Declining Incidence and Mortality Rates of Stroke in Finland From 1972 to 1991
Results of Three Population-Based Stroke Registers
Background and Purpose We aimed to determine trends in stroke incidence, mortality rates, case-fatality rates, and their relation in Finland.
Methods We compared the results of three population-based stroke registers that included first-ever strokes in people aged ≥15 years. Two registers were kept in Espoo-Kauniainen, the first in 1972 to 1973 (EK 72-73) and the second in 1978 to 1980 (EK 78-80). The present register of the Finnish Heart Association (FHA 89-91) was kept in four districts in Finland in 1989 to 1991.
Results The age-adjusted incidence rates were 240.9, 174.4, and 191.6, and the 1-year mortality rates were 121.9, 77.0, and 65.3 in the EK 72-73, EK 78-80, and FHA 89-91 registers, respectively. The overall decline from 1972 to 1991 was 20% in the stroke incidence rate and 46% in the stroke mortality rate. One-month case-fatality rates decreased from 34.8% to 29.4% in the EK 72-73 and EK 78-80 registers and to 23.3% in the present register.
Conclusions The decline in the stroke incidence rate during the 1970s stabilized during the late 1980s and early 1990s; however, the case-fatality rate is still decreasing. Their combined effects may explain the continuing decline in stroke mortality.
The incidence of first-ever stroke decreased from the early 1950s to the late 1970s1 but has since remained stable or tended to increase.2 3 4 5 6 In Finland a declining incidence was recorded in the 1970s.7 Stroke mortality rates have continued to decline in most Western countries,8 which should be a consequence of a decline in either incidence or case-fatality rates or both. This study examines trends in incidence, case-fatality rates, and mortality rates from 1972 to 1991 in Finland based on the results of three population-based stroke registers.7 9 10
Subjects and Methods
Study Area and Population
The first of the three population-based stroke registers, EK 72-73, started in January 1, 1972, and continued until December 31, 1973. The second register, EK 78-80, operated in the same area from April 1, 1978, to March 31, 1980. The aim of the third and present register, FHA 89-91, was to determine stroke outcome in four separate districts in Finland (Figure⇓), and it was kept from August 15, 1989, through August 14, 1991. The follow-up time was 1 year in all registers.
The study population consisted of 113 100, 136 850, and 134 804 inhabitants in the EK 72-73, EK 78-80, and FHA 89-91 registers, respectively (Table 1⇓). In the FHA 89-91 register the four districts comprised 21 rural municipalities. The population varied from 31 975 to 35 142 inhabitants between the study districts. In the two earlier registers 5.1% and 6.3% of the study population were aged ≥65 years, whereas in the present register 15.8% belong to this age group. The proportion of the Finnish population aged ≥65 years was 9.2%, 12.0%, and 13.4% in the 1970, 1980, and 1990 censuses, respectively.
Case Finding and Case Ascertainment
The EK 72-73 and EK 78-80 registers were kept according to the guidelines of the WHO Stroke Register.11 First-ever strokes were registered in patients aged ≥15 years. In the EK 72-73 and EK 78-80 registers, the patients were initially examined by one of the authors (K.A.) or study assistants.7 The case findings were based on personal contact with all hospitals and wards treating stroke patients. The classification of strokes into subgroups was based on WHO guidelines.12 To complete case ascertainment, all death certificates were scrutinized regularly.
The present register was also kept according to WHO guidelines.11 All patients with a first-ever stroke and aged ≥15 years were registered by the study assistant of the district. The case findings were based on weekly contact with the health centers and central hospitals. The primary stroke diagnosis was made in health centers by a general practitioner or in hospitals by an internist or neurologist. In addition to the health centers, one central hospital treated stroke patients and one university hospital treated neurosurgical patients in each district. The final classification of strokes and their subtypes was made by one of the authors (H.N.) according to the Classification of Cerebrovascular Diseases III.13 All available data were used to confirm complete case findings.
Incidence adjustments were made by the direct method, and confidence intervals were calculated from published tables based on the Poisson distribution.14 In the tables, adjustments were made according to the mean Finnish population in 1990.15 For comparison with other studies, the Swedish population in 197016 was also used. Case-fatality rates were compared by χ2 analysis with continuity correction. For patients in the FHA 89-91 register, we performed an analysis of variables to determine odds ratios for death 30 days and 1 year after stroke. The variables significant in univariate regression analysis were chosen for multiple logistic regression analysis (BMDPLR, version 1990), in which a maximum-likelihood stepwise forward elimination procedure was used. In these analyses we considered cardiovascular disease to be present if the patient had diagnosed coronary artery disease, heart failure, or medically treated hypertension. The Rankin score before the onset of stroke with a cutoff point >2 was used to measure independence in activities of daily living. For statistical analysis, subtype of stroke was dichotomized into brain infarcts and other strokes.
A total of 244, 255, and 594 patients with first-ever stroke were included in the EK 72-73, EK 78-80, and FHA 89-91 registers, respectively. The respective proportions of patients aged ≥65 years were 54%, 57%, and 73% in the EK 72-73, EK 78-80, and FHA 89-91 registers. The respective mean ages of males were 61.9, 61.7, and 67.7 years in the EK 72-73, EK 78-80, and FHA 89-91 registers, and the respective figures for females were 65.2, 67.0, and 74.6 years. The ratio of females to males increased from 1.09 in the EK 72-73 register to 1.18 in the FHA 89-91 register.
Eighty-nine percent of patients in the EK 72-73 register, 95% in the EK 78-80 register, and 86% in the FHA 89-91 register were admitted to the hospital. In the FHA 89-91 register, primary contact was made in 77% of cases with a general practitioner in health centers; the primary diagnosis of stroke was made in 23% of cases by a general practitioner, in 19% by an internist, and in 58% by a neurologist.
Brain CT was performed for 11% of patients in the EK 78-80 register and for 60% in the FHA 89-91 register. Of those who died, an autopsy was performed for 48%, 51%, and 11% of cases in the EK 72-73, EK 78-80, and FHA 89-91 studies, respectively. The frequencies of diagnostic categories according to sex are presented in Table 2⇓. Of the 368 patients in the FHA 89-91 register for whom the diagnosis was verified by either CT or autopsy, 73.4% had brain infarct, 18.2% ICH, 7.9% SAH, and 0.5% nonspecific stroke.
Incidence and Mortality Rates
The crude annual incidence rates for all strokes in different age categories are presented in Table 3⇓. The rate is generally greater in males aged <65 years than in females and later equalizes. After the rates were adjusted to the Finnish population of 1990, the annual incidence per 100 000 for males aged <65 years decreased from 84.2 in the EK 72-73 register to 63.3 in the EK 78-80 register but increased to 77.7 in the FHA 89-91 register. For females, the respective figures are 58.7, 50.0, and 44.7. In patients aged ≥65 years, the trend in incidence resembles that in the younger patients, decreasing in males from 1466.0 in the EK 72-73 register to 916.6 in the EK 78-80 register but increasing to 1033.0 in the FHA 89-91 register. Females have an identical trend from 1266.1 to 946.5 and to 1035.6, respectively.
Incidence rates according to stroke subtype are shown in Table 4⇓. The incidence rates of SAH, nonspecific stroke, and all strokes for both sexes during 1972 to 1991 decreased significantly. The proportional decrease in incidence of all strokes was 20% during this period. After adjustment of the incidence rates to the 1970 Swedish population, the total incidence rates were 246.4, 181.0, and 194.7 in the EK 72-73, EK 78-80, and FHA 89-91 registers, respectively.
Total mortality rate decreased 46% (Table 5⇓). The decrease was 42% in the group aged <65 years and 47% in the group aged ≥65 years between the EK 72-73 and FHA 89-91 registers.
Case-fatality rates at different times after stroke onset are shown in Table 6⇓. The greatest decrease in the case-fatality rate of ICH occurred in females (from 73.9% to 17.4% at 30 days). To estimate the causes of declining case-fatality rates in brain infarct, the proportions of patients who died during the first week (days 1 to 7) and after (days 8 to 30) were calculated. These were 36% and 64% (n=33), 52% and 48% (n=46), and 53% and 47% (n=91) in the EK 72-73, EK 78-80, and FHA 89-91 registers, respectively.
Odds Ratios for Death
On the basis of univariate analysis, six variables were chosen for multiple logistic regression analysis to determine the odds ratios for death in the FHA 89-91 register at 30 days and at 1 year after stroke. These variables were age, sex, subtype of stroke, Rankin score before stroke onset, diagnosed cardiovascular disease, and atrial fibrillation recorded at stroke onset. In addition to these factors, diabetes, musculoskeletal diseases, or the side involved in hemiparesis were shown to be insignificant risk factors in univariate analysis. In the stepwise model, age, diagnosis (infarct versus ICH, SAH, or nonspecific stroke), and a Rankin score >2 before stroke onset proved to be independent risk factors for death at 30 days (Table 7⇓). At 1 year, atrial fibrillation recorded at the onset of stroke was also a significant risk factor.
This study confirms the results of a previous study,7 namely, that the incidence rate of stroke has declined in Finland since the early 1970s. In the 1980s the incidence increased negligibly, but the mortality rate showed a continuing decrease. During the period 1972 to 1991, the incidence rate of first-ever strokes declined 20% and the mortality rate declined 46%.
We noted a marked difference in the crude incidence rates between the three registers. The most likely reason is the increasing proportion of people aged ≥65 years. This is due partly to the general rising proportion of older age groups in the Western world and partly to the study area, which included mainly rural areas in the FHA 89-91 register in contrast to southern urban areas in the other two registers.
The observed decrease in the age- and sex-adjusted incidence rates occurred mainly from the EK 72-73 to the EK 78-80 register, ie, in the same geographic area with the same referral pattern and case finding methods; for this reason the decrease is most likely real. No major change in incidence rates occurred between the last two registers, which took place in different geographic areas with equal incidence rates.
Case finding in the EK 72-73 and 78-80 registers was based chiefly on contact with hospitals that treated stroke patients. We believe that case finding in all of these registers has been complete because in Finland stroke patients are traditionally treated at hospitals. In the two EK registers most patients came directly to the hospital emergency wards, whereas in the FHA 89-91 register most patients were first seen by the municipal general practitioner in a local health center because the central hospitals treating patients with acute stroke were more distant. The quality of hospital care is very similar throughout Finland.
In a recent study,17 the trends in stroke incidence during the 1970s and 1980s were studied in a Finnish population aged 25 to 74 years. In this FINMONICA register (the Finnish contribution to the WHO MONICA register), the decrease in incidence rates between 1972 and 1989 was 22% for men and 43% for women, showing a continuing decline in the 1980s. In the present study the decline reversed during the 1980s, and a slight increment in the incidence rates was seen in both the group aged <65 years and the group aged ≥65 years.
Comparison of our incidence rates with those reported from Rochester shows an almost identical trend. In Rochester the decline from 1960-1964 to 1980-1984 was 22% after the rates were adjusted to the Swedish standard population,18 including an increase of 8% during the last decade. In our Finnish study, after adjustment to the same standard, the decline from the first register for 1972 to 1973 to the present one for 1989 to 1991 was 20%, with an increase of 8% from 1978-1980 to 1989-1991. The slight increase in incidence in Rochester was mainly seen in the population aged ≥75 years.3
In two Scandinavian studies, no changes in incidence rates were seen during 1976 to 19884 or 1971 to 19875 in people aged 15 to 65 years. On the other hand, an increment in the incidence of stroke in women aged 45 to 64 years was reported in another Swedish study during the periods 1975 to 1978 and 1983 to 1986.2 In the present study, women aged 55 to 64 years were also the only group in which incidence increased continuously.
The changes in incidence rates can only be incompletely explained by changes in risk factors. A decrease in the prevalence of smoking and hypertension accounted for 29% of the decline in the incidence of stroke in Finland from 1972 to 1977 in a cohort of the population aged 30 to 59 years.19 None of the data on primary prevention with antiplatelet drugs support the role of antiplatelet treatment as a cause of changes in the incidence rates, even though the decline in the 1970s was of the same magnitude as would be achieved with these drugs in symptomatic patients during 1 year.20
Most of the patients in all three registers were treated in the hospital in the acute phase of their strokes (89%, 95%, and 86%), and in all time periods all patients with severe stroke were hospitalized. Also, there were few differences in referral patterns between the registers that could explain the observed decrease in case-fatality rates, for which many explanations have been suggested.8 However, one cannot overlook the possibility that the differences simply reflect geographic differences or differences in technology. One reason for the declining case-fatality rate of ICH could be the detection of milder cases with the help of CT, although incidence rates have also slightly decreased. The fact that CT was used more often in the last register cannot explain the decline of incidence. The easier availability of CT would only affect the distribution of ischemic and hemorrhagic strokes. We noted that the proportion of brain infarcts in relation to total stroke incidence increased through the entire study, but the case-fatality rate did not decrease during the same time.
The management of stroke has become more active during the study period, as shown by the proportion of patients who had a CT examination and were treated by a neurologist and in the reduced proportion of patients with nonspecific stroke. The changing composition of stroke subgroups and the reduced proportion of ICH in particular may partly explain the decrease in case-fatality and mortality rates. The overall case-fatality rate at 30 days decreased by 33% and in patients with brain infarct by 13%. In patients with brain infarct there was a shift in 30-day case-fatality rates to earlier days within a 30-day period between the EK 72-73 register and the two later registers: in the latter two a greater proportion of patients died during the first week than afterward. This may reflect more effective prevention of pulmonary embolism and better care of patients with pneumonia after the first week, whereas those dying in the first week still do so mainly as a result of brain edema and herniation,21 for which there is no effective therapy. Improved hospital care and prevention of complications may have decreased the case-fatality rate during the study period. Thus, these findings emphasize the role of case-fatality rates in the continuing decline of mortality rates.
The results of the multiple regression analysis emphasize the significance of the premorbid functional status of patients, which in some cases was poor because of coronary heart disease, heart failure, pulmonary diseases, or joint problems in the lower extremities. In addition, stroke subtype was shown to be a prognostic factor. In most cases it is possible to reliably estimate whether the patient was independent in activities of daily living before stroke. With the exception of atrial fibrillation, the presence of other cardiovascular diseases had no independent role as a predictor in the analysis.
The trends in stroke incidence, mortality rates, and case-fatality rates in our study are in agreement with other studies from Scandinavia and with the Rochester data and suggest that the trends are real. Our results support the theory that both incidence and case-fatality rates have decreased and may explain the continuing decline in stroke mortality.
Selected Abbreviations and Acronyms
|EK 72-73||=||Espoo-Kauniainen 1972 to 1973|
|EK 78-80||=||Espoo-Kauniainen 1978 to 1980|
|FHA 89-91||=||Finnish Heart Association 1989 to 1991|
|MONICA||=||Monitoring Trends and Determinants in Cardiovascular Disease|
|WHO||=||World Health Organization|
This study was supported by the Finnish Heart Association. We would like to thank Associate Professor Seppo Sarna for statistical assistance and Jaana Immonen, Seija Keskiväli, Eeva Kröger, Pirjo Lyytikäinen, Marja-Liisa Niemi, Oona Tuomi-Nikula, and Markku Karjalainen for registering patients.
- Received July 11, 1995.
- Revision received May 20, 1996.
- Accepted May 20, 1996.
- Copyright © 1996 by American Heart Association
Terént A. Increasing incidence of stroke among Swedish women. Stroke. 1988;19:598-603.
Broderick JP, Stephen JP, Whisnant JP, O'Fallon WM, Bergstralh EJ. Incidence rates of stroke in the eighties: the end of the decline in stroke? Stroke. 1989;20:577-582.
Lindenstrøm E, Boysen G, Nyboe J, Appleyard M. Stroke incidence in Copenhagen. Stroke. 1992;23:28-32.
Harmsen P, Tsipogianni A, Wilhelmsen L. Stroke incidence rates were unchanged, while fatality rates declined, during 1971-1987 in Göteborg, Sweden. Stroke. 1992;23:1410-1415.
Jørgensen HS, Plesner A-M, Hübbe P, Larsen K. Marked increase of stroke incidence in men between 1972 and 1990 in Frederiksberg, Denmark. Stroke. 1992;23:1701-1704.
Kotila M. Declining incidence and mortality of stroke? Stroke. 1984;15:255-259.
Bonita R, Beaglehole R. Stroke mortality. In: Whisnant JP, ed. Stroke: Population, Cohorts, and Clinical Trials. Oxford, England: Butterworth-Heinemann; 1993:59-79.
Aho K. Incidence, Profile and Early Prognosis of Stroke. Helsinki, Finland: University of Helsinki; 1975. Dissertation.
Numminen H, Kaste M, Kotila M, Waltimo O. Stroke outcome in a population based rehabilitation programme. Can J Neurol Sci. 1993;20(suppl 4):S88. Abstract.
WHO. Cerebrovascular Diseases: Prevention, Treatment and Rehabilitation. Geneva, Switzerland: World Health Organization; 1971. WHO Technical Report Series No. 469.
Special Report from the National Institute of Neurological Diseases and Stroke. Classification of cerebrovascular diseases III. Stroke. 1990;21:637-676.
Schoenberg BS. Calculating confidence intervals for rates and ratios. Neuroepidemiology. 1983;2:257-265.
Structure of Population 1990. Helsinki, Finland: Central Statistical Office of Finland; 1991.
Statistical Abstracts of Sweden. Stockholm, Sweden: National Central Bureau of Statistics; 1971.
Terént A. Stroke morbidity. In: Whisnant JP, ed. Stroke: Population, Cohorts, and Clinical Trials. Oxford, England: Butterworth-Heinemann; 1993:37-58.
Tuomilehto J, Bonita R, Stewart A, Nissinen A, Salonen JT. Hypertension, cigarette smoking, and the decline in stroke incidence in eastern Finland. Stroke. 1991;22:7-11.
Antiplatelet Trialists' Collaboration. Collaborative overview of randomised trials of antiplatelet therapy, I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. Br Med J. 1994;308:81-106.