Abstract
Background
Lower physical activity has been cross-sectionally associated with greater perceived fatigability, defined as self-reported fatigue anchored to activity intensity and duration. The temporality of this relationship, and whether it differs by activity type or marital status, remains unclear.
Methods
In the Osteoporotic Fractures in Men Study (N = 1 759), self-reported total, exercise, and household activity were assessed using the Physical Activity Scale for the Elderly across 7 visits (2000–2016). The Pittsburgh Fatigability Scale (range: 0–50; higher scores = greater fatigability) measured physical (mean = 16.6 ± 9.7) and mental (mean = 7.8 ± 8.3) fatigability at Year 14. Least absolute deviation and linear regression were used to examine associations between baseline and change in activity over 14 years with subsequent fatigability. Models were adjusted for demographic, health, and lifestyle factors.
Results
After adjustment, lower baseline (β= −0.08, 95% confidence interval [CI]: −0.12, −0.04) and greater annual declines in total activity (β = −0.09, 95% CI: −0.14, −0.05) were prospectively associated with higher Pittsburgh Fatigability Scale (PFS) Physical scores. Associations were similar for mental fatigability (both p < .05). Lower baseline leisure exercise, but not baseline household activity, predicted higher PFS Physical scores (β = −0.10 vs −0.001). In contrast, greater declines in household activity, but not declines in exercise, were associated with higher PFS Physical scores (β = −0.09 vs −0.03). Lower baseline household activity predicted higher PFS Mental scores only for unmarried men (β = −0.15, 95% CI: −0.29, −0.01, interaction p = .019).
Conclusions
Baseline total activity and leisure exercise, and declines in total and household activity, were associated with higher subsequent perceived fatigability in older men. Marital status may mitigate the contribution of household activity to subsequent fatigability.
Keywords: Disablement process, Epidemiology, Exercise, Fatigue, Psychosocial
Fatigue, a subjective lack of physical or mental energy to complete tasks, is a common complaint among older adults (1). Both worsening fatigue and declines in physical activity contribute to functional limitations in later life (2–4). Yet, the mechanisms by which lower levels of physical activity contribute to fatigue and the temporality of this relationship remain unclear. Prior studies have found that fatigue may reduce the amount and alter patterns of physical activity in later life (5,6). Others have found that maintained physical activity may protect against fatigue (7,8). A limitation of these studies is that they rely on global measures of fatigue, which are prone to self-pacing bias that prevents meaningful comparisons between individuals of differing health status (1,9).
Perceived fatigability has emerged as a novel measure of whole-body or cognitive fatigue standardized to activities of fixed intensity and duration (1,3,10–12). Importantly, perceived fatigability minimizes self-pacing bias and facilitates comparison across samples with different functional health (1,9). Perceived fatigability has been shown to be a more sensitive measure of functional declines and clinical outcomes than other fatigue measures (3,10,13–15) and is therefore a key marker of physiologic aging (9).
Physical activity is associated with perceived fatigability (4,16,17), but few studies examined the directionality of the relationship. In a cross-sectional statistical mediation analysis of the Long Life Family Study, Qiao et al. (4) found that greater perceived physical fatigability explained the association between lower physical activity and slower gait speed for both older participants (ie, probands; 22.5%) and their offspring (39.5%). Declines in physical activity may therefore precede worsening fatigability in the broader disablement pathway (18). This relationship likely contributes to a vicious cycle (19), where individuals with worsening fatigability adaptively self-pace and further reduce levels of physical activity to maintain tolerable levels of fatigue. Further declines in physical activity may reduce aerobic capacity (ie, fitness), reducing total energy available and leading to more severe fatigue.
Importantly, several research gaps remain that could provide further insights into the direction and timing of the association between physical activity and perceived fatigability. First, longitudinal studies are needed to properly establish temporality of the relationship and replicate findings from cross-sectional studies (20). Furthermore, it is unknown whether physical activity also predicts mental fatigability. Perceived mental fatigability is a distinct subdomain of fatigability linked with poorer cognitive health and depressive symptoms (12,21), but like physical fatigability, has also been associated with mobility in later life (22).
There are also no studies that examine whether the association between physical activity and perceived fatigability differs by the type of activity, such as leisure exercise (eg, sports, recreation) and household activity (eg, household chores, yardwork). Older adults generally get more of their physical activity from lower intensity, daily tasks (eg, errands, housework) than formal recreation (23,24). It may therefore be clinically relevant to establish a prospective link between household activity and fatigability, as this type of activity may serve as an additional intervention target for community-dwelling older populations (25). Examining differences by activity type may further inform on the timing of the prospective association with fatigability, as declines in leisure exercise are likely to precede declines in household activity across the life course (24).
Finally, marital status may play an important modifying role in how physical activity relates to fatigability. Being married has been associated with lower fatigue in older adults (6). The mechanism for this protective relationship is still unknown but may be due in part to spousal support in motivating activity (26) or managing life stressors (27). For example, married men have been shown to receive more instrumental support in completing household tasks than married women (28). This spousal support may reduce energy utilized for household tasks for married men compared to unmarried men and thus attenuate the association between household activity and fatigability for the married group. To the best of our knowledge, no study has formally evaluated effect modification of marital status on the relationship between physical activity (including housework) and fatigue or fatigability.
To address these gaps, this study examined whether self-reported physical activity across 14 years of follow-up predicted subsequent perceived fatigability in older men and whether the relationship differed by activity type (eg, total, household, leisure exercise) and marital status. We hypothesized that lower self-reported physical activity would be associated with both greater perceived physical and mental fatigability. We further hypothesized that lower leisure exercise at baseline would have a stronger prospective association with subsequent fatigability than household activity, and that lower household activity would have a stronger prospective association with perceived fatigability for unmarried versus married men.
Method
Study Population
Participants were from the Osteoporotic Fractures in Men Study (MrOS), a prospective cohort study of ambulatory, community-dwelling men aged 65 years or older (29). There were 5 994 men enrolled at baseline from March 2000 through April 2002 across 6 study sites: Birmingham, AL; Minneapolis, MN; Palo Alto, CA; Pittsburgh, PA; Portland, OR; and San Diego, CA. Recruitment and baseline characteristics of the sample were previously reported (29,30). All participants provided informed consent and study protocols were approved by the Institutional Review Boards at each study site.
The current study included men who were assessed at Year 14 (Visit 4) from May 2014 through May 2016. There were 2 424 men who finished at least some part of the Year 14 examination with 1 841 completing the clinic visit and 583 only doing the self-assessment questionnaires. The final analytic sample consisted of 1 759 men with complete data on perceived fatigability and covariate measures at Year 14 (Figure 1).
Figure 1.
Participants included in analyses. MrOS = Osteoporotic Fractures in Men Study.
Measures
Physical activity
Physical activity was measured using the self-reported Physical Activity Scale for the Elderly (PASE) (31). The PASE assessed the frequency, intensity, and duration of physical activities during the previous 7 days. Leisure exercise activities included walking outside the home; light, moderate, and strenuous sports; and muscle strength or endurance exercise. Household activities included light or heavy housework/chores, home repairs, lawn work, gardening, and caring for another person. Occupational activity included working for pay or volunteering. The frequency and duration reported for each activity were multiplied by a standardized intensity weight, and scores were summed to produce a total score (31).
The PASE has high test–retest reliability and has been previously validated against objectively measured physical activity (32,33). The current analysis used the PASE total score, as well as leisure exercise and housework subscores as primary exposures. We standardized each PASE score and subscore to allow for comparison between them. PASE data from the following study visits, spanning a median of 14 years (range: 12.1–15.8) of total follow-up, were used for the current analysis: Year 0 (Visit 1: 2000–2002; n = 1 759), Year 3.5 (Sleep Visit 1: 2003–2005; n = 1 233), Year 5 (Visit 2: 2005–2006; n = 1 754), Year 7 (Visit 3: 2007–2009; n = 1 756), Year 9 (Interim Visit 2: 2009–2011; n = 1 752), Year 9.5 (Sleep Visit 2: 2009–2012; n = 695), and Year 14 (Visit 4: 2014–2016; n = 1 759).
Perceived fatigability
The Pittsburgh Fatigability Scale (PFS) was used to measure perceived fatigability at Year 14 (Visit 4) and was previously validated in community-dwelling older adults (11,12). The PFS is a 10-item scale where participants reported physical and mental fatigue (0 = “no fatigue” to 5 = “extreme fatigue”) they expected or imagined they would feel after performing activities of fixed intensity and duration (eg, “leisurely walk for 30 minutes”). Sums of responses were used to generate physical and mental fatigability subscale scores (range: 0–50, higher PFS scores = greater fatigability) (11). Established binary cutoffs for fatigability severity status were used to identify men with more severe versus less severe physical (PFS ≥15 vs <15) and mental (PFS ≥13 vs <13) fatigability (3,9,34). Scores for 8 men with missing data for 1–3 PFS items were imputed following an existing protocol (35). Scores were imputed using the mean of the individual’s valid responses, accounting for sample-specific fatigue associated with the activity and whether participants performed the activity.
Covariates
Several demographic, lifestyle, and health measures were included as potential confounders of the association between physical activity and perceived fatigability. Age, race, educational attainment, and study site were collected at baseline via questionnaire (MrOS Year 0). Marital status (ie, married, widowed, divorced, separated, or single/never married) and living arrangement (ie, living alone vs. with others) were assessed across visits (Supplementary Table 2). Men were stratified into married and unmarried (widowed, divorced, separated, single/never married) groups at baseline. Men who were unmarried at baseline (n = 235) were most likely to be either widowed (n = 84, 36%) or divorced (n = 89, 36%). In sensitivity analyses, men were further stratified by their baseline living arrangement into 2 groups: (a) living alone and (b) living with others (eg, spouse, child). There was a high concordance between marital status and living situation at baseline and across visits, where a vast majority of men who were not living alone were married (Supplementary Table 2).
All other covariates were measured at Year 14. Depressive symptoms were measured using the self-reported, short-form Geriatric Depression Scale (GDS) (36), and sleep quality was measured using the self-reported Pittsburgh Sleep Quality Index (PSQI) (37). Fall history was assessed using the item “during the past 12 months, have you fallen and landed on the floor or ground, or fallen and hit an object like a table or chair.” Total medical conditions consisted of a sum score of the following self-reported past or current doctor-diagnosed conditions: heart attack, angina, hypertension, stroke, diabetes, congestive heart failure, chronic obstructive pulmonary disease, Parkinson’s disease, and nonskin cancer (range: 0–8). Medications were brought to the clinic visit and matched to ingredients based on the Iowa Drug Information Service Drug Vocabulary (38). A sum of any medications (prescription, over the counter, or vitamins) taken within the past 30 days was calculated. The Teng Modified Mini-Mental State Examination (3MSE; range: 0–100) was administered during the clinic visit to assess global cognition (39). Height and weight were also assessed at the clinic visit to calculate body mass index (BMI, kg/m2).
Statistical Approach
Descriptive statistics for each covariate were generated for the overall sample and by binary physical and mental fatigability cutoffs. Differences by fatigability severity status were tested using Wilcoxon tests for continuous variables and chi-square tests for categorical variables. Spaghetti plots and plots of average PASE measures with a loess smoother across visits were used to explore longitudinal changes in physical activity by fatigability severity status.
Person-level intercepts and slopes were generated for each PASE measure using least absolute deviation regression. Least absolute deviation was used rather than least squares regression to account for large within-person fluctuations in activity during the follow-up period (40). To construct the models, we first performed a repeated-measures least absolute deviation regression for each participant separately to account for within-person correlations in PASE measures over time. Years of follow-up were included as the predictor and the PASE measure as the outcome. Models were conducted separately for each PASE measure (total, exercise, household). From these models, we estimated the within-person intercept and slope of the PASE measures for each participant. Given the PASE measurement at Year 14 overlapped with the measurement of perceived fatigability, in a sensitivity analysis, we also derived person-specific PASE intercepts and slopes without the Year 14 PASE measurement to better establish temporal precedence of physical activity changes on perceived fatigability.
Multiple linear regression was then used to evaluate the association between person-level PASE intercepts and slopes and PFS Physical and Mental scores. For each set of models, Model 1 included terms for baseline PASE (person-level intercept) or annual change in PASE (person-level slope). Model 2 was further adjusted for baseline age (years), study site, education (≤high school, some college/college degree, some graduate/graduate degree), race (White vs non-White), as well as Visit 4 marital status (married vs unmarried), number of medical conditions, number of medications, self-rated health (excellent/very good/good vs fair/poor), significant depressive symptoms (GDS >6), poor sleep quality (PSQI >5), 3MS, and BMI. Linear regression assumptions were assessed using QQ plots of residuals (ie, normality of errors) and scatterplots of residuals versus fitted values (ie, constant error variance) (41).
Moderation by marital status was evaluated by including interactions between PASE and marital status in multiple linear regressions. Model 1 included terms for baseline PASE (person-level intercept) or annual change in PASE (person-level slope) and their interactions with marital status at baseline (unmarried, married). Model 2 was further adjusted for the above covariates, excluding Visit 4 marital status. We also examined moderation by baseline living status using an identical modeling procedure.
Results
Differences in Participant Characteristics by Perceived Fatigability
Fifty-six percent (n = 989) of the sample had more severe physical fatigability (PFS ≥15), whereas 24% (n = 426) had more severe mental fatigability (PFS ≥13; Table 1). Men with more severe physical or mental fatigability were on average older, had lower global cognition (3MS), and had a higher number of medical conditions and medications (all p < .05 for both physical and mental fatigability). These men also had poorer sleep quality and were more likely to report significant depressive symptoms and poor or fair self-rated health (all p < .05 for both physical and mental fatigability). Men with more severe physical fatigability had a higher mean BMI. The prevalence of perceived physical but not mental fatigability differed across study sites (p < .05). Men with more perceived mental fatigability were also more likely to live alone (p < .05), but there were no differences in living status by perceived physical fatigability. There were no differences in education or race by perceived physical or mental fatigability severity status (Table 1). Perceived physical fatigability was moderately correlated with perceived mental fatigability (ρ = .45).
Table 1.
Participant Characteristics Stratified by Perceived Fatigability Status at Visit 4
| Physical Fatigability | Mental Fatigability | ||||||
|---|---|---|---|---|---|---|---|
| Measures | Overall (N = 1 759) | More (PFS ≥15) (n = 989) | Less (PFS <15) (n = 770) | More (PFS ≥13) (n = 426) | Less (PFS <13) (n = 1 333) | ||
| M (SD) | M (SD) | M (SD) | p | M (SD) | M (SD) | p | |
| PFS Physical score | 16.6 (9.7) | ||||||
| PFS Mental score | 7.8 (8.3) | ||||||
| Age (years) | 84.4 (4.2) | 85.1 (4.4) | 83.5 (3.7) | <.001 | 85.2 (4.6) | 84.1 (4.0) | <.001 |
| BMI | 26.9 (3.7) | 27.2 (3.9) | 26.5 (3.5) | <.001 | 27 (3.8) | 26.8 (3.7) | .463 |
| 3MSE | 92.0 (7.4) | 91.3 (7.9) | 92.8 (6.5) | <.001 | 90.7 (8.5) | 92.4 (6.9) | <.001 |
| Number of medical conditions | 1.5 (1.2) | 1.6 (1.3) | 1.3 (1.1) | <.001 | 1.7 (1.4) | 1.4 (1.2) | .002 |
| Number of medications | 9.1 (5.2) | 9.8 (5.6) | 8.3 (4.5) | <.001 | 10.2 (6.6) | 8.8 (4.7) | <.001 |
| N (%) | N (%) | N (%) | p | N (%) | N (%) | p | |
| White race | 1 596 (91) | 904 (91) | 692 (90) | .308 | 389 (91) | 1 207 (91) | .705 |
| Baseline marital status | |||||||
| Married | 1 524 (87) | 849 (86) | 675 (88) | .004 | 367 (86) | 1 157 (87) | .808 |
| Widowed | 84 (5) | 61 (6) | 23 (3) | 19 (4) | 65 (5) | ||
| Separated | 12 (1) | 10 (1) | 2 (0) | 4 (1) | 8 (1) | ||
| Divorced | 89 (5) | 44 (4) | 45 (6) | 21 (5) | 68 (5) | ||
| Single, never married | 50 (3) | 25 (3) | 25 (3) | 15 (4) | 35 (3) | ||
| Baseline living situation (living with others vs alone) | 1 594 (91) | 887 (90) | 707 (92) | .150 | 375 (88) | 1 219 (91) | .044 |
| Self-rated health (good/excellent vs poor/fair) | 1 570 (89) | 828 (84) | 742 (96) | <.001 | 346 (81) | 1 224 (92) | <.001 |
| Poor sleep quality (PSQI >5) | 708 (40) | 455 (46) | 253 (33) | <.001 | 232 (54) | 476 (36) | <.001 |
| Significant depressive symptoms (GDS >6) | 112 (6) | 93 (9) | 19 (2) | <.001 | 65 (15) | 47 (4) | <.001 |
| Fell in the past month* | 673 (38) | 436 (44) | 237 (31) | <.001 | 203 (48) | 470 (35) | <.001 |
Notes: PFS = Pittsburgh Fatigability Scale; M = mean; SD = standard deviation; GDS = Geriatric Depression Scale; PSQI = Pittsburgh Sleep Quality Index; 3MSE = Teng Modified Mini-Mental State Examination. P values are for Wilcoxon rank-sum tests (continuous variables) and chi-square tests (categorical variables) for differences between fatigability severity status.
*During the past 12 months, reported falling and landing on the floor or ground or falling and hitting an object like a table or chair.
Compared to excluded participants, individuals included in the analytic sample were more likely to be younger, White, better educated, and have fewer depressive symptoms (all p < .05; Figure 1; Supplementary Table 1). Included men also had lower prevalence of significant depressive symptoms and poor sleep quality compared to excluded men (all p < .05).
Longitudinal Associations Between Physical Activity and Perceived Fatigability
Figure 2 displays average estimates with 95% confidence intervals (CIs) for each physical activity measure across visits and stratified by fatigability severity status at Year 14. PASE total, leisure exercise, and household activity declined on average across follow-up. However, those who had more severe fatigability at Year 14 tended to have greater declines for each PASE measure across prior visits. These individuals also had lower average baseline scores for total activity and leisure exercise, but not for household activity. Differences in activity declines by fatigability severity status appeared to be greater for household activity compared to leisure exercise (Figure 2).
Figure 2.
Change in standardized PASE scores across visits stratified by perceived fatigability severity at Visit 4. Note: Plots are of mean standardized scores with a loess smoother across visits. PFS = Pittsburgh Fatigability Scale; PASE = Physical Activity Scale for the Elderly. Visit 1 (V1): n = 1 759, Sleep Visit 1 (SV1): n = 1 233, Visit 2 (V2): n = 1 754, Visit 3 (V3): n = 1 756, Interim Visit 2 (VI2): n = 1 742, Sleep Visit 2 (SV2): n = 695, Visit 4 (V4): n = 1 759.
In the adjusted model, each standard deviation (SD) higher baseline PASE total score was associated with −0.08 SD lower PFS Physical scores at Year 14 (95% CI: −0.12, −0.04, p < .001, Table 2). Furthermore, each SD increase in annual slope of PASE total score was associated with −0.09 SD lower PFS Physical scores (95% CI: −0.14, −0.05, p < .001). Associations were similar and in the same direction for perceived mental fatigability (baseline: B = −0.06, 95% CI: −0.11, −0.02, p = .005; Slope: B = −0.06, 95% CI: −0.11, −0.02, p = .008).
Table 2.
Associations Between Baseline and Annual Change in Physical Activity Across Visits and Perceived Fatigability at Visit 4
| Model 1: Unadjusted | Model 2: Adjusted | |||||
|---|---|---|---|---|---|---|
| B | 95% CI | p | B | 95% CI | p | |
| Physical fatigability | ||||||
| PASE total | ||||||
| Baseline* | −0.106 | −0.15, −0.06 | <.001 | −0.080 | −0.12, −0.04 | <.001 |
| Annual change† | −0.190 | −0.24, −0.14 | <.001 | −0.092 | −0.14, −0.05 | <.001 |
| PASE exercise | ||||||
| Baseline* | −0.123 | −0.17, −0.08 | <.001 | −0.104 | −0.15, −0.06 | <.001 |
| Annual change† | −0.104 | −0.15, −0.06 | <.001 | −0.028 | −0.07, 0.01 | .198 |
| PASE household | ||||||
| Baseline* | −0.010 | −0.06, 0.04 | .685 | −0.001 | −0.04, 0.04 | .978 |
| Annual change† | −0.166 | −0.21, −0.12 | <.001 | −0.090 | −0.13, −0.05 | <.001 |
| Mental fatigability | ||||||
| PASE total | ||||||
| Baseline* | −0.076 | −0.12, −0.03 | .001 | −0.063 | −0.11, −0.02 | .005 |
| Annual change† | −0.133 | −0.18, −0.09 | <.001 | −0.061 | −0.11, −0.02 | .008 |
| PASE exercise | ||||||
| Baseline* | −0.098 | −0.14, −0.05 | <.001 | −0.097 | −0.14, −0.05 | <.001 |
| Annual change† | −0.046 | −0.09, 0.00 | .056 | 0.016 | −0.03, 0.06 | .480 |
| PASE household | ||||||
| Baseline* | −0.025 | −0.07, 0.02 | .296 | −0.018 | −0.06, 0.03 | .423 |
| Annual change† | −0.094 | −0.14, −0.05 | <.001 | −0.041 | −0.09, 0.00 | .080 |
Notes: PASE = Physical Activity Scale for the Elderly. Person-level intercepts and slopes derived from least absolute deviation regression. Estimates reported from multiple linear regression on Pittsburgh Fatigability Scale (PFS) at Visit 4. Model 1 included terms for baseline PASE (person-level intercept) and annual change in PASE (person-level slope). Model 2 was further adjusted for baseline age, study site, education (≤high school, some college/college degree, some graduate/graduate degree), race (White vs non-White), Visit 4 marital status (married vs unmarried), number of medical conditions, number of medications, self-reported history of fall, self-rated health (excellent/very good/good vs fair/poor), significant depressive symptoms (Geriatric Depression Scale >6), poor sleep quality (Pittsburgh Sleep Quality Index >5), Modified Mini-Mental State (3MS), and BMI.
*Interpreted as the average SD difference in PFS at Visit 4 per 1 SD greater baseline PASE (ie, person-level intercept).
†Interpreted as the average SD difference in PFS per SD greater annual change in PASE score (ie, person-level slope).
Differences in Longitudinal Associations With Perceived Fatigability by Activity Type
Leisure exercise
In the adjusted model, each SD higher baseline PASE leisure exercise score was associated with −0.10 SD lower PFS Physical scores at Year 14 (95% CI: −0.15, −0.06, p < .001, Table 2). Yet, annual change in PASE leisure exercise scores was not significantly associated with PFS Physical scores (B = −0.03, 95% CI: −0.07, 0.01, p = .198). Associations were similar for perceived mental fatigability (baseline: B = −0.10, 95% CI: −0.14, −0.05, p < .001; Slope: B = 0.016, 95% CI: −0.03, 0.06, p = .480).
Household activity
In the adjusted model, baseline PASE household score was not associated with PFS Physical or Mental scores at Year 14 (p > .05 for both). Yet, each SD greater annual slope of PASE household scores was associated with a −0.09 SD lower PFS Physical scores at Year 14 (95% CI: −0.13, −0.05, p < .001). For perceived mental fatigability, this association was attenuated and not significant in the adjusted model (B = −0.04, 95% CI: −0.09, 0.00, p = .080). Results were similar after excluding PASE measured at Year 14 (Visit 4; Supplementary Table 3).
Differences in Longitudinal Associations With Perceived Fatigability by Marital Status
For men who were unmarried at baseline, each SD greater baseline PASE household score was associated with a −0.15 SD lower PFS Mental scores (95% CI: −0.29, −0.01) in the adjusted model (Table 3). This was significantly higher than for men who were married (interaction p = .019), for whom the association between baseline PASE household and PFS Mental scores was not significant (B = 0.004, 95% CI: −0.05, 0.06). A similar relationship was observed in models including baseline living situation, where only men living alone had a significant negative association between baseline PASE household score and PFS Mental scores (Supplementary Table 4 Model 2; Living alone: B = −0.21, 95% CI: −0.37, −0.05; Living with others: B = 0.005, 95% CI: −0.05, 0.06, interaction p = .005). There were no differences by baseline marital status in associations between PASE measures and PFS Physical scores (interaction p > .05 for all, Supplementary Tables 5 and 6).
Table 3.
Moderation of Baseline Marital Status on Associations Between Baseline and Annual Change in Physical Activity and Perceived Mental Fatigability at Visit 4
| Model 1: Unadjusted | Model 2: Adjusted | |||||
|---|---|---|---|---|---|---|
| B | 95% CI | p (interaction) | B | 95% CI | p (interaction) | |
| PASE total | ||||||
| Baseline* | ||||||
| Married | −0.056 | −0.11, 0.00 | .038* | −0.046 | −0.10, 0.01 | .045* |
| Unmarried | −0.207 | −0.36, −0.05 | −0.184 | −0.33, −0.04 | ||
| Annual change† | ||||||
| Married | −0.124 | −0.18, −0.07 | .269 | −0.054 | −0.11, 0.00 | .393 |
| Unmarried | −0.206 | −0.36, −0.05 | −0.114 | −0.26, 0.04 | ||
| PASE exercise | ||||||
| Baseline* | ||||||
| Married | −0.091 | −0.15, −0.03 | .383 | −0.089 | −0.14, −0.04 | .411 |
| Unmarried | −0.154 | −0.31, 0.00 | −0.146 | −0.29, 0.00 | ||
| Annual change† | ||||||
| Married | −0.042 | −0.10, 0.01 | .633 | 0.017 | −0.04, 0.07 | .867 |
| Unmarried | −0.078 | −0.24, 0.08 | 0.005 | −0.15, 0.16 | ||
| PASE household | ||||||
| Baseline* | ||||||
| Married | −0.001 | −0.06, 0.06 | .019* | 0.004 | −0.05, 0.06 | .019* |
| Unmarried | −0.164 | −0.31, −0.02 | −0.150 | −0.29, −0.01 | ||
| Annual change† | ||||||
| Married | −0.091 | −0.15, −0.03 | 0.743 | −0.041 | −0.10, 0.01 | .813 |
| Unmarried | −0.115 | −0.27, 0.04 | −0.024 | −0.18, 0.13 |
Notes: PASE = Physical Activity Scale for the Elderly. Person-level intercepts and slopes derived from least absolute deviation regression. Estimates reported from multiple linear regression on Pittsburgh Fatigability Scale (PFS) at Visit 4. Unmarried men included those who were widowed, separated, divorced, or never married. Model 1 included terms for baseline PASE (person-level intercept) and annual change in PASE (person-level slope), baseline marital status (married vs unmarried), and interactions between baseline marital status and both baseline PASE and annual change in PASE. Model 2 was further adjusted for baseline age, study site, education (≤high school, some college/college degree, some graduate/graduate degree), race (White vs non-White), current number of medical conditions, number of medications, self-reported history of fall, self-rated health (excellent/very good/good vs fair/poor), significant depressive symptoms (Geriatric Depression Scale >6), poor sleep quality (Pittsburgh Sleep Quality Index >5), Modified Mini-Mental State (3MS), and BMI.
*Interpreted as the average SD difference in PFS at Visit 4 per 1 SD greater baseline PASE (ie, person-level intercept).
†Interpreted as the average SD difference in PFS per SD greater annual change in PASE score (ie, person-level slope).
*p < .05.
Discussion
To the best of our knowledge, this study is the first to examine longitudinal associations between different types of physical activity and perceived fatigability in older men. We found that higher self-reported total physical activity, leisure exercise, and household activity were each independently associated with lower perceived physical and mental fatigability scores after adjusting for demographic, health, and lifestyle confounders. Furthermore, higher baseline and greater declines in total physical activity were associated with greater perceived fatigability scores up to 14 years later, and the results varied by activity type and marital status. Associations followed a similar pattern and were in the same direction for both perceived physical and mental fatigability.
Our findings extend prior cross-sectional work suggesting that physical activity declines may precede perceived fatigability in the broader disablement pathway (4,18). We found that self-reported physical activity was lower overall, but also declined more, across visits for those who went on to have more severe fatigability at Year 14. Qiao et al. (4) similarly found that lower physical activity was related to greater perceived fatigability in a cross-sectional statistical mediation analysis. Higher physical activity may decrease incident fatigability through multiple mechanisms. Higher physical activity is linked with greater cardiorespiratory fitness, which may provide a higher level of total energy to complete daily activities without reaching the threshold for fatigue (19,42). Greater physical activity may also reduce fat mass and promote muscle and mitochondrial functioning, leading to more efficient energy utilization (43,44). Although we adjusted for total chronic conditions in this study, perceived physical fatigability may also reflect the underlying disease burden that is not captured by a simple count of conditions (1). Thus, physical activity may also lead to lower fatigability through reducing the risk and severity of incident health conditions (eg, cardiovascular disease, metabolic syndrome, type 2 diabetes) (23,45).
Importantly, because fatigability was not measured at baseline in the current study, we cannot rule out a bidirectional relationship. Greater baseline fatigability may have driven further reductions in physical activity over time to preserve energy needed for homeostasis and resting metabolic rate (6,19), resulting in a vicious cycle of functional declines that culminates in impairments. Nevertheless, the longitudinal design and extensive follow-up of the current study established the temporal precedence of changes in physical activity on subsequent fatigability. Here we found that total physical activity is linked with subsequent fatigability in older men up to 14 years later.
The current study also revealed important differences by activity type in the timing and strength of their prospective associations with future perceived physical and mental fatigability. Lower baseline leisure exercise was more strongly related to future perceived physical and mental fatigability than lower baseline household activity for the overall sample. Yet, annual declines in household activity were more strongly related to future perceived fatigability than annual declines in leisure exercise. These findings suggest that leisure exercise (eg, sports, strength/conditioning) is a stronger early predictor of fatigability, but more proximate declines in household activities (eg, housework, yardwork) also signal greater fatigability.
The temporal ordering observed here parallels normative age-related changes in activity (23,24), where declines in leisure exercise typically precede declines in household activity, and implies that those who go on to develop more severe perceived physical and mental fatigability may have an accelerated rate of activity change. Slowing this change to affect fatigability and downstream functional outcomes may ultimately require interventions that are specific to an individual’s life course context. For example, interventions later in the life course may provide benefit by also targeting lighter intensity, nonexercise activity, rather than only leisure exercise (25). The current findings support this, as men with less decline in household activity across 14 years had lower subsequent fatigability.
Furthermore, we found that marital and living status at baseline modified the prospective association between household activity and perceived mental fatigability. Baseline household activity predicted mental fatigability only for men who were unmarried or living alone at baseline. Spouses or other cotenants may provide the emotional or motivational support needed to maintain activity of sufficient intensity and duration to mitigate future fatigability (46). Pettee et al. (26) found that older married men with high activity were 3 times as likely to have a spouse with similarly high activity compared to married men with low activity, suggesting the importance of a spouse in motivating physical activity for older men. The instrumental support provided by spouses or cotenants may also mask or compensate for lower levels of activity in older men. Prior work suggests a gender divide in activity engagement for older cohorts, where women typically perform more household activity than men, especially among married couples (28). If married men were more likely to receive support in doing household activities from their wives, it may especially explain why baseline household activity was not associated with perceived mental fatigability for this group. How spousal support may specifically affect physical activity to reduce fatigability remains an important future research area.
The current study was limited to primarily White, older men with higher education. Activity was also self-reported and may be subject to recall bias (47,48). Future studies in diverse samples (eg, including women, racial/ethnic minorities) and with objective measurement of activity via accelerometry are necessary to replicate these findings. We also could not characterize fatigability at baseline and therefore could not rule out a bidirectional relationship between physical activity and perceived fatigability, which is an important research gap to be addressed by further longitudinal studies. Future studies can also examine whether changes in marital status over time further moderate the association between physical activity and perceived fatigability.
This study also had several strengths. We used data from a large, well-characterized cohort of older men. We also examined associations separately across multiple types of activities and separately by marital status. Both are contextual factors that may be important to modify the impact of physical activity on perceived fatigability in later life. Lastly, the measure of perceived fatigability used here allowed for comparisons across samples with different health characteristics (9,12), which may improve the generalizability of the current findings to more diverse samples of older adults.
The current study demonstrated a prospective association between self-reported physical activity and both perceived physical and mental fatigability, independent of health and lifestyle factors. This provides further evidence for perceived fatigability as a potential intermediary factor between physical activity and functional impairments on the broader disablement pathway that may offer opportunities for interventions. Baseline differences were largest for leisure exercise, whereas declines were greater for household activity, suggesting a temporal ordering of activity change in relation to fatigability. Being married may further mitigate or mask the impact of household activity on future fatigability. Future work could clarify how marital and other social support mitigates the activity–fatigability relationship, and whether the relationship can be further modified by intervening on light-intensity activity.
Supplementary Material
Acknowledgments
We would like to thank the participants for giving their time and energy to this study.
Contributor Information
Kyle D Moored, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Yujia (Susanna) Qiao, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Robert M Boudreau, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Lauren S Roe, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Peggy M Cawthon, California Pacific Medical Center, San Francisco, California, USA; Department of Epidemiology and Biostatics, University of California San Francisco, San Francisco, California, USA.
Jane A Cauley, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Nancy W Glynn, Department of Epidemiology, University of Pittsburgh School of Public Health, Pittsburgh, Pennsylvania, USA.
Funding
This work and the Osteoporotic Fractures in Men Study (MrOS) are supported by the National Institutes of Health funding. The following institutes provide support: the National Institute on Aging (NIA), the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS), the National Center for Advancing Translational Sciences (NCATS), and NIH Roadmap for Medical Research under the following grant numbers: U01 AG027810, U01 AG042124, U01 AG042139, U01 AG042140, U01 AG042143, U01 AG042145, U01 AG042168, U01 AR066160, R01 AG066671, and UL1 TR000128. Additionally, the Claude D. Pepper Older Americans Independence Center, Research Registry and Developmental Pilot Grant (NIH P30 AG024827), and the Intramural Research Program, National Institute on Aging supported N.W.G. to develop the Pittsburgh Fatigability Scale. K.D.M. is supported by the Pittsburgh Epidemiology of Aging Training Program (NIA grant T32 AG000181).
Conflict of Interest
None declared.
Data Availability
MrOS data are publicly available at https://mrosonline.ucsf.edu. Analytic code is available upon request of the corresponding author (K.D.M.).
References
- 1. Eldadah BA. Fatigue and fatigability in older adults. PM R. 2010;2(5): 406–413. doi: 10.1016/j.pmrj.2010.03.022 [DOI] [PubMed] [Google Scholar]
- 2. Fielding RA, Guralnik JM, King AC, et al. Dose of physical activity, physical functioning and disability risk in mobility-limited older adults: results from the LIFE study randomized trial. PLoS One. 2017;12(8):e0182155. doi: 10.1371/journal.pone.0182155 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Simonsick EM, Schrack JA, Santanasto AJ, Studenski SA, Ferrucci L, Glynn NW. Pittsburgh Fatigability Scale: one-page predictor of mobility decline in mobility-intact older adults. J Am Geriatr Soc. 2018;66(11): 2092–2096. doi: 10.1111/jgs.15531 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Qiao YS, Gmelin T, Renner SW, et al. Evaluation of the bidirectional relations of perceived physical fatigability and physical activity on slower gait speed. J Gerontol A Biol Sci Med Sci. 2020;76(10):e237–e244. doi: 10.1093/gerona/glaa281 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Egerton T, Helbostad JL, Stensvold D, Chastin SF. Fatigue alters the pattern of physical activity behavior in older adults: observational analysis of data from the Generation 100 Study. J Aging Phys Act. 2016;24(4): 633–641. doi: 10.1123/japa.2015-0237 [DOI] [PubMed] [Google Scholar]
- 6. Moreh E, Jacobs JM, Stessman J. Fatigue, function, and mortality in older adults. J Gerontol A Biol Sci Med Sci. 2010;65(8):887–895. doi: 10.1093/gerona/glq064 [DOI] [PubMed] [Google Scholar]
- 7. Webel AR, Perazzo J, Decker M, Horvat-Davey C, Sattar A, Voss J. Physical activity is associated with reduced fatigue in adults living with HIV/AIDS. J Adv Nurs. 2016;72(12):3104–3112. doi: 10.1111/jan.13084 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Engberg I, Segerstedt J, Waller G, Wennberg P, Eliasson M. Fatigue in the general population—associations to age, sex, socioeconomic status, physical activity, sitting time and self-rated health: the northern Sweden MONICA study 2014. BMC Public Health 2017;17(1):1–9. doi: 10.1186/s12889-017-4623-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Schrack JA, Simonsick EM, Glynn NW. Fatigability: a prognostic indicator of phenotypic aging. J Gerontol A Biol Sci Med Sci. 2020;75(9):e63–e66. doi: 10.1093/gerona/glaa185 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Simonsick EM, Glynn NW, Jerome GJ, Shardell M, Schrack JA, Ferrucci L. Fatigued, but not frail: perceived fatigability as a marker of impending decline in mobility-intact older adults. J Am Geriatr Soc. 2016;64(6): 1287–1292. doi: 10.1111/jgs.14138 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Glynn NW, Santanasto AJ, Simonsick EM, et al. The Pittsburgh Fatigability Scale for older adults: development and validation. J Am Geriatr Soc. 2015;63(1):130–135. doi: 10.1111/jgs.13191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Renner SW, Bear TM, Brown PJ, et al. Validation of perceived mental fatigability using the Pittsburgh Fatigability Scale. J Am Geriatr Soc. 2021;69(5):1343–1348. doi: 10.1111/jgs.17017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Prochaska MT, Zhang H, Alavi C, Meltzer DO. Fatigability: a new perspective on and patient-centered outcome measure for patients with anemia. Am J Hematol. 2020;95(7):E166–E169. doi: 10.1002/ajh.25803 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Prochaska MT, Zhang H, Babajide R, Meltzer DO. The effect of red blood cell transfusion on fatigability after hospital discharge. Blood Adv. 2020;4(22):5690–5697. doi: 10.1182/bloodadvances.2020003364 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kim I, Hacker E, Ferrans CE, Horswill C, Park C, Kapella M. Evaluation of fatigability measurement: integrative review. Geriatr Nurs. 2018;39(1): 39–47. doi: 10.1016/j.gerinurse.2017.05.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Schrack JA, Kuo P-L, Wanigatunga AA, et al. Active-to-sedentary behavior transitions, fatigability, and physical functioning in older adults. J Gerontol A Biol Sci Med Sci. 2019;74(4):560–567. doi: 10.1093/gerona/gly243 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wanigatunga AA, Simonsick EM, Zipunnikov V, et al. Perceived fatigability and objective physical activity in mid-to late-life. J Gerontol A Biol Sci Med Sci. 2018;73(5):630–635. doi: 10.1093/gerona/glx181 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Verbrugge LM, Jette AM. The disablement process. Soc Sci Med. 1994;38(1):1–14. doi: 10.1016/0277-9536(94)90294-1 [DOI] [PubMed] [Google Scholar]
- 19. Alexander NB, Taffet GE, Horne FM, et al. Bedside-to-bench conference: research agenda for idiopathic fatigue and aging. J Am Geriatr Soc. 2010;58(5):967–975. doi: 10.1111/j.1532-5415.2010.02811.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Glass TA, Goodman SN, Hernán MA, Samet JM. Causal inference in public health. Annu Rev Public Health. 2013;34:61–75. doi: 10.1146/annurev-publhealth-031811-124606 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Salerno EA, Wanigatunga AA, An Y, et al. Longitudinal association between perceived fatigability and cognitive function in older adults: results from the Baltimore Longitudinal Study of Aging. J Gerontol A Biol Sci Med Sci. 2020;75(9):e67–e73. doi: 10.1093/gerona/glz287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Moored KD, Rosso AL, Gmelin T, et al. Life-Space mobility in older men: the role of perceived physical and mental fatigability. J Gerontol A Biol Sci Med Sci. 2021. doi: 10.1093/gerona/glab286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Chodzko-Zajko WJ, Proctor DN, Singh MAF, et al. Exercise and physical activity for older adults. Med Sci Sports Exerc. 2009;41(7):1510–1530. doi: 10.1249/MSS.0b013e3181a0c95c [DOI] [PubMed] [Google Scholar]
- 24. DiPietro L. Physical activity in aging: changes in patterns and their relationship to health and function. J Gerontol A Biol Sci Med Sci. 2001;56(suppl 2):13–22. doi: 10.1093/gerona/56.suppl_2.13 [DOI] [PubMed] [Google Scholar]
- 25. Nicklas BJ. No expiration date on the association between physical activity and mortality. J Am Geriatr Soc. 2018;66(5):850–852. doi: 10.1111/jgs.15243 [DOI] [PubMed] [Google Scholar]
- 26. Pettee KK, Brach JS, Newman AB, et al. Influence of marital status on physical activity levels among older adults. Med Sci Sports Exerc. 2006;38(3):541–546. doi: 10.1249/01.mss.0000191346.95244.f7 [DOI] [PubMed] [Google Scholar]
- 27. Cohen S, Wills TA. Stress, social support, and the buffering hypothesis. Psychol Bull. 1985;98(2):310–357. doi: 10.1037/0033-2909.98.2.310 [DOI] [PubMed] [Google Scholar]
- 28. South SJ, Spitze G. Housework in marital and nonmarital households. Am Sociol Rev. 1994; 59:327–347. doi: 10.2307/2095937 [DOI] [Google Scholar]
- 29. Orwoll E, Blank JB, Barrett-Connor E, et al. Design and baseline characteristics of the osteoporotic fractures in men (MrOS) study—a large observational study of the determinants of fracture in older men. Contemp Clin Trials. 2005;26(5):569–585. doi: 10.1016/j.cct.2005.05.006 [DOI] [PubMed] [Google Scholar]
- 30. Blank JB, Cawthon PM, Carrion-Petersen ML, et al. Overview of recruitment for the Osteoporotic Fractures in Men Study (MrOS). Contemp Clin Trials. 2005;26(5):557–568. doi: 10.1016/j.cct.2005.05.005 [DOI] [PubMed] [Google Scholar]
- 31. Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activity Scale for the Elderly (PASE): development and evaluation. J Clin Epidemiol. 1993;46(2):153–162. doi: 10.1016/0895-4356(93)90053-4 [DOI] [PubMed] [Google Scholar]
- 32. Washburn RA, Ficker J. Physical Activity Scale for the Elderly (PASE): the relationship with activity measured by a portable accelerometer. J Sports Med Phys Fitness. 1999;39(4):336. [PubMed] [Google Scholar]
- 33. Dinger M, Oman F, Taylor E, Vesely S, Able J. Stability and convergent validity of the Physical Activity Scale for the Elderly (PASE). J Sports Med Phys Fitness. 2004;44(2):186. doi: 10.1016/S0895-4356(97)00010-3 [DOI] [PubMed] [Google Scholar]
- 34. Cohen RW, Meinhardt AJ, Gmelin T, et al. Prevalence and severity of perceived mental fatigability in older adults: the Long Life Family Study. J Am Geriatr Soc. 2021;69(5):1401–1403. doi: 10.1111/jgs.17075 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Cooper R, Popham M, Santanasto AJ, Hardy R, Glynn NW, Kuh D. Are BMI and inflammatory markers independently associated with physical fatigability in old age? Int J Obes (Lond). 2019;43(4):832–841. doi: 10.1038/s41366-018-0087-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Almeida OP, Almeida SA. Short versions of the Geriatric Depression Scale: a study of their validity for the diagnosis of a major depressive episode according to ICD-10 and DSM-IV. Int J Geriatr Psychiatry. 1999;14(10):858–865. doi: 10.1002/(sici)1099-1166(199910)14:10<858::aid-gps35>3.0.co;2-8 [DOI] [PubMed] [Google Scholar]
- 37. Buysse DJ, Reynolds CF III, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193–213. doi: 10.1016/0165-1781(89)90047-4 [DOI] [PubMed] [Google Scholar]
- 38. Pahor M, Chrischilles E, Guralnik J, Brown S, Wallace R, Carbonin P. Drug data coding and analysis in epidemiologic studies. Eur J Epidemiol. 1994;10(4):405–411. doi: 10.1007/BF01719664 [DOI] [PubMed] [Google Scholar]
- 39. Teng EL, Chui HC. The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry. 1987;48(8):314–318. [PubMed] [Google Scholar]
- 40. Koenker R, Hallock KF. Quantile regression. J Econ Perspect. 2001;15(4):143–156. doi: 10.1257/jep.15.4.143 [DOI] [Google Scholar]
- 41. Berry WD, Feldman S, Stanley Feldman D.. Multiple Regression in Practice. Sage; 1985. doi: 10.4135/9781412985208 [DOI] [Google Scholar]
- 42. Schrack JA, Wanigatunga AA, Zipunnikov V, Kuo PL, Simonsick EM, Ferrucci L. Longitudinal association between energy regulation and fatigability in mid-to-late life. J Gerontol A Biol Sci Med Sci. 2020;75(9): e74–e80. doi: 10.1093/gerona/glaa011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Santanasto AJ, Glynn NW, Jubrias SA, et al. Skeletal muscle mitochondrial function and fatigability in older adults. J Gerontol A Biol Sci Med Sci. 2015;70(11):1379–1385. doi: 10.1093/gerona/glu134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Bogdanis GC. Effects of physical activity and inactivity on muscle fatigue. Front Physiol. 2012;3:142. doi: 10.3389/fphys.2012.00142 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Kannel WB, Sorlie P. Some health benefits of physical activity: the Framingham study. Arch Intern Med. 1979;139(8):857–861. doi: 10.1001/archinte.1979.03630450011006 [DOI] [PubMed] [Google Scholar]
- 46. Gellert P, Ziegelmann JP, Warner LM, Schwarzer R. Physical activity intervention in older adults: does a participating partner make a difference? Eur J Ageing. 2011;8(3):211. doi: 10.1007/s10433-011-0193-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181–188. doi: 10.1249/mss.0b013e31815a51b3 [DOI] [PubMed] [Google Scholar]
- 48. Brach JS, Kriska AM, Glynn NW, Newman AB. Physical activity and the older adult: measurement, benefits, and risks. Curr Cardiovasc Risk Rep. 2008;2(4):305–310. doi: 10.1007/s12170-008-0056-6 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
MrOS data are publicly available at https://mrosonline.ucsf.edu. Analytic code is available upon request of the corresponding author (K.D.M.).