Effects of multiple daily injection therapy with Humalog mixtures versus separately injected insulin lispro and NPH insulin in adults with type I diabetes mellitus

Abstract

Background: Injection of insulin lispro (LP) before meals provides a more physiologic insulin activity profile than regular human insulin, but the relatively short duration of action of LP may allow the blood glucose (BG) level to increase during the late postprandial period (4-7 hours after meals) unless basal insulin is optimally replaced. One approach to basal insulin optimization has been to combine small doses of NPH with LP before meals. When used in a similar fashion, premixed, fixed-ratio insulin preparations containing LP and NPL (an LP-based intermediate-acting insulin) may provide the basis for an optimized basal-bolus insulin regimen.

Objective: This study assessed mean late postprandial glycemic control during treatment with a premixed formulation consisting of a high proportion of LP (75% LP/25% NPL; H) and a premixed formulation consisting of a medium proportion of LP (50% LP/50% NPL; M). The H/M formulation was given before meals and was compared with treatment with preprandial LP + NPH (LP + N) in patients with type 1 diabetes mellitus (DM).

Methods: This multicenter, randomized, open-label, 2-period crossover study was conducted at 4 centers in Italy and 1 center in France. Patients eligible for the study had type 1 DM, were > or = 18 years of age, and had a glycosylated hemoglobin (HbA(1c)) <150% of the upper limit of normal. Patients were randomly assigned to 1 of 2 treatment sequences: LP self-mixed with NPH before meals plus NPH alone at bedtime for 8 weeks (LP + N) followed by preprandial H or M, plus NPH alone at bedtime for 8 weeks (H/M), or the opposite sequence. Assessments included 8-point self-monitored BG profiles, HbA(1c), and hypoglycemia (any sign or symptom of hypoglycemia or BG < 3.0 mmol/L [<54.0 mg/dL]). The primary outcome measure was the late postprandial BG value, calculated as the mean of the combined prelunch (late postbreakfast), predinner (late postlunch), and bedtime (late postdinner) values.

Results: A total of 89 patients with type 1 DM were enrolled (44 men, 45 women; mean [SD] age, 38.3 [12.8] years; mean [SD] body weight, 70.8 [11.6] kg; mean [SD] body mass index, 24.6 [3.0] kg/m(2); mean [SD] duration of diabetes, 17.8 [10.5] years; mean HbA(1c), 7.9% [0.88%]). The mean (SD) late postprandial BG values were similar between treatments (8.9 [2.1] mmol/L [160.3 (37.8) mg/dL] for H/M vs 9.0 [1.8] mmol/L [162.1 (32.4) mg/dL] for LP + N), as were the end point HbA(1c) values (7.8% [0.9%] for H/M vs 7.9% [0.8%] for LP + N). The rate of hypoglycemia was significantly higher during treatment with H/M, primarily because of episodes occurring between 12 PM and 6 PM, but was relatively low in both groups (mean/median rate per patient per 30 days: 2.87/2.14 for H/M and 2.11/1.07 for LP + N; P < 0.05).

Conclusions: In this population of patients with type 1 DM, preprandial H/M provided an effective alternative regimen for prandial and basal insulin replacement. Late postprandial BG control, an indicator of basal insulin sufficiency, was similar to that achieved with an intensified regimen of LP + N injected separately before meals, and the end point HbA(1c) was similar between the 2 treatments.