Does a Physical Therapist-Administered Physical Activity Intervention Reduce Sedentary Time after Total Knee Replacement: An Exploratory Study?

Abstract

Design:

Exploratory secondary data analysis of a pilot randomized control trial conducted between 2015 and 2017.

Setting:

Outpatient physical therapy clinic in an academic setting.

Participants:

43 participants were randomized to the intervention or control group. Both groups received standard physical therapy (PT) after total knee replacement (TKR).

Interventions:

Participants in the intervention group received a Fitbit, weekly step goals from their physical therapist, and up to six monthly phone calls after their discharge from PT (DC).

Main Outcome Measures:

The outcome, change in sedentary time (∆SED), was measured using a triaxial accelerometer (Actigraph GT3X) from initial evaluation (IE) to DC (short-term), and IE to 12 months (12M) after DC (long-term). Positive values represent a reduction in SED. We compared the short- and long-term ∆SED between the intervention and control groups using independent t-tests.

Results:

Of the 43 participants [mean(SD) age 67.0 (7.0) years, BMI 31.5 (5.9) kg/m², 53% female], 36 participants had data at IE and DC (18 intervention, 18 control) and 27 had data at IE and 12M (12 intervention, 15 control). The between group difference in short-term ∆SED was [mean(95%CI)] 32.9 (−14.4, 80.1) minutes/day. The between group difference in long-term ∆SED was 59.2 (8.6, 109.7) minutes/day.

Conclusion:

Our preliminary results indicate that a physical therapist-administered physical activity intervention may reduce SED in adults after TKR. However, a full-clinical trial is needed to establish the effect on SED.

Introduction:

Worldwide, over 250 million adults have knee osteoarthritis (OA), which is a serious disease given it is painful, causes functional limitation, and has no cure (Lee et al., 2012). All current treatments focus on reducing pain and functional limitation, with the definitive treatment for end-stage knee OA being total knee replacement (TKR). Despite improvements in pain and physical function after TKR, many adults continue to be highly sedentary, i.e., spending the majority of waking time sitting or lying on a couch, and physically inactive, i.e., engaging in little moderate-to-vigorous intensity physical activity (MVPA) (George, Ruiz, & Sloan, 2008; Hammett et al., 2018). High levels of sedentary time (SED) and little to no time in MVPA are independently associated with increased risk of chronic diseases, all-cause mortality, and disability (Biswas et al., 2015; Dunlop et al., 2015; Semanik et al., 2015).

To promote engagement in physical activity (PA) after TKR, a physical therapist-administered PA intervention was developed (Christiansen et al., 2019). The findings of this pilot trial indicated that adults who received the PA intervention during physical therapy (PT) after TKR increased their time in MVPA more than the control group, from initial evaluation (IE) to discharge (DC) from PT, and 12 months (12M) after DC from PT (Christiansen et al., 2019). However, change in SED (∆SED) following the intervention is not known. This investigation is important because people can have improvements in MVPA but continue to be sedentary. Therefore, the purpose of this study is to investigate the short- and long-term ∆SED following a physical therapist-administered PA intervention in people after TKR. We define short-term as IE to DC from PT, and long-term as IE to 12M after DC from PT. We hypothesized that the intervention group will have a greater reduction in SED compared to the control group in the short- and long-term.

Methods:

This study was an exploratory secondary data analysis of a pilot randomized controlled trial conducted between 2015 and 2017. The detailed description and findings of this trial have been previously published (Christiansen et al., 2019). We intended to recruit 72 participants (36 participants in each group), however, we had to stop the enrollment at 43 participants since we received funding for a larger trial to investigate the effectiveness of a physical therapist-administered PA intervention. Briefly, we included adults undergoing outpatient PT in an academic setting in the United States after a unilateral TKR who were ≥ 45 years of age, were not planning or did not have another lower extremity surgery for six months after TKR, and were interested in increasing their PA. After providing informed consent, participants were randomized to the intervention or control group. Participants were excluded if they had a comorbidity that would prevent them from participating in a PA intervention such as unstable angina. Both groups received standard PT after TKR and up to 6 monthly follow-up phone calls after DC from PT. The intervention group also received a Fitbit Zip, weekly step goals during PT, and a discussion within each follow-up phone call to set step goals after DC from PT.

We measured SED using a triaxial accelerometer, i.e., Actigraph GT3X (Pensacola, FL, USA), worn for 7 days at 3 timepoints: IE, DC from PT, and 12M from DC. Valid accelerometry data requires ≥10 hours/day of wear for ≥ 4 days (Troiano et al., 2008). SED was defined as time spent in <100 activity counts/minute. We standardized SED to 16 waking hours/day [(time spent in <100 activity counts/minute ÷ wear time/day) * 16 hours] to minimize variability in wear times between participants. Our primary outcome, ∆SED, was calculated as the difference in SED from IE to DC (short-term ∆SED) and IE to 12M (long-term ∆SED; positive values indicated a reduction in SED.

In this analysis, we included participants with valid accelerometry data at IE and DC for short-term ∆SED, and at IE and 12M for long-term ∆SED. To contextualize our findings, we also reported short- and long-term ∆MVPA for the participants included in our analysis. These data were previously reported in full as all data points were considered in the analysis (Christiansen et al., 2019). For ∆MVPA, positive values indicate an increase in time spent in MVPA.

Descriptive statistics were used to define the demographic characteristics of the sample. We computed means and 95% confidence intervals (95% CI) to quantify the short- and long-term ∆SED and ∆MVPA. We compared ∆SED and ∆MVPA between the intervention and control groups using independent t-tests. The analysis was conducted in SAS version 9.4 (SAS Institute Inc, Cary, NC, USA).

Results:

Of the 43 participants [mean(SD) age 67.0 (7.0) years, BMI 31.5 (5.9) kg/m², 53% female], 36 participants had valid accelerometry data at IE and DC (18 intervention, 18 control), and 27 had valid accelerometry data at IE and 12M (12 intervention, 15 control). At IE, participants in the intervention group has similar SED compared to those in the control group. About half of the total sample had one or more comorbidities including cardiovascular disease, COPD, stroke, cancer, history of falls, diabetes mellitus, or depression. The frequency of comorbidity was similarly distributed between the intervention and control groups.

In the short-term (IE to DC), SED decreased while time in MVPA increased for the intervention and control groups (Table 1). Further, participants in the intervention group demonstrated greater reductions in SED (mean difference 32.9 [−14.4, 80.1] minutes/day), and greater improvements in MVPA (mean difference 14.3 [2.2, 26.5] minutes/day) compared to the control group (Table 1).

Table 1:

Short- and long-term changes in standardized sedentary time (SED) and time in moderate-to-vigorous intensity activity (MVPA). Positive number indicates reduction in SED and increase in MVPA.

	Short-term				Long-term
	IE	DC	1Change	2Difference	IE	12M	1Change	2Difference
Number	18	18			12, 15**	12, 15**
	Mean (SD)	Mean (SD)	Mean (95%CI)	Mean (95%CI)	Mean (SD)	Mean (SD)	Mean (95%CI)	Mean (95%CI)
SED (minutes/day)
Intervention	742.6 (64.8)	651.6 (70.7)	91.1 (54.3, 127.8)		733.2 (69.7)	637.6 (67.6)	95.6 (58.8, 132.4)
Control	741.6 (93.6)	683.4 (72.9)	58.2 (25.7, 90.7)	32.9 (−14.4, 80.1)	731.7 (96.4)	695.3 (73.6)	36.4 (−0.8, 73.7)	59.2 (8.6, 109.7)*
MVPA (minutes/day)
Intervention	5.1 (5.4)	23.7 (24.9)	18.6 (6.9, 30.3)		4.6 (5.5)	19.4 (14.4)	14.7 (6.7, 22.8)
Control	2.1 (2.2)	6.4 (9.4)	4.3 (−0.4, 8.9)	14.3 (2.2, 26.5)*	2.1 (2.2)	8.2 (10.4)	6.1 (1.0, 11.3)	8.6 (−0.1, 17.3)

¹Change was calculated by subtracting values at DC and 12M time-points from IE

²Between group changes were tested using independent t-tests, (mean (95% CI)).

^*Denotes statistical significance (p <0.05)

^**Intervention n=12, Control n=15

In the long-term, (IE to 12M), SED decreased while time in MVPA increased for the intervention and control groups (Table 1). Participants in the intervention group demonstrated greater reductions in SED (mean difference 59.2 [8.6, 109.7] minutes/day), and greater improvements in MVPA (mean difference 8.6 [−0.1, 17.3] minutes/day) compared to the control group (Table 1).

Discussion:

These preliminary results indicate that participants who received the intervention reduced SED more than the control group in the short- and long-term. The difference in ∆SED between the two groups was above and beyond the difference in ∆MVPA in the short- and long-term.

The difference between the two groups was not significant in the short-term, which may be due to lack of power to observe difference in ∆SED. Yet, these results of the exploratory analysis suggest that a PT-led PA intervention during rehabilitation after TKR may also improve sedentary behavior, in addition to MVPA. These findings highlight the potential to further improve SED if strategies to specifically target a reduction in SED are included during PT.

The difference in ∆SED between the two groups was greater (i.e. 60 minutes/day) in the long-term, which can be partially attributed to participants in the control group spending more time in SED at the 12M timepoint. Additionally, we believe that greater reduction in SED in the intervention group is not fully reflected by an increase in MVPA, suggesting that instead there may have been an increase in light physical activity (LPA), (such as walking slowly or gardening,) which may have reduced SED. Future full-scale clinical trials are needed to fully explore the effects of the intervention on the interplay between different intensities of PA and sedentary behavior from a compositional perspective.

Study Limitations:

A strength of our analysis was that SED was measured objectively using a research-grade wearable accelerometer that did not provide biofeedback regarding SED to participants. Our study had several limitations. We had a small sample size with 25% loss to follow-up, or drop out, between DC and 12M. Further, missing data was not accounted for by multiple imputation methodology given the intent of this study was to observe the SED changes in short- and long-term descriptively among the participants who completed the study. Therefore, our study findings should be viewed conservatively. Lastly, this was a pilot study; these findings should not be considered conclusive, but instead, viewed as a first step for developing a full-scale randomized control trial to determine the effects of an intervention that combines increasing MVPA and decreasing SED to improve health outcomes following TKR.

Conclusion:

Based on this pilot study, participants who receive a PA intervention may also reduce SED compared to those in control group. However, future trials are needed to examine the short- and long-term effectiveness of a PA intervention that not only increases MVPA, but also reduces SED following TKR.