Statin Adherence After Ischemic Stroke or Transient Ischemic Attack Is Associated With Clinical Outcome
論文ID: CJ-15-0753
この記事には本公開記事があります。
詳細
論文ID: CJ-15-0753
Background: Statin therapy is beneficial for ischemic stroke patients, but little is known about whether statin adherence affects clinical outcome. We therefore evaluated the effect of statin adherence in patients with ischemic stroke or transient ischemic attack (TIA).
Methods and Results: From Taiwan Bureau of National Health Insurance database, we enrolled patients with no prior statin therapy admitted for ischemic stroke or TIA between January 2002 and December 2005. Patients were grouped based on statin adherence according to medication possession ratio (MPR): good adherence (MPR >80%; n=2,274), intermittent adherence (MPR=40–80%; n=3,710), and poor adherence (MPR <40%; n=9,424). The study endpoint was the composite outcome of recurrent ischemic stroke, hemorrhagic stroke, and acute coronary event 1 year after statin initiation. Follow-up data were obtained through December 2010. During follow-up, composite endpoints occurred in 5,354 patients (34.7%): good adherence, 798 patients (35.1%); intermittent adherence, 1,338 patients (36.1%); and poor adherence, 3,218 patients (34.1%). Compared with the good adherence group, patients in the poor adherence group and intermittent adherence group had higher risk of worse clinical outcome (adjusted HR, 1.26 and 1.16, respectively; 95% CI: 1.17–1.37 and 1.07–1.27, respectively).
Conclusions: Good statin adherence was associated with better clinical outcome in patients with acute ischemic stroke or TIA.
Stroke is one of the leading causes of death and disability worldwide,1,2 especially in industrialized countries, and accounts for approximately 2–4% of health-care costs.1 Ischemic stroke comprises approximately 80% of all stroke,3 and patients with ischemic stroke have a high risk of recurrent cardiovascular events during follow-up.4–6 Statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors), are powerful lipid-lowering agents. Over the past decade, several studies have shown that statins are highly efficacious for the primary and secondary prevention of coronary artery disease (CAD),7–13 and statin use is recommended in patients with CAD.14–17 Moreover, statin therapy is also associated with better clinical outcomes in patients with ischemic stroke and transient ischemic attack (TIA).18–20 Modern guidelines for stroke management from the American Heart Association/American Stroke Association (AHA/ASA) recommend that statin therapy be used in the vast majority of patients with a history of ischemic stroke and TIA.21 Furthermore, established pharmacoepidemiological studies emphasize the importance of statin adherence in clinical practice.22–24 Good statin adherence in patients with cardiovascular disease leads to the subsequent reduction of recurrent cardiovascular morbidity. Although early discontinuation of statin therapy is correlated with a higher mortality rate in patients with ischemic stroke,25 there is little evidence on the effect of adherence to statin therapy on recurrent cerebrovascular or cardiovascular events in ischemic stroke patients. Furthermore, statin therapy may increase the incidence of intracranial hemorrhage in ischemic stroke patients,20,26,27 but the safety of long-term statin use has never been formally assessed. Therefore, the aim of this study was to investigate the potential effect of adherence to statin use on clinical outcome in patients with ischemic stroke.
Editorial p ????
This cohort study used the population-based Taiwan National Health Insurance Research Database (NHIRD), which is derived from the claims data of the NHI program, initiated in 1995, from more than 99% of Taiwan’s approximately 23 million residents. The available data for each patient in the NHIRD includes demographic characteristics; medical diagnosis (up to 5 per inpatient; up to 3 per outpatient), procedure, expenditure, and all detailed prescriptions. Personal identity has been encrypted for privacy protection, but all data sets can be linked with the unique and anonymous identifiers created by data providers for research purposes.
CohortWe screened patients who were admitted to hospital for ischemic stroke or TIA between 1 January 2002 and 31 December 2005. Eligible patients were enrolled based on the principal diagnosis taken from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9 CM) codes recorded in the NHIRD. Those diagnosed with ICD-9 CM codes 433, 434, 436, 437.1, and 437.9 were classified as having ischemic stroke, and those with 435 as having TIA (n=231,880). Patients were excluded if they: (1) were aged <18 years (n=810); (2) had been admitted to hospital with the diagnosis of ischemic stroke or TIA within 1 year before index admission (n=4,698); (3) had statin therapy within 90 days before index admission (n=11,904); (4) had incomplete registration information (n=4,443); (5) had no available follow-up data after discharge (n=12,767); or (6) had no available data to calculate medication possession ratio (MPR; n=3,356).
Statin AdherenceAmong the remaining 193,902 patients, 15,408 patients had statin therapy initiated during hospitalization or within 3 months after discharge and were grouped based on statin adherence. The statin adherence of each patient was defined using MPR,28 which was calculated as the total days covered by prescribed statins in the 1 year after initiation of statin therapy divided by 365 days. Eligible patients were divided into 3 groups according to MPR: good adherence (n=2,274; mean age, 65.2 years; male, 49.9%; MPR >80%); intermittent adherence (n=3,710; mean age, 64.8 years; male, 50.8%; MPR=40–80%), and poor adherence (n=9,424; mean age, 65.1 years; male, 51.7%; MPR <40%; Figure 1).
Subject selection. TIA, transient ischemic attack.
The primary outcome of interest was a composite endpoint of recurrent ischemic stroke (ICD-9 codes 433, 434, 436, 437.1, 437.9), hemorrhagic stroke (ICD-9 codes 430, 431, 432), or acute coronary event (ICD-9 code 410), whichever came first. Secondary endpoints included individual components of the primary endpoint. Patients were identified as having incurred the composite endpoint based on the diagnosis recorded in the ambulatory and hospitalization claims of the NHIRD. All patients were followed up from the date of first statin prescription through their last medical record before the end of the study period (31 December 2010). Patients who did not reach the outcomes of interest were defined as censored. Information on ischemic stroke diagnosis in the NHIRD was validated in our previous study,29 and 97.85% of the ischemic stroke cases recorded were confirmed by medical records.
Statistical AnalysisBasic characteristics, comorbid disease, and medical history were recorded and analyzed from the data registered in the NHIRD. Conventional risk factors and comorbid diseases correlated with patients with ischemic stroke were retrieved based on ICD-9 CM codes (Table S1) from both inpatient and outpatient claims databases for 1 year before and during the index ischemic stroke hospitalization. The Charlson comorbidity index adjustment for ischemic stroke outcome was used to evaluate the severity of comorbidity for each ischemic stroke patient.30 Continuous variables are presented as mean±SD, and categorical variables are expressed as the number and percentage of patient. Student’s t-test was used to compare normally distributed continuous variables between each group, and chi-squared tests were used to compare categorical variables. Significance was set at P<0.05 (2-tailed). The event-free time to the composite endpoints is presented as a Kaplan-Meier survival curve (Figure 2) stratified according to statin adherence. Multivariate Cox proportional hazards model with stepwise selection was used to examine the relationship between endpoints and MPR. Hazard ratios (HR) and their 95% confidence intervals (CI) were calculated. SAS for Windows 9.2 (SAS Institute, Cary, NC, USA) was used for all data analyses.
Kaplan-Meier survival curve for the probability of being event-free 96 months after acute ischemic stroke or transient ischemic attack according to statin adherence (medication possession ratio).
There were no differences in age, Charlson index scores, or most underlying comorbid diseases between groups (Table 1). There was, however, a significantly lower proportion of patients with prescriptions for comorbid disease in the poor adherence group. Furthermore, more patients in the poor adherence group were male. Statin type and dosage prescribed in patients with ischemic stroke or TIA are listed in Table S2. Most patients in each group received low-moderate-intensity statin therapy.31
| Characteristics | Statin adherence | ||
|---|---|---|---|
| Good (n=2,274) | Intermittent (n=3,710) | Poor (n=9,424) | |
| Age (years) | 65.2±10.6 | 64.8±10.9 | 65.1±11.5 |
| Male | 1,135 (49.9) | 1,885 (50.8) | 4,876 (51.7)* |
| Charlson index | |||
| 0 | 719 (31.6) | 1,121 (30.2) | 2,818 (29.9) |
| 1 | 610 (26.8) | 1,049 (28.3) | 2,700 (28.6) |
| ≥2 | 945 (41.6) | 1,540 (41.5) | 3,906 (41.4) |
| Comorbidities | |||
| Hypertension | 1,727 (75.9) | 2,787 (75.1) | 6,956 (73.8)* |
| Diabetes mellitus | 1,029 (45.2) | 1,686 (45.4) | 4,184 (44.4) |
| Hyperlipidemia | 1,480 (65.1) | 2,514 (67.7)* | 6,210 (65.9) |
| Atrial fibrillation | 102 (4.5) | 167 (4.5) | 438 (4.6) |
| CAD | 632 (27.8) | 950 (25.6) | 2,534 (26.9) |
| Co-medication in previous 1 year | |||
| Anti-arrhythmic agents | 43 (1.9) | 60 (1.6) | 156 (1.7) |
| Anticoagulant agents | 41 (1.8) | 51 (1.4) | 93 (1.0)* |
| Antiplatelet agents | 1,344 (59.1) | 1,935 (52.2)* | 3,777 (40.1)* |
| Anti-diabetic agents | 701 (30.8) | 1,050 (28.3)* | 2,275 (24.1)* |
| ACEi/ARB | 853 (37.5) | 1,238 (33.4)* | 2,776 (29.5)* |
| β-blocker | 746 (32.8) | 1,160 (31.3) | 2,540 (26.9)* |
| CCB | 1,017 (44.7) | 1,548 (41.7)* | 3,335 (35.4)* |
| Diuretics | 521 (22.9) | 734 (19.8)* | 1,717 (18.2)* |
| Digoxin | 75 (3.3) | 121 (3.3) | 263 (2.8) |
| Other dyslipidemia agents† | 191 (8.4) | 284 (7.6) | 549 (5.8)* |
Data given as mean±SD or n (%). *P<0.05 compared with good adherence (reference) group (χ2 test for categorical data and independent t-test for continuous data). †Cholestyramine, fibrates, and niacin. ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CAD, coronary artery disease; CCB, calcium channel blocker.
During the follow-up period, 5,354 patients (34.7%) developed 1 of the primary composite endpoints: 798 (35.1%) were in the good adherence group, 1,338 (36.1%) in the intermittent adherence group, and 3,218 (34.1%) in the poor adherence group (Table 2). Patients in the poor adherence group had a significantly higher incidence of the primary endpoint (adjusted HR, 1.26; 95% CI: 1.17–1.37), compared with those in the good adherence group. Moreover, patients in the intermittent adherence group also had a higher rate of primary endpoints than those in the good adherence group (adjusted HR, 1.16; 95% CI: 1.07–1.27). The probability of events during the follow-up period was determined using Kaplan-Meier curve (Figure 2).
| Statin adherence | |||
|---|---|---|---|
| Good† (n=2,274) | Intermittent (n=3,710) | Poor (n=9,424) | |
| Follow-up (days) | 1,789±853 | 1,627±820 | 1,428±730 |
| Total follow-up (person-years) | 11,138 | 16,535 | 36,869 |
| Composite endpoints | 798 (35.1) | 1,338 (36.1) | 3,218 (34.1) |
| Crude incidence rate/1,000 person-years | 71.6 | 80.9 | 87.3 |
| Unadjusted HR (95% CI) | 1.00 (–) | 1.14 (1.04–1.25) | 1.24 (1.14–1.34) |
| Adjusted HR (95% CI) | 1.00 (–) | 1.16 (1.07–1.27) | 1.26 (1.17–1.37) |
Data given as mean±SD, n (%) or otherwise indicated. †Reference group. Adjusted variables: age, gender, Charlson index score, diabetes mellitus, anticoagulant agents, ACEi/ARB, calcium channel blockers, diuretics. Abbreviations as in Table 1.
Patients in the poor adherence group had a higher recurrence rate of ischemic stroke (adjusted HR, 1.20; 95% CI: 1.09–1.32), a secondary endpoint, than did those in the good adherence group (Table 3). They also tended to have a higher incidence of acute coronary events during the follow-up period (adjusted HR, 1.23; 95% CI: 0.96–1.58), but there was no significant difference in the incidence of hemorrhagic stroke between the 3 groups.
| Statin adherence | |||
|---|---|---|---|
| Good† (n=2,274) | Intermittent (n=3,710) | Poor (n=9,424) | |
| Follow-up (days) | 1,789±853 | 1,627±820 | 1,428±730 |
| Total follow-up (person-years) | 11,138 | 16,535 | 36,869 |
| Ischemic stroke | 524 (23.0) | 887 (23.9) | 2,110 (22.4) |
| Crude incidence rate/1,000 person-years | 47.0 | 53.6 | 57.2 |
| Unadjusted HR (95% CI) | 1.00 (–) | 1.13 (1.01–1.25) | 1.17 (1.07–1.29) |
| Adjusted HR (95% CI) | 1.00 (–) | 1.15 (1.03–1.28) | 1.20 (1.09–1.32) |
| Hemorrhagic stroke | 53 (2.3) | 68 (1.8) | 194 (2.1) |
| Crude incidence rate/1,000 person-years | 4.6 | 4.1 | 5.3 |
| Unadjusted HR (95% CI) | 1.00 (–) | 0.86 (0.60–1.23) | 1.08 (0.80–1.47) |
| Adjusted HR (95% CI) | 1.00 (–) | 0.85 (0.60–1.22) | 1.08 (0.80–1.47) |
| Acute coronary event | 78 (3.4) | 139 (3.7) | 300 (3.2) |
| Crude incidence rate/1,000 person-years | 7.0 | 8.4 | 8.1 |
| Unadjusted HR (95% CI) | 1.00 (–) | 1.21 (0.92–1.60) | 1.19 (0.92–1.53) |
| Adjusted HR (95% CI) | 1.00 (–) | 1.24 (0.94–1.64) | 1.23 (0.96–1.58) |
Data given as mean±SD, n (%) or otherwise indicated. †Reference group. Adjusted variables: age, gender, Charlson index score, diabetes mellitus, anticoagulant agents, ACEi/ARB, calcium channel blockers, diuretics. Abbreviations as in Table 1.
It is thought that ischemic stroke patients with good statin adherence might have good adherence to other medications, such as anti-platelet and anti-hypertensive agents, which contribute to better clinical outcome. For further evaluation of the beneficial effect of statin adherence, a subanalysis was conducted to exclude the influence of adherences to other medications on clinical outcome. We re-assessed the impact of statin adherence in ischemic stroke or TIA patients with different aspirin adherence (Table S3). In patients with poor aspirin adherence (aspirin MPR <40%), patients in the good statin adherence group had better clinical outcomes than those in the intermittent and poor statin adherence groups. Moreover, in patients with good aspirin adherence (aspirin MPR >80%), patients in the good statin adherence group had better clinical outcome than those in the intermittent adherence group. This shows that patients with good statin adherence had better clinical outcome, irrespective of adherence to other medication.
In the present cohort of 15,408 patients with ischemic stroke or TIA who had begun statin therapy while hospitalized or within 3 months after discharge, patients with good statin adherence had better clinical outcome than did patients with poor statin adherence.
Statins beneficially affect the primary and secondary prevention of cardiovascular and cerebrovascular disease by lowering lipid levels, decreasing oxidative stress and inflammation, and inhibiting the thrombogenic response, among other pleiotropic effects. Despite the benefits, however, a high proportion of patients discontinue their statin therapy. In 1 study of >34,000 elderly US patients, less than half were adherent users 1 year after beginning statin therapy; the most rapid decline occurred during the first half-year.32 Moreover, another study of >143,000 elderly Canadian patients reported that the 2-year statin adherence rate was only 40.1% in patients with acute coronary syndrome and 36.1% in patients with chronic CAD.33 Other studies reported that age >75 years old, female sex, concomitant cardiovascular medication, and higher medication copayments were predictors of poor statin therapy adherence.34,35 Non-adherent patients with CAD not only had a higher recurrence of cardiovascular events, but also a higher mortality rate during the year following their first cardiovascular event.22–24,36,37 The question of how to improve the statin therapy adherence rate in high-risk patients is a critical issue in preventive medicine.
Patients with cerebrovascular disease also have low adherence rates to statin therapy.25,38 One study of >600 ischemic stroke survivors reported that almost 39% of the patients discontinued statin therapy within 12 months of discharge (mean, 48.6±54.9 days).25 Older patients were more likely to discontinue statin therapy. In addition, statin discontinuation was an independent predictor of 12-month all-cause mortality (HR, 2.78; 95% CI: 1.96–3.72). A Canadian study of >112,000 patients on the effect of statin adherence on cerebrovascular disease in primary prevention found that patients with high statin adherence (MPR ≥80%) had a lower incidence of cerebrovascular disease than did those with low statin adherence (MPR <80%) during the follow-up period (RR, 074; 95% CI: 0.65–0.84).39 To the best of our knowledge, this is the first study to investigate the effect of statin adherence on patients with ischemic stroke or TIA in secondary prevention. We found that good statin adherence was associated with better clinical outcome, implying that patients with ischemic stroke or TIA should undergo long-term statin therapy. The mechanism underlying the beneficial effect of statin adherence on clinical outcome in such a population has not been elucidated, but in vivo and in vitro studies have reported that statin discontinuation led to vasoconstriction, pro-inflammatory and prothrombotic status, and endothelial dysfunction, which may partially contribute to the increasing recurrence of cardiovascular and cerebrovascular events.40
One major concern regarding statin therapy in patients with ischemic stroke is that statin therapy may be associated with an increasing incidence of hemorrhagic stroke. In the SPARCL study, patients with ischemic stroke taking atorvastatin had a 66% increase in hemorrhagic stroke compared with the control group.20 Similar findings were reported in post-hoc analyses of patients with prior cerebrovascular disease.26,27 These findings called into question the wide use of statins in patients with cerebrovascular disease, but we found no difference in the incidence of hemorrhagic stroke between the 3 groups in the present study, indicating the long-term safety of statin therapy in patients with cerebrovascular events. Furthermore, a recent Canadian study of >17,000 elderly patients (≥66 years old) who started statin therapy after an acute ischemic stroke re-assessed the correlation between statin use and the incidence of hemorrhagic stroke, and it was concluded that there was no association.41 Additional prospective controlled trials are needed to clarify this safety issue of statin therapy.
The strengths of the present study were the nationwide population datasets from the Taiwan NHI system, in which 99% of Taiwan’s residents are enrolled, and the longitudinal nature of the medical claims, which ensured sufficient information on diagnosis, health services utilized, and medications prescribed. The present assessment of the potential beneficial effects of statin adherence on clinical outcomes in patients with cerebrovascular diseases showed that long-term statin therapy did not increase the risk of hemorrhagic stroke.
There were several limitations in the present study. First, in contrast to the diagnosis of ischemic stroke and acute coronary event, which were validated,29,42 the diagnosis of hemorrhagic stroke was not validated in the NHIRD, and some events might have been misclassified. Based on the high accuracy rate, however, of the diagnosis of ischemic stroke in our previous study,29 which was also drawn from this population-based database, we believe that the chances of this occurring are negligible and non-differential. Second, we were unable to control for out-of-pocket purchase of statins, which could lead to misclassification of the exposure. Fortunately, this misclassification is unlikely because all statin products are reimbursable by Taiwan’s NHI. Third, serum cholesterol level was unavailable in the present study, therefore, we could evaluate neither cholesterol level after statin therapy nor the association between serum cholesterol level and the risk of hemorrhagic stroke. Patients who had taken statins before enrollment, however, were excluded from this study; baseline low-density lipoprotein was assumed to be ≥160 mg/dl, according to the reimbursement guidelines based on the National Cholesterol Education Program (NCEP) Adult Treatment Panel III guideline recommendation. Fourth, imaging to estimate the infarct size and status of hemorrhage transformation in the acute phase was lacking because of the limitations of the NHIRD. Fifth, we were not able to determine whether the patients actually took the statins when at home, but it is well known that statin is an effective lipid-lowering agent and has a beneficial effect on clinical outcome in patients with ischemic stroke. It was therefore thought that patients in the good adherence group would not have had a better outcome than those in the poor adherence group if they had not taken the statin. Finally, because of the inherent limitations of the claims data of the NHIRD, some individual characteristics (such as smoking status, body mass index, and dietary habits) that may contribute to the development of stroke were unavailable.
Patients with good statin adherence after an acute ischemic stroke or TIA had better clinical outcome than did those with poor statin adherence. Furthermore, long-term statin therapy did not increase the incidence of hemorrhagic stroke. The present findings indicate that long-term statin therapy in patients with acute ischemic stroke or TIA is essential. Further prospective randomized clinical trials are needed to confirm the present findings.
The authors declare no conflicts of interest.
Supplementary File 1
Table S1. ICD-9 CM codes
Table S2. Statin type and dosage in ischemic stroke or TIA patients
Table S3. Endpoints vs. aspirin adherence
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-15-0753
