In this cohort study, we investigated the association between biological ageing and all-cause mortality among individuals with tumors, as well as identifying potential factors which may influence this association. Our study indicated a positive correlation between biological ageing and all-cause mortality among the tumor-bearing population, consistent with current research findings. Individuals with lower education levels, higher incomes, or cardiovascular diseases showed a stronger association between biological ageing acceleration and all-cause mortality in the tumor-bearing population. Furthermore, among the non-tumor population, BMI may be a significant factor influencing this association.
In clinical trials concerning tumors, given the challenges of long-term follow-up, researchers have focused on five-year survival rats or disease-free survival (DFS), comparing middle adult to older adult population segments. Using a predictive model for tumor mortality based on cancer death certifications and population data from the World Health Organization and Eurostat databases, researchers found a continuous decline in colorectal cancer mortality across all ages, with the most significant reductions observed in individuals over 70(15). According to the American Cancer Society, between 2011 and 2020, annual colorectal cancer mortality decreased by 2% among individuals, while it increased by 0.5–3% among those under 50 and in native Americans younger than 65 years(16). Contrary to common trends, tumors manifested less aggressively in the older adult population(17). Although age is recognized as a crucial factor, it does not directly equal to the state of ageing. Few researchers have precisely explored the association between biological ageing and tumor prognosis in clinical trials.
Over the past 20 years, various epigenetic clocks and biological ageing indexes have been developed. Gregory Hannum(18) was the first to develop a quantitative predictive model of ageing based on DNA methylation, using measurements and analyses of human whole blood. The DNA methylation age predictor, developed by Steve Horvath, can be widely applied to multiple tissues and cell types(19). Ake T Lu(20) constructed the lifespan predictor called DNAm GrimAge (GrimAge) and the ageing acceleration named AgeAccelGrim, which adjusts for chronological age and has been proved to correlate closely with numerous age-related diseases. Beyond large cross-sectional studies, researchers also have recognized the importance of measuring ageing through longitudinal studies. Daniel W Belsky(21) attempted to measure ageing relying on biomarker scores and longitudinal studies. The former approach calculated biological ageing from cross-sectional data, while the latter assessed calculated individual ageing acceleration through longitudinal analysis of 18 biomarkers. Additionally, Timothy V. Pyrkov(22) developed the dynamic organism state indicator (DOSI) through longitudinal analysis of blood markers. The research team observed that the broadening of the DOSI distribution could be attributed to a gradual loss of physiological resilience, and they predicted that the human maximum lifespan ranges between 120 and 150 years. In our study, we selected PhenoAge and KDMAge as ageing biomarkers. First of all, both of them consider the influence of chronological age and describe the ageing status through a single variable. Secondly, compared to some biomarkers like DOSI, which require more time and contend with more recessive variables, PhenoAge and KDMAge are more accessible. Thirdly, PhenoAge also incorporates complete blood counts (CBC), a process akin to the DOSI procedure but with distinct features. Lastly, PhenoAge acceleration has been proved to be a useful predictor of multiple clinical risks, whereas DOSI still requires further validation through clinical and experimental studies. Our results indicated a positive correction between PhenoAge and KDMAge acceleration and all-cause mortality among individuals with tumors. Compared to KDMAge, PhenoAge demonstrates greater sensitivity in evaluating the relationship between ageing and tumor mortality.
In the tumor population, educational level and PIR may be potential factors affecting the association of biological ageing with all-cause mortality. A study on breast cancer revealed that, compared to women with low educational levels, college graduates exhibited the highest survival rates following diagnosis(23). Several researchers have argued that self-reported educational levels do not accurately measure health literacy; instead, they utilize the National Adult Literacy Survey (NALS) to further assess communication and understanding skills(24). Both low education levels and limited health literacy may adversely affect tumor prognosis by influencing treatment. As expected, patients’ income also affects diagnostic and treatment intervals(25). It is widely accepted that longer treatment interval correlates with poorer cancer outcomes including advanced stage at diagnosis and higher mortality rates(26, 27). Our study showed that higher income strengthens the association between biological ageing and mortality risk. High-income populations benefit from superior medical resources, allowing for timely detection of tumor lesions; furthermore, they are more likely to opt for aggressive treatment options due to greater financial affordability.
Besides education level and PIR, our study identified that cardiovascular disease intensifies the association of biological ageing with all-cause mortality in the tumor population; notably, cardiovascular disease is the second leading cause of death in tumor survivors(28). In vitro studies have demonstrated that the tumor antagonism mechanism of telomere attrition contributes to ageing by activating the DNA damage response(29). Age-related senescent cells (SNCs) accumulate in the cardiovascular system and ultimately promote cardiovascular disease in a bimodal manner(30). Moreover, previous study has observed that tumor patients face a higher risk of cardiovascular mortality compared to the general population, with this risk inversely correlated with the age at diagnosis(31). This may be attributed to the presence of severe cardiovascular diseases in patients who die prematurely. Following a definitive tumor diagnosis, treatments such as radiation therapy(32), chemotherapy(28) and immunotherapy(33) pose potential cardiotoxic risks; however, older patients tend to be more conservative in selecting treatment options.
In the non-tumor population, BMI may be a significant factor affecting the association between biological ageing and death risk. According to our results, the higher the degree of overweight, the stronger the association between biological ageing and mortality risk. This association may be due to the influence of BMI on ageing via insulin resistance (IR)(34). Research has shown that obesity can increase cellular senescence, which in turn promotes IR; subsequently, IR may further promote cellular senescence in human adipocytes and hepatocytes(35).
The strengths of this study are as follows. Firstly, our study benefits from a large, nationally representative sample size. Additionally, we compared tumor and non-tumor populations to investigate the robustness of the association of biological ageing with all-cause mortality in different subgroups. However, our study also has several limitations. Firstly, as the diagnoses were self-reported by participants, the results may lack complete reliability. Secondly, because biological ageing indicators were only measured at baseline, we could not capture their dynamic changes. Thirdly, although several covariates were controlled, the influence of potential confounders cannot be entirely ruled out. Finally, due to the cohort nature of this study, we cannot establish direct causal relationships between biological ageing and all-cause mortality across populations.