In Vino Veritas? Resveratrol for Diabetes

Your patients with diabetes may have heard through the grapevine about resveratrol to control blood sugar.  Indeed, a recent meta-analysis concluded that dietary supplements containing resveratrol significantly lowers blood glucose, A1c, and improves insulin sensitivity in patients with diabetes.¹  Dietary supplements are widely available and patients spend more than 30 billion (that’s billion with a “b”) dollars on them every year.  Health professionals should be prepared to analyze the evidence and discuss the potential benefits and harms of supplements, like resveratrol, with patients.²

Resveratrol—a naturally occurring polyphenolic compound—is found in the skin of red grapes and wine as well as some berries and nuts. Animal studies have suggested that resveratrol may lower blood glucose by activating the expression of sirutin 1 — an emerging drug target for regulating glucose and lipid metabolism — and by increasing GLUT4 — a glucose transporter that functions independent of insulin.³

All of the human resveratrol studies published to date have been quite small and not particularly compelling.¹  Thus, this meta-analysis combined the results of several studies in order to make a more conclusive determination regarding the effects of resveratrol on glucose control and insulin sensitivity. By pooling data from multiple studies we can glean a more precise estimate of the benefits and harms of a treatment, and thus have greater confidence making sound clinical decisions based on the data.⁴ However, the pooled estimate from a meta-analysis is only as good as the individual studies selected for inclusion in the analysis. This is especially important as we explore the details of this meta-analysis.

The literature search in a meta-analysis should be exhaustive to ensure that all relevant studies are identified and reviewed. In this respect the authors did an excellent job describing their search and study selection process.  Eligible studies were randomized controlled human trials employing a parallel or crossover design.  Baseline and follow up insulin, glucose, and A1C measures with accompanying statistical analysis were also required. Studies were not eligible if the intervention was part of a multi-component regimen. Individual study quality was assessed using the Jadad score with scores of 4 or greater on a 5 point scale considered to be high quality.

Of 131 articles identified and screened, only 11 studies were included in the analysis with data from 388 participants.  There were notable differences between these studies in terms of the populations enrolled, doses of resveratrol used, duration of exposure, and study quality.  For example, only three studies enrolled patients with diabetes. The remaining eight studies examined other populations including those with cardiovascular disease, obesity, metabolic syndrome, and healthy adults.  Doses used in these studies varied widely, ranging from 8 to 1500 mg per day. Trial duration ranged from as short as two weeks to six months.  Finally, less than half of the studies were considered to be high quality (e.g. a Jadad score 4 or more), and only one high quality trial was conducted in patients with diabetes.

Resveratrol use was not associated with changes in glucose, insulin concentrations, or A1C in patients without diabetes.  But in patients with diabetes, resveratrol demonstrated a significant reduction in fasting blood glucose (-35 mg/dL) and A1C (-0.79%). These results rival the A1C reduction seen with many antidiabetic medications! In addition resveratrol was also found to improve HOMA-IR, a marker for insulin resistance.  While these results are promising, it’s probably premature to recommend resveratrol to our patients.  

When interpreting a meta-analysis it is important to examine the heterogeneity of the data, or the variability of the results from individual studies included in the analysis.  The I² statistic tests for the magnitude of variability between studies and ranges between 0% and 100%. Lower percentages suggest that variability is more likely due to chance thereby increasing confidence in the findings. As the I² increases, variability increases and our confidence in the results of the analysis should decrease.⁴  In this meta-analysis, among the studies conducted in patients with diabetes, the I² statistic was 87.8%.  And when evaluating resveratrol’s effect on A1C in all patients, the I²  was 82.4%!

Other recently published literature suggests that resveratrol has little or no benefits on clinically important outcomes like cardiovascular disease, cancer, and death.⁵  Semba and colleagues recently published a prospective cohort study that enrolled 783 Italian citizens aged 65 or older.  They examined the relationship of urinary resveratrol metabolites—a surrogate for resveratrol dietary intake—and mortality. After 9 years of follow up, urinary resveratrol metabolite levels were not correlated with differences in mortality. In addition resveratrol intake was not associated with changes in secondary outcomes such inflammatory markers (CRP, IL-6, IL-8), incident cardiovascular disease, or cancer.

I believe health care professionals should not be routinely recommending resveratrol to their patients. In patients who do not have diabetes, no effect on blood sugar was found.  In patients with diabetes, there is a paucity of quality data and there was substantial heterogeneity among the trials.  We don’t know what the effective, much less optimal, dose is and we don’t have any long term safety data.  Moreover, I have concerns regarding the nutraceutical industry because the regulations are so lax.   Health care professionals need to discuss potential problems with product purity and consistency with their patients – and encourage them to use only those products that meet USP standards.

Well-designed, longer clinical trials are needed to determine the benefits and risks of resveratrol.  We need more data regarding the optimal dose as well as longer studies that include clinical endpoints, not just surrogate markers, with an appropriate safety analyses.  Unfortunately, funding for such research seem unlikely.  So what do you think? Are you ready to recommend resveratrol?  Or just kick back and enjoy a glass of wine?

FINAL NOTE:  iForumRx members are encourage to read:  
Murad MH, Montori VM, Ionannidis JPA, et al, How to read a systematic review and meta-analysis and apply the results to patient care.  JAMA 2014;312:171-179.
This is a great primer for students, residents, and clinicians on essential components of a meta-analysis and how to apply results into clinical practice.