Long-Term Risk and Predictors of Recurrent Stroke Beyond the Acute Phase
Background and Purpose—Previous studies have shown heterogeneous results on predictors and rates of stroke recurrence. This study set out to investigate the long-term risk and predictors of recurrent stroke in Northern Sweden 1995 to 2008.
Methods—In the population-based Monitoring Trends and Determinants of Cardiovascular Disease (MONICA) stroke incidence registry, stroke survivors of either ischemic stroke or intracerebral hemorrhage were followed for recurrent stroke or death. Cox regression was used to identify predictors of stroke recurrence.
Results—The study comprised 6700 patients and 26 597 person-years. During follow-up, 928 (13.9%) patients had a recurrent stroke. Comparison between the first time period (1995–1998) and the last (2004–2008) showed declined risk of stroke recurrence (hazard ratio, 0.64 [95% confidence interval, 0.52–0.78]). Previous myocardial infarction was less prevalent in the most recent cohort (P<0.001). Predictors of stroke recurrence were age (hazard ratio, 1.03 [95% confidence interval, 1.02–1.04]) and diabetes mellitus (hazard ratio, 1.34 [95% confidence interval, 1.15–1.57]). After an index intracerebral hemorrhage (n=815), a major part of recurrent events were ischemic (63%), and compared with the ischemic stroke group (n=5885), a tendency toward lower risk of recurrence was observed.
Conclusions—Despite declining recurrence rates in this relatively young stroke population, almost one third are either dead or have experienced a second stroke in 5 years.
Stroke recurrence rates have varied substantially across previous studies.1 Previously reported predictors of long-term risk of recurrence include advanced age, diabetes mellitus, previous myocardial infarction, smoking, and atrial fibrillation.2 After an intracerebral hemorrhage (ICH), recurrent ischemic and hemorrhagic events occur at about the same rate.3–5 This study set out to investigate the risk and predictors of stroke recurrence and survival in the population of Northern Sweden (1995–2008).
Material and Methods
In the Northern Sweden MONICA stroke incidence registry, stroke events in the ages 25 to 74 years are validated according to World Health Organization stroke criteria.6 Recurrences within 28 days of initial stroke are not recorded in the registry.
Between 1995 and 2008, we identified 6700 patients with a first-ever ischemic stroke (IS) or ICH (index stroke), who survived the first 28 days. Subjects were followed until they experienced a recurrent stroke, died, moved out of the study region, turned 75 years, or until study end date (December 31, 2008). Recurrent stroke included also subarachnoidal hemorrhage and unspecified stroke. Information on risk factors was retrieved at the time of the index stroke (for details, see online-only Data Supplement).
Three cohorts were defined: index stroke occurring between 1995 and 1998 (n=2210), 1999 and 2003 (n=2384), and 2004 and 2008 (n=2106). This study was approved by the Regional Ethical Review Board at Umeå University.
The χ2 test or t test was used for simple group comparisons. Kaplan–Meier survival curves and Cox proportional hazard regression were used to analyze time to and predictors of stroke recurrence. Patients who died, moved out of the study region, turned 75 years, or reached study end date without experiencing a recurrent stroke were censored.
The study comprised 26 597 person-years of follow-up. The mean follow-up time was 4.0 years (median, 3.1 years). There were 5885 patients with IS (87.8%) and 815 with ICH (12.2%). Patients with ICH tended to be younger and had a lower prevalence of traditional cardiovascular risk factors than patients with IS (Table 1).
There were 928 recurrent events overall, and the cumulative risk of stroke recurrence was 6% at 1 year, and the 5- and 10-year cumulative risk of stroke recurrence was 16% and 25%, respectively. Corresponding figures for a combined end point of recurrent stroke or death from any cause were 10%, 28%, and 45%, respectively (Figure). Adjusted annual risk of stroke recurrence is shown in Table I in the online-only Data Supplement.
Age and diabetes mellitus were associated with recurrent stroke in the total study population, hazard ratio 1.03 (95% confidence interval, 1.02–1.04) and hazard ratio 1.34 (95% confidence interval, 1.15–1.57), respectively. Predictors of a combined end point of recurrent stroke or death were age, atrial fibrillation, hypertension, previous myocardial infarction, sex, cohort, and diabetes mellitus. Patients in the most recent cohort (2004–2008) had a 36% lower risk of recurrent stroke than patients in the first cohort (1995–1998; Table 2; Figure I in the online-only Data Supplement). Previous myocardial infarction (P<0.001) was less prevalent in the most recent cohort, whereas other potential risk factors did not vary over time (Table II in the online-only Data Supplement).
Of all recurrent events after ICH, 62% (63 of 101) were ischemic. After IS, 9% (76 of 827) of recurrences were ICH (Table III in the online-only Data Supplement). There was a tendency toward lower risk of recurrence after ICH than after IS (hazard ratio, 0.86 [95% confidence interval, 0.70–1.06]; Figure II in the online-only Data Supplement.)
Based on uniform diagnostic criteria throughout the study period, we report a decline in long-term risk of recurrent stroke. The definition of recurrent stroke affects the reported recurrence rates,7 and our comparatively low figures are most likely because of the inherited feature of the registry that early recurrence is not recorded. Furthermore, the study population is relatively young.
The burden of cardiovascular risk factors in the ICH group was smaller, and they tended to be at lower risk of recurrence compared with patients with IS. This finding should be interpreted in the light of a high 28-day case fatality in ICH and needs confirmation.
The 2 most consistently published risk factors of recurrent stroke, age and diabetes mellitus, were also found in our study. When assessing a combined end point of death or stroke recurrence, additional cardiovascular risk factors showed significant association.
The 36% reduction in stroke recurrence over time is substantial, and this finding is in line with a meta-analysis by Mohan et al1 and with previous data from Northern Sweden.8 Apart from myocardial infarction before the index stroke, the distribution of assessed risk factors in our study did not change significantly over time. However, in Northern Sweden, there have been major improvements in cardiovascular risk factors, such as less smoking and lower lipid levels during the past 2 decades.9 Other factors possibly explaining the decline include improved secondary prevention, changes in acute treatment,10 and increased capacity for stroke unit care.11 All factors were not assessed in our study.
Assessment of the ICH population revealed that almost two thirds of recurrences were ischemic. This should be further explored because it may have implications for optimal secondary prevention measures.
Despite declining rates, stroke recurrence remains an important clinical problem in this relatively young stroke patient population. Our data show that among patients who have survived past day 28 after an index stroke event, almost one third are either dead or have experienced a second stroke in 5 years time.
We are thankful to the Northern Sweden MONICA study.
Sources of Funding
The MONICA stroke registry is funded by Västerbotten County Council. The study was also supported by Visare Norr and the Swedish Stroke Foundation.
Presented in part at the European Stroke Conference, Lisbon, Portugal, May 22–25, 2012, and was acknowledged with the Young Investigator Award.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.005060/-/DC1.
- Received February 25, 2014.
- Revision received March 31, 2014.
- Accepted April 2, 2014.
- © 2014 American Heart Association, Inc.
- Mohan KM,
- Wolfe CD,
- Rudd AG,
- Heuschmann PU,
- Kolominsky-Rabas PL,
- Grieve AP
- Zia E,
- Engström G,
- Svensson PJ,
- Norrving B,
- Pessah-Rasmussen H
- Hanger HC,
- Wilkinson TJ,
- Fayez-Iskander N,
- Sainsbury R
- Hill MD,
- Silver FL,
- Austin PC,
- Tu JV
- Coull AJ,
- Rothwell PM
- Rautio A,
- Eliasson M,
- Stegmayr B
- Eriksson M,
- Jonsson F,
- Appelros P,
- Asberg KH,
- Norrving B,
- Stegmayr B,
- et al