Impact of insulin lispro on HbA(1c) values in insulin pump users

Aim: To compare the therapeutic efficacy of the short-acting insulin analog ue insulin lispro (Humalog(R)) with that of buffered regular human insulin (Velosulin(R)) in patients on insulin pump therapy. Patients and Methods: Sixty-two (45 women and 17 men) young patients with t ype 1 diabetes using insulin pump therapy were compared while using buffere d regular human insulin for a mean +/- s.e.m. of 20.1 +/- 1.2 months or ins ulin lispro for a mean +/- s.e.m. of 19.7 +/- 0.5 months. The initial mean +/- s.e.m. age and duration of diabetes were 29.1 +/- 0.9 and 17.7 +/- 0.9 years, respectively. The mean HbA(1c) values, basal insulin dosages, premea l insulin dosages and number of low blood sugars were recorded during treat ment with both insulins. Results: Mean +/- s.e.m. HbA(1c) values were significantly lower (p < 0.001 ; paired Wilcoxon t-test) during insulin lispro treatment (7.4 +/- 0.1%) as compared to treatment with buffered regular human insulin (7.9 +/- 0.1%). Total units of insulin (mean +/- s.e.m.)/kg/day was significantly (p = 0.03 ) lower (0.61 +/- 0.02) during the insulin lispro treatment period as compa red to the buffered regular human insulin treated period (0.65 +/- 0.03). T otal mean +/- s.e.m. (U/kg/day) of basal insulin administered per day was h igher when patients received insulin lispro treatment (0.44 +/- 0.02 vs. 0. 42 +/- 0.01 for buffered regular human insulin treated period; p = 0.002). The pre-meal insulin boluses (mean +/- s.e.m.) for the two treatment groups were significantly different with less insulin required for the insulin li spro treatment period for all three meals (p < 0.001, t-test). The number o f mild/moderate and severe hypoglycaemic episodes were similar in the two g roups. Conclusions: We conclude that use of insulin lispro in pump therapy signifi cantly lowers HbA(1c) values in comparison to therapy with buffered regular human insulin insulin without increasing hypoglycaemic episodes.