Marital Disruption and Allostatic Load in Late Life

Abstract

Objective: This study examines the link between marital disruption and biological risk, and asks whether the association of this precarious life event with health is contingent on marital loss duration and history. Method: Data are drawn from the National Social Life, Health, and Aging Project (N = 1,414), and Poisson regressions are presented for allostatic load and logistic regressions for individual biomarkers. Results: The currently unmarried with more distant marital disruptions exhibit higher levels of allostatic load than the currently married, which is primarily driven by dysregulation of cardiovascular and metabolic indicators. Results also reveal the differing ways marital disruption “gets under the skin” with widowhood associated with compromised inflammatory, metabolic, and cardiovascular functioning, and divorce with cardiovascular and neuroendocrine markers. Discussion: Findings lend support for both the crisis and chronic strain models, and likely reflect normative expectations of the timing of life events, decrements in marital resources, and selection processes.

Keywords

allostatic load biomarkers divorce widowhood

Introduction

Recent shifts in marital transitions beg an array of questions concerning prospects for successful aging. Never marrying is becoming more common, the divorce rate for older adults is twice as high as it was in 1990, and because the population is aging, the percentage of those who have married and experienced marital disruption has almost doubled since 1960 (Brown & Lin, 2012; Kreider & Ellis, 2011; Wang, 2014; Wang & Parker, 2014). In addition to these shifts, an extensive body of research shows that the currently married fare better than the currently unmarried on numerous heath indicators (Carr & Springer, 2010; Fincham & Beach, 2010; Waite & Gallagher, 2000). Moreover, this health advantage of the married has been widening in recent years, especially in regard to the currently widowed (Liu & Umberson, 2008).

In addition to marital status, marital biography, which encompasses marital duration and number of prior marital transitions, is associated with health and well-being (Hughes & Waite, 2009). Most of this research has focused on mortality risk or self-reported indicators of health; however, more recent studies have paid attention to the direct biological consequence of marital transitions using national data (Das, 2013). In particular, the theory of allostatic load highlights the process whereby psychosocial conditions and stressors such as marital loss elicit responses in the body’s regular systems and increases the risk for poor health outcomes (Juster, McEwen, & Lupien, 2010). Although these responses are influenced by individual genetic and psychological vulnerability, there are likely general patterned responses to marital disruption.

The current study builds on previous work by linking biographical aspects of marital loss, including duration and history, to biological processes especially physiological and immunological pathways that link marital disruption to more tertiary outcomes. Data are drawn from the National Social Life, Health, and Aging Project (NSHAP), a population-based sample of U.S. adults aged 57 to 85. Understanding how, and to which extent, marital loss is associated with biological risk is the first step in identifying the most effective social, economic, and behavioral interventions to alleviate vulnerability to this potentially stressful life event.

Background

Marital Loss and Allostatic Load

The concept of allostasis suggests that to support homeostatic systems that are essential to living, physiological networks must adapt and change when faced with stressors (Sterling & Eyer, 1988). Stressful life circumstances, therefore, precipitate physiological changes in the sympathetic (and parasympathetic systems of the body (Crimmins, Johnston, Hayward, & Seeman, 2003; McEwen, 1998; Seeman, Singer, Rowe, Horwitz, & McEwen, 1997; Sterling & Eyer, 1988). In the short term, these changes are beneficial to functioning. However, when prolonged, these changes can tax and damage physiological systems. Increased demands along with the inability to modify the allostatic response results in allostatic load (Juster et al., 2010; McEwen, 2003; McEwen & Stellar, 1993).

Measures of allostatic load indicate functioning across interrelated physiological systems, including neuroendocrine, cardiovascular, metabolic, inflammatory, and immune response. Indices typically include primary mediators of the physiological stress response of the hypothalamic pituitary adrenal axis (HPA axis) and sympathetic nervous system (SNS) (e.g., cortisol, epinephrine, norepinephrine, dehydroepiandrosterone [DHEA]) as well as secondary outcomes (e.g., blood pressure, glycosylated hemoglobin, obesity) that reflect the activation of stress hormones and prime individuals for the flight or fight response (McEwen & Seeman, 1999). Allostatic load predisposes organisms to disease (McEwen & Stellar, 1993) and is associated with hypertension, diabetes, autoimmune disorders, and cancer (McEwen & Wingfield, 2003). Allostatic load indices predict mortality (Seeman et al., 2001), and may indicate frailty development (Gruenewald, Seeman, Karlamangla, & Sarkisian, 2009).

The existence and quality of social relationships, especially ties with family members, are linked to individual mediators and overall allostatic load (Brooks et al., 2014; Robles & Kiecolt-Glaser, 2003). The currently married, on average, have better cardiovascular and inflammatory functioning than the unmarried (Holt-Lunstad, Birmingham, & Jones, 2008; Sbarra, 2009); although, there is indication that gender conditions these associations among older adults (McFarland, Hayward, & Brown, 2013; Sbarra, 2009). Similarly, marital disruption is associated with elevated blood pressure and heart rate as well as compromised neuroendocrine and immune functioning (Das, 2013; Gersten, 2008; Kiecolt-Glaser, Bane, Glaser, & Malarkey, 2003; Kiecolt-Glaser et al., 1987). For allostatic load, data on older adults in Taiwan show that the unmarried, especially the recently widowed, have higher allostatic load scores than the currently married (Seeman et al., 2004; Weinstein, Goldman, Hedley, Yu-Hsuan, & Seeman, 2003). Therefore, marital status is associated with allostatic load and there is indication that marital biography is linked to individual mediators. Three models explain the linkage between marital status and health.

Marital Loss: Crisis, Chronic Strain, and Selection Models

The marital crisis model suggests that marital status differences in health are driven by the immediate stress and trauma of marital loss (Booth & Amato, 1991; K. Williams & Umberson, 2004). Within months of marital disruption, acclimatizing to new living arrangements, residence, or daily routines such as managing household and financial tasks can be difficult. In the short run, marital disruption tends to increase psychological distress (Hughes & Waite, 2009; Simon, 2002); yet, these effects vary by type of marital transition and prior marital functioning. Most research lends support to the crisis model for mental health in that elevated levels of depressive symptoms tend to taper off within 2 years of spousal loss (Sasson & Umberson, 2014). Divorce also increases psychological distress; although, there is evidence of better adjustment (Umberson, Liu, & Powers, 2009) likely due to exiting a dissatisfying relationship and the voluntary nature of the transition. Research also shows that strained or conflictual marriages impair immune response and cardiovascular functioning (Robles & Kiecolt-Glaser, 2003) and accelerate health decline especially at advanced ages (Umberson, Williams, Powers, Liu, & Needham, 2006). Exiting a dissatisfying marriage, therefore, can benefit health.

The chronic strain model indicates that marital disruptions prompt decrements in financial and social resources and set off more chronic and enduring strains (Umberson, Wortman, & Kessler, 1992), which lead to disability and elevated mortality risk. Lorenz, Wickrama, Conger, and Elder (2006) find that divorce increases disease burden through the accumulation of stressful life events. There is also evidence that the mortality disadvantage faced by the currently divorced is reflective of those divorced the longest having an especially high risk of early death (Sbarra & Nietert, 2009). Although the likelihood for a recently widowed person to die is highest immediately following spousal death, there is evidence of a “widowhood effect” for more than a decade after the transition (Shor et al., 2012). Long-term exposure to a marital loss status is also associated with low levels of social support and health compromising behaviors (Das, 2013). A decline in income and assets following marital disruption has implications for financial strain (Lavelle & Smock, 2012; Zick & Holden, 2000) and, therefore, overall health, especially among women.

Similarly, repeated marital disruptions may exert an additive or cumulative effect on health. Second and third marriages are less beneficial to well-being than first marriages (Barrett, 2000), and there is indication that marital history is more closely linked to physical than mental health (Hughes & Waite, 2009). There is also indication that previous marital disruptions are associated with compromised cardiovascular health in mid- and late life, and that number of previous marital disruptions can have lasting effects on cardiovascular functioning (Donoho, Seeman, Sloan, & Crimmins, 2015; Zhang & Hayward, 2006). Results from the NSHAP, in particular, show that past marital disruptions and being in a marital loss status for 10 or more years are risk factors for elevated cardiovascular and metabolic functioning (Das, 2013).

Finally, the marital selection model simply states that the health advantages of married adults reflect healthier and happier adults marrying in the first place (Hill, Reid, & Reczek, 2013; Johnson & Wu, 2002). Poor health status and low socioeconomic status (SES) can select individuals into earlier marital disruptions, longer widowhood and divorce durations, and multiple marital transitions (Carr & Utz, 2001). For example, poor mental health may select individuals into earlier marital disruptions, especially divorce. Remarriage likely selects the healthiest out of their marital disruption statuses, leaving those with poor health in the widowhood and divorce statuses the longest. In terms of gender, constrained opportunities and choices influence physiology, which results in gender differences in health (Bird & Rieker, 2008). Furthermore, Sasson and Umberson (2014) explain that because women are more likely than men to marry early, live longer, become and stay widowed, and are less likely to remarry, women are overrepresented at the high end of widowhood and divorce durations. For age, older adults are more likely to marry younger and have the potential to accumulate more time in a marital loss status than those who are younger. Below, includes a description of the current study and how it addresses some of these selection processes.

The Current Study

Supporting previous research the the current study includes the following hypothesis:

Hypothesis 1: The currently widowed and the currently divorced will have higher allostatic load than the currently married.

In accordance with the chronic strain theory:

Hypothesis 2: The currently widowed and currently divorced who have been in their marital status the longest will have higher allostatic load scores than the currently married.

Finally, given that additional marital disruptions tend to compromise health and mark repeated depletion of key resources:

Hypothesis 3: Number of prior marriages will be positively associated with allostatic load.

To try to address some of the selection factors mentioned previously, I specify age- and gender-specific tertiles from the continuous marital status duration indicators to create marital duration categories (similar to Barrett, 2005). These age- and gender-adjusted indicators are most suitable for the current project given gender and age differences in timing of marital transitions (discussed above) and allostatic load (Seeman, Singer, Ryff, Love, & Levy-Storms, 2002). Without these specifications, women and older adults would be overly represented in the longest marital loss duration categories.

Materials and Method

Data

The NSHAP is a population-based sample of individuals aged 57 to 85 years old living in the United States, with equal representation from the following age groups: 57- to 64-year-olds, 65- to 74-year-olds, and those 75 years and older. The principal investigators at the University of Chicago commissioned the National Opinion Research Center (NORC) to conduct 2-hr face-to-face interviews including biomarker collection in respondents’ homes during 2005 and 2006, yielding a sample of 3,005 adults (Smith et al., 2009). All respondents are asked to participate in the anthropometric and saliva assessments, whereas blood spot collection is randomized to 83% of the original sample. Overall, 1,451 participants have viable information on all biomarkers and 1,414 respondents have complete information on all study variables.

Measures

Individual biomarkers and allostatic load

Allostatic load is assessed through seven biological markers: body mass index (BMI), systolic blood pressure, diastolic blood pressure, pulse rate, c-reactive protein (CRP), DHEA, and glycosylated hemoglobin (HbA1c). BMI is based on body weight and height. Systolic blood pressure, diastolic blood pressure, and pulse rate are obtained through a digital blood pressure monitor applied to the respondent’s left arm, and based on the mean of two (or if the two varied greatly, three) readings. All respondents are asked to provide 2 mL of saliva to detect DHEA concentrations, and results are based on an average of two laboratory tests. For CRP and HbA1c, the interviewer applied a lancet to the respondent’s finger, the blood spot was transferred to filter paper, and then transported to laboratories for assessment (S. R. Williams & McDade, 2009).

High-risk quartile cutoff criterion is used rather than clinical cut points. For each biomarker, respondents who score in the upper quartile of the sample for the continuous markers are coded as 1 and all others as 0 (for cut points, see the appendix), except for DHEA, which is inversely related to age and based on scoring in the lowest quartile. Allostatic load is a sum of the number of biomarkers for which the respondent fell into the highest risk quartile with higher scores indicating more biological risk (similar to Crimmins et al., 2003).

Marital status

Respondents are asked whether they are currently married, living with a partner, divorced, separated, widowed, or never married. Because few separated and partnered respondents have complete information on all biomarkers, divorced and separated respondents are combined into one category and the currently partnered were dropped from the analysis leaving the following categories: married (reference), divorced (or separated), widowed, and never married.

Marital status duration

For widowed or divorced respondents, duration is the difference between the year of the interview and the year their most recent marital partner died or the year they stopped living with their spouse, respectively. To create duration categories, first, the sample is divided into the three cohorts based on the sampling design of the NSHAP (57- to 64-year-olds, 65- to 74-year-olds, and those 75 years and older) then these cohorts are divided by gender. Age- and gender-specific tertiles (similar to Barrett, 2005) are specified from the continuous duration indicators to obtain three categories for the currently widowed (i.e., widowed longest, widowed shortest, and those in between), and two categories for the currently divorced (i.e., below or above the age- and gender-specific means of the continuous duration indicator).

Number of marriages is based on past number of marriages. Currently married, divorced/separated, and widowed respondents are asked whether their current marriage is their first marriage. Then, respondents are asked “Altogether, how many times have you been married?” Number of marriages is a multiple dummy variable indicating never married, married once (reference), married twice, and married three or more times.

Control variables include gender, age, education, and race/ethnicity. Gender is coded 1 for female and 0 for males. Age is a continuous indicator ranging from 57 to 85. Education is multiple dummy coded indicating respondents with less than a high school degree, a high school degree or equivalent, some college, and those with a bachelor’s degree or more (reference category). Respondents’ self-report of racial/ethnic identification is coded as a multiple dummy variable indicating non-Latino Black, Latino, other, and non-Latino White (reference category).

Analytic Strategy

Table 1 presents the descriptive statistics for all study variables. The remaining tables examine whether allostatic load and individual biomarkers vary as a function of marital status (Table 2), marital status duration (Table 3), and number of prior number of marriages (Table 4) when controlling for background factors. Logistic models with odds ratios are presented for individual biomarkers, and incidence rate ratios (IRRs) for the Poisson models are presented for allostatic load. Poisson models are best suited for count outcomes. IRRs are exponentiated Poisson regression coefficients that are interpreted as the difference in the expected allostatic load count for each one-unit change in an independent variable, while all other variables in the model are held constant (Long & Freese, 2006). Because the model relies on the assumption that the dependent variable is equidispersed (M = variance), supplemental likelihood ratio tests support the use of this model over negative binomial regression, ${\bar{χ}}^{2}$ (01) = 0.00, p = .50.

Table 1.

Descriptive Statistics for Study Variables (N = 1,414).

	%	M (SD)	Range	n
Allostatic load		1.74 (1.33)	0.00-7.00	1,414
Current marital status and duration
Married	62			874
Duration married		37.01 (15.39)	0.00-69.00
Widowed	23			324
Duration widowed		12.25 (10.71)	0.00-57.00
Divorced	12			170
Duration divorced		21.28 (12.96)	0.00-54.00
Never married	3			46
Number of marital transitions
One marriage	68			960
Two marriages	21			303
Three+ marriages	7			105
Female	49			697
Age		69.43 (7.81)	57.00-85.00	1,414
Black	14			192
Latino	11			160
Other	2			29
Less than HS	22			311
High school degree	27			378
Some college	28			399

Source. National Social Life, Health, and Aging Project (NSHAP; 2005/2006).

Table 2.

Poisson and Logistic Regression of Allostatic Load and Individual Biomarkers on Marital Status (N = 1,414).

	Allostatic load	High BMI	High systolic BP	High diastolic BP	High pulse rate	Low DHEA	High CRP	High HbA1c
(Ref. married)
Widowed	1.22*** [1.09, 1.36]	1.45 [0.98, 2.16]	1.37 [0.91, 2.06]	1.58* [1.01, 2.49]	1.13 [0.76, 1.70]	0.78 [0.59, 1.05]	1.45 [0.99, 2.09]	1.66* [1.10, 2.51]
Divorced	1.08 [0.91, 1.27]	0.76 [0.44, 1.31]	1.12 [0.72, 1.72]	1.32 [0.81, 2.13]	1.64 [0.99, 2.70]	1.12 [0.69, 1.81]	0.90 [0.56, 1.47]	1.03 [0.62, 1.70]
Never married	1.29 [0.96, 1.72]	1.30 [0.56, 3.05]	1.67 [0.82, 3.39]	2.34* [1.07, 5.11]	1.99 [0.97, 4.09]	1.11 [0.50, 2.45]	0.84 [0.32, 2.20]	1.14 [0.42, 3.09]
LR χ²	77.27***	50.08***	21.51*	52.16***	31.49***	105.85***	75.24***	94.48***
Pseudo R²	.02	.03	.01	.03	.02	.07	.05	.06

Source. National Social Life, Health, and Aging Project (NSHAP; 2005/2006).

Note. Shown are incident rate ratios for allostatic load and odds ratios for individual biomarkers with 95% confidence intervals in parentheses. BMI = body mass index; BP= blood pressure; DHEA = dehydroepiandrosterone; CRP = c-reactive protein; HbA1c = glycosylated hemoglobin; NSHAP = National Social Life, Health, and Aging Project.

Significant at .05. **Significant at .01. ***Significant at .001.

Table 3.

Poisson and Logistic Regression of Allostatic Load and Individual Biomarkers on Marital Status Duration (N = 1,414).

	Allostatic load	High BMI	High systolic BP	High diastolic BP	High pulse rate	Low DHEA	High CRP	High HbA1c
(Ref. married)
Widowhood duration categories
Short term^a	1.13 [0.95, 1.35]	0.98 [0.55, 1.78]	1.64 [0.76, 3.55]	1.17 [0.57, 2.41]	0.96 [0.56, 1.64]	1.02 [0.66, 1.57]	1.40 [0.87, 2.26]	1.14 [0.62, 2.08]
Medium term^b	1.15 [0.98, 1.35]	1.65 [0.92, 2.96]	1.09 [0.66, 1.79]	1.63 [0.89, 3.00]	1.18 [0.65, 2.11]	0.63 [0.37, 1.06]	1.38 [0.84, 2.26]	1.29 [0.72, 2.31]
Long term^c	1.38*** [1.17, 1.63]	1.87* [1.11, 3.15]	1.40 [0.71, 2.79]	2.08* [1.20, 3.59]	1.31 [0.74, 2.31]	0.73 [0.43, 1.22]	1.57 [0.84, 2.93]	3.00*** [1.73, 5.19]
Divorce duration categories
Short term^d	0.97 [0.79, 1.20]	0.53 [0.25, 1.11]	0.93 [0.53, 1.63]	1.10 [0.59, 2.06]	1.26 [0.70, 2.29]	1.33 [0.69, 2.58]	0.85 [0.43, 1.67]	0.90 [0.46, 1.80]
Long term^e	1.21* [1.02, 1.42]	1.12 [0.56, 2.24]	1.37 [0.70, 2.67]	1.64 [0.91, 2.96]	2.21* [1.14, 4.28]	0.88 [0.45, 1.70]	0.98 [0.50, 1.92]	1.20 [0.60, 2.39]
Never married	1.29 [0.96, 1.72]	1.31 [0.56, 3.07]	1.66 [0.82, 3.38]	2.36* [1.08, 5.14]	2.00 [0.98, 4.10]	1.10 [0.50, 2.43]	0.84 [0.32, 2.20]	1.15 [0.43, 3.11]
LR χ²	85.71***	54.08***	21.94	57.84***	36.18**	109.06***	75.92***	105.64***
Pseudo R²	.02	.03	.01	.04	.02	.07	.05	.07

Source. National Social Life, Health, and Aging Project (NSHAP; 2005/2006).

Note. Controlling for gender, age, race/ethnicity, and education. Duration categories are age and gender adjusted. Shown are incident rate ratios for allostatic load and odds ratios for individual biomarkers with 95% confidence intervals in parentheses. BMI = body mass index; BP = blood pressure; DHEA = dehydroepiandrosterone; CRP = c-reactive protein; HbA1c = glycosylated hemoglobin; NSHAP = National Social Life, Health, and Aging Project.

n = 108.

n = 112.

n = 104.

n = 92.

n = 78.

Significant at .05. **Significant at .01. ***Significant at .001.

Table 4.

Poisson and Logistic Regression of Allostatic Load and Individual Biomarkers on Number of Marital Transitions (N = 1,414).

	Allostatic load	High BMI	High systolic BP	High diastolic BP	High pulse rate	Low DHEA	High CRP	High HbA1c
(Ref. currently in first marriage)^a
Currently in second marriage^b	1.04 [0.83, 1.31]	0.91 [0.53, 1.59]	1.26 [0.72, 2.19]	1.15 [0.68, 1.93]	1.23 [0.69, 2.20]	0.98 [0.61, 1.56]	1.17 [0.66, 2.08]	0.76 [0.50, 1.17]
Currently in third marriage^c	1.01 [0.83, 1.24]	0.55 [0.29, 1.04]	1.14 [0.54, 2.37]	1.29 [0.73, 2.27]	1.52 [0.88, 2.61]	1.16 [0.62, 2.16]	1.15 [0.59, 2.24]	0.62 [0.23, 1.69]
Currently widowed with one prior marriage^d	1.21** [1.06, 1.39]	1.36 [0.83, 2.22]	1.46 [0.94, 2.25]	1.59 [0.94, 2.70]	1.24 [0.77, 1.97]	0.84 [0.61, 1.17]	1.34 [0.90, 1.99]	1.47 [0.91, 2.35]
Currently widowed with multiple prior marriages^e	1.29** [1.08, 1.54]	1.46 [0.84, 2.52]	1.46 [0.67, 3.17]	1.94* [1.11, 3.39]	1.20 [0.68, 2.10]	0.65 [0.36, 1.17]	2.26* [1.14, 4.49]	1.68 [0.95, 2.99]
Currently divorced with one prior marriage^f	1.15 [0.98, 1.34]	0.95 [0.49, 1.84]	1.18 [0.65, 2.15]	1.53 [0.88, 2.65]	1.68 [0.88, 3.19]	1.17 [0.64, 2.14]	1.29 [0.73, 2.29]	0.85 [0.46, 1.57]
Currently divorced with multiple prior marriages^g	0.97 [0.75, 1.26]	0.38 [0.14, 1.00]	1.21 [0.68, 2.52]	1.18 [0.52, 2.70]	2.00 [0.97, 4.12]	1.05 [0.46, 2.40]	0.47 [0.19, 1.19]	1.08 [0.43, 2.69]
Never married	1.30 [0.99, 1.71]	1.24 [0.54, 2.86]	1.78 [0.87, 3.62]	2.48* [1.16, 5.32]	2.16* [1.08, 4.34]	1.11 [0.49, 2.51]	0.89 [0.34, 2.31]	1.04 [0.39, 2.75]
LR χ²	78.71***	55.55***	23.54***	54.32***	32.33***	97.87***	81.81***	96.64***
Pseudo R²	.02	.03	.01	.03	.02	.06	.05	.06

Source. National Social Life, Health, and Aging Project (NSHAP; 2005/2006).

Note. Controlling for gender, age, race/ethnicity, and education. Shown are incident rate ratios for allostatic load and odds ratios for individual biomarkers with 95% confidence intervals in parentheses. BMI = body mass index; BP= Blood pressure; DHEA = dehydroepiandrosterone; CRP = c-reactive protein; HbA1c = glycosylated hemoglobin; NSHAP = National Social Life, Health, and Aging Project.

n = 598.

n = 204.

n = 72.

n = 245.

n = 79.

n = 117.

n = 53.

Significant at .05. **Significant at .01. ***Significant at .001.

Results

Bivariate Statistics

Figure 1 presents the overall distribution of the allostatic load scores by marital status. Most of the sample is located at the lower end of the distribution. About 50% of the currently married have an allostatic load score of 0 or 1, followed by the currently widowed (40%), currently divorced (38%), and never married (32%). Differences in mean allostatic load scores by current marital status are statistically significant (not shown) with the currently married having the lowest allostatic load scores (M = 1.60), followed by the currently widowed and divorced (M = 1.94 for both), and the never married (M = 2.20; F = 8.62; p < .001).

Figure 1.

Percentage distributions of allostatic load by marital status.

Multivariate Statistics

Table 2 shows that the currently widowed, on average, exhibit an estimated 22% times greater rate in the expected allostatic load count than the currently married (p < .001) when controlling for background factors. Surprisingly, when holding constant background factors, the currently divorced and the never married do not significantly differ from the currently married adults on allostatic load. Supplemental Wald tests indicate differences are most pronounced between the currently widowed and the currently married; the currently widowed do not significantly differ from either the currently divorced or the never married on allostatic load. For individual biomarkers, marital status is unrelated to BMI, systolic blood pressure, pulse rate, and DHEA. Whereas the currently divorced do not significantly vary from the currently married, the currently widowed exhibit higher odds of cardiovascular (diastolic blood pressure) and metabolic over activity (HbA1c) relative to the currently married. Similarly, the never married have 2.34 times greater odds of falling in the high risk category for diastolic blood pressure than the currently married (p < .05).

Table 3 lends support to the chronic strain model. Column 1 shows that those who report recent marital disruptions do not significantly differ from the currently married on allostatic load or individual mediators. Furthermore, both the currently widowed and currently divorced who report being in their marital status for the longest time have a significantly greater likelihood of elevated allostatic load than the currently married. Respondents who have been widowed for the longest time have a higher likelihood of falling in the high risk category for BMI, diastolic blood pressure, and HbA1c relative than the currently married. Respondents who report being divorced for the longest time also have 2.21 greater odds of falling in the high risk category for pulse rate than married respondents (p < .05).

Table 4 takes into account both current marital status and number of prior marital transitions.¹ Findings reveal that the effect of number of prior marital transitions is dependent on current marital status. For the currently divorced, marital status duration is more important for biological risk than past number of marital transitions. For the currently widowed, both respondents with one prior marriage and respondents with two or more prior marriages exhibit a significantly greater rate in the expected allostatic load count than respondents in their first marriage (21% and 29%, respectively). The currently widowed with multiple prior marriages are also more likely to score in the high risk category for diastolic blood pressure and CRP than the currently married in their first marriage.

Supplemental Analyses

In these analyses, the sample is separated by marital status. Among the currently widowed, results support the chronic strain model: Those widowed for the longest amount of time have significantly higher allostatic load and greater odds of being in the high risk HbA1c category compared with the recently widowed. For the currently divorced, results support the crisis model: Respondents who report being divorced the longest amount of time have 63% lower odds of being in the high risk category for DHEA, a primary mediator of allostatic load, than respondents with more recent divorce transitions. Supporting the findings presented above, there are no significant differences in biological risk by number of marital transitions within each marital status group. Additional analyses examine the moderating role of gender for marital status differences in biological risk. Gender was not a significant factor in the relationship between marital status and allostatic load; however, divorced women were at greater risk for elevated systolic blood pressure and divorced women at risk for under activity of DHEA than their male counterparts.

Conclusion

Allostatic load or cumulative biological risk is shaped by many factors throughout the life course. The aim of this study was to explore the role of marital loss, and it was hypothesized that cumulative biological risk and individual biomarkers would vary by type of marital disruption, marital loss duration, and number of prior marriages. Partial support was found for Hypothesis 1 in that relative to the currently married, the currently widowed (but not divorced) have elevated allostatic load, mostly through over activity of cardiovascular and metabolic markers. This supports findings on individual mediators of allostatic load, which show that the currently unmarried tend to have compromised cardiovascular and metabolic functioning compared with the currently married (Das, 2013; McFarland et al., 2013). Adding to the literature, the current study demonstrates the close link between widowhood in particular and biological risk.

Support was also found for the chronic strain model for widowhood and divorce (Hypothesis 2), in that only those widowed or divorced for the longest time had elevated allostatic load relative to the current married. These findings can be attributed to (a) accumulation of life stressors, (b) the timing of life events, and (c) selection processes. First, accumulation of stressors is a potential mechanism for understanding cumulative exposure processes and marital status differences in health. Stressors in one life domain likely influence other life domains, and primary stressors lead to secondary stressors (Pearlin, Schieman, Fazio, & Meersman, 2005). As an illustration, widowhood can result in financial strain, which hinders access to health-promoting resources. This is especially evident among women. Over time, cumulative exposure to stressful life events or a depletion of marital resources may account for the poor health observed for those widowed for the longest duration.

Second, the timing of life events is a potential explanation for the current findings. In the results presented, respondents who have been widowed for a long time tend to have experienced this marital loss transition in midlife. Spousal death in midlife is less expected than at older ages. Unexpected or unanticipated stressors may be especially devastating to health and functioning (Carr & Utz, 2001; Pearlin, 1980), particularly at a critical time in the life course for maintaining multiple work and family responsibilities (e.g., children growing up and moving out, parents becoming sick, saving for retirement). At older ages, adults may be better able to mobilize social networks such as moving in with adult children to combat the effects of marital disruptions effects on well-being (Stone, Evandrou, & Falkingham, 2013), an important precursor to physiological functioning.

Third, these findings likely reflect selection processes. Theories of stress reactivity highlight the role of individual physiological susceptibility to life stressors (Cacioppo et al., 1998). Healthier adults or those with low stress reactivity may select into additional marriages, leaving those who are unable to recover from becoming divorced or widowed in their marital status for a longer time (McFarland et al., 2013). Conversely, high stress reactivity or persistent biological response to stressors coupled with accumulation of stressors and depletion of key resources is likely to compromise functioning. Therefore, (a) determining whether individual physiological susceptibility to life stressors or chronic strains influences remarriage and (b) uncovering the unfolding of life strains that increase vulnerability to biological risk following marital disruption are important areas for future research.

Findings among the currently divorced also lend support for the crisis model. The recently divorced are more likely to display under activity of salivary DHEA, a primary mediator of allostatic load and steroid hormone. Although previous research has shown that DHEA responds to acute psychosocial stressors (Lennartsson et al., 2012), less attention has been paid to DHEA and marital disruption. DHEA is associated with decreased mobility in late life (Goldman & Glei, 2007) and increased risk for cardiovascular mortality (Ohlsson et al., 2015). DHEA/S influences glucose metabolic, systematic inflammation, and vascular health (Ohlsson et al., 2015). Results presented in the current study indicate that this is a potential biological mediator linking divorce to disease endpoints and premature mortality.

Overall, support for the third hypothesis (Hypothesis 3) was limited. The current results support findings that remaining divorced rather than being divorced accounts for observed differences in health between the currently married and currently divorced (Sbarra & Nietert, 2009). This finding likely represent selection processes with the healthiest individuals being able to marry multiple times, and reestablish marital resources. For the currently widowed, relative to older adults in their first marriage, those with multiple prior marriages demonstrate an even greater risk for high allostatic load than the currently widowed with one marital disruption. Two of the main drivers are blood pressure and CRP, a marker of inflammation. Contributing to the findings presented to Sbarra (2009) showing that marriage was protective of CRP among men in the NSHAP, the current research indicates that multiple marital transitions along with spousal death are influential on inflammation. Psychological distress, perceived stress, or decrements in social resources, especially, social support may account for these findings.

Some limitations are worth mentioning. First, the data are cross-sectional and uncovering individual change in these processes will lead to more causal arguments between marital disruption and biological risk. Second, due to the size of the sample, modeling the effects of marital loss duration and number of marital transitions jointly on biological risk was not possible. For example, among those who are currently widowed, do previous marital losses along with longer widowhood durations put individuals at an even greater risk of cumulative physiological dysregulation? Third, the current indicator of allostatic load is heavily based on surrogate endpoints and cardiovascular components. More primary mediators of allostatic load used in prior research, for example, noradrenalin, adrenalin, and cortisol excretion are needed to better specify allostatic load.

Overall, this study contributes to the extant literature by highlighting the role of marital status, marital loss duration, and marital history for individual mediators and allostatic load in late life. In general, this study demonstrates that relationship between marital status and biological risk is highly dependent on contextual aspects of marital statuses, especially their duration. It also reveals the pathways through which divorce and widowhood “get under the skin” with long-term widowhood influencing inflammation, metabolic, and cardiovascular parameters, and divorce influencing neuroendocrine and cardiovascular processes. Given the changing societal norms around marriage, future research will need to incorporate contextual aspects of domestic partnership transitions to better understand successful aging processes.

Footnotes

Appendix

Cut Points for Upper Risk Quartiles.

BMI (body mass index):

>32.14 kg/m²

Systolic blood pressure:

>149.40 mmHg

Diastolic blood pressure:

>88.40 mmHg

Pulse rate:

>78.50 bpm

Dehydroepiandrosterone (DHEA):

<18.70 pg/mL

C-reactive protein (CRP):

>3.54 mg/L

Glycosylated hemoglobin (HbA1c):

>6.25%

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Barrett

A. E.

(2000). Marital trajectories and mental health. Journal of Health and Social Behavior, 41, 451-464.

Barrett

A. E.

(2005). Gendered experiences in midlife: Implications for age identity. Journal of Aging Studies, 19, 163-183.

Bird

C. E.

Rieker

P. P.

(2008). Gender and health: The effects of constrained choices and social policies. Cambridge, UK: Cambridge University Press.

Booth

Amato

(1991). Divorce and psychological stress. Journal of Health and Social Behavior, 32, 396-407.

Brooks

K. P.

Gruenewald

Karlamangla

Koretz

Seeman

T. E.

(2014). Social relationships and allostatic load in the MIDUS study. Health Psychology, 33, 1373-1381.

Brown

S. L.

Lin

I.-F.

(2012). The gray divorce revolution: Rising divorce among middle-aged and older adults, 1990-2010. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 67, 731-741.

Cacioppo

J. T.

Berntson

G. G.

Malarkey

W. B.

Kiecolt-Glaser

J. K.

Sheridan

J. F.

Poehlmann

K. M.

. . . Glaser

(1998). Autonomic, neuroendocrine, and immune responses to psychological stress: The reactivity hypothesis. Annals of the New York Academy of Sciences, 840, 664-673.

Carr

Springer

K. W.

(2010). Advances in families and health research in the 21st century. Journal of Marriage and Family, 72, 743-761.

Carr

Utz

(2001). Late-life widowhood in the United States: New directions in research and theory. Ageing International, 27(1), 65-88.

10.

Crimmins

E. M.

Johnston

Hayward

Seeman

(2003). Age differences in allostatic load: An index of physiological dysregulation. Experimental Gerontology, 38, 731-734.

11.

Crimmins

E, M.

Kim

J. K.

Seeman

T. E.

(2009). Poverty and biological risk: The earlier “aging” of the poor. The Journals of Gerontology, Series A: Biological Sciences & Medical Sciences, 64, 286-292.

12.

Das

(2013). Spousal loss and health in late life moving beyond emotional trauma. Journal of Aging and Health, 25, 221-242.

13.

Donoho

C. J.

Seeman

T. E.

Sloan

R. P.

Crimmins

E. M.

(2015). Marital status, marital quality, and heart rate variability in the MIDUS cohort. Journal of Family Psychology, 29, 290-295.

14.

Fincham

F. D.

Beach

S. R.

(2010). Marriage in the new millennium: A decade in review. Journal of Marriage and Family, 72, 630-649.

15.

Gruenewald

T. L.

Seeman

T. E.

Karlamangla

A. S.

Sarkisian

C. A.

(2009). Allostatic load and frailty in older adults. Journal of the American Geriatrics Society, 57, 1525-1531.

16.

Gersten

(2008). Neuroendocrine biomarkers, social relations, and the cumulative costs of stress in Taiwan. Social Science & Medicine, 66, 507-519.

17.

Goldman

Glei

D. A.

(2007). Sex differences in the relationship between DHEAS and health. Experimental Gerontology, 42, 979-987.

18.

Hill

T. D.

Reid

Reczek

(2013). Marriage and the mental health of Low-income urban women with children. Journal of Family Issues, 34, 1238-1261.

19.

Holt-Lunstad

Birmingham

Jones

B. Q.

(2008). Is there something unique about marriage? The relative impact of marital status, relationship quality, and network social support on ambulatory blood pressure and mental health. Annals of Behavioral Medicine, 35, 239-244.

20.

Hughes

M. E.

Waite

L. J.

(2009). Marital biography and health at mid-life. Journal of Health and Social Behavior, 50, 344-358.

21.

Johnson

D. R.

(2002). An empirical test of crisis, social selection, and role explanations of the relationship between marital disruption and psychological distress: A pooled time-series analysis of four-wave panel data. Journal of Marriage and Family, 64, 211-224.

22.

Juster

R.-P.

McEwen

B. S.

Lupien

S. J.

(2010). Allostatic load biomarkers of chronic stress and impact on health and cognition. Neuroscience & Biobehavioral Reviews, 35, 2-16.

23.

Karlamangla

A. S.

Singer

B. H.

Seeman

T. E.

(2006). Reduction in allostatic load in older adults is associated with lower all-cause mortality risk: MacArthur studies of successful aging. Psychosomatic Medicine, 68, 500-507.

24.

Kiecolt-Glaser

J. K.

Bane

Glaser

Malarkey

W. B.

(2003). Love, marriage, and divorce: Newlyweds’ stress hormones foreshadow relationship changes. Journal of Consulting and Clinical Psychology, 71, 176-188.

25.

Kiecolt-Glaser

J. K.

Fisher

L. D.

Ogrocki

Stout

J. C.

Speicher

C. E.

Glaser

(1987). Marital quality, marital disruption, and immune function. Psychosomatic Medicine, 49, 13-34.

26.

Kreider

Ellis

(2011). Number, timing, and duration of marriages and divorces: 2009 (Current Population Reports, P70-125). Retrieved from http://www.census.gov/prod/2011pubs/p70-125.pdf

27.

Lavelle

Smock

P. J.

(2012). Divorce and women’s risk of health insurance loss. Journal of Health and Social Behavior, 53, 413-431.

28.

Lennartsson

A. K.

Kushnir

M. M.

Bergquist

Jonsdottir

I. H.

(2012). DHEA and DHEA-S response to acute psychosocial stress in healthy men and women. Biological Psychology, 90, 143-149.

29.

Liu

Umberson

D. J.

(2008). The times they are a changin’: Marital status and health differentials from 1972 to 2003. Journal of Health and Social Behavior, 49, 239-253.

30.

Long

J. S.

Freese

(2006). Regression models for categorical dependent variables using Stata. College Station, TX: Stata Press.

31.

Lorenz

F. O.

Wickrama

Conger

R. D.

Elder

G. H.

(2006). The short-term and decade-long effects of divorce on women’s midlife health. Journal of Health and Social Behavior, 47, 111-125.

32.

McEwen

B. S.

(1998). Stress, adaptation, and disease: Allostasis and allostatic load. Annals of the New York Academy of Sciences, 840(1), 33-44.

33.

McEwen

B. S.

(2003). Interacting mediators of allostasis and allostatic load: Towards an understanding of resilience in aging. Metabolism, 52, 10-16.

34.

McEwen

B. S.

Seeman

(1999). Protective and damaging effects of mediators of stress: Elaborating and testing the concepts of allostasis and allostatic load. Annals of the New York Academy of Sciences, 896(1), 30-47.

35.

McEwen

B. S.

Stellar

(1993). Stress and the individual: Mechanisms leading to disease. Archives of Internal Medicine, 153, 2093-2101.

36.

McEwen

B. S.

Wingfield

J. C.

(2003). The concept of allostasis in biology and biomedicine. Hormones and Behavior, 43, 2-15.

37.

McFarland

M. J.

Hayward

M. D.

Brown

(2013). I’ve got you under my skin: Marital biography and biological risk. Journal of Marriage and Family, 75, 363-380.

38.

Ohlsson

Vandenput

Tivesten

(2015). DHEA and mortality: What is the nature of the association? The Journal of Steroid Biochemistry and Molecular Biology, 145, 248-253.

39.

Pearlin

L. I.

(1980). Life strains and psychological distress among adults. In Smelser

N. J.

Erikson

E. H.

(Eds.), Themes of work and love in adulthood (pp. 174-192). Cambridge, MA: Harvard University Press.

40.

Pearlin

L. I.

Schieman

Fazio

E. M.

Meersman

S. C.

(2005). Stress, health, and the life course: Some conceptual perspectives. Journal of Health and Social Behavior, 46, 205-219.

41.

Robles

T. F.

Kiecolt-Glaser

J. K.

(2003). The physiology of marriage: Pathways to health. Physiology & Behavior, 79, 409-416.

42.

Sasson

Umberson

D. J.

(2014). Widowhood and depression: New light on gender differences, selection, and psychological adjustment. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 69, 135-145.

43.

Sbarra

D. A.

(2009). Marriage protects men from clinically meaningful elevations in C-reactive protein: Results from the National Social Life, Health, and Aging Project (NSHAP). Psychosomatic Medicine, 71, 828-835.

44.

Sbarra

D. A.

Nietert

P. J.

(2009). Divorce and death forty years of the Charleston Heart Study. Psychological Science, 20, 107-113.

45.

Seeman

T. E.

Glei

Goldman

Weinstein

Singer

Lin

Y.-H.

(2004). Social relationships and allostatic load in Taiwanese elderly and near elderly. Social Science & Medicine, 59, 2245-2257.

46.

Seeman

T. E.

Singer

B. H.

Rowe

J. W.

Horwitz

R. I.

McEwen

B. S.

(1997). Price of adaptation—Allostatic load and its health consequences: MacArthur studies of successful aging. Archives of Internal Medicine, 157, 2259-2268.

47.

Seeman

T. E.

Singer

B. H.

Ryff

C. D.

Love

G. D.

Levy-Storms

(2002). Social relationships, gender, and allostatic load across two age cohorts. Psychosomatic Medicine, 64, 395-406.

48.

Shor

Roelfs

D. J.

Curreli

Clemow

Burg

M. M.

Schwartz

J. E.

(2012). Widowhood and mortality: A meta-analysis and meta-regression. Demography, 49, 575-606.

49.

Simon

R. W.

(2002). Revisiting the relationships among gender, marital status, and mental health. American Journal of Sociology, 107, 1065-1096.

50.

Smith

Jaszczak

Graber

Lundeen

Leitsch

Wargo

O’Muircheartaigh

(2009). Instrument development, study design implementation, and survey conduct for the national social life, health, and aging project. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 64(Suppl. 1), i20-i29.

51.

Sterling

Eyer

(1988). Allostasis: A new paradigm to explain arousal pathology. In Fisher

Reason

(Eds.), Handbook of life stress, cognition, and health (pp. 629-649). Oxford, UK: John Wiley.

52.

Stone

Evandrou

Falkingham

(2013). The transition to living alone and psychological distress in later life. Age and Ageing, 42, 366-372.

53.

Umberson

Liu

Powers

(2009). Marital status, marital transitions, and body weight. Journal of Health and Social Behavior, 50, 327-343.

54.

Umberson

Williams

Powers

D. A.

Liu

Needham

(2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47, 1-16.

55.

Umberson

Wortman

C. B.

Kessler

R. C.

(1992). Widowhood and depression: Explaining long-term gender differences in vulnerability. Journal of Health and Social Behavior, 33, 10-24.

56.

Waite

Gallagher

(2000). The case for marriage: Why married people are healthier, happier, and better-off financially. New York, NY: Doubleday.

57.

Wang

(2014). Record share of Americans have never married. Retrieved from http://www.pewsocialtrends.org/2014/09/24/record-share-of-americans-have-never-married/

58.

Wang

Parker

(2014). Record share of Americans have never married. Retrieved from http://www.pewsocialtrends.org/2014/09/24/record-share-of-americans-have-never-married/

59.

Weinstein

Goldman

Hedley

Yu-Hsuan

Seeman

(2003). Social linkages to biological markers of health among the elderly. Journal of Biosocial Science, 35, 433-453.

60.

Williams

Umberson

(2004). Marital status, marital transitions, and health: A gendered life course perspective. Journal of Health and Social Behavior, 45, 81-98.

61.

Williams

S. R.

McDade

T. W.

(2009). The use of dried blood spot sampling in the national social life, health, and aging project. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 64(Suppl. 1), i131-i136.

62.

Zhang

Hayward

M. D.

(2006). Gender, the marital life course, and cardiovascular disease in late midlife. Journal of Marriage and Family, 68, 639-657.

63.

Zick

C. D.

Holden

(2000). An assessment of the wealth holdings of recent widows. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 55, S90-S97.