As with prior analyses, CCB or statin patients who start on SPAA are more likely to be adherent to antihypertensive and statin therapy in the first six months than are patients who add a CCB to statin or a statin to CCB as 2 separate pills [10, 11]. As an extension of increased adherence due to single pill advantages, this study found that greater adherence to hypertension and dyslipidemia therapy appears to have translated into a lower risk of CV events over time compared to non-adherent patients.
Slightly over 56% of the 1537 SPAA patients had at least 80% PDC adherence in the six months following initiation of therapy, compared with 21% of the 17,910 patients prescribed both a CCB and a statin. These adherence rates are consistent with other studies of single and dual-pill treatment of naïve patients with antihypertensive or statin therapy. In a study by Jackson et al.,  the effect of additional pills was evaluated as to its impact on patient adherence to medication, specifically measured via the medication possession ratio (MPR). Findings from this study suggest that an inverse relationship exists between additional medication tablets (pills) and patient MPR, as measured among patients receiving antihypertensive therapy in a managed care setting. MPR values were reduced from 75.4% among patients with a 2-tablet amlodipine regimen to 60.5% among patients with a 3-tablet amlodipine regimen. In another study with similar adherence findings to this study, Gerbino et al.  also showed a positive relationship between utilization of the fixed dose regimen and patient adherence, with MPR-based adherence measured at nearly 20% less among patients with ACE inhibitors plus CCB versus patients with a fixed dose amlodipine-benazepril.
This study demonstrates that patients' risk of cardiovascular events was significantly decreased among adherent patients compared with non-adherent patients. Unadjusted results found that the 12-month cardiovascular event incidence rate was only 1.88 per 100 person-years for adherent patients compared with 2.47 per 100 person-years in non-adherent patients. Adherence to either of the regimens included in the study was associated with a significantly lower risk of CV event (HR = 0.77, p = 0.003) after adjusting for potentially confounding baseline characteristics in multivariable Cox proportional hazards models.
This association between adherence and cardiovascular risks is in agreement with previous studies, including the 2007 study by Munger et al.  In this review, medication nonadherence was found to be responsible for several adverse health and economic outcomes, including an increased risk of death among patients with a prior myocardial infarction, an estimated annual cost of $396 to $792 million, and 33% to 66% of medication-related hospital admissions. Sever et al. [34, 35] found that three years of atorvastatin therapy produced a 79% reduction in coronary heart disease related events, from 22.8 events per 1,000 patient years to 4.8 events per 1,000 patient years. That study also found benefits of amlodipine and atorvastatin in reducing nonfatal MI by 46% [hazard ratio 0.54, confidence interval (CI) 0.40-0.72, P < 0.0001], stroke by 37% [hazard ratio 0.63, CI 0.46-0.87, P = 0.004] and total cardiovascular events and procedures by 27% [hazard ratio 0.73, CI 0.63-0.86, P < 0.0001].
This study has several limitations worth noting. Since PDC calculations are based on the assumption that patients take all medications for which they have prescriptions filled, these measures may overestimate adherence. Additionally, these adherence calculations fail to account for the possibility that patients received medications from sources other than the pharmacies included in the database used in this study.
Because of the way adherence was calculated in this analysis (patients had to remain on both CCB and statin to be considered adherent), our adherence rates may appear low relative to what has been reported in the literature. However, given that the patients in this analysis were prescribed both drugs, we believe patients should be considered nonadherent for the purposes of this study if they discontinue either CCB or statin.
Adherence was measured in a time period separate and distinct from the period during which CV events were identified and recorded. Due to this fact, it is possible that CV events occurred in the 6-month adherence measurement period and were not accounted for in our analysis, or that patient adherence measured prior to CV event monitoring is not representative of refill behavior had adherence and events been measured concurrently. To the extent that patients' adherence to their index medications differed between the 6-month period immediately following initiation of therapy and the minimum 12-month subsequent period, the estimates for adherence may vary from the time-specific values.
Cardiovascular events were identified by healthcare claims containing specific ICD-9 diagnosis codes. Due to the potential for unreported, misreported or miscoded cardiovascular events, the estimates for CV event incidence may overestimate or underestimate the actual number of clinical events. Since this limitation is similar for both cohorts, we do not expect it to bias the analysis for or against a cohort.
Two final limitations exist relative to analysis using a retrospective cohort design and adjudicated healthcare claims. Findings in this study are representative only of the U.S. commercially insured population of patients, not the overall population of treated patients who may have other forms of healthcare coverage (Medicaid, Medicare, etc.) not captured through this study methodology. Additionally, factors associated with both patient adherence and the incidence of cardiovascular events are limited within this study to those elements available through health plan enrollment files and insurance claims. Unmeasured and unknown confounding factors related to both baseline characteristics and clinical outcomes may exist, and their effect on these results cannot be accurately quantified.