Impacts of Reducing Sitting Time or Increasing Sit-to-Stand Transitions on Blood Pressure and Glucose Regulation in Postmenopausal Women: Three-Arm Randomized Controlled Trial

Abstract

Background:

Public health and clinical guidelines identify the importance of sedentary behaviors for cardiovascular diseases, particularly among postmenopausal women. The goal of this trial was to compare the behavioral and physiological impacts of two distinct approaches to changing sedentary behaviors.

Methods:

Overweight/obese sedentary postmenopausal women (N=407) were randomly assigned to one of three study conditions for 3 months: (1) healthy living (control), (2) reduce sitting time (Sit Less), (3) increase sit-to-stand transitions (Sit-to-Stand). Each study arm received 7 individual health coach sessions across 12-weeks. At baseline and 3 months participants had fasting blood drawn, blood-pressure measured, and wore thigh (activPal) and hip (ActiGraph) accelerometers for 7 days. Linear mixed models evaluated each intervention arm compared to the control (Healthy Living) arm.

Results:

388 women (95%) completed the 3-month trial. The Sit Less arm reduced total sitting time by 58 min/day more than the Healthy Living arm (95% CI [−82.9, −33.6], p<0.001), but did not change sit-to-stand transitions (STST; −1 STST/day (95% CI [−9.4, 6.5], p=0.72)). Conversely, the Sit-to-Stand arm significantly increased STST by 26 STST/day more than the Healthy Living arm (95% CI [17.71, 33.64], p<0.001), but did not differ in change to sitting time (−10 min/day (95% CI [−34.6, 14.9], p=0.44)). The Sit-to-Stand arm had significant decreases in diastolic blood pressure compared to the Healthy Living arm (−2.24 mmHg (95% CI [−4.08, −0.40], p=0.02) and similar decreases in systolic blood pressure compared to the Healthy Living arm (−3.33 mmHg (95% CI [−6.32, −0.33], p=0.03), although it did not reach the a priori significance threshold of p<0.025. There were no significant intervention effects on blood pressure for the Sit Less arm and no intervention effects for the glucoregulatory outcomes for either arm.

Conclusions:

This trial demonstrated the feasibility of changing sedentary behaviors as well as the distinct nature of sitting time and STST. Increasing STST improved blood pressure in overweight and obese postmenopausal women within 3-months. Focusing on increasing STST may be an achievable behavioral target to reduce cardiovascular disease risk in postmenopausal women.

Background

Postmenopausal women spend most of their waking hours sitting,¹ increasing their risk of cardiovascular diseases, type 2 diabetes, cancer, and premature mortality.^2–7 Public health efforts have focused on moderate to vigorous physical activity (MVPA) to improve cardiovascular health,⁸ but only 3-6% of adults 60 or older meet MVPA guidelines,^{9, 10} and increasing MVPA may be a challenging goal.¹¹ Changing sedentary behaviors may offer feasible alternative behavioral targets for improving cardiovascular health. ^{12, 13}

Reducing sitting time and increasing sit-to-stand transitions (STST) have been recommended as promising interventions for sedentary behavior change.¹⁴ Although previous studies support the feasibility of older adults reducing their sedentary time,^15–20 and identify potential benefits for blood pressure, glucose control, and insulin sensitivity,²¹ these studies have generally been of short duration, with small samples, methodological limitations, and have not focused on postmenopausal women.²² This is critical, as cardiovascular risk increases after menopause,^{23, 24} and older women have higher morbidity, disability, and healthcare use than other groups.^{2, 25–27} Most sedentary behavior interventions emphasize reducing sitting time with limited focus on STST,²¹ which are an important aspect of sedentary behavior as they can increase postural blood flow and positively impact insulin and HOMA-IR.^{6, 28–30} Interrupting prolonged sitting time with postural changes that involve transitioning from sitting to standing may reduce blood pressure physiologically through greater sheer stress which could improve endothelial function.^{13, 31} Further, real-world interventions to reduce sitting behaviors can improve blood pressure,^{32, 33} though more evidence is needed.³⁴ It is important to understand how changing different facets of sitting behaviors can improve the health of older women given the lack of prior studies.

The Rise for Health Study is the first randomized controlled trial to evaluate two distinct approaches for changing sitting behaviors and examine the physiologic impacts among postmenopausal women. This 3-month, three-arm trial examined: (a) reducing overall sitting time; or, (b) increasing sit-to-stand transitions; in comparison to (c) a healthy living control group. The primary aim was to evaluate changes in blood pressure and glucoregulatory biomarkers (fasting plasma insulin and glucose, glycated hemoglobin (HbA1c), insulin resistance (homeostatic model assessment 2 of insulin resistance (HOMA2-IR)) for the two intervention arms compared to the attention-control condition. We hypothesized that, compared to those allocated to the attention-control arm, participants in the two intervention arms would have greater improvements in blood pressure and glycemic control.

Methods

Rise for Health tested two interventions to change sitting patterns in overweight and obese postmenopausal women over 3-months by either reducing sitting time or increasing STST, compared to an attention-control condition. Data were collected from June 2018 to August 2022 at the University of California San Diego (UCSD). This study is one of three projects within a Program Grant examining interrupting sitting to support healthy aging (P01AG052352). The UCSD Institutional Review Board approved study procedures and all participants provided written informed consent. Study recruitment, safety, and progress were reviewed semi-annually by an external independent NIH- appointed Data Safety Monitoring Board. Data sharing is available upon written request, approval of the study investigators, and execution of a data sharing agreement.

Eligibility criteria included: (1) age 55 year or older and postmenopausal, defined as no menstrual period in the last 12 months; (2) >=7 hours of sitting time and <=70 STST on a majority of activPAL-measured days; (3); body mass index (BMI) ≥ 25 kg/m² and < 45 kg/m² or if identified as Asian, BMI ≥ 23 kg/m² and < 45 kg/m²; (4) no safety contraindications, determined by the Short Physical Performance Battery (SPPB). Exclusion criteria included: (1) use of insulin; (2) uncontrolled diabetes defined as HbA1c > 10%; (3) uncontrolled blood pressure defined as SPB > 180 or DBP > 110; (4) history of deep vein thrombosis; and (5) participating in another research study or program that would impact the outcomes of this study.

Protocol

A detailed description of the study protocol was previously published.³⁵ Briefly, potential participants were telephone-screened for eligibility, followed by an in-person visit. At this Screening Visit, participants provided signed informed consent, height, weight, and blood pressures were taken. The SPPB was used to assess safety for balance and STST, and they were given an activPAL^™, a thigh-worn accelerometer to measure posture and sedentary behaviors, and an ActiGraph GT3X+, a hip-worn accelerometer to measure physical activity for the next 7 days. Participants returned to the clinic about 1 week later where devices were screened for eligibility. HbA1c eligibility was then checked via finger prick. Those who met all eligibility criteria then had a fasting venous blood draw and completed any remaining baseline measures. All baseline measures were repeated at the 3-month visit.

Participants were randomized (1:1:1) within REDCap to one of three study arms: Reduce sitting time (Sit Less arm), Increase STST (Sit-to-Stand arm), or control arm (Healthy Living arm). Randomization was stratified by BMI (overweight vs obese) and employment status (full-time vs. not full-time) using a permuted block randomization scheme created by the Biostatistics Core of the P01. Data collectors and the Principal Investigator were blinded to study group assignment.

Intervention Arms: Sit Less and Sit-to-Stand

Participants received five in-person, individual coaching sessions (weeks 1-4, 8) and 2 telephone sessions (weeks 6, 11). Intervention participants wore an activPAL^w during weeks 1-4 and 7-8 for behavioral monitoring, goal setting, and personalized feedback. A toolbox of resources was provided to support behavior change, which included a wrist worn device (e.g. Watchminder), timers, and cue cards. Participants in the Sit Less arm were also offered a standing desk or tray. The primary goal of the interventions was to interrupt current sitting patterns by targeting a specific behavior change: reducing the total amount of time spent sitting (Sit Less arm) or increasing the total number of STST each day (Sit-to-Stand arm). Both interventions utilized habit formation,³⁶ social cognitive theory,³⁷ and motivational interviewing techniques to support behavior change. In-person sessions lasted about 60 minutes and telephone sessions were about 30 minutes. For further details see Hartman et al.³⁵

Healthy Living Arm

Participants received an equal number of contacts on the same timeline as those in the intervention arms. The first session (week 1) was in-person and the remaining sessions (weeks 2-4, 6, 8, 11) were delivered over the telephone. To support retention, participants were provided with a menu of various healthy aging topics to choose from including safe driving, stress reduction, and healthy bones. At each session they could pick a topic from the list they wanted to discuss with their coach. Coaches reviewed the learning objectives for the topic, provided information and work sheets, and ended the session by setting a personalized goal and action plan related to the topic. The in-person session was about 60 minutes, and the telephone sessions were about 30 minutes each. For further details see Hartman et al.³⁵

Measures and Outcomes

The primary outcomes of glucose regulation and resting blood pressure were assessed at baseline and 3 months. Glucose regulation was assessed using measurements of fasting plasma insulin and glucose, HbA1c, and HOMA2-IR. HbA1c was measured real-time in fresh whole blood (DCA Vantage, Siemens). Fasting plasma glucose and insulin concentrations were measured after blood sample collection was completed. Fasting plasma glucose concentration was determined using a glucose oxidase method (YSI 2900D Biochemistry Analyzer, YSI Inc., Yellow Springs, OH). Fasting plasma insulin concentration was determined using a high sensitivity immunoassay (Meso Scale Discovery, Rockville, MD, catalog # K1516HK). Intraplate and interplate coefficients of variance (CV) for glucose and insulin were (4.0%, 5.4%) and (7.1%, 18.0%). HOMA2-IR, a model-derived estimate of insulin resistance, was calculated using fasting plasma glucose and insulin concentrations entered into the HOMA2 calculator.^{38, 39} Resting blood pressure was measured using the Dinamap / Accutor 7 or Dinamap V100 blood pressure monitor after participants had rested while seated for at least five minutes. Blood pressure was measured 3-4 times and the mean of all available measures was used.

The activPAL (PAL Technologies, Glasgow, UK), a triaxial thigh-worn accelerometer, was used to objectively measure sedentary behaviors, including sitting time and STST at baseline and 3-month assessments. Event files from the activPAL were extracted via the CREA classification algorithm (PALanalysis, v8), which was set to require ≥ 4 seconds for a new posture to be registered, and generated sleeping time for removal from analyses. Minutes spent in postures and movement (i.e., sitting, standing, STST, stepping time) were recorded continuously, and then aggregated to produce the participant-level sedentary behavior metrics.⁴⁰ A day was considered valid if it had at least 10 hours of awake wear time. Participants without 4 or more valid days were asked to re-wear the activPAL. The activPAL has demonstrated good reliability and validity for assessing these behaviors.^{41, 42}

The Actigraph GT3X+ (ActiGraph LCC, Pensacola, FL, USA), a triaxial hip worn accelerometer, was used to measure moderate-to-vigorous physical activity (MVPA) minutes. A day was considered valid if it had at least 600 minutes of awake wear time. ActiGraph data were processed with the normal filter and aggregated to 15-second epoch files of awake wear time via OPACH cutpoints which were specifically calibrated for older women,⁴³ defined as vector magnitude (VM) ≤ 18 (sedentary), VM 19 through VM 518 (light physical activity), and VM ≥519 (MVPA). Actigraph has been validated⁴⁴ with good reliability⁴⁵ for measuring MVPA in adults under free-living conditions. The 15-second epoch data were aggregated to produce participant-level daily summaries of MVPA minutes.

Body Mass Index (BMI) was calculated as weight (in kg)/height (in m) ² at the baseline and 3-month visits. Height was measured to the nearest 0.1 cm with a wall-mounted stadiometer and weight was measured to the nearest 0.1 kg on a digital medical scale. Participants’ self-reported demographics including age, education, income, race/ethnicity, and marital status at baseline, and they self-reported comorbid health conditions.

COVID-19 Modifications

Due to the COVID-19 pandemic, in-person visits were briefly paused. Participants scheduled for their 3-month visit during office closures completed a remote visit. They were mailed an automated blood pressure machine to measure blood pressure during a Zoom call with study staff and were mailed activPAL and Actigraph devices to wear for 7 days. They were offered extended phone sessions every other week until they were able to come to an in-person visit for their blood draw. If the in-person visit occurred more than 28 days after the remote visit, participants repeated the blood pressure measurement and wearing of the activPAL and Actigraph. For analyses, in-person data was used when both were available. Only 13 participants had a delayed in-person visit with both remote and in-person assessments, and 6 had a remote assessment only.

Sample size Considerations

Given the multiple correlated outcomes, we created two composite scores derived as a mean of Z-scores of each of the sets of (possibly transformed) glucoregulatory (insulin, glucose, HbA1c, HOMA2-IR) and blood pressure markers. We estimated that a target sample size of 405 participants (135 per arm) would have 80% power to detect a 0.4 effect-size for changes in composite glucoregulatory and blood pressure outcomes between the 2 intervention arms compared to the Healthy Living arm with significance level α=0.025 for 2 comparisons (each of the sitting interventions compared to control), and allowing for 10% drop-out /missing data rate (details provided in Hartman et al.³⁵)

Statistical Analyses

Baseline characteristics were compared across the three treatment arms using ANOVA (for continuous) and chi-square or Fisher exact tests (for categorical) variables. Study-related adverse events were tabulated by group using standard MEdDRA (Medical Dictionary for Regulatory Activities) body system categories.

The primary analysis comparing the Sit Less and Sit-to-Stand interventions to the Healthy Living controls used the intent-to-treat (ITT) principle. Linear mixed-models (LMM) were applied, allowing for the inclusion of partially complete records, thereby reducing complete case-analysis biases. The model included baseline and 3-month outcomes as dependent variable; time (baseline, 3-month), group (active, control), and the group*time interaction were fixed-effect independent (categorical) variables. A subject-specific (random) intercept was included to model individual heterogeneity in outcomes. A significant group*time interaction indicated that outcome changes differed between that intervention and the control arm; this difference of outcome changes between intervention and control arms represents the intervention effect.

To reduce the chance of false positives, we first fit the aforementioned LMMs for two primary composite outcomes: (i) glucoregulatory outcomes defined as the average of four Z-scores of insulin, glucose, HbA1c and HOMA2-IR, (ii) blood pressure composite defined as the average of two Z-scores of systolic and diastolic blood pressure. The Z-score for each participant for a specific marker and time-point was derived by subtracting the sample baseline mean from the participant’s marker value at that time-point, and dividing this difference by the sample baseline SD. This composite score analysis represents an omnibus test of intervention efficacy across multiple (correlated) outcomes -- rejection of the null hypothesis would suggest that the intervention (compared to controls) effected changes across components of the composite score. Following this composite score analysis, we compared group differences for each of the outcomes; we present intervention results for the individual markers.

Several secondary analyses were conducted. First, we evaluated intervention effects adjusted for baseline stratification variables (obesity status, employment status). Second, we evaluated group differences in outcomes between the two intervention arms. Third, we evaluated changes in physical activity (standing, stepping, Light PA, MVPA) by intervention arm. Fourth, we included an indicator variable for COVID-19 (yes/no) and evaluated 3-way COVID-19*group*time interaction (via LRTs) to test if changes in outcomes differed between participants with delayed 3-month assessment due to COVID-19 and those who received the original protocol. Fifth, we explored if baseline hypertension status was an effect modifier for blood pressure and if baseline glycemic status (HbA1c < or >=5.7%) was an effect modifiers for glucoregulatory outcomes. We also excluded participants who self-reported having type 2-diabetes at baseline and repeated the LMM analysis. Finally, we examined if changes in MVPA altered the intervention effect estimates by including MVPA at baseline and 3-month as a time-varying covariate, and the MVPA*group interaction in the LMMs.

Gaussian link functions were used in the mixed models and diagnostic plots were generated to evaluate model assumptions. Insulin, glucose, and HOMA2-IR were log-transformed to better approximate Gaussian distributions and results were presented as geometric means with 95% confidence intervals (CIs) derived from model estimates (β coefficients) at baseline, and 3 months, along with mean (95% CI) within group changes. For the other outcomes, sedentary behavior measures and blood pressure, mean (95%CI) at baseline and 3 months, as well as mean (95% CI) within group-change derived from model-parameters (β-coefficients) are presented. Treatment effects were calculated as the mean (95% CI) of the difference (compared with the control group) in absolute changes (for all outcomes) and percent changes (for log-transformed outcomes) from baseline to 3 months.

Analyses were implemented in the R statistical programming language and LMMs were executed via R package nlme.^{46, 47} All hypothesis tests were two-sided. We used significance level α = 0.025 for the primary analyses to adjust for multiple comparisons (2 intervention groups versus control). No multiple comparison adjustment was made for secondary analyses.

Results

A total of 2028 women were telephone screened for eligibility; of those, 914 were eligible, and 620 attended the screening visit. See CONSORT diagram (Figure 1). Most common reasons for ineligibility, based on telephone interview, included unable to sit or stand for short or long periods (n = 211), BMI too low (n=203), and self-reported insufficient sedentary time (n=131). After attending the in-person visit, 213 were deemed ineligible with primary reasons being: too many activPAL measured STSTs (n=56), too little activPAL measured sedentary time (n=53), not interested/declined to continue (n=28), no-showed for visit (n=11), and pain when standing or failed the SPPB (n=10). A total of 407 participants were randomized, with 136 randomized to the Sit Less arm, 136 to the Sit-to-Stand arm, and 135 to the Healthy Living arm. Adherence to the seven coaching sessions was high with 95% attendance in the Sit Less arm, 92% attendance in the Sit-to-Stand arm, and 93% attendance in the Healthy Living arm. For details on use of study tools see Supplemental Table 1. Overall retention rate was 95.3% (n=388) with 6 participants lost to follow up in the Sit Less arm, 8 lost to follow up in the Sit-to-Stand arm, and 5 lost to follow up in the Healthy Living arm. See CONSORT diagram (Figure 1). Six participants had a remote assessment only due to COVID-19 (Sit Less n=1, Sit-to-Stand n=1, Healthy Living n=4). In-person 3-month visits for 13 were delayed due to COVID-19 closures, and all 13 completed a remote visit and a subsequent in-person visit. Of these 13, four were in the Healthy Living Arm and all 4 had additional coaching calls (ranging from 1 to 2 extra calls), 5 were in the Sit Less arm and 3 had additional coaching calls (ranging from 1 to 4 extra calls), and 4 were in the Sit-to-stand arm and 2 had additional coaching calls (3 extra calls each).

Figure 1.

CONSORT diagram.

Baseline characteristics stratified by randomization arm are shown in Table 1. Overall, participants averaged 68 years of age (SD=7.23), ranging from 55 to 91 years, predominantly white (92%), non-Hispanic (91%), with a college education or greater (72%). BMI averaged 32 kg/m² (SD=4.9), ranging from 25 to 45 kg/m². At baseline, 49% self-reported a diagnoses of hypertension, 9% diabetes, 57% arthritis, and 36% reported having had 1 or more falls in the past 12 months. There were no significant differences between the three study arms in baseline demographic characteristics (p>0.05).

Table 1:

Participant baseline characteristics overall and by randomized arm (N=407)

	Overall	Healthy Living	Sit Less	Sit-to-Stand	p*
n	407	135	136	136
Age (mean (SD))	67.81 (7.23)	67.76 (7.43)	68.39 (7.66)	67.29 (6.55)	0.45
BMI (mean (SD))	32.13 (4.93)	32.33 (5.13)	32.00 (4.64)	32.06 (5.06)	0.85
Race (%)					0.87
White	371 (92.1)	120 (90.2)	126 (94.0)	125 (91.9)
Black or African-American	15 (3.7)	7 (5.3)	4 (3.0)	4 (2.9)
Asian	10 (2.5)	3 (2.3)	2 (1.5)	5 (3.7)
American Indian/Alaskan Native	2 (0.5)	1 (0.8)	0 (0.0)	1 (0.7)
More than one race	5 (1.2)	2 (1.5)	2 (1.5)	1 (0.7)
Latina/Hispanic (%)	35 (8.6)	12 (8.9)	13 (9.6)	10 (7.5)	0.82
Employed Full-time (%)	101 (24.8)	32 (23.7)	34 (25.0)	35 (25.7)	0.93
Education (%)					0.58
High School diploma or less	113 (27.8)	33 (24.4)	43 (31.6)	37 (27.2)
College degree	99 (24.3)	38 (28.1)	28 (20.6)	33 (24.3)
Graduate degree	195 (47.9)	64 (47.4)	65 (47.8)	66 (48.5)
Have arthritis (%)	232 (57.0)	75 (55.6)	81 (59.6)	76 (55.9)	0.76
Have hypertension (%)	201 (49.4)	68 (50.4)	72 (52.9)	61 (44.9)	0.40
Currently taking hypertension medication, of those with hypertension (%)	163 (81.1)	54 (79.4)	60 (83.3)	49 (80.3)	0.83
Have diabetes (%)	35 (8.6)	18 (13.3)	11 (8.1)	6 (4.4)	0.03
Currently taking diabetes medication, of those with diabetes (%)	24 (68.6)	14 (77.8)	8 (72.7)	2 (33.3)	0.19
Number of falls in the past year (%)					0.43
0	260 (63.9)	83 (61.5)	92 (67.6)	85 (62.5)
1	99 (24.3)	38 (28.1)	26 (19.1)	35 (25.7)
2	38 (9.3)	9 (6.7)	15 (11.0)	14 (10.3)
>=3	10 (2.5)	5 (3.7)	3 (2.2)	2 (1.5)

^p*:for continuous variables, one-way anova; for categorical variables, chi.sq test; fisher test for race and number of falls

There were no serious adverse events that were classified as related or possibly related to the study, and seven serious adverse events that were classified as not related (3 in the Healthy Living arm, 3 in the Sit Less arm, and 1 in the Sit-to-Stand arm). For nonserious adverse events that were related or possibly related to the study, the most common was a rash related to wearing the activPAL (58% of adverse events), with 16 in the Sit Less arm, 18 in the Sit-to-Stand arm, and 8 in the Healthy Living arm. The second most common related or possibly related event category was musculoskeletal issues (e.g., pain) representing 27% of adverse events with 12 events reported in the Sit Less arm and 8 reported in the Sit-to-Stand arm, followed by injury (7% of events), with 4 events reported in the Sit Less arm and 1 in the Sit-to-Stand arm. Other than skin irritation from the activPAL, there was one related adverse event in the Healthy Living arm which was a swollen finger from the finger prick at the Baseline Visit. Of all the adverse events possibly or definitely related to the study, 83.6% were classified as mild, 16.4% as moderate; and more than 75% were resolved by the study end.

Behavioral Outcomes

At baseline, participants averaged 661.77 min/day sitting (SD=102.15), 199.10 min/day standing (SD=75.70), and 44.16 sit-to-stand transitions (STST) per day (SD=10.99). There were no significant between-group differences at baseline (p>0.05). Adherence to wearing the activPAL and ActiGraph were high. At baseline, participants averaged 6.21 days (SD= 1.07) of valid activPAL wear and 7.32 days (SD=1.66) of valid ActiGraph wear. At 3-months, participants averaged 5.89 days (SD= 1.00) of valid activPAL wear and 6.90 days (SD=1.59) of valid ActiGraph wear. There were no significant differences between the 3 study arms for valid wear days for either device at either time point (p>0.05).

Sitting Behavior Metrics: Primary Behavioral Outcomes

The Sit Less arm had significantly greater reductions in total daily sitting time (654min/day baseline, 579min/day 3-months) compared to the Healthy Living arm (669min/day baseline, 652min/day 3-months); the intervention effect was −58min/day (95% CI [−82.92, −33.59], p<0.001). See Figure 2a. There were no significant differences in the Sit-to-Stand arm (662min/day baseline, 635min/day 3-months) compared to the Healthy Living arm for changes in total sitting time; the intervention effect was −10 min/day (95% CI [−34.56, 14.86], p=0.44). In contrast, the Sit-to-Stand arm had significantly greater increases in total daily STST (45 STST/day baseline, 71 STST/day 3-months) compared to the Healthy Living arm (43 STST/day baseline, 43 STST/day 3-months); the intervention effect was 26 STST/day (95% CI [17.71, 33.64], p<0.001). There was no significant difference in change in STST between the Sit Less arm (44 STST/day baseline, 43 STST/day 3-month) and the Healthy Living arm; the intervention effect was −1.5 STST/day (95% CI [−9.40, 6.50], p=0.72). See Figure 2b. Both groups significantly reduced minutes per day spent in long sitting bouts (>30min) (p<0.01), whereas only the Sit-to-Stand arm significantly reduced mean duration of each sitting bout (p<0.01) compared to the Healthy Living arm (See Supplemental Table 2).

Figure 2.

Baseline and 3 month values, by study arm, for minutes per day of sitting time (2a), number of sit-to-stand transitions per day (2b), minutes standing per day (2c), and minutes of moderate to vigorous physical activity (MVPA) per day (2d).

When comparing the two interventions, the Sit Less arm had significantly greater reductions in total daily sitting time compared to the Sit-to-Stand arm; the difference of the group changes was −48min/day (95% CI [−73.23, −23.58], p<0.001). The Sit-to-Stand arm had significantly greater increases in total daily STST compared to the Sit Less arm; the difference of the group changes was 27STST/day (95% CI [19.14, 35.12], p<0.001). The Sit-to-Stand arm had significantly greater reductions in mean sitting bout duration (p<.01) but there was no significant difference between the two intervention arms for change in long sitting bouts. See Supplemental Table 2 for full behavior change data and analyses.

Standing and Physical Activity Metrics

Differences in standing and physical activity were examined to determine changes to behaviors not directly targeted by the interventions. The Sit Less arm had significantly greater increases in total daily standing time compared to both Healthy Living arm and Sit-to-Stand arm. The intervention effect of Sit Less was 57min/day, (95% CI [38.83, 76.05], p<.001), and the difference between the two intervention group changes (Sit Less arm versus Sit-to-Stand arm) was 41min/day (95% CI [22.26, 59.72], p<.001). See Figure 2c. There were no significant differences over time in the Sit-to-stand arm compared to the Healthy Living arm for total standing time; the intervention effect was 16min/day (95% CI [−2.20, 35.10], p=0.09). For MVPA, the Sit-to-Stand arm significantly increased minutes of MVPA compared to the Healthy Living arm. The intervention effect was 6min/day (95% CI [0.85,11.29], p=0.02), but it did not significantly differ from the Sit Less arm, 4min/day (95% CI [−0.78, 9.62], p=0.10). See Figure 2d. There were no significant differences among the three groups for changes in BMI, Light PA, or stepping time (p>0.05). See Supplemental Table 2.

Physiological Outcomes

At baseline, HbA1c was significantly different between the three study arms (p=0.001); Healthy Living=mean 5.80 (SD=076), Sit Less=5.64 (SD=0.53), Sit-to-Stand=5.54 (SD=0.36). There were no significant differences between the groups at baseline for blood pressure or the glucose regulation outcomes.

Composite Outcomes

The intervention effects for the composite Z-score outcome for glucose regulation were null (mean intervention effect (SE) 0.02 (0.06), p=0.74 for the Sit Less arm, and −0.05 (0.07), p=0.46 for the Sit-to-Stand arm). For the composite blood pressure outcome, the Sit-to-Stand arm had significant decreases compared to the Healthy Living arm (−0.21 (0.08) p=0.01); the composite blood pressure score for the Sit-Less arm also decreased but did not reach the significance threshold (mean (SE) intervention effect (−0.15 (0.08), p=0.07).

Blood Pressure Outcomes

The Sit-to-Stand arm significantly decreased diastolic blood pressure compared to the Healthy Living arm. The intervention effect was −2.24 mmHg (95% CI [−4.08, −0.40], p=0.02). This significant effect for diastolic blood pressure remained (intervention effect=−2.11 mmHg, 95% CI [−4.01, −0.21], p=0.02) after adjusting for MVPA (at baseline and 3-months as a time-varying covariate) and the MVPA*group interaction in the model. Systolic blood pressure also decreased but did not meet the a priori statistical significance level of 0.025; the intervention effect was −3.33 mmHg (95% CI [−6.32, −0.33], p=0.03). MVPA-adjusted intervention effect for systolic blood pressure in the Sit-to-Stand arm was −3.08 mm Hg (95% CI [−6.15, −0.01], p=0.05). The Sit Less arm decreased systolic and diastolic blood pressure, although it did not reach statistical significance; the intervention effects were −2.23 mmHg (95% CI [−5.22, 0.76], p=0.14) and −1.70 mmHg (95% CI [−3.53, 0.14], p=0.07), respectively. There were no significant differences between the two intervention arms for systolic or diastolic blood pressure. See Table 2. Exploratory analyses stratifying by self-reported hypertension status found no significant effect modification (See Supplemental Table 3). Adjusting for MVPA did not alter the findings for the Sit-Less arm nor the between intervention results (See Supplementary Table 6).

Table 2:

Differences in blood pressure measures between the three arms from baseline to 3-months (N=407).

outcome		Healthy Living (n=135)	Sit Less (n=136)	Sit to Stand (n=136)	Sit Less compared to Healthy Living	Sit-to-Stand compared to Healthy Living	Sit-to-Stand compared to Sit Less
Diastolic blood pressure (mmHg)
	Baseline	76.11 (74.33,77.89)	75.73 (73.95,77.50)	77.52 (75.75,79.30)
	3 months	75.94 (74.14,77.74)	73.86 (72.06,75.66)	75.11 (73.31,76.92)
	Absolute Change	−0.17 (−1.47, 1.13)	−1.87 (−3.17, −0.57)	−2.41 (−3.72, −1.10)	−1.70 (−3.53, 0.14)	−2.24 (−4.08, −0.40)	−0.54 (−2.39, 1.30)
	P				p = 0.07	p = 0.02	p = 0.56
Systolic blood pressure (mmHg)
	Baseline	125.00 (122.45,127.55)	125.40 (122.86,127.94)	126.47 (123.93,129.01)
	3 months	125.99 (123.41,128.58)	124.16 (121.58,126.75)	124.13 (121.54,126.72)
	Absolute Change	0.99 (−1.12, 3.10)	−1.24 (−3.35, 0.88)	−2.34 (−4.46, −0.21)	−2.23 (−5.22, 0.76)	−3.33 (−6.32, −0.33)	−1.10 (−4.09, 1.90)
	P				p = 0.14	p = 0.03	p = 0.47

Values and absolute changes are estimated means and their changes (95%CI) from LMMs.

p-values from LMMS are for between group comparisons of the change. p < 0.025 signifies rejection of the null hypothesis based on a priori alpha level of 0.025 to adjust for two intervention versus control comparisons.

Glucoregulatory Outcomes

There were no statistically-significant intervention effects for any of the glucoregulatory markers (p > 0.05) for either intervention arm compared to the Healthy Living arm. Given the group differences at baseline, HbA1c was assessed via a linear model with 3-month value as the dependent variable, group as main effect adjusting for baseline HbA1c. There was no significant difference at 3-months between Sit Less versus Healthy Living (mean difference: 0.05%, 95%CI: [−0.02%, 0.12%], p=0.13) or between Sit-to-Stand versus Healthy Living (difference: 0.01%, 95%CI: [−0.05%,0.08%], p=0.68). There were also no significant differences between the two intervention arms at 3-months (mean difference: −0.04%, 95%CI: [−0.11%, 0.03%], p=0.26). Insulin, glucose and HOMA2-IR were log-transformed in the analysis, hence geometric means, and absolute and relative (percent) intervention effects are presented in Table 3. The Sit-to-Stand arm exhibited a small 3% decrease on average for insulin (mean relative change −2.84%; 95% CI: [−9.59%, 3.90%]) compared to an average 5% increase for the Healthy Living arm (mean relative change 4.94%; 95%CI: [−2.27%, 12.16%]). These changes translated to an approximate 8% (non-significant) between group difference (mean relative change difference −7.79%; 95%CI: [−17.66% , 2.09%]; p=0.12). Similar results were obtained for HOMA2-IR for the Sit-to-Stand arm compared to the Healthy Living arm, with an approximate 8% (non-significant) between group difference (mean relative change difference −8.46%; 95%CI: [−18.35%, 1.44%]; p=0.09). There were no significant differences between the two intervention arms for the glucoregulatory outcomes. These, and additional results (i.e., glucose, Sit Less arm), are presented in Table 3. Exploratory analyses excluding participants who self-reported having type 2 diabetes at baseline found no significant intervention effect on the glucoregulatory outcomes (See Supplemental Table 4). When stratified by baseline HbA1c <5.7% vs ≥5.7% there was a trend for intervention effects for Insulin with the Sit-to-Stand arm compared to Healthy Living arm, with greater reductions in Insulin among those with a baseline HbA1c ≥5.7% than those with a baseline HbA1c <5.7% (mean difference −21.6, 95%CI −41.5, −1.7, p=0.04). No other effect modifications approached significance (See Supplemental Table 5). Adjusting for MVPA did not alter the findings for the glucoregulatory markers (See Supplementary Table 6).

Table 3:

Differences in glucoregulatory measures between the three arms from baseline to 3-months (N=399).

outcome		Healthy Living (n=131)	Sit Less (n=133)	Sit to Stand (n=135)	Sit Less compared to Healthy Living	Sit-to-Stand compared to Healthy Living	Sit-to-Stand compared to Sit Less
Insulin (mU/L)
	Baseline	15.93 (14.25, 17.62)	15.68 (14.03, 17.33)	15.27 (13.68, 16.86)
	3 months	16.72 (14.94, 18.51)	15.85 (14.17, 17.53)	14.84 (13.26, 16.41)
	Absolute Change	0.79 (−0.34, 1.91)	0.17 (−0.92, 1.26)	−0.43 (−1.48, 0.61)	−0.62 (−2.19, 0.95)	−1.22 (−2.76, 0.31)	−0.60 (−2.11,0.91)
	Percent Change	4.94 (−2.27, 12.16)	1.07 (−5.92, 8.07)	−2.84 (−9.59, 3.90)	−3.87 (−13.92, 6.18)	−7.79 (−17.66, 2.09)	−3.92 (−13.63,5.80)
	P				p = 0.45	p = 0.12	p = 0.43
Glucose (mg/dL)
	Baseline	105.11 (101.95, 108.27)	101.80 (98.75, 104.86)	101.01 (98.00, 104.01)
	3 months	104.80 (101.60, 108.00)	104.00 (100.83, 107.15)	101.18 (98.10, 104.26)
	Absolute Change	−0.31 (−2.99, 2.37)	2.19 (−0.45, 4.83)	0.17 (−2.43, 2.77)	2.50 (−1.26, 6.26)	0.48 (−3.25, 4.22)	−2.02 (−5.72,1.69)
	Percent Change	−0.30 (−2.84, 2.25)	2.15 (−0.47, 4.77)	0.17 (−2.40, 2.75)	2.45 (−1.21, 6.10)	0.47 (−3.15, 4.09)	−1.98 (−5.65,1,70)
	P				p = 0.19	p = 0.80	p = 0.29
HOMA2-IR
	Baseline	2.08 (1.87, 2.30)	2.05 (1.84, 2.26)	2.00 (1.79, 2.20)
	3 months	2.20 (1.97, 2.43)	2.06 (1.85, 2.28)	1.94 (1.74, 2.14)
	Absolute Change	0.12 (−0.03, 0.27)	0.01 (−0.13, 0.15)	−0.06 (−0.19, 0.08)	−0.11 (−0.31, 0.10)	−0.17 (−0.37, 0.03)	−0.07 (−0.26,0.13)
	Percent Change	5.62 (−1.65, 12.90)	0.54 (−6.40, 7.49)	−2.83 (−9.54, 3.87)	−5.08 (−15.14, 4.98)	−8.46 (−18.35, 1.44)	−3.37 (−13.03,6.28)
	P*				p = 0.32	p = 0.09	p = 0.49

Values and absolute changes are estimated geometric means and their changes (95%CI, variance is from back-transforming the log scale by delta method).

Percentage changes are the absolute geometric mean changes relative to their baseline values.

p-values from LMMs are for between group comparisons of the change. p < 0.025 signifies rejection of the null hypothesis based on a priori alpha level of 0.025 to adjust for two intervention versus control comparisons.

Results did not change for any of the outcomes when the randomization stratification variables (obesity status, employment status) were included in the models. There were no significant COVID19 effects (3-way interaction COVID-19*group*time LRTs, p > 0.1), indicating that having received alternative protocols due to COVID-19 did not affect outcomes compared to participants who received the originally planned protocol.

Discussion

The interventions were effective at improving targeted indicators of sitting behavior. Participants in the Sit Less condition reduced their sitting time by about 58 minutes/day (relative to control) and increased their standing time commensurately. Participants in the Sit-to-Stand condition increased their STSTs by over 25 per/day and improved MVPA by 6 minutes/day (relative to control). Those in the Sit Less condition did not increase their STSTs and those instructed to increase STST did not decrease their sitting time. The distinct nature of these behaviors calls for targeted intervention approaches. Despite the behavior change achieved in both arms, glucose regulation metrics did not significantly improve in the 3-month intervention. However, both intervention arms had small, non-significant, improvements in systolic blood pressure, and the Sit-to-Stand arm significantly improved diastolic blood pressure compared to control, such that increasing daily STST by an average of 25 per day resulted in a mean 2.24mmHg reduction. While the change in diastolic blood pressure did not reach a clinically meaningful change (3-5mmHg)⁴⁸, the 2.24mmHg change achieved in just 3 months is a novel finding that could be important at a population level.

Prolonged sitting is hypothesized to reduce shear stress that can cause endothelial dysfunction, vasoconstriction, and reduced muscle activity, which can lead to insulin resistance.³¹ This can induce a low-grade inflammatory state, hypertension, and hyperlipidemia.³¹ Acute, laboratory-based studies suggest interruptions of prolonged sitting improve glucose regulation, but this free-living RCT did not show significant effects on glucose regulation. However, exploratory analyses suggest that increasing STST may be helpful for reducing insulin among those with prediabetes or diabetes. Greater behavior change over a longer period of time may be needed to impact HbA1c as evidenced by a worksite trial (mean age 40 years old) that reduced sedentary time at work by 90 minutes over six months from baseline.⁴⁹

The most notable finding was that the Sit-to-Stand arm had small but significant improvements in diastolic blood pressure. About half the sample had a hypertension diagnosis with the majority taking hypertension medication, which could have limited our ability to demonstrate more sizeable reductions in blood pressure. A prior sit less intervention in older adults that was 6-months long showed improvements in systolic blood pressure.²⁰ This 3-month month intervention may have been too short for physiologic adaptations to sitting reduction to occur. Consistent with this interpretation, another 3-month intervention that achieved about a 1 hour reduction in sitting time also found no improvements in blood pressure,³⁴ suggesting longer sitting interventions may be necessary.

The Sit Less arm’s 58 min/day greater reduction in sitting time than the Healthy Living arm is comparable to, or slightly greater than prior studies in older adults (60+) that observed ~45 minute reductions in sitting time than control.⁵⁰ Studies that achieved greater reductions in sitting have predominantly been conducted in worksite populations with average ages in their 40s.^{34, 49} Prior interventions have not targeted STST and have not found changes in STST when focusing on sitting time. By specifically targeting STST, using cues from wrist-worn timers and emphasizing multiple STST during one break, we achieved an increase of 25 STST/day compared to the Healthy Living arm. Future studies can build on this approach as the sit-stand therapeutic exercise is known to benefit lower extremity strength, which is an important ability to maintain with aging in order to prevent frailty.^{51, 52}

It may be useful to speculate why only the Sit-to-Stand arm had a significant diastolic blood pressure effect. One possibility is suggested by the pattern of secondary behavioral outcomes. The Sit Less arm increased standing time while the Sit-to-Stand arm decreased the average length of sitting time and increased their physical activity. Both the act of standing up and the increased activity observed require a higher intensity of muscle activation than does standing still.⁵³ Furthermore, sitting for long durations can increase blood pressure.⁵⁴ Greater muscle activation, reduced sitting bouts, and related physiological processes⁵⁵ might explain why favorable diastolic blood pressure effects were observed only in the Sit-to-Stand condition.

Limitations of our study include a lack of racial and ethnic diversity which limits generalizability. The intervention may have been too short to observe sizeable physiologic changes given the time needed for downstream outcomes to shift. Normotensive individuals and non-diabetes were included which could limit our ability to detect differences in glucoregulatory outcomes; decreases within the healthy range can be difficult to achieve.⁵⁶ Another limitation was the use of fasting glucoregulatory measures only which cannot provide insights into changes in postprandial glucose regulation. Future research utilizing continuous glucose monitoring could provide greater insights into the relationship between glucoregulatory outcomes and sedentary behaviors. Although this was a randomized trial, we did not have measures of diet or sleep to be able to assess any potential impact on results. Our inclusion criteria aimed at enrolling people who had high amounts of sitting and low amounts of transitions may limit generalizability to others not meeting the cutoffs. Strengths included the rigorous design, testing of multiple sedentary behavior targets, and use of device-assessed sedentary behavior and physical activity measures. The focus on postmenopausal women is a strength, as cardiovascular disease is the leading cause of death among women, but it remains unstudied and undertreated in women.⁵⁷ However, future studies are needed to confirm these results in men. The high retention rate and significant behavior changes demonstrate the feasibility, acceptability, and safety of targeting sitting and STST in postmenopausal overweight and obese women.

Conclusion

Postmenopausal women are at high-risk of engaging in large amounts of sitting time and cardiovascular diseases. The present RCT adds to existing evidence by demonstrating that within just 3 months, increasing STST can lower diastolic blood pressure. The Rise for Health Trial supports the effectiveness of interventions to modify the distinct dimension of sitting behaviors including sitting time and STST. Future studies can usefully extend these findings by examining extended intervention periods and follow-ups, replicating and comparing findings in older men and women, and focusing on clinical populations such as those with pre-hypertension or pre-diabetes where behavior change is often a primary intervention and health markers may be more amenable to change.

Clinical Perspective:

What Is New?

• In an RCT, increasing daily sit-to-stand transitions improved blood pressure in overweight and obese post-menopausal women.

• Decreasing total time spent sitting each day did not improve blood pressure or glucoregulatory measures (insulin, glucose, and HOMA2-IR).

• Sitting time and sit-to-stand transitions are distinct behaviors, and efforts to change either behavior did not have any carryover changes to the other behavior.

What are the clinical implications

• Encouraging and assisting overweight and obese postmenopausal women to increase the number of times they engage in sit-to-stand transitions each day could be beneficial for lowering blood pressure.

Non-Standard Abbreviations and Acronyms:

STST

sit-to-stand transitions

MVPA

moderate to vigorous physical activity

UCSD

University of California San Diego

RCT

randomized controlled trial