ORIGINAL INVESTIGATION. Lifestyle and 15-Year Survival Free of Heart Attack, Stroke, and Diabetes in Middle-aged British Men

ORIGINAL INVESTIGATION Lifestyle and 15-Year Survival Free of Heart Attack, Stroke, and Diabetes in Middle-aged British Men S. Goya Wannamethee, PhD; A. Gerald Shaper, FRCP; Mary Walker, MA; Shah Ebrahim, FRCP Background: To examine the relationship between modifiable lifestyle factors (smoking, physical activity, alcohol intake, and body mass index [BMI]) and the likelihood of 15-year survival free of major cardiovascular end points and diabetes in middle-aged men. Methods: A prospective study of 7142 men aged to 59 years at screening with no history of coronary heart disease, diabetes, and stroke drawn from 1 general practice in each of 24 British towns and followed up for 15 years. Main Outcome Measures: Death from any cause and a combined end point, including survival free of heart attacks or stroke or the development of diabetes over a follow-up of 15 years for each man. Results: During the 15-year follow-up, there were 164 deaths from all causes, 77 major heart attacks (fatal and nonfatal), 247 stroke events (fatal and nonfatal), and 252 cases of diabetes among the 7142 men. After adjustment for age and each of the other modifiable lifestyle factors, the risk of the combined end point (death or having a heart attack, stroke, or diabetes) went up significantly with increasing smoking levels and from BMI levels of 26 kg/m 2 or higher, and decreased significantly with increasing levels of physical activity up to levels of moderate activity with no further benefit thereafter (heavy smoking vs never: relative risk [RR] [odds], 2.5; 95% confidence interval [CI], 2.12-2.94; BMI 3 vs - 21.9 kg/m 2 : RR, 2.11; 95% CI, 1.71-2.62; moderate vs inactive: RR,.6; 95% CI,.5-.72). Light drinking (vs occasional) showed a relatively small but significant reduction in risk (RR,.84; 95% CI,.74-.96). Using Cox predictive survival models, the estimated probability of surviving 15 years free of cardiovascular events and diabetes in a man aged 5 years ranged from 89% in a moderately active man at BMI levels of to 24. kg/m 2 who had never smoked to 42% in an inactive smoker with BMI level of 3 kg/m 2 or higher. Conclusions: Modifiable lifestyles (smoking, physical activity, and BMI) in middle-aged men play an important role in long-term survival free of cardiovascular disease and diabetes. These findings should provide encouragement for public health promotion directed toward middle-aged men. Arch Intern Med. 1998;158:2433-24 From the Department of Primary Care and Population Sciences, Royal Free Hospital School of Medicine, London, England. THE ROLE of modifiable lifestyle factors (body mass index [BMI], which is a measure of weight in kilograms divided by the square of the height in meters, smoking, alcohol intake, and physical activity) in cardiovascular disease and all-cause mortality has been extensively studied. 1,2 Previous epidemiological studies have generally focused on specific lifestyle factors in relation to specific end points (mortality, coronary heart disease [CHD], stroke, or diabetes) and have usually presented results in terms of the relative risks (odds) of developing the disease. There is considerable evidence that each of the lifestyle factors contribute to the incidence of cardiovascular disease and to all-cause mortality, and some studies 3-5 have calculated survival estimates or average life expectancy. However, there is more to life than survival. The absence of major chronic disease in the years of retirement should be required as a desirable target. Our recent report 6 has focused on the use of a wider spectrum of end points in determining a healthy BMI, ie, the level associated with the lowest mortality and the lowest incidence of CHD, stroke, and diabetes mellitus. This article examines prospectively the effects that lifestyle factors have on risk of not only surviving for 15 years but also surviving free of major cardiovascular problems and diabetes, using all-cause mortality, major cardiovascular events, or diabetes as a combined end point. We aim to provide estimates of the chances of surviving 15 years free of major cardiovascular disease and diabetes in a middleaged male population, in a manner that is of meaning to public health professionals and to the general public. 2433

SUBJECTS AND METHODS SUBJECTS The British Regional Heart Study is a large prospective study of cardiovascular disease consisting of 7735 men aged to 59 years selected from the age-sex registers of 1 group general practice in each of 24 towns in England, Wales, and Scotland. The criteria for selecting the town, the general practice, and the subjects, as well as the methods of data collection, have been reported. 7 The practices selected for study in each town reflected the social class distribution of middle-aged men in that town, and the cohort as a whole was representative of the social class distribution of middle-aged men in Great Britain. Research nurses administered to each man a standard questionnaire that included questions on smoking habits, alcohol intake, physical activity, and medical history. Several physicalmeasurementsincludingheightandweightwereperformed and nonfasting blood samples were taken for measurement of biochemical and hematologic variables. The men were asked to recall a physician s diagnosis of angina, myocardial infarction (heart attack or coronary thrombosis), stroke, diabetes, and several other disorders listed on the questionnaire. SOCIAL CLASS The longest-held occupation of each man was recorded and then coded in accordance with the Registrar General s occupational classification, and men were grouped into 1 of 6 social classes: 1, 2, 3 nonmanual, and 3 manual, and 4 and 5 (manual). Those whose longest occupation was in the armed forces formed a separate group. LIFESTYLE FACTORS This presentation is concerned only with outcomes based on lifestyle variables measured at screening. Body Mass Index The BMI was used as an index of relative weight. Smoking Status The men were classified according to their current smoking status: those who had never smoked, former cigarette smokers, and current smokers at 4 levels (1-19,, 21-39, and cigarettes/day). Those who had only ever smoked pipe or cigars are grouped as never smoked. Former cigarette smokers who were currently pipe/cigar smokers are classified as former cigarette smokers. Alcohol Intake Alcohol consumption was recorded using questions on frequency, quantity, and type, similar to those used in the 1978 General Household Survey. 8 The men were classified into 5 groups according to their estimated current weekly alcohol intake: none, occasional ( 1 U/wk), light (1-15 U/ wk), moderate (16-42 U/wk), and heavy ( 42 U/wk). (A unit was defined as a glass of wine or a tot of spirit or half a pint of beer (1 U = 8-1 g of alcohol). 9 Occasional drinkers rather than nondrinkers are used as a baseline in analyses involving alcohol intake. 1-12 Physical Activity At screening the men were asked to indicate their usual pattern of physical activity, under the headings of regular walking or cycling, recreational activity, and sporting (vigorous) activity. Regular walking and cycling relate to weekday journeys that included travel to and from work. RESULTS During the follow-up period of the 7142 men with no history of diabetes, stroke, or heart attacks at screening, there were 164 deaths from all causes, 77 major CHD events (fatal and nonfatal), 247 strokes, and 252 incident cases of diabetes. LIFESTYLE FACTORS AND MORTALITY Table 1showstheage-adjusteddeathratesper1personyears and age-adjusted relative risk of death from all causes for the major lifestyle factors (smoking, physical activity, alcohol intake, and BMI). Smoking (positively) and physical activity(negatively) were strongly associated with risk of allcause mortality. Light drinking was associated with a small but significantly lower relative risk than occasional drinking. A U-shaped relationship was seen between BMI and allcause mortality with mortality only significantly increased in thin ( kg/m 2 ) and to a lesser extent in very obese men ( 3 kg/m 2 ). Since these factors are interrelated and are all associated with social class, we have examined the relationship between the lifestyle factors and risk of all-cause mortality adjusting for each of the other factors and then in addition for social class. The relationships seen for smoking, physical activity, and alcohol intake persisted after adjustment for each of the other factors. These adjustments attenuated the increased risk seen in thin men and increased the risk in very obese men. Further adjustment for social class made little difference to the associations seen. AGE AND COMBINED END POINT Of the 7142 men, 1766 men had died or developed a major cardiovascular event (heart attack or stroke) or diabetes. The 15-year probability of surviving free of these major events in the cohort was 75.3%. Figure 2 shows the 15-year survival free of the above major events by age at initial screening. The - to 44-year-old group had an 87.7% probability of surviving free of major events after 15 years compared with 61.% in the 55- to 59-year-old groups. LIFESTYLE FACTORS AND COMBINED END POINT Table 2 shows the age-adjusted rates per 1 personyears of dying or developing a major CHD or stroke event or diabetes (combined end point) by the major lifestyle fac- 2434

Recreational activity includes gardening, pleasure walking, and do-it-yourself jobs. Sporting activity includes running, golf, swimming, tennis, sailing, and digging. A physical activity (exercise) score was derived for each man based on the frequency and type (intensity) of physical activity. 13 The men were initially grouped into 6 broad categories based on their total score: inactive, occasional, light, moderate, moderately vigorous, and vigorous. Active was defined as men who were engaged in at least moderate levels of activity. FOLLOW-UP All men have been followed up for all-cause mortality, cardiovascular morbidity, and development of non insulin-dependent diabetes mellitus from the initial January 1978-July 198 screening up to December 1995, a mean period of 16.8 years (range, 15.5-18 years) and follow-up has been achieved for 99% of the cohort. 14 This analysis is based on a 15-year follow-up for each man. All men with recall of a physician diagnosis of CHD, stroke, or diabetes at screening have been excluded from the analysis (n = 593). Information on death was collected through the established tagging procedures provided by the National Health Service registers in Southport (England and Wales) and Edinburgh (Scotland). Evidence regarding nonfatal heart attacks and strokes and diabetes were obtained by reports from general practitioners, by biennial reviews of the patients notes through to the end of the study period, and from personal questionnaires to surviving subjects at the 5th year and 12th year after initial examination. Nonfatal stroke events were those that produced a neurologic deficit that was present for more than 24 hours. A nonfatal heart attack was diagnosed according to the World Health Organization criteria, which included any report of myocardial infarction accompanied by at least 2 of the following: a history of severe chest pain, electrocardiographic evidence of myocardial infarction, and cardiac enzyme changes associated with myocardial infarction. A diagnosis of diabetes was not accepted on the basis of self-completed questionnaire data unless confirmed in the primary care records. COMBINED END POINT In the analyses a combined end point is used, which is defined as death from any cause or the development of heart attack, stroke, or diabetes during follow-up. STATISTICAL METHODS Direct standardization was used to obtain age-adjusted rates per 1 person-years using the study population as the standard. Cox proportional hazards model was used to assess the relation between lifestyle factors and outcome and to estimate the 15-year predicted survival probability free of major cardiovascular events and diabetes. 15-17 Time to event was defined as the time to development of the first nonfatal event (heart attack, stroke, or diabetes) or death. The relative risks for each level of lifestyle factor were obtained fitting the factors as categorical variables BMI (6 levels), physical activity (6 levels), smoking (5 levels), and alcohol intake (5 levels). In the adjustment, age was fitted continuously. In Figure 1, the survival probabilities were obtained from a model based on age, physical activity, smoking, and BMI. The survival probability can be written as, S(t)=S (t)exp(b z) where b is the vector of coefficients; exp, exponential; z, the risk factor variables; and S (t), the underlying survival function corresponding to the average levels or baseline level of risk factor. 16,17 tors (smoking, physical activity, alcohol intake, and BMI). Smoking, physical activity, and BMI were significantly and strongly associated with the combined end point. Light drinking was associated with a small but significant reduction in risk. These relationships persisted after further adjustment for each of the other factors. Additional adjustment for social class made little difference to the relationships. Since the presence of other chronic disease may affect the lifestyle characteristic, we examined the data after exclusion of a further 2813 men who recalled at least 1 physician diagnosis of heart trouble, high blood pressure, peptic ulcer, gout, gallbladder disease, thyroid disease, arthritis, bronchitis, and asthma. This made small differences to the relative risks seen in Table 2. SMOKING AND LIFESTYLE FACTORS We have also examined the relationships with the combined end point separately by smoking status at baseline: never, former, and current cigarette smokers. Since men with BMI levels of to 23.9 kg/m 2 showed similar risks, these 2 groups have been combined and 6 groups are used. Similarly, the physical activity groups have been combined and 4 groups are used: inactive, occasional, light, and active (at least moderately active). Figure 3 shows the age-adjusted rates per 1 person-years and Table 3 shows the adjusted relative risk (adjusted for each of the other factors) for the combined end point for each of the lifestyle factors within each smoking category separately. Current smokers had higher rates than never smokers irrespective of BMI levels, physical activity, or alcohol intake. Former smokers showed intermediate rates. With the exception of alcohol intake, the relationships were fairly similar in all 3 smoking groups (Figure 3, Table 3). After adjustment, light or moderate drinking was associated with lower risk than occasional drinking in former smokers only. This may reflect the high prevalence of ill health in men who have given up smoking and given up or reduced their alcohol intake. 18 In never and current smokers, nondrinkers had the highest risk, but there was little difference between occasional and light or moderate drinking. CALCULATION OF LIFESTYLE AND 15-YEAR PREDICTED PROBABILITY Smoking, BMI, and physical activity are the major lifestyle factors that have a significant effect on health out- 2435

Inactive Active 1 Never Smokers 1 Former Smokers 1 Current Smokers Predicted 15-Year Survival Free, % 8 6 83 89 77 85 67 78 8 78 6 86 71 81 59 72 8 6 67 77 56 7 42 58-24 28 3+ -24 28 3+ -24 28 3+ BMI, kg/m 2 BMI, kg/m 2 BMI, kg/m 2 Figure 1. Predicted probability of 15-year survival free of coronary heart disease, stroke, and diabetes in a man aged 5 years for selected combinations of lifestyle levels. BMI indicates body mass index. Table 1. Mortality Rates and Adjusted RR of Death in 7142 Men Without Recall of CHD, Stroke, or Diabetes at Screening (Deaths = 164) by Lifestyle Factors* Variables No. of Cases/Men Rate/1 Person-Years Age-Adjusted RR (95% CI) Adjusted RR (95% CI) RR Adjusted in Addition for Social Class Smoking, cigarettes per day Never 138/1726 6.2 1. 1. 1. Former 31/2456 8.3 1.35 (1.1-1.65) 1.31 (1.6-1.6) 1.3 1-19 192/165 12.5 2.7 (1.66-2.58) 1.94 (1.54-2.42) 1.89 174/784 15.5 2.63 (2.1-3.29) 2.54 (2.2-3.19) 2.49 21 247/193 17.1 3.1 (2.45-3.72) 2.71 (2.18-3.36) 2.79 Physical activity Inactive 131/589 16. 1. 1. 1. Occasional 365/2121 11.8.73 (.6-.89).79 (.64-.96).77 Light 257/1642 1.5.64 (.52-.79).69 (.56-.86).67 Moderate 146/1133 9..53 (.42-.68).64 (.5-.81).62 Moderately vigorous 16/166 7.8.48 (.37-.62).63 (.48-.82).63 Vigorous 44/496 6.9.41 (.29-.58).54 (.38-.77).55 Alcohol intake None 81/399 13.2 1.22 (.95-1.57) 1.19 (.92-1.52) 1.15 Occasional 259/1699 1.8 1. 1. 1. Light 296/235 8.5.76 (.64-.9).84 (.71-.99).84 Moderate 35/1912 11.6 1.4 (.88-1.23).98 (.83-1.16).94 Heavy 122/775 12. 1.11 (.89-1.37).9 (.72-1.13).87 BMI 64/247 18.1 1.55 (1.13-2.11) 1. (1.2-1.92) 1.35 21.9 11/664 11.3 1. 1. 1. 22 23.9 18/1452 9.1.8 (.63-1.2).89 (.69-1.14).91 24 25.9 266/1929 9.9.87 (.69-1.9).99 (.78-1.25) 1.2 26 27.9 222/153 9.8.9 (.71-1.14) 1.2 (.8-1.29) 1.7 28 29.9 125/788 1.9.97 (.75-1.26) 1.9 (.84-1.43) 1.11 3 15/555 13.3 1.29 (.98-1.69) 1.42 (1.7-1.87) 1.53 *CHD indicates coronary heart disease; RR, relative risk; CI, confidence interval; and BMI, body mass index. Adjusted for age and each of the other factors. come. As an alternative means of expressing the impact of these lifestyle factors and to illustrate the combined influence of these factors, we have used Cox predictive survival models to estimate the probability of surviving for 15 years free of major cardiovascular events and diabetes mellitus based on a model that includes age, smoking, BMI, and physical activity. To illustrate the survival probability as predicted by these risk factors, we have presented selected combinations of BMI levels and physical activity for never, former, and current smokers for a man aged 5 years (Figure 1). The probability of surviving 15 years free of major cardiovascular events and diabetes in a man aged 5 years ranged from 89% in an active man with 2436

Table 2. Combined End Point Rates and Adjusted RR (95% CI) in Men Without Recall of CHD, Stroke, or Diabetes at Screening by Lifestyle Factors* Variables Cases Rate/1 Person-Years Age Adjusted RR (95% CI) RR (95% CI) RR Adjusted in Addition for Social Class Smoking, cigarettes per day Never 259 11.6 1. 1. 1. Former 548 15.6 1.34 (1.15-1.55) 1.33 (1.15-1.55) 1.31 1-19 312 21.6 1.89 (1.6-2.22) 1.92 (1.62-2.27) 1.87 26 25. 2.19 (1.84-2.6) 2.25 (1.88-2.68) 2. 21 382 28.3 2.52 (2.15-2.95) 2.5 (2.12-2.94) 2.43 Physical activity Inactive 7 27.6 1. 1. 1. Occasional 59.5.74 (.63-.87).77 (.66-.91).77 Light 423 18.5.66 (.56-.78).71 (.6-.84).71 Moderate 232 14.8.53 (.44-.64).6 (.5-.72).61 Moderately vigorous 199 14.2.53 (.44-.65).67 (.55-.82).68 Vigorous 89 14.1.5 (.39-.65).64 (.5-.83).65 Alcohol None 133 23.9 1.25 (1.3-1.52) 1.24 (1.2-1.52) 1.25 Occasional 433 19.2 1. 1. 1. Light 58 15.4.8 (.7-.9).84 (.74-.96).85 Moderate 484 19.2 1. (.87-1.13).9 (.79-1.3).91 Heavy 5 21.2 1.1 (.93-1.3).89 (.75-1.5).9 BMI 79 23.5 1.38 (1.5-1.82) 1.25 (.95-1.64) 1.24 21.9 146 16.8 1. 1. 1. 22 23.9 291 15..89 (.73-1.8).99 (.81-1.21).98 24 25.9 425 16.4.97 (.8-1.17) 1.9 (.9-1.32) 1.8 26 27.9 387 18.5 1.11 (.92-1.34) 1.28 (1.6-1.56) 1.27 28 29.9 225 21.1 1.25 (1.1-1.54) 1.41 (1.14-1.74) 1. 3 212 29.9 1.82 (1.47-2.24) 2.11 (1.71-2.62) 2.9 *RR indicates relative risks; CI, confidence intervals; CHD, coronary heart disease; and BMI, body mass index. Adjusted for age and each of the other factors. Survival Free, % 1 9 8 7 6 5-44 y 45-49 y 5-54 y 55-59 y 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 No. of Years of Follow-up Figure 2. Fifteen-year survival free of coronary heart disease, stroke, and diabetes by 4 age groups. BMI levels of to 24. kg/m 2 who had never smoked to 42% in an inactive smoker with BMI levels of 3 kg/m 2 or higher. For a 5-year-old former smoker who is active and had BMI levels within the range of to 24 kg/m 2, the predicted probability was nearly 86%. COMMENT Coronary heart disease, stroke, and non insulindependent diabetes mellitus are major life-threatening diseases occurring in the middle-aged and elderly populations and are associated with considerable disability and premature mortality. Although extensive work has been undertaken to examine the association between lifestyle factors and cardiovascular disease, diabetes, and all-cause mortality, previous studies have tended to focus on specific end points (CHD, diabetes, stroke, or all-cause mortality) and relationships have generally been expressed as relative risks (odds). Cigarette smoking, obesity, and lack of physical activity are all wellestablished risk factors for CHD and all-cause mortality. Cigarette smoking is a recognized risk factor for stroke 19 and has also been implicated as a risk factor for diabetes. Physical activity has also been associated with reduced risk of stroke 21,22 and diabetes, 23,24 and obesity is a recognized risk factor for diabetes. 24 While several studies 3-5 have estimated life expectancy associated with smoking, estimates of survival free of multiple major end points have seldom been assessed for modifiable lifestyle factors. Of 4 lifestyle factors examined, cigarette smoking, physical activity, and BMI are important modifiable factors that greatly influence survival free of these major end points in this study. Although light or moderate drinking is generally regarded as pro- 2437

Age-Adjusted Rate1-Person/Years 36 32 28 24 16 12 8 4 36 32 28 24 16 12 8 4 Current Former Never < 24 26 28 3 Inactive Occasional Light Active None Occasional Light Moderate Heavy BMI Groups Physical Activity Alcohol Intake Figure 3. Lifestyle and age-adjusted rate per 1 person-years for the combined end point by current smoking status. BMI indicates body mass index. 36 32 28 24 16 12 8 4 Table 3. Combined End Point: Adjusted* RR (95% CI) in Men Without Recall of CHD, Stroke, or Diabetes at Screening by Smoking Status at Screening Variables Never (255/177) RR (95% CI) Former (541/2431) RR (95% CI) Smoker (937/289) RR (95% CI) Physical activity Inactive 88 1. 181 1. 318 1. Occasional 433 1.19 (.73-1.95) 733.82 (.61-1.1) 948.69 (.56-.84) Light 383 1.12 (.68-1.83) 533.62 (.45-.85) 724.69 (.56-.86) Active 82.78 (.45-1.32) 1183.6 (.44-.81) 93.64 (.52-.78) Alcohol intake None 131 1.43 (.94-2.23) 115 1.22 (.86-1.75) 148 1.19 (.87-1.54) Occasional 468 1. 537 1. 672 1. Light 687 1. (.73-1.38) 874.7 (.56-.88) 756.88 (.74-1.9) Moderate 332 1.1 (.7-1.46) 669.7 (.56-.89) 878 1. (.84-1.) Heavy 88 1.19 (.71-2.1) 235.69 (.5-.95) 439.99 (.8-1.23) BMI 32 1.46 (.59-3.71) 1.63 (.87-3.4) 17 1.19 (.9-1.57) 21.9 521 1. 591 1. 979 1. 24 25.9 47 1.23 (.84-1.78) 686 1.1 (.85-1.43) 734 1.6 (.89-1.26) 26 27.9 36 1.63 (1.13-2.34) 62 1.22 (.94-1.59) 529 1.26 (1.4-1.52) 28 29.9 19 1.92 (1.26-2.92) 288 1.54 (1.14-2.7) 32 1.23 (.98-1.53) 3 141 2.71 (1.78-4.11) 223 2.37 (1.77-3.17) 179 1.88 (1.47-2.) *Adjusted for age and each of the other factors. Data on all covariates available in 728 men. RR indicates relative risk; CI, confidence interval; CHD, coronary heart disease; and BMI, body mass index. tective to some degree against CHD 25 and diabetes, 26 its influence on stroke 27,28 and all-cause mortality 29 is less certain. Using the combined end point, light drinking was associated with a small but significant reduction in risk compared with occasional drinking, but the benefit was small compared with the effects of other lifestyle factors and was apparently confined to former smokers. Occasional drinkers have been used as the base group, as nondrinkers comprise a variable mixture of lifelong teetotalers and former drinkers and have been shown to be unsuitable as a baseline group in studies on the effects of alcohol on health. 1-12 Both lifelong teetotalers and former drinkers have characteristics that effect morbidity and mortality outcomes. Occasional drinkers are a large and relatively stable group and constitute a more appropriate baseline for assessing the effects of alcohol. ESTIMATES OF HEALTHY SURVIVAL In this study, a 5-year-old man has on average an estimated 75.3% probability of surviving to 65 years without developing CHD, stroke, or diabetes. Although age is a critical factor in determining this probability, it is also greatly influenced by lifestyle factors. In a 5-year-old man who has never smoked, is physically active, and has BMI levels within the acceptable range (-24. kg/m 2 ), this probability increases to nearly 9%. At the other extreme, a 5-year-old man who smokes, is inactive, and very obese ( 3 kg/m 2 ) has only about a 42% chance. This clearly illustrates the difference that these lifestyle factors can have on health outcome. The estimates obtained were based on measurements of lifestyle factors measured at baseline without taking into account modification of lifestyle during follow- 2438

up. Given that some individuals give up smoking during the course of follow-up and are likely to reduce their risk, this would tend to underestimate the true impact of cigarette smoking assessed at baseline alone. Similarly, inclusion of men with other chronic diseases who may suffer weight loss may inflate the risk in the low BMI group, thus underestimating the true impact of obesity and overestimating the impact of physical activity. However, we have noted that exclusion of all men with a recall of a physician diagnosis of chronic disease made relatively minor differences to the outcome. While our estimates are conservative, we have demonstrated the substantial effect that lifestyle factors have on healthy survival and we suspect that the true impact of lifestyle factors is probably greater. CHOOSING AN END POINT We have recently demonstrated that mortality data alone are insufficient to assess the impact of obesity on health. 6 On the basis of all-cause mortality, there is little relationship between BMI and outcome within the broad range of a BMI of to 27 kg/m 2. However, when a wider spectrum of end points is used, the risk of surviving free of cardiovascular disease or diabetes is increased with a BMI from 24 kg/m 2 or higher. If all-cause mortality is used as an outcome instead of the combined end point, the corresponding probability of surviving 15 years ranged from 94.1% to 68.8%, a far less impressive range than when the wider spectrum of end points is used. We have focused only on major cardiovascular events and diabetes. There is also general concern regarding other causes of morbidity but smoking, obesity, and physical activity are generally related positively (for physical activity inversely) to other diseases, eg, osteoarthritis or chronic obstructive airways disease. This is in contrast to alcohol intake that although apparently beneficial for CHD has other adverse consequences. 3 Thus, in terms of overall health outcome the difference that smoking, physical activity, and BMI makes to health outcome would be even greater if other diseases were included in the outcome. HEALTHY LIFE EXPECTANCY Healthy life expectancy has been promoted as an index for monitoring the health status of populations 31 and our findings support suggestions that major gains in healthy life expectancy may be expected from reductions in diseases associated with modifiable lifestyle factors. 32 MODIFICATION OF LIFESTYLE IN LATER LIFE: EVIDENCE FROM OBSERVATIONAL STUDIES This study has not examined the effects of modification of lifestyle in later life during the follow-up period, but previous reports 12,33-35 on this cohort of men have commented on the effects of changes in cigarette smoking, BMI, and alcohol intake on outcomes such as major CHD events and all-cause mortality. There is overwhelming evidence that cessation of smoking is associated with a significant reduction in subsequent risk of all-cause mortality, CHD, and stroke although the risk may not revert to that of never smokers. 4,34-36 The benefits of weight reduction and increased physical activity are less certain. Most national health authorities recommend weight reduction in the overweight and obese individuals as a measure toward the reduction of cardiovascular disease, diabetes, and other weight-related disorders. However, the evidence that weight loss per se decreases the incidence of cardiovascular events is much less certain than the known effects of reducing cardiovascular risk factors such as blood pressure and blood lipids. 37 However, weight loss is often associated with underlying illness that could confound the relationships seen. 33 Few studies concerned with weight loss have information on the proportion of subjects losing weight who do so intentionally. Indeed, the only study 38 to report associations between intentional weight loss and outcome appears to be the American Cancer Prevention Study I. In this study, of more than 44 overweight never-smoking American white women aged to 64 years, modest intentional weight loss among women with obesity-related conditions was associated with substantial reduction in mortality. In overweight women with no preexisting illness, the association between weight loss and benefit was equivocal. Few studies 22,39, have examined the effects of increasing physical activity in later life and disease outcome, but the evidence suggests that taking up physical activity even in later life confers benefit on disease outcome. The clear benefits of leading a healthy lifestyle shown in this study emphazise the importance of finding effective means of achieving such behavioral change more widely and before the onset of disease. CONCLUSIONS Cigarette smoking, overweight, and physical inactivity singly or in combination constitute major hazards to survival free of cardiovascular disease and diabetes. It seems reasonable to assume that in the absence of these adverse lifestyles, survival free of these and a wide range of other disorders is far more likely. Involvement in healthy lifestyles from childhood and early adulthood are likely to lead to the maintenance of such behavior into later adult life with considerable reduction in the risk of heart attacks, stroke, and diabetes mellitus. There is strong evidence that cessation of smoking is associated with a reduced risk of all-cause mortality in middle-aged men, and this is clearly seen in our study of those former smokers who are lean and physically active and who have a high probability of surviving free of major cardiovascular disease and diabetes. Intentional weight reduction, particularly in those who have already manifest illness associated with overweight or obesity and increased physical activity in adulthood, appears to be beneficial. The promotion of regular, moderate levels of physical activity, the maintenance of an acceptable weight, and not smoking would greatly increase the likelihood of surviving into older age free of CHD, stroke, and diabetes mellitus. The benefits of light drinking are relatively small compared with the effects of other lifestyle factors, and we do not 2439

regard encouragement to drink in nondrinkers or occasional drinkers as a desirable public health recommendation. Our data provide estimates that can be conveyed to health professionals and the general public and should provide further encouragement for health promotion efforts directed toward middle-aged men. Accepted for publication April 2, 1998. The British Regional Heart Study is a British Heart Foundation Research Group and receives support from The Stroke Association and the Department of Health. Dr Wannamethee is a British Heart Foundation Research Fellow, London, England. Reprints: S. Goya Wannamethee, PhD, Department of Primary Care and Population Sciences, Royal Free Hospital School of Medicine, Rowland Hill St, London NW3 2PF, England. REFERENCES 1. Marmot M, Elliot P, eds. Coronary Heart Disease Epidemiology: From Aetiology to Public Health. New York, NY: Oxford University Press Inc; 1992. 2. Poulter N, Sever P, Thom S. Cardiovascular Disease: Risk Factors and Intervention. Oxford, England: Radcliffe Medical Press; 1993. 3. Doll R, Peto R, Wheatley K, Gray R, Sutherland I. Mortality in relation to smoking: years observations on male British doctors. BMJ. 1994;39:91-911. 4. Phillips AN, Wannamethee SG, Walker M, Thomson A, Davey Smith G. Life expectancy in men who have never smoked and those who have smoked continuously: 15 year follow-up of large cohort of middle-aged British men. BMJ. 1996; 313:97-98. 5. Fraser GE, Lindsted KD, Beeson WL. Effect of risk factor values on lifetime risk of and age at first coronary event. Am J Epidemiol. 1995;142:746-758. 6. Shaper AG, Wannamethee SG, Walker M. Body weight: implications for the prevention of coronary heart disease, stroke and diabetes mellitus in a cohort study of middle aged men. BMJ. 1997;314:1311-1317. 7. Shaper AG, Pocock SJ, Walker M, Cohen NM, Wale CJ, Thomson AG. British Regional Heart Study: cardiovascular risk factors in middle-aged men in 24 towns. BMJ. 1981;283:179-186. 8. Office of Population Censuses and Surveys, Social Survey Division. General Household Survey, 1978. London, England: HMSO; 198. 9. Shaper AG, Wannamethee G, Walker M. Alcohol and mortality in British men: explaining the U-shaped curve. Lancet. 1988;2:1268-1273. 1. Wannamethee G, Shaper AG. Men who do not drink: a report from the British Regional Heart Study. Int J Epidemiol. 1988;17:37-316. 11. Wannamethee SG, Shaper AG. Lifelong teetotallers, ex-drinkers and drinkers: mortality and incidence of coronary heart disease events in middle-aged British men. Int J Epidemiol. 1997;26:523-531. 12. Shaper AG, Wannamethee SG. The J-shaped curve and changes in drinking habit. In: Alcohol and Cardiovascular Disease. New York, NY: John Wiley & Sons Inc; 1998:173-192. Novartis Foundation Symposium No. 216. 13. Shaper AG, Wannamethee G, Weatherall R. Physical activity and ischaemic heart disease in middle-aged British men. Br Heart J. 1991;66:384-394. 14. Walker M, Shaper AG. Follow-up of subjects in prospective studies based in general practices. J R Coll Gen Pract. 1984;34:365-37. 15. Cox DR. Regression models and life tables (with discussion). J R Stat Soc (B). 1972;34:187-2. 16. Christensen E. Multivariate survival analysis using Cox s regression model. Hepatology. 1987;7:1346-1358. 17. SAS/STAT Software: The PHREG Procedure Version 6. Cary, NC: SAS Institute Inc; 1991. SAS Technical Report P-217. 18. Wannamethee G, Shaper AG. Changes in drinking habits in middle-aged British men. J R Coll Gen Pract. 1988;38:4-442. 19. Shinton R, Beevers G. Meta-analysis of relation between cigarette smoking and stroke. BMJ. 1989;298:789-794.. Rimm EB, Chan J, Stampfer MJ, Colditz GA, Willett WC. Prospective study of cigarette smoking, alcohol use and the risk of diabetes in men. BMJ. 1995;31: 555-559. 21. Wannamethee G, Shaper AG. Physical activity and risk of stroke in British middleaged men. BMJ. 1992;34:597-61. 22. Shinton R, Sagar G. Lifelong exercise and stroke. BMJ. 1993;37:231-234. 23. Helmrich SP, Ragland DR, Leung RW, Paffenbarger RS. Physical activity and reduced occurrence of non insulin dependent diabetes mellitus. N Engl J Med. 1991;325:147-152. 24. Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG. Prospective study of risk factors for development of non insulin dependent diabetes in middle-aged British men. BMJ. 1995;31:56-564. 25. Marmot M, Brunner E. Alcohol and cardiovascular disease: the status of the U- shaped curve. BMJ. 1991;33:565-568. 26. Holbrook TJ, Barrett-Connor E, Wingard DL. A prospective population-based study of alcohol use and non insulin dependent diabetes mellitus. Am J Epidemiol. 199;132:92-99. 27. Carmargo CA. Moderate alcohol consumption and stroke: the epidemiological evidence. Stroke. 1989;:1611-1626. 28. Wannamethee SG, Shaper AG. Patterns of alcohol intake and risk of stroke in middle-aged Brtish men. Stroke. 1996;27:133-139. 29. Shaper AG, Wannamethee G, Walker M. Alcohol and coronary heart disease: a perspective from the British Regional Heart Study. Int J Epidemiol. 1994;23: 482-494. 3. Verschuren PM, ed. Health Issues Related to Alcohol Consumption. Washington, DC: International Life Sciences Institute Press; 1993. 31. Bone M, Bebbington A, Jagger C, Morgan K, Nicolaas G. Health Expectancy and Its Uses. London, England: HMSO; 1995. 32. Robine J, Ritchie K. Healthy life expectancy: evaluation of a global indicator of change in population health. BMJ. 1991;32:457-46. 33. Walker M, Wannamethee G, Shaper AG, Whincup PH. Weight change and risk of coronary heart disease in the British Regional Heart Study. Int J Epidemiol. 1995;24:694-73. 34. Tang JL, Cook DG, Shaper AG. Giving up smoking: how rapidly does the excess risk of ischaemic heart disease disappear? J Smoking Relat Dis. 1992;3: 3-215. 35. Wannamethee SG, Shaper AG, Whincup PH, Walker M. Smoking cessation and the risk of stroke in middle-aged men. JAMA. 1995;274:155-16. 36. Surgeon General of the United States. The Health Benefits of Smoking Cessation. Washington, DC: US Dept of Health and Human Services; 199. Publication CDC 9-8416. 37. Goldstone DJ. Beneficial health effects of modest weight loss. Int J Obes. 1992; 16:397-415. 38. Williamson DF, Pamuk E, Thun M, Flanders D, Heath C. Prospective study of intentional weight loss and mortality in never-smoking overweight US white women aged -64 years. Am J Epidemiol. 1995;141:1128-1141. 39. Paffenbarger RS Jr, Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB. The association of changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J Med. 1993;328:538-545.. Wannamethee SG, Shaper AG, Walker M. Changes in physical activity, mortality, and incidence of coronary heart disease in older men. Lancet. 1998;351: 163-168. 24