Stroke is the third leading cause of death and the chief cause of long-term disability. It is estimated that there are 750 000 strokes per year in the United States¹ and 5.7 million stroke deaths per year worldwide.² As the world's population ages, the concomitant collective stroke risk is expected to create an even larger health burden. However, much progress has been made in the primary and secondary prevention of stroke, with some improvements in the treatment of acute stroke.³ Evaluating long-term trends may provide insights into the effects of both the aging population and the positive developments of improved prevention and treatment on stroke incidence, lifetime risk, severity, and 30-day mortality.

Subjects and methods

The Framingham Study is a well-known community-based cohort of 9152 men and women predominantly of European descent that was initially started in 1948. Its subjects undergo biennial assessment of vascular risk factors and are under active surveillance for incident stroke and cause-specific mortality. It has the unique advantage of having up to 50 years of follow-up and consistent diagnostic criteria and personnel for stroke ascertainment. Age-adjusted annual incidence, stroke severity, 30-day mortality, and mortality-adjusted lifetime risk of stroke were determined over 3 periods: 1950-1977 (before imaging was available), 1978-1989 (during which computed tomography [CT] imaging was available), and 1990-2004 (during which CT and magnetic resonance imaging [MRI] were available). Log-linear Poisson regression analysis was used to estimate age-adjusted incidence. We also used logistic regression analysis to estimate severity and 30-day mortality, chi-square and t test analysis to compare exposure to risk factors, and modified Kaplan-Meier methods to compute mortality-adjusted lifetime risk over the 3 periods.

Results

Over 174 917 person-years of follow-up, 1030 incident strokes occurred, 450 of which occurred in men and 629 of which were atherothrombotic brain infarctions. Age-adjusted annual incidence was 7.6, 6.2, and 5.3 per 1000 person-years for the 3 periods (P = .02 for trend) for men, respectively, and 6.2, 5.8, and 5.1 per 1000 person-years (P = .01 for trend) for women, respectively. Lifetime risk of stroke for men at age 65 was 19.5%, 15.1%, and 14.5% for the 3 periods (P = 0.11), respectively, a reduction of approximately 30%. For women at age 65, the lifetime risk of stroke was 18%, 16.4%, and 16.1% (P = .61), respectively. The percentage of strokes considered moderate-to-severe when adjusted for age was not significantly different between men and women. The 30-day mortality rate decreased in men, from 23% to 20% to 14% over the 3 periods (P = .01), respectively, but not in women, with a rate of 21% for the first 2 periods and 20% for the last period (P = .32).

Discussion

Previous studies examining temporal trends in stroke incidence and outcomes have shown either a stable or, more often, a decreasing age-adjusted incidence over time. Most studies have used shorter follow-up periods. These trends have often been attributed to better risk factor control.^4-10 The trend in decline was reported to diminish in the 1980s and 1990s in many studies and was thought to be secondary to advances in diagnostic sensitivity, resulting from the development of newer imaging techniques, such as CT and MRI.^11,12 Our study failed to show this attenuation, perhaps because of a longer follow-up period as well as an emphasis on and success in applying uniform clinical diagnostic criteria over the 50 years studied.

Lifetime risk, which was previously reported as 1 in 5 for men and 1 in 6 for women,¹³ showed a decreasing trend over the 3 periods but failed to reach statistical significance. We also assessed the 10-year cumulative incidence as an intermediate-term assessment of stroke risk and found that it had significantly declined over the past 50 years as well. These changes were not trivial, however, accounting for approximately a 37% lower risk in men from the first to the third periods. Given that annual incidence and 10-year cumulative risks are declining, the lack of a statistically significant trend or decline in lifetime risk is likely secondary to small numbers, the increasing life expectancy, and the resulting higher-risk population having strokes at an older age.

Stroke severity has also been reported to be decreasing in previous studies because of inclusion of transient ischemic attacks and better imaging in these studies.^5,6,8,12,14 Despite increasing numbers of subjects receiving imaging in our study, with 90.6% of subjects undergoing either MRI or CT scanning in the last period, we did not see a decrease in severity. The severity of stroke was stable in men over the 3 periods. In women, the percentage of moderate-to-severe stroke increased dramatically from the first to the second period and was stable through the third. After adjusting for age, however, this was found to be statistically nonsignificant. The rates for 30-day mortality significantly decreased in men but not in women over the 3 periods. Both the severity and mortality trends suggest that women had strokes at an older age with more attendant disability¹⁵ and greater mortality.

Conclusions

In the Framingham cohort of subjects of European descent, the incidence of initial stroke has decreased significantly in men and women over the past 50 years. Lifetime risk of stroke also declined, but failed to reach statistical significance. Severity of stroke remained unchanged, and the 30-day mortality improved in men only.

Improvements in the treatment of risk factors have resulted in a decrease in incidence of initial stroke over the past 50 years. This trend was likely offset by increasing life expectancy, particularly in women, whose greater longevity has resulted in more severe strokes at an older age with no change in overall 30-day mortality. Further efforts are needed to continue primary prevention to improve the lifetime risk, severity, and 30-day mortality of stroke.

References