Predicting stroke risk in hypertensive patients with coronary artery disease: a report from the INVEST

Abstract

Background and purpose: Our understanding of factors influencing stroke risk among patients with coronary artery disease is incomplete. Accordingly, factors predicting stroke risk in hypertensive, clinically stable coronary artery disease patients were determined with data from the INternational VErapamil SR-trandolapril STudy (INVEST).

Methods: The effect of baseline characteristics and on-treatment blood pressure (BP) were analyzed to determine the risk of stroke (fatal or nonfatal) among the 22 576 patients enrolled. Cox proportional-hazards models (unadjusted, adjusted, and time dependent) were used to identify predictors of stroke among subgroups with these characteristics present at entry and on-treatment BP.

Results: Excellent BP control (at 24 months, >70% <140/90 mm Hg) was achieved during 61 835 patient-years of follow-up, as 377 patients had a stroke (6.1 strokes/1000 patient-years) and 28% of those patients had a fatal stroke. Increased age, black race, US residency, and history of prior myocardial infarction, smoking, stroke/transient ischemic attack, arrhythmia, diabetes, and coronary bypass surgery were associated with an increased risk of stroke. Achieving a systolic BP <140 mm Hg and a diastolic BP <90 mm Hg was associated with a decreased risk of stroke. There was no statistically significant difference in stroke risk comparing the verapamil SR-based with the atenolol-based treatment strategy (adjusted hazard ratio=0.87; 95% CI, 0.71 to 1.06; P=0.17).

Conclusions: Among hypertensive patients with chronic coronary artery disease, stroke was an important complication associated with significant mortality. Black race, US residency, and conditions associated with increased vascular disease severity and arrhythmia predicted increased stroke risk, whereas achieving a BP <140/90 mm Hg on treatment predicted a reduced stroke risk.

Figure 1

Independent predictors of increased risk for stroke during follow-up. Baseline characteristics associated with stroke during follow-up were selected by the procedure and included in the model when P≤0.10. Two baseline revascularization variables were input into the stepwise model: CABG with or without PCI and PCU only.

Figure 2

Risk of stoke by category of SBP during follow up. HR for SBP <140 mm Hg from a stepwise Cox proportional-hazards model within each subgroup is shown. *Adjusted HR (95% CI) for SBP <140 mm Hg vs SBP ≥140 mm Hg in a Cox model with time-dependent SBP in each subgroup after stepwise selection, excluding SBP measurements within 6 weeks before stroke/censoring.