Associations of dietary patterns with risk of incident atrial fibrillation in the REasons for Geographic And Racial Differences in Stroke (REGARDS)

Atrial fibrillation (AF) is the most frequently encountered cardiac arrhythmia in clinical practice and more often seen in older populations. The overall prevalence of AF in the United States is 1% to 2% [1]. The prevalence of AF in European individuals over the age of 65 is estimated at 7.8% [2]. The presence of AF is associated with an over two-fold increased risk of stroke, heart failure, and cardiovascular mortality [3]. The risk of developing AF, however, is not evenly distributed amongst the various racial and ethnic groups. Despite underrepresented racial and ethnic groups experiencing a higher burden of traditional AF risk factors and a poorer AF-associated outcome profile, a paradox exists in that the overall incidence and prevalence of AF in these groups is only half of what is observed in White individuals [4, 5]. Continued investigations to improve our understanding of the mechanisms that underlie AF, as well as the differences in AF incidence observed across various racial and ethnic groups, are important.

Adherence to the Mediterranean or Dietary Approach to Stop Hypertension (DASH) diets are advocated by major medical societies due to their established effectiveness in reducing incident cardiovascular disease (CVD) [6‐10]. The beneficial effects of diet on CVD risk reduction are thought to be due to reduction of inflammation and oxidative stress, mechanisms that are both implicated in the pathogenesis of AF [11‐14]. A secondary analysis of the Prevención con Dieta Mediterránea (PREDIMED) trial found that individuals randomized to a Mediterranean diet, including the use of extra virgin olive oil (EVOO), experienced a substantially reduced risk of incident AF [15]. There is limited observational research, however, looking at associations of the Mediterranean diet and AF risk.

Similarly, prior study has also not evaluated whether the less-healthy Western dietary patterns that do not incorporate elements of a heart-healthy diet may place an individual at an increased risk for AF. Prior research in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study demonstrated that a Southern dietary pattern, one that is high in added fats, fried food, eggs and egg dishes, organ meats, processed meats, and sugar-sweetened beverages, was associated with an increased risk of coronary heart disease and sudden cardiac death [16, 17]. We examined whether the risk of incident AF is lower in participants who adhere to heart-healthy dietary patterns—Mediterranean and plant-based—and higher in participants who adhere to less heart-health dietary patterns, e.g., the—Southern, sweets, or convenience as previously derived in REGARDS [18].

Details of the methods of the REGARDS study have been published [19]. Briefly, REGARDS is a prospective cohort study designed to identify contributors to regional and Black-White disparities in stroke mortality. The study over-sampled Black persons and residents of the stroke belt (North Carolina, South Carolina, Georgia, Alabama, Mississippi, Tennessee, Arkansas, and Louisiana). Between January 2003 and October 2007, using postal mailings and telephone interviews, a total of 30,239 participants were recruited from a commercially available list of residents. Socio-demographic information and medical histories were obtained by a computer-assisted telephone interview (CATI). An in-home examination was performed 3–4 weeks after the telephone interview. Trained staff collected medication information, blood and urine samples, blood pressure readings, and a resting electrocardiogram (ECG). Approximately 10 years after the baseline assessment, 2013–2016, participants who were still alive and active completed a follow-up examination similar to the baseline visit. The institutional review boards at the collaborating centers approved the REGARDS study protocol, and all participants provided written informed consent.

There were 18,967 REGARDS participants with complete covariate data and dietary assessment at baseline and 8,977 of these were included in the prospective analyses because they were without prevalent AF and completed a follow-up examination. Baseline characteristics stratified by MDS category and dietary pattern quartiles are shown in Table 1 and Supplemental Table 1, respectively. Mean (SD) age was 63 (8.3) years, 30% were Black, and 56% female. 782 (8.7%) participants developed AF over a median (IQR) follow-up of 9.3 (1.4) years. Compared to participants in the lowest MDS category, those in higher categories were older, more likely to be male, be black, have a high school education, be more physically active and less likely to smoke, have a low annual income, have hypertension, and have diabetes (Table 1).

Risk of incident AF according to the various dietary patterns and MDS are reported in Fig. 1. Neither the MDS score (OR per SD increment = 1.03; 95% CI 0.95–1.11) or the plant-based dietary pattern (OR per SD increment = 1.03; 95% CI 0.94–1.12) were associated with AF risk. In age-, race-, region-, sex and income-adjusted models, greater adherence to the sweets dietary pattern was associated with a lower AF risk (OR per SD increment = 0.89; 95% confidence interval (CI): 0.81, 0.97; p-trend = 0.01), while greater adherence to the Southern pattern was associated with an increased risk of AF (OR per SD increment = 1.12; 95% CI 1.02, 1.22; p-trend = 0.01) (Fig. 1). After additional adjustment for covariates, only the association between the sweets dietary pattern and AF remained significant (OR per SD increment = 0.90; 95% CI 0.82, 0.99; p-trend = 0.03).

Measures of adiposity and hypertension were most responsible for attenuation of association between Southern pattern and incident AF (Supplemental Table 2). The convenience and alcohol and salads dietary patterns were not associated with incident AF risk. In cross-sectional analyses, no associations were observed for any of the dietary patterns or MDS with prevalent AF (Supplemental Fig. 1). Findings were similar for both prevalent and incident AF when evaluating participants with the MIND diet score (Supplemental Table 3).

In a biracial cohort of individuals participating in the ongoing REGARDS study, a stronger adherence to a healthier dietary pattern, defined as either plant-based or Mediterranean, was not associated with a reduced incident AF risk. In addition, higher consumption of unhealthy Western dietary patterns, defined here as convenience, sweets, or Southern, was not associated with an increased incident AF risk.

Inflammation and oxidative stress are considered to play a role in AF pathogenesis and protective benefits of the Mediterranean diet are thought to involve modification of these pathways [12, 27, 28]. Anti-inflammatory and anti-oxidative properties are well documented in principal components of the Mediterranean diet [29‐31]. A secondary analysis of the PREDIMED trial found that those randomized to a Mediterranean diet with EVOO experienced a 38% reduced risk of incident AF [15]. A similar association, however, was not found for participants in this trial randomized to the Mediterranean diet that included mixed nuts. No details were provided on the quantity of EVOO consumed or how much quantities differed amongst the EVOO, mixed nuts, and control groups, making it difficult to clinically interpret findings. Additional limitations of the PREDIMED trial included a sample size consisting of older White individuals at high cardiovascular risk and an unexpectedly low AF incidence (1.1% over 5 years) [15].

A “plant-based” dietary pattern is also thought to protect against AF development via similar mechanisms to those for the Mediterranean diet. Multiple large observational studies performed in biracial or multi-ethnic populations, however, reported no association between adherence to an American Heart Association recommended diet, which largely overlaps with a plant-based diet, and risk of AF [32‐34]. Adherence in these studies were also measured using a FFQ, and diet scores were calculated as the sum of the scores for each of 5 individual components: fruits and vegetables (≥ 4.5 cups per day), fish (≥ 2 3.5-oz servings per week), fiber-rich whole grains (≥ 3 1-oz-equivalent serving per day), sodium (< 1500 mg/day), and sugar-sweetened beverages (≤ 450 kcal/week). The Framingham Heart Study, a predominantly white cohort of European ancestry, separately evaluated two of these components, fiber-rich whole grains and fish intake, and also reported no association [35].

The lack of association of MDS or plant-based dietary pattern with the risk of incident AF observed is consistent with much of the published observational literature mentioned above. Dietary factors do not appear to confer substantially to AF risk, and it is possible that any association with AF may be through indirect pathways where poor dietary habits increase the susceptibility to comorbidities that enhance an individual's risk for AF. It has been previously demonstrated that adherence to a plant-based diet reduces the likelihood of primary AF risk factors, including hypertension, obesity, diabetes, and coronary heart disease [36‐41]. Blood pressure is reduced due to improved vasodilation and increased potassium intake [36, 42, 43]. Improvements in blood glucose control are thought to be due to the lower content of advanced glycation endproducts [44]. Higher fiber, lower fat content of plant‐based nutrition, and reduced energy density all may contribute to reduce obesity risk [39].

In this study, contrary to expectation, the sweets dietary pattern was independently associated with a reduction in AF risk. Although our findings run contrary to what we would have expected, a similar protective relationship was described in REGARDS for the sweets pattern with both sudden cardiac death in those with baseline coronary heart disease as well as incident stroke [16, 45]. It is well established, however, that a diet high in added sweets and saturated fats is associated with increased CVD incidence and cardiovascular mortality [46, 47].

A major limitation was that a reliable estimated energy requirement could not be calculated due to how physical activity was assessed, and adjustment for dietary under-reporting could not be performed, potentially introducing information bias. Additional limitations were also present. The majority of AF cases were by self-report at a single follow-up visit and participants may not have accurately recalled a diagnosis that occurred years earlier. More sensitive detection methods such as hospitalization diagnosis were not available. This likely led to the under ascertainment of AF events, especially those that were asymptomatic. This method of AF detection treated incident AF as a binary variable and did not allow use of longitudinal time to event analyses. The FFQ relied on recall of dietary intake, and this is prone to measurement error. Inaccuracies in reporting dietary intake may have resulted in misclassification, biasing results toward the null. Unfortunately, correction for potential measurement error was not possible. Diet was assessed at only one time point, so potential changes in diet could not be accounted for. Nearly 30% of participants did not complete a FFQ at baseline and these individuals were more likely to be Black race, less educated, and have a lower income. Finally, certain dietary pattern designations may not wholly depict each of the individual food components and it is possible that alternative naming options may be more representative; however, there are now multiple REGARDS publications using the names from the original design paper, so, for consistency, we have chosen to keep the names of the patterns the same.

In conclusion, in this large population-based sample of US adults, adherence to a variety of Western dietary patterns as well as to the Mediterranean diet were not associated with risk of incident AF. While specific dietary patterns have been strongly associated with AF risk factors, our findings suggest that they do not directly affect susceptibility to AF development.