VIENNA – Treatment with an insulin pump rather than insulin injections may help protect patients with type 2 diabetes from heart disease, a large registry study has determined.
Over a mean follow-up of about 7 years, the pump was associated with a 44% decrease in the risk of fatal cardiovascular disease. It also conferred a 29% decrease in the risk of death overall, Dr. Soffia Gudbjornsdottir said at the annual meeting of the European Association for the Study of Diabetes.
She used data extracted from the Swedish National Diabetes Registry, which was founded in 2005. It contains information on 95% of the country’s type 1 diabetes patients and is linked with national inpatient and death registries.
The study group comprised 18,168 patients, 2,441 of whom were using an insulin pump. There were some significant baseline differences between the pump users and the users of injectable insulin.
Pump users were younger (age at baseline 38 vs. 41 years), and more likely to be women. Those treated with the pump had better measures of blood pressure and lipids. They were more likely to exercise and less likely to smoke. Insulin injection users had higher rates of prior cardiovascular disease. However, a propensity matching score eliminated these differences and balanced the group, Dr. Gudbjornsdottir noted.
By 7 years, there had been few cardiovascular events or deaths – an expected finding, since the cohort was young. Fatal/nonfatal coronary heart disease had developed in 7% of the insulin injection users and 4% of the of pump users, a nonsignificant difference. Neither were there significant differences in fatal or nonfatal cardiovascular disease (8% vs. 5%) or noncardiovascular mortality (4% vs. 2%).
However, significant differences did appear in two other endpoints: fatal cardiovascular disease (3% vs. 1%) and total mortality (7% vs. 3%).
A multivariate regression model found several significant risk reductions associated with pump use, including fatal cardiovascular disease (hazard ratio, 0.56), and all-cause mortality (HR, 0.71).
There were no significant differences in the risks for fatal/nonfatal coronary heart disease, fatal/nonfatal cardiovascular disease, or noncardiovascular death.
During the discussion, Dr. Gudbjornsdottir addressed the concern of a treatment allocation bias due to some unknown variable in the group. “If we had such an unknown covariate, with, for example, a hazard ratio of 1.3, it would have had to be present in at least 80% of the insulin group and in none of the pump group to invalidate the results, so I would say our study findings are quite robust.”
She had no financial disclosures.