CRD summary

This review concluded that the evidence seemed to exclude any overall harmful effect of metformin on cardiovascular risk and suggested a possible benefit versus placebo/no treatment. Limitations with the included evidence and a lack of reporting of some aspects of the review process limit the reliability and generalisability of the results. The cautious conclusion and recommendations for research seem appropriate.

Authors' objectives

To assess the effects of metformin on the incidence of cardiovascular events and mortality.

Searching

MEDLINE, EMBASE and The Cochrane Library were searched for articles published in English from inception to October 2009; search terms were reported.

Study selection

Randomised controlled trials (RCTs) that compared metformin with placebo, active glucose-lowering therapies or no therapy were eligible for inclusion. Trials needed to last at least one year. Concurrent therapies had to be the same across trial arms. Cardiovascular events were defined as either fatal or non-fatal myocardial infarction, stroke, peripheral artery disease or other cardiovascular death; all-cause mortality and incidence of heart failure were also outcomes of interest.

Most studies were conducted in patients with Type 2 diabetes. Where reported, mean age ranged from eight to 75 years, males comprised zero to 79% of populations and glycated hemoglobin (HbA1c) was 5.9% to 12.4%. Metformin dose ranged from 425 to 3,000mg/day. Comparators included placebo, glibenclamide, insulin, glipizide, pioglitazone, rosiglitazone, sulphonylureas, vildagliptin and no treatment. Median trial duration was 112 weeks (range 52 to 343 weeks).

The number of reviewers who selected studies for the review was not reported.

Assessment of study quality

Trial quality was assessed using the Jadad scale. The number of reviewers who performed the quality assessment was not reported.

Data extraction

Data were extracted by two independent reviewers to produce odds ratios (OR) with 95% confidence intervals (CI). Disagreements were resolved by a third reviewer.

Methods of synthesis

Pooled odds ratios with 95% CI were calculated for cardiovascular events using a Mantel-Haenszel random-effects model. Heterogeneity was assessed using the Q statistic. Subgroup analyses were conducted for different comparators, metformin used as add-on to different therapies, for populations that included non-diabetic patients and patients aged less than 30 or more than 65 years. A meta-regression was performed to identify moderators of the effects of metformin on cardiovascular morbidity (metformin dose, trial duration, age, HbA1c and body mass index, lipid profile, fasting blood glucose and proportion of women enrolled). Publication bias was assessed using funnel plots and the Egger’s test.

Results of the review

Thirty-five trials met the inclusion criteria (7,171 participants who received metformin and 11,301 comparators). Studies that did not describe major cardiovascular events or reported no events were excluded from the analysis. Twelve trials were included in the analysis (5,455 participants who received metformin and 8,996 comparators). Randomisation was considered adequate in 19 trials. Allocation concealment was adequate in 18 trials. Blinding was adequate in 18 trials. Sixteen trials used an intention-to-treat analysis.

Overall, 451 patients who received metformin and 775 patients who received comparator treatments had cardiovascular events. Metformin did not produce any significant effect on cardiovascular events (OR 0.94, 95% CI 0.82 to 1.07). No significant heterogeneity was observed (p=0.46). Similar results were obtained for diabetic and non-diabetic populations and where metformin was used as an add-on therapy. Metformin was associated with a significant reduction in cardiovascular events when compared with placebo or no therapy (OR 0.79, 95% CI 0.64 to 0.98; eight trials), but not with active comparators.

Compared to other treatments metformin was associated with reduced cardiovascular morbidity in trials that enrolled patients aged less than 30 years (OR 0.76, 95% CI 0.61 to 0.95; four trials) or excluded patients aged more than 65 years (OR 0.78, 95% CI 0.60 to 1.00; three trials). There was no significant effect of metformin on all-cause mortality, cardiovascular mortality or incidences of myocardial infarction, stroke and heart failure. Further results were reported. There was no evidence of publication bias.

Authors' conclusions

Available evidence seemed to exclude any overall harmful effect of metformin on cardiovascular risk and suggested a possible benefit versus placebo/no treatment.

CRD commentary

This review addressed a clear review question supported by appropriate inclusion criteria. Three sources were searched. The review was restricted to articles published in English and unpublished studies were not sought, so language and publication biases could not be ruled out. Data extraction was conducted in duplicate; it was unclear whether similar methods were used to reduce error and bias during study selection and quality assessment.

Trial quality was assessed using established criteria and results were reported in full in an online appendix. A large proportion of trials were at a high risk of bias. Trial quality was not considered in the analyses. Although there was no statistically significant heterogeneity across all trials, there was substantial clinical heterogeneity; appropriate subgroup analyses were conducted.

Limitations with the included evidence and a lack of reporting of some aspects of the review process limit the reliability and generalisability of the results. The cautious conclusion and recommendations for research seem appropriate.

Implications of the review for practice and research

Practice: The authors stated that considering its safety, low cost and possible effects on non-cardiovascular endpoints, use of metformin as first-line agent in type 2 diabetes appeared justified.

Research: The authors stated that the cardiovascular effect of a pharmacological treatment should be verified through randomised trials with major cardiovascular events as the primary endpoint. Serious adverse events and other non-cardiovascular events could be recorded in trials. The detrimental effect of the combination with sulphonylureas needed further investigation.

Funding

No direct funding specified.

Bibliographic details

Lamanna C, Monami M, Marchionni N, Mannucci E. Effect of metformin on cardiovascular events and mortality: a meta-analysis of randomized clinical trials. Diabetes Obesity and Metabolism 2011; 13(3): 221-228. [PubMed: 21205121]

Original Paper URL

http://onlinelibrary.wiley.com/doi/10.1111/j.1463-1326.2010.01349.x/abstract

Indexing Status

Subject indexing assigned by NLM

MeSH

Cardiovascular Diseases /drug therapy /mortality /prevention & control; Diabetes Mellitus, Type 2 /complications /drug therapy /mortality; Diabetic Angiopathies /drug therapy /mortality /prevention & control; Female; Humans; Hypoglycemic Agents /administration & dosage /pharmacology; Male; Metformin /administration & dosage /pharmacology; Randomized Controlled Trials as Topic; Risk Assessment

AccessionNumber

12011001438

Database entry date

07/12/2011

Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.