- Email: [email protected]
Smoking and inequalities
Comment
Smoking and inequalities
www.thelancet.com Vol 368 July 29, 2006
focus on getting the smoking rates down in people of low status. The size of the estimate might, therefore, matter, which means it is worth pausing to consider how the current estimates are derived. Because of the scarcity of Whitehall-type cohort studies that can yield direct estimates of the contribution of smoking to social disadvantage in health, we need methods to make up the gap. The indirect method Jha and colleagues use takes lung cancer mortality of a population as a proxy measure of exposure to smoking. The method uses absolute lung cancer rates in various social strata to indicate the proportion of smoking-attributable deaths from certain other diseases. These proportions are based on the Cancer Prevention Study-II Cohort done by the American Cancer Society,9 and are assumed to apply universally. Mackenbach and colleagues10 used a variant of this indirect method to ask, in ten European countries, how much of the higher mortality associated with low education could be attributed to smoking. They pointed out that lung cancer mortality rates in non-smokers probably differ between countries and educational groups—thus violating the assumption that the American cohort data can simply be applied universally. The estimates of the contribution of smoking to socioeconomic differences in mortality must, of necessity, be crude when this indirect method is used. In the Mackenbach study, the contribution of smoking
Published Online July 14, 2006 DOI:10.1016/S01406736(06)68976-9 See Articles page 367
Photofusion
Smoking is a modern scourge. The WHO global burden of disease study suggested that, in developed countries, 26% of male deaths and 9% of female deaths could be attributed to smoking—the single most important risk factor.1 We must consider not just aggregate effects but social distribution. In many countries the likelihood of smoking is linked to socioeconomic position: lower status, more smoking. Hence, among other evils, smoking contributes to social inequalities in mortality. The size of that contribution is considerable. In today’s Lancet, Prabhat Jha and colleagues2 estimate that at least half of socioeconomic differences in mortality in men aged 35–69 years can be laid at the foot of smoking. An estimate of more than half of social differentials caused by smoking is larger than previous estimates. In the Whitehall study of British civil servants, a combination of smoking and other coronary risk factors accounted for 27% of the inverse social gradient in mortality from coronary heart disease.3 The British Regional Heart Study reached a similar conclusion: 24% of the higher all-cause mortality in manual compared with non-manual workers could be attributed to smoking.4 To put it differently: in the Whitehall II study, in which fewer than 20% of people smoked,5 there was still a clear social gradient in disease in people who had never smoked. The study by Jha and colleagues confirms previous findings that smoking is an important contributor to socioeconomic differences in mortality. Does it really matter if smoking is responsible for a quarter or a half of the excess? Well, yes, it might. Consider two positions— exaggerated to make a point. The social gradient in mortality is the result of differences in the social circumstances in which people live and work.6 Reduction of the health disadvantage of those in lower social positions will require action to improve those circumstances throughout the life course. This, indeed, was the conclusion of Acheson’s report Inequalities in Health.7 If, however, at least half of social differences in adult mortality are the result of differences in smoking, the conclusion might be to forget social conditions, neighbourhood deprivation, employment conditions, early childhood and subsequent adult disease;8 just
341
Comment
to excess death in men with less education varied from about 30% in England and Wales to under 5% in Madrid. In women, the contribution was less. The fact that smoking’s contribution to excess disease in the socially disadvantaged might be less than a half does not detract from the essential importance of Jha and colleagues’ study. If smoking were eliminated, average health would improve and socioeconomic differences in adult mortality would be less although still substantial. There are, then, two important public-health and scientific questions. How can we reduce social inequalities in smoking? And how can we take action on the other major social causes of inequalities in health? Both questions are actively being considered by the Commission on Social Determinants of Health.11
1
2
3
4
5
6 7 8 9
Michael Marmot International Institute for Society and Health, University College London, London WC1E 6BT, UK [email protected] I declare that I have no conflict of interest.
10
11
WHO. World health report 2002: reducing risks, promoting healthy life. 2002: http://www.who.int/whr/2002/en (accessed July 5, 2006). Jha P, Peto R, Zatonski W, Boreham J, Jarvis MJ, Lopez AD. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet 2006; 368: 367–70. van Rossum CTM, Shipley MJ, Van de Mheen H, Grobbee DE, Marmot MG. Employment grade differences in cause specific mortality: a 25 year follow up of civil servants from the first Whitehall study. J Epidemiol Community Health 2000; 54: 178–84. Emberson JR, Whincup PH, Morris RW, Walker M. Social class differences in coronary heart disease in middle-aged British men: implications for prevention. Int J Epidemiol 2004; 33: 289–96. Marmot MG, Smith GD, Stansfeld SA, et al. Health inequalities among British Civil Servants: the Whitehall II study. Lancet 1991; 337: 1387–93. Marmot M. Status syndrome. London: Bloomsbury, 2004. Acheson D. Inequalities in health: report of an independent inquiry. London: HM Stationery Office, 1998. Barker DJP. Mothers, babies and health in later life. Edinburgh: Churchill Livingtone, 1998. Steenland K, Henley J, Thun M. All-cause and cause-specific death rates by educational status for two million people in two American Cancer Society Cohorts, 1959–1996. Am J Epidemiol 2002; 156: 11–21. Mackenbach JP, Huisman M, Andersen O, et al. Inequalities in lung cancer mortality by the educational level in 10 European populations. Eur J Cancer 2004; 40: 126–35. Marmot M. Social determinants of health inequalities. Lancet 2005; 365: 1099–104.
Sex and anticoagulation in patients with idiopathic venous thromboembolism See Articles page 371
342
Men are thought to have a greater risk of recurrent venous thromboembolism than women. This issue of The Lancet includes a meta-analysis by Simon McRae and colleagues1 of studies on the long-term follow-up of patients with acute venous thromboembolism. This meta-analysis lends support to a significantly increased risk of recurrent venous thromboembolism in men (relative risk 1·6, 95% CI 1·2–2·0). This increased risk in men remains, even when the analysis is confined to the subgroup of patients with idiopathic venous thromboembolism (relative risk 1·5, 95% CI 1·1–1·9), which is the patients’ category still awaiting definition of the optimum duration of anticoagulation.2 McRae and colleagues’ findings are potentially valuable, because they could help identify patients with idiopathic venous thromboembolism who are at a greater risk of recurrent thromboembolism. However, interpretation of their findings requires caution. The authors were not able to retrieve information about the sex distribution of recurrent venous thromboembolism from 11 adequate studies. Moreover, a multivariate analysis accounting for important individual data
(such as age) could not be done. Finally, in all but two single studies dealing with idiopathic venous thromboembolism, there was no significant difference between men and women. On the basis of McRae and colleagues’ findings, the increase in risk of recurrent thromboembolism conferred by being male in patients with idiopathic venous thromboembolism is modest, and similar to that conferred by disorders such as the heterozygous carriage of factor V Leiden or prothrombin mutation, which are consistently reputed to play little, if any, part in determination of the risk of recurrent venous thromboembolism.3 The 95% CI around this risk (1·1–1·9) is wide, and an increased risk cannot be excluded. In a large series of 859 patients with idiopathic venous thromboembolism, consecutively enrolled at our centre and followed up for 10 years, the relative risk of recurrence of venous thromboembolism in men compared with women was 1·21 (95% CI 0·95–1·55).4 The cumulative 10-year incidence of recurrence of venous thromboembolism in women with idiopathic venous thromboembolism was about www.thelancet.com Vol 368 July 29, 2006
Smoking and inequalities
www.thelancet.com Vol 368 July 29, 2006
focus on getting the smoking rates down in people of low status. The size of the estimate might, therefore, matter, which means it is worth pausing to consider how the current estimates are derived. Because of the scarcity of Whitehall-type cohort studies that can yield direct estimates of the contribution of smoking to social disadvantage in health, we need methods to make up the gap. The indirect method Jha and colleagues use takes lung cancer mortality of a population as a proxy measure of exposure to smoking. The method uses absolute lung cancer rates in various social strata to indicate the proportion of smoking-attributable deaths from certain other diseases. These proportions are based on the Cancer Prevention Study-II Cohort done by the American Cancer Society,9 and are assumed to apply universally. Mackenbach and colleagues10 used a variant of this indirect method to ask, in ten European countries, how much of the higher mortality associated with low education could be attributed to smoking. They pointed out that lung cancer mortality rates in non-smokers probably differ between countries and educational groups—thus violating the assumption that the American cohort data can simply be applied universally. The estimates of the contribution of smoking to socioeconomic differences in mortality must, of necessity, be crude when this indirect method is used. In the Mackenbach study, the contribution of smoking
Published Online July 14, 2006 DOI:10.1016/S01406736(06)68976-9 See Articles page 367
Photofusion
Smoking is a modern scourge. The WHO global burden of disease study suggested that, in developed countries, 26% of male deaths and 9% of female deaths could be attributed to smoking—the single most important risk factor.1 We must consider not just aggregate effects but social distribution. In many countries the likelihood of smoking is linked to socioeconomic position: lower status, more smoking. Hence, among other evils, smoking contributes to social inequalities in mortality. The size of that contribution is considerable. In today’s Lancet, Prabhat Jha and colleagues2 estimate that at least half of socioeconomic differences in mortality in men aged 35–69 years can be laid at the foot of smoking. An estimate of more than half of social differentials caused by smoking is larger than previous estimates. In the Whitehall study of British civil servants, a combination of smoking and other coronary risk factors accounted for 27% of the inverse social gradient in mortality from coronary heart disease.3 The British Regional Heart Study reached a similar conclusion: 24% of the higher all-cause mortality in manual compared with non-manual workers could be attributed to smoking.4 To put it differently: in the Whitehall II study, in which fewer than 20% of people smoked,5 there was still a clear social gradient in disease in people who had never smoked. The study by Jha and colleagues confirms previous findings that smoking is an important contributor to socioeconomic differences in mortality. Does it really matter if smoking is responsible for a quarter or a half of the excess? Well, yes, it might. Consider two positions— exaggerated to make a point. The social gradient in mortality is the result of differences in the social circumstances in which people live and work.6 Reduction of the health disadvantage of those in lower social positions will require action to improve those circumstances throughout the life course. This, indeed, was the conclusion of Acheson’s report Inequalities in Health.7 If, however, at least half of social differences in adult mortality are the result of differences in smoking, the conclusion might be to forget social conditions, neighbourhood deprivation, employment conditions, early childhood and subsequent adult disease;8 just
341
Comment
to excess death in men with less education varied from about 30% in England and Wales to under 5% in Madrid. In women, the contribution was less. The fact that smoking’s contribution to excess disease in the socially disadvantaged might be less than a half does not detract from the essential importance of Jha and colleagues’ study. If smoking were eliminated, average health would improve and socioeconomic differences in adult mortality would be less although still substantial. There are, then, two important public-health and scientific questions. How can we reduce social inequalities in smoking? And how can we take action on the other major social causes of inequalities in health? Both questions are actively being considered by the Commission on Social Determinants of Health.11
1
2
3
4
5
6 7 8 9
Michael Marmot International Institute for Society and Health, University College London, London WC1E 6BT, UK [email protected] I declare that I have no conflict of interest.
10
11
WHO. World health report 2002: reducing risks, promoting healthy life. 2002: http://www.who.int/whr/2002/en (accessed July 5, 2006). Jha P, Peto R, Zatonski W, Boreham J, Jarvis MJ, Lopez AD. Social inequalities in male mortality, and in male mortality from smoking: indirect estimation from national death rates in England and Wales, Poland, and North America. Lancet 2006; 368: 367–70. van Rossum CTM, Shipley MJ, Van de Mheen H, Grobbee DE, Marmot MG. Employment grade differences in cause specific mortality: a 25 year follow up of civil servants from the first Whitehall study. J Epidemiol Community Health 2000; 54: 178–84. Emberson JR, Whincup PH, Morris RW, Walker M. Social class differences in coronary heart disease in middle-aged British men: implications for prevention. Int J Epidemiol 2004; 33: 289–96. Marmot MG, Smith GD, Stansfeld SA, et al. Health inequalities among British Civil Servants: the Whitehall II study. Lancet 1991; 337: 1387–93. Marmot M. Status syndrome. London: Bloomsbury, 2004. Acheson D. Inequalities in health: report of an independent inquiry. London: HM Stationery Office, 1998. Barker DJP. Mothers, babies and health in later life. Edinburgh: Churchill Livingtone, 1998. Steenland K, Henley J, Thun M. All-cause and cause-specific death rates by educational status for two million people in two American Cancer Society Cohorts, 1959–1996. Am J Epidemiol 2002; 156: 11–21. Mackenbach JP, Huisman M, Andersen O, et al. Inequalities in lung cancer mortality by the educational level in 10 European populations. Eur J Cancer 2004; 40: 126–35. Marmot M. Social determinants of health inequalities. Lancet 2005; 365: 1099–104.
Sex and anticoagulation in patients with idiopathic venous thromboembolism See Articles page 371
342
Men are thought to have a greater risk of recurrent venous thromboembolism than women. This issue of The Lancet includes a meta-analysis by Simon McRae and colleagues1 of studies on the long-term follow-up of patients with acute venous thromboembolism. This meta-analysis lends support to a significantly increased risk of recurrent venous thromboembolism in men (relative risk 1·6, 95% CI 1·2–2·0). This increased risk in men remains, even when the analysis is confined to the subgroup of patients with idiopathic venous thromboembolism (relative risk 1·5, 95% CI 1·1–1·9), which is the patients’ category still awaiting definition of the optimum duration of anticoagulation.2 McRae and colleagues’ findings are potentially valuable, because they could help identify patients with idiopathic venous thromboembolism who are at a greater risk of recurrent thromboembolism. However, interpretation of their findings requires caution. The authors were not able to retrieve information about the sex distribution of recurrent venous thromboembolism from 11 adequate studies. Moreover, a multivariate analysis accounting for important individual data
(such as age) could not be done. Finally, in all but two single studies dealing with idiopathic venous thromboembolism, there was no significant difference between men and women. On the basis of McRae and colleagues’ findings, the increase in risk of recurrent thromboembolism conferred by being male in patients with idiopathic venous thromboembolism is modest, and similar to that conferred by disorders such as the heterozygous carriage of factor V Leiden or prothrombin mutation, which are consistently reputed to play little, if any, part in determination of the risk of recurrent venous thromboembolism.3 The 95% CI around this risk (1·1–1·9) is wide, and an increased risk cannot be excluded. In a large series of 859 patients with idiopathic venous thromboembolism, consecutively enrolled at our centre and followed up for 10 years, the relative risk of recurrence of venous thromboembolism in men compared with women was 1·21 (95% CI 0·95–1·55).4 The cumulative 10-year incidence of recurrence of venous thromboembolism in women with idiopathic venous thromboembolism was about www.thelancet.com Vol 368 July 29, 2006