Network Meta-analysis Of Glucagon-like Peptide-1 Receptor Agonists Versus Insulin Glargine In Type 2 Diabetes Suboptimally Controlled With Oral Antidiabetic Drugs

Objective:In clinical practice,type 2 diabetes mellitus?T2DM?patients usually consider the syringe hypoglycemic drug after suboptimally controlled with oral antidiabetic drugs.Now,the most widely used needle antidiabetic drugs are Glucagon-like peptide-1 receptor agonists?GLP-1RAs?and Insulin glargine?IGlar?.Since 2005,several GLP-1RAs have been approved to treat T2DM.These agents are alternative at this point in the treatment of insulin glargine?IGlar?.Among the 7 kinds of GLP-1RAs approved by FDA,only exenatide and lixisenatide are used on the market in China,and their effect and clinical significance have been gradually highlighted.When the treatment of the traditional drugs has drawbacks,the use of the remaining GLP-1RAs drugs on the market in China will also be an inevitable trend.Therefore,this evaluation of this kind of medicine is the research emphasis,and a network meta-analysis will assess the clinical efficacy and safety between GLP-1RAs and IGlar to help inform treatment decision.Methods:Embase,Medline,the Cochrane Library and clinicaltrails.gov databases were searched.Randomized clinical trials?RCTs?lasting 26±6 weeks that compared GLP-1RAs?the long-acting agents:semaglutide,dulaglutide,liraglutide,exenatide extended-release,albiglutide and taspoglutide;the short-acting agents:exenatide and lixisenatide?and IGlar were included.Eligibility participants were adult patients in type 2 diabetes suboptimally controlled with oral antidiabetic drugs for three months.The Cochrane collaboration's tool would assess risk of bias in included studies.Data about clinical efficacy and safety outcomes were analyzed using the gemtc package of R statistical software?3.4.1?.The node-split method would assess the consistency between the direct and indirect outcome for each comparison.Based the study sample or the limited data resource of treatments,sensitivity analysis should be assessed from excluding some trials to confirm the reliable.The comparison-adjusted funnel plot would assess the reporting bias.Results:A total of 45 RCTs?11390 patients?were included.Mean patient age is 56 years,44%is female,and mean disease duration is 8.3 year.All trails report detail random sequence generation,69%for allocation concealment,31%for blinding of patients and providers,40%for blinding of outcome assessment,only 6%for no dropping-out,51%for assessing data with the last-observation-carried-forward imputation for incomplete data outcome,and 11%for selecting report.Compared with placebo,all agents reduced the HbA_1c level,ranging from the high-dose semaglutide?-1.61%,[-1.96,-1.25]?to the high-dose lixisenatide?-0.39%,[-0.74,-0.04]?.All agents?except the low-dose lixisenatide?reduced the FPG level,ranging from the high-dose semaglutide?-2.94 mmol/L,[-3.95,-2.02]?to the high-dose lixisenatide?-0.78 mmol/L,[0.98,-0.57]?.Compared with IGlar,high-dose semaglutide?-0.81%,[-1.14,-0.48]?,the low-dose semaglutide?-0.38%,[-0.71,-0.05]?,the high-dose dulaglutide?-0.47%,[-0.64,-0.31]?,the low-dose dulaglutide?-0.26%,[-0.42,-0.1]?,the high-dose liraglutide?-0.27%,[-0.42,-0.12]?and exenatide extended-release?-0.21%,[-0.38,-0.04]?reduced more in HbA_1cc level,and the low-dose liraglutide,taspoglutide,albiglutide and the short-acting agents were similar.GLP-1RAs had an advantage of reducing body weight?-6.32 to-2.22 kg?,low risk of hypoglycaemia?except semaglutide,OR=0.22 to 0.57?,LDL?except albiglutide,-0.24 to-0.15 mmol/L?and SBP?except lixisenatide,-3.46 to-2.33 mmHg?level,but increasing the gastrointestinal symptoms,nausea?OR=44.38 to 8.16?,vomiting?OR=22.33 to 2.31?and diarrhea?OR=5.17 to 2.2?.Among GLP-1RAs,the long-acting agents semaglutide?-1.21%to-0.83%,-2.37 to-1.08 mmol/L?,the high-dose dulaglutide?-0.87%to-0.49%,-1.48 to-0.88 mmol/L?and liraglutide?-0.67%to-0.29%,-1.55 to-0.77 mmol/L?reduced more HbA_1c and FPG levels than the short-acting agents;across these three agents,semaglutide had the lowest risk of nausea?OR=0.38?and vomiting?OR=0.37?,and the low-dose liraglutide had the lowest risk of diarrhea?OR=0.46?.Among GLP-1RAs,semaglutide had the highest efficacy in reducing body weight?-4.11 to-1.7 kg?;the high-dose exenatide had the highest risk of hypoglycaemia?OR=3.1 to 1.77?;excepting the discontinued agent taspoglutide,the high-dose dulaglutide had the highest risk of the gastrointestinal symptoms,nausea?OR=4.85 to 1.78?,vomiting?OR=4.32 to 2.53?and diarrhea?OR=2.51 to 1.92?.Not clinically meaningful differences were found in blood pressure,lipid profile,upper respiratory tract infection,nasopharyngitis and serious adverse events.The consistency between direct and indirect evidence had small difference,which reduced the reliable across studies.The sensitivity analysis showed no import change.Observing the funnel plot,the comparison between IGlar and the high-dose exenatide may exist publication bias.Conclusion:?1?Compared with placebo,excepting the low-dose lixisenatide,all agents reduce the HbA_1cc and FPG levels.?2?Compared with IGlar,the long-acting agents semaglutide,dulaglutide,the high-dose liraglutide and exenatide extended-release reduced more in HbA_1c level.GLP-1RAs had an advantage of reducing body weight,low risk of hypoglycaemia,LDL and SBP level,but increasing the gastrointestinal symptoms.?3?Among GLP-1RAs,the long-acting agents semaglutide,the high-dose dulaglutide and liraglutide reduced more HbA_1c and FPG levels.Of these agents,semaglutide had the lowest risk of nausea and vomiting,and the low-dose liraglutide had the lowest risk of diarrhea.Whilst,semaglutide showed a greater reduction of Hb A_1c,FPG and body weight and the lower risk of hypoglycaemia,nausea and vomiting.