Abstract
Background and Purpose
Black Americans and residents of the Southeastern United States, are at increased risk of stroke. Diet is one of many potential factors proposed that might explain these racial and regional disparities.
Methods
Between 2003–2007, the REasons for Geographic and Racial Differences in Stroke (REGARDS) cohort study enrolled 30,239 black and white Americans aged 45 years or older. Dietary patterns were derived using factor analysis and foods from food frequency data. Incident strokes were adjudicated using medical records by a team of physicians. Cox proportional hazards models were used to examine risk of stroke.
Results
Over 5.7 years, 490 incident strokes were observed. In a multivariable-adjusted analysis, greater adherence to the Plant-based pattern was associated with lower stroke risk (HR=0.71; 95% CI=0.56–0.91; ptrend=0.005). This association was attenuated after addition of income, education, total energy intake, smoking, and sedentary behavior. Participants with a higher adherence to the Southern pattern experienced a 39% increased risk of stroke (HR=1.39; 95% CI=1.05, 1.84), with a significant (p = 0.009) trend across quartiles. Including Southern pattern in the model mediated the black-white risk of stroke by 63%.
Conclusions
These data suggest that adherence to a Southern style diet may increase the risk of stroke while adherence to a more plant-based diet may reduce stroke risk. Given the consistency of finding a dietary impact on stroke risk across studies, discussing nutrition patterns during risk screening may be an important step in reducing stroke.
Keywords: stroke, race, region, dietary patterns
Background
The risk of having and dying from a stroke is higher in the Southeastern United States (US) than in other regions of the country1–3. Many causes for this regional disparity in stroke have been suggested, including demographic factors, socio-economic status, and lifestyle differences between regions3–8. Among the most commonly proposed causes is a geographic difference in diet, potentially differentially influencing stroke risk through a multi-faceted impact on stroke risk factors including obesity, hypertension, diabetes, oxidative stress, and inflammation9–12. Because traditional risk factors like hypertension do not fully explain geographic variability in stroke risk, behavioral factors like diet are ideal targets to further explore13, 14. Indeed, while dietary differences are frequently hypothesized to be the cause of this disparity, very few prospective data are available to support the existence of regional differences in diet or whether potential differences contribute to the excess risk of stroke in the Southeastern US. Using dietary patterns to measure dietary exposures has become popular since foods and nutrients are not eaten in isolation15. Two approaches are commonly used: a priori defined dietary scores like the Healthy Eating Index or Mediterranean diet score16; and a posteriori identified patterns using factor or cluster analysis17. A data-driven method, factor analysis, measures eating patterns in specific populations without making judgments about which foods are commonly consumed together or which foods provide health benefit. Employing such empirically defined dietary patterns may help identify social or cultural influences that are more nuanced when it comes to food selection in a specific population and provides a novel method for examining diet in a heterogeneous population4.
An additional challenge in examining regional differences on the impact of diet on stroke risk is the potential confounding of race on regional differences. Black Americans have a greater risk of stroke than their white counterparts and also represent a larger proportion of the population in the Southeast compared to other regions in the US18 hence, it is possible that geographic disparities are confounded with racial disparities.19 As such a large, geographically diverse population sample is needed to be able to first identify and confirm the dietary patterns, and then to test whether any potential association of diet with stroke risk is homogeneous across both race and region. The REasons for Geographic and Racial Differences in Stroke (REGARDS) study provides an ideal population for examining this association in US blacks and whites.
Methods
Study Participants
The REGARDS study is specifically designed to examine racial and regional differences in stroke in the lower 48 continental states, and as such oversampled both black Americans and persons residing in the Southeast, an area of the country known as the stroke belt (includes Louisiana, Arkansas, Mississippi, Alabama, Tennessee, Georgia, North Carolina, and South Carolina)20. A total of 30,239 participants were recruited for the study. The cohort comprised 56% from the stroke belt and 42% black, 55% women21. Participants were recruited using commercially available lists from Genesys, Inc. (City, State), which is the same list used by the Behavioral Risk Factor Surveillance Systems (BRFSS) in the United States20. REGARDS participants were initially contacted through a personalized mailing and brochure to describe the study and inform them of an upcoming phone call. REGARDS staff then conducted a 45-minute telephone interview to obtain verbal consent and to collect data on demographics, socio-economic status, stroke risk factor characterization, and medical history. The telephone response rate was 33% and cooperation rate was 49%, similar to other cohort studies22, 23. Following the telephone call, a trained health professional went to the participant’s home to obtain written consent and collect blood and urine specimens. During this visit an electrocardiogram was performed and blood pressure, waist circumference, height and weight were measured. The study was approved by the Institutional Review Board at all participating universities and written informed consent was obtained from all participants.
Stroke Ascertainment
REGARDS participants are contacted via telephone every six months to ascertain vital status and obtain information on reasons for hospitalization including stroke, transient ischemic attack (TIA), and stroke symptoms. Medical records are pursued if the participant reports seeking medical care for stroke or TIA and/or was hospitalized for stroke symptoms or unknown reason. Once a medical record is received, it is reviewed by a committee of stroke neurologists to verify that a stroke occurred. Stroke is defined: 1) according to the World Health Organization definition of focal neurologic deficit lasting greater than 24 hours; or 2) nonfocal neurologic symptoms with imaging consistent of stroke21. The National Death Index is queried annually to identify stroke deaths that might not have been hospitalized. For this analysis, strokes occurring through September 1, 2011 were included.
Dietary Assessment
At the conclusion of the in-home visit, the health professional left the participant with self-administered forms including the Block98 Food Frequency Questionnaire (FFQ). The Block 98 version developed by Block Dietary Data Systems [Berkeley, CA] and distributed by NutritionQuest is an 8-page paper-and-pencil form with more than 150 multiple-choice questions based on 107 food items that can be completed in about 30–40 minutes. The Block98 allows for identification of higher and lower fat versions of food and added sugars in the diet and has been validated in populations similar to REGARDS24, 25. Participants were asked to recall usual dietary intake over the past year. The participant completed the form and mailed it back to the operations center within three months of the in-home visit. NutritionQuest processed the forms to assess the mean daily intake (grams) of each individual food on the FFQ.
Dietary Pattern Identification
Pattern identification was based upon the estimated quantity of the 107 individual food items. Prior to analysis, we combined some FFQ items based on similarity of food content. For example, “beverages containing some juice like Hi-C” was grouped with “sugar-sweetened beverages” due to nutritional content. Some individual food items were separated into two items due to hypothesized cardiovascular impact, such as high- and low-fat milk. Fried items for fish, potatoes and chicken were separated from the non-fried items due to the high fat content of fried foods and anticipated differences of use across race and regionally-defined populations. Some items like “Chinese food” were left as a standalone food group due to the uniqueness of the item26. Five factors were retained from principal components analysis (PCA) based on the eigenvalue (scree plot) and the solution providing the optimal congruence across region, gender, and race. Final factor loadings were derived in the full population using PCA with varimax rotation. Descriptive names were assigned to the patterns based on foods that contributed most highly to each pattern: Convenience, Plant-based, Sweets/fats, Southern, Alcohol/salads (Appendix 1). A score was given to each participant reflecting their “adherence” to a specific pattern (higher scores reflecting closer adherence to the eating pattern).
Appendix 1.
Final factor loadings derived in the entire REGARDS population (showing only those with absolute value > 0.15 for simplicity)
| Convenience | Healthy | Sweets/Fats | Southern | Alcohol/Salads | |
|---|---|---|---|---|---|
| 100% fruit juice | 0.25 | 0.17 | −0.17 | ||
| Added fats | 0.40 | 0.38 | 0.25 | ||
| Beans | 0.36 | 0.38 | |||
| Beer | −0.16 | 0.23 | |||
| Bread | 0.47 | 0.37 | |||
| Bread - Whole Grain | 0.30 | 0.18 | |||
| Butter | 0.17 | 0.32 | |||
| Candy | 0.40 | ||||
| Cereal | 0.38 | −0.20 | |||
| Cereal - High Fiber | 0.24 | −0.25 | |||
| Chinese food | 0.44 | ||||
| Chocolate | 0.46 | ||||
| Coffee | 0.22 | −0.16 | 0.30 | ||
| Condiments | 0.25 | 0.31 | 0.29 | ||
| Desserts | 0.20 | 0.53 | −0.17 | ||
| Eggs and egg dishes | 0.42 | 0.29 | |||
| Fish | 0.27 | 0.38 | 0.21 | ||
| Fried food | 0.24 | 0.56 | |||
| Fried potatoes | 0.37 | 0.28 | 0.16 | ||
| Fruit | 0.58 | ||||
| Highfat dairy | 0.18 | 0.37 | 0.21 | ||
| Liquor | 0.31 | ||||
| Lowfat dairy | 0.20 | −0.19 | |||
| Margarine | 0.37 | ||||
| Mexican dishes | 0.48 | ||||
| Milk alternatives | 0.18 | ||||
| Milk - Highfat | 0.18 | 0.24 | |||
| Milk - Lowfat | 0.16 | −0.42 | |||
| Miscellaneous sugar | 0.54 | 0.19 | |||
| Mixed dishes with meat | 0.61 | ||||
| Organ meat | 0.17 | 0.47 | |||
| Pasta dishes | 0.59 | 0.17 | |||
| Pizza | 0.45 | −0.18 | 0.20 | ||
| Potatoes | 0.36 | 0.26 | |||
| Poultry | 0.29 | 0.31 | |||
| Processed meats | 0.25 | 0.26 | 0.45 | 0.22 | |
| Red meat | 0.45 | 0.18 | 0.26 | 0.26 | |
| Refined grains | 0.31 | 0.17 | 0.20 | 0.20 | |
| Salad dressing/sauces | 0.30 | 0.55 | |||
| Salty snacks | 0.32 | 0.30 | |||
| Seeds, nuts | 0.26 | 0.19 | 0.19 | ||
| Shell fish | 0.28 | 0.23 | 0.24 | ||
| Soda | −0.23 | 0.24 | |||
| Soup | 0.44 | 0.32 | |||
| Sugar-sweetened beverages | 0.37 | −0.15 | |||
| Sweet breakfast foods | 0.19 | 0.39 | |||
| Tea | 0.31 | ||||
| Vegetable - cruciferous | 0.59 | ||||
| Vegetable - dark yellow | 0.41 | −0.17 | |||
| Vegetable - green leafy | 0.16 | 0.49 | −0.22 | 0.48 | |
| Vegetable - other | 0.48 | ||||
| Vegetable - tomato | 0.32 | 0.27 | |||
| Vegetable mixed dishes | 0.35 | 0.31 | −0.25 | ||
| Water | 0.32 | ||||
| Wine | 0.36 | ||||
| Yogurt | 0.31 | −0.25 |
Statistical Analyses
The five dietary patterns were grouped into quartiles in order to examine the association of diet with stroke. Descriptive statistics (Chi-squared, ANOVA, and log rank) were used first to examine the association of each pattern with demographic, socio-economic, behavioral and nutrient variables. Cox-proportional hazards analysis was used to examine the risk of stroke associated with each dietary pattern and all participants were censored at date of stroke, date of withdrawal from the study, date of death, or September 1, 2011. Schoenfeld residuals were examined to ensure proportional hazards assumptions were met. Models were built sequentially by first adding demographic variables (age, race, gender, age-race interaction). An age-race interaction term was found in previous work to be necessary as there is a large difference in stroke incidence between blacks and whites at younger ages that does not exist in older ages21. Following the addition of demographic variables, income and education were then added to the model followed by total energy intake. The final set of models included smoking history and television viewing (defined as watching more than 4 hours of TV per day). The proportion of the racial and regional disparities in stroke risk attributable to adherence to specific diets was assessed using mediation analyses, where the beta coefficient changes in the Cox models for both race and region were assessed before and after adjustment for diet adherence. Women and white participants were more likely to return the FFQ as were those who graduated college, therefore we performed a sensitivity analysis to examine stroke risk factors in the nutrition sub-sample and the full REGARDS cohort.
Results
After excluding participants with a self-report of stroke (n=2032), 28,151 were available for the present analysis. Food frequency questionnaires were available on 20,251 (72%). After a median of 5.7 years of follow-up, 490 strokes were observed (12% were hemorrhagic (n=59), 82% were ischemic(n=402) and 6% were not classified). Participants who reported a stroke were more likely to be over age 65, have a history of hypertension, diabetes, heart disease and atrial fibrillation and were less likely to be obese than the general REGARDS dietary cohort.
Pattern adherence (defined by greater proportion of participants in the fourth quartile compared to the first) to the Convenience dietary pattern was more common in younger, male, white participants, and those residing outside the Southeast (Table 1). In contrast, adherence to the Plant-based pattern was more common in older, female, black participants and in those who graduated college. Income was not associated with the Plant-based dietary pattern but income less than $20,000 per year was associated with decreased adherence to the Convenience pattern. Those who were higher consumers of the Sweets/Fats and Southern patterns were more likely to reside in the Southeast but the Sweets/Fats pattern was more likely to be consumed by white participants, while the Southern pattern was more likely to be consumed by black participants.
Table 1.
Demographic characteristics by quartile of dietary pattern in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study
| Q1 n (%) |
Q2 n (%) |
Q3 n (%) |
Q4 n (%) |
||
|---|---|---|---|---|---|
| Convenience | Age <65 | 2016 (40) | 2452 (48) | 2731 (53) | 3183 (63) |
| Black | 2255 (45) | 1879 (37) | 1402 (27) | 1147 (23) | |
| Male | 1975 (39) | 2416 (47) | 2764 (54) | 3068 (62) | |
| Residence in the Southeast | 3068 (62) | 2925 (58) | 2828 (55) | 2589 (51) | |
| College graduate | 1567 (31) | 1865 (37) | 2080 (41) | 2200 (43) | |
| Income < $20,000 | 1003 (20) | 799 (16) | 641 (13) | 621 (12) | |
| Current Smoker | 658 (13) | 678 (13) | 653 (13) | 726 (14) | |
| Obese (BMI≥30kg/m2) | 1774 (36) | 1813 (36) | 1774 (35) | 1978 (39) | |
| Sedentary(> 4 hours of TV/day) | 635 (13) | 629 (12) | 552 (11) | 655 (13) | |
| Plant Based | Age <65 | 3105 (61) | 2530 (50) | 2363 (46) | 2384 (47) |
| Black | 1374 (27) | 1687 (34) | 1747 (34) | 1875 (37) | |
| Male | 2688 (53) | 2257 (45) | 2080 (41) | 1853 (37) | |
| Residence in the Southeast | 2829 (56) | 2922 (58) | 2842 (56) | 2817 (55) | |
| College graduate | 1616 (32) | 1792 (36) | 2048 (40) | 2256 (44) | |
| Income < $20,000 | 768 (15) | 766 (15) | 779 (15) | 751 (15) | |
| Current Smoker | 1163 (23) | 683 (14) | 492 (10) | 377 (7) | |
| Obese (BMI≥30kg/m2) | 1874 (37) | 1831 (36) | 1835 (36) | 1799 (35) | |
| Sedentary (> 4 hours of TV/day) | 842 (17) | 660 (13) | 540 (11) | 429 (8) | |
| Sweets/Fats | Age <65 | 2732 (54) | 2458 (49) | 2522 (50) | 2670 (53) |
| Black | 2256 (44) | 1640 (33) | 1401 (28) | 1386 (27) | |
| Male | 1985 (39) | 2280 (45) | 2709 (54) | 2490 (48) | |
| Residence in the Southeast | 2704 (53) | 2787 (55) | 2844 (56) | 3075 (61) | |
| College graduate | 2020 (40) | 2057 (41) | 1959 (39) | 1676 (33) | |
| Income < $20,000 | 797 (16) | 714 (14) | 709 (14) | 844 (17) | |
| Current Smoker | 623 (12) | 588 (12) | 659 (13) | 845 (17) | |
| Obese (BMI≥30kg/m2) | 1917 (38) | 1803 (36) | 1790 (35) | 1829 (36) | |
| Sedentary (> 4 hours of TV/day) | 549 (11) | 546 (11) | 571 (11) | 805 (16) | |
| Southern | Age <65 | 2601 (50) | 2527 (50) | 2465 (49) | 2789 (56) |
| Black | 483 (9) | 1217 (24) | 2018 (40) | 2965 (60) | |
| Male | 1904 (37) | 1985 (39) | 2280 (45) | 2709 (54) | |
| Residence in the Southeast | 2490 (48) | 2752 (54) | 2971 (59) | 3197 (64) | |
| College graduate | 2621 (51) | 2138 (42) | 1720 (34) | 1233 (25) | |
| Income < $20,000 | 424 (8) | 617 (12) | 858 (17) | 1165 (23) | |
| Current Smoker | 411 (8) | 548 (11) | 736 (15) | 1020 (20) | |
| Obese (BMI≥30kg/m2) | 1362 (26) | 1690 (33) | 1937 (39) | 2350 (47) | |
| Sedentary (> 4 hours of TV/day) | 414 (8) | 541 (11) | 665 (13) | 851 (17) | |
| Alcohol/Salads | Age <65 | 2240 (45) | 2507 (50) | 2677 (52) | 2958 (57) |
| Black | 2456 (50) | 1904 (38) | 1338 (26) | 985 (19) | |
| Male | 1777 (36) | 2006 (40) | 2404 (47) | 2691 (52) | |
| Residence in the Southeast | 2903 (59) | 2976 (59) | 2887 (56) | 2644 (51) | |
| College graduate | 1417 (29) | 1704 (34) | 2051 (40) | 2540 (49) | |
| Income < $20,000 | 1228 (25) | 822 (16) | 603 (12) | 411 (8) | |
| Current Smoker | 545 (11) | 652 (13) | 748 (15) | 770 (15) | |
| Obese (BMI≥30kg/m2) | 1841 (37) | 1856 (37) | 1811 (35) | 1831 (35) | |
| Sedentary (> 4 hours of TV/day) | 720 (15) | 670 (13) | 555 (11) | 526 (10) |
For all survival analyses, the first quartile (lowest adherence to specific diet pattern) was used as the reference group. In age-, race-, region-, sex and age-race interaction-adjusted models, greater adherence to the Plant-based and Sweets/Fats dietary patterns were associated with a reduction in stroke risk, while greater adherence to the Southern pattern was associated with an increased risk of stroke (Table 2). The Southern pattern was associated with a 39% increased risk of stroke (HR=1.39; 95% CI=1.05, 1.84) when comparing the highest quartile to the lowest quartile, (p for trend=0.009). This association was attenuated with the addition of covariates but the direction of the association remained significant (p=0.05). In the Sweets/Fats pattern, the risk of stroke was lower in those with the greatest adherence than in those with the lowest adherence (HR=0.80; 95% CI=0.61, 1.04) and this association was strengthened and became significant with the adjustment for covariates. The risk of stroke was lowest in the Plant-based pattern (HR=0.71; 95% CI=0.56–0.91) (p for trend=0.005). This association was attenuated and no longer significant after the addition of income, education, total energy intake, smoking, and sedentary behavior but the direction of association persisted. Neither the Convenience pattern nor the Alcohol/Salads pattern was associated with risk of stroke (p > 0.05).
Table 2.
Risk of stroke by quartile of dietary pattern in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study
| Q1 | Q2 | Q3 | Q4 | p for trend | ||
|---|---|---|---|---|---|---|
| Convenience | n strokes/n total | 144/4981 | 127/5082 | 116/5106 | 103/5082 | |
| Model 1 | ref | 0.97 (0.76, 1.24) | 0.93 (0.72, 1.19) | 0.98 (0.75, 1.27) | 0.75 | |
| Model 2 | ref | 0.98 (0.77, 1.25) | 0.95 (0.74, 1.22) | 0.99 (0.76, 1.30) | 0.67 | |
| Model 3 | ref | 0.99 (0.78, 1.26) | 0.97 (0.75, 1.25) | 1.08 (0.81, 1.45) | 0.72 | |
| Model 4 | ref | 0.97 (0.76, 1.24) | 0.96 (0.75, 1.25) | 1.09 (0.82, 1.46) | 0.69 | |
| 0.97 (0.76, 1.24) | 0.93 (0.72, 1.19) | 0.98 (0.75, 1.27) | ||||
| Plant Based | n strokes/n total | 135/5056 | 122/5033 | 111/5086 | 122/5076 | |
| Model 1 | ref | 0.74 (0.58, 0.95) | 0.65 (0.50, 0.83) | 0.71 (0.56, 0.91) | 0.005 | |
| Model 2 | ref | 0.76 (0.59, 0.97) | 0.67 (0.52, 0.87) | 0.75 (0.58, 0.96) | 0.02 | |
| Model 3 | ref | 0.76 (0.60, 0.98) | 0.68 (0.52, 0.88) | 0.78 (0.59, 1.02) | 0.04 | |
| Model 4 | ref | 0.80 (0.62, 1.02) | 0.74 (0.57, 0.96) | 0.85 (0.65, 1.12) | 0.20 | |
| Sweets/Fats | n strokes/n total | 125/5076 | 136/5043 | 126/5065 | 103/5067 | |
| Model 1 | ref | 1.01 (0.79, 1.30) | 0.93 (0.73, 1.20) | 0.80 (0.61, 1.04) | 0.03 | |
| Model 2 | ref | 1.00 (0.78, 1.28) | 0.91 (0.71, 1.17) | 0.75 (0.58, 0.98) | 0.07 | |
| Model 3 | ref | 1.00 (0.78, 1.27) | 0.90 (0.70, 1.18) | 0.74 (0.54, 1.02) | 0.05 | |
| Model 4 | ref | 0.99 (0.78, 1.27) | 0.88 (0.68, 1.14) | 0.72 (0.53, 0.99) | 0.04 | |
| Southern | n strokes/n total | 109/5156 | 107/5101 | 136/5017 | 138/4977 | |
| Model 1 | ref | 0.98 (0.75, 1.28) | 1.22 (0.94, 1.59) | 1.39 (1.05, 1.84) | 0.009 | |
| Model 2 | ref | 0.96 (0.73, 1.25) | 1.15 (0.89, 1.51) | 1.27 (0.96, 1.70) | 0.05 | |
| Model 3 | ref | 0.93 (0.71, 1.22) | 1.15 (0.88, 1.50) | 1.37 (1.02, 1.84) | 0.04 | |
| Model 4 | ref | 0.93 (0.71, 1.22) | 1.12 (0.86, 1.47) | 1.30 (0.97, 1.76) | 0.05 | |
| Alcohol/Salads | n strokes/n total | 142/4943 | 121/5019 | 122/5119 | 105/5170 | |
| Model 1 | ref | 0.91 (0.72, 1.17) | 0.96 (0.75, 1.24) | 0.86 (0.66, 1.12) | 0.37 | |
| Model 2 | ref | 0.93 (0.73, 1.19) | 1.00 (0.78, 1.29) | 0.94 (0.72, 1.23) | 0.79 | |
| Model 3 | ref | 0.93 (0.73, 1.19) | 1.01 (0.79, 1.30) | 0.97 (0.74, 1.27) | 0.96 | |
| Model 4 | ref | 0.91 (0.71, 1.17) | 0.97 (0.75, 1.24) | 0.93 (0.71, 1.23) | 0.71 | |
Model 1 adjusts for age, race, region, sex, and age-race interaction
Model 2 adds income and education to the above
Model 3 adds total energy to the above
Model 4 adds smoking and sedentary behavior (watching > 4 hours of television per day) to the above
As a primary aim for the parent study is to identify causes for disparities in stroke, the effect of diet on racial and regional differences in stroke was also considered. Since the Southern dietary pattern was the only pattern associated with increased risk of stroke, mediation analyses were only done for this pattern. Examining the change in the beta coefficients for race, the Southern dietary pattern was associated with a 63% reduction in the association of black race with stroke risk in participants under the age of 65 where the racial disparity is most prevalent. Southern dietary pattern did not mediate regional disparities in stroke.
Discussion
Our findings suggest that racial and regional differences in diet do exist, and that these differences may contribute to disparities in stroke risk for black Americans. We did not find evidence that regional differences in diet were associated with increased stroke risk; however, adherence to a Southern style diet is more common both among residents of the Stroke Belt and among black participants in the study. This eating pattern is also associated with increased risk of stroke, suggesting that it is a potential contributor to the racial but not regional disparities in stroke. We have further shown that a Southern dietary pattern mediates the racial disparity in stroke by 63%. Previous work by our group demonstrated that traditional risk factors such as hypertension and atrial fibrillation mediated only 50% of the excess stroke in black Americans14 which further supports the important role for diet in determining stroke risk.
Using factor analysis to derive dietary patterns is a well-accepted method for studying the relationship between diet and cardiovascular disease risk17, 27, 28. Yet most studies have considered stroke as a composite outcome with other cardiovascular events thus not allowing identification of dietary effects that may be unique to stroke but not heart disease. Only one study (Nurses’ Health Study) has used factor analysis to examine dietary patterns and stroke risk29. Even though the previous study was limited to women, similar results were observed29. Our study extends the findings from the Nurses Health Study to a diverse population of black and white Americans across a wide range of income and education levels suggesting that the impact of diet on stroke risk goes beyond race and socioeconomic status. One main difference between the two studies was in the number of dietary patterns observed. REGARDS is not the first study to observe more than two dietary patterns. The Multi-Ethnic Study of Atherosclerosis (MESA) investigators also identified more than two dietary patterns and have published extensively on these patterns and cardiovascular risk factors26, 30–33. However, MESA is restricted to only four geographic regions and may not be reflective of national dietary patterns.
Although few studies have used factor analysis to examine diet and stroke, other methods have been used. A recent review found that strong evidence exists to recommend increasing potassium, fatty fish, fruits, vegetables, whole grains and limiting sodium intake for stroke prevention10. The Plant-based pattern identified in this paper as being protective in terms of stroke risk includes many of the foods (fish, fruit, vegetables) listed above. In the Northern Manhattan Study (NOMAS) examining a priori dietary patterns (Mediterranean diet score), the Mediterranean diet score was not associated with risk of stroke34 but was associated with a lower degree of white matter hyperintensity volume (a marker of small vessel damage) in the brain35. Though not identical to the Mediterranean diet score, the Plant-based pattern observed in our study was associated with reduction in stroke risk and had more than double the number of events observed in the NOMAS study. Diet quality was recently tested in the interventional study PREVIMED in which participants were randomized to a Mediterranean style diet36. Those randomized to the Mediterranean style diet had significantly lower risk of stroke than those eating the control diet suggesting that dietary change has a strong effect on stroke outcomes. The consistency across cohort studies and this randomized clinical trial is striking and suggests discussing dietary change as part of a general health screening has the potential to reduce the burden of stroke.
The Sweets/Fats pattern was associated with a reduction in stroke risk that was strengthened by the addition of covariates. This was contrary to our hypothesis given that a diet high in added sweets and saturated fats is considered to be adverse in terms of cardiovascular health 37. It is possible that adherence to a pattern such as this is associated with an increased risk of cancer or heart disease which might lead to death before a stroke could occur 38, 39. As additional data are collected, future studies of these data are needed to the mechanism behind the observed protective effect.
REGARDS is a large national sample of black and white Americans in the United States with excellent ascertainment of stroke outcomes and provides a unique opportunity to examine dietary components as stroke risk factors. Since the group used in this analysis is a sub-sample of the full REGARDS population, selection bias needs to be considered. The hazard ratios were not significantly different for known stroke risk factors in the dietary sub-sample and the full REGARDS sample. This would indicate that any bias introduced in the sample does not alter conclusions regarding risk of stroke in the dietary sub-sample. This limitation is further overcome in that there were over 20,000 participants from 48 states available for analysis.
Examining dietary patterns as the main exposure may be accounting for cultural or social practices that influence stroke risk from a young age. Previous studies have indicated that birth in the Southeast is associated with a greater risk of stroke than birth outside the Southeast4. It is possible that diet is modeled from birth and/or from practices observed during childhood. Additionally, dietary patterns associated with cultural foodways are more likely to persist in the region where they originated and become less likely to persist as individuals or groups migrate from a region40. Future studies examining the complex interactions between region of residence, migration, and dietary pattern adherence will help tease out these cultural-diet interactions.
Dietary patterns explain some of the racial disparity in stroke risk observed in the United States. Additional analyses are needed with these patterns to determine how they are associated with other outcomes like cardiovascular disease and cancer and to determine to what degree dietary patterns are inherited from childhood. Data from this and future studies would help identify culturally specific interventions in diet that would help reduce racial and geographic disparities in stroke risk.
Acknowledgments
We would like to acknowledge the coordinating center and survey research unit at the University of Alabama Birmingham for excellent work in putting together this rich dataset.
Funding Sources: This research project is supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, NIH, Department of Health and Human Services and also an American Reinvestment and Recovery Act supplement. Additional funding was received from General Mills to code the dietary data. The content is solely the responsibility of the authors and does not necessarily represent the official views and positions of the National Institute of Neurological Disorders and Stroke or the National Institutes of Health. Representatives of the funding agency were involved in the review of the manuscript prior to submission for publication. Dr. Judd acknowledges funding from the NIH loan repayment program.
Footnotes
Disclosures: None.
References
- 1.Casper ML, Wing S, Anda RF, Knowles M, Pollard RA. The shifting stroke belt. Changes in the geographic pattern of stroke mortality in the united states, 1962 to 1988. Stroke. 1995;26:755–760. doi: 10.1161/01.str.26.5.755. [DOI] [PubMed] [Google Scholar]
- 2.Lanska DJ, Kuller LH. The geography of stroke mortality in the united states and the concept of a stroke belt. Stroke. 1995;26:1145–1149. doi: 10.1161/01.str.26.7.1145. [DOI] [PubMed] [Google Scholar]
- 3.Liao Y, Greenlund KJ, Croft JB, Keenan NL, Giles WH. Factors explaining excess stroke prevalence in the us stroke belt. Stroke. 2009;40:3336–3341. doi: 10.1161/STROKEAHA.109.561688. [DOI] [PubMed] [Google Scholar]
- 4.Glymour MM, Avendano M, Berkman LF. Is the ‘stroke belt’ worn from childhood?: Risk of first stroke and state of residence in childhood and adulthood. Stroke. 2007;38:2415–2421. doi: 10.1161/STROKEAHA.107.482059. [DOI] [PubMed] [Google Scholar]
- 5.Howard G. Why do we have a stroke belt in the southeastern united states? A review of unlikely and uninvestigated potential causes. The American journal of the medical sciences. 1999;317:160–167. doi: 10.1097/00000441-199903000-00005. [DOI] [PubMed] [Google Scholar]
- 6.Howard G, Cushman M, Prineas RJ, Howard VJ, Moy CS, Sullivan LM, et al. Advancing the hypothesis that geographic variations in risk factors contribute relatively little to observed geographic variations in heart disease and stroke mortality. Prev Med. 2009;49:129–132. doi: 10.1016/j.ypmed.2009.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lackland DT, Egan BM, Jones PJ. Impact of nativity and race on “stroke belt” mortality. Hypertension. 1999;34:57–62. doi: 10.1161/01.hyp.34.1.57. [DOI] [PubMed] [Google Scholar]
- 8.Nahab F, Le A, Judd S, Frankel MR, Ard J, Newby PK, et al. Racial and geographic differences in fish consumption: The regards study. Neurology. 2011;76:154–158. doi: 10.1212/WNL.0b013e3182061afb. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hajjar I, Kotchen T. Regional variations of blood pressure in the united states are associated with regional variations in dietary intakes: The nhanes-iii data. J Nutr. 2003;133:211–214. doi: 10.1093/jn/133.1.211. [DOI] [PubMed] [Google Scholar]
- 10.Ding EL, Mozaffarian D. Optimal dietary habits for the prevention of stroke. Seminars in neurology. 2006;26:11–23. doi: 10.1055/s-2006-933305. [DOI] [PubMed] [Google Scholar]
- 11.Lopez-Garcia E, Schulze MB, Fung TT, Meigs JB, Rifai N, Manson JE, et al. Major dietary patterns are related to plasma concentrations of markers of inflammation and endothelial dysfunction. Am J Clin Nutr. 2004;80:1029–1035. doi: 10.1093/ajcn/80.4.1029. [DOI] [PubMed] [Google Scholar]
- 12.Meyer KA, Sijtsma FP, Nettleton JA, Steffen LM, Van Horn L, Shikany JM, et al. Dietary patterns are associated with plasma f(2)-isoprostanes in an observational cohort study of adults. Free radical biology & medicine. 2013;57:201–9. doi: 10.1016/j.freeradbiomed.2012.08.574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kuo F, Gardener H, Dong C, Cabral D, Della-Morte D, Blanton SH, et al. Traditional cardiovascular risk factors explain the minority of the variability in carotid plaque. Stroke; a journal of cerebral circulation. 2012;43:1755–1760. doi: 10.1161/STROKEAHA.112.651059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Howard G, Cushman M, Kissela BM, Kleindorfer DO, McClure LA, Safford MM, et al. Traditional risk factors as the underlying cause of racial disparities in stroke: Lessons from the half-full (empty?) glass. Stroke. 2011;42:3369–3375. doi: 10.1161/STROKEAHA.111.625277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hu FB, Rimm E, Smith-Warner SA, Feskanich D, Stampfer MJ, Ascherio A, et al. Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am J Clin Nutr. 1999;69:243–249. doi: 10.1093/ajcn/69.2.243. [DOI] [PubMed] [Google Scholar]
- 16.Abiemo EE, Alonso A, Nettleton JA, Steffen LM, Bertoni AG, Jain A, et al. Relationships of the mediterranean dietary pattern with insulin resistance and diabetes incidence in the multi-ethnic study of atherosclerosis (mesa) The British journal of nutrition. 2012:1–8. doi: 10.1017/S0007114512003339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Newby PK, Tucker KL. Empirically derived eating patterns using factor or cluster analysis: A review. Nutrition reviews. 2004;62:177–203. doi: 10.1301/nr.2004.may.177-203. [DOI] [PubMed] [Google Scholar]
- 18.Howard G, Labarthe DR, Hu J, Yoon S, Howard VJ. Regional differences in african americans’ high risk for stroke: The remarkable burden of stroke for southern african americans. Annals of epidemiology. 2007;17:689–696. doi: 10.1016/j.annepidem.2007.03.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yang D, Howard G, Coffey CS, Roseman J. The confounding of race and geography: How much of the excess stroke mortality among african americans is explained by geography? Neuroepidemiology. 2004;23:118–122. doi: 10.1159/000075954. [DOI] [PubMed] [Google Scholar]
- 20.Howard VJ, Cushman M, Pulley L, Gomez CR, Go RC, Prineas RJ, et al. The reasons for geographic and racial differences in stroke study: Objectives and design. Neuroepidemiology. 2005;25:135–143. doi: 10.1159/000086678. [DOI] [PubMed] [Google Scholar]
- 21.Howard VJ, Kleindorfer DO, Judd SE, McClure LA, Safford MM, Rhodes JD, et al. Disparities in stroke incidence contributing to disparities in stroke mortality. Ann Neurol. 2011;69:619–627. doi: 10.1002/ana.22385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Morton LM, Cahill J, Hartge P. Reporting participation in epidemiologic studies: A survey of practice. Am J Epidemiol. 2006;163:197–203. doi: 10.1093/aje/kwj036. [DOI] [PubMed] [Google Scholar]
- 23.Jackson R, Chambless LE, Yang K, Byrne T, Watson R, Folsom A, et al. Differences between respondents and nonrespondents in a multicenter community-based study vary by gender and ethnicity. J Clin Epidemiol. 1996;49:1441–1446. doi: 10.1016/0895-4356(95)00047-x. [DOI] [PubMed] [Google Scholar]
- 24.Block G, Woods M, Potosky A, Clifford C. Validation of a self-administered diet history questionnaire using multiple diet records. J Clin Epidemiol. 1990;43:1327–1335. doi: 10.1016/0895-4356(90)90099-b. [DOI] [PubMed] [Google Scholar]
- 25.Caan BJ, Slattery ML, Potter J, Quesenberry CP, Jr, Coates AO, Schaffer DM. Comparison of the block and the willett self-administered semiquantitative food frequency questionnaires with an interviewer-administered dietary history. Am J Epidemiol. 1998;148:1137–1147. doi: 10.1093/oxfordjournals.aje.a009598. [DOI] [PubMed] [Google Scholar]
- 26.Nettleton JA, Polak JF, Tracy R, Burke GL, Jacobs DR., Jr Dietary patterns and incident cardiovascular disease in the multi-ethnic study of atherosclerosis. Am J Clin Nutr. 2009;90:647–654. doi: 10.3945/ajcn.2009.27597. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Nettleton JA, Schulze MB, Jiang R, Jenny NS, Burke GL, Jacobs DR., Jr A priori-defined dietary patterns and markers of cardiovascular disease risk in the multi-ethnic study of atherosclerosis (mesa) Am J Clin Nutr. 2008;88:185–194. doi: 10.1093/ajcn/88.1.185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Fung TT, Willett WC, Stampfer MJ, Manson JE, Hu FB. Dietary patterns and the risk of coronary heart disease in women. Archives of internal medicine. 2001;161:1857–1862. doi: 10.1001/archinte.161.15.1857. [DOI] [PubMed] [Google Scholar]
- 29.Fung TT, Stampfer MJ, Manson JE, Rexrode KM, Willett WC, Hu FB. Prospective study of major dietary patterns and stroke risk in women. Stroke. 2004;35:2014–2019. doi: 10.1161/01.STR.0000135762.89154.92. [DOI] [PubMed] [Google Scholar]
- 30.Nettleton JA, Matijevic N, Follis JL, Folsom AR, Boerwinkle E. Associations between dietary patterns and flow cytometry-measured biomarkers of inflammation and cellular activation in the atherosclerosis risk in communities (aric) carotid artery mri study. Atherosclerosis. 2010;212:260–267. doi: 10.1016/j.atherosclerosis.2010.04.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nettleton JA, Diez-Roux A, Jenny NS, Fitzpatrick AL, Jacobs DR., Jr Dietary patterns, food groups, and telomere length in the multi-ethnic study of atherosclerosis (mesa) Am J Clin Nutr. 2008;88:1405–1412. doi: 10.3945/ajcn.2008.26429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Nettleton JA, Steffen LM, Ni H, Liu K, Jacobs DR., Jr Dietary patterns and risk of incident type 2 diabetes in the multi-ethnic study of atherosclerosis (mesa) Diabetes care. 2008;31:1777–1782. doi: 10.2337/dc08-0760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Nettleton JA, Steffen LM, Mayer-Davis EJ, Jenny NS, Jiang R, Herrington DM, et al. Dietary patterns are associated with biochemical markers of inflammation and endothelial activation in the multi-ethnic study of atherosclerosis (mesa) Am J Clin Nutr. 2006;83:1369–1379. doi: 10.1093/ajcn/83.6.1369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gardener H, Wright CB, Gu Y, Demmer RT, Boden-Albala B, Elkind MS, et al. Mediterranean-style diet and risk of ischemic stroke, myocardial infarction, and vascular death: The northern manhattan study. Am J Clin Nutr. 2011;94:1458–1464. doi: 10.3945/ajcn.111.012799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gardener H, Scarmeas N, Gu Y, Boden-Albala B, Elkind MS, Sacco RL, et al. Mediterranean diet and white matter hyperintensity volume in the northern manhattan study. Archives of neurology. 2012;69:251–256. doi: 10.1001/archneurol.2011.548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, et al. Primary prevention of cardiovascular disease with a mediterranean diet. The New England journal of medicine. 2013;368:1279–1290. doi: 10.1056/NEJMc1806491. [DOI] [PubMed] [Google Scholar]
- 37.Zhang Q, Zhou Y, Gao X, Wang C, Zhang S, Wang A, et al. Ideal cardiovascular health metrics and the risks of ischemic and intracerebral hemorrhagic stroke. Stroke. 2013;44:2451–2456. doi: 10.1161/STROKEAHA.113.678839. [DOI] [PubMed] [Google Scholar]
- 38.Agurs-Collins T, Rosenberg L, Makambi K, Palmer JR, Adams-Campbell L. Dietary patterns and breast cancer risk in women participating in the black women’s health study. Am J Clin Nutr. 2009;90:621–628. doi: 10.3945/ajcn.2009.27666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Satagopan JM, Ben-Porat L, Berwick M, Robson M, Kutler D, Auerbach AD. A note on competing risks in survival data analysis. Br J Cancer. 2004;91:1229–1235. doi: 10.1038/sj.bjc.6602102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Yang Y, Buys DR, Judd SE, Gower BA, Locher JL. Favorite foods of older adults living in the black belt region of the united states. Influences of ethnicity, gender, and education. Appetite. 2013;63:18–2. doi: 10.1016/j.appet.2012.12.007. [DOI] [PMC free article] [PubMed] [Google Scholar]