Abstract
Objective:
The Mediterranean diet is associated with lower risks for type 2 diabetes (T2D) and cardiovascular disease in certain populations, although data among diverse groups are limited. We evaluated cross-sectional and prospective associations between a novel South Asian Mediterranean-style (SAM) diet and cardiometabolic risk among US South Asians.
Methods:
We included 891 participants at baseline in the MASALA study. We grouped culturally relevant foods into 9 categories to construct the SAM score. We examined associations between this score, cardiometabolic risk factors, and incident T2D.
Results:
At baseline, higher adherence to the SAM diet was associated with lower HbA1c (−0.43 ± 0.15% per 1-unit increase in SAM score; P=0.004) and lower pericardial fat volume (−1.22 ± 0.55 cm3; P=0.03), as well as a lower likelihood of obesity (odds ratio [OR] 0.88, 95% confidence interval [CI] 0.79 to 0.98) and fatty liver (OR 0.82, 95% CI 0.68 to 0.98). Over the follow-up (~5 y), 45 participants developed T2D; each 1-unit increase in SAM score was associated with a 25% lower odds of incident T2D (OR 0.75, 95% CI 0.59 to 0.95).
Conclusions:
A greater intake of a South Asian Mediterranean-style diet is associated with favorable adiposity measures and a lower likelihood of incident T2D.
Keywords: Mediterranean diet, nutrition, South Asians, cardiometabolic risk, diabetes, health disparities
INTRODUCTION
South Asians represent one quarter of the global population and include individuals with ancestry from Bangladesh, Bhutan, India, the Maldives, Nepal, Pakistan, and Sri Lanka. They are one of the most rapidly growing ethnic groups in the United States with a population size of 5.4 million individuals as of 2017.(1) Research has consistently shown that South Asians have a disproportionally increased risks of type 2 diabetes (T2D) and cardiovascular disease (CVD) compared with other populations, including White and other Asian American groups.(2–5)
Various healthy dietary patterns are associated with lower risks of CVD and T2D in the broader population, and therefore may be a compelling means to mitigate these consistently reported adverse health conditions among South Asians. One such dietary pattern is the Mediterranean diet, a primarily plant-based dietary pattern that is characterized by higher intakes of vegetables, whole grains, olive oil, nuts, beans and other legumes, and lower intakes of animal proteins (with fish being the preferred source), as well as moderate alcohol intake. It is associated with marked health benefits, including reduced risks of CVD,(6) T2D,(7) certain cancers,(8) and overall mortality.(9) However, most studies investigating the health benefits of the Mediterranean diet have been conducted among those living in Mediterranean or European countries or among predominantly US-based White populations, while similar data among more diverse groups are limited.(10)
Therefore, in the current study we aimed to (1) adapt an existing Mediterranean diet score to be inclusive of foods consumed by the US South Asian population using data from a validated, ethnic-specific food frequency questionnaire (FFQ), and (2) examine the cross-sectional and prospective associations of this score with a broad panel of cardiometabolic markers among those enrolled in the Mediators of Atherosclerosis in South Asians Living in America (MASALA) study.
METHODS
Study Population
The MASALA study is a community-based prospective cohort of South Asian men and women living in the San Francisco Bay Area and greater Chicago area. A description of the MASALA study eligibility, methods, and measures has been published in detail previously.(11) Briefly, eligible cohort participants had South Asian ancestry, were ages 40 to 84 at the time of enrollment, and were able to communicate in English, Hindi, or Urdu. Those with cardiovascular disease at baseline were excluded. A total of 906 participants were enrolled and underwent the baseline clinical examination between October 2010 and March 2013. Participants returned for the second clinical examination between September 2015 and March 2018 (n=749). In the current analysis, we excluded participants who reported implausible energy intakes (<800 kcal/d or >4000 kcal/d for men and <500 kcal/d or >3500 kcal/d for women) for a total of 891 and 735 participants at baseline and follow-up, respectively.
Dietary Assessment and Mediterranean-Style Diet Score
Usual dietary intake over the previous year was assessed using the previously validated Study of Health Assessment and Risk in Ethnic Groups FFQ.(12) This ethnic-specific FFQ contains 163 food items (61 of which are unique to South Asian cuisine), and participants were asked to indicate their consumption frequency (per day, per week, per month, per year, or never) and serving size (either small or large relative to the specified average serving size) for each item.
Based on the previously developed Alternate Mediterranean Diet Score,(13) we constructed a South Asian Mediterranean-style (SAM) diet score by grouping foods into nine pre-defined categories (vegetables, fruits, legumes, nuts, whole grains, red/processed meats, fish, alcohol, and monounsaturated to saturated fat ratio). Supplemental Table 1 details what foods were included in each category. Participants consuming greater than the median intake received one point per category; those consuming between 5 and 25 g/day of alcohol received one point and red/processed meats were reversely scored. We summed these points for a possible SAM score ranging from 0–9, with higher scores reflecting greater adherence to a Mediterranean-style South Asian diet.
Ascertainment of Cardiometabolic Risk Factors
At baseline, we collected measures of subclinical atherosclerosis, glycemia, dyslipidemia, blood pressure, uric acid, anthropometry, and computed tomography (CT) derived ectopic and subcutaneous fat depots. At the follow-up (approximately 5 years later), we collected updated measures of body weight, fasting glucose, glycated hemoglobin (HbA1c), and uric acid.
Subclinical Atherosclerosis
Participants underwent high-resolution B-mode ultrasonography (University of California, San Francisco [UCSF]: Vivid 7 ultrasound, GE Healthcare; Northwestern University [NWU]: Acuson Sequoia C256, Siemens Healthcare, Mountain View, CA) to measure right and left internal and common carotid artery intima-media thickness (IMT). The radiographic protocols for these measures have been described in detail previously.(11, 14) Cardiac CT scans were performed with gated-cardiac CT scanners (UCSF: 16D scanner, Philips Medical Systems, Andover, MA, or the MSD Aquilion 64 model, Toshiba Medical Systems, Tustin, CA; NWU: Sensation Cardiac 64 scanner, Siemens Medical Solutions, Malvern, PA). Coronary calcification in each of the 4 major coronary arteries was quantified using the Agatston score, and the sum of the unadjusted score was used in our analysis.(15)
Glycemia Measures
Fasting blood samples were collected by certified phlebotomists or nurses. At baseline and follow-up, fasting plasma glucose was measured using the hexokinase method (Ortho Clinical Diagnostics, Johnson & Johnson). HbA1c was measured using the immunoturbidimetry assay. Fasting serum insulin was measured using the sandwich immunoassay method (Roche Elecsys 2010, Roche Diagnostics).
At baseline, participants who were not taking diabetes medications underwent a 75g oral glucose tolerance test. Blood samples of glucose and insulin were collected after two hours. The Homeostatic Model Assessment for Insulin Resistance (HOMA-IR; a measure of insulin resistance) was calculated as [Glucose (mmol/L) x Insulin (uIU/mL)/22.5]. β-cell function (a surrogate measure for insulin sensitivity) was estimated using the oral disposition index which was calculated as [(Δinsulin0–30/Δglucose0–30)*(1/fasting insulin)].
Participants were classified as having T2D if they were using a glucose-lowering medication and/or had a fasting plasma glucose ≥7.0 mmol/l and/or had a 2-hour post-challenge glucose ≥11.1 mmol/l. Participants without T2D at baseline but who met the criteria at follow-up were defined as incident T2D cases.
Dyslipidemia, Inflammatory Markers, and Uric Acid
Plasma concentrations of high-density lipoprotein (HDL), low-density lipoprotein (LDL), and triglycerides were measured at baseline and follow-up using enzymatic methods (Quest Diagnostics, San Jose, CA). Baseline high sensitivity C-reactive protein (hsCRP) was measured using the BNII nephelometer (N High Sensitivity CRP; Siemens Healthcare Diagnostics, Deerfield, IL). Baseline serum total adiponectin was measured using Millipore Luminex adipokine panel A (EMD Millipore, Billerica, MA). Uric acid level was measured at baseline and follow-up using spectrophotometry (Quest Diagnostics, San Jose, CA). We defined hyperuricemia as a uric acid level ≥7.0 mg/dL.
Body Composition
Body weight, height, and waist circumference were measured by centrally trained study staff using a standard protocol.(11) Body mass index (BMI) was defined as the participant’s weight in kilograms divided by their height in meters squared (kg/m2). We defined overweight as BMI ≥23.0 kg/m2 and obesity as BMI ≥27.5 kg/m2 in accordance with the current World Health Organization recommendation for Asian populations,(16) as has been done previously in this cohort.(17) Subcutaneous and visceral fat area were measured with abdominal CT using standard protocols.(11) Pericardial fat volume and hepatic fat attenuation were both obtained using non-contrast cardiac CT imaging. We defined fatty liver as <40 Hounsfield units (HU).(18)
Other Cardiometabolic Risk Factors
Seated resting blood pressure was evaluated 3 times using an automated blood pressure monitor (V100 Vital Signs Monitor; GE Healthcare, Fairfield, CT). An average of the final two readings was used for analysis. We defined hypertension as either the self-reported use of an antihypertensive medication or blood pressure ≥140/90 mmHg.
Covariate Assessment
All participants provided information on personal history, demographics, socioeconomic status, medical history, medication use, and family history of relevant diseases at the baseline visit. Physical activity (PA) was measured as total metabolic equivalent (MET) minutes per week using the Typical Week’s Physical Activity Survey (TWPAS), which was adapted from the Cross-Cultural Activity Participation Study.(19) Cultural beliefs and behaviors were assessed using a multi-dimensional measure of acculturation developed specifically for this cohort.(20) This 7-item traditional cultural beliefs scale is scored from 7 to 35, with higher values indicating weaker traditional South Asian beliefs, and has previously shown good reliability and validity.(20) Routine experiences of interpersonal discrimination and unfair treatment were measured using the Everyday Discrimination Scale, a valid and reliable 9-item scale scored from 9 to 54, with higher values indicating more discrimination.(21) Total energy intake was assessed from the FFQ as kilocalories per day. All questionnaires were translated into Hindi and Urdu.
Statistical Analysis
We compared baseline characteristics of study participants across quartiles of the SAM score using general linear regression adjusted for age, sex, and total energy intake for continuous variables and chi-square tests for categorical variables. We tested for linear trends by assigning the median value to each quartile and treating this as a continuous variable in the regression model.
We evaluated the associations between the SAM score and cardiometabolic risk factors using multivariable general linear regression for continuous outcomes and logistic regression for binary outcomes. We tested the assumptions of linear regression by examining the normal probability plots of the residuals and the residual vs. predicted plots. Where necessary, we log-transformed the outcome variables and re-examined each diagnostic plot.
In the first multivariable model, we adjusted for age, sex, and total energy intake (model 1). We further adjusted for education level, family income, number years lived in the United States, physical activity, perception of discrimination, the traditional cultural beliefs scale, tobacco pack years of smoking, family history of diabetes (parents and siblings), use of diabetes medication, cholesterol-lowering medication, and hypertension medication in model 2. Based on prior literature, we hypothesized a priori that measures of discrimination and cultural beliefs could confound our associations of interest.(20, 22–24) Finally, because these associations could potentially be mediated by BMI, we adjusted for it separately in model 3. For all linear models, results are presented as the unit change (or percent change for log-transformed variables) in outcome per 1-unit increase in SAM score. To minimize potential confounding by prevalent disease, we excluded participants with T2D in analyses where glycemia measures were the outcome. For all prospective associations, we adjusted for the baseline value of the covariates and of the corresponding outcome measure. Finally, for the association between SAM score and incident T2D, we tested for potential effect modification by age, sex, diabetes family history, obesity status, and physical activity level by including cross-product terms between these variables and the SAM score variable. All statistical tests were 2-sided and were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC).
Ethics
The MASALA study protocol was approved by the Institutional Review Boards of the University of California San Francisco and Northwestern University. The current analysis protocol was approved by the Institutional Review Board of Brigham and Women’s Hospital, Boston, MA. All participants provided written informed consent.
RESULTS
Baseline Characteristics
Participant characteristics at baseline are shown in Table 1. Compared with participants in the lowest quintile of SAM score, those in the uppermost quintile were older (56.9 vs. 52.4 years in Q4 and Q1 respectively; P for trend <0.0001), more often held a Bachelor’s degree (91.0 vs. 83.7%; P for trend=0.02), reported fewer experiences of discrimination (score of 14.5 vs. 16.1; P for trend=0.03), were less likely to be smokers (1.9 vs. 5.9%; P for trend=0.04), and had a lower BMI (25.3 vs. 26.6 kg/m2; P for trend=0.005). Those in the highest quintile of SAM score also reported greater intakes of carbohydrates, dietary fiber, and total energy, but lower intakes of protein and saturated fat.
Table 1.
Baseline characteristics of participants enrolled in the MASALA cohort according to quartile of the South Asian Mediterranean (SAM)-style score, 2010–2013
| Characteristic* | South Asian Mediterranean-Style Diet Score |
P-value | |||
|---|---|---|---|---|---|
| Quartile 1 | Quartile 2 | Quartile 3 | Quartile 4 | ||
|
|
|||||
| N | 153 | 355 | 172 | 211 | |
| Mean score | 1.7 | 3.5 | 5.0 | 6.5 | |
|
| |||||
| Age, years | 52.4 (0.8) | 55.1 (0.5) | 55.9 (0.7) | 56.9 (0.7) | <0.0001 |
| Female, % | 45.8 | 50.4 | 44.2 | 44.6 | 0.38 |
| Duration lived in US, years | 27.2 (0.8) | 26.3 (0.5) | 27.3 (0.7) | 27.7 (0.7) | 0.48 |
| Bachelor’s degree or more, % | 83.7 | 86.5 | 90.1 | 91.0 | 0.02 |
| Family income of at least $75,000, % | 78.0 | 67.5 | 73.2 | 80.6 | 0.11 |
| Sum of cultural tradition measures | 14.4 (0.5) | 13.6 (0.3) | 13.6 (0.5) | 14.8 (0.5) | 0.67 |
| Everyday discrimination scale | 16.1 (0.5) | 15.0 (0.3) | 14.8 (0.5) | 14.5 (0.4) | 0.03 |
| Dine out at least once per week, % | 46.4 | 43.9 | 44.2 | 42.2 | 0.47 |
| Current smokers, % | 5.9 | 3.1 | 2.3 | 1.9 | 0.04 |
| Physical activity, MET-min/week | 10,091 (347) | 10,098 (213) | 9751 (306) | 10,585 (293) | 0.47 |
| Body mass index, kg/m2 | 26.6 (0.3) | 26.2 (0.2) | 25.7 (0.3) | 25.3 (0.3) | 0.005 |
| Hyperlipidemia, % | 81.1 | 79.5 | 82.7 | 76.1 | 0.36 |
| Hypertension, % | 36.0 | 41.7 | 41.9 | 38.4 | 0.85 |
| Dietary factors, per day | |||||
| Energy (kcal) | 1278 (35) | 1600 (23) | 1781 (33) | 2007 (30) | <0.0001 |
| Carbohydrates (g) | 240.0 (2.2) | 245.2 (1.3) | 248.0 (1.9) | 251.5 (1.8) | <0.0001 |
| Protein (g) | 63.8 (0.8) | 61.6 (0.5) | 61.9 (0.7) | 60.7 (0.7) | 0.007 |
| Monounsaturated fat (g) | 20.7 (0.4) | 21.6 (0.2) | 22.7 (0.4) | 22.9 (0.3) | <0.0001 |
| Polyunsaturated fat (g) | 10.7 (0.2) | 11.8 (0.1) | 12.7 (0.2) | 13.2 (0.2) | <0.0001 |
| Saturated fat (g) | 18.0 (0.3) | 15.6 (0.2) | 13.8 (0.3) | 12.4 (0.3) | <0.0001 |
| Dietary fiber (g) | 16.8 (0.3) | 19.7 (0.2) | 22.1 (0.3) | 23.9 (0.3) | <0.0001 |
| Alcohol (g) | 2.5 (0.5) | 3.2 (0.3) | 1.8 (0.4) | 2.6 (0.4) | 0.58 |
| Food groups, servings per day | |||||
| Fish and seafood | 0.16 (0.02) | 0.12 (0.01) | 0.12 (0.02) | 0.13 (0.02) | 0.34 |
| Red/processed meat | 0.27 (0.03) | 0.19 (0.02) | 0.10 (0.02) | 0.04 (0.02) | <0.0001 |
| Whole grains | 1.6 (0.1) | 1.9 (0.1) | 2.1 (0.1) | 2.3 (0.1) | <0.0001 |
| Legumes | 1.0 (0.1) | 1.3 (0.0) | 1.6 (0.1) | 1.8 (0.1) | <0.0001 |
| Fruit | 2.0 (0.1) | 2.4 (0.1) | 2.8 (0.1) | 3.3 (0.1) | <0.0001 |
| Vegetables | 5.7 (0.2) | 6.7 (0.1) | 7.9 (0.2) | 8.5 (0.2) | <0.0001 |
| Nuts | 0.37 (0.05) | 0. 65 (0.03) | 1.01 (0.04) | 1.24 (0.04) | <0.0001 |
| Dairy products | 4.3 (0.2) | 3.8 (0.1) | 3.7 (0.1) | 3.4 (0.1) | <0.0001 |
| Sweets | 0.74 (0.05) | 0.63 (0.03) | 0.46 (0.04) | 0.48 (0.04) | <0.0001 |
Values are reported as either age, sex, and calorie adjusted mean (standard error) or percentages using ANOVA for continuous variables and chi-square tests for categorical variables.
Cross-sectional Analysis
All cross-sectional associations between the SAM score and measures of cardiometabolic risk at baseline are shown in Table 2. At baseline, each 1-unit increase in SAM score was associated with a lower HbA1c in the fully adjusted model (−0.43 ± 0.15%; P=0.004). No other measures of glycemia were significantly associated with the SAM score at baseline. A higher SAM score was significantly and inversely associated with a lower hsCRP concentration after adjusting for age, sex, and total energy intake, but this association was attenuated after multivariable adjustment.
Table 2.
Cross-sectional associations (β [standard error]) between the South Asian Mediterranean (SAM)-style diet score and measures of cardiometabolic risk among MASALA study participants at baseline
|
Model 1 Age, sex, and energy adjustment |
Model 2 Multivariable adjustment* |
Model 3 Multivariable adjustment* + BMI |
|
|---|---|---|---|
| SUBCLINICAL MEASURES OF ATHEROSCLEROSIS | |||
|
| |||
| Coronary artery calcium score† | −2.59 (4.57) | −2.78 (4.63) | −2.17 (4.63) |
| Common carotid IMT, mm† | −0.47 (0.48) | −0.43 (0.51) | −0.29 (0.51) |
| Internal carotid IMT, mm† | −0.56 (0.66) | −0.50 (0.68) | −0.33 (0.68) |
|
| |||
| GLYCEMIA MEASURES ** | |||
|
| |||
| Fasting glucose (mg/dL)† | 0.01 (0.31) | 0.13 (0.33) | 0.16 (0.33) |
| HbA1c, %† | −0.43 (0.15)§ | −0.45 (0.15)§ | −0.43 (0.15)§ |
| Beta cell function†‡ | 0.00 (0.12) | 0.01 (0.12) | 0.01 (0.12) |
| HOMA-IR† | −2.23 (1.70) | −1.05 (1.76) | −0.84 (1.53) |
|
| |||
| PLASMA LIPIDS | |||
|
| |||
| Triglycerides, mg/dL† | −0.54 (1.02) | 0.16 (1.07) | 0.52 (1.05) |
| HDL-C, mg/dL | 0.06 (0.28) | −0.14 (0.28) | −0.23 (0.28) |
| LDL-C, mg/dL | −1.12 (0.72) | −0.51 (0.70) | −0.42 (0.70) |
|
| |||
| INFLAMMATION AND ADIPOKINES | |||
|
| |||
| C-reactive protein, µg/mL† | −6.09 (2.42)§ | −3.95 (2.53) | −2.46 (2.30) |
| Adiponectin, mg/dL† | 0.76 (1.42) | 0.80 (1.46) | 0.53 (1.45) |
|
| |||
| OTHER METABOLIC MEASURES | |||
|
| |||
| Uric acid, mg/dL | 0.00 (0.04) | 0.04 (0.04) | 0.06 (0.03) |
|
| |||
| BODY COMPOSITION MEASURES | |||
|
| |||
| Body mass index, kg/m2 | −0.24 (0.09)§ | −0.16 (0.09) | -- |
| Weight, kg | −0.47 (0.26) | −0.22 (0.27) | -- |
| Waist circumference, cm | −0.49 (0.22)ǁ | −0.29 (0.23) | 0.02 (0.13) |
| Subcutaneous fat area, cm2 | −1.98 (2.22) | −1.05 (2.28) | 1.39 (1.72) |
| Visceral fat area, cm2 | −3.07 (1.20)ǁ | −2.83 (1.25)ǁ | −1.35 (1.01) |
| Pericardial fat volume, cm3 | −2.37 (0.63)¶ | −1.81 (0.64)§ | −1.22 (0.55)ǁ |
| Hepatic fat attenuation, HU | 0.72 (0.23)§ | 0.60 (0.24)ǁ | 0.45 (0.22)ǁ |
All values are presented as β (standard error) per 1-unit increase in SAM score.
Adjusted for age, sex, total energy intake, years lived in the United States, education level, family income, physical activity, perception of discrimination, the traditional cultural beliefs scale, tobacco pack years of smoking, family history of diabetes, use of diabetes medication (except for glycemia measures), cholesterol-lowering medication, and hypertension medication.
Participants with diabetes were excluded from these four measures.
Values were log-transformed to obtain a normal distribution of the residuals. For outcomes that were log-transformed, values represent % increase in outcome variable for every 1 unit increase in SAM score.
Measured using the oral disposition index.
P<0.05
P<0.01
P<0.001
P<0.0001
For body composition measures, each 1-unit higher SAM score was associated with a lower BMI (−0.24 ± 0.09 kg/m2; P=0.008) after adjusting for age, sex, and energy intake, although this association was attenuated after multivariable adjustment. When we examined the likelihood of overweight and obesity, after multivariable adjustment, each 1-unit increase in SAM score was associated with 12% lower odds of obesity (OR 0.88, 95% confidence interval [CI] 0.79 to 0.98) but not overweight (OR 0.96, 95% CI 0.86 to 1.08) (Figure 1). A higher SAM score was not significantly associated with subcutaneous fat area but was initially associated with a lower visceral fat area (−2.83 ± 1.25 cm2; P=0.02). However, this did not remain significant after further adjustment for BMI. Finally, in our fully adjusted models, each 1-unit increase in SAM score was significantly associated with lower pericardial fat volume (−1.22 ± 0.55 cm3; P=0.03), higher hepatic fat attenuation (0.45 ± 0.22 HU; P=0.045), and lower odds of fatty liver disease (OR 0.82, 95% CI 0.68 to 0.98) (Figure 1). We found no evidence of an association between the SAM score and measures of subclinical atherosclerosis, blood lipids, adiponectin, hyperuricemia, or hypertension in any of the models.
Figure 1. Likelihood (odds ratio [95% confidence interval]) of cardiometabolic risk and incident type 2 diabetes per 1-unit increase of South Asian Mediterranean (SAM)-style diet score among MASALA study participants.
Fatty liver is defined as liver-spleen attenuation <40 Hounsfield units. Hypertension is defined using the NCEP criteria as blood pressure ≥140/90 mm Hg or use of medication. Obesity is defined as a body mass index ≥27.5 kg/m2. Overweight is defined as a body mass index ≥23 kg/m2. Hyperuricemia is defined as a serum uric acid level ≥7 mg/dL.
All models were adjusted for age, sex, total energy intake, years lived in the United States, education level, family income, physical activity, perception of discrimination, the traditional cultural beliefs scale, tobacco pack years of smoking, family history of diabetes, use of diabetes medication (except for incident diabetes outcome), cholesterol-lowering medication, and hypertension medication (except for hypertension outcome), and body mass index (except for overweight and obesity outcomes).
Prospective Analysis
In prospective analyses (~5 years after baseline), we identified 45 cases of incident diabetes. After multivariable adjustment, for each 1-unit higher SAM score the odds of incident T2D were 25% lower (OR 0.75, 95% CI 0.59 to 0.95) (Figure 1). We found no evidence of effect modification of this association according to age, sex, family history of diabetes, obesity status, or physical activity (all P for interaction > 0.05). Finally, we found no evidence of an association between baseline diet scores and follow-up measures of fasting glucose, HbA1c, triglycerides, HDL cholesterol, LDL cholesterol, or measures of anthropometry (Table 3).
Table 3.
Prospective associations (β [standard error]) between the South Asian Mediterranean (SAM)-style diet score and measures of cardiometabolic risk among MASALA study participants at the follow-up visit
|
Model 1 Age, sex, and energy adjustment |
Model 2 Multivariable adjustment* |
Model 3 Multivariable adjustment* + BMI |
|
|---|---|---|---|
| GLYCEMIA MEASURES ** | |||
|
| |||
| Fasting glucose, (mg/dL)† | −0.18 (0.29) | −0.21 (0.30) | −0.19 (0.30) |
| HbA1c, %† | 0.14 (0.15) | 0.05 (0.16) | 0.05 (0.16) |
|
| |||
| PLASMA LIPIDS | |||
|
| |||
| Triglycerides, mg/dL† | −0.10 (0.87) | −0.51 (0.90) | −0.39 (0.90) |
| HDL-C, mg/dL | −0.04 (0.21) | −0.06 (0.21) | −0.06 (0.21) |
| LDL-C, mg/dL | −0.86 (0.72) | −0.72 (0.74) | −0.78 (0.74) |
|
| |||
| OTHER METABOLIC MEASURES | |||
|
| |||
| Uric acid, mg/dL | 0.02 (0.03) | 0.02 (0.03) | 0.03 (0.03) |
|
| |||
| BODY COMPOSITION MEASURES | |||
|
| |||
| Body mass index, kg/m2 | −0.02 (0.04) | −0.01 (0.05) | -- |
| Weight, kg | −0.12 (0.10) | −0.12 (0.11) | -- |
| Waist circumference, cm | −0.29 (0.17) | −0.26 (0.16) | −0.20 (0.15) |
All values are presented as β (standard error) per 1-unit increase in SAM score.
Adjusted for age, sex, total energy intake, years lived in the United States, education level, family income, physical activity, perception of discrimination, the traditional cultural beliefs scale, tobacco pack years of smoking, family history of diabetes, use of diabetes medication (except for glycemia measures), cholesterol-lowering medication, hypertension medication, and baseline value of the corresponding cardiometabolic risk marker.
Participants with diabetes were excluded from these two measures.
Values were log-transformed to obtain a normal distribution of the residuals. For outcomes that were log-transformed, values represent % increase in outcome variable for every 1 unit increase in SAM score.
All P-values >0.05.
DISCUSSION
In this community-based prospective cohort study of South Asians, we found that a greater intake of a Mediterranean-style diet inclusive of foods commonly consumed by this population was associated with more favorable baseline measures of body composition (including pericardial fat volume and fatty liver) and HbA1c. Moreover, we found that this dietary pattern was associated with a lower risk of incident T2D after approximately 5 years of follow-up. These associations were independent of known confounders such as age, sex, years lived in the US, measures of socioeconomic status, lifestyle factors, comorbidities, medication use, and BMI. These data suggest that a Mediterranean-style diet may be a useful dietary strategy to help prevent T2D in this high-risk population.
The inverse associations between a higher SAM score and measures of ectopic fat observed in our study deserve comment. Ectopic fat is a major independent risk factor for cardiometabolic disease including T2D.(25–29) Compared with other ethnic groups, South Asians tend to have a less favorable body composition profile that is characterized by greater levels of ectopic fat depots including intermuscular and hepatic fat,(30) which may partly explain the disparities in CVD and T2D risk among this population.(31–34) Given the rising cardiometabolic public health concerns, adherence to a healthy dietary pattern such as a Mediterranean-style diet is a readily implementable strategy to mitigate these disparities. Further experimental studies are warranted to confirm these observational findings.
The Mediterranean diet has been previously shown to be associated with health benefits, including lower risks of T2D and CVD,(6, 7) although data among South Asians are very limited. Nevertheless, our findings are consistent with previous studies that examined the association between a Mediterranean-style diet (or components of this diet) and cardiometabolic risk among members of this population. An India-based randomized trial evaluated the potential cardioprotective effect of an Indo-Mediterranean diet (which encouraged high intakes of fruits, vegetables, nuts, whole grains, legumes, and mustard seed or soy bean oil) among individuals considered to be at high risk for coronary artery disease (CAD).(35) Participants consuming the Indo-Mediterranean diet had significant decreases in LDL cholesterol, triglycerides, fasting blood glucose, blood pressure, and BMI, as well as fewer non-fatal myocardial infarctions and sudden cardiac deaths.(35) Nigam and colleagues conducted a randomized trial among Asian Indian men with non-alcoholic fatty liver disease to examine the effect of consuming olive or canola oil (both high in monounsaturated fatty acids [MUFAs], the cornerstone and main source of fat in the Mediterranean diet) and reported improvements in BMI, insulin resistance, blood lipids, and fatty liver grading after 6 months.(36) Our study is the first to document an inverse association between a Mediterranean-style diet and selected cardiometabolic factors among a broader cohort of South Asian immigrants in the US and the first to document an inverse association between this dietary pattern and incident T2D among South Asians.
Prior research suggests that a Mediterranean-style diet is enjoyable and sustainable in the long term. For example, participants in the Dietary Intervention Randomized Controlled Trial (DIRECT) who were randomized to consume a Mediterranean diet lost ~50% more weight at the end of the 2-year follow-up than those consuming a low-fat diet.(37) When these participants were followed for an additional 4 years, those in the Mediterranean diet group regained the least amount of weight lost in the original trial (1.4 kg vs. 2.7 kg and 4.1 kg in the low-fat and low-carbohydrate groups, respectively; P=0.004 for all comparisons).(38) A post-experimental qualitative study of middle-aged UK adults found that participants considered the Mediterranean diet to be “enjoyable” and a “better quality of food”.(39) However, these individuals also felt that it was initially challenging to adapt to this eating pattern and that it resulted in greater food costs.(39) Similar studies would be useful to help inform potential barriers and further guide implementation among South Asians in the US.
A 2019 consensus report highlighted the importance of integrating cultural factors and preferences to individualize nutrition therapy for adults with diabetes or prediabetes.(40) To date, the majority of research on the Mediterranean diet has been conducted among predominantly White populations living in Europe or the US without taking into account food items and dishes that may be consumed by different ethnic groups.(10) To our knowledge, our study is the first to develop a comprehensive Mediterranean-style diet score that incorporates culturally relevant foods consumed by South Asians, which may facilitate the uptake of this healthful dietary pattern among this population. Indeed, prior research found that a sample of British Pakistani and Indian individuals reduced the overall amount of South Asian foods they consumed after receiving a T2D diagnosis.(41) Many of these participants were advised by healthcare providers to stop eating their traditional diets which drove their belief that these foods were collectively “damaging”.(41) Similarly, participants in a UK qualitative study conducted among ethnically diverse older adults also described receiving negative educational messages about their traditional foods.(42) Participants noted that they were often advised to stop consuming these foods altogether, which was viewed as unrealistic as food is tied to cultural identity.(42) A Canadian qualitative study among African Nova Scotians and Punjabi British Columbians also described a close relationship between eating patterns and identity, and these participants reported that they were unlikely to avoid eating the foods that they are used to.(43) Taken together, these data underscore the importance of studying and integrating culturally specific foods and dishes into dietary recommendations and lifestyle interventions. Indeed, pilot data from a community-based and culturally salient lifestyle intervention among South Asians was found to be effective and maintain full participant retention, and continued research in this area is warranted.(44)
The 2020–2025 Dietary Guidelines for Americans described three dietary patterns (a healthy US-style eating pattern, vegetarian pattern, and Mediterranean-style pattern) which offer greater flexibility in food choices across a set of shared core elements (vegetables, fruits, grains, low-fat dairy, protein foods, and oils).(45) Accordingly, our findings add to the growing literature of dietary patterns that are associated with a favorable metabolic risk profile among South Asians. An earlier study conducted among MASALA cohort participants found that a dietary pattern rich in fruits, vegetables, nuts, and legumes (a “prudent” pattern) was associated with lower odds of hypertension and metabolic syndrome as well as lower insulin resistance.(46) A vegetarian diet pattern was associated with lower BMI, fasting glucose, and insulin resistance as well as lower odds of fatty liver.(47) Those in the vegetarian group did not consume seafood or alcohol,(47) which differentiates the diet from the SAM score constructed in the current study (which positively scored both seafood and moderate alcohol intake). Finally, a plant-based eating pattern that specifically emphasized greater intakes of healthy plant foods was found to be associated with lower odds of both T2D and fatty liver.(48) These findings are consistent with our study of a Mediterranean-style diet that also emphasizes healthy plant foods. Taken together, these studies among a US South Asian population provide several options for tailored risk reduction to suit comorbidity profiles among this population.
The strengths of our study deserve comment. First, we used a previously validated ethnic-specific FFQ to construct a Mediterranean-style diet score that is inclusive of foods commonly consumed among the South Asian community. This allowed us to better capture the usual diet of study participants and also has practical implications for supporting the adoption of this dietary pattern in the community. Next, the MASALA study is the only prospective cohort of US South Asians with detailed data on cardiometabolic risk factors (including subclinical atherosclerosis and body composition obtained using imaging methods) and other potential confounders such as a measure of cultural traditions and experiences of discrimination, which allowed us to adjust for a comprehensive list of confounders. However, our study is not without its limitations. As in any observational study, residual confounding is a possibility, although we adjusted for a large number of known and potential confounders. We did not collect data on food preparation methods for all items on the FFQ, which may lead to some misclassification of diet. Further, the FFQ did not collect detailed data on specific oil intake which is of interest to a Mediterranean style diet; however, both MUFA and PUFA intake tended to increase with increasing SAM score. Next, we did not have data on all cardiometabolic risk factors at the follow-up visit. Moreover, we used a baseline assessment of diet in our prospective analysis; therefore, future studies of longer-term habitual intake in this population would be valuable. Finally, our findings can be extrapolated most directly to Asian Indian adults who are living in the US because MASALA included only small proportions of immigrants from Pakistan, Bangladesh, and other countries.
In conclusion, our findings provide the first prospective evidence that a South Asian Mediterranean-style diet is associated with a lower risk of incident diabetes, as well as favourable cross-sectional measures of body composition, glycated hemoglobin, and fatty liver. This dietary pattern may serve as a useful dietary strategy to prevent T2D in this high-risk population.
Supplementary Material
STUDY IMPORTANCE QUESTIONS.
1. What is already known about this subject?
South Asians have a disproportionately high risk for cardiometabolic disease.
The Mediterranean diet is associated with a reduced risk for type 2 diabetes (T2D) and cardiovascular disease in certain populations, but data among diverse groups are scarce.
2. What are the new findings in your manuscript?
Our findings provide the first prospective evidence that a Mediterranean-style diet that incorporates traditional South Asian foods is associated with a lower risk of incident diabetes.
This dietary pattern is also associated with favorable cross-sectional measures of body composition, glycated hemoglobin, and fatty liver.
3. How might your results change the direction of research or the focus of clinical practice?
Prior research has underscored the importance of studying and integrating culturally specific foods and dishes into dietary recommendations and lifestyle interventions to facilitate implementation among diverse communities.
Accordingly, this Mediterranean-style dietary pattern that incorporates traditional South Asian foods may serve as a useful dietary strategy to prevent T2D in this high-risk population.
Acknowledgements
The authors thank the other investigators, the staff, and the participants of the MASALA study for their valuable contributions.
Funding
The project described was supported by National Institutes of Health grants 5R01HL093009, and 2R01HL093009, and at the UCSF site with grants UL1RR024131 and UL1TR001872.
Dr. Rai was supported by a Doctoral Foreign Study Award from the Canadian Institutes of Health Research.
The funding agencies had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Footnotes
Disclosures
Dr. Bhupathiraju is a scientific consultant to LayerIV for work outside the submitted work.
Prior presentation of data statement
The results of the current study were presented as a poster at the AHA EPI Lifestyle 2020 scientific sessions (May 20–21, 2021), held virtually due to the COVID-19 pandemic.
Data sharing statement
Data described in the article, code book, and analytic code will not be made publicly available. Further information including the procedures to obtain and access data from the MASALA study is described at https://www.masalastudy.org/for-researchers.
REFERENCES
- 1.South Asian Americans Leading Together (SAALT) Strengthening South Asian Communities in America. Demographic snapshot of South Asians in the United States April 2019. https://saalt.org/wp-content/uploads/2019/04/SAALT-Demographic-Snapshot-2019.pdf (accessed 15 March 2021).
- 2.Al Rifai M, Cainzos-Achirica M, Kanaya AM, et al. Discordance between 10-year cardiovascular risk estimates using the ACC/AHA 2013 estimator and coronary artery calcium in individuals from 5 racial/ethnic groups: Comparing MASALA and MESA. Atherosclerosis, 2018. 279: 122–129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kanaya AM, Herrington D, Vittinghoff E, et al. Understanding the high prevalence of diabetes in U.S. south Asians compared with four racial/ethnic groups: the MASALA and MESA studies. Diabetes Care, 2014. 37(6): 1621–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chiu M, Austin PC, Manuel DG, and Tu JV Comparison of cardiovascular risk profiles among ethnic groups using population health surveys between 1996 and 2007. CMAJ, 2010. 182(8): E301–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gujral UP, Pradeepa R, Weber MB, Narayan KMV, and Mohan V Type 2 diabetes in South Asians: similarities and differences with white Caucasian and other populations. Annals of the New York Academy of Sciences, 2013. 1281(1): 51–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rosato V, Temple NJ, La Vecchia C, et al. Mediterranean diet and cardiovascular disease: a systematic review and meta-analysis of observational studies. European Journal of Nutrition, 2019. 58(1): 173–191. [DOI] [PubMed] [Google Scholar]
- 7.Schwingshackl L, Missbach B, König J, and Hoffmann G Adherence to a Mediterranean diet and risk of diabetes: a systematic review and meta-analysis. Public Health Nutrition, 2015. 18(7): 1292–1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Schwingshackl L, Schwedhelm C, Galbete C, and Hoffmann G Adherence to Mediterranean Diet and Risk of Cancer: An Updated Systematic Review and Meta-Analysis. Nutrients, 2017. 9(10): 1063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sofi F, Cesari F, Abbate R, Gensini GF, and Casini A Adherence to Mediterranean diet and health status: meta-analysis. BMJ, 2008. 337: a1344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sotos-Prieto M and Mattei J Mediterranean Diet and Cardiometabolic Diseases in Racial/Ethnic Minority Populations in the United States. Nutrients, 2018. 10(3): 352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kanaya AM, Kandula N, Herrington D, et al. Mediators of Atherosclerosis in South Asians Living in America (MASALA) study: objectives, methods, and cohort description. Clin Cardiol, 2013. 36(12): 713–720. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kelemen LE, Anand SS, Vuksan V, et al. Development and evaluation of cultural food frequency questionnaires for South Asians, Chinese, and Europeans in North America. J Am Diet Assoc, 2003. 103(9): 1178–84. [DOI] [PubMed] [Google Scholar]
- 13.Fung TT, McCullough ML, Newby PK, et al. Diet-quality scores and plasma concentrations of markers of inflammation and endothelial dysfunction. Am J Clin Nutr, 2005. 82(1): 163–73. [DOI] [PubMed] [Google Scholar]
- 14.Kandula NR, Kanaya AM, Liu K, et al. Association of 10-year and lifetime predicted cardiovascular disease risk with subclinical atherosclerosis in South Asians: findings from the Mediators of Atherosclerosis in South Asians Living in America (MASALA) study. Journal of the American Heart Association, 2014. 3(5): e001117–e001117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kanaya AM, Vittinghoff E, Lin F, et al. Incidence and Progression of Coronary Artery Calcium in South Asians Compared With 4 Race/Ethnic Groups. J Am Heart Assoc, 2019. 8(2): e011053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet, 2004. 363(9403): 157–63. [DOI] [PubMed] [Google Scholar]
- 17.Deol R, Lee KA, Kandula NR, and Kanaya AM Risk of Obstructive Sleep Apnoea is Associated with Glycaemia Status in South Asian Men and Women in the United States. Obes Med, 2018. 9: 1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kodama Y, Ng CS, Wu TT, et al. Comparison of CT methods for determining the fat content of the liver. AJR Am J Roentgenol, 2007. 188(5): 1307–12. [DOI] [PubMed] [Google Scholar]
- 19.Ainsworth BE, Irwin ML, Addy CL, Whitt MC, and Stolarczyk LM Moderate physical activity patterns of minority women: the Cross-Cultural Activity Participation Study. J Womens Health Gend Based Med, 1999. 8(6): 805–13. [DOI] [PubMed] [Google Scholar]
- 20.Kanaya AM, Wassel CL, Mathur D, et al. Prevalence and Correlates of Diabetes in South Asian Indians in the United States: Findings From the Metabolic Syndrome and Atherosclerosis in South Asians Living in America Study and the Multi-Ethnic Study of Atherosclerosis. Metabolic Syndrome and Related Disorders, 2009. 8(2): 157–164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Williams DR, Yan Y, Jackson JS, and Anderson NB Racial Differences in Physical and Mental Health: Socio-economic Status, Stress and Discrimination. J Health Psychol, 1997. 2(3): 335–51. [DOI] [PubMed] [Google Scholar]
- 22.Rodrigues YE, Fanton M, Novossat RS, and Canuto R Perceived racial discrimination and eating habits: a systematic review and conceptual models. Nutr Rev, 2022. [DOI] [PubMed]
- 23.Pascoe EA and Smart Richman L Perceived discrimination and health: a meta-analytic review. Psychol Bull, 2009. 135(4): 531–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Reddy S, Anitha M. Culture and its Influence on Nutrition and Oral Health. Biomed Pharmacol J 2015;8(October Spl Edition). [Google Scholar]
- 25.Speliotes EK, Massaro JM, Hoffmann U, et al. Fatty liver is associated with dyslipidemia and dysglycemia independent of visceral fat: the Framingham Heart Study. Hepatology, 2010. 51(6): 1979–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lim S and Meigs JB Ectopic fat and cardiometabolic and vascular risk. Int J Cardiol, 2013. 169(3): 166–76. [DOI] [PubMed] [Google Scholar]
- 27.Kim TH, Yu SH, Choi SH, et al. Pericardial fat amount is an independent risk factor of coronary artery stenosis assessed by multidetector-row computed tomography: the Korean Atherosclerosis Study 2. Obesity (Silver Spring), 2011. 19(5): 1028–34. [DOI] [PubMed] [Google Scholar]
- 28.Rosito GA, Massaro JM, Hoffmann U, et al. Pericardial fat, visceral abdominal fat, cardiovascular disease risk factors, and vascular calcification in a community-based sample: the Framingham Heart Study. Circulation, 2008. 117(5): 605–13. [DOI] [PubMed] [Google Scholar]
- 29.Okamura T, Hashimoto Y, Hamaguchi M, et al. Ectopic fat obesity presents the greatest risk for incident type 2 diabetes: a population-based longitudinal study. International Journal of Obesity, 2019. 43(1): 139–148. [DOI] [PubMed] [Google Scholar]
- 30.Shah AD, Kandula NR, Lin F, et al. Less favorable body composition and adipokines in South Asians compared with other US ethnic groups: results from the MASALA and MESA studies. International journal of obesity (2005), 2016. 40(4): 639–645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lear SA, Chockalingam A, Kohli S, Richardson CG, and Humphries KH Elevation in cardiovascular disease risk in South Asians is mediated by differences in visceral adipose tissue. Obesity (Silver Spring), 2012. 20(6): 1293–300. [DOI] [PubMed] [Google Scholar]
- 32.Garg SK, Lin F, Kandula N, et al. Ectopic Fat Depots and Coronary Artery Calcium in South Asians Compared With Other Racial/Ethnic Groups. J Am Heart Assoc, 2016. 5(11). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Shah AD, Kandula NR, Lin F, et al. Less favorable body composition and adipokines in South Asians compared with other US ethnic groups: results from the MASALA and MESA studies. Int J Obes (Lond), 2016. 40(4): 639–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Flowers E, Lin F, Kandula NR, et al. Body Composition and Diabetes Risk in South Asians: Findings From the MASALA and MESA Studies. Diabetes Care, 2019. 42(5): 946–953. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Singh RB, Dubnov G, Niaz MA, et al. Effect of an Indo-Mediterranean diet on progression of coronary artery disease in high risk patients (Indo-Mediterranean Diet Heart Study): a randomised single-blind trial. Lancet, 2002. 360(9344): 1455–61. [DOI] [PubMed] [Google Scholar]
- 36.Nigam P, Bhatt S, Misra A, et al. Effect of a 6-month intervention with cooking oils containing a high concentration of monounsaturated fatty acids (olive and canola oils) compared with control oil in male Asian Indians with nonalcoholic fatty liver disease. Diabetes Technol Ther, 2014. 16(4): 255–61. [DOI] [PubMed] [Google Scholar]
- 37.Shai I, Schwarzfuchs D, Henkin Y, et al. Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. N Engl J Med, 2008. 359(3): 229–41. [DOI] [PubMed] [Google Scholar]
- 38.Schwarzfuchs D, Golan R, and Shai I Four-Year Follow-up after Two-Year Dietary Interventions. New England Journal of Medicine, 2012. 367(14): 1373–1374. [DOI] [PubMed] [Google Scholar]
- 39.Middleton G, Keegan R, Smith MF, Alkhatib A, and Klonizakis M Brief Report: Implementing a Mediterranean Diet Intervention into a RCT: Lessons Learned from a Non-Mediterranean Based Country. J Nutr Health Aging, 2015. 19(10): 1019–22. [DOI] [PubMed] [Google Scholar]
- 40.Evert AB, Dennison M, Gardner CD, et al. Nutrition Therapy for Adults With Diabetes or Prediabetes: A Consensus Report. Diabetes Care, 2019. 42(5): 731–754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Lawton J, Ahmad N, Hanna L, et al. ‘We should change ourselves, but we can’t’: accounts of food and eating practices amongst British Pakistanis and Indians with type 2 diabetes. Ethn Health, 2008. 13(4): 305–19. [DOI] [PubMed] [Google Scholar]
- 42.Asamane EA, Greig CA, Aunger JA, and Thompson JL Perceptions and Factors Influencing Eating Behaviours and Physical Function in Community-Dwelling Ethnically Diverse Older Adults: A Longitudinal Qualitative Study. Nutrients, 2019. 11(6): 1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Ristovski-Slijepcevic S, Chapman GE, and Beagan BL Engaging with healthy eating discourse(s): ways of knowing about food and health in three ethnocultural groups in Canada. Appetite, 2008. 50(1): 167–78. [DOI] [PubMed] [Google Scholar]
- 44.Kandula NR, Dave S, De Chavez PJ, et al. Translating a heart disease lifestyle intervention into the community: the South Asian Heart Lifestyle Intervention (SAHELI) study; a randomized control trial. BMC Public Health, 2015. 15(1): 1064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.U.S. Department of Agriculture and U.S. Department of Health and Human Services. Dietary Guidelines for Americans, 2020–2025 9th Edition. [Google Scholar]
- 46.Gadgil MD, Anderson CAM, Kandula NR, and Kanaya AM Dietary patterns are associated with metabolic risk factors in South Asians living in the United States. The Journal of nutrition, 2015. 145(6): 1211–1217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Jin Y, Kanaya AM, Kandula NR, Rodriguez LA, and Talegawkar SA Vegetarian Diets Are Associated with Selected Cardiometabolic Risk Factors among Middle-Older Aged South Asians in the United States. J Nutr, 2018. 148(12): 1954–1960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Bhupathiraju SN, Sawicki CM, Goon S, et al. A healthy plant-based diet is favorably associated with cardiometabolic risk factors among participants of South Asian ancestry. Am J Clin Nutr, 2022. 116(4): 1078–1090. [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.
Supplementary Materials
Data Availability Statement
Data described in the article, code book, and analytic code will not be made publicly available. Further information including the procedures to obtain and access data from the MASALA study is described at https://www.masalastudy.org/for-researchers.