Published online before print November 13, 2003, doi: 10.1161/01.STR.0000101751.20118.C1
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2003;34:2876.)
© 2003 American Heart Association, Inc.
Original Contributions |
Predictors of Death Among Long-Term Stroke Survivors
From the Institute of Community Medicine (T.E., E.A.) and Department of Geriatric Medicine (M.V.), University of Tromsø, Tromsø, Norway; and Department of Geriatric Medicine, University of Turku, Turku, Finland (M.V.).
Correspondence to Torgeir Engstad, Faculty of Medicine, Institute of Community Medicine, University of Tromsø, N-9037 Tromsø, Norway. E-mail Torgeir.Engstad{at}ism.uit.no
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Background and Purpose— We evaluated the risk factors for death among long-term stroke survivors compared with stroke-free subjects.
Methods— In 1997 we investigated 221 stroke survivors (mean, 9.4 years after index stroke) and 243 stroke-free subjects; both groups were recruited from a population-based health study. During the subsequent 5 years, all deaths (51 and 21 in the stroke and stroke-free groups, respectively) were registered.
Results— The age- and sex-adjusted total mortality rate for the 5-year follow-up was 21.0% in the stroke group and 7.9% in the stroke-free group (P<0.0001), depending on different rates of cardiovascular deaths (P<0.0001). Better physical and social functioning (P<0.0001) and moderate use of alcohol (P
0.004), the latter compared with no use, decreased risk of death, irrespective of stroke status. A myocardial infarction risk score had no impact on death in the stroke group, in contrast to the increased risk seen among the stroke-free group (P=0.0001).
Conclusions— The long-term stroke survivors had significant excess risk of death compared with stroke-free subjects. Better physical and social functioning and moderate use of alcohol were associated with decreased risk, whereas a myocardial infarction risk score increased risk for death only in the stroke-free group.
Key Words: mortality • risk factors • stroke
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Cerebrovascular accidents are the second leading cause of death in the world and the third in most industrialized countries.1 The case fatality rates of stroke vary over time and geographic areas,2 depending on different patterns of risk factors,3 stroke subtypes,4 severity of stroke,5 and diagnostic procedures associated with sensitivity and thereby severity of diagnosed stroke. Long-term stroke survivors have been exposed to risk factors over many years,6,7 and some of them may have additionally suffered a recurrent stroke and/or coronary heart disease. The main cause of death after a stroke is a cardiovascular event.8–10
The purpose of this population-based study was to examine the factors predicting death among long-term stroke survivors compared with stroke-free subjects.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Subjects
The subjects were recruited from a health survey of the Tromsø population in 1994, of which 27 159 subjects attended, which is equivalent to 77% of the eligible population. In 1997, 269 of 362 invited individuals (74.3%) with a self-reported history of stroke were examined together with 262 age- and sex-matched subjects who reported no stroke in 1994.11 Among these, a stroke was verified in 221 individuals, of which 20 had a subarachnoid hemorrhage, 14 an intracerebral hemorrhage, and 30 at least 1 recurrent stroke. In comparison, 243 had suffered neither a stroke nor a transient ischemic attack.11 The mean elapsed time from stroke onset to the examination in 1997 was 9.4 years (SD=7.2).
The Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate approved the study, and written consent was obtained from all participants.
Medical History
Questions about medical history, self-perceived health, urinary incontinence, use of drugs, marital status, economy, smoking habits, and consumption of alcohol were asked at the examination in 1997. Consumption of coffee was registered at the 1994 survey. An index characterizing the severity of urinary incontinence was calculated by multiplying the reported frequency by the quantity.12 Additional information was obtained from either a family member or a caregiver or by reading the hospital medical record when memory problems were suspected.
Measurements
The mean blood pressure was based on the average values measured in 1994 and 1997 by trained assistants using an automatic device (Dinamap).13 Change in blood pressure from 1994 to 1997 was calculated by subtracting the former from the latter. Body mass index was calculated as a proportion of weight divided by the height squared. Change in body mass index was the difference between the 1994 and 1997 measurements. Levels of total cholesterol in 1994 and 1997 and HDL cholesterol in 1994 were determined by an enzymatic colorimetric method (Hitachi 737). Fibrinogen was measured by an automatic clotting method (ACL 3000), and folic acid was measured by an immunochemical method (Immuno-1). All analyses were performed at the Department of Clinical Chemistry, University Hospital, in northern Norway.
Functional and Cognitive Status
A nurse assessed physical and social activities using a modified Frenchay Activity Index (FAI)14 and depressive symptoms using the Montgomery Aasberg Depression Rating scale.15 The variable of paid work in the FAI was excluded since the majority of the attendees had retired. A physician (T.E.) assessed global cognitive function using the Mini-Mental State Examination (MMSE).16
Myocardial Infarction Risk Score
The individual risk of suffering a myocardial infarction was computed by multiplying the risk value of systolic blood pressure, serum cholesterol, cigarettes smoked, sex, and familial myocardial infarction, an algorithm based on previous Norwegian follow-ups of healthy individuals.17,18 The purpose was to rank individuals within the same age group according to their risk of suffering a myocardial infarction during the next few years.17 This risk score has been used in several Norwegian health surveys19,20 and is also discussed in a Norwegian article21 and doctoral thesis.22 It confirms an almost linear correlation between increasing total risk score and mortality of myocardial infarction and sudden death.
Deaths During 5-Year Follow-Up
All deaths from August 1997 to January 2003 were identified by the national population register, in which all deaths are registered and updated with a delay of 14 days. A copy of the death certificate of each deceased person was found by searching the archives in the Community Health Service of Tromsø. The causes of death were validated and classified according to the International Statistical Classification of Diseases, 10th Revision by 2 physicians (T.E., M.V.), who read and discussed all available information from the medical records, including the death certificates.
Statistical Analysis
ANOVA tested differences between baseline characteristics in the stroke and the stroke-free groups, with adjustments for age and sex. Pearson correlation analysis was performed to exclude highly correlated predictors (r>0.5). Different distributions of the main diagnoses of death in the stroke and stroke-free groups were tested by 
2 test. The impact of the explanatory variables on death was assessed with a Cox proportional hazards regression model, with adjustment first for age and sex and then for all variables in the model.
All calculations were performed with SAS software.23
Baseline Characteristics
Table 1 shows that the stroke-free subjects reported more frequently excellent or very good health and less frequently impaired health the last year (P=0.0001). They also reported more frequently an optimistic attitude (P=0.0002) and eagerness for work (P<0001) and had greater MMSE (P<0.0001) and FAI (P<0.0001) scores compared with the stroke survivors. The stroke group had a more frequent use of self-reported and registered use of antihypertensive drugs (P=0.0001) as well as occurrence of myocardial infarction or diabetes mellitus (P=0.03). The stroke group also had a greater myocardial infarction risk score (P=0.02) and a higher level of mean systolic (P=0.05) and diastolic blood pressure (P=0.0003), white blood cell count (P=0.002), and fibrinogen (P<0.0001) than the stroke-free group.
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
During the 5-year follow-up, 51 of 221 stroke survivors (23.1%) and 21 of 243 stroke-free subjects (8.6%) died. The age- and sex-adjusted mortality rates were 21.0% and 7.9% (P<0.0001), respectively (Table 2). The different total mortality rates are mainly attributed to cardiovascular deaths in the stroke and stroke-free groups (13.1% and 3.4%, respectively), whereas the rates of death due to malignancy were similar in the 2 groups (P=0.5). Other causes of death included pneumonia, sudden death of unknown origin, dementia, and traumatic head injuries. The distribution of main diagnostic groups (Table 2) was also different between those with and without stroke (P=0.009).
|
A history of stroke increased the risk of death significantly (hazard ratio [HR]=2.66; 95% CI, 1.60 to 4.43; adjusted for age and sex), although this became insignificant when adjusted for all variables in the model (HR=1.34; 95% CI, 0.73 to 2.46) (Table 3). Better physical and social functioning, by means of a 5-point increased FAI score, reduced the risk of death similarly in the stroke (HR=0.71; 95% CI, 0.62 to 0.82) and the stroke-free groups (HR=0.59; 95% CI, 0.47 to 0.75). Moderate use of alcohol also had a preventive effect in these groups (HR=0.62; 95% CI, 0.45 to 0.85; and HR=0.32; 95% CI, 0.15 to 0.69, respectively). Only 8 individuals (1.7%) reported a daily use of alcohol, whereas the mean frequency during 1 month was 6.1 (SD=0.7) for the stroke group and 4.6 (SD=0.4) for the stroke-free group, indicating a moderate consumption among users. The myocardial infarction risk score had no influence on death in the stroke group (HR=1.00; 95% CI, 0.76 to 1.30) in contrast to results for the stroke-free group (HR=2.58; 95% CI, 1.60 to 4.16). A high level of fibrinogen or white blood cell count was associated with increased risk of death in both the stroke and stroke-free groups, reaching statistical significance when the groups were analyzed together. In contrast, a high level of folic acid decreased the risk of death. These risk scores were mainly unchanged when we adjusted for all variables in the model.
|
In the stroke group, the estimated risk value of male versus female sex, with adjustment for age, increased when adjusted for all variables in the model (HR=2.15 and HR=3.29, respectively) because of a reinforcing effect of the myocardial infarction risk score and use of alcohol. Possible confounders tested were marital status, socioeconomic factors, coffee consumption, self-perceived health, use of drugs, blood pressure, atrial fibrillation, urinary incontinence, depressive symptoms, and homocysteine values. The estimated risk values did not change either by adding these variables to the model or by excluding the group with subarachnoid hemorrhage from the analysis.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Long-term stroke survivors had an increased risk of death compared with a stroke-free group (HR=2.66). Because of the low number of verified hemorrhagic strokes (34 of 221 [15.4%]), the findings in this study are valid primarily for those with thromboembolic strokes.
Physical and Social Activities
The FAI score and risk of death were inversely correlated. This index measures outdoor physical and social and indoor domestic functioning and seems to be a better predictor of death than a history of myocardial infarction and/or diabetes mellitus; these assessments were based on changes in likelihood functions in a Cox proportional hazards model. Compared with the MMSE score, which reflects global cognitive function, the FAI score yielded a better model by means of a greater decrease in log likelihood function (P=0.01 and P=0.000005, respectively) when added separately to the other variables. The FAI score may reflect the broader consequences of diseases as well as aging and cognitive and environmental factors and appears to be a more sensitive predictor of death than deteriorating general cognition,24 although they are intercorrelated (r=0.5). The various items in the FAI score have different influences on death, depending on stroke status and sex. A sum score of the indoor domestic items did not significantly influence death among men, irrespective of stroke status, whereas a 5-point increased score was associated with reduced death among women in both the stroke (HR=0.51; 95% CI, 0.32 to 0.84) and stroke-free groups (HR=0.40; 95% CI, 0.18 to 0.89). Social and physical activities, which were intercorrelated for both sexes (r
0.6), were independent predictors of death irrespective of stroke status. Physical activity, which is related to mortality in older women,25 could be the main contributor in the FAI score, or this may simply be related to mobility.26 However, the association between a reduced FAI score and death does not necessarily mean that improving one’s FAI score will extend survival.
Alcohol
Stroke and stroke-free subjects who drink alcohol 4 times a month had significantly lower risk of death than abstainers (P0.004, with adjustment for age and sex). Previous studies have examined the prestroke use of alcohol and risk of stroke on the basis of either the frequency of alcohol consumption27 or the amount of alcohol consumed.28 The latter study reported that mainly women with alcohol consumption of <12 g/d show a decreased risk of having an ischemic stroke. Moderate alcohol consumption, probably no more than 2 to 6 drinks per week,29 is also reported to reduce the risk of fatal and nonfatal vascular events among healthy individuals.30 In the present study the stroke survivors who used alcohol were younger (P=0.005) and had higher scores on both the FAI and MMSE tests (P=0.001 for both) than abstainers. However, use of alcohol remained a statistically significant predictor of survival when we adjusted for all available confounders (Table 3).
Myocardial Infarction Risk Score
In the stroke group, cardiovascular events were the dominant cause of death, whereas a calculated risk of suffering a myocardial infarction within the next few years had no impact (P=1.0). The latter finding was in contrast to the effect seen for the stroke-free group (P=0.0001). Of the 5 separate factors constituting the risk score of myocardial infarction, systolic blood pressure had no influence on subsequent death in the stroke group but yielded a modest increased risk in the stroke-free group, a finding that is in accordance with studies on healthy elderly subjects.31 An increased cholesterol level had a modest preventive effect, whereas cigarette smoking and male sex increased the risk of death. It is a paradox that an estimated risk score for future myocardial infarction had no impact on death in the stroke group as long as this was the prevailing cause of death. We can speculate that survivors of a previous stroke may have reached a level of total risk that overshadows the impact of an additional myocardial infarction risk score. Additionally, a fatal coronary event may be attributed to the influence of thrombotic risk factors at this advanced stage of atherosclerosis.
Inflammation
The increased risk of death associated with a high level of fibrinogen or white blood cell count is interpreted as an inflammatory response linked to atherosclerosis.32
Limitations
The number of events, especially in the stroke-free group, permits no analysis of the various causes of death, including cardiovascular deaths. Likewise, the results of this study are valid mainly for thromboembolic strokes because of the low number of cerebral hemorrhages.
Conclusions
The mortality rate among long-term stroke survivors is significantly higher than that of stroke-free individuals (HR=2.66; 95% CI, 1.60 to 4.43). Better physical and social functioning and moderate use of alcohol are associated with lower risk in both groups, whereas the myocardial infarction risk score has no significant impact on death in the stroke group, in contrast to results for the stroke-free group.
|
Acknowledgments |
|---|
This study was supported by the Research on the Elderly in Tromsø program and was financed by the Ministry of Health and Social affairs and the Norwegian Research Council.
Received May 22, 2003; revision received June 30, 2003; accepted August 8, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
1. Murray CJL, Lopez AD. Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet. 1997; 349: 1269–1276.[CrossRef][Medline] [Order article via Infotrieve]
2. Sarti C, Rastnyte D, Cepaitis Z, Tuomilehto J. International trends in mortality from stroke, 1968 to 1994. Stroke. 2000; 31: 1588–1601.
3. Stroke incidence and mortality correlated to stroke risk factors in the WHO MONICA project: an ecological study of 18 populations. Stroke. 1997; 28: 1367–1374.
4. Eriksson S-E, Olsson J-E. Survival and recurrent strokes in patients with different subtypes of stroke: a fourteen-year follow-up study. Cerebrovasc Dis. 2001; 12: 171–180.[Medline] [Order article via Infotrieve]
5. Peltonen M, Stegmayr B, Asplund K. Time trends in long-term stroke survival after stroke: the Northern Sweden Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) Study, 1985–1994. Stroke. 1998; 29: 1358–1365.
6. Sacco RL, Wolf PA, Kannel WB, McNamara PM. Survival and recurrence following stroke. Stroke. 1982; 13: 290–295.
7. Viitanen M, Eriksson S, Asplund K, Wester PO, Winblad B. Determinants of long-term mortality after stroke. Acta Med Scand. 1987; 221: 349–356.[Medline] [Order article via Infotrieve]
8. Viitanen M, Winblad B, Asplund K. Autopsy-verified causes of death after stroke. Acta Med Scand. 1987; 222: 401–408.[Medline] [Order article via Infotrieve]
9. Brønnum-Hansen H, Davidsen M, Thorvaldsen P. Long-term survival and causes of death after stroke. Stroke. 2001; 32: 2131–2136.
10. Dennis MS, Burn JPS, Sandercock PAG, Bamford JM, Wade DT, Warlow CP. Long-term survival after first-ever stroke: the Oxford Community Stroke Project. Stroke. 1993; 24: 796–800.
11. Engstad T, Bønaa KH, Viitanen M. Validity of self-reported stroke: the Tromsø Study. Stroke. 2000; 31: 1602–1607.
12. Sandvik H, Hunskaar S, Seim A, Hermstad R, Vanvik A, Bratt H. Validation of a severity index in female urinary incontinence and its implementation in an epidemiological survey. J Epidemiol Community Health. 1993; 47: 497–499.
13. Bønaa KH, Thelle DS. Association between blood pressure and serum lipids in a population. Circulation. 1991; 83: 1305–1314.
14. Holbrook M, Skilbeck CE. An activities index for use with stroke patients. Age Ageing. 1983; 12: 166–170.
15. Montgomery SA, Asberg MA. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979; 134: 382–389.
16. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental State": a practical method for grading cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12: 189–198.[CrossRef][Medline] [Order article via Infotrieve]
17. Nordic Council for Arctic Medical Research. The Cardiovascular Study in Finnmark 1974–75. OULU; 1979: 21–23, 25.
18. Westlund K, Nicolaysen R. Ten-year mortality and morbidity related to serum cholesterol. Scand J Clin Lab Invest Suppl. 1972; 30 (suppl 127): 3–24.
19. Bjartveit K, Wøien G. Cardiovascular Disease Risk in Norway: Results From Surveys in 18 Counties. Oslo, Norway: National Health Screening Service; 1997: 3–22.
20. Holme I, Helgeland A, Hjermann I, Leren P, Lund-Larsen PG. Four and two-thirds years incidence of coronary heart disease in middle-aged men: the Oslo Study. Am J Epidemiol. 1980; 112: 149–160.
21. Holme I, Håheim LL, Hjermann I. Predicting development of myocardial infarctions for 21 years based on the Oslo Study 1972–73 [in Norwegian]. Tidsskr Nor Laegeforen. 2003; 123: 1050–1053.[Medline] [Order article via Infotrieve]
22. Tverdal A. A Mortality Follow-up of Persons Invited to a Cardiovascular Disease Study in Five Areas in Norway. Oslo, Norway: National Health Screening Service; 1989.
23. SAS Institute Inc. SAS/STAT User’s Guide, Version 8. Cary, NC: SAS Institute Inc; 1999.
24. St John PD, Mongomery PR, Kristjansson B, McDowell I. Cognitive scores, even within the normal range, predict death and institutionalization. Age Ageing. 2002; 31: 373–378.
25. Gregg EW, Cauley JA, Stone K, Thompson TJ, Bauer DC, Cummings SR, et al. Relationship of changes in physical activity and mortality among older women. JAMA. 2003; 289: 2379–2386.
26. Khokhar SR, Stern Y, Bell K, Anderson K, Noe E, Mayeux R, et al. Persistent mobility deficit in the absence of deficits in activities of daily living: a risk factor for mortality. J Am Geriatr Soc. 2001; 49: 1539–1543.[CrossRef][Medline] [Order article via Infotrieve]
27. Berger K, Ajani UA, Case CS, Gaziano JM, Burning JE, Glynn RJ, et al. Light-to-moderate alcohol consumption and the risk of stroke among US male physicians. N Engl J Med. 1999; 341: 1557–1564.
28. Reynolds K, Lewis LB, Nolen JDL, Kinney GL, Sathya B, He J. Alcohol consumption and risk of stroke. JAMA. 2003; 289: 579–588.
29. Camargo CA, Hennekens CH, Gaziano JM, Glynn RJ. Prospective study of moderate alcohol consumption and mortality in US male physicians. Arch Intern Med. 1997; 157: 79–85.
30. Jamrozik K, Broadhurst RJ, Forbes S, Hankey GJ, Anderson C. Predictors of death and vascular events in the elderly: the Perth Community Stroke Study. Stroke. 2000; 31: 863–868.
31. Ellekjær H, Holmen J, Vatten L. Blood pressure, smoking and body mass index in relation to mortality from stroke and coronary heart disease in the elderly: a 10-year follow-up in Norway. Blood Pressure. 2001; 10: 156–163.[CrossRef][Medline] [Order article via Infotrieve]
32. Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med. 1999; 340: 115–126.
This article has been cited by other articles:
 |
|
 |
|
  N. U. Weir, A. Gunkel, M. McDowall, and M. S. Dennis Study of the Relationship Between Social Deprivation and Outcome After Stroke Stroke, April 1, 2005; 36(4): 815 - 819. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. Anderson, K. N. Carter, W. J. Brownlee, M. L. Hackett, J. B. Broad, and R. Bonita Very Long-Term Outcome After Stroke in Auckland, New Zealand Stroke, August 1, 2004; 35(8): 1920 - 1924. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||