Nathan DM, et al. 3-CT-SY18. Presented at: American Diabetes Association Scientific Sessions; June 25-29, 2021 (virtual meeting).
Kahn reports he serves on advisory panels for Bayer AG, Boehringer Ingelheim, Eli Lilly, Merck, Novo Nordisk and Pfizer. Matthews reports he serves on advisory panels for Novartis Pharmaceuticals Canada and Novo Nordisk. Wexler reports she is a board member for Novo Nordisk. Nathan and Tiktin report no relevant financial disclosures.
In a head-to-head study of four classes of glucose-lowering drugs, the GLP-1 receptor agonist liraglutide and insulin glargine were more effective for maintaining HbA1c below 7% than a sulfonylurea or DPP-IV inhibitor.
David M. Nathan
In addition, the DPP-IV inhibitor sitagliptin (Januvia, Merck) was the least effective for maintaining HbA1c targets in the Comparative Effectiveness Study of Major Glycemia-lowering Medications for Treatment of Type 2 Diabetes (GRADE).
“GRADE was designed to help guide the choice of glucose-lowering medications that would be added to metformin based on effect on HbA1c, duration of effectiveness, effect on complications, tolerability and adverse effects,” David M. Nathan, MD, professor of medicine at Harvard Medical School and director of the Diabetes Center and the Clinical Research Center at Massachusetts General Hospital (MGH), said during a virtual presentation at the American Diabetes Association Scientific Sessions. “The study results would facilitate individualization of therapy — personalized medicine — and the results would address cost-benefit.”
For GRADE, researchers randomly assigned 5,047 adults with type 2 diabetes for less than 10 years taking metformin monotherapy to receive one of four additional commonly prescribed diabetes drugs with different mechanisms of action, used according to their labeling: the sulfonylurea glimepiride (n = 1,254), the DPP-IV inhibitor sitagliptin (n = 1,268), the GLP-1 receptor agonist liraglutide (Victoza, Novo Nordisk; n = 1,262) or insulin glargine (Lantus, Sanofi; n = 1,263). Mean age of the study cohort was 57 years, mean diabetes duration was 4.2 years, 65.7% were white adults, 19.8% were Black adults, 18.4% were Hispanic/Latino adults and 14.5% identified as other racial groups. Participants had baseline HbA1c between 6.8% and 8.5% (mean, 7.5%); mean BMI was 34.3 kg/m2.
Based on baseline characteristics, about 9.1% of the U.S. population with diabetes would have qualified for GRADE, according to presenter Margaret Tiktin, DNP, CNP, MBA, a nurse practitioner at the VA Northeast Ohio Healthcare System.
Participants were followed for a mean of 4.8 years between March 2013 and May 2021. HbA1c was recorded every 3 months, and lipids and serum creatinine were measured annually. Participants also underwent periodic oral glucose tolerance tests, neurocognitive exams and ECGs.
The primary outcome was time to a confirmed HbA1c of at least 7%. Secondary outcome was time to a confirmed HbA1c greater than 7.5%; basal insulin was then added to the regimen for participants in the non-insulin groups or intensified for those in the glargine group. Tertiary outcome was time to a confirmed HbA1c greater than 7.5% after the secondary outcome.
“Overall, the majority — 71% — of participants reached the primary outcome, with more than 50% doing so within 3 years. This metabolic outcome was reached despite attentive diabetes care and glucose-lowering medications provided free of charge,” Steven E. Kahn, MB, ChB, professor of medicine at the VA Puget Sound Health Care System and University of Washington in Seattle, said during a presentation.
Liraglutide was more effective than sitagliptin and glimepiride with respect to the primary outcome, with a 31% lower RR vs. sitagliptin (P .001) and a 13% lower RR vs. glimepiride (P = .007). The RR was slightly but nonstatistically significantly lower for liraglutide vs. glargine.
Participants in the liraglutide group took a mean of 185 days longer to reach a confirmed HbA1c of at least 7% than those in the sitagliptin group, 72 days longer than the glimepiride group and 21 days longer than the glargine group.
Results were similar in subgroup analyses based on race or ethnicity, sex and age. However, RR reduction was slightly greater for glargine than liraglutide vs. sitagliptin for participants in the highest tertile of baseline HbA1c, 7.8% to 8.5%.
Glargine was significantly more effective than the other agents with respect to the secondary outcome. Participants in the glargine group took a mean of 158 days longer to reach an HbA1c of at least 7.5% vs. sitagliptin, 72 days longer vs. glimepiride and 33 days longer than the liraglutide group.
Significantly more participants in the sitagliptin and glimepiride groups reached the tertiary outcome vs. glargine and liraglutide, which were similarly effective.
CV events, adverse effects
Microvascular events, including nephropathy and neuropathy, were similar across the drug groups. In preliminary results, the liraglutide group had significantly less composite cardiovascular disease, defined as three-point major adverse CV events, hospitalization for heart failure, transient ischemic attack, revascularization and unstable angina. No significant between-group differences were observed with respect to major adverse CV events, hospitalization for heart failure or mortality.
No significant differences in adverse effects were observed between the groups.
Participants in the liraglutide and sitagliptin groups lost significantly more weight than those in the glimepiride group, whereas the glargine group had stable weight. Gastrointestinal adverse effects were more common in the liraglutide group vs. the other groups. About 2.3% of the glimepiride group experienced severe hypoglycemia vs. 1.4% with glargine, 0.9% with liraglutide and 0.7% with sitagliptin.
Promise of personalized medicine
The GRADE analyses will eventually allow clinicians to tailor diabetes treatment to individual patients, presenter Deborah J. Wexler, MD, MSc, associate professor of medicine at Harvard Medical School and clinical director of the MGH Diabetes Center, told Healio.
Deborah J. Wexler
“We see the slight differences in outcomes among each of the treatment arms,” Wexler said. “We’re going to be looking a lot at the detailed predictors of response within groups by genetics, by phenotypic characteristics. GRADE is going to be pivotal in allowing us to personalize treatment for our patients with diabetes.”
One criticism of the GRADE study raised by David R. Matthews, DPhil, FRCP, professor emeritus of diabetic medicine at University of Oxford, during his independent commentary was the failure to include an SGLT2 inhibitor arm.
In addition to the much larger participant population needed to include a fifth study arm, Wexler said, SGLT2 inhibitors were not approved by the FDA when GRADE was designed and canagliflozin (Invokana, Janssen) had only just been approved at the study’s start.
“GRADE was designed to be comparative effectiveness trial using commonly approved combinations of medications,” Wexler told Healio. “In 2014, nobody knew what the SGLT2 inhibitors were going to turn out to be. The medications that we use in GRADE are still extremely commonly used and will continue to be extremely commonly used.”