Type 2 diabetes mellitus (T2DM) is a growing health problem worldwide. Whether sulphonylureas show better, equal or worse therapeutic effects in comparison with other antidiabetic interventions for patients with T2DM remains controversial.
To assess the effects of sulphonylurea monotherapy versus placebo, no intervention or other antidiabetic interventions for patients with T2DM.
We searched publications in The Cochrane Library, MEDLINE, EMBASE, Science Citation Index Expanded, LILACS and CINAHL (all until August 2011) to obtain trials fulfilling the inclusion criteria for our review.
We included clinical trials that randomised patients 18 years old or more with T2DM to sulphonylurea monotherapy with a duration of 24 weeks or more.
Two authors independently assessed the risk of bias. The primary outcomes were all-cause and cardiovascular mortality. Secondary outcomes were other patient-important outcomes and metabolic variables. Where possible, we used risk ratios (RR) with 95% confidence intervals (95% CI) to analyse the treatment effect of dichotomous outcomes. We used mean differences with 95% CI to analyse the treatment effect of continuous outcomes. We evaluated the risk of bias. We conducted trial sequential analyses to assess whether firm evidence could be established for a 10% relative risk reduction (RRR) between intervention groups.
We included 72 randomised controlled trials (RCTs) with 22,589 participants; 9707 participants randomised to sulphonylureas versus 12,805 participants randomised to control interventions. The duration of the interventions varied from 24 weeks to 10.7 years. We judged none of the included trials as low risk of bias for all bias domains. Patient-important outcomes were seldom reported.First-generation sulphonylureas (FGS) versus placebo or insulin did not show statistical significance for all-cause mortality (versus placebo: RR 1.46, 95% CI 0.87 to 2.45; P = 0.15; 2 trials; 553 participants; high risk of bias (HRB); versus insulin: RR 1.18, 95% CI 0.88 to 1.59; P = 0.26; 2 trials; 1944 participants; HRB). FGS versus placebo showed statistical significance for cardiovascular mortality in favour of placebo (RR 2.63, 95% CI 1.32 to 5.22; P = 0.006; 2 trials; 553 participants; HRB). FGS versus insulin did not show statistical significance for cardiovascular mortality (RR 1.36, 95% CI 0.68 to 2.71; P = 0.39; 2 trials; 1944 participants; HRB). FGS versus alpha-glucosidase inhibitors showed statistical significance in favour of FGS for adverse events (RR 0.63, 95% CI 0.52 to 0.76; P = 0.01; 2 trials; 246 participants; HRB) and for drop-outs due to adverse events (RR 0.28, 95% CI 0.12 to 0.67; P = 0.004; 2 trials; 246 participants; HRB).Second-generation sulphonylureas (SGS) versus metformin (RR 0.98, 95% CI 0.61 to 1.58; P = 0.68; 6 trials; 3528 participants; HRB), thiazolidinediones (RR 0.92, 95% CI 0.60 to 1.41; P = 0.70; 7 trials; 4955 participants; HRB), insulin (RR 0.96, 95% CI 0.79 to 1.18; P = 0.72; 4 trials; 1642 participants; HRB), meglitinides (RR 1.44, 95% CI 0.47 to 4.42; P = 0.52; 7 trials; 2038 participants; HRB), or incretin-based interventions (RR 1.39, 95% CI 0.52 to 3.68; P = 0.51; 2 trials; 1503 participants; HRB) showed no statistically significant effects regarding all-cause mortality in a random-effects model. SGS versus metformin (RR 1.47; 95% CI 0.54 to 4.01; P = 0.45; 6 trials; 3528 participants; HRB), thiazolidinediones (RR 1.30, 95% CI 0.55 to 3.07; P = 0.55; 7 trials; 4955 participants; HRB), insulin (RR 0.96, 95% CI 0.73 to 1.28; P = 0.80; 4 trials; 1642 participants; HRB) or meglitinide (RR 0.97, 95% CI 0.27 to 3.53; P = 0.97; 7 trials, 2038 participants, HRB) showed no statistically significant effects regarding cardiovascular mortality. Mortality data for the SGS versus placebo were sparse. SGS versus thiazolidinediones and meglitinides did not show statistically significant differences for a composite of non-fatal macrovascular outcomes. SGS versus metformin showed statistical significance in favour of SGS for a composite of non-fatal macrovascular outcomes (RR 0.67, 95% CI 0.48 to 0.93; P = 0.02; 3018 participants; 3 trials; HRB). The definition of non-fatal macrovascular outcomes varied among the trials. SGS versus metformin, thiazolidinediones and meglitinides showed no statistical significance for non-fatal myocardial infarction. No meta-analyses could be performed for microvascular outcomes. SGS versus placebo, metformin, thiazolidinediones, alpha-glucosidase inhibitors or meglitinides showed no statistical significance for adverse events. SGS versus alpha-glucosidase inhibitors showed statistical significance in favour of SGS for drop-outs due to adverse events (RR 0.48, 95% CI 0.24 to 0.96; P = 0.04; 9 trials; 870 participants; HRB). SGS versus meglitinides showed no statistical significance for the risk of severe hypoglycaemia. SGS versus metformin and thiazolidinediones showed statistical significance in favour of metformin (RR 5.64, 95% CI 1.22 to 26.00; P = 0.03; 4 trials; 3637 participants; HRB) and thiazolidinediones (RR 6.11, 95% CI 1.57 to 23.79; P = 0.009; 6 trials; 5660 participants; HRB) for severe hypoglycaemia.Third-generation sulphonylureas (TGS) could not be included in any meta-analysis of all-cause mortality, cardiovascular mortality or non-fatal macro- or microvascular outcomes. TGS versus thiazolidinediones showed statistical significance regarding adverse events in favour of TGS (RR 0.88, 95% CI 0.78 to 0.99; P = 0.03; 3 trials; 510 participants; HRB). TGS versus thiazolidinediones did not show any statistical significance for drop-outs due to adverse events. TGS versus other comparators could not be performed due to lack of data.For the comparison of SGS versus FGS no meta-analyses of all-cause mortality, cardiovascular mortality, non-fatal macro- or microvascular outcomes, or adverse events could be performed.Health-related quality of life and costs of intervention could not be meta-analysed due to lack of data.In trial sequential analysis, none of the analyses of mortality outcomes, vascular outcomes or severe hypoglycaemia met the criteria for firm evidence of a RRR of 10% between interventions.
There is insufficient evidence from RCTs to support the decision as to whether to initiate sulphonylurea monotherapy. Data on patient-important outcomes are lacking. Therefore, large-scale and long-term randomised clinical trials with low risk of bias, focusing on patient-important outcomes are required.