Impact of fixed-dose and multi-pill combination dyslipidemia therapies on medication adherence and the economic burden of sub-optimal adherence

Objective: To compare medication adherence between patients initiating fixed-dose combination versus multi-pill combination dyslipidemia therapies and assess the association between optimal adherence (MPR >= 80%) and cardiovascular disease (CVD)-associated total healthcare resource utilization (THR) and costs (THC). Research design and methods: The HealthCore Integrated Research Database was used to identify patients >= 18 years newly initiating fixed-dose combination [niacin extended-release (NER) and lovastatin (NERL)] or multi-pill combination therapies [NER and simvastatin (NER/S) or lovastatin (NER/L)] between 1/1/2000 and 6/30/2006 (index date), with minimum 18 months of follow-up. Adherence was measured using medication possession ratio (MPR). Three multivariate models were developed controlling for demographic and clinical characteristics. A logistic model evaluated the association between study cohorts and optimal adherence, while negative binomial and gamma models estimated the association between optimal adherence and CVD-associated THR and THC, respectively. Results: In all, 6638 NERL, 1687 NER/S, and 663 NER/L patients were identified. Fixed-dose combination patients were younger [mean (SD) ages of 51.9 (10.5) vs. 56.0 (9.4) [NER/S] and 56.1 (10.6) [NER/L]; p < 0.01], had lower comorbidity (Deyo-Charlson Index 0.50 +/- 0.9 vs. 0.7 +/- 1.1 and 0.6 +/- 1.1, p < 0.01 and p < 0.05) and comprised fewer males (73.1 vs. 83.0% and 77.7%; p < 0.01 and p = 0.1). Fixed-dose combination patients had higher average 1-year MPR versus NER/S and NER/L patients (0.54 +/- 0.35 vs. 0.50 +/- 0.35 and 0.47 +/- 0.34, p < 0.01). NER/S and NER/L patients were 31.3% (95% CI: 22.9-39.5%) and 39.1% (95% CI: 26.7-49.4%) less likely to be optimally adherent than fixed-dose combination patients (p < 0.01). Additionally, optimally adherent patients had 8% and 40% decreases in annual CVD-attributable THR [0.920 (95% CI: 0.857-0.989); p = 0.023] and THC [0.601 (95% CI: 0.427-0.845); p = 0.003] versus sub-optimally adherent patients. Key limitations of the study include the limited ability of MPR to analyze the continuity of medication usage, inability to capture data on other key variables including race, income, and clinical characteristics such as smoking history, absence of laboratory values on all study patients, inability to capture over-the-counter fills of niacin, and inability to show causality of results obtained. Conclusions: Adherence was significantly higher among patients initiating fixed-dose combination versus multi-pill combination dyslipidemia therapies in this managed-care population. Additionally, patients with optimal adherence had a significantly lower CVD-associated THR and THC versus patients with sub-optimal adherence.