VANCOUVER, B.C. – When patients with poorly controlled type 1 diabetes all receive good training in self-management, pump therapy is not superior to insulin injections in terms of glycemic control, according to findings from the REPOSE trial.
But the two differ somewhat on other outcomes.
A total of 317 adult patients in the United Kingdom were studied. Among the majority with a hemoglobin A1c level at baseline indicating poor control, the reduction at 24 months did not differ significantly between those assigned to insulin pumps and those assigned to multiple daily injection (MDI) therapy, investigators reported at the World Diabetes Congress.
However, pump therapy was associated with a greater improvement in certain measures of diabetes-specific quality of life, and MDI therapy was associated with a lower rate of diabetic ketoacidosis.
Dr. Simon Heller
“We believe that these results support a care pathway for adults with type 1 diabetes that starts with structured training using MDI but with pumps offered later to those in whom the limitations of MDI interfere with effective self-management,” commented lead investigator Dr. Simon Heller, professor of clinical diabetes at the University of Sheffield (England).
“Structured training reduces the risk of severe hypoglycemia and leads to modest but importantly long-lasting benefits in HbA1c, and should be delivered much more widely, even in the U.K.,” he added. “But there is also no doubt that further work is needed … to develop effective approaches to support people in achieving tighter glucose targets while we wait for the artificial pancreas to become generally available.”
Dr. Remi Rabasa-Lhoret
In an interview, session moderator Dr. Remi Rabasa-Lhoret, an endocrinologist with the Montreal Diabetes Research Center, commented, “I am overall disappointed. I would say my internal bias was that there would be a larger benefit with the pump.”
The fact that the trial enrolled a relatively unselected population may have played a role, he suggested. In particular, the patients studied had very high HbA1c levels at baseline, with a mean exceeding 9%, which in some countries would be a contraindication to getting a pump. Thus, they may have been unlikely to benefit even with good training.
“Another message is that maybe you need CGM [continuous glucose monitoring],” Dr. Rabasa-Lhoret added. “But then you are having huge cost issues and access issues to really benefit from the pump.”
Dr. Heller, in introducing the trial, noted that meta-analyses have suggested that pumps may have an edge over MDI when it comes to glycemic control. But “one of the limitations in the existing evidence is that people allocated to pumps have received more training and attention, compared to MDI, and very few trials have compared pumps to MDI with comparable training and insulin adjustment,” he noted.
To be eligible for the REPOSE (Relative Effectiveness of Pumps Over MDI and Structured Education) trial, patients had to have had type 1 diabetes for at least 1 year. Also, they had to be willing to undertake flexible intensive insulin therapy with self-monitoring of blood glucose and carbohydrate counting, and had to have no clinical indication or strong preference for a pump.
The patients were randomized to MDI (using rapid and twice-daily long-acting insulin analogues) or to pumps that provided continuous subcutaneous infusion of insulin. All were given 38 hours of high-quality structured education through the Dose Adjustment for Normal Eating (DAFNE) program, a skills-based program in flexible intensive insulin therapy and self-management.
The results for the entire trial population showed that among the large majority with baseline HbA1c levels of at least 7.5%, the mean change at 24 months was –0.64%.
In an intent-to-treat analysis, the mean change was –0.84% in the pump group and –0.42% in the MDI group. After adjustment, the difference between groups was not significant (–0.23%; P = .121).
In a per-protocol analysis, however, there was a difference between groups in favor of the pump (–0.84% vs. –0.32%; difference, –0.34%; P = .018).
“Importantly, HbA1c remained well above recommended national and international targets,” Dr. Heller said, with only about a fifth of patients in each group achieving an HbA1c of less than 7.5%.
The trial population as a whole had a reduction in the annual number of episodes of severe hypoglycemia per patient from 0.17 before baseline to 0.10 during the trial (incidence rate ratio, 0.46; P = .021). Again, there was no difference between groups.
The groups were also statistically indistinguishable with respect to diabetes-specific quality of life, general measures of well-being, anxiety and depression, and fear of hypoglycemia. But the pump group had greater improvements than the MDI group in scores for diabetes-specific diet restrictions (P = .004), daily hassles and function (P = .006), and treatment satisfaction (P less than .001).