Incidence of Type II Diabetes Mellitus and Hyperlipidemia in Patients Prescribed Dasatinib or Nilotinib As First or Second Line Therapy for Chronic Myelogenous Leukemia (CML)

Background: Metabolic changes such as Type II Diabetes Mellitus (T2DM) and hyperlipidemia (HLD) have been reported in patients receiving BCR-ABL tyrosine kinase inhibitor (TKI) therapy. Incidence of metabolic outcomes in this population has not been evaluated in large, real-world studies. The purpose of this study was to determine incidence and risk for T2DM and HLD in commercially and Medicare insured US patients prescribed dasatinib or nilotinib as first or second line therapy for CML.

Methods: A retrospective cohort study using the MarketScan healthcare claims was conducted. Patients ≥18 years with ≥2 inpatient or outpatient claims for CML (ICD-9 205.10-205.12) separated by at least 30 days and ≥1 prescription for dasatinib or nilotinib from July 2006- December 2014 were eligible. Patients were required to have continuous enrollment for ≥6 months prior to the first TKI prescription (index date) and no exposure to a TKI other than imatinib during the pre-index period. No minimum follow-up time was required. First line TKI users were defined as those patients with no imatinib prescription during the pre-index period. Patients with a prescription for imatinib during the pre-index period were considered second line. Two cohorts of patients were studied. Cohort 1 was evaluated for the occurrence of T2DM, defined as ≥2 claims with an ICD-9 code for T2DM or 1 diagnosis claim and an anti-diabetic medication claim. Patients who had a diagnosis of T2DM or received medications for T2DM prior to the index date were excluded from cohort 1. Cohort 2 was evaluated for the occurrence of HLD, defined as ≥2 claims with an ICD-9 code for HLD, or 1 diagnosis claim and an anti-HLD medication claim. Patients with a diagnosis of HLD or medications for HLD prior to the index date were excluded from cohort 2. Patients were evaluated for the outcome of interest (T2DM or HLD) from index date until the outcome of interest occurred, discontinuation of therapy, switch in therapy, end of study (12/31/2014), or end of enrollment, whichever occurred first. To estimate the incidence rates of both outcomes of interest, person years were enumerated from index date until the end of follow up. Incidence rates were computed as the number of events divided by the sum of the person years (PY) at risk for all subjects in the cohort and were reported with associated 95% confidence interval (CI) assuming a Poisson distribution. Multivariate Cox proportional hazards models were used to estimate hazard ratios (HR) for outcomes while adjusting for age, gender, region, insurance type, index year, Charlson Comorbidity Index (CCI), comorbidities, concomitant medications, and line of therapy.

Results: 2,004 patients met inclusion criteria for cohort 1 (n=1,272 dasatinib, n=732 nilotinib), while 1,280 patients met cohort 2 inclusion criteria (n=845 dasatinib, n=435 nilotinib). Baseline characteristics were largely similar between dasatinib and nilotinib patients in each cohort. Significant differences existed in index year, proportion of patients with certain comorbidities and baseline concomitant medications, and percentage of patients receiving the TKI for first line vs second line therapy. Among cohort 1, mean age was 53.4 years and 54.1% were male; median follow-up time was 216 days for dasatinib and 245 days for nilotinib. Incidence rate of T2DM was 17.58 per 1,000 PY (95% CI: 11.14, 26.38) for dasatinib, and 40.36 per 1,000 PY for nilotinib (95% CI: 27.60, 56.98). The HR for occurrence of T2DM (dasatinib as reference) was 2.77 (95% CI: 1.58, 4.86, P=0.0004), indicating patients on nilotinib had a 2.77 times greater risk of developing T2DM while adjusting for baseline characteristics. The mean age of cohort 2 was 48.6 years and 50.3% were male; median follow-up time was 213 days for dasatinib and 206 days for nilotinib. Incidence rate of HLD was 46.41 per 1,000 PY (95% CI: 33.00, 63.45) for dasatinib, and 74.63 per 1,000 PY for nilotinib (95% CI: 50.70, 105.94). The HR for occurrence of HLD (dasatinib as reference) was 1.75 (95% CI: 1.07, 2.87, P=0.025), indicating patients treated with nilotinib had a 1.75 times greater risk of developing HLD after adjusting for baseline characteristics.

Conclusions: Patients receiving nilotinib for CML had a higher incidence rate and significantly greater adjusted risk for developing T2DM or HLD than patients receiving dasatinib.