Background: Multicenter long-term studies of predictors for the effectiveness of continuous subcutaneous insulin infusion (CSII) in clinical practice are lacking. We hypothesized that there are substantially greater reductions in hemoglobin A1c (HbA1c) in patients with poor glycemic control and that other predictors may also exist.
Subjects and methods: We used data from 10 outpatient diabetic clinics in Sweden and studied CSII treatment over 5 years. Patients with HbA1c values available ≤ 6 months before starting CSII and at 5 years were included (n = 272, 82% of CSII patients) along with 2,437 contemporaneous controls on multiple daily insulin injections (MDI). Baseline variables evaluated were age, sex, diabetes duration, insulin dose, body mass index (BMI), HbA1c at baseline, and outpatient clinical care unit.
Results: At 5 years, significantly greater reductions in HbA1c by CSII compared with MDI were found for patients with higher baseline HbA1c (P = 0.032) and lower baseline BMI (P = 0.013). For baseline HbA1c levels of 7.0%, 8.0%, and 9.0% and a BMI of 25 kg/m(2), the reduction in HbA1c level by CSII was 0.08% (difference not significant), 0.16% (95% confidence interval, 0.03-0.29%), and 0.25% (95% confidence interval, 0.11-0.39%), respectively. Corresponding analyses for the change in HbA1c level from start to 1 and 2 years revealed a significant interaction of insulin pump therapy only with baseline HbA1c levels (P < 0.001 and P = 0.030, respectively). The interaction term between outpatient clinical care unit and CSII treatment was statistically significant for some care units, with some care units demonstrating a benefit from CSII and others demonstrating a detriment.
Conclusions: Patients with high HbA1c levels have a greater probability of improved HbA1c after initiating pump therapy, but effects remain relatively modest even for patients with poor control. Factors predicting successful insulin pump use need further study.