点击显示 收起
the Institute of Clinical Neuroscience (P.H.), Sahlgrenska University Hospital
the Cardiovascular Institute (G.L., L.W.), Section of Preventive Cardiology, Sahlgrenska University Hospital
the Department of Medicine (A.R.), Sahlgrenska University Hospital/stra, Gteborg, Sweden.
Abstract
Background and Purpose— To estimate the predictive value of risk factors for stroke measured in midlife over follow-up extending through 28 years.
Methods— A cohort of 7457 men 47 to 55 years of age and free of stroke at baseline year 1970 were examined. Risk of stroke was analyzed for the entire period and for 0 to 15, 16 to 21, and 22 to 28 years of follow-up using age-adjusted and multiple Cox regression analyses.
Results— Age, diabetes, and high blood pressure were independently associated with increased risk of stroke for the entire 28 years and for each of the periods. Previous transient ischemic attacks, atrial fibrillation, history of chest pain, smoking, and psychological stress were independently related to stroke for the entire follow-up period and also during the first 1 or 2 successive periods. Family history of stroke or of coronary disease carried no independent prognostic information, nor did serum cholesterol. Elevated body mass index predicted stroke during the later part of the follow-up and so did (almost) low physical activity during leisure time, together with antihypertensive medication at baseline.
Conclusions— High blood pressure and diabetes retain their importance as stroke risk factors also over an extended follow-up into old age. A family history of cardiovascular disease was not significantly related to outcome. Transient ischemic attacks, atrial fibrillation, stress, smoking, and a history of chest pain were associated with outcome only for the first or the first 2 periods. High body mass index and antihypertensive medication at baseline emerged as risk factors in the second and third decades.
Key Words: risk factors stroke
Introduction
Risk factors for stroke have been identified in many studies, mostly from cohorts with <10 years of follow-up. However, estimates of risk might differ when cohorts are followed over longer time periods1–5 or when risk factor levels are studied consecutively over time. Some studies have shown the effect of risk factors over very long periods of time as well as in successive time periods.6,7 One study reported that hypertension, the single most important risk factor for stroke, has a diminishing impact with increasing age.8 However, few studies have considered to what extent risk factors measured in midlife continues to be predictive over a much extended follow-up.
We reported in 1990 on stroke risk factors over a 12-year follow-up period in a cohort of 7494 middle-aged men in Gteborg.9 The men in this cohort have now been followed altogether for 28 years. The present study analyzes the predictive value of 14 baseline factors on the development of stroke events (nonfatal and fatal) during the whole period of follow-up and during 3 consecutive time periods. A similar study of coronary events in the same cohort of men was published recently.10
Subjects and Methods
The Multifactor Primary Prevention Study started in Gteborg in 197011 and was originally an intervention trial against smoking, hypercholesterolemia, and hypertension at predefined levels in an intervention group of 10 000 men, a random third of all men in the city born between 1915 and 1925 except for men born in 1923. There were 2 control groups of 10 000 men each from the same age groups of the city. A first screening examination took place between 1970 and 1973 and a second during 1974 to 1977. At follow-up 10 years after the first examination, serum cholesterol (S-Chol), smoking, and blood pressure had decreased in all 3 groups, and no significant differences were detected in the pattern of risk factors or in outcome between the intervention and control groups.12 Thus, any changes brought about by the intervention took place in the general population as well, and the present study group (the intervention group) is therefore considered to be representative of the background population in the city.
The present study uses baseline information from the examination in 1970 to 1973. Responders to a postal questionnaire were invited to a screening examination in hospital. In the questionnaire, information was collected on personal and family history of heart disease and stroke, previous myocardial infarction, stroke, and diabetes mellitus. Of the participating 7494 men (75% of the sample), 37 reported a previous stroke at baseline, leaving 7457 men who form the study population of the present study. A history of transient ischemic attacks (TIAs) was considered present if previous focal neurological symptoms lasting <24 hours were reported. Chest pain on exertion was considered present if triggered by climbing 2 flights of stairs. Smoking habits were coded in 5 levels. In the present analysis, never smokers and former smokers were coded together as nonsmokers, and all current smokers were considered together. Psychological stress was assessed by using 1 question on self-perceived stress and rated from 1 through 6, with 5 and 6 defined as continuous stress during 1 or 5 years, respectively. Physical leisure time activity was coded as: (1) sedentary, (2) moderate, or (3) strenuous and regular. Socioeconomic class was coded according to the Swedish socioeconomic classification system (socioeconomic index), with 5 occupational classes: (1) unskilled and semiskilled workers, (2) skilled workers, (3) foremen in industrial production and assistant nonmanual employees, (4) intermediate nonmanual employees, and (5) employed and self-employed professionals, higher civil servants, and executives. The men also reported ongoing medication for hypertension.
Screening examinations were performed in the afternoon. Blood pressure was measured after 5 minutes of rest to the nearest of 2 mm Hg with the subjects seated. In a subset, blood pressure was checked in the morning, at a more relaxed situation, and was then found to be considerably lower. Body height was measured to the nearest centimeter and body weight to the nearest 0.1 kg. Body mass index (BMI) was calculated as weight (kg)/height (m).2 S-Chol concentrations (from a sample taken after fasting for 2 hours) were determined according to standard laboratory procedures. For continuous variables, systolic blood pressure (SBP), BMI, and S-Chol quintile 5 was compared with quintile 1 as reference. All men with medical treatment for hypertension at baseline were defined as treated hypertension. An ECG was done and used for diagnosing atrial fibrillation (AF). The screening examination was concluded with a structured interview by a physician regarding previous and ongoing health problems, and followed by a check-up of hospital records.
The follow-up extended through 1998. End points of first-ever stroke in participants free of previous stroke were registered from several sources. Stroke events up to 65 years of age were recorded using specified criteria in a special stroke register.13 In addition, 1 nurse and 1 medical technician checked case records manually for all diagnoses of inpatients in the hospitals in the city from the start of the study. Further, a hospital discharge register has operated on a nationwide basis since 1987, but all discharges from Gteborg hospitals have been entered in the national register since 1970 (except 1976 because of a legislative change for that single year). A file of the men of the study was run against this register, thereby detecting men admitted outside the city (23% of all admissions). A manual check of hospital admissions because of myocardial infarction, AF, and stroke revealed that the national register missed 3% of the events. The Swedish cause of death register was also matched against the file of the men in the study. In 1987, the International Classification of Diseases (ICD) was changed from the eighth to the ninth revision, and in 1997, there was a change to ICD-10. These changes were taken into account in the present analyses. The ICD-8 codes were changed to ICD-9 below. For the present analyses, the following principal or contributory diagnoses at hospitalization or death were used: stroke, nonfatal and fatal, ICD-9 codes 430–434 and 436 and ICD-10 codes I60, I61, I63, and I64.
Statistical Methods
The main objective of the study was to examine to what extent risk factors measured at baseline at 47 to 55 years of age predicted stroke during long-term follow-up. Men with a stroke before the baseline examination were excluded. In previous analyses, it was found that, despite the narrow age range, increasing age at baseline was associated with increased risk of stroke during follow-up. Therefore, the analyses are age adjusted in Tables 3 and 4. Subjects with missing values for the risk factors were excluded from the analyses. In a first analysis, we considered the entire follow-up period, and in a second analysis, 3 successive periods of time: 0 to 15 years, 16 to 21 years, and 22 to 28 years. These intervals were selected to obtain reasonably similar numbers of events for each period.
Psychological stress here denotes grading 5 to 6 compared with grades 1 to 4. When continuous variables were analyzed, the fifth quintile was compared with quintile 1 in the tables, but each quintile was compared with the first quintile in the multivariate analyses. The fifth quintile cutoff value for SBP was 167 mm Hg, for BMI 28.0 kg/m2, and for S-Chol 7.31 mmol/L. The association between the hazard for stroke and the risk factors were modeled through Cox regression14 for the whole period and for the 3 periods separately. The same set of covariates was entered into all regression models.
Results
For the whole period of 28 years, 1019 cases of first-ever stroke occurred. Table 1 gives the number of events in each time period and the person years of exposure together with the age ranges.
Table 2 shows the distribution of risk factors at baseline. An analysis of participants with missing information on family history of acute myocardial infarction and stroke or stress showed that their presence of baseline factors were similar to all others, although they had a slightly higher prevalence of diabetes, high BMI, high S-Chol, and low physical activity.
Table 3 shows age-adjusted hazard ratios for stroke for individual risk factors during the whole follow-up period. With the exception of myocardial infarction in either parent, elevated S-Chol, and socioeconomic class, all factors predicted stroke. All significant factors retained their significance in multivariate analysis, except stroke in either parent and low leisure time physical activity.
The risks associated with baseline factors were not similar across all 3 time periods, as shown in Table 4. A history of TIAs or chest pain increased risk significantly in the first 1 or 2 periods only, and so did a history of AF, as well as psychological stress. However, in multivariate analysis, the elevated risk over the first period associated with a history of TIAs was no longer significant, although it appeared to be so during the second period. A history of chest pain was significant in both univariate (age-adjusted) and multivariate analysis during the first period. A history of diabetes mellitus and smoking at baseline were independently associated with increased risk during the first 2 periods but not during the third. High blood pressure carried a decreasing but significantly elevated risk through all 3 periods, also when adjusted for other risk factors, although just under significance in the third period. Family history of stroke, coronary events, or elevated S-Chol were not independently predictive for any time period. BMI did not predict stroke until after the first 15 years but was an independent predictor for the last 2 periods. Low physical activity predicted stroke during the last period only but not significantly so after adjustment for other factors. Finally, antihypertensive medication at baseline did, by itself, significantly increase the risk of stroke and so even for the third time period after adjustment for all other factors.
Discussion
Nearly all of the baseline factors, including TIAs, AF, diabetes, psychological stress, chest pain, smoking, high blood pressure, and high body mass as well as treatment for hypertension at baseline were independent predictors over the entire 28-year period. High blood pressure almost predicted stroke across all 3 periods, whereas high BMI and treatment for hypertension emerged as significant risk factors during the later periods of follow-up. AF was associated with a very high risk but only for the first 15 years.
For TIAs, the immediate risk of stroke is high,15 and an increased risk continues for many years.16 Few men in our study had a history of TIAs, and in addition, TIA was strongly associated with other adverse factors, particularly chest pain, and this probably explains why TIA was not constantly predictive of stroke in the study. For patients with AF, there is a continuing risk of stroke, supporting current recommendations on anticoagulants in patients with AF.17 But in our study, the stroke risk associated with AF decreased in later time periods, either because treatment prevented stroke or because the number of persons at risk decreased markedly through death or stroke. Further, among the men in the cohort without AF at baseline, 1315 experienced AF and received treatment in hospital during follow-up. An analysis of stroke events among these men showed that the majority of strokes occurred during the first 1 to 3 years after an episode of AF, but there was also an aggregation of stroke events the year before AF, indicating that AF may have been intermittent and only diagnosed after an embolic stroke. The relative risk of stroke for all these men together was 2.07 (CI, 1.70 to 2.54) through 28 years.
For diabetes, the relative risk would probably have increased with time if new cases of diabetes successively had been detected in the population. Another 671 men free of diabetes at baseline were treated in hospital with a principal or contributory diagnosis of diabetes during follow-up, but data on nonhospitalized new diabetes events are not known. There was an aggregation of stroke events during the first year of diabetes onset that slowly decreased during the following 10 years. The relative risk of stroke for these men, including the men with diabetes at baseline, was 2.42 (CI, 1.94 to 3.02) through 28 years.
Some risk factors were statistically significant in some periods but not in others. We think this could be, in part, caused by some risk factors playing a role clinically in some periods and not in others or even could become statistically insignificant in some periods because of covariation. We did formal tests for time interaction for TIA and found that the hazard decreased with later years, whereas for diabetes and smoking, time interaction only slightly decreased with time.
The importance of psychological stress was apparent through the whole period of 28 years. Smaller numbers at risk for the 3 successive time periods may account for the loss of numerical significance. Others have found self-reported stress to be a stroke risk, although only for fatal stroke,18 but reports have varied about the relation of stress to stroke.19
High BMI significantly predicted stroke in time periods 2 and 3, consistent with recently published findings from the same study population20 and other studies.21,22 Truncal fat may be more important than BMI,23 but more sophisticated methods of determining body fatness were not commonly used in epidemiologic studies at the time, and measures of central obesity such as waist circumference were not available. In a recent study of risk of coronary disease during 28 years of follow-up of the same cohort of men, a high BMI was also more predictive during the last period.10
Physical activity has a protective effect against stroke and other cardiovascular disease.24,25 Our results for low physical activity support this but not independently for the last period, probably because of the strong associations with other factors.
The number of men at risk decreased successively and could cause the diminishing effect of smoking in the third time period. Findings were similar for coronary disease10 and could as well be attributable to the decreasing numbers of smokers in the same cohort over time.
Treatment of hypertension reduces the risk of stroke.26 But high blood pressure carries an increase in risk of stroke even when treated.27,28 In our study, hypertensive persons at baseline were recruited for and treated at a special hypertension unit at the university hospital as part of the project.27 Still, in our study, high blood pressure at baseline persisted as the most important risk factor throughout all 3 periods, and being on treatment at baseline appeared to be a risk in itself. This increase in risk could not be ascribed to high levels of the other risk factors within this group of 406 men on treatment at baseline because they had only very slightly elevated mean levels of SBP, BMI, and S-Chol when compared with the rest of the cohort.
The importance of socioeconomic conditions have been shown before to relate to cardiovascular disease as well as to stroke.29 Here, we did not find low social class to significantly predict stroke.
In conclusion, to a large extent, our present results are similar to those reported in 1990.9 At 28 years of follow-up, apparently most risk factors present at baseline still carry a stroke risk, although only hypertension was significant through all 3 time periods when correcting for all other variables. However, to this is added that some items suspected of being stroke risk factors in fact developed into carrying significant risks after 15 to 22 years: low physical activity and high BMI at baseline. Also, medication for hypertension at baseline developed into an independent risk during the later years of follow-up. Because numbers at risk diminish in time and exposed cases occur in the population that was unexposed at baseline, results from other long-term follow-ups might be valuable.
Acknowledgments
Sources of Funding
This study was supported by the Bank of Sweden Tercentenary Fund, the Swedish Research Council, and the Swedish Heart Lung Foundation.
Disclosures
None.
References
Ohlson LO, Svardsudd K, Welin L, Eriksson H, Wilhelmsen L, Tibblin G, Larsson B. Fasting blood glucose and risk of coronary heart disease, stroke, and all-cause mortality: a 17-year follow-up study of men born in 1913. Diabet Med. 1986; 3: 33–37.
Welin L, Svardsudd K, Wilhelmsen L, Larsson B, Tibblin G. Analysis of risk factors for stroke in a cohort of men born in 1913. N Engl J Med. 1987; 317: 521–526.
Nakayama T, Date C, Yokoyama T, Yoshiike N, Yamaguchi M, Tanaka H. A 15.5-year follow-up study of stroke in a Japanese provincial city: the Shibata Study. Stroke. 1997; 28: 45–52.
Rastenyte D, Tuomilehto J, Domarkiene S, Cepaitis Z, Reklaitiene R. Risk factors for death from stroke in middle-aged Lithuanian men: results from a 20-year prospective study. Stroke. 1996; 27: 672–676.
Tuomilehto J, Rastenyte D, Jousilahti P, Sarti C, Vartiainen E. Diabetes mellitus as a risk factor for death from stroke: prospective study of the middle-aged Finnish population. Stroke. 1996; 27: 210–215.
Seshadri S, Wolf PA, Beiser A, Vasan RS, Wilson PW, Kase CS, Kelly-Hayes M, Kannel WB, D’Agostino RB. Elevated midlife blood pressure increases stroke risk in elderly persons: the Framingham Study. Arch Intern Med. 2001; 161: 2343–2350.
Hart CL, Hole DJ, Smith GD. Risk factors and 20-year stroke mortality in men and women in the Renfrew/Paisley study in Scotland. Stroke. 1999; 30: 1999–2007.
Whisnant JP, Wiebers DO, O’Fallon WM, Sicks JD, Frye RL. Effect of time since onset of risk factors on the occurrence of ischemic stroke. Neurology. 2002; 58: 787–794.
Harmsen P, Rosengren A, Tsipogianni A, Wilhelmsen L. Risk factors for stroke in middle-aged men in Goteborg, Sweden. Stroke. 1990; 21: 223–229.
Wilhelmsen L, Lappas G, Rosengren A. Risk of coronary events by baseline factors during 28 years follow-up and three periods in a random population sample of men. J Intern Med. 2004; 256: 298–307.
Wilhelmsen L, Tibblin G, Werko L. A primary preventive study in Gothenburg, Sweden. Prev Med. 1972; 1: 153–160.
Wilhelmsen L, Berglund G, Elmfeldt D, Tibblin G, Wedel H, Pennert K, Vedin A, Wilhelmsson C, Werko L. The multifactor primary prevention trial in Goteborg, Sweden. Eur Heart J. 1986; 7: 279–288.
Harmsen P, Tibblin G. A stroke register in Gteborg, Sweden. Acta Med Scand. 1972; 191: 463–470.
Cox DR. Regression models and life-tables. J R Stat Soc. 1972; 34: 187–220.
Lovett JK, Dennis MS, Sandercock PA, Bamford J, Warlow CP, Rothwell PM. Very early risk of stroke after a first transient ischemic attack. Stroke. 2003; 34: e138–e140.
Clark TG, Murphy MF, Rothwell PM. Long-term risks of stroke, myocardial infarction, and vascular death in "low risk" patients with a non-recent transient ischemic attack. J Neurol Neurosurg Psychiatry. 2003; 74: 577–580.
Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med. 1999; 131: 492–501.
Truelsen T, Nielsen N, Boysen G, Gronbak M. Self-reported stress and risk of stroke. The Copenhagen City Heart Study. Stroke. 2003; 34: 01.
Iso H, Date C, Yamamoto A, Toyoshima H, Tanabe N, Kikuchi S, Kondo T, Watanabe Y, Wada Y, Ishibashi T, Suzuki H, Koizumi A, Inaba Y, Tamakoshi A, Ohno Y. Perceived mental stress and mortality from cardiovascular disease among Japanese men and women: the Japan Collaborative Cohort Study for Evaluation of Cancer Risk Sponsored by Monbusho (J Am Coll Cardiol Study). Circulation. 2002; 106: 1229–1236.
Jood K, Jern C, Wilhelmsen L, Rosengren A. Body mass index in mid-life is associated with a first stroke in men: a prospective population study over 28 years. Stroke. 2004; 35: 2764–2769.
Li C, Engstrom G, Hedblad B, Berglund G, Janzon L. Risk factors for stroke in subjects with normal blood pressure: a prospective cohort study. Stroke. 2005; 36: 234–238.
Tanne D, Medalie JH, Goldbourt U. Body fat distribution and long-term risk of stroke mortality. Stroke. 2005; 36: 1021–1025.
Kragelund C, Omland T. A farewell to body-mass index Lancet. 2005; 366: 1589–1591.
Rosengren A, Wilhelmsen L. Physical activity protects against coronary death and deaths from all causes in middle-aged men. Evidence from a 20-year follow-up of the primary prevention study in Goteborg. Ann Epidemiol. 1997; 7: 69–75.
Lee IM, Hennekens CH, Berger K, Buring JE, Manson JE. Exercise and risk of stroke in male physicians. Stroke. 1999; 30: 1–6.
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. Part 1, Prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet. 1990; 335: 765–774.
Almgren T, Persson B, Wilhelmsen L, Rosengren A, Andersson OK. Stroke and coronary heart disease in treated hypertension—a prospective cohort study over three decades. J Intern Med. 2005; 257: 496–502.
Li C, Engstrom G, Hedblad B, Berglund G, Janzon L. Blood pressure control and risk of stroke: a population-based prospective cohort study. Stroke. 2005; 36: 725–730.
Thrift AG, Dewey HM, Sturm JW, Paul SL, Gilligan AK, Srikanth VK, Macdonell RAL, McNeil JJ, Macleod MR, Donnan GA. Greater incidence of both fatal and nonfatal strokes in disadvantaged areas: the Northeast Melbourne Stroke Incidence Study. Stroke. 2006; 37: 877–882.