The Effect of Vascular Occlusive Events on Discontinuation and Cost of Care in Patients Treated with Tyrosine Kinase Inhibitors in Chronic Myeloid Leukemia

Introduction: Vascular occlusive events (VOE) are an important late complication among chronic myeloid leukemia (CML) patients treated with tyrosine kinase inhibitors (TKI). This study examines the impact of VOE on risk of discontinuation and cost of care among patients treated with TKIs outside of research protocol settings.

Methods: TKI treatment episodes for adult patients with CML diagnoses from 1/1/2009 to 1/31/2015 were identified in a large commercial insurance claims, Medicaid, and Medicare database in the US. Patients were required to have ≥6 months of enrollment prior to each TKI episode and ≥1 medical or pharmacy claims on or after the first CML diagnosis date for one of the TKIs: imatinib (IM), dasatinib (DAS), nilotinib (NIL), bosutinib (BOS), and ponatinib (PON).

VOE include one or more of the following: myocardial infarction, congestive heart failure, thrombotic events, acute coronary syndrome, peripheral vascular, cerebrovascular, or coronary artery diseases.

Cox proportional hazard model was used to evaluate time to discontinuation, as measured from the start date of each TKI episode to the earliest of the following: 1) switched TKI treatment, 2) stopped taking TKI defined as no refills for 90 days past end of drug supply, 3) had a bone marrow or stem cell transplant, or 4) death.

Gamma log-link regression was used to model total all-cause healthcare costs for each treatment episode. Costs were measured in 2014 US dollars per patient per month (PPPM), and included patient out-of-pocket costs, insurance-paid costs for CML drugs, other outpatient prescriptions, inpatient services, emergency room visits, office visits, and other outpatient services.

For both models, first occurrence of a treatment-emergent VOE within the TKI episode was included as a time-dependent variable. Other independent explanatory variables for the models included: age, gender, baseline Charlson Comorbidity Index from the National Cancer Institute (CCI NCI), type of TKI agent, and line of therapy.

Results: A total of 4,541 TKI treatment episodes (IM 1,982; DAS 1,321; NIL 1,059; BOS 114; PON 65) for CML were identified (mean age: 53.1 years and males: 51.9%). IM was the most commonly used first line TKI (54%). DAS was used 1st line in 26% and 2nd line in 33%. NIL was used 1st line in 19% and 2nd line in 25%. BOS and PON were predominately used in 3rd line or later.  947 (20.9%) VOEs occurred in all treatment episodes with 371 (21.1%), 357 (22.3%), and 219 (18.6%) occurring in first, second, and third plus lines, respectively. Occurrence of a VOE was associated with a significantly increased risk of TKI discontinuation (adjusted hazard ratio: 1.4 [1.2-1.6]). Compared to BOS, adjusted risk of discontinuation was not significantly different for treatment episodes using IM, DAS, or NIL, but was significantly higher for PON (Table 1). Cost of care for patients experiencing a VOE was also significantly higher (mean $3,702 PPPM higher) than for patients not experiencing an event (p=0.0454). Treatment with PON was independently associated with significantly higher cost of care compared to BOS (p=0.0026; Table 1).

Conclusions: Vascular Occlusive Events appear to be significantly associated with a higher risk of discontinuation and increased health care costs. Additionally, compared to BOS, PON appears to be independently associated with a significantly higher risk of discontinuation and higher costs. BOS appears comparable to IM, DAS, and NIL in duration of treatment and cost of care. It will be important to determine if factors other than VOE might underlie the higher discontinuations and costs observed with PON. Sensitivity analyses should also be conducted with deaths and transplants as censoring or true treatment discontinuation events.

 

 

 

Sponsorship Statement: This study was sponsored by Pfizer, Inc.