Dr. Christian S. Frandsen, from Hvidovre Hospital, Hvidovre, Denmark, and colleagues conducted a 12-week randomized, double-blind, placebo-controlled, study in 40 patients with poorly controlled type 1 diabetes.
The objective was to investigate the safety and efficacy of once-daily liraglutide 1.2 mg versus placebo as an add-on to insulin treatment. The primary end point was change in HbA1c.
Patients, ages 18-70, were outpatient Caucasians with a body mass index of 18-28, HbA1c at least 8%, and no residual beta-cell function. Diabetes was diagnosed between the ages of 5 and 40. Their only treatment was basal/bolus insulin.
The study failed to meet its primary end point. After 12 weeks, the change in HbA1c from baseline was a nonsignificant -0.6% with liraglutide and -0.5% with placebo.
Adverse events, mostly gastrointestinal, occurred in 90% of the liraglutide patients and in 65% of the placebo patients. The five serious adverse events were judged not study related.
The authors wrote that insulin treatment was continually optimized during the study without a run-in period, possibly causing a substantial effect on HbA1c.
Dr. Eleftheria Maratos-Flier, of Harvard Medical School and Beth Israel Deaconess Medical Center in Boston, agreed.
In a telephone interview with Reuters Health, Dr. Maratos-Flier, who was not involved in the study, said that it was confounding to adjust the insulin levels.
To do a study like this, she said, the insulin dose needs to stay the same to see the effect of the add-on drug.
"In someone who has really good glucose control and is adjusting their insulin on a regular basis, liraglutide is unlikely to have an effect," she said.
Dr. Adrian Vella, of the Division of Endocrinology and Metabolism at the Mayo Clinic in Rochester, Minnesota, also agreed.
Dr. Vella, who was not involved in the study, told Reuters Health by email that beyond the benefits of optimizing insulin dosing, indiscriminate use of glucagon-like peptide-1 receptor agonists in addition to basal/bolus insulin is unlikely to be helpful in unselected patients with type 1 diabetes.
"Very often people with type 1 diabetes can improve their HbA1c with appropriate supervision and instruction. It's possible that an effect was missed given the improvement of glycemic control in the placebo arm," she said.
It's also possible that an effect was missed due to the lack of a subanalysis in the liraglutide patients who lost significant amounts of weight.
Body weight decreased in the liraglutide group from -3.1 compared with an increase of +1.1 kg in the placebo group (p<0.0001).
According to Dr. Maratos-Flier, it would have been useful to know if the primary endpoint might have been met if the researchers looked at the subset of weight loss responders.
"Not everyone who takes liraglutide loses weight, so a subanalysis in the group of people who did might have shown if they had more improvement overall than those who didn't," she said.
The clinical trial is still of value, she said, noting that the failure to meet a primary endpoint needs to be put in context.
"These are important studies," she said. "They address post-drug release issues in the community and give us a better idea of what the benefits of a particular drug might be."
The author did not respond to requests for comments.
This research had a number of supporters. Six coauthors reported relationships with pharmaceutical companies.
Diabetes Care 2015 Nov.
www red DiabetologNytt