| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2002;33:1624.)
© 2002 American Heart Association, Inc.
Original Contributions |
From Lund Stroke Center, Department of Clinical Neuroscience, Division of Neurology, University Hospital, Lund, Sweden.
Correspondence to Björn Hallström, MD, Lund Stroke Center, Department of Clinical Neuroscience, Division of Neurology, University Hospital, S-221 85 Lund, Sweden. E-mail bjorn.hallstrom{at}skane.se
| Abstract |
|---|
|
|
|---|
Methods Patients living within the Lund-Orup, Sweden, hospital district and fulfilling the diagnostic criteria for first-ever stroke during 19831985 (n=998) and 19931995 (n=1318) were followed up concerning survival status at 28 days, 1 year, and 3 years. Age and sex adjustments were performed. The possible influence of Oxfordshire Community Stroke Project (OCSP) stroke subtypes on survival was also analyzed.
Results Overall survival improved between the study periods (Cox proportional hazards regression: hazard ratio, 0.84; 95% confidence limits, 0.75 to 0.94; P=0.0019). The 28-day case fatality was 15% for stroke patients from both study periods. One-year case fatality was 31% for 19831985 patients and 27% for 19931995 patients. The corresponding figures at 3 years were 44% and 40%, respectively. In the group aged <75 years, there were no significant changes in overall survival, but survival improved significantly among patients aged
75 years beyond 28 days after stroke onset. OCSP stroke subtype was an independent predictor of death (P<0.0001).
Conclusions The recently observed increase in stroke incidence among patients aged <75 years was not accompanied by changing survival in that age group. However, among patients aged
75 years, survival improved beyond 28 days after stroke. The causes of this change in survival are unknown but may be related to improved long-term care of elderly stroke patients.
Key Words: elderly epidemiology stroke survival Sweden
| Introduction |
|---|
|
|
|---|
75 years. This study reports survival with a minimum follow-up of 3 years to determine whether the observed changes in incidence rates were accompanied by changes in survival. | Subjects and Methods |
|---|
|
|
|---|
75 years was 6.1% in 1983 and 6.6% in 1995. In Lund-Orup 13.8% of the population was
65 years in 1995, and in the European Union the corresponding figure was 15.9%.6 The fertility rate was 1.59 (Lund) and 1.85 (Orup) in 19831985 and 1.65 (Lund) and 2.02 (Orup) in 19931995. During 19831985, 31 053 persons moved away and 32 077 moved into Lund-Orup. During 19931995, the corresponding figures were 38 898 and 43 181, respectively. Life expectancy figures are similar to those for Sweden in general. There is only 1 emergency hospital within the area. Acute stroke patients in our hospital district are predominantly treated within the Department of Neurology at Lund University Hospital. There is a strong tradition of in-hospital care of acute stroke patients in Lund-Orup as well as in Sweden in general.7
Case Ascertainment
The methods for detecting patients with first-ever stroke have been described previously.5 Stroke was defined according to the World Health Organization stroke definition as rapidly developing signs of focal (or global) disturbance of cerebral function, lasting >24 hours or leading to death, with no apparent nonvascular cause.8 Patients aged
15 years living within the hospital district of Lund-Orup at the onset of first-ever stroke between January 1, 1983, and December 31, 1985, or between January 1, 1993, and December 31, 1995, were included in the study. The authors retrieved (B.H.) and individually scrutinized (A.L.) the medical records for patients admitted to the departments of neurology, neurosurgery, and internal medicine, with the diagnoses 430 to 438 (according to the International Classification of Diseases, Ninth Revision), to determine whether the patients had suffered from first-ever stroke. The diagnoses screened include transient ischemic attack (435) and late effects of cerebrovascular disease (438). At the Department of Neurology, where most patients were treated (during 1993 1995: 87.7%), patients were evaluated at least twice: first at admission and then the next day by a senior neurologist. In our institutional-based study we also screened medical records at the departments of pathology and forensic medicine for possible stroke diagnoses. If it was unclear whether the patient had prior stroke, medical records for earlier hospital treatments were scrutinized to exclude that. Only first-ever-in-a-lifetime strokes were included in the study. In total, 4463 patients with possible stroke were screened. There were 998 patients (519 men and 479 women) with first-ever stroke in 1983 1985 and 1318 (680 men and 638 women) in 19931995.5
Survival status of the patients was determined from the Swedish population register during mid-1999. The exact dates of death were unknown for 19 patients who died within 28 days. These patients were randomly assigned a survival time within that time span.
Stroke Classification
Patients fulfilling the diagnostic criteria for first-ever stroke were classified into main stroke types as having cerebral infarction (CI), intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), or undetermined pathological type.5 Patients with CI, ICH, or undetermined pathological type were classified according to the Oxfordshire Community Stroke Project (OCSP) classification system, as having lacunar syndrome (LACS), total anterior circulation syndrome (TACS), partial anterior circulation syndrome (PACS), or posterior circulation syndrome (POCS).5,9 For patients with medical records that were incomplete to determine the OCSP subtype, the subtype was classified as unknown.
Statistical Methods
Case fatality at 28 days, 1 year, and 3 years was calculated in relation to sex, age group, main type of stroke, OCSP stroke subtype, and study period. Day 0 was defined as the day of stroke onset. A Kaplan-Meier graph was calculated to show the cumulative survival up to 3 years after stroke onset for both study periods and also for the groups aged <75 years and
75 years. The cutoff point at 75 years was chosen because this represents approximately the median age for the stroke patients in our study.5 The log-rank test was used to examine whether there were survival differences between the 2 periods. To test whether the estimated difference in the risk of dying after stroke was confounded by different distributions of age group (15 to 34, 35 to 54, 55 to 64, 65 to 74, 75 to 84, and
85 years), sex, and OCSP subtype, we performed a Cox proportional hazards regression analysis, stratifying for those factors. Hazard ratios with 95% confidence limits (CL) were shown. The
2 test was used for comparison of survival outcome for categorical variables. The Statistical Product and Service Solutions (SPSS) software package was used for statistical analysis. A value of P<0.05 was considered statistically significant.
| Results |
|---|
|
|
|---|
|
|
Figure 1 shows a Kaplan-Meier graph estimating the cumulative survival during the first 3 years after stroke onset for patients from both study periods. The survival was better during the latter period (P=0.0013, log-rank test). The improvement of case fatality noted between the 2 time periods was also statistically significant after adjustment for age, sex, and OCSP subtype of stroke, as shown in Table 2.
|
We also made Kaplan-Meier graphs for the 2 groups aged <75 years and
75 years (Figure 2), showing that the improved survival was confined to the elderly patients (P=0.0018, log-rank test in the group aged
75 years; P=0.29, log rank test in the group aged <75 years). The change among patients aged
75 years remained statistically significant after adjustment for age, sex, and OCSP subtype (hazard ratio, 0.80; 95% CL, 0.70 to 0.92; P=0.001, Cox proportional hazards regression), whereas for patients aged <75 years the corresponding hazard ratio was 0.95 (95% CL, 0.77 to 1.17; P=0.65). The change in survival among patients aged
75 years was confined to the period beyond 28 days after stroke onset (data not shown).
|
Sixty-two patients were lost to follow-up during the first study period and excluded from the case fatality analyses beyond 28 days, although they were included in Cox proportional hazards regression analyses. These patients with first-ever stroke during 19831985, whose personal code numbers were not registered, constituted 6.2% of the patients during 19831985. Among these patients, 31% had died within 28 days (median age, 82 years; range, 60 to 94 years; 3 patients <75 years and 16 patients
75 years). The 43 patients surviving at 28 days had a median age of 74 years, ranging from 56 to 88 years.
Main Types of Stroke in 19931995
The proportion of patients undergoing CT, MRI, or autopsy was 57% during 19831985 and 90% during 19931995. The 28-day case fatality in 19931995 was 11% for CI, 29% for ICH, 32% for SAH, and 22% for undetermined pathological type. At 1 year the corresponding figures were 22%, 41%, 37%, and 41% and at 3 years 35%, 53%, 41%, and 61%, respectively. In a univariate analysis, the case fatalities differed significantly between the different main types of stroke (P< 0.0001,
2 test).
OCSP Subtypes and Case Fatality
There were no significant changes among the OCSP stroke subtypes in case fatality between the 2 study periods except declining 3-year case fatality for POCS. In a univariate analysis, the case fatalities differed significantly between the different OCSP stroke subtypes (P<0.0001,
2 test). As mentioned before, the case fatality was significantly higher for patients with TACS.
| Discussion |
|---|
|
|
|---|
75 years, whereas among patients aged <75 years the case fatality was stable. However, the change in case fatality among patients aged
75 years was not explained by increased stroke survival during the first month. It was mainly the survival beyond the first month that improved. One possible explanation for the increased survival could be the introduction of a stroke unit at our hospital. Stroke units have been reported to increase short-term as well as long-term survival after stroke. 10 It has also been shown that stroke units lower case fatality in routine clinical practice. 11 At Lund University Hospital, stroke unit care was introduced in 1991, ie, between the 2 time periods we have studied. However, it is noteworthy that no change in survival was noted during the first month and that the increased survival was not seen in the younger age group. This might indicate that the stroke unit care did not influence stroke survival, but another possible explanation is that stroke care at Lund University Hospital before 1991 had much in common with the principles for stroke care in a stroke unit. The strategies for secondary stroke prevention have also changed over the studied period. Trials have shown the benefit of antiplatelet therapy after ischemic stroke,12 anticoagulation for patients with atrial fibrillation,13 antihypertensive treatment in elderly patients,14 and carotid endarterectomy in symptomatic patients with high-grade carotid stenosis.15,16 It would be interesting to know the cause of death among the patients in our study as well as the functional outcome among the survivors, but these data are not available. Short-term case fatality might be seen as an indirect indicator of stroke severity.17 In our study the finding of unchanging 28-day case-fatality might indicate unchanged severity of stroke between the 2 study periods.
Comparisons With Other Studies
The 28-day and 1-year case fatalities in Lund-Orup are low compared with other studies (based on unadjusted figures). In other studies during the 1990s, the total 28-day (or 30-day) case fatality of first-ever stroke (all ages) has been reported to be between 13% and 34%,1833 and the corresponding figure for the total 1-year case fatality is between 27% and 41%.19,21,26,29,3133 Three-year case fatality has been reported to be 43% to 50%.34,35 Much of the differences in case fatality can probably be explained by differences between studies regarding age distribution of the studied stroke patients. The proportion of patients
75 years of age in other studies ranged from 20% to 64%.18,23 Such large differences in age structure may strongly influence the case fatality. When compared with other studies (see overview of Marini et al36), our study population is older than some other study populations (Dijon, Auckland, Rochester, Oxfordshire) and similar to others (Perth, Aosta). There are also other studies with older study populations (Belluno, Umbria, LAquila, Innherred, and Frederiksberg). When incidence studies are compared, it is therefore important to adjust to different age and sex distributions. As noted in the aforementioned article, age and sex adjustments do not completely eliminate these differences. Differences in age structure are, however, probably not the only reason why the case fatalities in the various studies differ. Other possible explanations are incomplete case ascertainment and differences in stroke severity. There are also indications of differences between countries concerning stroke management and use of secondary preventive therapies that might influence the case fatalities in different studies.32
Time trends of case fatality have shown different results. In Finland,37 Göteborg (Sweden), 38 Auckland (New Zealand), 20 Tartu (Estonia), 27 and Shanghai (China) 39 and in the Northern Sweden MONICA Project, 40 decreasing 28-day case fatalities have been reported. However, stable or increasing 28-day case fatalities have been reported in Rochester (Minnesota) 41 during 19801984 and 19851989, Söderhamn (Sweden),42 Kaunas (Lithuania),43 and Novosibirsk (Russia).24 Long-term survival improved in the Northern Sweden MONICA Project, 44 and similar results have been reported from Finland29 and Denmark.33 Long-term survival also improved in Sö derhamn (Sweden)45 and in the Minnesota Heart Survey.46 However, no significant short- or long-term trend was seen in Rochester (Minnesota) during 19751989.47
The 28-day (or 30-day) and 1-year case fatality of the different main types of stroke varied between studies during the 1990s. Twenty-eight day (or 30-day) case fatality has been reported to be 11% to 26% for CI, 22% to 57% for ICH, 8% to 54% for SAH, and 31% to 82% for undetermined pathological type.21,23,25,26,28,29,35,48 One-year case fatality has been reported to be 30% to 32% for CI, 36% to 58% for ICH, 38% to 58% for SAH, and 46% to 88% for undetermined pathological type.21,26,29 The corresponding case fatalities in Lund-Orup were within the lower range compared with most other studies. As mentioned above, variations in case fatality may depend partly on differences in age structure of the studied stroke populations, but differences in case ascertainment and stroke severity probably also contribute to those differences. The case fatalities (28 or 30 days and 1 year) for the different OCSP subtypes in Lund-Orup were comparable to other studies.9,49
Methodological Aspects
There may be some potential sources of error in this retrospective and institutional-based study. Theoretically, a higher hospitalization rate of milder strokes during the latter period might have resulted in decreasing case fatality trends. However, it has been estimated that <5% of all first-ever stroke survivors in Sweden are treated out of hospital (for review, see Norris7), and between the 2 study periods there were no major changes in hospital admission policy in Lund-Orup.5 We did not screen routinely for nonhospitalized stroke patients in our study. Because full coverage of every single possible nonhospitalized stroke patient was not performed, it is thus possible that some of these patients may have remained undetected (eg, patients with sudden fatal stroke that were not subject to autopsy). However, according to Swedish regulations, autopsy should be performed in cases of sudden death without known cause occurring outside hospital. Therefore, we believe that we detected most cases of sudden fatal stroke, using our method of screening possible stroke diagnoses at the departments of pathology and forensic medicine. We believe that the number of undetected patients was low and did not influence our results more than marginally.
Survival status was obtained 6 months after the end of follow-up, ie, in mid-1999. We believe that the Swedish population register is sufficiently accurate after this time because death certificates must be sent in without delay to the authorities according to official Swedish regulations.
In part, increased long-term survival may reflect an improved survival in the general population that has occurred during the last decades in Sweden as well as in most other countries. The relative contribution of this factor for patients with stroke is unknown and needs to be addressed in future studies.
The 62 patients during 19831985 lost to follow-up had higher short-term case fatality, and they may have had higher long-term case fatality as well. However, this would not influence the main results of the present study.
In conclusion, the recently reported increase in stroke incidence among patients aged <75 years was not accompanied by changes in case fatality in that age group. However, among patients aged
75 years, the survival after first-ever stroke increased. The reasons for this change are unknown but may be related to an improved long-term care of elderly stroke patients.
Received December 18, 2001; revision received March 13, 2002; accepted March 15, 2002.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
B. Hallstrom, A.-C. Jonsson, C. Nerbrand, B. Norrving, and A. Lindgren Stroke Incidence and Survival in the Beginning of the 21st Century in Southern Sweden: Comparisons With the Late 20th Century and Projections Into the Future Stroke, January 1, 2008; 39(1): 10 - 15. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Carandang, S. Seshadri, A. Beiser, M. Kelly-Hayes, C. S. Kase, W. B. Kannel, and P. A. Wolf Trends in Incidence, Lifetime Risk, Severity, and 30-Day Mortality of Stroke Over the Past 50 Years JAMA, December 27, 2006; 296(24): 2939 - 2946. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. J. Bainbridge, M. W. J. Davie, and M. J. Haddaway Bone loss after stroke over 52 weeks at os calcis: influence of sex, mobility and relation to bone density at other sites Age Ageing, March 1, 2006; 35(2): 127 - 132. [Abstract] [Full Text] [PDF] |
||||
![]() |
R Fogelholm, K Murros, A Rissanen, and S Avikainen Long term survival after primary intracerebral haemorrhage: a retrospective population based study J. Neurol. Neurosurg. Psychiatry, November 1, 2005; 76(11): 1534 - 1538. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Eriksson, K. Asplund, E.-L. Glader, B. Norrving, B. Stegmayr, A. Terent, K. H. Asberg, P.-O. Wester, and for The Riks-Stroke Collaboration Self-Reported Depression and Use of Antidepressants After Stroke: A National Survey Stroke, April 1, 2004; 35(4): 936 - 941. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Terent Trends in Stroke Incidence and 10-Year Survival in Soderhamn, Sweden, 1975-2001 Stroke, June 1, 2003; 34(6): 1353 - 1358. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Stroke Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2002 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |