What would happen to health inequalities if smoking were eliminated?

Socioeconomic inequalities in mortality are a stubborn phenomenon. During the 20th century, despite increases in prosperity and various policies to take the sharper edges off social inequality, the gap in survival between those with a lower and a higher socioeconomic position has not disappeared. On the contrary, many countries in western Europe and North America have seen a widening of this gap during the past 40 years.1

In many of these countries, smoking now is one of the most important mediators of the effect of low socioeconomic position on mortality. Among men, and in some countries among women, smoking has been declining for several decades. Because this decline has been most prominent in more highly educated people and those with higher incomes, smoking has become relatively more common in lower socioeconomic groups.2 3 The linked cohort study by Hart and colleagues (doi:10.1136/bmj.d3785) is a welcome opportunity to see what could happen to health inequalities if smoking were eliminated.4

It is not unusual to find that smoking alone accounts for more than a quarter of inequalities in mortality, depending on the population studied and the methods used, especially in men.5 6 Smoking has therefore become one of the main targets of policies to reduce health inequalities, particularly in the United Kingdom, where smoking is strongly socially patterned.7

But what would happen to health inequalities if smoking were eliminated and could no longer mediate the effect of a low socioeconomic position on mortality? Smoking is not the fundamental cause of health inequalities, and if underlying inequalities in access to material and immaterial resources remain unchanged, other risk factors may replace smoking as a mediator, so that the effect on health inequalities may in the end be small.8

Hart and colleagues studied the survival of 3613 women who, when this study in the west of Scotland started in 1972, had never smoked. They found that the death rate in women in lower social classes was a third higher than for those in higher social classes (social class IIIM, relative risk 1.35, 95% confidence interval 1.16 to 1.57; social class IV and V, 1.34, 1.17 to 1.55; both relative to social classes I and II) and that this was partly mediated by their higher body mass index.4

A previous study by the same authors found that although mortality in never smokers in this cohort is much lower than that in smokers, substantial inequalities in survival are present among never smokers and current smokers. Surprisingly, in this cohort inequalities in mortality are larger among never smokers. For example, in women who had never smoked, mean 28 year survival ranged from 56% in the lowest occupational classes to 65% in the highest (9% points difference). The corresponding figures for women who currently smoked were 41% and 35%, respectively (6% points difference).9

The current study quantifies the contribution of a risk factor that seems to have replaced smoking as a mediator in women who never smoked—overweight and obesity. As the authors note, their results show an interesting similarity with recent findings in southern European women. In this group the smoking epidemic is in its early stages, with smoking still being uncommon among middle aged and elderly women, and more prevalent in higher socioeconomic groups. Overweight is currently more prevalent in poorly educated southern European women in the age range in which most deaths occur.10

In women in the west of Scotland cohort who had never smoked, overweight and obesity explain about a quarter of the inequalities in mortality.4 In the case of overweight and smoking, it could be argued that one can replace the other as a mediator of the effect of low socioeconomic status on mortality. Overweight is more common among never smokers, perhaps because of the effects of smoking on appetite and resting metabolic rate,11 and the higher prevalence of overweight and smoking in lower socioeconomic groups may partly result from higher exposure to psychosocial stressors.12

But it is important not to forget that smoking is a much stronger risk factor for mortality than most other risk factors, including obesity. In this cohort study, current cigarette smoking doubled the risk of (all cause) mortality, whereas moderate obesity increased the risk of mortality by only a third.4 9 Exchanging smoking for obesity is a good bargain, but inequalities in mortality will not necessarily become smaller. Inequalities in mortality persist among those who have never smoked, partly because obesity takes over the role of smoking, but they persist at a much lower level, and that is good news for whoever wants to reduce health inequalities.