Secondary prevention of peripheral arterial disease includes administration of statins regardless of the patient’s serum cholesterol level. Our study aimed to identify patient-associated risk factors for statin non-persistence and comparison of the explanatory power of models based on clusters of patient-associated characteristics.

Our study cohort (n = 8330) was assembled from the database of the largest health insurance provider in the Slovak Republic. Statin users aged ≥ 65 years in whom peripheral arterial disease was diagnosed during 2012 were included. Patients were followed for 5 years; those with a treatment gap period of at least 6 months without statin prescription were classified as “non-persistent”. The risk factors for non-persistence were identified within six models (sociodemographic, cardiovascular events, comorbid conditions, statin-related characteristics, cardiovascular co-medication and full model) using Cox regression. The explanatory power of models was assessed using Harrell’s C-index.

At the end of the follow-up, 35.7% of patients were found to be non-persistent. The full model had the highest explanatory power (C = 0.632). Female sex, atorvastatin and rosuvastatin as initially administered statins, being a new statin user and an increasing co-payment were associated with an increased risk for non-persistence. Increasing age, history of ischaemic stroke, diabetes mellitus, general practitioner as index prescriber, increasing overall number of medications and co-administration of certain cardiovascular co-medications were associated with a lower likelihood for non-persistence.

Patients identified as high risk for non-persistence require special attention aimed at the improvement of their persistence with statin treatment.