Dr. Hood is Associate Professor and Chair, Clinical Foundations Department, Ross University School of Medicine, Barbados, West Indies
Dr. Selfridge is Professor, Clinical Foundations Department, Ross University School of Medicine, Barbados, West Indies
Dr. Hood and Dr. Selfridge report no financial relationships relevant to this field of study.
• Measurements of plasma vitamin C and carotenoids serve as reliable objective surrogates for self-reported consumption of fruits and vegetables.
• There is an inverse association between vitamin C and carotenoid levels and the development of incident type 2 diabetes (T2D).
• Increasing daily servings of fruits and vegetables reduces the risk of T2D regardless of baseline consumption.
• Vitamin supplements alone do not reduce risk.
SYNOPSIS: A multicenter, multinational prospective case-cohort study concludes that an increase in dietary consumption of fruits and vegetables is beneficial in reducing the risk of diabetes mellitus type 2 regardless of the current level of consumption.
SOURCE: Zheng JS, Sharp SJ, Imamura F, et al. Association of plasma biomarkers of fruit and vegetable intake with incident type 2 diabetes: EPIC-InterAct case-cohort study in eight European countries. BMJ 2020;370:m2194.
The estimated risk of developing diabetes continues to increase throughout the world. The average risk of developing type 2 diabetes (T2D) for persons born in the United States in 2000 is 32.8% for males and 38.5% for females. The estimated risk for Hispanic persons born in the United States is 45.4% for men and 52.5% for women, the highest in the country.1 Any effective strategy for risk reduction and prevention has significant public health implications. Citing inconsistent, weak evidence from prospective studies and sparse randomized control trials linking fruit and vegetable intake with T2D risk reduction, Zheng et al aimed to examine the association of baseline levels of circulating vitamin C and carotenoids with incident T2D. The investigators used data collected in the European Prospective Investigation into Cancer and Nutrition (EPIC), an ongoing, multi-center prospective cohort study designed to investigate the links between nutrition and cancer risk in more than 500,000 community-based adult subjects from eight European countries.2
According to the case-cohort study design, and after exclusions for missing or inadequate blood samples, the authors included data from a population of 22,833 subjects from the EPIC-InterAct subcohort (total > 340,000 participants) nested within the EPIC study. The population consisted of 9,754 participants with incident T2D and 13,662 randomly selected sub-cohort participants with an average follow-up of 9.7 years across 26 study centers in eight European countries. Nonfasting blood samples, collected from subjects at the EPIC study's baseline visit, were used to obtain plasma levels of vitamin C, and several carotenoids (alpha-carotene, beta-carotene, lycopene, lutein, zeaxanthin, and beta-cryptoxanthin) were measured via the high-performance liquid chromatography-ultraviolet method.
Additional baseline data included weight, height, and waist circumference. Baseline physical activity, smoking status, sociodemographic factors, and medical history were obtained by study staff from self-administered questionnaires. The self-reported diabetic status of subjects was ascertained by primary and secondary care registers, drug registers, and hospital and mortality records.
Using averages of the standardized values of vitamin C and individual carotenoid levels in the subcohort, a composite biomarker score was calculated and applied to analyses, along with analyses of the effect of individual biomarker levels. The composite biomarker score was correlated with subject self-report of fruit and vegetable intake; differences in total intake were determined for each one standard deviation increase in composite biomarker score using linear regression. Adjustments in all analyses were made for confounding factors, including age, sex, sociodemographic factors (marital status and education), physical activity, smoking, body mass index (BMI), energy intake, alcohol, waist circumference, and diet (cereals, potatoes, soft drinks, legumes, nuts, eggs, fish, red meat, and vitamin supplementation).
Composite biomarker scores were divided into five categories, correlating with lowest consumption of fruits and vegetables (Group 1) to highest consumption (Group 5). These categories then were correlated with incident T2D, continuing to adjust for other confounding risk factors, including obesity, family history of T2D, insulin resistance, cardiovascular disease, cancer, stroke, menopausal status, and diet quality. Additionally, in order to investigate the association between incident T2D and current “five a day” recommendations for fruit and vegetable serving consumption, the authors used composite biomarker levels to identify subjects in two groups: those meeting or exceeding five servings a day (> 400 g) and those not meeting current recommendations.
Results, summarized in Table 1, focused on incident T2D and composite biomarker scores from Groups 2, 3, 4, and 5 compared to Group 1. They were applied in a primary reference statistical model, 1a, adjusting for age, sex, and research center. Two additional models adjusted for confounding risk factors: model 1b, further adjusting for physical activity, smoking, alcohol, total energy intake, sociodemographic factors, high-density lipoprotein and low-density lipoprotein levels; and model 2, additionally adjusting for BMI and adiposity. Although adjustments in these models attenuated the magnitude of the inverse association trends, all remained statistically significant (P < 0.001).
Table 1. Associations Between Composite Biomarker Scores and Incident Type 2 Diabetes
|
|
Hazard Ratio (95% Confidence Interval)
|
| |
Group 1
|
Group 2
|
Group 3
|
Group 4
|
Group 5
|
For Each SD
|
P Value
|
|
Median daily fruit and vegetable intake (g)
|
274
|
357
|
396
|
452
|
508
|
|
|
|
Median composite biomarker score
|
-0.66
|
-0.31
|
-0.05
|
0.23
|
0.74
|
|
|
|
Type 2 diabetes cases/person years
|
2,752/10,909
|
1,719/13,249
|
1,249/14,624
|
1,047/15,582
|
770/17,471
|
|
|
|
Model 1a
|
1.0 (reference)
|
0.61
(0.53-0.70)
|
0.43
(0.35-0.53)
|
0.32
(0.24-0.43)
|
0.22
(0.17-0.30)
|
0.55
(0.48-0.63)
|
< 0.001
|
|
Model 1b
|
1.0 (reference)
|
0.67
(0.59-0.77)
|
0.52
(0.44-0.62)
|
0.42
(0.33-0.52)
|
0.31
(0.25-0.39)
|
0.61
(0.55-0.69)
|
< 0.001
|
|
Model 2
|
1.0 (reference)
|
0.77
(0.68-0.87)
|
0.66
(0.54-0.80)
|
0.50
(0.40-0.62)
|
0.50
(0.40-0.62)
|
0.75
(0.67-0.83)
|
< 0.001
|
|
SD: standard deviation; model 1a: age, sex, and research center; model 1b: 1a plus physical activity, smoking, employment, marital status, education, alcohol intake, total energy intake, and high-density lipoprotein and low-density lipoprotein levels; model 2: 1b plus adiposity (body mass index and waist circumference)
|
Subjects with incident T2D had lower mean concentrations of plasma vitamin C and total carotenoids compared to the subcohort population. There was an inverse association between incident T2D and composite biomarker scores, as well as total vitamin C and total carotenoids.
Additionally, an inverse association with incident T2D was noted for all individual plasma carotenoids, except for zeaxanthin. The hazard ratios comparing Groups 2, 3, 4, and 5 of the composite biomarker score with Group 1 were 0.77 (95% confidence interval [CI], 0.68-0.87), 0.66 (95% CI, 0.54-0.80), 0.59 (95% CI, 0.48-0.72), and 0.50 (95% CI, 0.40-0.62), respectively, for model 2, which included adjustments for all confounding risk factors. Analysis comparing “five or more servings per day” to fewer than five servings daily and incident T2D resulted in a hazard ratio of 0.69 (95% CI, 0.63-0.76). A single standard deviation difference in the composite biomarker score, equivalent to approximately 66 g difference in daily fruit and vegetable intake, was associated with a hazard ratio of 0.75 (95% CI, 0.67-0.83; P < 0.001).
COMMENTARY
The strengths of this study include its case-cohort design, the large number of subjects with complete data sets from the EPIC database, the use of quantitative measures of fruit and vegetable consumption biomarkers that correlated well with subject self-report of daily intake, and the care the researchers took to statistically adjust dietary association results for other confounding risk factors for incident T2D. This study supports the current nutrition recommendation by several national and international organizations (American Heart Association, United States Department of Agriculture, Centers for Disease Control and Prevention, World Health Organization, et al) to consume five total servings of fruits and vegetables daily. It also provides evidence for encouraging patients with low intake to add daily servings toward meeting these recommendations, with each increase of about 66 g in daily consumption of fruits and vegetables appearing to significantly reduce the risk of developing T2D.
For patients who adhere to this recommendation, evidence exists for further risk reduction when consumption exceeds the five recommended servings per day. Further, as noted in previous studies, vitamin supplements do not have the same associations with risk reduction as whole fruits and vegetables. Thus, it is unlikely that these vitamins identified as biomarkers for fruits and vegetables are solely bio-physiologically responsible for risk-mitigating effects.
How can clinicians use this information? Patients can be confidently encouraged to increase their intake of fruits and vegetables and assured that even small increments make a difference. The 66-g increment associated with a standard deviation in biomarker plasma level amounts to about half of a medium apple, one medium raw carrot, one cup of chopped broccoli, or one cup of chopped kale. When patients ask if they can just take a vitamin supplement, we can cite this study as ongoing evidence that the value of whole food is much more than the vitamins contained therein. The common wisdom reflected in author Michael Pollan’s aphorism, now more than a decade old, rings loud and clear in these study results: “Eat food, not too much, mostly plants.”3
REFERENCES
- Narayan KMV, Boyle JP, Thompson TJ, et al. Lifetime risk for diabetes mellitus in the United States. JAMA 2003;290:1884-1890.
- Riboli E, Hunt KJ, Slimani N, et al. European Prospective Investigation in Cancer and Nutrition (EPIC): Study populations and data collection. Public Health Nutr 2002;5:1113-1124.
- Pollan M. In Defense of Food: An Eater’s Manifesto. Penguin Press;2009.
