From Heartwire — a professional news service of WebMD

October 28, 2008 — The widely prescribed metformin was alone among oral antidiabetic agents to show a significant effect on cardiovascular (CV) risk in a meta-analysis of 40 select controlled trials reported in the October 27, 2008 issue of the Archives of Internal Medicine [1]. It reduced CV mortality by 26% relative to any other antidiabetic agent or placebo, but there was no effect on all-cause mortality.

Rosiglitazone was the only such drug to show a possible risk increase, and it did so for all evaluated end points, none significantly, however.

"Our meta-analysis suggested that . . . metformin appeared moderately protective against cardiovascular effects and that rosiglitazone was possibly harmful, but a lack of power prohibited firmer conclusions," write the authors, led by Dr Elizabeth Selvin (Johns Hopkins Bloomberg School of Public Health, Baltimore, MD).

The group is up-front about the limitations of available data on the subject; in particular, the trials generally had glycemia-based end points and were too short to sufficiently evaluate CV risks. Few were longer than six months.

"I think that the study was well done and that they did the best they could with the available data. Meta-analyses can't be better than the data that underlie them," Dr David M Nathan (Massachusetts General Hospital, Boston) told heartwire . The author of an accompanying editorial [2], Nathan observed that until the potential CV risks of thiazolidinediones were known, diabetic trials didn't focus much on CV end points. "Sometimes they were collected haphazardly or without uniform definitions. So the data are what they are. They're okay, but they aren't definitive or conclusive in any way."

The included studies had been culled from 142 published randomized, controlled trials of oral drug therapy in type 2 diabetes. They were limited to those looking at second-generation sulfonylureas, biguanides, thiazolidinediones, and meglitinides that reported CV outcomes. Only two of the studies, however, included CV outcomes as prospectively defined end points.

Odds ratios (95% CI) for the effect of metformin vs any oral-antidiabetic comparator regimen or placebo in meta-analysis

*Fatal and nonfatal myocardial infarction (MI) or stroke.

The meta-analysis had excluded heart-failure (HF) outcomes except when they couldn't be distinguished from other events in combined end points. That was done, Selvin explained to heartwire , "to isolate the possible risk related to fatal and nonfatal myocardial infarction or stroke," events that defined the third main end point of the meta-analysis, "cardiovascular morbidity."

Without consideration of HF end points, the analysis could not show whether the oral drugs might promote heart failure. However, Nathan observed, the thiazolidinedione effect on fluid retention and possible heart failure and the lack of such an effect with the other antidiabetic agents are well defined.

Given the limitations of the current meta-analysis and others that have recently looked at the CV safety of antidiabetic agents, according to Nathan's editorial, "the conclusions drawn will be disappointing for healthcare practitioners who want a clear answer to the question 'Is it safe?' "

The meta-analysis was supported by the Agency for Healthcare Research and Quality and the National Institute of Diabetes and Digestive and Kidney Diseases; neither its coauthors nor Dr. Nathan has disclosed any relevant financial relationships.

Sources

• Selvin E, Bolen S, Hsin-Chieh Yeh HC, et al. Cardiovascular outcomes in trials of oral diabetes medications: A systematic review. Arch Intern Med. 2008;168:2070-2080.
• Nathan DM. Glycemic management of type 2 diabetes: How tight is right and how to get there. Arch Intern Med. 2008;168:2064-2066.

The complete contents of Heartwire , a professional news service of WebMD, can be found at www.theheart.org, a Web site for cardiovascular healthcare professionals.

Clinical Context

Study Highlights

• The investigators identified articles by searching MEDLINE, EMBASE, and the Cochrane Central Register of Controlled Trials (1966-2005).
• Included were single-therapy or combination-therapy (2 agents) studies with original data in adults with CV outcomes reported.
• Oral agents covered included metformin, second-generation sulfonylureas, biguanides, meglitinides, and thiazolidinediones.
• Excluded were studies with 3 or more agents in combination and the alpha-glucosidase inhibitors and studies that did not include CV outcomes.
• First-generation sulfonylureas were excluded.
• 2 reviewers independently extracted data, and the trials were summarized both qualitatively and quantitatively.
• 434 trials were relevant to the study questions.
• Only 40 trials from the United States and the United Kingdom included CV data and were included.
• Mean age of participants was 52 to 69 years, mean hemoglobin A_1c level at baseline was 6.2% to 10.2%, and two thirds of studies were less than 1 year in duration.
• More than half of studies reported support from the pharmaceutical industry.
• Qualitative analysis
  • In most studies, CV outcomes were reported as adverse events and not primary outcomes.
  • In many studies, a history of CV events was an exclusion factor.
  • The Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive) study and the United Kingdom Prospective Diabetes Study (UKPDS) were the largest studies.
  • The UKPDS provided the largest database, with more than 4000 participants and a mean of 10.7 years of follow-up.
  • A borderline 22% reduction in MI was seen with glyburide vs conventional treatment (P = .06) in the UKPDS 33 study.
  • The PROactive study had 5000 participants who were observed for 34.5 months.
  • It showed a nonsignificant 10% CV composite outcome reduction, and a significant 16% reduction in stroke, all-cause mortality, and nonfatal MI (P = .03) for oral antidiabetic agents.
• Quantitative analysis
  • The mean Jadad quality score was 3 of 5.
  • 8 of 40 trials that included CV events were reported as a primary or secondary outcome in 8 trials.
  • Results from these trials were pooled quantitatively with the Mantel-Haenszel method.
  • Comparisons of metformin vs other medications were the most robust, with 7 trials that included 11,986 patients vs other agents, diet, or placebo.
  • The pooled relative risk (RR) for metformin was 0.74 for CV mortality and 0.85 for CV morbidity.
  • For metformin, 9 trials contributed to all-cause mortality, and the pooled RR estimate was 0.81.
  • No other significant associations of oral antidiabetic agents with fatal or nonfatal CV events or all-cause mortality were found overall.
  • Rosiglitazone was the only oral agent associated with increased CV risk for mortality and morbidity and increased all-cause mortality with RRs of more than 1.0.
  • However, these estimates were not statistically significant.
  • The authors concluded that better reporting of CV outcomes in short-term studies were needed and that more robust research was also needed to compare long-term CV outcomes of oral antidiabetic agents.

Pearls for Practice

• Among the oral antidiabetic agents, reduction in CV morbidity and mortality rates has not been clearly demonstrated except for metformin.
• Use of rosiglitazone for type 2 diabetes may be associated with increased CV mortality and morbidity rates.

CME/CE Test