Home / Resources / Articles / ADA 2021:  Effectiveness of Liraglutide and Insulin Based on GRADE Study

ADA 2021:  Effectiveness of Liraglutide and Insulin Based on GRADE Study

Jul 10, 2021
Editor: Steve Freed, R.PH., CDE

Author: Alan Martinez, PharmD Candidate, University of South Florida Taneja College of Pharmacy

The GRADE Study is the most comprehensive study comparing multiple diabetic medications to assess efficacy in type 2 diabetes. So which drug therapy reigned supreme?  

Major advancements in diabetes management occurred within the 15 years before the  Glycemia Reduction Approaches in Diabetes—A Comparative Effectiveness (GRADE) Study began in 2013. Then, the only staple in the regimen for type 2 diabetes was metformin, with limited data on the selection of add-on therapies. This predicament led to the GRADE study to compare the effectiveness of multiple medications in patients with type 2 diabetes long-term.  


The GRADE study was an interventional, randomized, parallel-assigned, open-label clinical trial that recruited 5000 participants on metformin with a mean age of 57.2 years. The study consisted of a diverse population, including white (n=3314), black (n=1000), Asian (n=182), and Hispanic/ Latino ethnicity patients (n=929), among others. The four intervention groups, glimepiride, sitagliptin, liraglutide, and insulin glargine, each consisted of 1250 participants. The primary and secondary metabolic endpoints were the time to a confirmed HbA1c ≥ 7% and >7.5%, respectively. In a secondary metabolic endpoint event, initiate glargine in non-glargine groups, and the time to an HbA1c of >7.5% following the initiation of glargine, would be the tertiary metabolic endpoint. Secondary endpoints were microvascular outcomes, cardiovascular disease (CVD) outcomes, and adverse effects. 

In the end, 93.7% of the participants completed the study with an average participation of five years. Liraglutide and glargine had fewer participants develop primary metabolic endpoint compared to glimepiride and sitagliptin. Glargine’s relative risk reduction was better than glimepiride by 11% (P-value=0.016) and sitagliptin by 29% (P-value<0.001) but was similar to liraglutide (P-value=0.611). Liraglutide’s relative risk reduction was better than glimepiride by 13% (P-value=0.007) and sitagliptin by 31% (P-value<0.001). Within the first year, liraglutide and glimepiride were the most effective in decreasing HbA1c below 7%. However, in year 4, all interventions rose past the primary metabolic endpoint. The primary metabolic endpoint was longest with liraglutide and glargine, unlike glimepiride and sitagliptin. Differences by baseline HbA1c revealed that glargine, glimepiride, and liraglutide effectiveness increased as the HbA1c increased. It also concluded that sitagliptin was the least effective intervention in this study. Glargine was the most effective in not reaching the secondary metabolic endpoint, followed by liraglutide. Finally, liraglutide and glargine had fewer tertiary metabolic endpoints compared to glimepiride and sitagliptin.  

The microvascular outcomes consisted of moderate (≥30 mg/gm) or severe (>300 mg/gm) albuminuria, eGFR <60 mL/min/1.73m2, and distal sensory polyneuropathy. The results showed no differences between interventions for these parameters. CVD outcomes were Any CVD, total mortality, major adverse cardiovascular event (MACE), and Non- MACE outcomes. MACE outcomes included myocardial infarction (MI), stroke, or cardiovascular death. Non-MACE outcomes were heart failure requiring hospitalization and unstable angina requiring hospitalization or revascularization. There were no significant differences for MACE, Non-MACE, and mortality between interventions. For Any CVD, there was a significant difference, with liraglutide having lower cumulative incidence (P-Value=0.048) compared to the other interventions. For the last secondary endpoint, serious adverse events were similar among all interventions (P-Value=0.139). Severe hypoglycemia was observed frequently in glimepiride compared to the other medications (P-Value=0.003). Liraglutide had the highest weight loss within four years (P-Value <0.001) and the highest association with G.I. side effects (P-Value<0.001).  

Intervention: Primary Metabolic Endpoint Reached, Number of Patients 

  • Glargine: 67%, n=852 
  • Liraglutide: 68%, n=860 
  • Glimepiride: 72%, n=908 
  • Sitagliptin: 77%, n=981  

Intervention: Secondary Metabolic Endpoint Reached, Number of Patients 

  • Glargine: 39%, n=498 
  • Liraglutide: 46%, n=582 
  • Glimepiride: 50%, n=632 
  • Sitagliptin: 55%, n=696  

Intervention: Tertiary Metabolic Endpoint Reached, Number of Patients 

  • Glargine: 26%, n=332 
  • Liraglutide: 26%, n=332 
  • Sitagliptin: 30%, n=375 
  • Glimepiride: 31%, n=389 

Mean Time to Primary Metabolic Endpoint (Days)  

  • Sitagliptin: 697 
  • Glimepiride: 810 
  • Glargine: 861 
  • Liraglutide: 882 

Mean Time to Secondary Metabolic Endpoint (Days)  

  • Sitagliptin: 1030 
  • Glimepiride: 1116 
  • Liraglutide: 1154 
  • Glargine: 1188 

HbA1c Endpoint: Cumulative Incidence of Endpoints Over 6 Years 

  • Primary Metabolic Endpoint: 80%, P-Value < 0.001 
  • Secondary Metabolic Endpoint: 60%, P-Value <0.001 
  • Tertiary Metabolic Endpoint: 40%, P-Value=0.018 

 To summarize, the GRADE study has managed to compare these different interventions over a long period to assess their effectiveness in patients with type 2 diabetes. Liraglutide and glargine stood out as the most effective therapies during this trial. Secondary outcomes also gave an insight into a possible additional benefit to liraglutide, which was weight loss in the first four years. However, liraglutide’s gastrointestinal side effects appear to be heavily prominent compared to the others. To address the exclusion of TZDs and SGLT-2 inhibitors in the GRADE study: TZDs had safety concerns at the time, and the first SGLT-2 was approved after the study had begun. The clinical implications of this study are to provide clinical recommendations on the use of metformin with either glargine or liraglutide for patients with type 2 diabetes to promote excellent individualized medication therapy.  

Practice Pearls 

  • Liraglutide and glargine were the most effective treatments in managing HbA1c below 7%. 
  • Liraglutide had the additional health benefit of patients experiencing weight loss. 
  • There were minimal severe adverse events in all interventions, but liraglutide was associated with higher gastrointestinal side effects.  


David, Nathan M, et al. “Results of the Glycemia Reduction Approaches in Diabetes—A Comparative Effectiveness (GRADE) Study.” American Diabetes Association 81st Scientific Sessions. (ADA Symposium login required) 

Rationale and design of the glycemia reduction approaches in diabetes: a comparative effectiveness study (GRADE) Diabetes Care, May 20, 2013. 


Alan Martinez, PharmD Candidate, University of South Florida Taneja College of Pharmacy