Age-Stratified Competing Risk Analysis of Smoking, Metabolic Abnormalities, and Socioeconomic Factors on Cancer Mortality: Evidence from NHANES 2005-2018

Background: Cancer remains a major global public health concern, with increasing mortality rates especially in aging populations. While cigarette smoking is a well-established carcinogen, its impact on cancer mortality may vary by age and metabolic status. In addition, metabolic disorders—such as obesity, hypertension, hyperglycemia, and dyslipidemia—are increasingly implicated in cancer progression. Socioeconomic disparities further compound these risks by limiting access to screening and timely treatment. However, few studies have comprehensively examined the combined and age-stratified effects of smoking, metabolic abnormalities, and socioeconomic status on cancer-specific mortality within a large, population-based cohort. Methods: We conducted a retrospective cohort study using data from the National Health and Nutrition Examination Survey (NHANES) 2005–2018, with mortality follow-up through December 31, 2019. A total of 38,785 adults aged 20 years or older were included. Primary outcomes included cancer-specific mortality, competing mortality (non-cancer causes), and survival. Cox proportional hazards and Fine–Gray competing risk models were employed to estimate associations between smoking, metabolic factors (e.g., diabetes, blood pressure, dyslipidemia), and socioeconomic indicators (e.g., income, education, employment) with mortality risks. Participants were stratified into three age groups (20–50, 50–80, ≥80 years) to assess effect modification. We also evaluated dose–response relationships using pack-years of smoking exposure. Results: Among the cohort, 7.8% died from cancer, 2.5% from other causes, and 87.4% were alive at the end of follow-up. Age emerged as the strongest determinant of mortality risk, with individuals aged ≥80 years having a 20.72-fold higher risk of death than those under 50. Metabolic abnormalities significantly contributed to increased mortality risk. For instance, participants with diabetes and hypertension had higher mortality, while those with normal blood pressure had a 27% lower risk (HR = 0.73). Socioeconomic disadvantage—reflected in low income, unemployment, and limited education—was strongly associated with higher mortality risk; unemployed individuals had a 2.82-fold increased risk of death. Smoking demonstrated a clear dose–response effect. Current smokers had significantly higher mortality risk compared to never-smokers (HR = 1.51), and this association was more pronounced in older age groups. Age-stratified competing risk models revealed that never smokers had the lowest cancer mortality risk across all age groups. The greatest protective effect of smoking cessation was observed among younger adults (HR = 0.373 for former vs. current smokers). Cumulative incidence function (CIF) analysis indicated that cancer mortality risk accelerated sharply for current smokers after age 50, peaking at 20% in the ≥80 group. Dose–response analysis with pack-years showed a steep increase in risk beyond 200 pack-years in the 50–80 age group, while a paradoxical decline was seen in the ≥80 group, possibly due to survivor bias. Conclusion: This study underscores the multifactorial nature of cancer mortality and highlights strong age-dependent heterogeneity in risk. Smoking, metabolic dysfunction, and socioeconomic disadvantage independently and synergistically contribute to cancer and all-cause mortality. Early-life interventions targeting smoking cessation and metabolic control, along with policies aimed at reducing social inequities, are critical to reducing the burden of cancer mortality. Personalized, age-stratified prevention strategies are warranted, especially for middle-aged and socioeconomically vulnerable populations. The use of competing risk models offers a more accurate understanding of mortality dynamics and supports data-driven public health planning.