Physical performance changes as clues to late-life blood pressure changes with advanced age: the osteoporotic fractures in men study

Abstract

Objectives:

This study examined whether changes in late-life physical performance are associated with contemporaneous changes in blood pressure (BP) in older men.

Design:

prospective cohort study over 7 years.

Setting and Participants:

Physical performance (gait speed, grip strength, chair stand performance) and clinic-measured BP at baseline and at least one follow-up (year 7 or 9) were assessed in 3,135 men aged ≥65 y enrolled in the Osteoporotic Fractures in Men Study (MrOS).

Methods:

Generalized estimating equation analysis of multivariable models with standardized point estimates (β [95% CI]) described longitudinal associations between physical performance and BP changes in participants overall, and stratified by baseline cardiovascular disease (CVD), antihypertensive medication use (none, ≥1), and enrollment age (<75 years; ≥75 years).

Results:

Overall, positive associations (z-score units) were found between each increment increase in gait speed and systolic (SBP) (0.74 [0.22, 1.26]) and grip strength (0.35 [0.04, 0.65]) or gait speed (0.55 [0.24, 0.85]) with diastolic (DBP). Better grip strength and chair stand performance over time were associated with 1.83 [0.74, 2.91] and 3.47 [0.20, 6.74] mmHg higher SBP, respectively in men with CVD at baseline (both interaction P <.05). Gait speed increases were associated with higher SBP in men without CVD (0.76 [0.21, 1.32]), antihypertensive medication non-users (0.96 [0.30, 1.62]), aged <75 years (0.73 [0.05, 1.41]) and ≥75 years (0.76 [0.06, 1.47]). Similar positive, but modest associations for DBP were observed with grip strength in men with CVD, antihypertensive medication non-users, and aged <75 years, and with gait speed in men without CVD, aged <75 years, and irrespective of antihypertensive medication use.

Conclusion:

In older men, better physical performance is longitudinally associated with higher BP. Mechanisms and implications of these seemingly paradoxical findings, which appears to be modified by CVD status, antihypertensive medication use, and age, requires further investigation.

1. Introduction

Hypertension disproportionately affects older adults (≥65 years), contributing significantly to cardiovascular disease (CVD) morbidity and mortality [1,2]. The substantial growth of an older adult population undoubtedly contribute to the steep rise in both the incidence and prevalence of hypertension among older individuals [3]. A related, but distinct explanation for elevated BP and hypertension rates during old age is the impact of functional markers of biological age, rather than chronological age itself, and their dynamic changes over time. These factors may affect vascular function, subsequently influencing BP levels in late life [3–5].

Among the many pathophysiologic consequences of cardiovascular aging is the decline in physical performance in older adults [6–8]. Grip strength, chair stand performance, and gait speed, in particular, represent integrated measures of function across multiple organ systems, including the cardiovascular and musculoskeletal systems, that play a vital role in regulating BP control [9]. Physical performance deficits in old age have often been studied as a consequence of chronic conditions, including hypertension [10] and CVD [11,12]. There is additional evidence underscoring the prognostic relevance of physical performance, such as walking pace, in predicting future CVD risk or events [8,13]. Prior observational evidence suggests that well-functioning older adults with chronically elevated systolic BP (SBP) are more likely to have lower grip strength [14], chair stand performance, or gait speed [10,15,16] or experience faster gait speed decline [10]. However, whether age-related changes in physical performance precipitate changes in BP in old age is not well-studied but has significant public health significance given the aging population. In our previous study of older women, we observed a positive relationship between better physical performance and higher SBP and diastolic BP (DBP) that appeared to be modified by CVD history, antihypertensive medication use, and age [17].

The present study extends this research, by examining the longitudinal relationship between change in physical performance and change in BP over time in older well-functioning men enrolled in the Osteoporotic Fractures in Men Study (MrOS). We hypothesized that better performance would be associated with lower SBP, whereas the relationship between performance and DBP would vary by CVD status, and age group, as previously observed [17,18]. The evaluation of this relationship is important given that physical performance, modifiable by lifestyle intervention, could have significant impact on clinical decision-making for BP management in older populations, which currently remains controversial [19,20].

2. Materials and methods

2.1. Study population

MrOS is a prospective cohort study of 5,994 community-dwelling men ≥65 years old at enrollment (2000–2002) at six geographic areas of the United States [21–23]. Of 5,994 MrOs men, both SBP and DBP were measured in 3,135 who enrolled in the ancillary Outcomes of Sleep Disorders in Older Men Study (MrOS Sleep Study) between 2003 and 2005. Participation in the MrOS Sleep study required men do not use mechanical devices during sleep or oxygen therapy.

For the analysis, all 3,135 MrOS Sleep Study participants were included from Sleep Visit 1, henceforth ‘baseline’ visit, during which clinical SBP and DBP and the three physical performance measures (gait speed, grip strength, chair stand performance) were obtained. Depending on the variables included in different stepwise models, the number of observations may vary, but all available samples were utilized in the analysis without additional exclusion criteria. Of the total sample, 3,133 had at least one of the performance measures and SBP or DBP at baseline. Among them, 2,615 men had at least one follow-up assessment of one or more physical performance measures and BPs at MrOS clinic study visits during Year 7 (March 2007-March 2009) and Year 9 (May 2014-May 2016). The mean follow-up time was 6.80 (SD 4.15) years. In the longitudinal models, samples that have only baseline observations were included to take account of their effect on the correlation test between the performance measures and BP outcomes at baseline. All participants gave written informed consent, and the study was performed in accordance with the Declaration of Helsinki.

2.2. Physical performance measures

Physical performance included gait speed, grip strength, and repeated chair stands. Gait speed (meters per second) measured the time it took in seconds to complete a 6-meter walking course performed at usual pace, using ambulatory aids as needed. The test was repeated, and the faster of the two measured times was included in the analysis. Grip strength, a test of voluntary muscle strength, was measured in kilograms (kg) using a Jamar dynamometer (Sammons Preston Rolyan, Bolingbrook, IL, USA). Participants completed two trials for each hand, and the maximum effort across the trials was used for analyses. Chair stands evaluated whether the participant was able to stand at least once, without using hands or arms, from a standard chair. Men who were able to stand once were asked to repeat the task 5 times, and the number of chair stands completed was recorded. Chair stand speed was calculated and used to create a variable equal to the estimated number of chair stands per 10 s, with those who were unable to complete the test (n = 46) coded as zero.

2.3. Blood pressure

At both the baseline and the follow-up visits, SBP and DBP were measured in the seated position after rest and calculated as the average of two measurements. Auscultatory methods using a stethoscope and mercury sphygmomanometer on the right arm was used at baseline, but a BP Tru automated blood pressure monitor (model BMP-300; Coquitlam, British Columbia, Canada) was used at follow-up. The BP Tru device has been shown to have high quality and comparable accuracy as auscultatory manual mercury methodology [24].

2.4. Hypertensive medications

All prescribed medications including antihypertensive drugs taken on a regular basis in the past 30 days to the clinic visit and duration of use (<1 month, 1 month to 1 year, 1–3 years, 3–5 years, >5 years, and don’t know) were ascertained by trained clinic interviewers at baseline and follow-up visits. All prescription medications recorded by the clinics were stored in an electronic medications inventory database (San Francisco Coordinating Center, CA). Each medication was matched to its ingredient (s) based on the Iowa Drug Information Service Drug Vocabulary [25]. In the present analyses, men are classified as a user or non-user of antihypertensive medication coded according to their therapeutic classes based on the medication inventory at the screening: ACE inhibitors, Angiotensin II Receptor Antagonist, thiazides, β-blockers, calcium channel blockers. These medications were specifically chosen based on plausible physiological mechanisms shared between skeletal muscle and vascular systems [26,27]. In analytical models, maximum duration was used instead of number of medication, given this approach accounts for medication class and strength, allowing for more plausible assumptions about medication dosage. Maximum duration was calculated for each visit year as the maximum duration of use among all antihypertensive medications used per drug class through the Year 9 clinic visit.

2.5. Other covariates

Self-administered questionnaires were used to collect information on demographic, lifestyle behaviors, and medical history, including self-reported diabetes and CVD (self-reported physician diagnosis of myocardial infarction, stroke or heart failure), osteoarthritis, chronic obstructive pulmonary disease (COPD), kidney disease/failure, osteoarthritis, and history of falls in the 12 months prior to the clinic visit. The short form-12 questionnaire physical component subscale (PCS-12) was administered to assess self-rated physical health status, with scores ranging from 0 to 50 points and higher scores indicating better physical health status [28]. Cognitive function was assessed using the Teng 3 modified mini-mental state (3MS) examination, with a scale of 0–100, and cutpoint of 80 indicating cognitive impairment [29]. Physical activity was assessed by self-report using the Physical Activity Scale for the Elderly (PASE) questionnaire [30]. Weight was measured on a balance beam or digital scale, and height by wall-mounted stadiometers. Body mass index (BMI) was calculated as weight (kg)/height² (m²).

2.6. Statistical methodology

Characteristics of the participants were summarized by means and standard deviations (SD) for continuous variables and counts and percentages for categorical variables. SBP and DBP were the primary outcomes of interest, and grip strength, chair stands, and gait speed were the primary exposure variables, evaluated separately.

Data were checked for outliers, homoscedasticity of residuals, and skewness and kurtosis were assessed to confirm linearity and normality for all variables. Pearson’s correlation (r) was used to examine bivariate correlations between physical performance measures and SBP and DBP at baseline and follow-up years 7 and 9. Generalized Estimating Equations with an identifying link function was conducted to determine the estimated correlation coefficient between each performance measure and BP outcome, with age as the index of time. Each performance measure was converted into a z-score to standardize correlations with BP outcomes allowing for direct comparisons of the regression coefficients (β) across the three different units of physical performance measures. Adjustment for model covariates was considered for variables known to be associated with BP, or may potentially confound the relationship between performance and BP, based on previous literature [14,31–33]. A base model (Model 1) adjusted for age, race and ethnicity, MrOS clinic site, and education level; Model 2: additionally adjusted for smoking status, alcohol use, Teng 3MS score, SF12, PASE score and BMI (both expressed as time-varying covariates using data from baseline, year 7 and 9 follow-up time points), prior fall, and presence of at least one comorbidity (osteoarthritis, COPD, kidney disease/failure, diabetes); Model 3 was based on Model 2 plus history of CVD diagnosis, and duration of antihypertensive medication usage (a proxy of dosage and severity of hypertension), expressed as a time-varying covariate, calculated as the maximum value of a duration at each time point. Multicollinearity between predictors was assessed by reviewing standard errors of each parameter estimates and estimated correlation matrix, with no significant concerns of large standard error or correlation coefficient detected.

Given that CVD, hypertension medication use, and advanced age may contribute to differences in physical performance and BP change, multivariable analyses were repeated stratified according to history of CVD diagnosis (no, yes), hypertension medication use (no, yes), and age (median split; <75 vs ≥75 years) at enrollment.

Statistical analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC); all reported P-values are two-sided under significance level 0.05.

3. Results

3.1. Study population characteristics

Characteristics of the analytic cohort of 3,135 men (mean age 76.4, SD 5.6 years) at baseline are presented in Table 1. Most participants were White and not of Hispanic origin and had an average BMI 27.2 ± 3.9 kg/m². Nearly 23% had CVD at baseline. Mean SBP 127.2 ± 16.4/DBP 67.7 ± 9.5 met the criteria for prehypertension under JNC 7 guidelines used at MrOS enrollment [31]. More than half of the cohort reported receiving treatment for hypertension, and 25% had a maximum duration of antihypertensive medication use of 5 or more years.

Table 1.

Descriptive characteristics of MrOS study cohort (N = 3,135).

Variables	Mean ± SD or n (%)
Mean age at visit, y (range)	76.4 ± 5.6 (67–96 y)
Race or ethnicity
Non-Hispanic White	2816 (89.8)
African American	121 (3.9)
Hispanic or Latino	59 (1.9)
Asian/Pacific Islander	101 (3.22)
Other race/ethnicity	38 (1.21)
Education (some college and beyond)	2469 (78.7)
Weight (kg)	82.1 ± 13.2
Height (cm)	173.7 ± 6.8
BMI kg/m2	27.2 ± 3.9
PASE score	145.6 ± 71.9
SF-12 Physical Summary Scale score (0–50)	48.6 ± 10.2
Alcohol use-n (%)
Alcohol (none)	1082 (34.7)
Alcohol (<6 drink/week)	1250 (40.1)
Alcohol (6+ drinks/week)	786 (25.2)
Current smoker	64 (2.0)
Teng 3MS score	92.6 ± 6.4
Fall in last year	964 (30.8)
History of COPD	164 (5.23)
History of Diabetes	417 (13.3)
History of osteoarthritis	736 (23.5)
History of kidney disease/failure	36 (1.15)
Diagnosis of cardiovascular disease*	719 (22.9)
Diagnosis for Hypertension	1560 (49.8)
Baseline Antihypertensive medication use†
0	1332 (42.5)
1	906 (28.9)
2	628 (20.0)
≥3	268 (8.6)
Systolic BP, mmHg	127.0 ± 16.4
Diastolic BP, mmHg	67.7 ± 9.5
Grip strength	40.7 ± 8.5
Walking speed (m/s)	1.19 ± 0.23
Chair stand performance (n performed in 10 s)	4.6 ± 8.5

Notes: Baseline values taken at Sleep Visit 1. Mean ± SD or N (%), as appropriate. Missing *SV1: height (n = 6); BMI (n = 4); PASE score (n = 2); current smoker (n = 2); alcohol intake (n = 17); number of antihypertensive medication, (n = 1); history of fall in past 12 months (n = 2); history of diabetes (n = 2); history of COPD (n = 2); history of osteoarthritis (n = 3); history of kidney disease/failure (n = 2); CVD history (n = 2); systolic BP (n = 4); diastolic BP (n = 4); grip strength (n = 75); gait speed (n = 88); chair stand performance (n = 4).

^*History of cardiovascular disease defined as reported history of either myocardial infarction, stroke, or congestive heart failure.

^†Use of antihypertensive medication coded according to therapeutic classes based on the medication inventory at screening: ACE inhibitors, Angiotensin II Receptor Antagonist, β-blockers, thiazide diuretics, calcium channel blockers.

3.2. Missing data or loss to follow-up

Men with missing Year 7 physical function (n = 319) or BP outcome data (n = 317) had lower levels of physical performance at baseline compared to those who had the data available, including weaker grip strength (35.93 vs. 40.70 kg), fewer chair stands in 10 s (3.33 vs. 4.56), and had a slower gait speed (1.08 vs. 1.19 m/s) (all P < 0.0001). No significant differences in baseline SBP were observed; however, men who were missing DBP data at Year 7 had significantly lower DBP by 1.54 mm Hg compared to those with Year 7 BP data (P < 0.05). On average, men with missing data at Year 7 were older by 3.4 year, had slightly lower BMI, self-reported physical function, and cognitive function, were more likely have experienced a fall in the previous 12 months, have CVD, and report a longer duration of antihypertensive medication use by 0.3 year (all P < 0.05). Comparing men with and without Year 9 data showed similar results. Specifically, men without Year 9 data had lower levels of grip strength (38.22 kg), gait speed (1.12 m/s), slightly higher SBP and lower DBP by 1.00 mm Hg each (all P < 0.05).

As the result of simple linear regression to test overall trend across visit years, all performance measures and SBP showed significant decreases, while DBP showed significant increases over time (all slopes P < .0001; Supplement Table S1)

Across baseline and follow-up visits at years 7 and 9, physical performance showed weak correlations with BP outcomes (Supplemental Table S2). Among the significant bivariate findings, the direction of correlation between performance measures and SBP outcomes varied at each follow-up visit. Most correlations with gait speed were negative, while correlations with chair stand performance were primarily positive. For DBP, correlations with all three performance measures were generally positive at all follow-up visits.

Supplemental Figures 1–3 describe the unadjusted linear relationship between the distribution of SBP or DBP outcomes and each performance measure for the overall cohort, and according to CVD diagnosis, antihypertensive medication use, or age at baseline. Among the whole cohort, and similarly in models stratified by antihypertensive medication use, or baseline age, non-significant associations were observed between each performance measure and SBP and DBP. However, associations of grip strength or chair stand performance with SBP and DBP appeared to be influenced by baseline CVD status (all P < 0.05). Despite the observed positive associations, the estimated SBP and DBP levels across the range of each performance measure were within the 5th through 95th percentiles for BP and remained below JNC 7 hypertension thresholds as well as the 2018 hypertension targets for SBP and DBP [1] (Supplemental Figures 1–3).

Table 2 presents the sequential covariate adjustment and beta coefficients (β; 95% CI) using one z-score increment in performance measures in multivariable models of BP outcomes. In the overall cohort, several patterns were observed: first, each z-score increment in grip strength was associated with 0.88 mmHg (95% CI 0.37, 1.40) higher SBP, and 0.69 mmHg (0.39, 0.98) higher DBP; these estimates were attenuated though remained statistically significant after accounting for CVD diagnosis and duration of antihypertension medication use; second, no significant associations were observed between chair stand performance and either SBP or DBP; and third, an increase of one z-score in gait speed, i.e., faster speed, was associated with higher SBP (0.61; 95% CI 0.12, 1.10) and DBP (0.71; 95% CI 0.42, 0.99) and this remained significant with little attenuation in base and fully adjusted Model 3.

Table 2.

Associations between the change in performance measures and blood pressure after sequential covariate adjustment.

	Outcomes
	Systolic blood pressure β (95% CI)	Diastolic blood pressure β (95% CI)
Model 1
Grip strength	0.88 (0.37, 1.40) ‡	0.69 (0.39, 0.98) ∥
Chair stand	0.22 (−0.02, 0.47)	0.07 (−0.03, 0.17)
Gait speed	0.61 (0.12, 1.10) †	0.71 (0.42, 0.99) ∥
Model 2
Grip strength	0.62 (0.09, 1.15) †	0.41 (0.10, 0.71) ‡
Chair stand	0.18 (−0.04, 0.40)	0.01 (−0.05, 0.07)
Gait speed	0.78 (0.26, 1.30) ‡	0.59 (0.28, 0.90) ∥
Model 3
Grip strength	0.59 (0.06, 1.11) †	0.35 (0.04, 0.65) †
Chair stand	0.17 (−0.05, 0.38)	0.00 (−0.06, 0.06)
Gait speed	0.74 (0.22, 1.26) ‡	0.55 (0.24, 0.85) ‡

Notes:

^*point estimates express the association between one standardized unit change in each performance measure and the associated BP change.

Significance

^†p<0.05;

^‡p<0.01;

^∥p<0.0001.

Model 1: age, race and ethnicity, clinic site, education.

Model 2: Model 1+ smoking status, alcohol use, cognitive function (Teng 3MS), self-rated health (SF12), time-varying physical activity (PASE score) and BMI (calculated using data from baseline, year 7 and 9 follow-up time points), prior fall, presence of at least one comorbidity (yes/no, osteoarthritis, COPD, kidney disease/failure, diabetes).

Model 3: Model 2 + CVD + time-varying duration of antihypertensive medication usage (calculated as the maximum value of a duration at each time point).

There was some evidence of effect modification by baseline CVD diagnosis with both BP outcomes. Significant interactions were observed between grip strength and chair stand performance and SBP and/or DBP (Fig. 1). Among men with baseline CVD, better grip strength and greater chair stand performance were significantly associated with a 1.83 mmHg (95% CI 0.74, 2.91) and 3.47 mmHg (0.20, 6.74) higher SBP, respectively over time (all P_interaction <.05). Also, among men with CVD, per unit increases in grip strength was associated with a 0.62 mmHg (95% CI 0.02, 1.23) higher DBP. In men without CVD, only faster gait speed was associated with 0.76 mmHg (95% CI 0.21, 1.32) higher SBP and 0.58 (95% CI 0.25, 0.92) higher DBP over time.

Fig. 1.

Associations between physical performance and blood pressure change over time by baseline CVD status.

Notes: CVD, cardiovascular disease; CI, confidence interval; DBP, diastolic blood pressure; SBP, systolic blood pressure.

*Point estimates express the association between one standardized unit change in each performance measure and the associated BP change in NoCVD (n = 2,414) and Yes CVD (n = 719) men.

Statistical significance †p<0.05; ‡p<0.01.

p-interaction represents the overall effect of the performance*CVD interaction in the model, by GEE.

Fully adjusted Model 3 was used to estimate all effect magnitude (β): age, race and ethnicity, clinic site, education, smoking status, alcohol use, cognitive function (Teng 3MS), self-rated health (SF12), time-varying physical activity (PASE score) and BMI (calculated using data from baseline, year 7 and 9 follow-up time points), prior fall, presence of at least one comorbidity (yes/no, osteoarthritis, COPD, kidney disease/failure, diabetes), time-varying duration of antihypertensive medication usage (calculated as the maximum value of a duration at each time point).

There was no significant interaction observed between physical performance and either antihypertension medication use, or age. Further, no clear patterns of effect modification by antihypertensive medication use or age for the performance-BP models were observed. Despite a lack of statistical evidence of interaction, in men reporting antihypertensive medication use at baseline, greater chair stand performance was slightly and significantly associated with higher SBP by 0.25 (95% CI 0.06, 0.43) mmHg, while faster gait speed was associated with 0.56 mmHg (95% CI 0.15, 0.98) higher DBP over time (Table 3). In men reporting no baseline antihypertensive medication use, each unit increase in grip strength and gait speed was associated with a 0.77 mmHg (95% CI 0.07, 1.45) and 0.96 mmHg (95% CI 0.30, 1.62) higher SBP, and a 0.43 mmHg (95% CI 0.03, 0.83) and 0.50 mmHg (95% 0.09, 0.90) higher DBP, respectively over time.

Table 3.

Associations between physical performance and blood pressure change over time by baseline antihypertensive medication status^*.

	SBP	DBP	p-interaction
	β (95% CI)	β (95% CI)	Systolic BP	Diastolic BP
Grip strength
No HTN med use	0.77 (0.07, 1.45) †	0.43 (0.03, 0.83) †	0.78	0.62
Yes HTN med use	0.64 (−0.05, 1.32)	0.30(−0.09, 0.69)
Chair stand
No HTN med use	0.08 (−0.21; 0.38)	0.02 (−0.09, 0.13)	0.37	0.69
Yes HTN med use	0.25 (0.06, 0.43) †	−0.01 (−0.10, 0.08)
Gait speed
No HTN med use	0.96 (0.30, 1.62) ‡	0.50 (0.09, 0.90) †	0.51	0.80
Yes HTN med use	0.75 (−0.05, 1.35)	0.56 (0.15, 0.98) ‡

Notes: HTN, hypertension medication use; CI, confidence interval; DBP diastolic blood pressure; SBP, systolic blood pressure.

Bold print indicates Statistical significance

^†p≤0.05;

^‡p<0.01.

p-interaction represents the overall effect of the performance*antihypertensive medication use interaction in the model, by GEE.

Fully adjusted Model 3 was used to estimate all effect magnitude (β): age, race and ethnicity, clinic site, education, smoking status, alcohol use, cognitive function (Teng 3MS), self-rated health (SF12), time-varying physical activity (PASE score) and BMI (calculated using data from baseline, year 7 and 9 follow-up time points), prior fall, presence of at least one comorbidity (yes/no, osteoarthritis, COPD, kidney disease/failure, diabetes), and CVD.

^*Point estimates express the association between one standardized unit change in each performance measure and the associated BP change.

In men aged ≥75 years, one-unit increase in grip strength and gait speed was associated with 0.82 (95% CI 0.11, 1.54) mmHg and 0.76 (95% CI 0.06, 1.47) higher SBP, respectively (Table 4). In men <75 years, greater chair stands performed, and faster gait speed were associated with higher SBP by 0.24 (95% CI 0.10, 0.38) and 0.73 (95% CI 0.05, 1.41) mmHg, respectively over time. Faster gait speed was associated with higher DBP over time in men <75 years (0.72 mmHg; 95% CI 0.31, 1.13) but not in men aged ≥75 years.

Table 4.

Associations between physical performance and blood pressure change over time by median age at baseline^*.

	SBP	DBP	p-interaction
	β (95% CI)	β (95% CI)	Systolic BP	Diastolic BP
Grip strength
<75 y	0.29 (−0.40, 0.97)	0.59 (0.19, 0.99) ‡	0.23	0.21
≥75 y	0.82 (0.11, 1.54) †	0.27 (−0.12, 0.67)
Chair stand
<75 y	0.24 (0.10, 0.38) ‡	0.09 (−0.05, 0.23)	0.31	0.13
≥75 y	0.06 (−0.24, 0.37)	−0.07 (−0.20, 0.07)
Gait speed
<75 y	0.73 (0.05, 1.41) †	0.72 (0.31, 1.13) ‡	0.95	0.18
≥75 y	0.76 (0.06, 1.47) †	0.34 (−0.07, 0.75)

Notes: CVD, cardiovascular disease; CI, confidence interval; DBP, diastolic blood pressure; SBP, systolic blood pressure.

Bold print indicates Statistical significance

^†p<0.05;

^‡p<0.01;

^ǂborderline significant at p = 0.06; p-interaction represents the overall effect of the performance* enrollment age interaction in the model, by GEE.

Fully adjusted Model 3 was used to estimate all effect magnitude (β): race and ethnicity, clinic site, education, smoking status, alcohol use, cognitive function (Teng 3MS), self-rated health (SF12), time-varying physical activity (PASE score) and BMI (calculated using data from baseline, year 7 and 9 follow-up time points), prior fall, presence of at least one comorbidity (yes/no, osteoarthritis, COPD, kidney disease/failure, diabetes), CVD, time-varying duration of antihypertensive medication usage (calculated as the average of relevant follow up time points).

^*Point estimates express the association between one standardized unit change in each performance measure and the associated BP change in men aged < 75 y (n = 1,338) and ≥ 75 y (n = 1,797).

4. Discussion

This study found a positive association between changes in physical performance and SBP and DBP over time in older community-dwelling men. As physical performance test scores demonstrated greater levels of physical functioning, measures of SBP and DBP were higher. However, the nature of this association differed according to CVD diagnosis, antihypertensive medication use, and age at baseline. Additionally, clear evidence of effect modification by baseline CVD status was observed for both BP outcomes and grip strength, with significant positive associations with SBP and DBP observed only in men with diagnosed CVD. Conversely, gait speed was positively associated with SBP and DBP in men without CVD at baseline. While chair stand performance showed a weak, non-significant relationship with BP outcomes in the whole cohort, it displayed a robust association with SBP among men with pre-existing CVD.

A considerable amount of evidence has reported an inverse relationship between high SBP and grip strength [34], chair stand or gait speed [10,15,16] in well-functioning older populations. Contrary to these anticipated findings and our hypothesis, the results of our multivariable analyses suggest that, despite an overall declining trend in physical performance, better performance, particularly higher grip strength and faster gait speed, were associated with higher SBP and DBP over time. Although few studies to date have contemporaneously examined changes in BP in relation to change in physical performance, our findings are consistent with a recent study in older women from our group [17], as well as the Leiden 85-plus Study, whereby higher BP was associated with greater hand grip strength in those aged over 85 years [14]. The present study extends these findings to a population of older men, controlling for relevant factors such as time-varying BMI, PASE score, and duration of antihypertensive medication intake, and uses multiple objective performance tests that capture different aspects of muscular strength, balance, and functional capacity, over time [9,12,35,36].

Our findings of a positive relationship between performance and BP are potentially attributed to factors associated with older age. Conceivably, low levels of any performance measure may reflect early manifestations to disability, coexisting subclinical disease or poor health [11] associated with hypertension, increased CVD outcomes, and mortality [12,18,37]. Conversely, higher levels in performance measures may serve as important markers of physiological vigor, particularly in older adults with declining cardiovascular function, leading to hemodynamic adaptations that provide BP control. In our models specific to CVD, each unit increase in grip strength (kg) and chair stands performed was associated with higher SBP by nearly 2 mmHg and 3.5 mmHg, respectively over time. Despite the counterintuitive positive association, SBP and DBP levels across a range of performance levels remained below hypertension thresholds over time, regardless of factors such as CVD diagnosis, antihypertensive medication use, or enrollment age (data not shown). These findings suggest that maintaining physical performance potentially may help stabilize or prevent BP from dropping too low in advanced age [19,38].

The positive association of DBP with grip strength and gait speed, independent of baseline CVD status or antihypertensive medication use was striking. Although effect sizes were modest, these findings support the hypothesis that higher physical performance levels may help to attenuate declines in DBP [14,39], which could have important implications on future health outcomes, especially in those with CVD or hypertension [17,19,40]. This aligns with a study reporting that the BP-mortality association diminishes with age, partly due to changes in physical performance and frailty that occur with age [39,41]. Declining physical performance may accelerate mobility limitations and trigger vascular maladaptation, including impaired endothelial function skeletal muscle perfusion, and increased arterial stiffness contributing to poor vascular compliance [32,42,43]. Therefore, the positive association between physical performance and BP in late life may reflect a reactive compensatory adaptation to preserve vascular function and perfusion to vital organs, such as the heart, particularly during diastole [37,38]. While caution is warranted in interpreting the potential mechanisms, these findings underscore the importance of comprehensive geriatric assessments of physical performance and associated morbidities, such as frailty, in the decisional process for managing elevated BP and hypertension in older adults, particularly those with pre-existing CVD [19,20].

4.1. Strengths and limitations

Noteworthy strengths of this study include the large sample size, longitudinal study design, collection of clinic-measured physical performance measures and BP over 7 years, and inclusion of several potential confounding factors on BP control, physical performance, and aging. Limitations of this study include its limited generalizability to women and other race and ethnic groups, as the source cohort comprised of older male adults, mostly non-Hispanic white, and community-dwelling. Further, participants were typically healthy, more educated than the general population, with high functioning based on performance tests at the baseline visit. Our analyses included healthier men who attended either the Year 7 or 9 clinic visit. Compared to men included in the analyses, approximately 10% of men had missing Year 7 or 9 clinic visit data. These men had significantly lower functioning levels and modestly lower DBP, and they carried a greater burden from CVD or other comorbidities. Thus, the requirement to have repeat measures may have introduced selection bias for the study cohort. Despite taking several measures to control for potential residual confounding such as time-varying duration of antihypertensive medication use and stratification by baseline CVD status, we also cannot dismiss the likelihood of issues related to confounding by indication or structural confounding, which may explain some of the statistically significant but not clinically significant performance-BP associations observed. Further, there was a notable mismatch between those who reported having hypertension and those taking at least one antihypertensive medication at the baseline visit. This limitation is compounded by the lack of details regarding medication dosage, primary indication for medication initiation/change, or medication adherence. These factors are crucial as higher doses of certain medications could lead to greater reductions in BP, and the long term use of certain medication classes, particularly ACE inhibitors, could yield better or worse performance [44,45]. Given the multi-factorial nature of age and disease-related functional decline, we cannot rule out the possibility of residual confounding from underlying hypertension (treated or untreated) or CVD [10,19,31]. While the relationship between performance and BP is likely bidirectional, it remains unclear whether physical performance has a causal impact on BP, or vascular mechanisms underlying BP control versus the underlying change in comorbidity burden.

5. Conclusion

In conclusion, this study demonstrated that better performance is positively associated with higher SBP and DBP over time, with the most robust associations demonstrated in those with baseline CVD. Considering the complexity of BP regulatory control mechanisms and its interplay with physical performance during aging, further research is warranted to better characterize the longitudinal relationship between performance and BP, including the underlying mechanisms involved in older adults with varying levels functional status or frailty. Given the lack of consensus from various guidelines on optimal BP targets in older adults, particularly in those with CVD [19,20], some researchers recommend that geriatric assessments with one or more of these performance measures be incorporated in the clinical decision making for the management of hypertension in old age [19]. Future research that appraises therapeutic interventions that target physical performance is needed to examine the clinical implications on SBP and DBP and hypertension management in older adults.

Supplementary Material

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jnha.2024.100317.

Abbreviations:

BMI

body mass index

BP

blood pressure

COPD

chronic obstructive pulmonary disease

CVD

cardiovascular disease

DBP

diastolic blood pressure

MrOS

Osteoporotic Fractures in Men Study

PASE

Physical Activity Scale for the Elderly

PCS-12

The short form-12 questionnaire physical component subscale

SBP

systolic blood pressure