View More 2021, Scientific Sessions, Session Coverage

Head-to-head outcomes trial finds few differences between common type 2 diabetes medications


Estimated Read Time:

6 minutes

Steven E. Kahn, MB, ChB
Steven E. Kahn, MB, ChB

Preliminary findings from the first major head-to-head outcomes trial comparing four of the most commonly used type 2 diabetes agents showed few significant differences in clinical outcomes or adverse events in patients with established type 2 diabetes.

Trial investigators presented findings from Results of the Glycemia Reduction Approaches in Diabetes—A Comparative Effectiveness (GRADE) Study on Monday, June 28, at the Scientific Sessions. The symposium can be viewed by registered meeting attendees at through September 29, 2021. If you haven’t registered for the Virtual 81st Scientific Sessions, register today to access all of the valuable meeting content.

GRADE compares the clinical outcomes of glimepiride, a sulfonylurea; sitagliptin, a dipeptidyl peptidase-4 (DPP-4) inhibitor; insulin Lantus; and liraglutide, a glucagon-like peptide-1 (GLP-1) receptor agonist. The study, which includes 5,047 type 2 diabetes patients who are also taking metformin, does not include sodium-glucose cotransporter-2 (SGLT2) inhibitors, which were approved by the U.S. Food and Drug Administration after GRADE enrollment began.

David M. Nathan, MD
David M. Nathan, MD

“A majority of participants, 71%, reached the primary metabolic endpoint of A1C ≥7 and more than half of them did so within three years,” reported Steven E. Kahn, MB, ChB, Professor of Medicine, University of Washington School of Medicine and the Veterans Affairs Puget Sound Health Care System. “The main differences between the treatment groups occurred in the first year, [with the prevalence of A1C ≥7] being higher in sitagliptin, followed by glimepiride, and less frequently in the glargine- and liraglutide-treated groups. Subsequently, the cumulative incidence curves were parallel, progressing at a similar rate in all groups.”

Participants taking liraglutide had fewer overall cardiovascular events, but there were no significant differences in mortality, microvascular complications, or serious adverse events between the agents. The liraglutide group reported more gastrointestinal (GI) complaints than the other groups.

Dr. Kahn said a planned cost-effectiveness analysis of the four agents will be released at a later date.

John M. Lachin, ScD
John M. Lachin, ScD

“GRADE was conceived as a comparative-effectiveness study with the potential for immediate translation since it used approved medications according to FDA labeling,” said study chair David M. Nathan, MD, Professor of Medicine, Harvard Medical School, and Director of the Massachusetts General Hospital Clinical Research Center and Diabetes Center. “It included metformin as the basal medication regardless of what other drugs were used. The treatment goal in type 2 diabetes is A1C less than 7—the question was how best achieve this goal.”

Metabolic outcomes

The primary metabolic outcome was time to A1C progression ≥7. Time to A1C >7.5 was a secondary outcome, followed by the addition of glargine to the treatment regimen or more intensive dosing for those already on insulin. The tertiary outcome was A1C rebound to >7.5.

The mean time to A1C ≥7 was 697 days for sitagliptin, 810 days for glimepiride, 861 days for glargine, and 882 days for liraglutide, reported John M. Lachin, ScD, Research Professor of Biostatistics and Bioinformatics, George Washington University Biostatistics Center.

John B. Buse, MD, PhD
John B. Buse, MD, PhD

When analyzed by race, ethnicity, baseline A1C, body mass index (BMI), age, sex, and diabetes duration, baseline A1C was the only heterogeneous factor.

When stratified by tertials—A1C 6.8 to 7.2, 7.3 to 7.7, and 7.8 to 8.5—participants in each tertial progressed faster and higher. Glargine, glimepiride, and liraglutide were more effective than sitagliptin and their relative benefits increased as A1C increased.

Glargine had the longest time to exceed A1C 7.5 at 1,188 days, followed by liraglutide at 1,154 days, glimepiride at 1,116 days, and sitagliptin at 1,030 days. Glimepiride and sitagliptin had more tertiary metabolic outcomes, A1C >7.5, after the addition of glargine. Glargine and liraglutide had fewer and similar tertiary outcomes.

CV outcomes

John B. Buse, MD, PhD
Deborah J. Wexler

With just 378 total cardiovascular (CV) events, GRADE was not powered for CV outcomes.

“Our participants were at relatively low cardiovascular risk, at least for a diabetes population,” said John B. Buse, MD, PhD, Professor of Medicine, University of North Carolina School of Medicine.

The preliminary analysis looked at major adverse cardiac events (MACE), a composite of myocardial infarction, stroke, or cardiovascular death. Liraglutide showed a statistically significant reduction in any cardiovascular disease, with glargine, glimepiride, and sitagliptin showing similar results. Liraglutide showed numerical advantages in MACE, heart failure, and overall mortality, Dr. Buse said, but differences were not statistically significant. About 10% of cardiovascular events have not yet been adjudicated, he added. Reported findings are preliminary.

Microvascular outcomes

Microvascular outcomes included albuminuria, estimated glomerular filtration rate (eGFR), and peripheral neuropathy. There were no significant differences between the four treatments, reported Deborah J. Wexler, MD, MSc, Associate Professor of Medicine, Harvard Medical School, and Medical Director, Massachusetts General Medical School.

Mary Larkin, MS, RN
Mary Larkin, MS, RN

Between 11% and 12% of participants developed moderately increased albuminuria, while 4.7% to 5.3% developed severely increased albuminuria. Between 12% and 15% developed eGFR <60, Dr. Wexler reported.

Most participants, between 69% and 72%, developed peripheral polyneuropathy.

Adverse events

“The portion of individuals who developed serious adverse events was low overall and there was no difference among the treatment groups,” reported Mary Larkin, MS, RN, Assistant Professor, Harvard Medical School, and Nurse Director, Massachusetts General Hospital Diabetes Center. “Severe hypoglycemia was more common with glimepiride, but was remarkably infrequent and not significantly different among the other treatment groups.”

Margaret A. Tiktin, DWP, CNP, MBA
Margaret A. Tiktin, DNP, CNP, MBA

GI side effects were most common with liraglutide, reported by 60% of participants, which was 10% higher than the other treatment groups.

Trial design

Planning for GRADE began in 2009, the study was funded in late 2012, and patient recruitment began in 2013. Data collection ended in May 2021, although 10% of the CV events remain to be adjudicated.

More than 11,000 patients were screened for GRADE based on electronic medical records, and 5,000 were randomized to the four treatment groups, said Margaret A. Tiktin, DNP, CNP, MBA, Director of the Clinical Research Center, Cleveland Veterans Affairs Medical Center. All study participants had type 2 diabetes for less than 10 years at enrollment and were diagnosed as adults.

At baseline, all were on metformin monotherapy of 1,000 mg/day to 2,000 mg/day, with an A1C of 6.8 to 8.5. Key exclusions were evidence of type 1 diabetes, any recent glucose-lowering medication other than metformin, any major CVD event in the past year, a history of pancreatitis, congestive heart failure, or new diagnosis or treatment for cancer within the past five years.

Heidi Krause-Steinrauf, MS
Heidi Krause-Steinrauf, MS

The cohort was 65.7% white, 19.8% Black, and 14.5% other, including Asian, American Indian/Alaska Native, and Hawaiian/Pacific Islander. The cohort was 36.4% female and the mean age was 57 years.

Participants were generally obese, with a mean BMI of 34, mean blood pressure of 128/77 mmHg, mean A1C of 7.6, and mean eGFR of 95. A comparison with National Health and Nutrition Examination Survey data from 2011 to 2014 suggests that 9.1% of the U.S. population with type 2 diabetes would qualify for the study, Tiktin said.

Study retention was very good, reported Heidi Krause-Steinrauf, MS, Lead Research Scientist, George Washington University Biostatistics Center. Only 1.5% of participants were lost to follow-up and 94% attended the close-out study visit with no differences in retention between treatment groups.

“Completion of scheduled visits remained high during COVID,” she said. “There was an immediate shift to phone visits in March 2020, with a shift back to in-person visits in July.”


Already registered?
View this presentation at


Not yet registered?
Register now to access all presentations from the Virtual 81st Scientific Sessions