Vegetarian diet and diabetes risk

In our prospective analysis, both consuming a vegetarian diet and switching to a vegetarian diet are associated with substantial reduction in risk of diabetes. This trend is consistent across sex, baseline BMI categories, metabolic syndrome, impaired fasting glucose, and HDL-C statuses. To the best of our knowledge, this is the first prospective study that examines the impact of consuming a vegetarian diet and switching to a vegetarian diet on diabetes risk.

Plant based dietary patterns and diabetes

The magnitude of protective effect of a vegetarian diet in our study is comparable to the Adventist Health Study – 2 ⁽³⁹⁾, and consistent with those reported in US nurses and health

professionals ⁽¹⁰⁴⁾. All these studies showed dose-dependent protective effect with increasing degree of plant based diet in conjunction with decreasing animal based foods, independent of BMI. Vegetarians in our cohort consumed more whole grains and vegetables than

nonvegetarians, and these may protect against diabetes through higher fiber and magnesium ⁽¹⁴²⁾. In addition, soy is a major source of protein for Taiwanese vegetarians, and soy has been shown to improve insulin resistance when replacing meat in randomized controlled trials ^(83,84). Increase in soy and legume consumption is inversely associated with risk of diabetes in a Chinese cohort

(102). A vegetables-fruits-soy dietary pattern is also inversely associated with diabetes incidence

in Singaporean Chinese ⁽¹⁰¹⁾.

Meat and diabetes risk

Although the protective effect is likely caused by various plant components, it may also be influenced by the simultaneous elimination of meat. Meat is high in saturated fats, and saturated

fat have been shown to trigger human β-cell apoptosis ⁽¹⁸⁾. Fatty acids from meat have also been adversely associated with insulin secretion, and Disposition Index (β cell function accounting

for insulin sensitivity) ⁽⁸⁹⁾. In our study, meat consumption is associated with a nonsignificant increase in diabetes risk (per 30g serving of meat: HR = 1.15, 95% CI: 0.90 – 1.46). The statistical insignificance may be related to small sample size.

Red meat and processed meat appear to be the most diabetogenic^(188,189), whereas the role of

other types of meat is less clear. Heme iron found in red meat and nitrites from processed meat may exacerbate insulin resistance and damage β-cell(90,91,159,160).

However, the effect of meat on diabetes is equivocal among Asian women: meat was not associated with diabetes risk in Japanese women ⁽¹⁹⁰⁾, and was associated with protection among normal weight Chinese women in Shanghai ⁽¹⁹¹⁾. None of these studies actually included a diet range of complete meat avoidance, and it is possible that even the lowest quantile in these cohorts did not consume low enough meat (and high enough healthy plant foods) to observe maximum protection. Figure 5 – 3 demonstrates a potential non-linear association between meat intake and diabetes risk. There may be a threshold of risk (triangle) above which, risk of

diabetes increases (range of meat intake of Western populations). On the other end of spectrum, there may be a turning point of accelerated protection (star), below which risk drastically

reduces (a range of meat intake our study, TCHS). Many Asian populations may have diet range in-between the threshold of risk and the turning point, and thus unable to detect the meat – diabetes association.

Figure 5 – 3. Potential nonlinear relationship between meat consumption and diabetes risk. TCHS = Tzu Chi Health Study (current study).

The inverse association between meat intake and diabetes in the Shanghai Women’s Health Study may be confounded by unmeasured social economic factor and possibly early life food insecurity. Those in the lowest quintile differ greatly from the highest quintiles in education (37% vs 10% with no eduction), income (21.5% vs 13.7% with income <10000), occupation (63% vs 37% retired or housewife), and were on average 5 years older⁽¹⁹¹⁾, suggesting they may have come from different social economic classes and from different birth cohorts, possibly implying different degree of exposure to famine in early life. Early life undernutrition could trigger epigenetic changes to induce diabetes risk ⁽¹⁹²⁾.

Interaction between meat and metabolic risk factors

In the Shangahi Women’s Health Study ⁽¹⁹¹⁾ and in Japanese Americans within the

Multiethnic Cohort ⁽¹⁹⁰⁾, effect of meat or meat-fat dietary pattern appeared to be modified by BMI status, where meat is associated with diabetes risk among those with higher BMI, but not those with normal BMI. In our study, vegetarian diet is protective across statuses of BMI, metabolic syndrome, HDL-C, and impaired fasting glucose. However, when stratified by fatty liver status, the protective effect of vegetarian diet appears to be more protective among those with fatty liver at baseline. It is possible that the insulin sensitizing effect of a vegetarian diet helps ameliorate insulin resistance associated with fatty liver, thereby lowering risk of diabetes.

Although our cohort may be too small to detect significant interaction, it is possible that the effect of vegetarian diet is not due solely to either the minimization of animal product or the higher functional plant ingredient, but the combined effect of both (more discussion later).

Fish and diabetes

Fish and sea food intake has been shown to increase risk for diabetes in American

populations, but decrease risk for some Asian populations in previous meta-analyses of cohort studies^(193,194). The Singapore Chinese Health Study found that it is the plant omega-3 (ALA), not the marine omega-3 (which corresponds to fish intake), that exert the protective effect for diabetes⁽¹⁹⁵⁾. A Japanese cohort found the protective effect of fish only in men, not women⁽¹⁹⁶⁾. In our study, fish consumption was associated with marginal increase in diabetes risk (per 30g increase in fish intake: HR: 1.17, 95% CI: 1.00 – 1.37) among those who did not change dietary

patterns (excluded the reverted and the converted).

Fish is known to be contaminated with mercury and other contaminants in Taiwan and abroad^(197,198), and vegetarians living in contaminated area in Taiwan have been found to have lower blood level of dioxin compared with their nonvegetarian counterparts⁽⁵⁾. The lower

exposures to these environmental toxins may reduce insulin resistance and lessen the damage to β-cell function and thereby protect against diabetes^(199,200). In addition, a trial showed that while

plant polyphenol improves glucose metabolism, fish omega-3 decreases insulin secretion and postrandial GLP-1 ⁽⁹²⁾.

Eggs and diabetes

We observed a non-significant association between eggs and risk of diabetes. Egg consumption was associated with increased risk of diabetes in Physician’s Health Study I and Women’s Health Study⁽²⁰¹⁾, but not in the Cardiovascular Health Study that enrolled those≧65

years old⁽²⁰²⁾. Egg is rich in cholesterol and choline. Egg yolk-enriched high cholesterol diet has been shown to increase in plasma glucose in rats⁽²⁰³⁾. Choline may be metabolized to produce trimethylamine N-oxide (TMAO) via intestinal microbes and liver⁽²⁰⁴⁾, and higher TMAO has been associated diabetes⁽²⁰⁵⁾. More research is needed in this topic.

Conversion to vegetarian diet

Our finding that the converted experienced a strong protection also suggests that diabetes risk or protection may be influenced by recent diets. In US nurses and health professionals, increase in red meat consumption over 4 years has been associated with diabetes risk, independent of baseline red meat intake and BMI ⁽¹⁸⁸⁾. Trials using vegetarian diet had also observed improvement in glycemic control in weeks ⁽¹⁴⁵⁾. Switching to a complete plant based diet can increase intestinal microbes that ferment fiber to produce butyrate in a matter of days

(72). Butyrate may induce incretin secretion, contributing to β-cell proliferation ⁽⁷⁰⁾. Microbiome

screening showed F. prausnitizii (a butyrate producing bacteria) to be low in diabetes ^(206,207) and high in vegetarians ⁽²⁰⁸⁾, suggesting a potential diet-microbiome-disease link.