Physical robustness and resilience among long-lived female siblings: a comparison with sporadic long-livers

Introduction

Long-lived individuals are central in studies of the determinants of a long and healthy life. Both genetic and environmental factors contribute to longevity, and research is increasingly more focused on the genetic part [1–4]. As highlighted in several Scandinavian twin studies, longevity clusters in families, which suggests the existence of genetic variants for survival [5–7]. These studies showed that lifespan is partly heritable, and that the heritability increases with age. One study indicated a genetic influence on human longevity, underscoring minimal genetic effects on lifespan for people under age 60 and then moderate genetic effects for people aged 60 and over [6]. However, few genetic or other pro-longevity factors have been identified, possibly due to “phenocopies” - those individuals that live long by chance. Some studies reported better health in long-lived siblings compared to sporadic long-livers, including a lower prevalence of Alzheimer’s disease and related disorders, diabetes, depression, heart failure and osteoporosis [8, 9]. One Dutch longitudinal study found better survival in long-lived siblings compared to sporadic long-livers [10]. However, these studies did not evaluate disease risk and survival in the same population and, therefore, were not able to examine the interplay between indicators of robustness and resilience, which is the focus of this study.

Physical robustness could be defined as the ability to resist a deviation from the normal physical state and avoid an adverse health event [11]. In this framework, disease risk can be viewed as a proxy indicator of the whole-body robustness [12]. Physical resilience could be defined as the ability to bounce back, and quickly and completely recover after an adverse health event [11, 12]. The ability to survive to very old age could be used as a proxy indicator of the whole-body resilience [11, 12]. Physical resilience universally declines with age, increasing vulnerability to death as people grow older. Robustness also generally declines with age – manifested by increased risks of many health disorders (e.g. Alzheimer’s disease, stroke, renal and heart failure, fractures, pneumonia), and disabilities in the elderly; however, robustness may improve in some health domains, reflected in declining risks of certain chronic diseases (e.g. several cancers, diabetes, asthma) towards extreme ages [11–13]. This indicates that the mechanisms underlying changes in physical robustness and resilience during aging are not necessarily (or entirely) the same. The same factor may have beneficial effects on resilience and adverse effects on robustness, and vice versa. For example, a chronically suppressed apoptosis in the body may contribute to both increased risk of cancer (lower robustness) in middle-old life and better survival at extreme ages (higher resilience) [14].

The long-lived individuals may live longer than the general population for various reasons. They may be more robust than the general population because they are able to avoid major diseases, or they may be more resilient because they are able to better survive after disease onset, or both. They may also live longer simply ‘by chance’. This study investigates whether long-lived female siblings are more robust and/or resilient than sporadic (“non-familial”) long-lived women of the same age, and, specifically, whether they have a better ability to (i) avoid the common diseases of the elderly at ages 65+ (proxy for robustness), and/or (ii) survive to extreme ages (proxy for resilience). We compared the cumulative incidence of the 20 most prevalent chronic conditions from age 65 between long-lived female siblings and sporadic long-lived Danish women. We also evaluated the difference in overall survival at the oldest-old ages between the familial and the sporadic long-livers, and examined factors contributing to this difference. Due to a scarce number of long-lived males, we used only data on long-lived females from three studies enrolling Danish long-lived siblings.

Results

Discussion

People may live long for various reasons including better robustness or better resilience, or both. They may also live longer lives simply by chance. In this study, we examined whether the long-lived female siblings are more robust and/or more resilient than sporadic long-lived women. Our results indicate that the long-lived siblings may be more robust to some health conditions (hypertension, cerebrovascular disease, depression), and less robust to some other (cancer); though only the risk of depression (OR=0.72; p-value=0.001) remained statistically significant after adjustment for multiple testing [15]. Our results strongly support the possibility that the long-lived female siblings are more resilient than sporadic long-livers because they show significantly better survival at extreme ages, even after controlling for all covariates/comorbidities. This indicates that familial longevity could be mainly related to a better resilience as the ability to overcome various life and health problems, rather than to a simply good health, and that being not depressed may be a key factor supporting organism’s robustness in advanced years of life.

Findings from earlier US studies that used samples of LLFS data suggest that long-lived siblings could be more robust to some diseases. Ash et al. [8] found lower risks of depression, Alzheimer’s disease, diabetes, heart failure, and some other conditions, in the long-lived LLFS siblings compared to the sporadic long-lived Medicare beneficiaries [8]. Unlike Ash et al., we found no evidence for a significantly different risk of dementia between the long-lived female siblings and controls [8]. However, Ash et al. studied a combination of both sexes, whereas our study focused only on females, which might lead to the difference in results. Similar to our findings, Newman et al. reported a lower risk of hypertension in LLFS compared to the Framingham Heart Study (FHS), but not to the Cardiovascular Health Study (CHS) and the New England Centenarian Study (NECS) participants [9]. That study also reported mixed results regarding stroke, with a higher risk in long-lived LLFS siblings compared to CHS and FHS, and no significant difference for NECS participants [9], while our results indicated a lower risk of cerebrovascular disease in the long-lived female siblings in the age-adjusted analysis.

In our study, we found a marginally significant higher risk of cancer among the long-lived female siblings compared to sporadic long-livers. Ukraintseva et al. earlier suggested potential biological mechanisms of trade-offs between extreme longevity and cancer risk, including antagonistic pleiotropic role of some genetic and non-genetic factors in aging and cancer development, which could be applicable to these results [14]. Other studies did not report a significant difference in the risk of cancer between long-lived LLFS siblings and controls [8, 9], but this may be due to differences in research design. A higher risk of hearing loss in the long-lived female siblings compared to sporadic long-livers did not remain significant after adjustment for age. So, we may assume that this difference was related to the age difference between siblings and controls. If multiple testing was considered using the conservative Bonferroni correction or the FDR test, only the lower risk of depression among siblings remained statistically significant [15].

Regarding survival to extreme ages, we observed a better overall survival of female long-lived siblings compared to sporadic long-livers. Similar findings were reported in a Dutch study that also showed better survival of the long-lived siblings at high ages, as compared to long-lived sporadic controls [10]. The association remained statistically significant after adjustment for age, marital status, comorbidities and/or medication.

It should be emphasized that none of the above studies considered indicators of physiological robustness and resilience in the same population, whereas in our large population-based study, we investigated both disease risks and survival to extreme ages in the same individuals. Our results suggest that better physiological robustness (manifested by lower disease risks) does not entirely explain better resilience (manifested by a higher survival to extreme ages) of female members of the long-lived families, as compared to sporadic long-livers. And adjustment for the above diseases, to which the long-lived female siblings were more robust, did not change the association, or did so only marginally.

We showed that longevity clustered in families can lead to a significantly lower mortality risk at extreme ages compared to sporadic long-livers. This suggests that resilience has similar biological mechanisms in the long-lived siblings influenced by common genetic and other familial factors. Indeed, in the first part of their lives, siblings share many environmental factors (e.g., socioeconomic status, place of residence, lifestyle). In addition, studies have highlighted the influence of genetic effects on lifespan at older ages, indicating that genetic factors could play a major role in the familial longevity [6]. Genetic factors are more likely than other factors to be related to resilience in long-lived siblings compared to sporadic long-livers, although other factors should also be explored.

There are some limitations to this study. First, long-lived female siblings were, on average, slightly but significantly older than the sporadic long-lived women even after matching on age. However, our analyses were adjusted for age. This adjustment did not lead to different results in the survival analyses. In fact, the adjustment for age only changed the HR slightly, and the association became stronger. Second, the condition or disease for which the medication was prescribed (indication code) was not available [16]. Not considering the indication codes could lead to a potential information bias with an overestimation of some chronic conditions. Nonetheless, if there was information bias, it was non-differential with an overestimation in long-lived siblings and controls. Third, only chronic conditions diagnosed through medication or in hospitals were considered. Consequently, some conditions could not be included. However, we focused on the 20 most prevalent chronic conditions in Danes aged 75 and over [17] and defined them based on validated definitions [16]. Finally, the initial study, which in 2004 identified the long-lived families that were included in the present study, did not use current definitions of longevity such as top sex and birth cohort survival percentiles [18, 19]. In our study population, 99.5% of the recruited families included at least two siblings who survived to age 90 years, whereas for the remaining families, one sibling survived well past 90 years and at least one other sibling survived to age 89 years. With our definition of longevity, a large sibship – everything else equal – has a higher probability of becoming a long-lived family compared to a sibship of say two. Recognizing this bias, we have been collecting evidence to address the potential size of the bias. We found that the long-lived siblings in our study were from sibships with an average of 7.2 siblings compared to 6.6 in 358 control families that were selected among families with at least two children. A difference this size seems unlikely to be of importance for late life disease and survival. Moreover, we cannot exclude that some controls were from non-identified long-lived families, but they would be rare. Also, males have higher mortality throughout life compared to females. Therefore, a study of similar long-lived men would have been an even more selected study sample and bring additional information to our findings. However, in our study we were not able to identify a sufficient number of male controls in the 5% sample of the Danish population to allow meaningful analyses.

The main strength of this work lies in the study design consisting of long-lived female siblings matched on age with female controls from a representative sample of the Danish population. Thus, the registry-based study design permitted avoidance of selection bias. Long-lived siblings and controls were all identified in high-quality National Danish registers, also leading to a large sample size. Finally, the definition of chronic conditions was based on both hospital and medication data [16]. In this way, we identified individuals with different disease levels, although not all individuals.

In conclusion, long-lived female siblings demonstrated better robustness to some health conditions (especially, depression), while increased vulnerability to some other diseases (cancer). Physiological resilience (manifested in higher chances of survival to extreme ages) was consistently better in the long-lived female siblings than in their age-peers from general population. This indicates that the ability to overcome deleterious life events may be more important for extreme longevity than a good health alone, and that avoiding depression is major factor of maintaining physical robustness in familial longevity. A consistently better survival of the long-lived siblings also suggests that resilience may have stronger genetic component in familial longevity, warranting further investigation. Overall, results of this study indicate that long-lived siblings are excellent candidates for healthy longevity studies, and that familial longevity cases could be more informative than sporadic cases for studying mechanisms of longevity. Since health phenotypes such as discharge diagnoses of chronic conditions and medication could not explain better resilience of the long-lived female siblings, a next step can be to focus on other factors that might explain their better resilience, such as response to acute health events, dynamic changes in functional status, cognitive functioning, or psychological factors [11; 20–22]. Cognitive and physical functioning are known to be highly predictive of survival in the very old individuals [20]. In addition, it would be reasonable to explore last year of life events, or the cause of death in long-lived siblings compared to sporadic long-livers to examine whether they experience different life events compared to sporadic long-lived women.

Materials and Methods

Supplementary Materials

Supplementary Table 1