Reviews/Commentaries/ADA Statements META-ANALYSIS A Systematic Review and Meta-Analysis of Hypoglycemia and Cardiovascular Events A comparison of glyburide with other secretagogues and with insulin 1,2 2 AZIM S. GANGJI, MD, FRCPC CHARLES H. GOLDSMITH, PHD 1–2% (6). Differences in chemical struc- 2,3 1,2 TALI CUKIERMAN, MD CATHERINE M. CLASE, MB, FRCPC ture, pharmacokinetic, and pharmacody- 2,3 HERTZEL C. GERSTEIN, MD, FRCPC namic properties between sulfonylureas may lead to differences in the rates of hy- poglycemic reactions. Glyburide (called OBJECTIVE — Glyburide is the most widely used sulfonylurea but has unique pharmaco- glibenclamide in Europe), the most dynamic properties that may increase harm. We hypothesized that glyburide causes more hy- widely used sulfonylurea (7), has a rela- poglycemia and cardiovascular events than other secretagogues or insulin. tively long terminal half-life in chronic dosing compared with other sulfonyl- RESEARCH DESIGN AND METHODS — Data sources were Medline, Embase, ureas, owing to its high affinity for the Cochrane, and three other web-based clinical trial registers (1966–2005). Parallel, randomized, ␤ controlled trials in people with type 2 diabetes comparing glyburide monotherapy with mono- -cell sulfonylurea receptor and the accu- therapy using secretagogues or insulin were selected. Outcomes were hypoglycemia, glycemic mulation of active metabolites that are ex- control, cardiovascular events, body weight, and death. Titles and abstracts of 1,806 publications creted through the kidney (7). Several were reviewed in duplicate and 21 relevant articles identified. Data on patient characteristics, observational studies have reported in- interventions, outcomes, and validity were extracted in duplicate using predefined criteria. creased rates of hypoglycemia with the use of glyburide compared with other sul- RESULTS — Glyburide was associated with a 52% greater risk of experiencing at least one fonylureas (3,4). A systematic review of episode of hypoglycemia compared with other secretagogues (relative risk 1.52 [95% CI 1.21– randomized controlled trials (RCTs) evalu- 1.92]) and with 83% greater risk compared with other sulfonylureas (1.83 [1.35–2.49]). Gly- buride was not associated with an increased risk of cardiovascular events (0.84 [0.56–1.26]), ating the risk for hypoglycemia associated death (0.87 [0.70–1.07]), or end-of-trial weight (weighted mean difference 1.69 kg [95% CI with the use of glyburide has not, to our Ϫ0.41 to 3.80]) compared with other secretagogues. Limitations included suboptimal reporting knowledge, been previously conducted. of original trials. Loss to follow-up exceeded 20% in some studies, and major hypoglycemia was The University Group Diabetes Pro- infrequently reported. gram (UGDP) RCT (8) noted excess car- diac deaths in patients treated with CONCLUSIONS — Glyburide caused more hypoglycemia than other secretagogues and tolbutamide; whether this was attribut- other sulfonylureas. Glyburide was not associated with an increased risk of cardiovascular able to higher baseline cardiac risk in the events, death, or weight gain. patients allocated to tolbutamide or to a Diabetes Care 30:389–394, 2007 true biological effect has been widely de- bated. Consistent with the findings of the UGDP study, experimental laboratory he global prevalence of diagnosed risk for severe hypoglycemia, which is as- data have suggested that sulfonylureas, type 1 and 2 diabetes was estimated sociated with mortality and morbidity particularly glyburide, might increase the to be 2.8% in 2000 and projected to (3,4). T risk of cardiovascular events. During cor- be 4.4% by 2030 (1). The UK Prospective Sulfonylurea drugs bind the sulfonyl- ϩ onary angioplasty, with each subsequent Diabetes Study (UKPDS) showed that im- urea receptor, an ATP-sensitive K chan- balloon dilatation, the extent of ST seg- proving glycemic control reduced long- nel, and inhibit potassium efflux, which ment depression decreases and the time term microvascular complications (2). facilitates insulin secretion (5). Compared However, intensive therapy increases the with placebo, they reduce A1C levels by to onset of ST depression and the time to onset of angina increase. This phenome- ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● non, known as ischemic preconditioning, From the 1Division of Nephrology, McMaster University and St. Joseph’s Healthcare, Hamilton, Ontario, is thought also to occur in acute coronary Canada; the 2Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, On- syndromes. Sulfonylurea-induced potas- tario, Canada; and the 3Division of Endocrinology & Metabolism and Population Health Research Institute, sium efflux has been shown to reduce McMaster University and Hamilton Health Sciences, Hamilton, Ontario, Canada. cardiac ischemic preconditioning in ani- Address correspondence and reprint requests to Catherine M. Clase, 708-25 Charlton Ave. East, Hamil- ton, Ontario L8P 3P7, Canada. E-mail: [email protected]. mal studies (9). In humans undergoing Received for publication 23 August 2006 and accepted in revised form 12 November 2006. coronary angioplasty, the infusion of gli- Additional information for this article can be found in an online appendix at http://dx.doi.org/10.2337/ mepiride has the same effect on ST de- dc06-1789. pression and time to onset of angina as Abbreviations: FPG, fasting plasma glucose; RCT, randomized controlled trial; UKPDS, UK Prospective Diabetes Study; WMD, weighted mean difference. placebo, whereas the infusion of gly- A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion buride leads to a reduction in each of the factors for many substances. effects of preconditioning (10). DOI: 10.2337/dc06-1789 This systematic overview of random- © 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby ized controlled trials in people with type 2 marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. diabetes was conducted to determine DIABETES CARE, VOLUME 30, NUMBER 2, FEBRUARY 2007 389 Hypoglycemia: glyburide vs. secretagogues/insulin tent of agreement between reviewers. Dis- agreements were resolved by consensus. Validity assessment We assessed validity in duplicate using the following criteria: 1) method of ran- domization, 2) presence of allocation con- cealment, 3) blinding, 4) loss to follow- up, and 5) reporting of an intention-to- treat analysis. Data abstraction For each study, we abstracted, in dupli- cate: 1) inclusion and exclusion criteria; 2) baseline characteristics for the different treatment arms, including the number of participants at the start of the study; 3) the intervention and comparator (including dose, frequency, target A1C, and A1C achieved); 4) follow-up period and num- ber of participants at study completion; and 5) the definitions used to report hy- poglycemia and cardiovascular events. For each treatment arm we abstracted: 1) Figure 1—Study flow diagram. AG, Azim S. Gangji; TC, Tali Cukierman. all episodes of hypoglycemia (number of participants with one or more episodes and number of episodes per unit of per- whether people taking glyburide are at in- tially appropriate articles and each was re- son-time), 2) number of episodes of major creased risk for hypoglycemia or cardio- viewed independently for eligibility. and minor hypoglycemia, 3) number of vascular events compared with those cardiovascular events, 4) number of taking other secretagogues (other sulfo- Study selection deaths from any cause, and 5) weight nylureas and meglitinides) or those taking Eligible studies 1) described people with change and end-of-trial weight. When the insulin. For completeness of our analysis type 2 diabetes; 2) compared glyburide study included more than two arms, we of important harms, we also examined monotherapy with monotherapy using chose the comparator with the largest weight gain. other sulfonylureas, meglitinides, or insu- number of people. lin; 3) reported one or more of the follow- RESEARCH DESIGN AND ing outcomes: hypoglycemia (major, Data analysis METHODS — We followed the rec- minor, or all), cardiovascular events, or We summarized studies that compared ommendations of the Quality of Report- weight change; 4) described a parallel de- glyburide with other secretagogues sepa- ing of Meta-Analyses (QUOROM) sign RCT; and 5) were written in English. rately from studies that compared gly- conference (online appendix 1 [available For hypoglycemia and cardiovascular buride with insulin. Because of the at http://dx.doi.org/10.2337/dc06-1789]) events, we accepted the definition or out- unique pharmacokinetic and pharmaco- (11). come cluster reported in the original dynamic properties of glyburide, we pre- manuscript. Cardiovascular events in- specified a subgroup analysis comparing Search strategy cluded incident myocardial infarction, glyburide with other sulfonylureas. We searched biomedical databases (Med- stroke, amputation, episodes of conges- We assessed patient characteristics, line, Embase, the Cochrane library, clini- tive heart failure, or cardiovascular death. interventions, and outcomes for clinical caltrials.gov, controlled-trials.com, and Where multiple outcomes were reported, heterogeneity and used the I2 statistic to the U.K. national register of RCTs) and we selected the cluster that most closely quantify the