children not headed by a single female (25). The American Healthy Eating Index (AHEI)-2010 is a dietary index based on the daily intake of several foods or constituents of foods, including vegetables, fruits, whole grains, and red/processed meat, that have been associated with the risk of some chronic diseases; the index does not include dairy products (26). Based on reported intake from the FFQs, we computed AHEI-2010 scores; a higher score, which indicates a diet higher in grains and vegetables and lower in red and processed meats, has been strongly inversely associated with diabetes (26).
Analytic cohort
Follow-up for the present analyses began in 1995. We excluded women who, at baseline, reported diabetes (n = 2648) at age 30 y or older; diabetes diagnosed before age 30 y (n = 89); a history of myocardial infarction, stroke, or cancer (n = 2217); were pregnant (n = 972); had unknown height or weight (n = 664); or had implausible energy intake (<500 or ≥3800 kcal/d) or 10 or more missing items on the 1995 FFQ (n = 5133), which left 47,227 women in the analytic cohort (Supplemental Figure 1). During follow-up from 1995 to 2015, 8061 participants aged 30 y or older reported a diagnosis of incident diabetes.
Statistical methods for analyses of yogurt and dairy consumption in relation to incident T2D
We assessed 1995 dietary intake in relation to T2D diagnosed until 2001 and the average of intake reported on the 1995 and 2001 FFQs in relation to T2D diagnosed from 2001 to 2015. Participants for whom 2001 FFQ data was missing were assigned their 1995 FFQ data values for 2001. In addition, in a sensitivity analysis we assessed an update of dietary intake between 1995 and 2001 rather than the cumulative average. We also considered intake as a continuous variable.
Of the 47,227 women in the analytic cohort, all of whom provided data on yogurt consumption on the 1995 FFQ, 30,003 (64%) provided yogurt consumption data on the 2001 FFQ as well. The mean of total yogurt servings per day was 0.17 in 1995 and 0.18 in 2001 (r = 0.30, P <0.001). For low-fat dairy other than yogurt, the mean of servings per day was 0.44 in 1995 and 0.34 in 2001 (r = 0.31, P <001). The mean of servings per day of regular dairy was 0.84 in 1995 and 0.86 in 2001 (r = 0.39, P <0.001).
Cox proportional hazards regression was used to estimate HRs and 95% CIs for yogurt intake (categorized according to the number of servings) in relation to incident T2D adjusted for age (1-y intervals), 2-y time period (1995–1997, ….2013–2015), and energy intake (quintiles). The FFQs asked about food consumption in categories of servings: <1/mo, 1/mo, 2–3/mo, 1/wk, 2/wk, 3–4/wk, 5–6/wk, 1/d, and ≥2/d. For yogurt consumption, because there were insufficient numbers for an informative analysis of ≥2 servings/d, we assessed the combined category of 1/d and ≥2/d; we also combined categories of low consumption (e.g., <1/mo with 2–3/mo) for which there has been no evidence of an association with risk of T2D. Person-years of follow-up were calculated from entry to the study in 1995 until diagnosis of diabetes, death, loss to follow-up (year of last questionnaire completed), or end of the 2015 follow-up, whichever came first. We estimated HRs with control for age, calendar time, and energy intake in quintiles. In multivariable analyses, we controlled in addition for established or potential risk factors for T2D, using the Andersen-Gill data structure to update time-varying factors for which data was collected across questionnaire cycles (27): BMI (kg/m2 continuous), vigorous exercise (<1 h/wk,1–2 h/wk, ≥3 h/wk), neighborhood SES score (quintiles), years of education (≤12, 13–15, 16, ≥17), smoking status (current, not current), pack years of smoking (<10, ≥10), and alcohol consumption (never drank, ex-drinker, current drinker of 1–3 drinks/wk, current drinker of 4–6 drinks/wk, current drinker of ≥7 drinks/wk). We repeated the multivariable analyses controlling in addition for AHEI-2010 score (quintiles) and consumption of unprocessed and processed red meat (servings/d continuous), fruits and vegetables (servings/d continuous), cereal fiber (g/d continuous), sugar-sweetened beverages (sodas, fruit drinks, and Kool-aid) (servings/d continuous), coffee (servings/d continuous), low-fat dairy other than yogurt (servings/d continuous), and regular dairy (servings/d continuous). Similar analyses were carried out for low-fat dairy intake other than yogurt and for regular dairy intake. In sensitivity analyses in which we excluded the first 2 y of follow-up, results were unchanged. We tested for trend across categories of yogurt and dairy consumption by treating the median value in each of the consumption categories as a continuous variable in the models.
We investigated the risk of T2D of substituting 1 serving/d of another food item for a serving per day of yogurt by including both variables in continuous form in the same multivariate model. The difference in their regression coefficients was used to estimate the HR for the substitution and the covariance to estimate the 95% CI (28).
Secondary analyses to assess potential bias from undiagnosed diabetes
We used blood samples collected from 10,777 BWHS participants during 2013–2018 (29) who had not been diagnosed with T2D to assess whether reporting of yogurt consumption differed between participants with and without undiagnosed T2D. Hemoglobin (Hb) A1c concentrations, used clinically to identify individuals who may have undiagnosed diabetes, had been measured in the blood samples (30). All of these participants had completed the 1995 FFQ. We compared reporting of yogurt consumption of women with concentrations of HbA1c <6.5% to those of women with concentrations ≥6.5%.
Results
The 47,277 participants were followed for a mean of 18 y; 8061 incident cases of T2D were reported. Characteristics of BWHS participants at baseline in the analytic cohort according to total intake of yogurt (nonfrozen and frozen) in 1995 are shown in Table 1. Yogurt consumption was greater among women who were older, lived in neighborhoods with a higher neighborhood SES score, had higher AHEI-2010 scores, had higher caloric intake, more often exercised vigorously, more often did not smoke, and had greater educational attainment. In addition, women who consumed more yogurt also consumed more fruits and vegetables, more cereal fiber, and more low-fat dairy other than yogurt. As shown in Supplemental Table S1, low-fat dairy intake other than yogurt in 1995 was greater among women with many of the same characteristics as those who consumed more yogurt. By contrast, as shown in Supplemental Table S2, intake of regular dairy was greater among women who had lower neighborhood SES scores and lower AHEI-scores, were less likely to exercise vigorously, were current smokers, had lower educational attainment, were current alcohol drinkers, and consumed more red meat, more cereal fiber, more sugar-sweetened beverages, and more coffee.
TABLE 1.
Characteristics of Black Women's Health Study participants in 1995 by yogurt1 consumption (n = 47,277)
| Servings of yogurt | ||||||
|---|---|---|---|---|---|---|
| Characteristic2,3 | <1/mo (n = 19,392) | 1–3/mo (n = 11,315) | 1–2/wk (n = 8491) | 3–4/wk (n = 4080) | 5–6/wk (n = 1827) | ≥1/d (n = 2172) |
| Age, y | 37.8 ± 10.2 | 37.8 ± 10.0 | 38.4 ± 10.3 | 39.2 ± 10.5 | 39.7 ± 11.3 | 40.2 ± 11.0 |
| Neighborhood SES score | −0.05 ± 0.9 | 0.04 ± 0.9 | 0.1 ± 0.9 | 0.1 ± 0.9 | 0.2 ± 0.6 | 0.2 ± 1.0 |
| AHEI-2010 score | 36.9 ± 9.4 | 38.9 ± 9.3 | 40.3 ± 9.4 | 41.6 ± 9.5 | 42.5 ± 9.6 | 43.7 ± 10.5 |
| Total energy intake, kcal/d | 1386 ± 644 | 1405 ± 635 | 1499 ± 637 | 1599 ± 639 | 1712 ± 642 | 1971 ± 702 |
| Current BMI, kg/m2 | 27.4 ± 6.5 | 27.8 ± 6.4 | 28.0 ± 6.5 | 27.7 ± 6.1 | 28.0 ± 6.5 | 28.2 ± 6.6 |
| Vigorous exercise, % | ||||||
| <1 h/wk | 37 | 29 | 24 | 23 | 20 | 22 |
| 1–2 h/wk | 36 | 41 | 41 | 39 | 37 | 36 |
| ≥3 h/wk | 23 | 26 | 32 | 36 | 39 | 39 |
| Smoking status, % | ||||||
| Never smoked | 63 | 67 | 68 | 68 | 67 | 68 |
| Past smoker | 18 | 19 | 20 | 21 | 22 | 22 |
| Current smoker | 19 | 14 | 12 | 11 | 10 | 10 |
| Family history of diabetes, % | 36 | 36 | 37 | 37 | 37 | 36 |
| Education, % | ||||||
| ≤12 y | 21 | 16 | 14 | 13 | 12 | 13 |
| 13–15 y | 38 | 36 | 35 | 34 | 32 | 35 |
| ≥16 y | 41 | 47 | 52 | 53 | 56 | 52 |
| Alcohol intake, % | ||||||
| Never drank | 56 | 57 | 57 | 56 | 56 | 56 |
| Past drinker | 13 | 14 | 13 | 14 | 14 | 14 |
| Current drinker | 30 | 29 | 29 | 29 | 30 | 29 |
| Red meat,4 s/d | 0.7 ± 0.7 | 0.6 ± 0.6 | 0.6 ± 0.6 | 0.6 ± 0.6 | 0.5 ± 0.6 | 0.6 ± 0.7 |
| Fruits and vegetables, s/d | 1.8 ± 1.6 | 2.1 ± 1.7 | 2.4 ± 1.7 | 2.8 ± 1.8 | 3.0 ± 1.9 | 3.6 ± 2.4 |
| Cereal fiber,5 g/d | 3.5 ± 2.7 | 3.9 ± 2.7 | 4.4 ± 2.9 | 4.7 ± 3.0 | 5.0 ± 3.2 | 5.9 ± 3.7 |
| SSB,6 s/d | 2.0 ± 1.7 | 1.9 ± 1.6 | 1.8 ± 1.5 | 1.8 ± 1.6 | 1.8 ± 1.6 | 1.8 ± 1.6 |
| Coffee, s/d | 0.6 ± 1.3 | 0.6 ± 1.3 | 0.6 ± 1.3 | 0.7 ± 1.3 | 0.7 ± 1.3 | 0.7 ± 1.5 |
| Low-fat dairy,7 s/d | 0.3 ± 0.7 | 0.4 ± 0.8 | 0.5 ± 0.9 | 0.6 ± 1.0 | 0.7 ± 1.0 | 0.9 ± 1.3 |
| Regular dairy, s/d | 0.9 ± 1.4 | 0.9 ± 1.3 | 0.9 ± 1.3 | 1.0 ± 1.4 | 1.0 ± 1.3 | 1.0 ± 1.5 |
Sum of nonfrozen and frozen yogurt.
All variables are continuous unless % is specified; for continuous variables, means ± SDs are given.
Values are standardized to the age distribution of the 47,277 participants at baseline.
s/d is servings/d.
g/d is grams/d.
Sugar-sweetened beverages.
Low-fat dairy excludes yogurt.
AHEI, American Healthy Eating Index; BMI, body mass index; SES, socioeconomic status.
As shown in Table 2, the multivariable HR for ≥1 serving of yogurt/d (nonfrozen plus frozen) (based on 288 cases in that exposure category) relative to <1 serving/mo (3070 cases) was 0.90 (95% CI: 0.79, 1.02) after controlling for all potential confounders except dietary variables. When AHEI-2010 score and intake of red meat, fruits and vegetables, cereal fiber, sugar-sweetened beverages, coffee, low-fat dairy other than yogurt, and regular yogurt were also controlled, the HR was 0.99 (95% CI: 0.87,1.13; P trend = 0.65). All HRs for the other categories of consumption were also close to 1.0 and compatible with 1.0. Yogurt consumption exceeding 1 serving/d was infrequent, with only 21 cases reporting ≥2 servings of yogurt/d. In a sensitivity analysis using updating of yogurt consumption from 1995 to 2001 rather than a cumulative average, the HRs for yogurt servings of ≥1/d, 5–6/wk, 3–4/wk, 1–2/wk, and 1–3/mo, relative to <1/mo, were 0.96 (95% CI: 0.86, 1.08), 0.93 (95% CI: 0.82, 1.06), 1.01 (95% CI: 0.93, 1.06), 1.00 (95% CI: 0.94, 1.07), and 1.03 (95% CI: 0.97,1.09), respectively, similar to those from the cumulative average analysis. An analysis of yogurt servings as a continuous variable also yielded results similar to those from the cumulative average analysis: P trend = 0.85. In an analysis in which consumption of nonfrozen yogurt only was assessed, based on 228 cases consuming ≥1 serving/d relative to <1 serving/mo (3352 cases), the HR after controlling all confounders was 0.96 (95% CI: 0.83, 1.10); the estimates for servings of 5–6/wk, 3–4/wk, 1–2/wk, and 1–3/mo were 0.97 (95% CI: 0.92, 1.08), 1.00 (95% CI: 0.92, 1.08), 1.00 (95% CI: 0.94, 1.07), and 1.02 (95% CI: 0.97, 1.08), respectively.
TABLE 2.
Consumption of yogurt,1 low-fat dairy other than yogurt, and regular dairy in relation to incidence of type 2 diabetes (T2D)
| Servings of yogurt | ||||||||
|---|---|---|---|---|---|---|---|---|
| <1/mo | 1–3/mo | 1–2/wk | 3–4/wk | 5–6/wk | ≥1/d | P trend | ||
| T2D cases | 3070 | 1927 | 1550 | 903 | 323 | 288 | ||
| Person-years | 328,287 | 199,949 | 164,791 | 93,208 | 35,359 | 31,439 | ||
| HR2 (95% CI) | Ref | 1.02 (0.96, 1.08) | 0.95 (0.89, 1.01) | 0.91 (0.85, 0.98) | 0.84 (0.75, 0.94) | 0.82 (0.73, 0.93) | <0.0001 | |
| HR3 (95% CI) | Ref | 1.02 (0.96, 1.08) | 0.97 (0.91, 1.03) | 0.97 (0.90, 1.05) | 0.89 (0.79, 1.00) | 0.90 (0.79, 1.02) | 0.01 | |
| HR4 (95% CI) | Ref | 1.04 (0.98, 1.10) | 1.00 (0.94, 1.07) | 1.03 (0.95, 1.12) | 0.96 (0.85, 1.08) | 0.99 (0.87, 1.13) | 0.65 | |
| Servings of low-fat dairy other than yogurt | ||||||||
| <1/mo | 1–3/mo | 1–2/wk | 3–4/wk | 5–6/wk | 1/d | ≥2/d | P trend | |
| T2D cases | 2171 | 1501 | 1362 | 1217 | 649 | 830 | 331 | |
| Person-years | 249,638 | 159,667 | 143,106 | 126,432 | 56,273 | 83,390 | 34,528 | |
| HR2 (95% CI) | Ref | 1.09 (1.02, 1.17) | 1.08 (1.01, 1.16) | 1.08 (1.00, 1.15) | 1.25 (1.14, 1.37) | 1.07 (0.99, 1.16) | 0.98 (0.87, 1.10) | 0.92 |
| HR3 (95% CI) | Ref | 1.02 (0.95, 1.09) | 0.97 (0.90, 1.03) | 0.98 (0.91, 1.05) | 1.10 (1.01, 1.21) | 0.96 (0.88, 1.04) | 0.88 (0.78, 0.99) | 0.06 |
| HR5 (95% CI) | Ref | 1.03 (0.97, 1.10) | 0.99 (0.92, 1.06) | 1.01 (0.93, 1.09) | 1.15 (1.04, 1.28) | 1.01 (0.90, 1.14) | 0.97 (0.75, 1.27) | 0.74 |
| Servings of regular dairy | ||||||||
| <1/mo | 1–3/mo | 1–2/wk | 3–4/wk | 5–6/wk | 1/d | ≥2/d | P trend | |
| T2D cases | 678 | 912 | 1206 | 1372 | 924 | 1834 | 1135 | |
| Person-years | 78,327 | 103,073 | 140,608 | 144,402 | 90,299 | 188,109 | 108,216 | |
| HR2 (95% CI) | Ref | 1.08 (0.98, 1.19) | 1.09 (0.99, 1.20) | 1.18 (1.07, 1.29) | 1.23 (1.12, 1.37) | 1.13 (1.03, 1.23) | 1.12 (1.01, 1.23) | 0.39 |
| HR3 (95% CI) | Ref | 1.02 (0.92, 1.12) | 0.99 (0.90, 1.09) | 1.06 (0.97, 1.17) | 1.12 (1.01, 1.24) | 1.03 (0.95, 1.13) | 1.05 (0.95, 1.16) | 0.37 |
| HR6 (95% CI) | Ref | 1.00 (0.90, 1.10) | 0.96 (0.87, 1.05) | 1.02 (0.93, 1.13) | 1.08 (0.97, 1.20) | 1.00 (0.90, 1.11) | 1.06 (0.91, 1.25) | 0.36 |
Yogurt is sum of nonfrozen and frozen yogurt.
Adjusted for age, energy intake, and questionnaire cycle.
Adjusted for factors in HR1 plus BMI, alcohol consumption, family history of T2D, neighborhood SES, education, vigorous exercise, smoking status, pack years.
Adjusted for factors in HR2 plus AHEI-2010 score and intake of red meat, fruits and vegetables, cereal fiber, sugar-sweetened beverages, coffee, low-fat dairy other than yogurt, and regular dairy.
Adjusted for factors in HR2 plus AHEI-2010 score and intake of red meat, fruits and vegetables, cereal fiber, sugar-sweetened beverages, coffee, yogurt, and regular dairy.
Adjusted for factors in HR2 plus AHEI-2010 score and intake of red meat, fruits and vegetables, cereal fiber, sugar-sweetened beverages, coffee, yogurt, and low-fat yogurt other than dairy.
AHEI, American Healthy Eating Index; SES, socioeconomic status; T2D, type 2 diabetes.
As shown in Table 2, the HR for ≥2 servings/d of low-fat dairy other than yogurt (331 cases) relative to <1 serving per/mo (2171 cases) was 0.88 (95% CI: 0.78, 0.99; P trend = 0.06) with control for all confounding factors except dietary factors and 0.97 (95% CI: 0.75, 1.27, P trend = 0.74) after control in addition for dietary factors. For regular dairy, the HR for ≥2 servings/d (1135 cases) relative to <1 serving/mo (678 cases) was 1.05 (95% CI: 0.95, 1.16, P trend = 0.37) before control for dietary factors and 1.06 (95% CI: 0.91, 1.25, P trend = 0.36) after control for dietary factors.
For the substitution of 1 serving of yogurt/d with red meat, there was a 15% higher risk of T2D (HR = 1.15, 95% CI: 1.06, 1.25). There was no risk of T2D when yogurt was substituted with fruits and vegetables (HR = 0.95, 95% CI: 0.88, 1.30), cereal fiber (HR = 0.94, 95% CI: 0.87, 1.02), sugar-sweetened beverages (HR = 0.98, 95% CI: 0.91, 1.06), and coffee (HR = 0.91, 95% CI: 0.85, 1.00).
HbA1c concentrations of ≥6.5% are suggestive of undiagnosed T2D (30). Reporting of yogurt consumption was similar for women with blood concentrations of HbA1c of <6.5% and ≥6.5%: in the former group, 1.9% reported 1 or more servings/d and 19.3% reported 1–6 servings/wk, compared with 2.1% and 20.7% in the ≥6.5% group.
Discussion
In this study of black women, consumption of a serving or more a day of yogurt (the sum of intake of nonfrozen and frozen yogurt) was associated with a 10% reduction in incidence of T2D after control for T2D risk factors but before allowance for other dietary intake. After control for AHEI-2010 score and foods associated with risk of T2D and yogurt consumption, the HR was 0.99 (95% CI: 0.87, 1.13, P trend = 0.65). Nonfrozen yogurt comprised most reported yogurt consumption. Frozen yogurt may not contain the probiotics or other constituents of nonfrozen yogurt that might be related to a reduced risk of T2D (31). However, a sensitivity analysis that excluded frozen yogurt yielded HRs similar to those for total yogurt.
In 2014, Chen et al. (6) meta-analyzed results on yogurt and dairy consumption in relation to T2D from 9 prospective studies that included over 400,000 participants. The HR from the meta-analysis for 1 serving of yogurt/d was 0.82 (95% CI: 0.70, 0.96). A subsequent meta-analysis in 2016 by Gijsbers et al. (7), which added several small prospective studies, estimated a HR of 0.86 (95% CI: 0.83, 0.90). A later review of more studies by Alvarez-Bueno concluded that yogurt was related to reduced risk (32). Recently, an assessment of changes in dairy consumption over the course of follow-up in the Nurses’ Health Studies and the Health Professionals Follow-up Study suggested that increased yogurt consumption over time was associated with reduced T2D risk (33). The vast majority of subjects in the studies of yogurt and T2D have been white. Thus, there is consistent evidence that yogurt consumption is associated with a decrease in T2D risk in white populations. To our knowledge, the present results on yogurt and T2D are the first based on a black population. Before control for AHEI-2010 score and dietary confounders, yogurt intake was associated with a 10% reduction in risk of T2D. Yogurt consumption in the BWHS was correlated with a higher AHEI-2010 score and higher consumption of fruits, vegetables, and whole grains and lower consumption of red meats, all of which are associated with a lower risk of T2D (26). Control for healthy eating behaviors associated with yogurt consumption weakened the initial inverse association between yogurt and T2D, resulting in no association.
Because calcium, magnesium, vitamin D, whey protein, and certain fatty acids in dairy products might lower body fat and insulin resistance, there has been interest in the relation of both low-fat and regular dairy intake to risk of T2D (7, 10, 11). The meta-analysis of Chen et al. found an overall HR of 0.98 (95% CI: 0.96, 1.01) for a serving a day of dairy as a whole; in a separate analysis of the 2 Nurses’ Health Studies of women and the Health Professionals study of men, there was no association with low-fat dairy (6). Some meta-analyses have found inverse associations with dairy intake, such as those of Aune et al. (12) and Alvarez-Bueno et al. (32) and low-fat dairy has been associated with reduced risk more often than regular dairy. There has been considerable variability in results across the various studies. Of note, Chen et al. (6) and Gijbers et al. (7) reported that studies with longer follow-up were more likely to find null associations. Recent findings in the Nurses’ Health Studies and Health Professionals studies provide some evidence against an inverse association with dairy consumption, as increases in low-fat and regular dairy consumption over the course of follow-up were unrelated to T2D risk (33), whereas the same study linked increased yogurt intake to decreased T2D risk. In the present study, 2 or more servings/d of low-fat dairy other than yogurt was associated with a 12% reduction in risk of T2D before controlling for other dietary factors; after control for the healthier eating habits of low-fat dairy consumers, there was no association. Regular dairy consumption was also not associated with risk.
Based on national data collected from 1994–2000, Fulgoni et al. reported that black Americans ate less dairy than nonblacks and that blacks were more likely to fail to meet recommended intake (34). A higher prevalence of lactose intolerance among African Americans might account in part for their lower dairy intake (35). The intake of dairy products by BWHS participants in 1995 was greater than that for black women and similar to that for nonblack women in the report by Fulgoni et al. (34), perhaps reflecting the higher SES of BWHS participants compared with black women nationally (36).
With regard to study strengths and limitations, prospective collection of the dietary data in the current study reduced the possibility of biased reporting. Risk factors for T2D identified in the BWHS have been consistent with the literature (19, 37–41), and important risk factors for T2D were controlled in the analyses. We also controlled for dietary factors associated with yogurt and dairy consumption. None of the risk factors can be reported or specified perfectly and therefore complete control is not possible.
The number of cases of T2D reported by participants was large, providing adequate statistical power. A validation study indicated that T2D is reported by BWHS participants with a high degree of accuracy (19). Our finding that reporting of yogurt consumption was similar regardless of HbA1c status suggests that undiagnosed T2D did not bias the results.
With regard to the estimation of dietary intake, a validation study in the BWHS indicated adequate reporting of a number of food items on the FFQs, but yogurt was not assessed specifically (21). The mean number of total dairy servings estimated by the 1995 FFQ and the average of the 3 dietary recalls were similar (1.0 ± 1.3 from the FFQ and 1.1 ± 1.0 from the dietary recalls) (24) and it was reassuring that yogurt consumption and low-fat dairy intake in the present study were associated with other healthy behaviors, as expected, whereas regular dairy intake was not. However, the FFQs were administered at only 2 points during follow-up in the BWHS and some women will have changed their eating habits over the course of follow-up, leading to imprecision in the effect estimates. In some of the studies that observed an inverse association between yogurt consumption and T2D, diet was assessed at multiple time points (6).
The BWHS was not a probability sample of US black women. However, the participants were drawn from multiple states across the USA and represent a range of educational and socioeconomic levels (17, 42, 43). Thus, results are likely generalizable to many black women.
In conclusion, the present study does not support an association of yogurt consumption or other dairy intake with lower risk of T2D. More studies of African Americans are needed to clearly establish the association of yogurt and other dairy consumption with T2D in this population group.
Supplementary Material
Acknowledgments
The authors’ contributions were as follows—LR and JRP: designed and conducted the research; LR and JRP: designed and oversaw data analyses; YPN: conducted data analyses; SL: advised on analyses and provided computer code; EAR-N: advised on analyses; LR and JRP: wrote the manuscript; LR: had primary responsibility for final content; and all authors: read and approved the final manuscript. LR, YPR, SL, EAR-N, and JRP report no conflicts of interest.
Notes
Supported by R01CA058420, UM1CA164974, and U01CA164974 from the National Cancer Institute.
Supplemental Figure 1 and Supplemental Tables 1 and 2 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.
Abbreviations used: AHEI: American Healthy Eating Index; BWHS, Black Women's Health Study; SES, socioeconomic status; T2D, type 2 diabetes.
Contributor Information
Lynn Rosenberg, Slone Epidemiology Center at Boston University, Boston, MA, USA.
Yvonne P Robles, Slone Epidemiology Center at Boston University, Boston, MA, USA.
Shanshan Li, Slone Epidemiology Center at Boston University, Boston, MA, USA.
Edward A Ruiz-Narvaez, Department of Nutritional Services, University of Michigan School of Public Health, Ann Arbor, MI, USA.
Julie R Palmer, Slone Epidemiology Center at Boston University, Boston, MA, USA.
References
- 1. National Diabetes Statistics Report, 2017 [Internet]. Atlanta, GA: Centers for Disease Control and Prevention, Department of Health and Human Services;2017. Available from: https://www.cdc.gov/diabetes/data/statistics/statistics-report.html. (Accessed 24 April, 2020). [Google Scholar]
- 2. Geiss LS, Wang J, Cheng YJ, Thompson TJ, Barker L, Li Y, Albright AL, Gregg EW. Prevalence and incidence trends for diagnosed diabetes among adults aged 20 to 79 years, United States, 1980–2012. JAMA. 2014;312(12):1218–26. [DOI] [PubMed] [Google Scholar]
- 3. Ejtahed HS, Mohtadi-Nia J, Homayouni-Rad A, Niafar M, Asghari-Jafarabadi M, Mofid V. Probiotic yogurt improves antioxidant status in type 2 diabetic patients. Nutrition. 2012;28(5):539–43. [DOI] [PubMed] [Google Scholar]
- 4. Ejtahed HS, Mohtadi-Nia J, Homayouni-Rad A, Niafar M, Asghari-Jafarabadi M, Mofid V, Akbarian-Moghari A. Effect of probiotic yogurt containing Lactobacillus acidophilusand Bifidobacterium lactis on lipid profile in individuals with type 2 diabetes mellitus. J Dairy Sci. 2011;94(7):3288–94. [DOI] [PubMed] [Google Scholar]
- 5. Parvez S, Malik KA, Ah Kang S, Kim HY. Probiotics and their fermented food products are beneficial for health. J Appl Microbiol. 2006;100(6):1171–85. [DOI] [PubMed] [Google Scholar]
- 6. Chen M, Sun Q, Giovannucci E, Mozaffarian D, Manson JE, Willett WC, Hu FB. Dairy consumption and risk of type 2 diabetes: 3 cohorts of US adults and an updated meta-analysis. BMC Med. 2014;12:215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Gijsbers L, Ding EL, Malik VS, de Goede J, Geleijnse JM, Soedamah-Muthu SS. Consumption of dairy foods and diabetes incidence: a dose-response meta-analysis of observational studies. Am J Clin Nutr. 2016;103(4):1111–24. [DOI] [PubMed] [Google Scholar]
- 8. Salas-Salvado J, Guasch-Ferre M, Diaz-Lopez A, Babio N. Yogurt and diabetes: overview of recent observational studies. J Nutr. 2017;147(7):1452S–61S. [DOI] [PubMed] [Google Scholar]
- 9. Alvarez-Bueno C, Cavero-Redondo I, Martinez-Vizcaino V, Sotos-Prieto M, Ruiz JR, Gil A. Effects of milk and dairy product consumption on type 2 diabetes: overview of systematic reviews and meta-analyses. Adv Nutr. 2019;10(suppl_2):S154–S63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Rice BH, Cifelli CJ, Pikosky MA, Miller GD. Dairy components and risk factors for cardiometabolic syndrome: recent evidence and opportunities for future research. Adv Nutr. 2011;2(5):396–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jakubowicz D, Froy O. Biochemical and metabolic mechanisms by which dietary whey protein may combat obesity and type 2 diabetes. J Nutr Biochem. 2013;24(1):1–5. [DOI] [PubMed] [Google Scholar]
- 12. Aune D, Norat T, Romundstad P, Vatten LJ. Dairy products and the risk of type 2 diabetes: a systematic review and dose-response meta-analysis of cohort studies. Am J Clin Nutr. 2013;98(4):1066–83. [DOI] [PubMed] [Google Scholar]
- 13. Kell KP, Judd SE, Pearson KE, Shikany JM, Fernandez JR. Associations between socio-economic status and dietary patterns in US black and white adults. Br J Nutr. 2015;113(11):1792–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hiza HA, Casavale KO, Guenther PM, Davis CA. Diet quality of Americans differs by age, sex, race/ethnicity, income, and education level. J Acad Nutr Diet. 2013;113(2):297–306. [DOI] [PubMed] [Google Scholar]
- 15. Kirkpatrick SI, Dodd KW, Reedy J, Krebs-Smith SM. Income and race/ethnicity are associated with adherence to food-based dietary guidance among US adults and children. J Acad Nutr Diet. 2012;112(5):624–35.e6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Beydoun MA, Gary TL, Caballero BH, Lawrence RS, Cheskin LJ, Wang Y. Ethnic differences in dairy and related nutrient consumption among US adults and their association with obesity, central obesity, and the metabolic syndrome. Am J Clin Nutr. 2008;87(6):1914–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rosenberg L, Adams-Campbell L, Palmer JR. The Black Women's Health Study: a follow-up study for causes and preventions of illness. J Am Med Womens Assoc. 1995;50(2):56–8. [PubMed] [Google Scholar]
- 18. Russell C, Palmer JR, Adams-Campbell LL, Rosenberg L. Follow-up of a large cohort of black women. Am J Epidemiol. 2001;154(9):845–53. [DOI] [PubMed] [Google Scholar]
- 19. Palmer JR, Boggs DA, Krishnan S, Hu FB, Singer M, Rosenberg L. Sugar-sweetened beverages and incidence of type 2 diabetes mellitus in African American women. Arch Intern Med. 2008;168(14):1487–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Block G, Hartman AM, Naughton D. A reduced dietary questionnaire: development and validation. Epidemiology. 1990;1(1):58–64. [DOI] [PubMed] [Google Scholar]
- 21. Kumanyika SK, Mauger D, Mitchell DC, Phillips B, Smiciklas-Wright H, Palmer JR. Relative validity of food frequency questionnaire nutrient estimates in the Black Women's Health Study. Ann Epidemiol. 2003;13(2):111–18. [DOI] [PubMed] [Google Scholar]
- 22. Hu FB, Stampfer MJ, Manson JE, Ascherio A, Colditz GA, Speizer FE, Hennekens CH, Willett WC. Dietary saturated fats and their food sources in relation to the risk of coronary heart disease in women. Am J Clin Nutr. 1999;70(6):1001–8. [DOI] [PubMed] [Google Scholar]
- 23. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med. 2011;364(25):2392–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Mitchell D, Lancaster K, Smiciklas-Wright H. Comparison of food group intakes assessed by food-frequency questionnaire vs. 24-hour dietary recalls in the Black Women's Health Study dietary substudy. The 4th International Conference on Dietary Assessment Methods; 17–20 Sept, 2000; Tucson, Arizona (abstr).
- 25. Palmer JR, Boggs DA, Wise LA, Adams-Campbell LL, Rosenberg L. Individual and neighborhood socioeconomic status in relation to breast cancer incidence in African-American women. Am J Epidemiol. 2012;176(12):1141–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Chiuve SE, Fung TT, Rimm EB, Hu FB, McCullough ML, Wang M, Stampfer MJ, Willett WC. Alternative dietary indices both strongly predict risk of chronic disease. J Nutr. 2012;142(6):1009–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Therneau T. Extending the Cox model. In: Lin DY, Fleming TReditors. Proceedings of the First Seattle Symposium in Biostatistics: Survival Analysis. New York: Springer Verlag; 1997. p.51–84. [Google Scholar]
- 28. Song M, Giovannucci E. Substitution analysis in nutritional epidemiology: proceed with caution. Eur J Epidemiol. 2018;33(2):137–40. [DOI] [PubMed] [Google Scholar]
- 29. Palmer JR, Gerlovin H, Bethea TN, Bertrand KA, Holick MF, Ruiz-Narvaez EN, Wise LA, Haddad SA, Adams-Campbell LL, Kaufman HW et al. Predicted 25-hydroxyvitamin D in relation to incidence of breast cancer in a large cohort of African American women. Breast Cancer Res. 2016;18(1):86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Gillett MJ. International Expert Committee report on the role of the A1c assay in the diagnosis of diabetes. Diabetes Care. 2009;32(7):1327–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Aryana KJ, Olson DW. A 100-year review: yogurt and other cultured dairy products. J Dairy Sci. 2017;100(12):9987–10013. [DOI] [PubMed] [Google Scholar]
- 32. Alvarez-Bueno C, Cavero-Redondo I, Sotos-Prieto M, Gil A, Martinez-Vizcaino V, Ruiz JR. Milk and dairy product consumption and risk of mortality: an overview of systematic reviews and meta-analyses. Adv Nutr. 2019;10(suppl_2):S97–S104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Drouin-Chartier JP, Li Y, Ardisson Korat AV, Ding M, Lamarche B, Manson JE, Rimm EB, Willett WC, Hu FB. Changes in dairy product consumption and risk of type 2 diabetes: results from 3 large prospective cohorts of US men and women. Am J Clin Nutr. 2019;110(5):1201–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Fulgoni V 3rd, Nicholls J, Reed A, Buckley R, Kafer K, Huth P, DiRienzo D, Miller GD. Dairy consumption and related nutrient intake in African-American adults and children in the United States: continuing survey of food intakes by individuals 1994–1996, 1998, and the National Health and Nutrition Examination Survey 1999–2000. J Am Diet Assoc. 2007;107(2):256–64. [DOI] [PubMed] [Google Scholar]
- 35. Bailey RK, Fileti CP, Keith J, Tropez-Sims S, Price W, Allison-Ottey SD. Lactose intolerance and health disparities among African Americans and Hispanic Americans: an updated consensus statement. J Natl Med Assoc. 2013;105(2):112–27. [DOI] [PubMed] [Google Scholar]
- 36. Educational attainment in the United States. March 1998 (update). Current population reports. p.20–543. U.S. Department of Commerce [Internet]. Available from:. https://www.census.gov/data/tables/1998/demo/educational-attainment/p20-513.html. (Accessed 24 April, 2020).
- 37. Krishnan S, Cozier YC, Rosenberg L, Palmer JR. Socioeconomic status and incidence of type 2 diabetes: results from the Black Women's Health Study. Am J Epidemiol. 2010;171(5):564–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Krishnan S, Rosenberg L, Djousse L, Cupples LA, Palmer JR. Overall and central obesity and risk of type 2 diabetes in U.S. black women. Obesity. 2007;15(7):1860–6. [DOI] [PubMed] [Google Scholar]
- 39. Krishnan S, Rosenberg L, Palmer JR. Physical activity and television watching in relation to risk of type 2 diabetes: the Black Women's Health Study. Am J Epidemiol. 2009;169(4):428–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Krishnan S, Rosenberg L, Singer M, Hu FB, Djousse L, Cupples LA, Palmer JR. Glycemic index, glycemic load, and cereal fiber intake and risk of type 2 diabetes in US black women. Arch Intern Med. 2007;167(21):2304–9. [DOI] [PubMed] [Google Scholar]
- 41. Vimalananda VG, Palmer JR, Gerlovin H, Wise LA, Rosenzweig JL, Rosenberg L, Ruiz Narvaez EA. Night-shift work and incident diabetes among African-American women. Diabetologia. 2015;58(4):699–706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Bethea TN, Palmer JR, Rosenberg L, Cozier YC. Neighborhood socioeconomic status in relation to all-cause, cancer, and cardiovascular mortality in the Black Women's Health Study. Ethn Dis. 2016;26(2):157–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Cozier Y, Albert M, Castro-Webb N, Coogan P, Ridker P, Kaufman H, Palmer J, Rosenberg L. Neighborhood socioeconomic status in relation to serum biomarkers in the Black Women's Health Study. J Urban Health. 2016;93(2):279–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.