Risk Factors of Stroke Incidence and Mortalit

Transcription

1 1484 Risk Factors of and Mortalit A 12-Year Follow-up of the Oslo Study Lise Lund Haheim, BDS; Ingar Holme, PhD; Ingvar Hjermann, MD; Paul Leren, MD Background and Purpose: The objective of this study was to determine the risk factors of stroke incidence and mortality. Methods: Our data are based on a prospective cohort study of men aged 40 to 49 years after 12 years of follow-up. Results: In age-adjusted Cox proportional-hazards regression analysis of healthy men, diastolic blood pressure was a stronger predictor for stroke incidence and mortality than systolic blood pressure. Smoking was a stronger predictor of mortality than of incidence. However, there was no dose-response relation among smokers by increased cigarette consumption. Total serum cholesterol was a significant (P<.05) risk factor for stroke mortality and of borderline significance (P=.08) for stroke incidence. Increased physical activity at leisure was associated with reduced stroke incidence but not mortality. The myocardial infarction risk score comprising systolic blood pressure, total serum cholesterol, and daily cigarette smoking was a strong predictor of mortality and incidence. Body mass index, triglycerides, blood glucose, and physical activity at work were not found to be risk factors for stroke. Conclusions: Reduction of blood pressure, cessation of smoking, lowered cholesterol, and increased physical activity at leisure are individual measures to reduce the risk of stroke. (Stroke. 1993;24: ) KEY WoRDs * incidence * mortality * Norway * risk factors Stroke is the third most common cause of death in men in Norway after myocardial infarction and cancer (1990 data). During the 10-year period from 1980 to 1990, stroke mortality in men declined by 4%.' The incidence and mortality of stroke vary greatly among populations.2-11 During the last decades stroke mortality has steadily declined in many countries, including the United States, Canada, Japan, and European countries other than Eastern European countries.12 However, there have been reports of a change in trend toward an increase in stroke incidence.13,14 Although new poststroke treatment strategies reduce mortality and hospitalizations,15 increased knowledge of risk factors is still important in the prevention of stroke. Major risk factors in stroke for men are hypertension, clinical evidence of cardiac disease, transient ischemic attacks, high hematocrit, diabetes mellitus, elevated blood cholesterol and lipids, cigarette smoking, excessive alcohol intake, physical inactivity, and obesity.16 In this article we present the results after 12 years of follow-up with regard to the incidence and mortality of stroke and the long-term predictivity of certain risk factors for men with no known symptoms or diseases of cardiovascular origin or diabetes. Received March 1, 1993; final revision received May 3, 1993; accepted May 24, From the Life Insurance Companies Institute of Medical Statistics (L.L.H., I. Holme) and the Department of Medicine (I. Hjermann, P.L.), Ulleval Hospital, Oslo, Norway. Reprint requests to Lise Lund H?aheim, Life Insurance Companies Institute of Medical Statistics, Pb 6, Ulleval Hospital, N-0407 Oslo, Norway. Subjects and Methods During 1972 and 1973 all men in Oslo aged 40 to 49 years and a 7% sample of men aged 20 to 39 years were invited to a screening. This study reports on the men aged 40 to 49 years. A total of men attended the screening from May 1972 to December The screening'7 included answering a questionnaire on medical history with regard to cardiovascular symptoms or diseases, smoking habits, physical activity, and stress factors. The men had their blood pressure measured with a mercury sphygmomanometer; the first and the fifth Korotkoffs sounds determined the systolic blood pressure (SBP) and diastolic blood pressure (DBP), respectively. A blood sample was taken in the nonfasting state to determine total serum cholesterol, triglycerides, and glucose. Height and weight were measured. The risk factor analysis is based on the baseline measurements. Also included in this report are the men who later participated in the dietary and smoking intervention trial (n= 1232)18 and the hypertension trial (n=785).19 Our institute collected the data on stroke morbidity from all Oslo hospitals and two hospitals in the neighboring county. It was not feasible to collect morbidity data on men leaving the Oslo area to attend other hospitals. Data on all deaths were supplied from the Central Bureau of Statistics. Mortality data are complete except for those who emigrated. The eighth revision of The International Classification of Diseases was used (stroke codes 430 through 438). The diagnoses were not reviewed by the investigators. We included those men with stroke as major event. The follow-up period ended on December 31, For comparative reasons the men were grouped in accordance with other Norwegian studies20,21 as follows:

2 HLaheim et al Risk Factors of and Mortality 1485 TABLE 1. Rate and Rate Ratio During 12-Year Follow-up of (Nonfatal and Fatal) and Mortality in the Oslo Study, Norway, 1972 Stroke Mortality At Risk, Group N n Rate (Cl) RR (Cl) n Rate (Cl) RR (Cl) (0.41, 0.55) (0.14, 0.18) (0.34,1.86) (0.99, 4.49) (0.11, 0.61) (0.45, 7.62) (1.37, 5.33) (4.61,10.07) (0.29,1.11) (1.74,10.14) (1.12,1.66) (2.21, 3.82) (0.45, 0.67) (2.25, 5.97) Total (0.84,1.06) (1.55, 2.56) (0.29, 0.37) (1.37, 3.43) Results are presented for all men invited to screening, aged years, subdivided into groups 1-4: group 1, healthy; group 2, symptoms of cardiovascular disease (including men [n=36] reporting previous stroke at screening) or diabetes; group 3, cardiovascular disease or diabetes; group 4, nonattenders. Stroke incidence includes nonfatal and fatal cases of first occurring event. N indicates number of men; n, number of cases; RR, rate ratio relative to group 1; and Cl, 95% confidence interval. group 1, men with no symptoms or diseases of cardiovascular origin or diabetes; group 2, men with symptoms of angina or claudication of the lower limbs; group 3, men with cardiovascular diseases or diabetes; group 4, nonattenders. Incidence and mortality are presented as rate ratio of number of cases per 1000 person-years. The relation of risk of disease to risk factors is presented by graphs of the rate at quintile recordings of the risk factors. Age-adjusted Cox proportional-hazards regression analysis is presented for the total serum cholesterol, triglycerides, SBP, DBP, height, weight, body mass index, and blood glucose. Triglycerides and glucose were adjusted for time since last meal. We adjusted for time since last meal by adding this variable to the risk factor model to be analyzed. Physical activity at work or leisure was recorded as sedentary, moderate, intermediate, or great. Cigarette smoking was presented as daily versus nondaily cigarette smoking, the latter group later termed as nonsmokers. The myocardial infarction risk score is a multiplicative score of total serum cholesterol, SBP, and daily cigarette consumption.2' Results The stroke incidence rate (nonfatal or fatal first stroke) in 277 subjects was 0.95 per 1000 person-years for all invited to the screening (Table 1). The rate of healthy men was The rate of the nonattenders was close to that of men with cardiovascular symptoms but twice that of all attenders. The rate of men with cardiovascular diseases, which includes men with a history of stroke previous to screening, was nearly seven times that of the healthy men. A total of 98 men died of stroke, and the mortality rate was 0.33 per 1000 person-years. The nonattenders tended to have a higher rate compared with healthy men than men with symptoms of cardiovascular diseases. The number of events in some groups was low. The nonattenders had a rate ratio of 2.85 compared with all attenders. The mean values of the risk factors SBP, DBP, daily cigarette smoking, and total cholesterol together with the TABLE 2. Mean±SD Values of Risk Factors for Stroke of Men Healthy at Screening and Men Reporting Previous Stroke at Screening in the Oslo Study, Norway, 1971 Men Healthy at Screening Men Reporting Previous No Stroke Stroke Mortality Stroke at Screening Risk Factor (N=14 322) (n=81) (n=26) (n=36) Cholesterol, mmol/l 6.38± ± Systolic BP, mm Hg ± ± ±22.5 Diastolic BP, mm Hg 86.2± ± ± Ml risk score 10.07± ± Daily cigarette smoking, % (Cl) (44.4, 46.0) (59.1, 79.1) (70.7, 98.5) (59.2, 85.2) N indicates number of men per group; n, number of cases; BP, blood pressure; Ml, myocardial infarction; and Cl, 95% confidence interval.

3 1486 Stroke Vol 24, No 10 October 1993 TABLE 3. Age-Adjusted Univariate Proportional-Hazards Regression Analysis of Risk Factors for Incidence of First Stroke (Nonfatal and Fatal) and Mortality of Stroke for Healthy Men in the Oslo Study, Norway, 1972 (n=81) Stroke Mortality (n=26) Variable Coeff (SE) St Coeff RR Coeff (SE) St Coeff RR Age, y * (0.0394) (0.0677) Cholesterol, mmol/l * (0.0773) (0.0967) Triglycerides, mmol/l (0.0006) (0.0009) Systolic BP, mm Hg t t (0.0054) (0.0091) Diastolic BP, mm Hg t t (0.0091) (0.0146) Glucose, mmol/l (0.3864) (0.6160) Height, cm (0.0173) (0.0307) Weight, kg (0.0108) (0.0196) Body mass index, kg/m (0.3547) (0.6766) Ml risk score * (0.0028) (0.0028) BP indicates blood pressure; Ml, myocardial infarction. Results are presented as regression coefficient (Coeff), standard error (SE), standardized regression coefficient (St Coeff) (St Coeff=coeffxSD), and relative risk (RR) by an increase of 1 SD. Triglyceride and glucose values are adjusted for time since last meal. *P<.05; tp<.001; tp< Stroke mortality Rate per 1,000 person-year Rate per 1,000 peraon-year 1 if r 0.6 F n_ Diastolic blood pressure, mmhg ( vu Quintiles: Incidence: N: Person-year: ( D Mortality: N: Person-year: ( FIG 1. Line graph shows 12-year incidence rates of nonfatal and fatal stroke and mortality rates of stroke per 1000 person-years by quintile values of diastolic blood pressure t 0.2 Stroke incidence + Stroke mortality - -L Systolic blood pressure, mmhg Quintiles: Incidence: N: Person-year: Mortality: N: Person-year: FIG 2. Line graph shows 12-year incidence rates of nonfatal and fatal stroke and mortality rates of stroke per 1000 person-years by quintile values of systolic blood pressure. 160

4 Haheim et al Risk Factors of and Mortality , Stroke Incik Rate per 1,000 person-year ol Chc dence Stroke mortality rate of mortality from the first quintile to the second. In the lowest quintile of the SBP distribution (Fig 2), no cases of stroke mortality occurred. The incidence risk was of the same magnitude for SBP and DBP for the lowest quintile. The risk leveled off for the fourth and fifth quintiles. Cholesterol was a significant risk factor for mortality, but for incidence it was only of borderline significance W(P=.08) (Fig 3). Increasing age was a significant risk factor for incidence but not for mortality in the group, aged 40 to 49 years. The myocardial infarction risk score was highly predictive of stroke incidence and mortality Alesterol, mmoil/l with only one case of mortality in each of the first and Quintiles: second quintiles. There was no difference in prediction Incidence: N: of incidence and mortality for triglycerides and height. Person-year: Height was inversely related to risk. The results of the Mortality: age-adjusted Cox analysis were largely confirmed by the N: 3 Person-year: interquintile analysis. Comparison of daily smokers versus nonsmokers re- rates 12-yearincidence of nonfatal sulted in a rate ratio of 2.75 for incidence and 6.74 for FIG 3. Line graph showvs and fatal stroke and nnortality rates of stroke per 1000 mortality (Table 4). The cigarette consumption of men person-years by quintile v'alues of total serum cholesterol. who smoked daily seemed to result in dose to an increased risk. The X2 tests for a linear trend were myocardial infarction riisk score are presented in Table 2. nonsignificant (P>.10). The values increased between the groups with increasing Because of the small number of cases, the intermedior predisposing symptoms or dis- ate and great activity levels were grouped together with severity of end points eases. The highest value.s were found in cases of mortality. regard to physical activity at work and at leisure. The number of men simoking was significantly lower in Physical activity at leisure but not at work was found to men who had not sufferied a stroke. Seventy percent of the be inversely related to stroke, indicating a dose response incidence cases and 85 oof the mortality cases were daily (Table 5). The x2 tests for a linear trend were not smokers at the time of screening compared with 45% in significant for mortality (P>.10) or incidence (P=.09). men not suffering a str( oke. The proportional-hazards multiple regression analy- Increasing levels of SBP and DBP at screening were sis for incidence gave a somewhat different result from associated with increa.,sed rates of incidence and mor- that of mortality (Table 6). DBP and daily cigarette tality. DBP was found to be a highly significant predic- smoking were independent risk factors for both end tor for both incidence and mortality in the age-adjusted points. Age and physical activity at leisure were inderds regression analysis (Table 3). pendent risk factors for incidence but not mortality. Cox proportional-haza The second most impsortant risk factor for both inci- Total cholesterol was of no significance for either end dence and mortality was SBP. The coefficient value point. gives the relative predictability of risk factors when comparing incidence a,nd mortality. The coefficients for Discussion SBP and DBP were greater for mortality than for The mortality rate of all men invited to screening in incidence. For DBP (lfig 1) the trend showed a falling this study during a 12-year period was 0.20 compared TABLE 4. Twelve-Year and Mortality in Cigarette Smokers and Nonsmokers in the Oslo Study, Norway, 1972 Stroke Mortality Variable n RR (Cl) n RR (Cl) Nonsmoker Cigarette smoker (1.75, 4.32) (2.69,16.92) 1-9 per day (1.36, 5.47) (1.64, 33.34) per day (1.60, 4.99) (1.47, 23.39) 225 per day (1.15, 5.48) (2.10, 42.65) n Indicates number of cases; RR, rate ratio per 1000 person-years relative to non-cigarette smokers; and Cl, 95% confidence interval.

5 1488 Stroke Vol 24, No 10 October 1993 TABLE 5. Rate Ratios of Increasing Physical Activity at Work and at Leisure for Group 1 of Healthy Men in the Oslo Study, Norway, 1972 Stroke Mortality n RR (Cl) n RR (Cl) Physical activity at work Sedentary Moderate (0.34, 1.23) (0.33, 2.69) Intermediate+great (0.95, 2.75) (0.46, 3.81) Physical activity at leisure Sedentary Moderate (0.38,1.08) (0.33, 2.35) Intermediate+great (0.15, 0.80) (0.03,1.51) n Indicates number of cases; RR, rate ratio of cases per 1000 person-years relative to sedentary; and Cl, 95% confidence interval. with the Bergen study22 rate of 0.55 in men aged 40 to 49 years followed for 20 years. The difference in rate is due most likely to different durations of follow-up. Men from this cohort have been participating in two intervention trials, in which a total of 1010 men were in the treated groups. There was a significant difference in stroke events after 5 years of follow-up in the hypertension trial (intervention, 0; control, 7) but no difference in the dietary and smoking trial (intervention, 3; control, 3). Consequently, the effect of treatment is evident in the results. Results of this study indicate that DBP is a stronger predictor than SBP. This is at variance with the results of the British regional heart study.2 SBP and DBP both predict mortality better than incidence. The association between a single baseline measurement of DBP and the risk of stroke is underestimated due to measurement error.23 Assuming a universal correlation between withinperson measurements of DBP, the method presented by MacMahon et a123 by correcting for regression dilution of 60% can be applied to this study. The increased risk in DBP by an increase of 1 SD of 10.6 mm Hg after adjusting for age, total cholesterol, daily cigarette smoking, and physical activity at leisure is 2.14 (1.61) for stroke incidence and 2.96 (1.97) for stroke mortality. We have not controlled for any treatment effect by antihypertensive medication during the follow-up period. Treatment of hypertension may be the reason that SBP TABLE 6. Proportional-Hazards Multiple Regression Analysis of Risk Factors for Incidence of First Stroke (Nonfatal and Fatal) and mortality of stroke for Healthy Men in the Oslo Study, Norway, 1972 Variables in Model Regression Coefficient SE RR Stroke incidence (n=81) Age, y (SD=2.91) * Cholesterol, mmol/l (SD=1.296) Diastolic BP, mm Hg (SD=10.6) * Daily cigarette smoking (no=0, yes=1) * Physical activity at leisure (1-4) * Stroke mortality (n=26) Age, y Cholesterol, mmol/l Diastolic BP, mm Hg * Daily cigarette smoking (no=0, yes= 1) * Physical activity at leisure (1-4) n Indicates number of cases; RR, relative risk by an increase of 1 SD for a continuous variable and 1 unit change for a coded variable; and BP, blood pressure. *P<.05.

6 Hdheim et al Risk Factors of and Mortality 1489 and DBP did not show the expected exponential increase in the fourth and fifth quintiles (Figs 1 and 2). Age, which is known to be a risk factor for stroke, was not a significant risk factor for stroke mortality. Possible reasons are treatment effect, the narrow age span of 10 years under study, and the young age group of 40 to 49 years with regard to the occurrence of stroke. Total serum cholesterol was found to be a significant risk factor for mortality but not for incidence in ageadjusted univariate analysis. Cholesterol is no longer significant when controlling for age and blood pressure. The importance of cholesterol might be age dependent, as indicated by Tell et al.24 Many studies have provided evidence that there is an increased risk of stroke among cigarette smokers.25 There is a higher proportion of smokers among subjects in this study than others.58 The lack of dose response by daily cigarette smoking is comparable to other cohort studies226 but in contrast to a meta-analysis quoting cohort studies, case-control studies, and an intervention study.25 We found that daily cigarette smoking was a stronger predictor of mortality than of incidence. Physical activity at leisure was inversely related to stroke. The group with an activity level classified as great had no events, which emphasized the preventive aspects of physical activity at leisure. Physical activity is known to exert positive effects on established risk factors such as hypertension and cholesterol. Physically active men are also known to smoke less than inactive men. The results of the univariate analysis were largely confirmed in the multiple regression analysis, which found age, DBP, daily cigarette smoking, and physical activity at leisure to be independent risk factors for stroke incidence. DBP and daily cigarette smoking were found to be independent risk factors for stroke mortality. Because the number of fatal cases was low, no conclusions can be firmly drawn from the results that age and physical activity at leisure were of no significance in the multiple regression analysis of stroke mortality. This study failed to find body mass index, height, weight, glucose, triglycerides, and physical activity at work as risk factors in an age-adjusted univariate analysis for stroke in healthy middle-aged men. In conclusion, this study confirmed that increased blood pressure and daily cigarette smoking were the main risk factors for stroke. Total serum cholesterol was of less importance, possibly because of young age. The infarction risk score was highly predictive of stroke incidence and mortality. Acknowledgments We thank Mona Eggen for her conscientious work in collecting and computerizing the data on morbidity. We thank Professor Egil Arnesen at the Institute of Social Medicine at the University of Troms0 for assistance in the statistical analysis. References 1. Central Bureau of Statistics of Norway. Causes of Death. 1980, Universitetsforlaget. 2. Shaper AG, Phillips AN, Pocock SJ, Walker M, MacFarlane PW. Risk factors for stroke in middle aged British men. Br Med J. 1991;302: Ueshima H, lida M, Shimamoto T, Konishi M, Tsujioka K, Tanigabi M, Nakanishi N, Ozawa H, Kojima S, Komachi Y. Multivariate analysis of risk factors for stroke. Prev Med. 1980;9: Reed DM. The paradox of high risk of stroke in populations with low risk of coronary heart disease. Am J Epidemiol. 1990;131: Wolf PA, D'Agostine RB, Kannel WB, Bonita R, Belanger AJ. Cigarette smoking as a risk factor for stroke. JAMA. 1988;259: Harmsen P, Rosengren A, Tsipogianni A, Wilhelmsen L. Risk factors for stroke in middle-aged men in Gdteborg, Sweden. Stroke. 1990;21: Salonen JT, Puska P, Tuomilehto J, Homan K. Relation of blood pressure, serum lipids, and smoking to the risk of cerebral stroke: a longitudinal study in Eastern Finland. Stroke. 1982;13: Iso H, Jacobs DR, Wentworth D, Neaton JD, Cohen JD. Serum cholesterol levels and six-year mortality from stroke in 350,977 men screened for the Multiple Risk Factor Intervention Trial. N Engl J Med. 1989;320: Westlund K, Nicolaysen R. Ten-year mortality and morbidity related to serum cholesterol. Scand J Clin Lab Invest. 1972; 30(suppl 127): Boysen G, Nyboe J, Appleyard M, S0rensen PS, Boas J, Somnier F, Jensen G, Schnohr P. Stroke incidence and risk factors for stroke in Copenhagen, Denmark. Stroke. 1988;19: Welin L, Svardsudd K, Wilhelmsen L, Larsson B, Tibblin G. Analysis of risk factors for stroke in a cohort of men born in N Engl J Med. 1987;317: Whelton PK, Klag MJ. Recent trends in the epidemiology of stroke: what accounts for the persistent decline in stroke mortality in Western nations? Curr Opin Cardiol. 1987;2: J0rgensen HS, Plesner A-M, Hubbe P, Larsen K. Marked increase of stroke incidence in men between 1972 and 1990 in Fredriksberg, Denmark. Stroke. 1992;23: Alfredsson L, von Arbin M, de Faire U. Mortality from and incidence of stroke in Stockholm. Br Med J. 1986;292: Indredavik B, Bakke F, Solberg R, Rokseth R, Lund Haheim L, Holme I. Benefit of a stroke unit: a randomized controlled trial. Stroke. 1991;22: Dyken ML, Wolf PA, Barnett HJM, Bergan JJ, Hass WK, Kannel WB, Kuller L, Kurtze JF, Sundt TM. Risk factors in stroke: a statement for physicians by the Subcommittee on Risk Factors and Stroke of the Stroke Council. Stroke. 1984;15: Leren P, Askevold EM, Foss OP, Fr0ili A, Grymyr D, Helgeland A, Hjermann I, Holme I, Lund-Larsen PG, Norum KR. The Oslo Study: cardiovascular disease in middle-aged and young men. Acta Med Scand. 1975;199(suppl 588): Hjermann I, Byre K, Holme I, Leren P. Effect of diet and smoking intervention on the incidence of coronary heart disease: a randomised trial in healthy men: the Oslo Study. Lancet. 1981;2: Helgeland A. Treatment of mild hypertension: a five year controlled drug trial. Am J Med. 1980;69: Thelle DS, F0rde OH, Try K, Lehmann EL. The Troms0 Heart Study: methods and main results of the cross-sectional study. Acta Med Scand. 1976;200: Bjartveit K, Foss OP, Gjervig T, Lund-Larsen PG. The cardiovascular disease study in Norwegian counties: background and organization. Acta Med Scand. 1979;(suppl 634): Selmer R. Blood pressure and twenty-year mortality in the city of Bergen, Norway. Am J Epidemiol. 1992;136: MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease, I: prolonged differences in blood pressure: prospective studies corrected for the regression dilution bias. Lancet. 1990;335: Tell GS, Crouse JR, Furberg CD. Relation between blood lipids, lipoproteins, and cerebrovascular atherosclerosis: a review. Stroke. 1988;19: Shinton R, Beevers G. Meta-analysis of relation between cigarette smoking and stroke. BMJ. 1989;298: Abbott RD, Yin Y, Reed M, Yano K. Risk of stroke in male cigarette smokers. N Engl J Med. 1986;315: