A Prospective Study of Red Meat Consumption and Type 2 Diabetes in Middle-Aged and Elderly Women: The Women’s Health Study

The WHS is a randomized, double-blind, placebo-controlled trial designed to evaluate the balance of benefits and risks of low-dose aspirin and vitamin E in the primary prevention of cardiovascular disease and cancer (10). We randomized a total of 39,876 female health professionals aged ≥45 years who were free of coronary heart disease, stroke, and cancer (other than nonmelanoma skin cancer). Of them, 98% provided detailed information about their diet, completing a 131-item semiquantitative food frequency questionnaire (SFFQ) in 1993 (10). We excluded individuals with >70 items left blank in their SFFQ and with energy intake outside the range of 2,514 kJ (600 kcal) and 14,665 kJ (3,500 kcal), with reported diabetes at baseline, and with completed data on meat consumption, which left 37,309 women for the analysis. This study has been conducted according to the ethical guidelines of Brigham and Women’s Hospital. Written informed consent was obtained from all participants.

For each food, a commonly used unit or portion size was specified, and each participant was asked how often she had consumed that amount, on average, over the previous year. Nine possible responses ranging from “never” to “six or more times per day” were recorded. Intake of each meat item was calculated by multiplying the intake frequency of individual items in those food categories by their weights, estimated from the specified portion size. Nutrient intakes were computed by multiplying the frequency of consumption of each unit of food from the SFFQ by the nutrient content of the specified portion size according to food composition tables from the Harvard Food Composition Database (11). Each nutrient was adjusted for total energy using the residual method (12). Red meat intake was considered to be the sum of hamburger, beef, or lamb as a main dish, pork as a main dish, beef, pork, or lamb as a sandwich or mixed dish, and all processed meat. Total processed meat was considered to be the sum of hot dogs, bacon, and other processed meat. Other processed meat included sausage, salami, and bologna. This SFFQ has demonstrated reasonably good validity as a measure of long-term average dietary intakes in women (13). The correlations between the SFFQ and two 1-week diet records were 0.48–0.68 for total and specific types of fat intakes, 0.35–0.45 for meat, and 0.33–0.53 for processed meat (13,14).

The status of type 2 diabetes was evaluated at baseline, and women with a history of diagnosed diabetes were excluded. Thereafter, all of the participants were asked annually whether and when they had been diagnosed with diabetes since completing the previous questionnaire. Two complementary approaches have been used to confirm self-reported type 2 diabetes in the WHS. First, we attempted to contact 473 women with self-reported diabetes who provided a blood sample as part of a nested case-control study of diabetes, to verify the self-reported diagnosis (15). Using the American Diabetes Association (ADA) diagnostic criteria (16), the self-reported diagnosis of diabetes was confirmed in 406 (91%) of 446 women who responded via telephone interview (15). Second, a random sample of 147 women with self-reported diabetes was mailed a supplemental diabetes questionnaire, also using the ADA criteria to parallel the telephone interview. Among 136 respondents, 124 (91%) women were classified as having type 2 diabetes by the supplemental questionnaire. In addition, 113 of the 124 women gave permission to contact their primary care physician. Ninety-seven of the 113 physicians responded, of whom 90 provided adequate information to apply the ADA criteria. For these 90 women, 89 (99%) were confirmed to have type 2 diabetes on the basis of the combined information from the supplemental questionnaire and physician information. Thus, we believe that self-reported type 2 diabetes is valid in the WHS.

We first categorized intakes of red and processed meats into quintiles of intake. We also categorized meat consumption by aggregating nine possible responses from SFFQ into four categories for red meat and total processed meat (<1/week, 1/week, 2–4/week, ≥5/week). Subtype of red meat and processed meat were categorized into three categories (<1/week, 1/week, ≥2/week). We calculated the incidence rates of type 2 diabetes for each category of baseline meat intake by dividing the number of incident cases by the person-years of follow-up from 1993 to 2003. After testing the proportional hazard assumption, we used Cox proportional hazards models to estimate the rate ratios (described as relative risks [RRs]) and 95% CIs of developing type 2 diabetes for each category of meat item compared with the lowest category. The initial model was adjusted for age and total energy intake. In multivariate models, we adjusted for age (continuous), BMI (continuous), total energy intake (quintiles), smoking status (current, past, and never), exercise (rarely/never, <1/week, 1–3/week, and ≥4/week), alcohol intake (rarely/never, 1–3 drinks/month, 1–6 drinks/week, ≥1 drink/day), and family history of diabetes (yes/no) and then further adjusted for dietary factors, including intakes of dietary fiber (quintiles), glycemic load (quintiles), total fat (quintiles), and magnesium (quintiles). Tests of linear trend across increasing categories of intake were conducted by assigning the medians of intakes in quintiles treated as a continuous variable. A likelihood ratio test was used to assess the significance of interaction terms. We also did separate analyses for intakes of type of fat, protein, dietary heme iron, and total iron. All statistical analyses were conducted using SAS (version 8.0; SAS Institute, Cary, NC).

At baseline in 1993, the differences comparing the highest intake to the lowest intake among 37,309 women were ∼10-fold in total red meat intake (Table 1). High consumers of red meat tended to be current smokers and less likely to exercise, use postmenopausal hormones, or use supplements of multivitamins than those who rarely consumed red meat. Women with higher intake of red meat also had a higher BMI and were more likely to have a history of hypertension and a family history of diabetes. Furthermore, total red meat intake was positively associated with total energy intake, intakes of all fatty acids, cholesterol, and protein but inversely associated with dietary carbohydrate, fiber, magnesium intakes, and glycemic load. Similar associations were observed for total meat and processed meat.

During an average of 8.8 years of follow-up (326,876 person-years), we documented 1,558 incident cases of type 2 diabetes. In the age- and energy-adjusted models, total red meat and processed meat were significantly associated with an increased risk of type 2 diabetes (Table 2). After adjusting for age, total energy intake, BMI, smoking, alcohol intake, physical activity, and family history of diabetes, these associations were attenuated but remained statistically significant for red meat and processed meat (RRs comparing two extreme quintiles 1.28, 95% CI 1.07–1.53, P for trend < 0.001 for red meat and 1.23, 1.05–1.45, P for trend = 0.001 for processed meat). Additional adjustment for dietary fiber, glycemic load, magnesium intake, and total fat intake did not change these significant results. As compared with the lowest quintile of intake, the multivariate-adjusted RRs of type 2 diabetes across increasing quintiles were 1.00, 0.90, 1.07, 1.08, and 1.24 for red meat intake (P = 0.005 for trend) and 1.00, 0.96, 0.95, 1.08, and 1.19 for total processed meat intake (P for trend = 0.007).

The multivariate-adjusted RRs when frequency of red meat and processed meat consumption were modeled as four categories (<1/week, 1/week, 2–4/week, and ≥5/week) were 1.00, 1.15, 1.08, and 1.29, respectively (P = 0.01 for trend), for red meat and 1.00, 1.05, 1.20, and 1.43 for total processed meat (P < 0.001 for trend). After further adjustment for nutrient factors, only total processed meat intake remained significantly associated with diabetic risk. As major items of processed meat, bacon and hot dogs also showed consistently positive associations with risk of type 2 diabetes (Table 3). Comparing women who consumed ≥2/week to those who consumed <1/week, the multivariate-adjusted RRs of type 2 diabetes after controlling for dietary nutrients were 1.17 (95% CI 1.02–1.35, P = 0.02 for trend) for bacon and 1.24 (1.05–1.45, P = 0.02 for trend) for hot dogs. Subtypes of red meat appeared to increase diabetes risk in the multivariate models, but 95% CIs were wider and included the null.

Finally, we found significantly positive associations for dietary intakes of cholesterol, animal protein, and heme iron, which are the primary constituents of red meat (Table 4). The multivariate-adjusted RRs comparing two extreme quintiles were 1.35 for cholesterol (95% CI 1.11–1.64, P = 0.003 for trend), 1.44 for animal protein (1.16–1.78, P = 0.001 for trend), and 1.46 for heme iron (1.20–1.78, P < 0.0001 for trend).

In this large prospective study, we found that a higher consumption of red meat, especially total processed meat, was associated with an increased risk of developing type 2 diabetes in middle-aged and older U.S. women, independent of known diabetes risk factors.

The prospective design and high follow-up rates in our study minimized the possibility of selection bias or recall bias. As any measurement errors from baseline SFFQ were unlikely to be related to subsequent diabetes end points, misclassification of dietary assessment would most likely be nondifferential and would have attenuated the true associations. Thus, our observed associations may not be explained by such errors and were, on the contrary, somewhat conservative. Our results were also less likely to be influenced by misclassification of type 2 diabetes because of a high accuracy of self-reported diabetes as assessed by our validation study. Moreover, because we adjusted for age, BMI, and other important risk factors for type 2 diabetes, the residual confounding due to these known factors was probably modest.

Our findings are consistent with recent evidence from both the Health Professionals Follow-Up Study of male health professionals and the Nurses’ Health Study II of younger and middle-aged U.S. women (8,9). Processed meat appeared entirely responsible for the elevated diabetes risk associated with total red meat in these two large cohort studies (8,9).

Red meat, especially processed meat, contains certain types of preservatives, additives, or other chemicals arising from meat preparation, including preservation, packaging, and cooking. These compounds include nitrates and nitrites added in meat processing as well as a variety of heterocyclic amines and polycyclic aromatic hydrocarbons formed in red meat, especially when cooked well done (17). These compounds can be converted to N-nitrosamines (17), which were found to be toxic to pancreatic β-cell (18). Consumption of foods with a high content of nitrites and nitrosamines has been associated with type 1 diabetes (19–21). Also, advanced glycation and lipoxidation end products produced during the cooking or processing of meat have been associated with insulin resistance and diabetes-related complications in animal models (22) and human subjects (23,24). Therefore, such specific compounds mainly present in processed meat might largely explain the observed significant association between processed meat intake and type 2 diabetes.

Nevertheless, red meat is also a major source for saturated fat, cholesterol, animal protein, and heme iron. It has been shown that certain types of fat from red meat may play a major role in the development of type 2 diabetes (9,25). As noted in a recent review, epidemiological evidence for the relevance of dietary fats and risk of type 2 diabetes seem to be inconsistent (25). The present study did not show any positive associations between intakes of saturated fat or trans fatty acid and risk of type 2 diabetes. Likewise, we found no evidence of decreased risk of diabetes with increased intake of monounsaturated or polyunsaturated fatty acids or vegetable fat. However, cholesterol intake tended to be positively related to an elevated risk of diabetes. Cholesterol intake from red meat may thus explain, at least in part, the observed association between red meat intake and type 2 diabetes.

Recently, body iron overload has been postulated to promote insulin resistance and increase type 2 diabetes risk (26–30). However, the positive association between heme iron intake and type 2 diabetes should be interpreted with caution because the high correlation between red meat intake and heme iron intake substantially limited our statistical capability to separate the independent effect of heme iron from other components of red meat. The lack of association between total iron intake and risk of type 2 diabetes might reflect the fact that body iron status is not well regulated by intakes of total dietary iron or heme iron.

Our study has several limitations. First, we cannot completely exclude the possibilities of residual confounding from unmeasured or incompletely measured underlying lifestyle factors even though we have adjusted for many major risk factors for type 2 diabetes. Second, participants might change their diets after developing some diseases. However, these associations persisted when we carried out secondary analyses after excluding participants who had a history of hypertension or high cholesterol levels, which allows for elimination of dietary change related to these diagnoses. Third, because of a high degree of statistical collinearity, our ability to reliably distinguish the effect of red meat from intakes of its major components such as animal fat, animal protein, and heme iron was limited. Fourth, limited variation of intakes for each subtype of total red meat or other processed meat in our cohort could lead to insufficient statistical power to detect significant association. Finally, we were also unable to assess levels of specific chemicals added or produced in different food preparation methods and thus could not address the relationship between these specific chemicals and diabetic risk.

In conclusion, our study indicates that higher consumption of total red meat, especially various processed meats, may increase risk of developing type 2 diabetes in women. However, the underlying mechanisms by which consumption of red meat or processed meat influence type 2 diabetes risk are still not well understood and require further investigation.

This study was supported by the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-02767.

We are indebted to the 39,876 dedicated and committed participants of the WHS.