Study published in the American Journal of Clinical Nutrition
(November 2004)
A new epidemiological study suggests an increased risk of cardiovascular disease (CVD) events in postmenopausal diabetic women taking higher doses of vitamin C. The paper was published in the American Journal of Clinical Nutrition by scientists from the US, Korea and Singapore.
The evaluation was carried out on a subgroup of 1,923 women from a total of 41,836 recruited for the Iowa Women’s Health Study (1986-2000) (1). Vitamin C intake was assessed with a food-frequency questionnaire (FFQ) at baseline and classified into quintiles with ≥ 360 mg/d as the highest quintile. The quality of patients’ adherence to this classification seems to be mediocre at best, considering the medium correlations of 0.76 and 0.53 between the FFQ and the 24-hour dietary recalls during follow-up. Moreover, the diabetes subgroup was not pre-defined and the diagnosis and duration of diabetes were based on self-reports. Therefore, the study leaves considerable doubts about the quality of the statistical evaluation (see box below for statistical details) because other studies came to different conclusions. Even in the large complementary subgroup of 32,569 women who did not have diabetes at baseline total vitamin C intake was not associated with disease outcome. There was rather a non-significant trend for risk reduction.
The study is further refuted by a recent report from Walter Willet’s group at Harvard University in 85,118 women from the Nurses’ Health Study (2). Women taking vitamin C supplements had a significant lower risk of incident coronary heart disease (CHD) than women who did not take supplements. The risk reduction was 27 % in the highest quintile of vitamin C intake (median 704 mg/d) comprising vitamin C supplement users only, compared to the lowest quintile (median 70 mg/d). Interestingly, vitamin C from the diet alone (median 209 mg/d) in this well-nourished cohort did not reduce CHD risk. The inverse association between vitamin C supplement use and risk of CHD was even stronger among women with diabetes (-43 %). This is corroborated by previous reports from the NHANES I study (3) and human intervention studies where vitamin C supplementation reduced blood pressure (4), improved arterial function in type 2 diabetes (5) and reduced symptoms of pre-eclampsia (6).
In conclusion, the totality of evidence form epidemiological and supplemental human studies supports that vitamin C reduces the risk of CVD and is safe.
General comments on the statistical analysis of the study
Statistical analyses on several subgroups cause the issue of multiple testing and are prone to bias. Indeed, in a storied example, the ISIS-2 trial, the overwhelming treatment effect of aspirin in the total study population was even slightly reversed in patients with birth signs of Gemini and Libra (7). According to guidance in Statistics (8) subgroups have to be identified by interaction analysis with the whole study population. The p values for interaction between prevalent diabetes and total vitamin C intake and supplement use were 0.08 and 0.03, respectively, not significant when compared to a Bonferroni-adjusted level of 0.025 for two comparisons. When the subgroups are pre-specified in the protocol, the total number of subgroup analyses is known and a proper adjustment can be made. However, the number of statistical tests performed in this study was out of control, and thus, the results have a low degree of confidence and are purely exploratory. Further, covariates to be included in the primary analysis must be pre-specified in the protocol. However, in this study data driven covariate selection for adjustment seems likely to generate some significant or lower p values.
References
Lee DH, Folsom AR, Harnack L, Halliwell B, Jacobs DR Jr. Does supplemental vitamin C increase cardiovascular disease risk in women with diabetes? Am J Clin Nutr 80:1194-2000, 2004.
Osganian SK, Stampfer MJ, Rimm E, Spiegelman D, Hu FB, Manson JE, Willett WC. Vitamin C and risk of coronary heart disease in women. J Am Coll Cardiol 42:246-252, 2003.
Enstrom JE, Kanim LE, Klein MA. Vitamin C intake and mortality among a sample of the United States population.
Epidemiology 3:194-202, 1992.Duffy SJ, Gokce N, Holbrook M, Huang A, Frei B, Keaney JF Jr, Vita JA. Treatment of hypertension with ascorbic acid.
Lancet 354:2048-2049, 1999.Mullan BA, Young IS, Fee H, McCance DR. Ascorbic acid reduces blood pressure and arterial stiffness in type 2 diabetes.
Hypertension 40:804-809, 2002.Chappell LC, Seed PT, Briley AL, Kelly FJ, Lee R, Hunt BJ, Parmar K, Bewley SJ, Shennan AH, Steer PJ, Poston L. Effect of antioxidants on the occurrence of pre-eclampsia in women at increased risk: a randomised trial.
Lancet 354:810-816, 1999.Collins R, Peto R, Gray R, Parish S. Large-Scale Randomized Evidence: Trials and Overviews, 3rd Ed.; Oxford University Press: Oxford, UK, 1996
The European Agency for the Evaluation of Medicinal Products. ICH Topic E 9, Statistical Principles for Clinical Trials (1998)