The Effect of Vitamin E on Endothelial Function of Micro- and Macrocirculation and Left Ventricular Function in Type 1 and Type 2 Diabetic Patients

For this study, 32 type 1 diabetic and 57 type 2 diabetic subjects, ages 21–80 years, were recruited. Diabetes was defined according to the recommendations of the American Diabetes Association’s Expert Committee on the Classification and Diagnosis of Diabetes (10).

The following exclusion criteria were applied: serious long-term diabetes complications, unstable coronary artery disease, cardiac arrhythmia, congestive heart failure, uncontrolled hypertension, recent stroke, lower extremities bypass surgery for peripheral vascular disease, liver disease, renal disease (including macroalbuminuria), severe dyslipidemia (triglycerides >600 mg/dl or cholesterol >300 mg/dl), history of smoking during the previous 6 months, serious chronic disease requiring active treatment, ingestion of vitamin E during the previous 2 months, and current treatment with lipid-lowering agents, glucocorticoids, antineoplastic agents, psychoactive agents, or bronchodilators. The protocol was approved by the Institutional Review Board at the Joslin Diabetes Center and the Beth Israel Deaconess Medical Center, and all participants gave written informed consent. Volunteers for the study were recruited through local advertisement.

This was a prospective, randomized, double-blind, placebo-controlled study of a total duration of 12 months. The baseline visit included physical examination, blood tests, vascular reactivity evaluation of micro- and macrocirculation, retinal photography, and LV function measurements. Participants were seen for a middle visit 6 months after recruitment and for an exit visit at 12 months. The 6-month visit included a physical examination, blood tests, and vascular reactivity measurements. The exit visit included all the above plus retinal photography and LV function measurements. Subjects fasted overnight before all visits.

Laser Doppler perfusion imaging measurements were used before and after the iontophoresis of acetylcholine chloride (Ach; endothelium-dependent vasodilation) and sodium nitroprusside (endothelium-independent vasodilation), as previously described (11). Flow-mediated dilation (FMD) was measured using a high-resolution ultrasound with a 10.0-MHz linear array transducer and the HDI Ultramark 9 system (Advanced Technology Laboratories, Bothel, WA), according to standard guidelines (12). The results were analyzed before the codes were broken in a blinded fashion. Information about the coefficient of variation of these techniques, which is <15%, has been provided elsewhere (1).

LV functions were evaluated using transthoracic echocardiogram according to the American Society of Echocardiography guidelines (13,14). The following measurements were obtained: LV end-diastolic and end-systolic diameters, LV fractional shortening, the ejection fraction, and LV mass.

Evaluation of retinopathy was performed at the Eye Unit of the Joslin Diabetes Center. The photographs were scored according to the following scale: 0 = no signs of retinopathy; 1 = very mild nonproliferative; 2 = mild to moderate nonproliferative; 3 = mild and other; 4 = evidence of laser treatment; 5 = evidence of laser treatment plus additional abnormalities; 6 = evidence of overall retinal vascular tortuosity; 7 = severe nonproliferative; 8 = mild proliferative; 9 = mild proliferative plus additional abnormalities; 10 = moderate proliferative; 11 = moderate to severe proliferative; and 12 = severe plus additional abnormalities.

Participants were randomized to 1,800 IU of vitamin E (in the form of 600-IU softgel capsules, all-racemic form) or corresponding placebo (485 mg of soybean oil). All capsules were taken as one dose in the morning before food. Compliance was evaluated by counting the returned capsules. The α-tocopherol levels were measured in six randomly selected vitamin E–treated and six placebo-treated patients. Baseline α-tocopherol levels were 24.9 ± 4.9 vs. 32.9 ± 4.5 μmol/l in the placebo and vitamin E groups, and posttreatment levels were 23.7 ± 6.9 and 56.2 ± 30.7 μmol/l, respectively.

The Minitab statistical package (Minitab, State College, PA) and the Statistical Analysis System (SAS, Cary, NC) were used for statistical analysis. The primary hypothesis of this study was to determine if there were micro- or macrocirculation changes through the use of vitamin E. Therefore, the primary analysis examined the difference in mean change between placebo- and vitamin E groups of these measures. This analysis was also done in each subset of diabetes type. The distribution of the change variables for baseline versus 6 month and baseline versus exit were tested for normality using Kolmogorov-Smirnov tests. Variables that did not meet the criteria for normality under these tests were analyzed with nonparametric methods (Kruskal-Wallis two-sample test) and are presented as medians with the 25th−75th percentile.

Because initial studies examining antioxidative agents have suggested a positive effect of vitamin E on circulation, preliminary power calculations were done estimating a 20% improvement in circulation measures in the vitamin E group. These were done using a power of 80% and a 0.05 level of significance to test the a priori hypothesis that vitamin E should increase circulation-related parameters by ≥20% in comparison with the placebo group. In the final analysis of the data, the authors felt that a significance threshold of P < 0.025 for significance would be a better criterion considering the number of measures being examined, but still appreciating the initial estimates of the power calculations.

The change between baseline and exit visit or baseline and 6-month visit in each group was evaluated using the paired t test for parametrically distributed data and Wilcoxon’s matched pair signed-rank test for nonparametrically distributed data. The t test was used to compare baseline characteristics between active and placebo groups. Correlation between variables was tested using both univariate and multivariate analyses (Pearson’s correlation and Spearman’s correlation analysis were used for parametrically and nonparametrically distributed data and analysis and multiple stepwise regression analysis). The results are presented as means ± SD or medians with the 25th−75th percentile.

The baseline demographics of the whole group of enrolled subjects and type 1 and type 2 diabetic patients separately are shown in Table 1. Type 1 diabetic patients were younger than type 2 diabetic patients (P < 0.0001), had a longer duration of diabetes (P < 0.0001), had lower BMI (P < 0.0001), had lower systolic (P < 0.05) and diastolic (P < 0.05) blood pressure, had higher HDL (P < 0.0001), and had lower triglycerides (P < 0.0001). The results of vascular reactivity and biochemical markers of endothelial function are shown in Table 2. Compared with type 1 diabetic patients, type 2 diabetic patients had higher levels of endothelin (P < 0.05), tumor necrosis factor-α (TNF; P < 0.05), PAI-1 (P = 0.01), and tissue plasminogen activator (tPA; P < 0.0001). No differences existed between active and placebo groups regarding treatment of diabetes, hypertension, or dyslipidemia, and glycemic control. There were no significant changes in the medications the participants were receiving during the study.

Of those initially enrolled, 26 subjects (9 type 1 and 17 type 2) did not complete the study; 5 patients did not finish the study because of nonserious adverse events not related to the study medication, and 21 subjects withdrew from the study for personal reasons or were lost to follow-up. Of these 26 subjects, 23 also did not return for the 6-month visit, whereas the remaining 3 returned for the 6-month visit but did not return for the exit visit. The results presented regarding the 6-month visit include data from the three subjects who were seen during this visit, but subsequently dropped out of the study. The subjects who did not complete the study were not different in terms of their average age, BMI, HbA_1c, FMD, nitroglycerin-induced dilation (NID), or Ach dilation from those who completed the study. In addition, the distribution of sex and use of a multivitamin was not different among those who did not complete the study compared with those who did.

The measurements that showed a significant difference in the changes between vitamin E and placebo groups are shown in Table 3. Because of multiple comparisons, significance was assumed at a level of P < 0.025. During the first 6 months of the study, improvement in endothelium-dependent skin vasodilation (P = 0.02) (Fig. 1) and a reduction in the endothelin (P = 0.01) were observed in the placebo compared with the vitamin E group. In addition, compared with the vitamin E group, the placebo group had a marginal improvement in systolic blood pressure (P = 0.04) and cholesterol (P = 0.04) and a marginal reduction in tPA (P = 0.04). During the entire 12-month period of the study, when compared with the control group, the vitamin E group showed a marginal increase in systolic blood pressure (P = 0.05). The FMD (endothelium dependent) was marginally different between the groups, mainly due to an improvement in the control group (P = 0.04). A difference was also observed in the NID (endothelium independent), mainly due to a reduction in the vitamin E group (P = 0.02) (Fig. 2). In addition, a marginal reduction was also observed in the C-reactive protein (CRP) levels in the vitamin E group (P = 0.05).

Changes in the endothelial function measurements between baseline and the 6-month visit are shown in Table 4. Compared with baseline, the placebo-treated patients showed an improvement in response to the iontophoresis of Ach (endothelium-dependent vasodilation) and a reduction in endothelin levels. A reduction in HbA_1c, total cholesterol, and HDL was also observed in the placebo group. The changes between baseline and exit visit in the active and placebo groups regarding demographic characteristics, endothelial function, and biochemical markers of endothelial function are shown in Table 5. No changes were observed in weight or BMI between the placebo and vitamin E groups (data not shown). The responses to vitamin E and placebo treatment were similar in type 1 and type 2 diabetic patients

The results of the eye examination and LV function are shown in Table 6. There was no difference in the changes during the study between the vitamin E and placebo groups. However, both groups showed a similar deterioration during the 12-month period of the study. No differences were observed in the changes in LV function from baseline to the 12-month visit between the active and control groups.

No correlations were observed between the changes in vascular reactivity of micro- and macrocirculation and any changes in glycemic control, lipids, or other biochemical markers of endothelial function.

The main findings of this study were that vitamin E does not have any beneficial effect on the vascular reactivity and ventricular function in patients with long-duration type 1 or type 2 diabetes when given at rather high dosages for a relative prolonged time period. In addition, the actively treated group showed a worsening in certain measurements of vascular reactivity compared with the placebo group. This is a novel finding that is contrary to our original hypothesis and one that has never been described in previous studies.

A considerable number of large clinical trials have been conducted over the last decade regarding the effect of vitamin E on preventing cardiovascular disease in at-risk populations, such as in subjects with hypercholesterolemia or a history of myocardial infarction and postmenopausal women, but the results are inconclusive (15–17). Thus, despite initial positive results in the CHAOS (Cambridge Heart Antioxidant Study), most subsequent studies have shown negative results (18). Furthermore, recent studies have also suggested that administration of vitamin E may prevent the beneficial effects of statins on HDL and the progression of coronary artery stenosis or increase the morbidity and mortality in postmenopausal women with coronary artery disease (19,20).

With regard to microvascular outcomes, no beneficial effects were detected in the diabetic patients who participated in the HOPE (Heart Outcomes Prevention Evaluation) trial and were treated with 400 IU of vitamin E for an average of 4.5 years (21). In contrast, a previous study that used the same dosage as this study showed that vitamin E increased retinal blood flow to normal levels in type 1 diabetic patients who were considerably younger and had a shorter duration of diabetes compared with the patients in the present study (22). The differences in age and diabetes duration may have been the main reason for the observed discrepancies between the two studies. Furthermore, in previous, short-duration studies, vitamin E decreased PAI-1 and CRP levels and the albumin excretion rate and improved nerve electrophysiological parameters (23–26). Regarding macrocirculation, most short-term studies have also reported beneficial effects of vitamin E on endothelial function, but these results have not been confirmed in studies of longer duration (27–32). In diabetic patients, 3- to 6-month studies have shown negative results in type 2 diabetes and positive results in type 1 diabetes with vitamin E (33–35).

In the present study, we followed patients for 12 months, the longest-duration study of vitamin E treatment in diabetes. Our results indicate that by the end of the study, the vitamin E group showed no changes in FMD, whereas a small improvement was noticed in the placebo group, resulting in a marginal difference in the change between the two groups (Table 3). Vitamin E–treated patients also showed a deterioration in NID, whereas the control subjects remained stable. Furthermore, both groups showed a minor increase in systolic blood pressure, but this increase was more pronounced in the vitamin E group, resulting in a marginal difference between the two groups. It is also of interest that, in contrast to previous studies, we observed the same trend in both type 1 and type 2 diabetic patients, namely, the lack of any beneficial effect of vitamin E

We have also investigated changes in skin microcirculation, an end point that has not been previously examined. A difference in endothelium-dependent vasodilation at the 6-month visit was found between the active and placebo groups, but this difference did not persist to the end of the study. These changes were accompanied by a similar change, although relatively small to be characterized as clinically significant, in the endothelin levels. These results suggest that these changes in endothelin may be related to the changes in endothelium-dependent vasodilation. The finding of no change in endothelium-dependent vasodilatation and endothelin levels at the end of the 12-month study period indicates that these abnormalities were transient, but further studies will be required before any solid conclusions can be made.

Another interesting finding in this study was the slight improvement in HbA_1c, a minor lowering of total and LDL cholesterol, an improvement in the Ach response, and a reduction of endothelin levels in the control group. Finally, when comparisons were made between the placebo and active groups, a marginal improvement in HbA_1c and total cholesterol levels was noticed during the first 6 months in the control group, whereas tPA was higher in the active group (Table 3). The reasons for these findings are not clear, but were not related to any additional care in the placebo group. In addition, as the placebo pills contained only a very small amount of soybean oil, it seems that the observed results are related mostly to a placebo nonspecific effect rather than a specific effect of the soybean oil.

Vitamin E had no influence in LV function measurements. To our knowledge, this is the first study to examine this end point. As no major changes were observed between the baseline and end of the study in the active and placebo groups, these results indicate a slow progression of ventricular changes in diabetes. This is in contrast to the retinal evaluation, which showed a deterioration of diabetic retinopathy in both groups, with no effect of the vitamin treatment.

Acute administration of vitamin E can prevent the postprandial increase of PAI-1 and inflammatory cytokines such as TNF-α, intracellular adhesion molecule, vascular CAM, and CRP that is related to the ingestion of a high-fat meal (36). In this study, we observed a small reduction in CRP levels after 12 months of treatment. The possible anti-inflammatory effects of vitamin E should not be overlooked, given the mounting evidence that CRP may be an important factor in the pathogenesis of atherosclerosis (37).

In conclusion, the results of this study indicate that long-term treatment with 1,800 IU of vitamin E daily has no beneficial effects on endothelial or LV function in diabetic patients. Because vitamin E–treated patients had a worsening in some vascular reactivity measurements when compared with control subjects, the use of high dosages of vitamin E cannot be recommended.

The study was supported by a research grant from the Juvenile Diabetes Research Foundation International (to A.V.), in part by National Institutes of Health Grant RR-01032 to the Beth Israel Deaconess Medical Center General Clinical Research Center, a William Randolph Hearst Foundation Fellowship and a Mary K. Iacocca Foundation Fellowship (to P.A.E.). Vitamin E capsules and the corresponding placebo were provided free of charge by Henkel Corporation (Countryside, IL).