Elsevier

Experimental Gerontology

Volume 43, Issue 12, December 2008, Pages 1052-1057
Experimental Gerontology

Sex-specific health deterioration and mortality: The morbidity–mortality paradox over age and time

https://doi.org/10.1016/j.exger.2008.09.007Get rights and content

Abstract

The traditional sex morbidity–mortality paradox that females have worse health but better survival than males is based on studies of major health traits. We applied a cumulative deficits approach to study this paradox, selecting 34 minor health deficits consistently measured in the 9th (1964) and 14th (1974) Framingham Heart and 5th (1991–1995) Offspring Study exams focusing on the 55–78 age range. We constructed four deficit indices (DIs) using all 34 deficits as well as subsets of these deficits characterizing males’ (DIM) and females’ (DIF) health disadvantages, and no relative sex-disadvantages. The DI34-specific age patterns are sex-insensitive within the 55–74 age range. The DI34, however, tends to selectively increase the risk of death for males. The DIF-associated health dimension supports the traditional morbidity paradox, whereas the DIM-associated dimension supports the inverse paradox, wherein males have worse health but better survival than females. The traditional paradox became less pronounced, whereas the inverse paradox became more pronounced from the 1960s to the 1990s. The sex-specific excess in minor health deficits may vary according to particular set of deficits, thus providing evidence for traditional and inverse morbidity paradoxes. The time-trends suggest the presence of a strong exogenous effect modifier affecting the rate of health deterioration and mortality risk.

Introduction

Females tend to have worse health than males, but they live longer than males. This traditional sex morbidity–mortality paradox was noted in the mid 1970s (Nathanson, 1975, Verbrugge, 1982). From that time, this paradox has been further studied by focusing on various aspects of health and the effect of socio-demographic factors (see, e.g., (Arber and Cooper, 1999, Case and Paxson, 2005, Idler and Benyamini, 1997, Lawlor et al., 2001, MacIntyre et al., 1999, Verbrugge and Wingard, 1987a, Verbrugge and Wingard, 1987b)).

The most common explanations of this paradox are based on differences in health and health behaviors, i.e., biological risks; the risks acquired through life-style-related behaviors, illness behavior, health reporting behavior, and differences in the health care utilization (MacIntyre et al., 1999, Verbrugge, 1989). Biological vulnerability and acquired risks are typical explanations when individuals are examined in clinical or laboratory settings (Verbrugge, 1982). The other three potential explanations listed above are typically used to explain sex differentials observed in surveys wherein data on health behaviors are self-reported.

Numerous studies (see e.g., (Gorman and Read, 2006) and references therein) suggest that sex differences may vary by health dimensions reflecting the process of population health deterioration (a simplified pathway of health worsening can be sketched as: pre-disease → pathology → loss of functioning → disability → death (Crimmins, 2004, Verbrugge and Jette, 1994)). Most population-based studies have focused on major health-related dimensions including self-perceived health, functional limitations, selected diseases, as well as health-care utilization and hospitalization (see, e.g., (Case and Paxson, 2005, Gorman and Read, 2006, Idler and Benyamini, 1997, Verbrugge, 1982)). Explanations of males’ excessive mortality typically emphasize greater biological vulnerability and lower health care utilization (MacIntyre et al., 1999, Verbrugge, 1989). The excessive morbidity of females is usually attributed to illness (e.g., diabetes, heart diseases, hypertension) and health reporting behaviors (Verbrugge, 1982, Verbrugge, 1989). Despite these studies, however, the nature of this paradox is still uncertain.

Further insights into the nature of the morbidity–mortality paradox can be gained by analyzing “minor-effect” health traits associated with the least severe health dimensions, i.e., with the pre-disease state (e.g., signs, symptoms) and certain diseases. Such analyses can shed light on possible sex differentials in pre-disease conditions that may lead to sex-targeted interventions to prevent the development of major health problems. Relatively few population-based studies, however, have focused on such traits (see MacIntyre et al., 1999, Verbrugge, 1982 and references therein). These studies basically examined the prevalence of problem-oriented symptoms (e.g., symptoms related to headaches Celentano et al., 1990). They found that rates of morbidity for minor-effect health traits tend to be higher in females (Emslie et al., 1999), although the sex-specific excess in such traits may vary according to the particular setting of problems (MacIntyre et al., 1996). These studies, however, rarely examined associations of those minor traits with mortality. The challenge facing studies of such associations is the large number of various minor health traits and the small, insignificant, or inconsistent effect of each on mortality risks. The aggregate effect of several such minor traits, however, might be more informative. This is the underlying paradigm of recent developments of a promising new instrument, which is called a frailty index (Goggins et al., 2005, Mitnitski et al., 2001, Rockwood and Mitnitski, 2007) or an index of cumulative deficits (Kulminski et al., 2007c, Yashin et al., 2007a). The concept of a cumulative health deficits index (DI) appears to be useful for examining possible sex health differentials on the level of minor health traits.

This study uses the concept of the DI to address the following three major research questions: (i) Are there differences in the prevalence of morbidity associated with minor health traits (e.g., signs, symptoms, health conditions with non-significant effect on mortality, etc.) between the sexes? (ii) How do those minor health dimensions affect the mortality risks for each sex? and (iii) Are there time trends in the morbidity paradox? To address these questions, we studied the health status of a sample of adult and elderly individuals participating in the Framingham Heart Study [FHS] and Offspring [FHSO] using the new instrument, the DI, which aggregates minor health-related variables routinely collected during the 1960–1990 period.

Section snippets

The FHS and FHSO data

Beginning in 1948, 5209 respondents (46% male) aged 28–62 years residing in Framingham, Massachusetts were enrolled in the FHS. The FHSO dataset consists of a sample of 3514 biological descendants of the FHS cohort, 1576 of their spouses and 34 adopted offspring for a total sample of 5124 subjects; 48% male. The FHSO subjects were enrolled in 1971–1975 using research protocols similar to those of the FHS so that comparisons of the results from the FHSO and the FHS could be made. Selection

Results

For the pooled sample of participants of the 9th and 14th FHS and 5th FHSO exams, Fig. 1a shows no significant differences in the mean values of the DI constructed using a set of all 34 deficits (DI34) except in the last age group (75–78 years). For certain age groups, however, the DI34 for males tend to be higher than for females. Disaggregation of the DI34 into the DIs associated with males’ health disadvantages (i.e., DIM; prevalence of individual deficits is significantly higher for males

Discussion and conclusions

The results show the capacity of the cumulative deficits approach for characterizing sex differentials in morbidity and mortality outcomes when considering the effect of minor health-related traits. The aggregation of such minor traits into an integrative or cumulative measure (i.e., the DI) establishes a background for reliable conclusions which are often difficult to achieve considering individual minor traits.

Analysis of the pooled sample of the participants of the 9th and 14th FHS and 5th

Acknowledgements

The research reported in this paper was supported by the National Institute on Aging Grants 1R01 AG028259, 1R01-AG-027019, 5R01-AG-030612, and 5P01-AG-008761. The Framingham Heart Study (FHS) is conducted and supported by the NHLBI in collaboration with the FHS Investigators. This Manuscript was prepared using a limited access dataset obtained from the NHLBI and does not necessarily reflect the opinions or views of the FHS or the NHLBI.

References (26)

  • T.R. Dawber et al.

    An approach to longitudinal studies in a community: the Framingham Study

    Annu. NY Acad. Sci.

    (1963)
  • W.B. Goggins et al.

    Frailty index as a measure of biological age in a Chinese population

    J. Gerontol. A Biol. Sci. Med. Sci.

    (2005)
  • B.K. Gorman et al.

    Gender disparities in adult health: an examination of three measures of morbidity

    J. Health Soc. Behav.

    (2006)
  • Cited by (41)

    • A quantile frailty index without dichotomization

      2021, Mechanisms of Ageing and Development
      Citation Excerpt :

      In many studies men are measured as “healthier” despite having greater prevalence of negative outcomes (Gordon and Hubbard, 2018). Selecting clinical-level health deficits based on sex-dependent prevalence affects sex-dependent mortality prediction of composite measures (Kulminski et al., 2008). When biomarker measurements are used the prevalence of each deficit can be tuned by the dichotomization threshold.

    • The potential for complex computational models of aging

      2021, Mechanisms of Ageing and Development
    • Modulation of frailty syndrome by diet: A review of evidence from mouse studies

      2019, Mechanisms of Ageing and Development
      Citation Excerpt :

      Moreover, CR reduced FI in a sex dependent manner with males having a significant reduction in FI when measured at 18mo of age (14mo on CR), while significance was not reached in females (Kane et al., 2016b). Such sexual dimorphism is common but poorly understood, including the morbidity-mortality paradox in humans in which females have better survival despite worse health than males (Kulminski et al., 2008), or an apparently opposite phenomenon in B6 females on 40% CR showing improved healthspan without lifespan extension (Mitchell et al., 2016a). While this ground-breaking study showed the relative ease with which non-invasive frailty measurements could be implemented into ongoing mouse studies, it also sets the stage for much needed mechanistic and interventional studies, including testing the interesting possibility of FI as a measure of biological rather than chronological age.

    View all citing articles on Scopus
    View full text