Trajectories of adherence to secondary prevention pharmacotherapy after coronary angiography: a retrospective population-based study

Abstract Background Medical therapy improves the outcomes of coronary artery disease (CAD), but medication adherence remains a notable challenge. Current guidelines recommend statins, beta-blockers and renin-angiotensin system inhibitors (RASi), as well as temporary use of P2Y12 inhibitors, for most patients without contraindications. However, there is limited understanding of how adherence developes over time and when patients are at risk of non-adherence. Purpose We assessed long-term adherence patterns to secondary prevention medications following coronary catheterization using real-world data. Methods This retrospective cohort study included individuals undergoing first-time coronary angiography or intervention with 90 days of follow-up and at least one dispensed prescription for the respective medications, from 2010-2020. Clinical data were derived from internal hospital registries. Medication data were derived from a central registry containing all dispensed prescriptions in the country. Medication availability was estimated by dispensed prescriptions in 30-day intervals, up to one year for P2Y12i and up to three years for other medications. A latent class growth model was used to estimate trajectories of adherence and non-adherence to medication. Cox regression was used to evaluate outcomes of high vs. low adherence trajectories. Logistic regression was used to identify variables associated with non-adherence. Propensity score matching was used to generate covariate balance. Results Out of 14337 individuals, 11385 (94.9%) met the inclusion criteria and collected at least one medication of interest after the index procedure, distributed as 9607 (84%), 8892 (78%), 7402 (65%), and 5459 (48%) for statins, beta-blockers, RASi and P2Y12i, respectively. We identified 5 to 6 distinct trajectories for each individual medication, with 80%, 72%, 74% and 73% having an adherent trajectory for statins, beta-blockers, RASi and P2Y12i, respectively. Trajectories reflecting low adherence to statins (HR 1.3, CI 1.1-1.5), RASi (HR 1.5, CI 1.3-1.7), and P2Y12i (HR 1.2, CI 1.1-1.4) within three years were associated with an increased risk of long-term mortality. Conclusions Non-adherence was prevalent in patients undergoing first-time coronary catheterization for CAD. Our comprehensive approach of classifying medication adherence by distinct trajectories of medication use revealed associations with various patient characteristics and worse outcomes. These findings could provide important considerations for targeted clinical intervention.Latent class analysis of medication use High vs. low adherence group outcomes