Clinical impact of medication adherence on 10‐year cardio‐cerebrovascular mortality in newly diagnosed hypertensive patients

Abstract The purpose of this study is to evaluate the impact of medication adherence on cardio‐cerebrovascular (CCV) mortality in newly diagnosed hypertensive patients. The authors retrospectively reviewed data from 20,836 patients who newly diagnosed hypertension from January 1, 2003 to December 31, 2005. Medication adherence was calculated from the compliance ratio (CR) during the first year after the diagnosis of hypertension. CCV mortality for 10 years was assessed according to the presence or absence of complications of hypertension. The risk of CCV death was significantly reduced in the CR ≥ 70% group than in the CR < 70% group (hazard ratio, 0.70; p = .004) for 10 years. In the patients without complications, the risk of CCV death was significantly lower in the CR ≥ 70% group than in the CR < 70% group (hazard ratio, 0.56; p = .014). However, in patients with complications, there was no significant difference in risk of CCV death between the CR ≥ 70% group and the CR < 70% group (hazard ratio, 0.79; p = .100). Only the CR ≥ 90% group had a significantly lower risk of CCV death (hazard ratio, 0.56; p < .001) for those with complications. Medication adherence is significantly associated with CCV mortality during 10 years in newly diagnosed hypertensives patients. Patients with complications of hypertension have to continue a high adherence rate (CR ≥ 90) for better long‐term clinical outcomes.

F I G U R E 1 Flow diagram of the study population. NHIS, National Health Insurance Service; HTN, hypertension; CCV, cardio-cerebrovascular controlled studies 1,2 and compared with low adherent patients (proportion of days covered, ≤ 40%), high adherers (proportion of days covered, ≥ 80%) showed a decreased risk of cardiovascular event. 3 It was also reported that patients with poor medication adherence (cumulative medication adherence, < 50%) had worse mortality from ischemic heart disease (IHD) (hazard ratio (HR), 1.64; 95% CI, 1.16-2.31), cerebral hemorrhage (HR, 2.19; 95% CI, 1.28-3.77), and cerebral infarction (HR, 1.92; 95% CI, 1.25-2.96) than those with good adherence (cumulative medication adherence, ≥ 80%). 4 Adherence to antihypertensive medication is an indicator that can evaluate the degree of control of hypertension and has a characteristic that it varies depending on various factors such as the passage of time and the occurrence of complications. Nevertheless, early medication compliance after the initial diagnosis of hypertension is one of the main factors that can predict the risk of CCV death.
However, there has been little long-term study to evaluate the impact of medication adherence on CCV mortality according to the presence and absence of hypertension complications. The purpose of this study is to evaluate the effect of medication adherence on CCV mortality in newly diagnosed hypertensive patients.

Data source
This study analyzed sample cohort data of 1,108,369 people randomly extracted from the database of the Korean National Health Insurance  (Table S1).  Study patients were divided into two groups according to their medication adherence to compare cerebrovascular mortality. We assessed how much decreased mortality in patients with adherence above a certain value compared to patients with adherence below a certain value.
We also divided patients into two groups according to whether they had CCV disease. In our study, 9,792 patients had a CCV disease and 11,044 patients didn't have the CCV disease. Previously, the relative reduction in mortality due to high adherence was evaluated. In the second analysis, whether the patient had CCV disease was additionally considered ( Figure 3).

Assessment of medication adherence
The proportion of days covered (PDC), is the recently preferred method of measuring medication adherence, 5,6 and we used compliance ratio (CR), which is a modified PDC for the purpose of our study.
The PDC is calculated by the ratio of the number of days the patient is covered by the medication to the number of days in a uniform period. ited the hospital after a longer period than the prescribed number of days, it was assumed that the patient took all the drugs during the prescribed number of days and did not take them for the rest of the period.
If the patient visited the hospital without completing the doctor's prescription days, it was assumed that the patient had taken all the medications by the day of the visit, and the medication adherence was not allowed to exceed 100%.  (Table S3).

Definitions of CCV disease
If CCV disease occurred during the observation period, the patient was considered to have had CCV disease. CCV disease was defined accord-

Statistical analysis
The characteristics of patients who were clinically diagnosed with hypertension for the first time and who took antihypertensives for more than a year were described using frequency and percentage. We In the Cox regression analysis, we adjusted for the age, sex, type of medical insurance, comorbidity (DM, dyslipidemia), family history (hypertension, strokes, heart disease), and whether CCV disease occurred during the observation period. We defined comorbidity as chronic diseases with high incidence or prevalence among those related to hypertension and included diabetes mellitus (ICD: E10-E14) and dyslipidemia (ICD: E78). Based on health examination data, the family history of hypertension, strokes, and heart disease was adjusted.
All statistical analyses were performed with SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Baseline characteristics
Baseline characteristics of 20,836 hypertensive patients according to 1 year of adherence to antihypertensive medication in this study were summarized in Table 1

Clinical outcomes
The median follow-up duration after the diagnosis of hypertension was 9.64 ± 1.26 years. The HR of CCV death was significantly lower in the CR ≥ 70% group than in the CR < 70% group (HR, 0.70; p = .004; Table 2). Similarly, the CR ≥ 80% and CR ≥90% groups had TA B L E 3 Association between adherence to antihypertensive medication within 1 year after diagnosis and CCV death in patients who didn't have or had CCV disease Each of HR of CR ≥70, CR ≥80, and CR ≥90 was calculated based on the mortality of the CR < 70, CR < 80, and CR < 90 group as a reference.
In the patients who didn't occur CCV disease during the observation period, the HR of CCV death was significantly lower in the CR ≥ 70% group than in the CR < 70% group (HR, 0.56; p = .014; Table 3). However, in patients with CCV disease during that period, there was no significant difference in risk of CCV death between the CR ≥ 70% group and the CR < 70% group (HR, 0.79; p = .100; Table 3). Of the patients with CCV disease, only the CR ≥ 90% group had a significantly lower risk of CCV death (HR, 0.56; p < .001; Table 3). Most research has shown differences in clinical outcomes between patients of poor adherence to antihypertensive therapy and good adherence. However, our study showed that better adherence to antihypertensive medication has a greater relative risk reduction in reducing CCV mortality.

LIMITATIONS
This study has several potential limitations. Because this study is a retrospective cohort study, we showed that there is a relationship between high adherence to antihypertensive medication and a reduction in the rate of CCV mortality. However, it was unclear whether there is a causal relationship between them. The CR is one of the indirect methods to measure adherence to medication using an electronic pharmacy database. 19 However, since there is no guarantee that the patients take all their prescribed medicine, actual adherence to antihypertensive medicine can be different from the CR. The recorded diagnosis using the International Classification of Diseases (ICD) from the database of the Korean NHIS is used for confirming the diagnosis of patients instead of medical records. This method can be different from the actual diagnosis. According to several studies, there is a possibility of ascertainment bias for the diagnosis of stroke due to the high fatality rate. 20 And a diagnosis of acute myocardial infarction has a sensitivity ranging from 60% to 80%. 21 We used the sample cohort extracted from the database of the Korean NHIS. There may be selection bias because there are a lot of missing data. In this study, patient factors such as psychological factors and level of knowledge of disease were not considered. Finally, chronic kidney disease is one of the major complications of hypertension and associated with a poor prognosis. 22 However, from the data of this study, we couldn't confirm the information on this. Since the study population was newly diagnosed with hypertension, it seems that the number of patients with chronic kidney disease may account for a very small proportion of the total population.

CONCLUSIONS
In conclusion, high adherence to antihypertensive medication can reduce the rate of CCV mortality. Furthermore, patients presented with the presence of complications at initial diagnosis of hypertension need more intensive control of adherence to antihypertensive medication (CR ≥ 90) for lowering CCV mortality than those presented without complications.