(CSPOT01) Medication non-adherence and risk of mortality among patients with incident heart failure: a cohort study using the Clinical Practice Research Datalink (CPRD) Aurum database

Background: Clinical and cost effectiveness evaluation of health technologies for heart failure assume that medication adherence in the real world is consistent with guideline-directed recommendations. However, this may not always be realised in practice, limiting the effectiveness of these technologies.

Objectives: To evaluate heart failure medication non-adherence in general practice and its association with mortality among patients with heart failure.

Methods: A retrospective observational cohort study of 127,927 patients aged ≥18 years with incident heart failure diagnosed during 01/01/2009 - 31/12/2018 in England. Adherence at 24 months was based on proportion of days covered (PDC) averaged across the treatment classes angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, β-blockers, and mineralocorticoid receptor antagonists, to create a summary PDC. Non-adherence was defined as PDC < 80%. Outcomes were all-cause, cardiovascular (CVD) and heart failure (HF) mortality. Hazard ratios for all cause and cause-specific mortality were estimated using Cox proportional hazard models and Fine-Gray subdistribution hazard models for competing risks respectively.

Results: About 44% of patients were non-adherent to cardioprotective medication at 24 months. Non-adherent patients were more likely to be male, slightly older (73.3 years, standard deviation (SD) 13.4 versus 72.4 years, SD:12.3) and have ≥5 comorbidities (34.4% vs 30.6%) than adherent patients. During a median follow-up of 2.4 years (range; 0.0–9.2 years), there were 45,115 (35.2%) deaths from any cause, and 16,841 (13.3%) and 1,251 (1%) from CVD and HF respectively. The crude mortality rates for all cause, CVD and HF mortality were 119.6/1,000, 44.6/1,000 and 3.3/1,000 person-years respectively. In multivariable analysis, non-adherence was associated with a higher risk of all-cause mortality (hazard ratio 1.31, (95% confidence interval (CI):1.28-1.33), CVD mortality (subdistribution hazard ratio (SHR) 1.19, 95% CI:1.16-1.23), and HF related death (SHR 1.18, 95% CI:1.05-1.32), compared with summary PDC ≥80%.

Conclusions: Non-adherence to heart failure medication is associated with increased mortality risk. Our findings may inform clinical decision-making and underscore the importance of considering treatment adherence in cost effectiveness evaluations of health technologies for heart failure.