Adherence to a traditional Mexican diet and non-communicable disease-related outcomes: secondary data analysis of the cross-sectional Mexican National Health and Nutrition Survey

This study evaluated the association between adherence to a traditional Mexican diet (TMexD) and obesity, diabetes and CVD-related outcomes in secondary data analysis of the cross-sectional Mexican National Health and Nutrition Survey 2018–2019. Data from 10 180 Mexican adults were included, collected via visits to randomly selected households by trained personnel. Adherence to the TMexD (characterised by mostly plant-based foods like maize, legumes and vegetables) was measured through an adapted version of a recently developed TMexD index, using FFQ data. Outcomes included obesity (anthropometric measurements), diabetes (biomarkers and diagnosis) and CVD (lipid biomarkers, blood pressure, hypertension diagnosis and CVD event diagnosis) variables. Percentage differences and OR for presenting non-communicable disease (NCD)-related outcomes (with 95 % CI) were measured using multiple linear and logistic regression, respectively, adjusted for relevant covariates. Sensitivity analyses were conducted according to sex, excluding people with an NCD diagnosis and using multiple imputation. In fully adjusted models, high, compared with low, TMexD adherence was associated with lower insulin (−9·8 %; 95 % CI (−16·0, −3·3)), LDL-cholesterol (−4·3 %; 95 % CI (−6·9, −1·5)), non-HDL-cholesterol (−3·9 %; 95 % CI (−6·1, −1·7)) and total cholesterol (−3·5 %; 95 % CI (−5·2, −1·8)) concentrations. Men and those with no NCD diagnosis had overall stronger associations. Effect sizes were smaller, and associations weakened in multiple imputation models. No other associations were observed. While results may have been limited due to the adaptation of a previously developed index, the results highlight the potential association between the TMexD and lower insulin and cholesterol concentrations in Mexican adults.

Not included (not evaluated in food frequency questionnaire) a Toasted or fried tortillas or tortilla pieces. b Item not measured in food frequency questionnaire used. c Traditional process where maize dough has been soaked in an alkaline solution, cooked, and dried. d Hot beverage prepared with maize dough. e Water blended with fruit/flowers, with or without sugar. f This recommended quantity did not reach the consensus in the study conducted to develop the index, it was selected based on plurality of votes (i.e., agreement by a large portion of the sample but less than 50%). g Using the standard portion of processed meats (30 g). h Type of curd cheese, like fresh cheese. i Type of maize-based Mexican snack. j Soup made with maize kernels, meat, chile, and seasonings. k Includes maize-based Mexican snacks like sopes, quesadillas, tlacoyos, gorditas, enchiladas. l Includes maize-based Mexican snacks like tacos, quesadillas, tlacoyos, enchiladas, gorditas. m Refers to a traditional open-air market, which occurs on certain days of the week. Table S2. Foods measured in the original traditional Mexican diet index but omitted in the present study, as these are not measured in the Mexican National Health and Nutrition Survey
b High adherence reflects individuals with higher scores in the traditional Mexican diet index.
* Model 1: unadjusted model. † Model 2: adjusted for age, sex, socioeconomic status, education level, region of the country, area of residence, physical activity, smoking. Diabetes, blood lipid, and blood pressure outcomes were additionally adjusted for family history of disease and use of medication. ‡ Model 3: model 2 plus total energy intake. § Model 4: model 3 plus overweight/obesity status (≥25 kg/m2).
|| Significance assessed at P < 0.004 using the Bonferroni correction. ¶ Percent of variance explained by the model. Table S4. Percentage difference in non-communicable disease-related outcomes a in the highest tertile versus the lowest tertile of adherence b to the traditional Mexican diet, in adults without an NCD diagnosis and adults not dieting after an NCD diagnosis.

Participants without NCD diagnosis Participants not dieting after NCD diagnosis
Model NCD, non-communicable disease; CI, confidence interval; NA, non-applicable; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.
a All analyses were conducted through multiple linear regression.
b High adherence reflects individuals with higher scores in the traditional Mexican diet index.
* Model 1: unadjusted model. † Model 2: adjusted for age, sex, socioeconomic status, education level, region of the country, area of residence, physical activity, smoking. Diabetes, blood lipid, and blood pressure outcomes were additionally adjusted for family history of disease and use of medication. ‡ Model 3: model 2 plus total energy intake. § Model 4: model 3 plus overweight/obesity status (≥25 kg/m2).
|| Significance assessed at P < 0.004 using the Bonferroni correction. ¶ Percent of variance explained by the model.  a All analyses were conducted through multiple linear regression.
b High adherence reflects individuals with higher scores in the traditional Mexican diet index.
* Model 1: unadjusted model. † Model 2: adjusted for age, sex, socioeconomic status, education level, region of the country, area of residence, physical activity, smoking. Diabetes, blood lipid, and blood pressure outcomes were additionally adjusted for family history of disease and use of medication. ‡ Model 3: model 2 plus total energy intake. § Model 4: model 3 plus overweight/obesity status (≥25 kg/m2).
|| Significance assessed at P < 0.004 using the Bonferroni correction. Table S6. Odds ratio for having non-communicable disease-related outcomes a in adults in the highest tertile versus the lowest tertile of adherence b to the traditional Mexican diet, women compared to men.

Women Men
Model e Defined as having a previous medical diagnosis of heart attack, angina, or heart failure; total number of cases: 332.
* Model 1: unadjusted model. † Model 2: adjusted for age, sex, socioeconomic status, education level, region of the country, area of residence, physical activity, smoking. Diabetes, blood lipid, and blood pressure outcomes were additionally adjusted for family history of disease and use of medication. ‡ Model 3: model 2 plus total energy intake. § Model 4: model 3 plus overweight/obesity status (≥25 kg/m2).
|| Significance assessed at P < 0.004 using the Bonferroni correction. Table S7. Odds ratio for having non-communicable disease-related outcomes a in the highest tertile versus the lowest tertile of adherence b to the traditional Mexican diet, in adults without an NCD diagnosis and adults not dieting after an NCD diagnosis.

Participants without NCD diagnosis Participants not dieting after NCD diagnosis
Model e Defined as having a previous medical diagnosis of heart attack, angina, or heart failure; total number of cases: 332.
* Model 1: unadjusted model. † Model 2: adjusted for age, sex, socioeconomic status, education level, region of the country, area of residence, physical activity, smoking. Diabetes, blood lipid, and blood pressure outcomes were additionally adjusted for family history of disease and use of medication. ‡ Model 3: model 2 plus total energy intake. § Model 4: model 3 plus overweight/obesity status (≥25 kg/m2).
|| Significance assessed at P < 0.004 using the Bonferroni correction. Table S8. Odds ratio for having non-communicable disease-related outcomes a in 10,087 adults in the highest tertile versus the lowest tertile of adherence b to the traditional Mexican diet, data analysed with multiple imputation. e Defined as having a previous medical diagnosis of heart attack, angina, or heart failure; total number of cases: 332.