Human lifespan and sex-specific patterns of resilience to disease: a retrospective population-wide cohort study

Aging is the main risk factor for developing highly prevalent diseases including cancer, cardiovascular diseases, diabetes mellitus, neurodegenerative diseases, and chronic respiratory diseases, among others [1]. By 2050, the number of people aged 65 or older in Europe and the USA will have increased from 20 to 25% [2], and two theories have been put forward to address the effect of this increase in lifespan on the prevalence of diseases and disability of the population. The “compression of morbidity” theory claims that the onset of morbidity is postponed with the increase in lifespan, resulting in an extended health span [3]. In contrast, the “expansion of morbidity” theory suggests that health span is not increased, resulting in additional years of life accompanied by morbidity and disability [4]. However, to discern the interplay among lifespan, health span and morbidity, further insight, at the individual level, of the relationship between lifespan and the age of onset of the diseases and the ability to avoid them is needed. Apart from these two opposing theories, a more recent individual-centered theory suggests that there is a continuum of individual paces of aging (from accelerated to deaccelerated) that results in differing onset of symptoms and diseases depending on individual vulnerability [5].

The aging process affects all organs and tissues, leading to changes like altered pancreatic function [6], shifts in adipose tissue distribution [7], loss of cognitive and motor abilities in the nervous system [8], circulatory system issues [9], reduced respiratory capacity [10], liver volume and blood flow decline [11], diminished nutrient absorption in the gastrointestinal tract [12], skeletal muscle mass and functionality reduction [13], and kidney structural alterations [14]. It has previously been reported that the inheritable factor of human lifespan could explain from 20 to 40% of the variability (with the remaining being related to environmental traits) and that extreme lifespan clusters in specific families [15], leaving room for modulating external factors to increase lifespan. Previous studies [16‐18] reported that although the changes related to aging are common in all persons, the speed of loss of organ and tissue functionality differs between individuals, even at early ages, implying that aging has an inter-individual effect and remarking the differences between chronological and biological age. Specifically, the Dunedin study, carried out in a population of 869 young adults followed up from birth to age 45, showed that individuals who aged more rapidly according to their biological age were less physically able, had greater cognitive decline and brain aging, and presented worse self-perceived health [16, 17]. The objective of this study was to characterize biological aging in the early stages of adulthood and the relationship between the individual aging process and the development of chronic diseases. Death in this study population is yet to be determined. Similarly, the Baltimore Longitudinal Study of Aging [19] revealed, in a smaller cohort (968 participants) that slower paces of aging were related to better health status and lower risk of death. In this line, a study from the UK Biobank [20] calculated biological age specifically for each organ system and used this to predict the development of chronic diseases and premature death with high accuracy. However, the maximum lifespan of the participants in this study was 83 years, and extreme lifespan was not assessed. Centenarians and supercentenarians from the New England Centenarians Study showed that the prevalence of escapers (defined as individuals with onset of the first studied disease after age 100 years) was higher in centenarians, semisupercentenarians, and supercentenarians, compared with non-centenarians [21]. In this case, participants who died before 90 were considered as a single group, and shorter lifespans were not addressed. Other projects such as the Global Burden of Disease Study [22] investigated death and disability across all ages but either focused only on the causes of death or on chronological age and population alive at each age rather than the biological age or lifespan of the participants. In this sense, evaluating the interplay between longevity and diseases, including all the range of possible lifespans in a population-wide cohort should be crucial to discern the patterns that define human lifespan.

Until recently, the main increase in lifespan was related to progress in the prevention and treatment of infectious diseases (COVID-19 aside), but this tendency has now shifted and this increase is attributable to the improvement in chronic disease care and the decline in death rates after 65 [23]. In this sense, the characterization of age-related diseases (onset, progression, time between diagnosis and death) in large populations and the definition of the diseases more implicated in the human lifespan are crucial to continue delaying their onset. This knowledge should contribute to defining health strategies to decrease the pace of aging, promote P4 (predictive, preventive, personalized, and participatory) medicine, and increase lifespan and health span.

In the present study, we offer a comprehensive study, at a population-wide level, of the sex-specific pathological patterns, age of onset, multisystem involvement, and years free of disease and their association with human lifespan. We evaluated, in a cohort of 482,058 participants (50.4% women) with a mean lifespan of 83 years (ranging from 50 to 112), survival models for the onset of disease in multiple organ systems, the ability to avoid disease, and the percentage of life free of disease, all of which were evaluated according to the lifespan of the individuals. These results were specifically evaluated for each sex, considering the organ systems independently and accounting for multisystem involvement. The results show novel relationships among the age of onset of disease, health span, lifespan, and how these are adapted in escapers, showing system-specific and disease-specific trends that are highly relevant in explaining lifespan variability. Furthermore, there are notable differences between women and men at the system level in the determination of health span.

In this study, we analyzed the relationship between lifespan and age of onset, avoidance, and percentage of life free of diseases to define their contribution to lifespan. We describe a gradual delay in the age of onset of all diseases as lifespan increases, both single and multisystem, indicating that globally, individuals with greater lifespan have a specific resistance to developing diseases and can delay their onset. The delay was observed in both lethal (e.g., neoplasms) and non-lethal diseases (e.g., diseases of the musculoskeletal system and connective tissue) indicating a globally slower pace of aging in longer-lived individuals. These results agree with those reported in extremely long-lived individuals (from 100 to 119 years old) that demonstrated that those individuals present a later onset of frailty and diseases such as cancer, cardiovascular disease, dementia, and stroke, as well as cognitive and functional decline [21], and reinforce the effectiveness of disease prevention for increasing lifespan rather than curing diseases. In the same line, long-lived families can delay or avoid age-related diseases to a greater extent than shorter-lived families [28]. Furthermore, our results show that the HR gradually decreases as the number of systems affected increases, indicating that this resistance is enhanced for multisystem involvement.

When considering the prevalence of escapers, we observed that this prevalence was lower in lifespans around life expectancy (in most diseases, between lifespans of 70 and 90 years) depending on the system and that individuals with the longest lifespans had a similar prevalence of escapers than those with the shortest ones. These results could be explained by a higher severity of specific diseases in individuals with shorter lifespans, which may prevent the onset of other diseases before death [29], in contrast with a higher protection against developing diseases in longer-lived individuals, as suggested by other authors [30, 31]. We also observed how globally, the number of systems free of disease decreased until the ages of death 87–88 and started increasing thereafter, indicating that up to these ages (87–88) individuals die with more systems affected as their lifespan increases whereas after these ages, individuals progressively die with less systems affected as their lifespan increases. Similarly, a previous study performed with the SIDIAP database described how before 85 years of age frailty is mainly associated with the number of concurrent diseases, and that from 85 years onwards frailty is associated with disability and other signs and symptoms [32]. These results support the existence of a major survival hurdle at 87–88 years old, defining two main groups of individuals in terms of their lifespan: those who die before and those who die after. Before these ages, our results confirm the age-dependent loss of homeostasis in a system-specific fashion. Individuals that die after 87–88 years of age may have greater protection against suffering age-related diseases, which may be the result of optimized homeostatic protective mechanisms. Until now, there has been a consensus that considers centenarians as chronological age-based models of successful aging [30], together with other individual-centered models such as the one proposed by Rowe and Kahn [33]. However, based on our data, we propose that centenarians are the most extreme case of successful aging, and individuals with lifespans over 87–88 years are also relevant for the study of extreme lifespan. Consequently, a better characterization of the individual pace of aging in these subjects should provide valuable information on how to delay the onset of chronic diseases and increase health span and lifespan in the whole population. In this sense, a deep molecular characterization of the biology of aging in human tissues is crucial to better understand the meaning of human longevity and should serve as a reference to identify longevous and non-longevous individuals among middle-aged adults to better define individualized health strategies.

Focusing on sex-specific patterns, we found remarkably low HR and higher prevalence of escapers in women for diseases of the respiratory system and neoplasms at lifespans around life expectancy, as well as remarkably high HR and lower prevalence of escapers for diseases of the musculoskeletal system and connective tissue and for diseases of the nervous system. Neoplasms and diseases of the respiratory system (excluding infectious diseases) are the second and fourth leading causes of death in Spain [34]. Advanced pulmonary age has been described as the strongest predictor of mortality in the UK Biobank study [20], as well as an important mortality risk factor in epidemiological studies [35]. In contrast, diseases of the nervous system are the third cause of death, and diseases of the musculoskeletal system and connective tissue are the last cause of mortality among the studied diseases [34]. Both diseases of the nervous system and diseases of the musculoskeletal system have significantly higher survival times [36] and are among the main causes of frailty [25, 37]. These results could be explained by the differences in metabolic regulation in women and men during the aging process. For instance, previous results from our group described a higher involvement of gut-derived aromatic amino acids in women, which are strongly related to cognitive function and could lead to a higher risk of developing age-related nervous system diseases, as described in this work [38]. Furthermore, our study reveals that multisystem involvement is lower in women than in men at longer lifespans, contrarily to most of the current literature, which describes higher multimorbidity in women, specifically at older ages [32, 39]. One possible explanation is that we evaluated the number of systems affected instead of global multimorbidity. These differences imply that men and women achieve the same lifespan through different mechanisms. In women, the capacity to delay the onset of more lethal diseases and multisystem involvement could be greater, while men may be able to survive them for longer times. This could partly explain why life expectancy in women is greater, as mortality is closely associated with disease-related frailty [32]. In line with this hypothesis, we previously described sex-specific metabolic regulation during aging [38], that could provide women a better adaptation to the aging process and a higher resistance to multisystem diseases. Notably, the hazard rate for neoplasms in women at younger ages of death was higher than in men, in line with the increase in the incidence of prostate cancer in men after 65 years of age compared with the decrease in breast cancer after this age in women [40].

When analyzing the percentage of life free of disease in non-escapers, we found important differences related to the system that could explain human lifespan. Our results suggest that in people who suffer neoplasms, lifespan is strongly related to survival. In contrast, in the case of diseases of the respiratory system (men and women), and endocrine, nutritional, and metabolic diseases (men), lifespan is related to the ability of individuals to delay these diseases. In contrast with other chronic age-related diseases such as diabetes mellitus, Alzheimer’s disease, hypertension, or chronic obstructive pulmonary disease, to name some of the most prevalent diseases from other systems, neoplasms present an earlier onset, higher short or mid-term mortality and, although considered as a chronic disease, could be a one-time event [41]. These results highlight the importance of understanding the underlying mechanisms leading to each pathology to better comprehend their etiology.

Integrating the variables related to escaping the disease and the percentage of life spent free of disease at the time of death in an MFA, we found that these variables alone can contribute to explaining variability in individual lifespan.

This study has some limitations: (i) It was based on real-world data and specifically EHR, so it may include inconsistencies derived from data entry during clinical practice. However, these are expected to be small in number and minimized by the size and representativeness of the study population. (ii) Data before 2005 were collected in paper records, and some of the data, such as completeness of the registers and dates of diagnosis, might be less accurate. However, the fact that the results are sex- and system-dependent imparts a trustworthiness of there not being a generalized inconsistency. (iii) The analysis did not include confounding variables. This was because the main confounders that may be used can be interpreted as the cause of accelerated or decelerated aging rather than purely confounding variables. (iv) The severity of the disease was not taken into account. The wide scope of the present study rendered it impossible to specify each disease. Further system-specific studies discriminating between specific diseases must be performed. (v) The study population exhibited bias in terms of race and ethnicity, as 80% of the individuals were from Spain, 5.1% from other European regions, 6.2% from America, 4.9% from Africa, 3.1% from Asia, and less than 0.1% from Oceania. Consequently, this fact may limit the generalizability of the results to all regions worldwide.

The key findings of this study are specified in Fig. 6 and can be summarized as (i) the age of onset of disease, both single and multisystem, is gradually delayed as lifespan increases, (ii) the prevalence of escapers is lower in ages around life expectancy, (iii) the number of systems free of diseases decreases until ages of death 87–88 years old and increases onwards, (iv) long-lived women are less prone to multisystem involvement, (v) health span-lifespan associations are system-dependent, (vi) disease and percentage of life free of disease at death contribute to explaining the variability of lifespan, and (vii) there are system-specific differences between women and men.

Health interventions focused on delaying aging and age-related diseases should be the most effective in increasing not only lifespan but also health span. The findings of this research highlight the relevance of EHR in studying the aging process and open up new possibilities in age-related disease prevention that should assist primary care professionals in devising individualized care and treatment plans.