Despite annualized rates of decline in smoking prevalence, smoking remains the second leading risk factor for early death and disability worldwide, contributing to 11.5% of global deaths (6.4 million), of which more than half occur in four countries (1).

Evidence from several extensive contemporary studies illustrates sex and age differences in the health outcomes of tobacco exposure, suggesting that the comprehensive risk of smoking is considerably more significant in women than in men, and in young adults than in older adults (2–8). For example, in 2014, a cohort study with an extended follow-up indicated that the full hazards of light smoking (1–14 cigarettes per day) at baseline were more significant for women than for men, while the effect decreased with increasing extent of smoking, suggesting a dose–response relationship (9). The risks of mortality and health outcomes vary according to the age at smoking initiation. A prospective cohort study including a contemporary US population showed that delayed smoking initiation reduced the mortality risk of cancer (7).

A limitation in these studies is that the quantification of the extent of smoking was subjective and self-reported, including the pack-year index and daily cigarette consumption. Self-reports are frequently affected by memory and emotion-expression biases (10). Meanwhile, the adverse health impact of actual exposure to smoke toxicants is determined by multiple factors, including the number of smoked cigarettes, smoking years, cigarette brands, cigarette emissions, and even smoking topography (11–14). Therefore, cigarette consumption may not fully reflect the extent of the harmful effects of smoking. Moreover, it is complicated to quantify the extent of passive smoking by the frequency of smoking and the number of cigarettes consumed. The methods of tobacco exposure quantification may have resulted in inaccuracies in these previous studies regarding the sex and age differences in smoking. Hence, an objectively quantifiable indicator is required to evaluate these differences in smoking.

Cotinine is the most predominant metabolite of nicotine, which is similarly a tobacco-specific biomarker (15–17). Serum cotinine levels measured at a single time point are positively associated with the degree of tobacco smoke exposure, considered one of the most reliable indicators of both smoking intensity and exposure to environmental tobacco smoke, which includes first-hand, second-hand, and third-hand smoke exposure (18, 19). Meanwhile, cotinine-assessed smoking status is generally more accurate than self-reported smoking status (20). Smoking is associated with mortality from various diseases, including cardiovascular diseases and cancer (1). A National Health and Nutrition Examination Survey (NHANES) of 20,175 self-reported non-smokers aged ≥20 years revealed that serum cotinine levels were significantly associated with death from lung cancer, all cancers, and heart diseases (21).

These findings support the assertion that cotinine is an effective indicator to evaluate smoking-related mortality risk. Another related question arises from these results, namely, whether there are thresholds and differences among different diseases in sex-and age-related differences in smoking. In particular, there is a need for long-term prospective studies to completely confirm these differences by measuring serum cotinine with standard statistical methods. Therefore, this study aimed to investigate sex-and age-related differences in smoking-related mortality by measuring serum cotinine levels, emphasizing subgroup analyses and possible nonlinear relationships and threshold saturation effects.

This study explored the influence of sex and age on the impact of cigarette exposure on all-cause, cardiovascular disease, and cancer mortality. To the best of our knowledge, no previous study has explored sex-and age-related differences in the association between cigarette exposure and mortality by measuring serum cotinine levels. In summary, our significant findings were that females and younger adults had higher risks of all-cause cardiovascular disease and cancer mortality associated with cigarette exposure. Sex differences were significant only in all-cause and cancer mortality. Threshold saturation effects were observed in all-cause mortality among both males and females, cancer mortality among females, and cardiovascular disease mortality among males. Age differences were significant only in all-cause and cardiovascular disease mortality. Threshold saturation effects were found in cardiovascular disease mortality among younger adults and cancer mortality among all-age population. Furthermore, the nomogram and ROC curves demonstrated the accuracy and reliability of the established models.

Inconsistent with the decreasing trend in overall smoking prevalence worldwide, smoking prevalence in more places changed minimally or increased among women (22). A major finding of this study was that younger adults demonstrated a higher proportion of all-cause, cardiovascular disease, and cancer mortality associated with cigarette exposure than older adults. These age differences are consistent with those reported in previous studies. Hiroyasu et al. demonstrated that the relative risk of smoking-induced cardiovascular disease mortality in Japanese women was greater in persons aged 40–64 years than in older persons (23). Furthermore, many studies have found that early smoking-onset age indicates a high risk of smoking-related mortality. A retrospective study that included 4,499 current or former smokers in a median 7.02 years of follow-up suggested that each year of delay in early smoking-onset age was inversely associated with cardiovascular mortality in the group aged ≤12 years and 4% lower risk of all-cause mortality in all groups (8, 24). However, our study revealed more comprehensive findings that the relationship between serum cotinine and cardiovascular disease in younger adults, and cancer mortality in all-age population showed threshold saturation effects. Additionally, the positive correlation was only significant in cancer mortality among older adults. For cardiovascular diseases mortality, the risk in younger adults was increased with increasing serum cotinine, while when serum cotinine increased to 306 ng/mL, this dose–response trend was no longer significant. There were, however, no dose–response trends in older adults. For cancer mortality, the risk was increased with increasing serum cotinine in all ages and the increased risk was early weakened in older adults after serum cotinine increased beyond the threshold, compared to younger adults. This implies that there is a dose-effect relationship in smoking-related carcinogenesis among all ages, while smoking-related atherosclerosis does not present a dose-effect relationship in older adults. Our results suggested that a delay in early smoking-onset age reduced the risk of smoking-induced cardiovascular disease mortality but did not significantly decrease the risk of smoking-induced cancer mortality.

Females have greater health consequences associated with tobacco exposure. For example, earlier large-scale cohort studies in the Asia-Pacific region reported that women presented a two-fold higher risk of lung cancer mortality associated with current smoking in Australia and New Zealand than men (2). Similarly, increased all-cause mortality in females from myocardial infarction was previously reported (5). Hurley found that for women, the all-cause mortality risk induced by light smoking may be approximately three times that for men, whereas no such relationship was observed for medium/heavy smoking (9), which was consistent with our study. We found that females had a higher risk of all-cause and cancer mortality than males before serum cotinine increased to the threshold. However, for cardiovascular diseases mortality, the risk was increased simultaneously with increasing serum cotinine in males and females and the increased risk was early weakened in males after serum cotinine rose beyond the threshold. This indicates that mild and moderate tobacco exposure resulted in a higher risk of all-cause and cancer mortality in females and heavy tobacco exposure brought little risk of mortality in males and females. Interestingly, the harm of cardiovascular diseases from mild and moderate tobacco exposure in females and males is approximately equal.

One of the underlying mechanisms of the sex difference may be that due to the anti-estrogenic effect of smoking (25). Women smokers suffer from the twice “struck” phenomena: smoking directly increases the risk of ischemic heart disease by inducing atherosclerosis in both women and men, and in women, smoking may substantially attenuate the protective role of estrogen in atherosclerosis, and gastric and colon cancers (26–28). Another underlying mechanism for their higher risk of all-cause mortality is because women are more vulnerable to chronic obstructive pulmonary disease (29). Additionally, smoking cessation is more difficult in females than males (30, 31). Females experience more weight gain related to smoking cessation worries and suffer severe side effects from smoking cessation medications and craving, leading to fewer antismoking medications (32–35). Cepeda-Benito et al. reported that nicotine replacement therapy is less effective in females than males (35). Difficulty in quitting smoking increases the lasting detrimental effects of smoking in women who attempt to quit.

The age difference effect is relatively easy to understand. A prospective study reported that younger smokers have a greater mortality risk from cancer than those who start later (7). Similar results have been reported for coronary heart disease (36). The initiation of premature smoking is typically associated with more difficult cessation. This study suggests that younger adults are at a higher risk of smoking-related mortality. To interpreting this result, possibility of a cumulative effect need be considered; younger adults with the same serum cotinine level as the older adults may have a longer duration of smoking exposure (37).

In the present study, dose–response relationships and saturation effects of tobacco exposure showed sex and age differences, while sex differences in cardiovascular disease and age differences in cancer mortality were not significant; these results were not found in the previous study. These results suggest that and sex and age differences and corresponding saturation effects in smoking-related mortality are inconsistent in different diseases. Finally, this study included several laboratory parameters and demographic data. Most underlying diseases from the more representative and multiracial samples of the US population were excluded, which improved the reliability and accuracy of the study. The major study limitations were the long-term prognostic outcome, and a single point detection of serum cotinine concentration may not completely reflect the degree of smoking exposure during the follow-up period.

Taken together, our study demonstrated sex and age differences in smoking risk by measuring serum cotinine levels. The findings of dose–response relationships and threshold saturation effects provide evidence that mild and moderate tobacco exposure brought a higher risk of all-cause and cancer mortality in females. It has been noted that sex differences in cardiovascular disease and age differences in cancer mortality were insignificant, indicating inconsistency in different diseases. Cotinine could be an effective indicator to reveal novel findings for sex and age differences in smoking risk.