Incidence of hypoglycemia and associated costs in Type 1and Type 2 diabetes patients treated with insulin analogs vs. human insulin: A retrospective cohort study

Human insulin and insulin analogs are registered for the insulin treatment of Type 1 (T1DM) and Type 2 Diabetes Mellitus (T2DM). However, there is an ongoing debate concerning the hypoglycemic events and associated costs in both treatment options. For this reason, a systematic retrospective study was performed to compare hypoglycemic events and associated costs of both human insulin and insulin analogs treatment options. Relevant articles were identified by a methodical search through the electronic medical databases (MEDLINE, EMBASE, CURRENT CONTENTS, BIOSIS) from 1990 thru 2010. This was a retrospective cohort study of patients' data in the claims database from 2007 thru 2010. First year of the study period was used as a baseline period to meet inclusion criteria of patients being insulin naive for at least one year prior to the index period. Following three years from the index date 2008 to 2010 was considered as the follow-up period to analyze hypoglycemic events and associated costs. Descriptive analysis of the data suggested that differences in hypoglycemic events existed between the two cohorts of patients treated with insulin analogs and human insulin.;A total of 486 patients treated with Human Insulin (HI) and 6,599 patients treated with Insulin Analogs (IA) fulfilled all the inclusion and exclusion criteria. Medical and prescription claims related records were studied for these patients for hypoglycemic events and associated costs. HI and IA groups were further divided into sub-groups of Human Insulin Regular (HIR, n=106), Human Insulin - Neutral Protamine Hagedorn (NPH) (HIN, n=112), Human Insulin Pre-mixed (HIMIX, n=268); Insulin Analogs Fast Acting (IAFA, n=2,894), Insulin Analogs Long Acting (IALA, n=2,934) and Insulin Analogs Pre-mixed (IAMIX, n=771).;Systematic analysis showed that treatment with IA in T2DM patients compared to HI resulted in significantly better hypoglycemic rates when adjusted for relevant covariates. However in T1DM patients, the treatment with HI resulted in slightly better hypoglycemic rates. Furthermore, comparison of sub-groups (IAFA vs. HIR; IALA vs. HIN; IAMIX vs. HIMIX) showed that treatment with IAFA and IAMIX in T2DM patients resulted in better hypoglycemic rates when adjusted for relevant covariates. However in T1DM patients, the treatment with HIR resulted in slightly better hypoglycemic rates. For T1DM patients treated with HIN and HIMIX, due to extremely small patient population in these sub-groups the algorithm did not converge and the estimate was not established.;Cost analysis showed that treatment with IA in T2DM patients compared to HI resulted in lower cost when adjusted for relevant covariates. However in T1DM patients, treatment with IA resulted in lower cost when adjusted for some covariates. Furthermore, comparison of sub-groups (IAFA vs. HIR; IALA vs. HIN; IAMIX vs. HIMIX) indicated that treatment with IAFA (T1&T2;) and IAMIX in T2DM patients compared to HIR and HIMIX resulted in lower cost when adjusted for relevant covariates. IALA in T2DM patients compared to HIN resulted in slightly higher cost. For T1DM patients in IALA vs. HIN and IAMIX vs. HIMIX, the model did not converge.;There is significant evidence to conclude that T2DM patients treated with IA group and sub-group IAFA showed better hypoglycemic rates and associated costs compared to HI and HIR respectively. For other sub-groups, especially in T1DM patients this study cannot provide any statistically significant conclusion due to data limitations. Further research in these sub-groups is warranted using different datasets.