Ascorbic acid (vitamin C), a water-soluble antioxidant vitamin, plays a pivotal role in several biologic processes, including readily scavenging reactive oxygen and nitrogen species. Accumulating evidence from animal and human studies suggests that ascorbic acid could also play a role in glucose homeostasis by preventing abnormal insulin secretion attributable to pancreatic β-cell damage from oxidative stress,1,2 inhibiting hyperglycemia-induced glycation of insulin in the pancreatic β-cell,3 and improving insulin transport and sensitivity.4–6
A relatively large body of evidence supports the thesis that dietary antioxidants, particularly vitamin C, could have the capacity to influence glucose tolerance, suggesting that modification of diet can reduce the prevalence and incidence of type 2 diabetes.7–10 For instance, Feskens et al.,9 in a prospective study of 338 middle-aged Finnish and Dutch men, reported that dietary intake of vitamin C was inversely related to glucose tolerance measured after a 20-year follow-up period.
The role of dietary factors in the etiology of gestational diabetes mellitus has been given very limited attention, despite the fact that there are considerable overlaps between gestational diabetes and type 2 diabetes risk factors.11,12 Furthermore, it has been long known that women with a history of gestational diabetes, as compared with their counterparts without such a reproductive history, are at increased risk of developing type 2 diabetes overall.13,14
In a case-control study of 67 women with gestational diabetes and 260 controls, we observed that maternal intrapartum plasma ascorbic acid concentrations were 31% lower for cases than controls on average (mean ± standard error = 36.5 ± 2.0 vs. 53.0 ± 1.0 μmol/L). We also found that women with plasma ascorbic acid concentrations below 42.6 μmol/L experienced a 13-fold (95% CI = 3.5–46.2) increased risk of gestational diabetes as compared with women with values above 63.3 μmol/L.15 However, inferences from case-control studies are hampered by possible recall bias of self-reported dietary vitamin C intake and an ambiguous temporal relationship between plasma ascorbic acid concentrations and gestational diabetes. We therefore sought to further test our hypothesis by using available information and plasma specimens from an ongoing prospective cohort study of women initiating prenatal care before 16 weeks gestation. Our a priori hypothesis was that lower maternal plasma ascorbic acid concentrations in early pregnancy are associated with a subsequent increased risk of gestational diabetes.
The OMEGA Study
The population for this present analysis was drawn from participants of the ongoing OMEGA Study, a prospective cohort study of maternal dietary risk factors of preeclampsia. The study population comprises women attending prenatal care clinics affiliated with Swedish Medical Center (Seattle) and Tacoma General Hospital (Tacoma), both in Washington State. We began recruiting for the study in December 1996. Women eligible for inclusion into the study were those who initiated prenatal care before 16 weeks gestation. Women were ineligible if they were younger than 18 years of age, did not speak and read English, did not plan to carry the pregnancy to term, did not plan to deliver at either of the 2 research hospitals, or were past 16 weeks gestation. We primarily approached and enrolled nulliparous women.
Enrolled participants were invited to participate in a 45- to 60-minute interview during which trained research personnel used a structured questionnaire to elicit information regarding maternal sociodemographic characteristics and lifestyle habits, as well as medical and reproductive histories. Participants were also provided with a 121-item semiquantitative food frequency questionnaire (FFQ) and a self-addressed stamped envelope, along with instructions for completing and returning the questionnaire to our research offices. Nonfasting blood and urine samples were collected in early pregnancy. Samples were processed and stored in continuously monitored freezers until biochemical analyses were performed in batches. Pregnancy outcome information was ascertained by reviewing participants’ hospital labor and delivery medical records or clinic records after the estimated delivery date. The procedures used in this study were approved by the Institutional Review Boards of Swedish Medical Center and Tacoma General Hospital. All participants provided written informed consent.
The study population for this report is derived from information and specimens collected from those participants who enrolled in the OMEGA study from 1996 through 2000. During this period, 1096 eligible women were approached, and 898 (approximately 82%) agreed to participate. Interviews were completed for 99% (n = 885) of enrolled participants, and 94% (n = 841) of participants provided blood and urine samples. Women who experienced a spontaneous abortion or other fetal demise or who had an induced abortion were excluded (n = 24). Also excluded were those women for whom the outcome of pregnancy was unknown because the woman moved, delivered elsewhere, or was missing medical records (n = 47). An additional 15 women with pregestational diabetes mellitus were excluded. Hence, a cohort of 755 women remained for analysis.
Blood Collection, Processing, and Storage and Ascorbic Acid Analysis
Maternal nonfasting blood samples were collected in EDTA 10-mL Vacutainer tubes at 13 weeks gestation, on average (interquartile range = 11–16 weeks gestation). Samples were protected from ultraviolet light, kept on wet ice, and processed within 20 minutes of phlebotomy. The median time between participants’ last meal and phlebotomy was 2.5 for both cases and controls. Plasma decanted into cryovials was preserved with metaphosphoric acid/dithiothreitol solution and frozen at −80°C until analysis. Laboratory analyses were performed within 24 months of blood collection. Ascorbic acid concentrations have been shown to be stable for more than 57 months under these storage conditions.16,17 Plasma ascorbic acid (ie, total ascorbate) concentrations were determined on a Roche Cobas (Branchburg, NJ) Mira Plus Chemistry Analyzer using a colorimetric procedure described previously.18 The intra- and interassay coefficients of variation for the assay used were both less than 10%. All assays were performed without knowledge of pregnancy outcome.
Maternal and infant medical records were reviewed approximately 7 to 9 months after participants were enrolled in the study to collect detailed information concerning antepartum, labor and delivery characteristics, and conditions of the newborn. From medical records, we abstracted laboratory results from participants’ 50-g 1-hour oral glucose tolerance tests and from the diagnostic 100-g 3-hour oral glucose tolerance test (OGTT). Women were classified as having a pregnancy complicated by gestational diabetes if results from their diagnostic test met the then-current National Diabetes Data Group (NDDG) criteria.19 Briefly, women were classified as having gestational diabetes mellitus if 2 or more of the following plasma glucose concentrations were abnormal according to these criteria: fasting ≥105 mg/dL, 1-hour ≥190 mg/dL, 2-hour ≥165 mg/dL, 3-hour ≥145 mg/dL.19
Using structured questionnaires, interviewers collected information on maternal sociodemographic, behavioral, and medical characteristics. Covariate information included maternal age, height, prepregnancy weight, reproductive and medical histories, and medical histories of first-degree family members. We also collected information on maternal educational attainment, annual household income, occupation and prenatal vitamin supplementation use, as well as smoking and alcohol consumption before and during pregnancy. Maternal age (in years) at the time of interview was determined at the time of interview. Parity was reported as the number of previous pregnancies lasting beyond 20 weeks gestation. All participants reported their maximum height and weight 3 months before the index pregnancy. Maternal race and educational attainment were based on self-reports made during the interview. Prepregnancy body mass index (BMI), used as a measure of overall maternal adiposity, was calculated as weight (kg) divided by height (m) squared. Detailed information about maternal habitual dietary intake during the periconceptional period and early pregnancy was ascertained at 12 weeks gestation, on average, using the self-administered, 121-item semiquantitative food frequency questionnaire developed for the Women's Health Initiative Clinical Trial.20 Food composition values for vitamin C and other nutrients were obtained from the University of Minnesota Nutrition Coding Center nutrient database (Nutrition Coordinating Center, Minneapolis, MN).21
Continuous variables are presented as mean ± standard error of mean (SE). Women were grouped according to quartiles determined by the distribution of plasma ascorbic acid concentrations among the entire cohort. We also created a dichotomous variable for dietary vitamin C intake using the Institute of Medicine estimated average requirement for pregnant women (<70 vs. ≥70 mg/day) as the cutoff value.22 We examined frequency distributions of maternal sociodemographic characteristics and medical and reproductive histories according to plasma ascorbic acid concentrations. We fitted generalized linear models, using a log-link function, to derive risk ratios (RRs) and 95% confidence intervals (CIs).23,24 We also explored the possibility of a nonlinear relation between plasma ascorbic acid concentrations and diabetes risk using generalized additive modeling (GAM) procedures25 by using S-PLUS (version 6.1; Insightful Corp., Seattle, WA). In multivariate analyses, we evaluated linear trends in risk by treating the 4 quartiles of plasma ascorbic acid concentrations as a continuous variable after assigning a score (1, 2, 3, and 4) as its value.26 We considered the following covariates as possible confounders in this analysis: maternal age, race, age, smoking during pregnancy, first-degree family history of type 2 diabetes, prepregnancy adiposity, and annual household income. We adjusted dietary vitamin C consumption for total energy intake using the procedures previously described.27 To assess confounding, we entered variables into a generalized linear model one at a time, and then compared the adjusted and unadjusted risk ratios. Final generalized linear models included covariates that altered unadjusted risk ratios by at least 10%, as well as those covariates of a priori interest (eg, advanced maternal age and prepregnancy adiposity). Whenever appropriate, we used the most parsimonious variable specification that achieved the greatest control of confounding. For instance, to control potential confounding from prepregnancy BMI, we expressed prepregnancy BMI as a continuous variable, categorical variable (<20.0, 20.0–24.9, 25.0–29.9, >30.0 kg/m2) and grouped linear variable, respectively, in the regression model. A grouped linear variable for prepregnancy BMI is one in which the 4 categories are given a score (eg, 1, 2, 3, 4) and then is modeled as a continuous variable.26 The greatest amount of control for confounding was achieved when BMI was expressed as a grouped linear variable. All reported P values are 2-tailed, and confidence intervals were calculated at the 95% level.
Analyses were also conducted to examine the extent to which an observed association between maternal plasma ascorbic acid and gestational diabetes risk would change depending on the diagnostic criteria used to define diabetes case status. For this analysis, women were classified as having gestational diabetes if 2 or more of the following plasma glucose concentrations obtained during the 100-g 3-hour OGTT test were abnormal according to the NDDG criteria: fasting ≥95 mg/dL, 1-hour ≥180 mg/dL, 2-hour ≥155 mg/dL, and 3-hour ≥140 mg/dL.28 Application of this less stringent criteria resulted in the classification of an additional 8 women as developing diabetes; hence, 41 diabetes cases were available for these analyses.
Table 1 shows the characteristics of members of the study cohort according to quartiles of plasma ascorbic acid concentrations. Overall, participants included in this analysis were primarily white, well-educated, and employed during pregnancy. Over 97% of participants reported taking multivitamins during pregnancy. Mean maternal plasma ascorbic acid concentrations varied approximately 2-fold between the highest and lowest quartiles. Comparing the extreme quartiles, women with plasma ascorbic acid concentrations in the lowest quartile were more likely to be Asian, multiparous, and obese, and to have lower daily dietary intake of vitamin C and a family history of type 2 diabetes mellitus. Maternal age, marital status, educational attainment, as well as calen-dar month and gestational age at blood collection did not vary appreciably across quartiles of plasma ascorbic acid concentrations.
Approximately 4% of the women (33 of 755) in this cohort developed gestational diabetes mellitus. Plasma ascorbic acid concentrations were inversely associated with the risk of diabetes (P for trend = 0.023). Women with plasma ascorbic acid concentrations less than 55.9 μmol/L (lowest quartile) experienced a 3.4- fold increased risk of diabetes (CI = 1.2–9.2) as compared with women whose plasma concentrations were 74.6 μmol/L or higher (highest quartile; Table 2). The association was only slightly attenuated after adjusting for maternal age, race, parity, prepregnancy adiposity, family history of type 2 diabetes, and annual household income (relative risk [RR] = 3.1; CI = 1.0–9.7). This association was not confounded by maternal dietary fiber intake.
To further evaluate the association between gestational diabetes risk and extremely low plasma ascorbic acid concentrations, we restricted the analysis to those women with concentrations falling in the lowest decile of the cohort distribution (75 subjects, of whom 10 developed diabetes). For this analysis, women with ascorbic acid concentrations in the highest quartile were used as the referent group. Women with extremely low plasma ascorbic acid concentrations (<48.3 μmol/L) experienced a 4.3-fold increased risk of gestational diabetes (95% CI = 1.2–14.9) as compared with women with concentrations in the upper quartile (data not shown).
We modeled the risk of gestational diabetes in relation to maternal plasma ascorbic acid concentrations expressed as a continuous variable using a generalized additive model (GAM). From these analyses, we noted an approximately inverse linear relation between the log-odds of diabetes risk and plasma ascorbic acid (Fig. 1A), particularly among those 708 women (94% of the cohort) with concentrations between 40 and 100 μmol/L (Fig. 1B). On the basis of this observation, we modeled plasma ascorbic acid concentrations expressed as a continuous variable in a generalized linear model after restricting the study population to 708 subjects with plasma ascorbic acid concentrations between 40 and 100 μmol/L. In this subgroup analysis, we found that a 10-μmol/L increase in plasma ascorbic acid was associated with a 42% reduction in diabetes risk (adjusted RR = 0.58; CI = 0.40–0.83).
We evaluated the relation between low maternal dietary vitamin C consumption as defined by the Institute of Medicine22 (<70 mg daily) and risk of diabetes (Table 3). In this population, 21% (ie, 153 of 729 women who completed the FFQ) of women reported daily dietary vitamin intake that was below 70 mg. Women whose daily vitamin C intake was <70 mg, as compared with those consuming more vitamin C, experienced a 2.2-fold increased risk of diabetes (95% CI = 1.0–4.6). This association was attenuated slightly after adjusting for total energy intake, maternal age, race, parity, prepregnancy adiposity, family history of type 2 diabetes, and annual household income (odds ratio [OR] = 1.8; CI = 0.8–4.4).
Inferences regarding the relation between maternal plasma ascorbic acid concentrations and gestational diabetes risk were largely similar in direction and magnitude when analyses were repeated using the less stringent diagnostic criteria that is currently being advocated by the NDDG28 (Table 4).
In this prospective cohort study of pregnant women, we observed an inverse relation between plasma ascorbic acid concentrations and incidence of gestational diabetes mellitus. Women with extremely low plasma ascorbic acid concentrations (ie, <48.3 μmol/L, the lowest decile) experienced a 4.3-fold increased risk of gestational diabetes (95% CI = 1.2–14.9) as compared with women with concentrations ≥74.6 μmol/L. This inverse association was independent of established sociodemographic risk factors of gestational diabetes such as maternal age, race, prepregnancy adiposity, parity, family history of type 2 diabetes, and annual household income. Women who consumed less than 70 mg per day of vitamin C (below the Institute of Medicine's22 average requirement for pregnant women) experienced a 1.8-fold increased risk of diabetes.
Vitamin C is an important component of the antioxidant defense against oxidative stress because it readily scavenges reactive oxygen and nitrogen species. A growing body of evidence from animal and human studies suggests that antioxidants, including vitamin C, could also play a part in glucose homeostasis through several biologic processes. For instance, diets high in vitamin C could favorably influence glucose homeostasis by preventing pancreatic β-cell cellular damage from oxidative stress secondary to the progressive increase in hyperglycemia.2 Some investigators have noted that vitamin C could play a role in maintaining insulin sensitivity and alleviating insulin resistance by increasing biologic potency of circulating insulin through inhibiting hyperglycemia-induced glycation of insulin in pancreatic β-cells.3 Others have noted that vitamin C improves insulin action by ameliorating the impact of oxidative stress on endothelium-dependent vasodilation and the transport of insulin.4–6
Our findings are generally consistent with a relatively large body of literature documenting inverse associations of plasma ascorbic acid and/or vitamin C intake with hyperglycemia,29 insulin resistance,30,31 glucose intolerance,9 and type 2 diabetes32 in nonpregnant women and men. For instance, Sargeant et al.29 reported lower mean plasma ascorbic acid concentrations in patients with diabetes as compared with normoglycemic controls (men: 41.8 vs. 46.2 μmol/L; women: 49.0 vs. 57.9 μmol/L). The authors also noted that glycemic control (as indicated by HbA1c concentrations) was inversely related with plasma ascorbic acid concentrations in men and women, respectively. These results are in agreement with those from large observational and intervention studies (conducted in ethnically and geographically diverse populations) that demonstrate inverse associations between vegetable and fruit intake (major dietary sources of vitamin C) and risk of type 2 diabetes.10,33,34
There are no published reports concerning the risk of gestational diabetes mellitus in relation to maternal antioxidant status or dietary intake. Given the considerable overlap between gestational diabetes and type 2 diabetes risk factors11 and the growing body of evidence supporting the thesis that antioxidants, particularly vitamin C, could play a role in glucose homeostasis, we analyzed available data from a case-control study.15 From that exercise, we found that women who consumed <70 mg of vitamin C daily during pregnancy, as compared with those who consumed more, experienced a 3.7-fold increased risk of gestational diabetes (95% CI = 1.7–8.2). We also noted an inverse relation between diabetes risk and maternal plasma ascorbic acid concentrations. However, as a result of the cross-sectional design of that study, we could not, with certainty, determine the temporal relationship between plasma ascorbic acid and the occurrence of diabetes; nor could we eliminate concerns about selection bias and biased reporting of habitual dietary intake.
The likelihood of bias, particularly selection and recall bias, is minimized in this present study because of the prospective study design and the high follow-up rate. Determination of maternal ascorbic acid status using plasma collected in early pregnancy also served to clarify the temporal relationship between low maternal plasma ascorbic acid concentrations in pregnancy and subsequent risk of gestational diabetes. Nevertheless, several limitations of the present study merit discussion. First, a single measurement of plasma ascorbic acid is not likely to provide a time-integrated measure of maternal vitamin status, and it is likely that some women in our study could have been misclassified with respect to their long-term vitamin C intake. Such errors, however, are unlikely to be conditional on pregnancy outcome. Hence, misclassification of maternal vitamin C intake (or plasma ascorbic acid concentrations) in the present study could have resulted in diluting their associations with diabetes. Second, because of the near-universal use of prenatal multivitamins (>97%) in our study population, we were not able to assess the independent effect of multivitamin use and diabetes risk. Third, although we adjusted for several known and suspected confounders, we cannot exclude the possibility of residual confounding. Fourth, given the relatively small size of the study cohort, estimated effect measures were imprecise, as reflected by their relatively wide 95% confidence intervals. The small sample size also impeded our efforts to examine the risk of diabetes in relation to maternal dietary intake patterns in this study. In an earlier case-control study with a larger number of diabetes cases,15 we observed that the positive association between low vitamin C intake and increased diabetes risk persisted across different subgroups of individuals categorized by high/low consumption of dietary fiber, total carbohydrate, total fat, and vegetable protein. Another concern relates to the misclassification of maternal anthropometric status. Weights reported by participants in other studies, however, have been shown to be valid.35,36 Lastly, the present cohort study focused primarily on nulliparous women, and cohort members were predominantly white, well-educated women. Therefore, the generalizability of results could be limited.
In summary, findings from this prospective study are generally consistent with those found in our previous case-control study of vitamin C and gestational diabetes risk.15 Our findings are also consistent with a much larger body of evidence from experimental, clinical, and epidemiologic investigations of vitamin C in relation to risk of hyperglycemia,28,34 insulin resistance,30,31 glucose intolerance,9,33 and type 2 diabetes. Khaw et al.,37 in a population-based prospective cohort study, reported recently that one additional serving of vitamin C-rich fruits and vegetables to the daily diet can result in as much as a 20-μmol/L increase in plasma ascorbic acid in nonpregnant women and men. Taken together, these converging lines of evidence suggest that current efforts to encourage populations to consume diets rich in fruits and vegetables38 could also benefit pregnant women and result in substantial reductions in the occurrence of gestational diabetes mellitus. However, concerns remain about the paucity of empiric evidence concerning dietary patterns associated with optimal pregnancy outcomes. Confirmation is needed by intervention studies and larger cohort studies with assessments of dietary intake of foods rich in vitamin C, use of dietary supplements, as well as serial measurements of maternal plasma ascorbic acid concentrations, insulin secretion, and glycemic controls. Results from such studies are likely to have practical significance in developing primary prevention efforts for gestational diabetes and for motivating the adoption (particularly by reproductive-aged women) of dietary strategies for which there is already substantial evidence of their effectiveness in preventing type 2 diabetes mellitus.
We are indebted to the participants of the OMEGA Study for their cooperation. We are also grateful for the technical expertise contributed by Zelalem Ashiny, Emilio Beltran, Lauren Braam, Julie Foster, Lois Meryman, Raymond Miller, Chidinma Okoro, Chun-fang Qiu, Kathy Ramsey, and Trudi Witt.
1. Wells WW, Dou CZ, Dybas LN, et al. Ascorbic acid is essential for the release of insulin from scorbutic guinea pig pancreatic islets. Proc Natl Acad Sci U S A
2. Dou C, Xu DP, Wells WW. Studies on the essential role of ascorbic acid in the energy dependent release of insulin from pancreatic islets. Biochem Biophys Res Commun
3. Abdel-Wahab YH, O'Harte FP, Mooney MH, et al. Vitamin C supplementation decreases insulin glycation and improves glucose homeostasis in obese hyperglycemic (ob/ob) mice. Metabolism
4. Paolisso G, D'Amore A, Volpe C, et al. Evidence for a relationship between oxidative stress and insulin action in non-insulin-dependent (type II) diabetic patients. Metabolism
5. Paolisso G, D'Amore A, Balbi V, et al. Plasma vitamin C affects glucose homeostasis in healthy subjects and in non-insulin-dependent diabetics. Am J Physiol
6. Paolisso G, Giugliano D. Oxidative stress and insulin action: is there a relationship? Diabetologia
7. Colditz GA, Manson JE, Stampfer MJ, et al. Diet and risk of clinical diabetes in women. Am J Clin Nutr
8. Will JC, Byers T. Does diabetes mellitus increase the requirement for vitamin C? Nutr Rev
9. Feskens EJ, Virtanen SM, Rasanen L, et al. Dietary factors determining diabetes and impaired glucose tolerance. A 20-year follow-up of the Finnish and Dutch cohorts of the Seven Countries Study. Diabetes Care
10. Williams DE, Wareham NJ, Cox BD, et al. Frequent salad vegetable consumption is associated with a reduction in the risk of diabetes mellitus. J Clin Epidemiol
11. Williams MA, Mittendorf R. Maternal morbidity. In: Goldman MB, Hatch M, eds. Women and Health
. London: Academic Press Inc; 2000:172–181.
12. Pendergrass M, Fazioni E, Defronzo RA. Non-insulin-dependent diabetes mellitus and gestational diabetes mellitus: same disease, another name? Diabetes Rev
13. O'Sullivan JB. Body weight and subsequent diabetes mellitus. JAMA
14. Kjos SL, Buchanan TA. Gestational diabetes mellitus. N Engl J Med
15. Zhang C, Williams MA, Sorensen TK, et al. A case-control study of vitamin C and the risk of gestational diabetes mellitus (GDM)—results from dietary questionnaire and plasma assay. J Reprod Med
16. Comstock GW, Norkus EP, Hoffman SC, et al. Stability of ascorbic acid, carotenoids, retinol, and tocopherols in plasma stored at -70 degrees C for 4 years. Cancer Epidemiol Biomarkers Prev
17. Margolis SA, Davis TP. Stabilization of ascorbic acid in human plasma, and its liquid-chromatographic measurement. Clin Chem
18. Lee W, Roberts SM, Labbe RF. Ascorbic acid determination with an automated enzymatic procedure. Clin Chem
19. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care
20. Patterson RE, Kristal AR, Tinker LF, et al. Measurement characteristics of the Women's Health Initiative food frequency questionnaire. Ann Epidemiol
21. Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient database. J Am Diet Assoc
22. Institute of Medicine, Food and Nutrition Board. Dietary reference intake for vitamin C, vitamin E, selenium and carotenoids. A Report of the Panel on Dietary Antioxidants and Related Compounds, Subcommittees on Upper References Levels of Nutrients and Interpretation and Uses of Dietary References Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes.
Washington, DC: National Academy Press; 2000.
23. Baker RJ, Nelder JA. The Generalized Linear Interactive Modeling System
, release 3.77. Oxford, UK: Oxford Numerical Algorithms Group, 1985.
24. Hardin JW, Hilbe J. Generalized Linear Models and Extensions
. College Station, TX: Stata Press; 2001.
25. Hastie TJ, Tibshirani RJ. Generalized Additive Models
. London: Chapman-Hall; 1990.
26. Rothman KJ, Greenland S. Modern Epidemiology,
2nd ed. Philadelphia: Lippincott-Raven Publishers; 1998;311–316.
27. Willett W. Nutritional Epidemiology
, 2nd ed. New York: Oxford University Press; 1998;290–291.
28. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care
. 2002;25(suppl 1):S5–S20.
29. Sargeant LA, Wareham NJ, Bingham S, et al. Vitamin C and hyperglycemia in the European Prospective Investigation into Cancer–Norfolk (EPIC-Norfolk) study: a population-based study. Diabetes Care
30. Ceriello A. Oxidative stress and glycemic regulation. Metabolism
31. Hirai N, Kawano H, Hirashima O, et al. Insulin resistance and endothelial dysfunction in smokers: effects of vitamin C. Am J Physiol Heart Circ Physiol
32. Will JC, Ford ES, Bowman BA. Serum vitamin C concentrations and diabetes: findings from the Third National Health and Nutrition Examination Survey, 1988–1994. Am J Clin Nutr
33. Pan XR, Li GW, Hu YH, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care
34. Sargeant LA, Khaw KT, Bingham S, et al. Fruit and vegetable intake and population glycosylated haemoglobin levels: the EPIC-Norfolk Study. Eur J Clin Nutr
35. Willett WC, Stampfer MJ, Bain C, et al. Cigarette smoking, relative weight, and menopause. Am J Epidemiol
36. Troy LM, Hunter DJ, Manson JE, et al. The validity of recalled weight among younger women. Int J Obes Relat Metab Discord
37. Khaw KT, Bingham S, Welch A, et al. Relation between plasma ascorbic acid and mortality in men and women in EPIC–Norfolk prospective study: a prospective population study. European Prospective Investigation into Cancer and Nutrition. Lancet
38. US Department of Agriculture, US Department of Health and Human Services. Nutrition and Your Health: Dietary Guidelines for Americans.
Washington, DC: US Government Printing Office; 1995.