Medication taking behaviour of allopurinol treatment among patients with gout: A retrospective cohort study in the clinical practice research datalink (CPRD)

Background: Lowering serum uric acid concentrations is effective to control gout attack and prevent formation of tophi. Persistence with and adherence to uric acid lowering therapy are important to achieve benefit from medication. Therefore, insights into medication taking behaviour and factors influencing it, are extremely important. Objectives: To describe patterns of persistence with and adherence to allopurinol treatment among gout patients in the UK, and explore determinants of nonpersistence and non-adherence. Methods: A retrospective cohort study using the UK Clinical Practice Research Datalink (CPRD) was conducted. Patients with a first ever diagnosing coding for gout between 1st January 1987 and 30th June 2014 were included. Start of follow-up was defined as the first prescription of allopurinol after the start of valid data collection. Patients were excluded if they were younger than 40 years, were (ever) users of febuxostat and probenecid, or had a follow-up of less than 90 days after completion of the first prescription. Determinants included patient characteristics (at baseline), medication use (6 months prior to index-date) and other comorbidities (ever before index-date). Medication taking behaviour was described by non-persistence (occurrence of a first gap of 30 or 90 days) and non-adherence (proportion of days covered [PDC] over observation period) to allopurinol treatment. Kaplan Meier survival and multivariable Cox- and logistic regression were used to estimate the median time until discontinuation, and the strength of the association between determinants with non-persistence (90-day gap) and non-adherence, respectively. Results: A total of 47.744 gout patients (75.6% men; mean age 63.9 years) received allopurinol exclusively. After 5.3 years (SD 4.6) of follow-up 77% had a gap of 30-days and 54% had a gap of 90-days, and were non-persistence. Median survival time until discontinuation was 229 days (95% confidence interval [CI] 224 - 235) for a 30-day and 1059 days (CI 1012 - 1107) for a 90-day gap. The median PDC was 0.67 (IQR: 0.64). Over half of the patients (61%) were non-adherent (PDC