An analysis conducted in the United States reveals that the apparent increase in mortality among smokers over the decades is largely explained by the aging of the smoking population. Smoking rates have decreased among younger individuals, and smokers are on average older than they were several decades ago, leading to higher mortality rates, which increase with age. These results do not challenge the health risks associated with smoking. Despite prevention and awareness efforts, smoking remains one of the leading causes of preventable death worldwide. Concerns have been raised in recent years regarding the apparent increase in the association between smoking and mortality over the past decades. Large-scale studies, such as the Cancer Prevention Studies in the United States or the British Doctors Study in the United Kingdom, have shown that during the 20th century, the association between smoking and mortality increased significantly. More recent data in the United States suggest that similar trends continued from the 1980s to 2006. Various hypotheses have been proposed to explain this increase, including increased smoking intensity. Another hypothesis suggests that the harmful effects of tobacco may have intensified due to changes in the chemical composition of cigarette smoke. It has also been suggested that the physiological characteristics of smokers might have changed over time. Smokers may be increasingly likely to live in areas with other health risks or adopt other health-harming behaviors. On the other hand, reports indicate that socioeconomic disparities in smoking have increased: individuals without a high school diploma represent an increasing proportion of smokers. These hypotheses suggest that, aside from potential physiological and chemical differences, the increase in the association between smoking and mortality could also be correlated with a wide range of underlying factors. However, these hypotheses have rarely been tested. In a detailed study published in the journal PNAS Nexus, researchers provide new insights into the reasons for this trend in the United States. The authors highlight that this analysis could have methodological implications for the study of other risk factors, particularly in countries where the population composition is changing. National data indicate that in the United States, the difference in five-year mortality risk between current smokers and people who have never smoked was 3.4% in 1987. This figure increased to 4.9% in 2013. However, paradoxically, the prevalence of smoking has generally decreased, dropping notably from 52% among men and 34% among women in 1965 to 13% among men and 10% among women in 2022, according to the new study. To understand this apparently paradoxical development, the team analyzed a large medical database among Americans between 1987 and 2013, as part of the National Health Interview Survey (NHIS). The researchers divided the NHIS data into three periods and estimated mortality by age for each of them. They found that the difference in death risk between smokers and non-smokers remains roughly the same at a given age over time. The authors concluded that the change in age structure explains largely the increase in the statistical association between smoking and mortality. This would be explained by the fact that the American population is aging and, at the same time, smoking rates have decreased among younger cohorts, probably because smoking is less socially accepted. Thus, current smokers are, on average, older than they were a few decades ago, and the mortality risks associated with smoking increase with age. "These results illustrate how changes in the composition of the population can influence estimates of mortality risks related to exposures such as smoking and have implications for the understanding of temporal variations in other risk factors such as obesity," conclude the authors.