Antihypertensive Drug Use and Adherence After Stroke

Data Sources and Study Population

Using the Registry of the Canadian Stroke Network, we identified all patients admitted to hospital with acute ischemic or hemorrhagic stroke at 11 stroke centers in Ontario, Canada (July 2003 through March 2006). Participating centers, tertiary care institutions with stroke care resources and expertise, represent different geographic regions and account for ≈20% of all stroke admissions for the province. The registry prospectively identifies all consecutive patients presenting to the emergency department or admitted to hospital with a diagnosis of stroke or transient ischemic attack.¹³ Trained research nurses collect data by chart review and interview with hospital care providers. This registry provides detailed baseline information including age, sex, comorbid conditions (including preexisting hypertension, depression, tobacco use, and hyperlipidemia), stroke type, stroke severity based on the validated Canadian Neurological Scale,¹⁴ and baseline systolic and diastolic blood pressure. Inter-rater reliability of these variables in a random sample of patient charts was moderate to excellent (κ 0.6 to 1.0).¹⁵

Prescribing Data

Using unique patient identifiers, stroke patients identified in the registry were linked to the Ontario Drug Benefits Database, which contains information on antihypertensive drug prescriptions, including the quantity and dates of drugs dispensed as well as the number of days supplied from each prescription, for patients ≥65 years of age. Residents may fill prescriptions at any outpatient pharmacy in Ontario with a maximum copayment of $6.11 (Canadian) for each prescription after a yearly $100 (Canadian) deductible. Low income seniors have a $2 (Canadian) maximum copayment with no yearly deductible. Using postal codes, patients in the registry are also linked to data from the 2006 Canada Census to determine median neighborhood income.

For the prescribing analysis, we determined whether patients ≥66 years of age filled a prescription for a diuretic, ACE inhibitor, or ACE inhibitor/diuretic combination before admission for acute stroke, within 3 months and 1 year after discharge from hospital. We also determined prescribing and adherence to angiotensin receptor blockers (ARBs), beta blockers, and calcium channel blockers, antihypertensive agents that have fewer data to support their use in secondary prevention of stroke. Patients were excluded from the main analysis if they were nonresidents of Ontario, if they had been transferred to a rehabilitation or palliative care facility (because drug data for these patients are not captured in the Ontario Drug Benefits Database), or if they had died within 3 months of discharge from hospital. Patients transferred to a rehabilitation facility who were subsequently discharged home were analyzed separately.

Adherence

One-year adherence was assessed for those who filled a prescription for any antihypertensive drug within 3 months of discharge from hospital for acute stroke. We analyzed patients who survived 1 year beyond this 3-month prescribing window to allow for a uniform assessment of adherence. We also separately evaluated adherence among patients who had died during this 1-year period. Adherence was assessed using prescription claims data that highly correlate with home inventory, pill count, as well as serum measures of drug presence.^16–18 One-year proportion of days covered (PDC) was determined for each patient and each drug class. One year PDC represents the number of days a patient had a medication available during the year divided by 365 days. For patients who had died during this time period, we used survival time since the first prescription fill as the denominator. PDC values were categorized as ≥0.8 to reflect high adherence and <0.8 to reflect suboptimal adherence. PDC values <0.4 indicate low adherence. The PDC cut point of ≥0.8 is associated with improved blood pressure control and reduced mortality relative to other lower levels of adherence among patients with cardiovascular disease.^19–21

Statistical Analysis

Differences in baseline characteristics, 1 year prescribing, and levels of adherence according to patient sex were assessed using χ² testing for categorical values and t tests for continuous variables. Missing data were excluded from the analysis (area level income data were missing in 45 [2.6%] men and 78 [4.2%] women). We used multivariate logistic regression modeling to evaluate the independent effect of patient sex on adherence. We used backward selection modeling with removal of nonsignificant cofactors one by one from the cofactors with largest nonsignificant P value. Eventually, only significant cofactors (P<0.05) were retained in the final models to give adjusted odds ratios with 95% CIs. Cofactors included variables known to be associated with adherence: age, previous use of antihypertensive agents, comorbid conditions, total number of baseline drugs used, history of depression, and socioeconomic status. Given that severity of stroke may also impact drug adherence, we also considered type of stroke and Canadian Neurological Scale score for each antihypertensive drug class. All assumptions of logistic regression modeling were satisfied. Statistical significance was defined as a 2-tailed P<0.05. All analyses were performed using SAS software version 9.1 (SAS Institute Inc).

Baseline Characteristics

There were 5870 patients ≥66 years of age identified with acute stroke, of whom 1482 were discharged to rehabilitation facilities (53% women). After exclusions, there were 3571 patients for the main analysis (Figure). Women were more likely to be older and have a previous diagnosis of hypertension with higher baseline systolic blood pressure, whereas men were more likely to have coexisting vascular disease. Women were more likely to present with subarachnoid hemorrhage and greater stroke severity compared with men (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Patient flow for prescribing and adherence study cohorts.

Prescribing of Antihypertensive Agents

Before admission to hospital for acute stroke, women were more likely than men to be receiving ≥1 antihypertensive agent (76% versus 69%; P<0.001). After stroke, 84.7% of men and 87.6% of women received ≥1 antihypertensive agent (Table 2). Most patients who received antihypertensive therapy received ≥2 antihypertensive agents rather than monotherapy. The most common monotherapy for both sexes was calcium channel blocker. Men were more often prescribed ACE inhibitor monotherapy compared with women. The most common combination antihypertensive regimens were ACE inhibitor and diuretic-based. There was no sex difference in prescribing of ACE inhibitor/diuretic combination, but women were more likely to be prescribed thiazide diuretics, beta blockers, calcium channel blocker, and ARBs relative to men. This prescribing pattern was generally similar among patients discharged to rehabilitation facilities. Patients who died within 1 year of filling their antihypertensive prescription were less likely to be prescribed ACE inhibitor (41.7% versus 50.4%; P<0.001) or ARB (6.3% versus 9.6%; P=0.025) compared with those who survived 1 year. However, there were no sex differences in prescriptions among those who died.

Adherence to Antihypertensive Agents

Among the 3142 patients who survived 1 year from the initial prescribing period, the proportion of patients highly adherent to antihypertensive drugs ranged from 61.8% to 75.8% (Table 3). There were no significant sex differences in adherence to any of the antihypertensive agents, even after adjusting for baseline variables. Similarly, there were no sex differences in adherence in patients discharged to rehabilitation facilities. Adherence was lowest among patients using ACE inhibitor/diuretic combination and indeed was significantly lower for patients using combination antihypertensive therapy in general compared with monotherapy (P<0.001). Adherence was also lower among those with newly initiated antihypertensive therapy after stroke: 58% highly adherent in the diuretic group, 66% for ACE inhibitors, and 67% for beta blockers. In the adjusted analysis, previous use of the medication was a significant predictor of adherence in all models. Greater stroke severity (Canadian Neurological Scale score <8) was significantly associated with high adherence in the ACE inhibitor and calcium channel blocker models. Factors such as depression, living alone, socioeconomic status, and total number of baseline medications were not significantly associated with adherence.

Adherence was lower among those 429 patients who died within 1 year of stroke compared with those who survived 1 year even when accounting for differences in follow-up time and patient sex (high PDC ≥0.8 63% versus 67%; low PDC <0.4 7.1% versus 4.4%; P=0.04). Low PDC (<0.4) was also associated with an increased risk of death within 1 year compared with higher PDC levels (odds ratio, 1.75; 95% CI, 1.07 to 2.87; P=0.025). There were no sex differences in adherence to antihypertensive therapy in patients who died within 1 year of stroke.

Conclusion

Previous studies identified that women have greater stroke deaths compared with men. However, in our study, elderly women were just as or more likely than men to receive evidence-based antihypertensive therapy. Despite experiencing a devastating event such as stroke, adherence to therapies known to substantially reduce morbidity and mortality after stroke was similarly poor in women and men. Low adherence to antihypertensive therapy was associated with increased risk of death after stroke. Greater efforts are required to improve antihypertensive drug adherence among women and men, including implementation of follow-up programs after stroke, academic detailing, and educational strategies.