The Association between Sulfonylurea Use and All-Cause and Cardiovascular Mortality: A Meta-Analysis with Trial Sequential Analysis of Randomized Clinical Trials

Abstract

Background: Sulfonylureas are an effective and inexpensive treatment for type 2 diabetes. There is conflicting data about the safety of these drugs regarding mortality and cardiovascular outcomes. The objective of the present study was to evaluate the safety of the sulfonylureas most frequently used and to use trial sequential analysis (TSA) to analyze whether the available sample was powered enough to support the results.

Methods and findings: Electronic databases were reviewed from 1946 (Embase) or 1966 (MEDLINE) up to 31 December 2014. Randomized clinical trials (RCTs) of at least 52 wk in duration evaluating second- or third-generation sulfonylureas in the treatment of adults with type 2 diabetes and reporting outcomes of interest were included. Primary outcomes were all-cause and cardiovascular mortality. Additionally, myocardial infarction and stroke events were evaluated. Data were summarized with Peto odds ratios (ORs), and the reliability of the results was evaluated with TSA. Forty-seven RCTs with 37,650 patients and 890 deaths in total were included. Sulfonylureas were not associated with all-cause (OR 1.12 [95% CI 0.96 to 1.30]) or cardiovascular mortality (OR 1.12 [95% CI 0.87 to 1.42]). Sulfonylureas were also not associated with increased risk of myocardial infarction (OR 0.92 [95% CI 0.76 to 1.12]) or stroke (OR 1.16 [95% CI 0.81 to 1.66]). TSA could discard an absolute difference of 0.5% between the treatments, which was considered the minimal clinically significant difference. The major limitation of this review was the inclusion of studies not designed to evaluate safety outcomes.

Conclusions: Sulfonylureas are not associated with increased risk for all-cause mortality, cardiovascular mortality, myocardial infarction, or stroke. Current evidence supports the safety of sulfonylureas; an absolute risk of 0.5% could be firmly discarded.

Review registration: PROSPERO CRD42014004330.

Fig 1. Study flowchart.

Fig 2. Forest plot for all-cause mortality.

For studies with multiple treatment groups, the group being compared is presented in parentheses.

Fig 3. Forest plots for all-cause mortality of sulfonylureas according to comparator (placebo/diet or active comparators).

For studies with multiple treatment groups, the group being compared is presented in parentheses.

Fig 4. Forest plots for cardiovascular mortality of sulfonylureas according to comparator (placebo/diet or active comparators).

For studies with multiple treatment groups, the group being compared is presented in parentheses.

Fig 5. Forest plots for all-cause and cardiovascular mortality of sulfonylureas as an add-on to metformin.

For studies with multiple treatment groups, the group being compared is presented in parentheses.

Fig 6. TSA for all-cause and cardiovascular mortality.

TSA discarded harm with sulfonylurea use with an α of 5%, a β of 80%, and an absolute difference of 0.5% between the groups (sulfonylurea and comparator). The blue line represents the Z curve (cumulative effect size), the red dashed lines represent the harm, benefit, and futility boundaries and the estimated optimal sample size adjusted to sample size and repeated analysis, and the black lines represent the conventional confidence intervals. The black number and marking in the x-axis represent the number of patients accrued until that point. (A) Sulfonylureas overall for all-cause mortality. Futility and optimal sample size boundaries were crossed. (B) Sulfonylureas overall for cardiovascular mortality. Futility and optimal sample size boundaries were crossed. (C) Sulfonylureas as add-on to metformin for all-cause mortality. Futility and optimal sample size boundaries were crossed.