Joint association of moderate-to-vigorous intensity physical activity and sedentary behavior with incident functional limitation: data from the Osteoarthritis Initiative (OAI)

Abstract

Objective:

To examine joint association of moderate-to-vigorous intensity physical activity (MVPA) and sedentary behavior with the risk of developing functional limitation 4 years later in adults with knee OA.

Methods:

Using 48-month (baseline) accelerometry data from Osteoarthritis Initiative,, we classified participants as Active-Low Sedentary (≥1 10-minute bout/week of MVPA, lowest tertile for standardized sedentary time), Active-High Sedentary (≥1 10-minute bout/week of MVPA, top two tertiles for standardized sedentary time), Inactive-Low Sedentary (0 10-minute bout/week of MVPA, lowest tertile for standardized sedentary time) and Inactive-High Sedentary (0 10-minute bout/week of MVPA, top two tertiles for standardized sedentary time) groups. Functional limitation was defined as >12sec for five repetition sit-to-stand test (5XSST) and <1.22m/sec gait speed during 20-meter walk test. To investigate the association of exposure groups with risk of developing functional limitation 4 years later, we calculated risk ratios(aRR) adjusted for potential confounders.

Results:

Of 1,091 and 1,133 participants without baseline functional limitation, based on 5XSST and 20-meter walk test, respectively, 15% and 21% developed functional limitation 4 years later. Active-High Sedentary group didn’t have, while Inactive-Low Sedentary, and Inactive-High Sedentary groups had increased risk of developing functional limitations compared to Active-Low Sedentary group. Inactive-Low Sedentary group had 72% (aRR[95% confidence interval]1.72[1.00, 2.94]) and 52% (1.52[1.03, 2.25]) more risk of developing functional limitation based on 5XSST and 20-meter walk test, respectively, compared to Active-Low Sedentary group.

Conclusion:

Regardless of sedentary category, being inactive (0 10-minute bouts/week in MVPA) may increase risk of developing functional limitation in adults with knee OA.

Introduction:

Over 14 million Americans have symptomatic knee osteoarthritis (OA)(1), which is a leading cause of functional limitation, such as difficulty getting up from a chair or slow walking.(2-6) Regular participation in physical activity is recommended as a first-line treatment for adults with knee OA.(7-9) The intensity of physical activity is particularly relevant for knee OA.(10-12) Different intensities of physical activity include moderate-to-vigorous intensity physical activity (MVPA), defined as any activity with ≥ 3 metabolic equivalents like brisk walking, and sedentary behavior, defined as any activity with ≤ 1.5 metabolic equivalents like sitting on a couch.(13) Both intensities have been linked to health outcomes. Replacing time in sedentary behavior with time in MVPA is important to reduce the risk of developing future functional limitation(11) and pain that interferes with work.(12) At the same time, prolonged time in sedentary behavior increases the risk of disability(14), the decline in physical function(15), and lower quality of life(16) and physical functioning(17) in adults with knee OA.

At present, the joint association between MVPA and sedentary behavior with functional limitation is unclear, though important since engaging in MVPA and prolonged time in sedentary behavior coexist throughout the day. For instance, adults may walk briskly for 10 mins for exercise in the morning, and then spend the rest of the day sitting for work. We are particularly interested in functional limitation, i.e., the restriction in the performance of physical function, or difficulty in the execution of a task or action.(18) We focus on established categories that represent clinically meaningful restrictions in physical functioning or functional limitation in adults with knee OA.(19-21)

The purpose of this study was to investigate the joint association of MVPA and sedentary behavior with incident functional limitation over four years in adults with or at risk of knee OA. Additionally, we investigated the stability of this association among adults with radiographic and symptomatic knee OA given the presence of knee OA and/or symptoms may affect daily walking(22) and physical function.(23) In this study, we used a similar approach that was employed by prior studies to examine the joint association of MVPA and sedentary behavior with health markers in population-based samples from the United States and England.(24, 25) Specifically, in this study, a 4-level mutually exclusive exposure variable was created based on thresholds for time spent in sustained (bouted) MVPA to classify being inactive vs. active in adults with arthritis(26-28) and distribution-based approach for time spent in sedentary behavior to classify being more vs. less sedentary. Based on the findings of prior studies by Loprinzi et al,(24) and Bakrania et al,(25), we hypothesize that regardless of time spent in sedentary behavior, adults with knee OA who are classified as being inactive will have higher risk of functional limitation compared to those who are active. This investigation is important for assessing to what extent the different combinations of activity, i.e., MVPA and sedentary behavior are associated with functional limitation in adults with knee OA.

Methods:

Study participants

We used publicly available de-identified data from the Osteoarthritis Initiative (OAI), a large prospective observational cohort study of 4,796 adults with or at risk of knee OA. Participants with risk of knee OA were determined based on body mass index (BMI) being at least overweight, knee pain, and prior knee injury or knee surgery. Detailed description of the study protocol can be found the NIH website (https://nda.nih.gov/oai/). Participants were recruited from four clinical sites (Maryland, Pennsylvania, Rhode Island and Ohio). Institutional review board approval was obtained from all OAI sites. Adults with rheumatoid or inflammatory arthritis, bilateral end-stage disease defined as severe joint space narrowing or total knee replacements in both knees, and those who used ambulatory aids other than a cane at baseline were excluded from the OAI. For this study, we used data from 48-month and 96-month clinic visits. We considered the 48-month visit as our study baseline because it was the first visit at which device-assessed physical activity data was collected. We considered the 96-month visit as the 4-year follow-up. We excluded participants who did not have valid device-assessed physical activity data, and who had functional limitation at baseline. Supplemental figures 1 and 2 provide a summary of the number of participants at the baseline and reasons for exclusion from our analytical sample.

Study Outcome

Incident functional limitation.

Our primary outcome was incident functional limitation at the 4-year follow-up. We defined incident functional limitation using two performance-based clinical tests as separate study outcomes since they represent difficulty in different activities, i.e. sit-to-stand vs. walking over a short distance.

Five repetition sit to stand test (5XSST).

First definition of functional limitation was taking ≥12 sec to complete the 5XSST, as this threshold is associated with increased risk for falls in older adults (29) and the inability to be walk at least 6000 steps per day in adults with knee OA.(20) Each participant was instructed to fold their arms across their chest, stand from a chair and return to sitting five times as quickly as possible. Digital stopwatch was used to measure total time in seconds (sec) needed to complete one trial of 5XSST. Time was recorded as the average of two trials. The 5XSST has high test-retest reliability in older adults with symptomatic hip or knee OA.(30, 31) Participants who were classified as having functional limitation at baseline were removed from the analytic sample. We defined participants as having the outcome if they developed functional limitation at the 4-year follow-up.

Gait speed.

Second definition of functional limitation was gait speed <1.22 meters/second (m/sec), as this threshold is the minimum gait speed needed to safely cross timed crosswalks in the community(32) and is associated with lower odds of walking at least 6000 steps per day (20) and increased mortality risk (21) in adults with knee OA. Each participant was instructed to walk at their usual speed over a marked, 20-m course in an unobstructed and dedicated corridor. Gait speed was calculated by dividing the total distance, i.e., 20 m, by total time (sec) needed to complete the test. Measuring gait speed using the 20-m walk test has high test-retest reliability in adults with knee OA.(33) Participants who were classified as having functional limitation at baseline were removed from the analytic sample. We defined participants as having the outcome if they developed functional limitation at the 4-year follow-up.

Study Exposures

Time spent in MVPA and sedentary behavior were collected at baseline using a uniaxial accelerometer (Actigraph GT1M). The Actigraph GT1M is a valid device for measuring physical activity in free-living conditions.(34, 35) Though uniaxial accelerometers cannot classify body positions, they provide valid time spent in overall sedentary behavior that is shown to be associated with health outcomes in population-based studies.(36, 37) Subjects were fitted with the accelerometer above the right hip and were instructed to wear during waking hours for seven consecutive days. Previously published methods were used to determine valid physical activity data.(38) Briefly, we defined a valid wear day as days with ≥10 hours of wear time and included participants with ≥4 valid wear days, as this is the minimum time for a reliable estimate of physical activity behavior.(26) We identified and eliminated non-wear time, i.e., time when the Actigraph was not worn, if the accelerometer had registered >90 minutes of consecutive activity of <100 counts/minute.(39) This non-wear criterion was applied in accordance with previously published study validating such criteria for the Actigraph.(39)

Based on threshold recommended by National Cancer Institute, MVPA was defined as intensity ≥2020 activity counts/minute.(38) This threshold is intended to capture ambulatory activities that reach >3 metabolic equivalents, including brisk walking. Based on national survey data from the United States, adults with arthritis who did not engage in at least 10 continuous minutes of MVPA over a week were considered as being inactive.(28) Therefore, we classified adults who did not accumulate any 10-minute bout/week in MVPA as being inactive and those with at least 1 bouts/week in MVPA as being active. The bouts allowed for an interruption of 1 or 2 minutes. Accumulating 0 10-minute bout/week in MVPA was previously used to define adults with knee OA who were inactive.(26)

Sedentary behavior was defined as <100 activity counts/minute.(36) We standardized the time in sedentary behavior to a 16-hour wear day since minimal wear time for a valid day was 10 hours, yet average wear time ranged from 10 hours to 19 hours. This standardization approach has been used by prior studies to account for typical waking period during the day.(16, 40, 41) At present, there is no known threshold of time spent in sedentary behavior that is predictive of poor health outcomes. Therefore, a conservative distribution based approach was used to classify adults who were more versus less sedentary. We classified adults as more sedentary if they resided in the top two tertiles of the standardized time spent in sedentary behavior. Adults were classified as less sedentary if they resided in the lowest tertile of the standardized time spent in sedentary behavior.

We formulated categories of activity by combining the MVPA and sedentary behavior classifications to create a 4-level mutually exclusive exposure variable. Active-Low Sedentary were defined as those who were active and less sedentary (≥1 10-minute bout/week of MVPA, lowest tertile for standardized sedentary time). Active-High Sedentary were defined as those who were active and more sedentary (≥1 10-minute bout/week of MVPA, top two tertiles for standardized sedentary time). This group had more time in sedentary behavior, but the similar time in bouted MVPA over a week as Active-Low Sedentary group. Inactive-Low Sedentary were defined as those who were inactive and less sedentary (0 10-minute bout/week of MVPA, lowest tertile for standardized sedentary time). This group had a similar time in sedentary behavior, but spent no time in bouted MVPA over a week as Active-Low Sedentary group. Inactive-High Sedentary were defined as those who were inactive and more sedentary (0 10-minute bout/week of MVPA, top two tertiles for standardized sedentary time). This group had more time in sedentary behavior and spent no time in bouted MVPA over a week compared with Active-Low Sedentary group.

Potential Confounders

We considered the following factors as potential confounders based on their known association with physical activity and physical function(23, 42): Age, sex (female versus male), race (white versus non-white), education (less than college graduate versus at least college graduate), BMI (kg/m²) computed from weight and height assessment, comorbidity using the modified Charlson comorbidity index(43) (>1 comorbidity versus no comorbidity), presence of knee pain, aching or stiffness on most days in past month during the previous year in either right or left knee (Yes versus No), and presence of radiographic knee OA defined as Kellgren–Lawrence (KL) grade ≥ 2 on x-ray in one or both knees (Yes versus No). All potential confounders were measured at the study enrollment or baseline, i.e., a 48-month visit by interview via questionnaire and/or direct measurement, as appropriate.

Statistical Analysis

We described the study sample using means and standard deviations (sd) for continuous variables and percentages for categorical variables. To examine the joint association of MVPA and sedentary behavior with the risk of incident function limitation at 4-year follow-up, we calculated risk ratios and 95% confidence intervals [RR (95%CI)] using regression models with a log-link function and robust standard errors adjusted for potential confounders.(44) We repeated the analyses restricting our sample to adults with radiographic knee OA only and symptomatic knee OA only. Symptomatic knee OA was defined as the presence of radiographic knee OA and the presence of pain, aching, or stiffness on most days of a month during the previous year in either knee. The intent of these subgroup analyses weres to investigate the generalizability of the study findings across adults with varying degrees of disease severity. All analyses were performed using SAS software, version 9.4 (Cary, NC).

Results:

Of 1,927 participants with valid accelerometer data, 1,091 and 1,133 were without baseline functional limitation and completed the 4-year follow-up, defined by 5XSST and gait speed, respectively (Supplemental Figures 1 and 2).

Incident functional limitation as defined by 5XSST.

Of the 1,091 participants free of functional limitation based on 5XSST at baseline, 15% of the sample developed functional limitation at the 4-year follow-up (Table 1). The sample was (mean±sd) 63.1±8.4 years old with BMI 28.0±4.6 kg/m², and over half were women (51%). The majority were white (88%) and graduated from college (71%).

Table 1:

Characteristics of study participants for the 5 repetition sit to stand test (5XSST) analytic sample at baseline

	All	1Active-Low Sedentary	2Active-High Sedentary	3Inactive-Low Sedentary	4Inactive-High Sedentary
Total sample	1091	293	412	106	280
Demographics
Age, years [Mean±sd (n)]	63.1±8.4 (1091)	61.4±7.2 (293)	62.3±8.5 (412)	62.1±8.3 (106)	66.4±8.7 (280)
Women [% (n)]	50.6 (552)	50.9 (149)	41.3 (170)	72.6 (77)	55.7 (156)
Race, white [% (n)]	87.9 (947)	85.8 (248)	93.4 (383)	83.0 (88)	83.8 (228)
Education, at least college graduate [% (n)]	71.1 (774)	74.0 (216)	74.0 (216)	79.8 (327)	53.8 (57)
BMI, kg/m2 [Mean±sd (n)]	28.0±4.6 (1091)	27.0±4.2 (293)	27.9±4.3 (412)	28.6±5.4 (106)	29.0±4.8 (280)
Presence of knee pain, aching or stiffness: more than half the days of a month, past 12 months (Y/N) (n)
Right [% (n)]	24.0 (261)	24.7 (72)	24.0 (99)	22.6 (24)	23.9 (66)
Left [% (n)]	23.6 (257)	26.3 (77)	21.8 (90)	19.8 (21)	24.8 (69)
Comorbidity [% (n)]	23.9 (259)	21.9 (64)	21.0 (86)	22.6 (24)	30.6 (85)
Time in bouted MVPA/day
minutes [Mean±sd (n)]	10.8±15.5 (1091)	19.2±18.9 (293)	14.8±14.3 (412)	0±0 (106)	0±0 (280)
Time in SED 5 /day
minutes [Mean±sd (n)]	627.6±79.4 (1091)	543.8±49.8 (293)	667.8±41.2 (412)	545.3±58.3 (106)	687.1±48.3 (280)
Baseline performance on 5XSST
sec [Mean±sd (n)]	9.1±1.7 (1091)	9.0±1.7 (293)	9.0±1.7 (412)	9.3±1.6 (106)	9.4±1.7 (280)
4-year incident function limitation (outcome)
5XSST>12 sec [%(n)]	14.9 (162)	10.6 (31)	11.9 (49)	18.9 (20)	22.1 (62)

¹Active-Low Sedentary were people who at least had 1 bout of moderate to vigorous intensity physical activity (MVPA) during the week and did not spend the majority of time in sedentary behavior

²Active-High Sedentary were people who were at least had 1 bout of MVPA during the week but spent the majority of time in sedentary behavior

³Inactive-Low Sedentary were people who did not have any bouts of MVPA during the week but did not spend the majority of time in sedentary behavior

⁴Inactive-High Sedentary were people who did not have any bouts of MVPA during the week and spent the majority of time in sedentary behavior

⁵Time in sedentary behavior was standardized to average wear time over 16 hours

Of participants classified as Active-Low Sedentary, 10.6% developed function limitation at the 4-year follow-up. Of participants classified as Inactive-Low Sedentary, 18.9% developed functional limitation at 4-year follow-up and had 72% greater risk (adjusted RR [95%CI]; 1.72 [1.00, 2.94]) of developing the outcome compared to the Active-Low Sedentary group. Of participants classified as Inactive-High Sedentary, 22.1% developed functional limitation at 4-year follow-up and had a 50% greater risk (1.50 [0.98, 2.31]) of incident functional limitation by the 4-year follow-up compared to Active-Low Sedentary group. Active-High Sedentary group had a similar risk of incident functional limitation as Active-Low Sedentary group (Table 3). We found similar trends in the incidence of the outcome when we restricted the sample to adults with radiographic knee OA only, as well as with symptomatic knee OA (Supplementary Table 1). However, the effect estimates did not reach statistical significance.

Table 3:

Joint association of MVPA and sedentary behavior with incident functional limitation at the 4-year follow-up as measured by the 5 repetition sit to stand test (5XSST) and gait speed test among all the study participants.

	Baseline Mean±SD	Incident Function Limitation/total	%	Unadjusted RR (95% CI)	*Adjusted RR (95% CI)
5XSST (>12 sec) (N = 1091)
1Active-Low Sedentary	9.0±1.7	31/293	10.6	1.00 [Reference]	1.00 [Reference]
2Active-High Sedentary	9.0±1.7	49/412	11.9	1.13 [0.74, 1.72]	1.08 [0.70, 1.67]
3Inactive-Low Sedentary	9.3±1.6	20/106	18.9	1.79 [1.07, 2.99]	1.72 [1.00, 2.94]
4Inactive-High Sedentary	9.4±1.7	62/280	22.1	2.10 [1.41, 3.12]	1.50 [0.98, 2.31]
Gait speed (<1.22 m/sec) (N = 1133)
1Active-Low Sedentary	1.46±0.15	45/311	14.5	1.00 [Reference]	1.00 [Reference]
2Active-High Sedentary	1.46±0.16	75/434	17.3	1.20 [0.86, 1.68]	1.25 [0.88, 1.77]
3Inactive-Low Sedentary	1.41±0.14	31/120	25.8	1.79 [1.19, 2.68]	1.52 [1.03, 2.25]
4Inactive-High Sedentary	1.40±0.14	85/268	31.7	2.20 [1.59, 3.03]	1.52 [1.09, 2.14]

^*Adjusted for baseline age, education, race, sex, BMI, the presence of > 1 comorbidities on Charlson Comorbidity Index, radiographic knee OA, and pain, aching or stiffness on most days in a past month during the previous year in either right or left knee

¹Active-Low Sedentary were people who at least had 1 bout of moderate to vigorous intensity physical activity (MVPA) during the week and did not spend the majority of time in sedentary behavior

²Active-High Sedentary were people who were at least had 1 bout of MVPA during the week but spent the majority of time in sedentary behavior

³Inactive-Low Sedentary were people who did not have any bouts of MVPA during the week but did not spend the majority of time in sedentary behavior

⁴Inactive-High Sedentary were people who did not have any bouts of MVPA during the week and spent the majority of time in sedentary behavior

Incident functional limitation as defined by gait speed.

Of the 1,133 participants free of functional limitation based on gait speed at baseline, 21% of the sample developed the outcome at the 4-year follow-up (Table 2). The sample was (mean±sd) 62.9±8.4 years old, with a BMI of 28.0±4.5 kg/m², and over half were women (51%). The majority were white (89%) and had graduated from college (73%).

Table 2:

Characteristics of study participants for the gait speed analytic sample at baseline

	All	1Active-Low Sedentary	2Active-High Sedentary	3Inactive-Low Sedentary	4Inactive-High Sedentary
Total sample	1133	311	434	120	268
Demographics
Age, years [Mean±sd (n)]	62.9±8.4 (1133)	61.5±7.2 (311)	62.5±8.6 (434)	61.8±8.4 (120)	65.9±8.8 (268)
Women [% (n)]	50.9 (577)	53.1 (165)	40.6 (176)	73.3 (88)	55.2 (148)
Race, white [% (n)]	88.8 (995)	87.9 (269)	92.8 (401)	81.7 (98)	86.6 (227)
Education, at least college graduate [% (n)]	72.5 (819)	74.5 (231)	80.8 (349)	52.5 (63)	65.7 (176)
BMI, kg/m2 [Mean±sd (n)]	28.0±4.5 (1133)	27.1±4.1 (311)	27.8±4.2 (434)	28.8±5.2 (120)	28.9±4.9 (268)
Presence of knee pain, aching or stiffness: more than half the days of a month, past 12 months (Y/N) (n)
Right [% (n)]	26.9 (303)	24.6 (76)	27.0 (117)	30.0 (36)	27.9 (74)
Left [% (n)]	25.4 (287)	24.8 (77)	22.9 (99)	25.8 (31)	30.0 (80)
Comorbidity [% (n)]	23.4 (263)	21.0 (65)	20.0 (86)	22.7 (27)	32.1 (85)
Time in bouted MVPA/day
minutes [Mean±sd (n)]	11.0±15.7 (1133)	18.9±18.8 (311)	15.2±14.9 (434)	0±0 (120)	0±0 (268)
Time in SED 5 /day
minutes [Mean±sd (n)]	624.5±80.2 (1133)	543.2±50.0 (311)	668.5±41.7 (434)	541.7±59.3 (120)	684.7±49.4 (268)
Baseline Gait speed
m/sec [Mean±sd (n)]	1.44±0.15 (1133)	1.46±0.15 (311)	1.46±0.16 (434)	1.41±0.14 (120)	1.40±0.14 (268)
4-year incident function limitation (outcome)
<1.22 m/ sec [%(n)]	20.8 (236)	14.5 (45)	17.3 (75)	25.8 (31)	31.7 (85)

¹Active-Low Sedentary were people who at least had 1 bout of moderate to vigorous intensity physical activity (MVPA) during the week and did not spend the majority of time in sedentary behavior

²Active-High Sedentary were people who were at least had 1 bout of MVPA during the week but spent the majority of time in sedentary behavior

³Inactive-Low Sedentary were people who did not have any bouts of MVPA during the week but did not spend the majority of time in sedentary behavior

⁴Inactive-High Sedentary were people who did not have any bouts of MVPA during the week and spent the majority of time in sedentary behavior

⁵Time in sedentary behavior was standardized to average wear time over 16 hours

Of participants classified as Active-Low Sedentary, 14.5% developed function limitation at 4-year follow-up. Of participants classified as Inactive-Low Sedentary, 25.8% developed function limitation at 4-year follow-up and had 52% greater risk (1.52 [1.03, 2.25]) of incident functional limitation at 4-year follow-up compared to those who were classified as Active-Low Sedentary s (Table 3). Of participants classified as Inactive-High Sedentary, 31.7% developed function limitation at 4-year follow-up and had 52% greater risk (1.52 [1.09, 2.14]) of incident functional limitation by the 4-year follow-up compared to those were defined as Active-Low Sedentary (Table 3). Active-High Sedentary group had a similar risk of incident functional limitation as Active-Low Sedentary group (Table 3). We found similar trends in the incidence of the outcome when we restricted the sample to adults with radiographic knee OA only, as well as with symptomatic knee OA (Supplementary Table 1). However, the effect estimates did not reach statistical significance.

Discussion:

We observed that adults with or at high risk of knee OA who were inactive had a greater risk of developing functional limitation compared to those who were minimally active, irrespective of their sedentary category. These findings were consistent for both definitions of incident functional limitation. We found similar effects after restricting the sample to adults with radiographic knee OA and symptomatic knee OA. Our findings suggest that healthcare professionals should encourage adults with knee OA to avoid inactivity in order to minimize the risk of incident functional limitation.

Our findings highlighting the importance of MVPA for reducing risk of functional limitation, in terms of limited sit to stand ability or slower gait speed, are consistent with previous literature.(10, 11) Studies in the general population have shown that engaging in MVPA results in favorable health measures, such as lower BMI and waist circumference even in those adults who spend prolonged time in sedentary behavior.(24, 25) Engaging in MVPA preserves or increases lower extremity muscle strength(45), and reduces pain interference with work(12). These benefits likely contribute to reducing the risk of functional limitation and disability.(46-48)

Based on this study, engaging in at least 10 minutes of sustained MVPA/exercise once a week may be a reasonable starting point for adults with knee OA who may spend the majority of their time in sedentary behavior. This finding aligns with guideline recommendations that “some activity is better than none” and studies which found that adverse effects of sedentary behavior can be attenuated by engaging in MVPA.(49) In this study, we used device-assessed MVPA to accurately define inactivity(26-28), which is a separate construct compared to time spent in sedentary behavior. Further, the goal of this study was not to investigate an optimal threshold or to investigate a dose-response relationship between activity and functional limitation. Rather, we were interested in investigating the effects of being inactive among participants with knee OA who were less sedentary as well as the effects of being active among those who were more sedentary.

We did not find our definition of low sedentary behavior to be related to functional limitation in this analysis. Previous studies have found a modest association of sedentary behavior with physical function. For example, more sedentary time was found to be associated with a decline in gait speed and longer time to complete 5 repeated chair stands, adjusting for MVPA using data from the OAI.(15) As well, spending more time in sedentary behavior was associated with lower quality adjusted life years in adults from the OAI who spent >11.6 hours/day being sedentary, though this association was not present in those who spent 10.7 to 11.6 hours/day being sedentary.(16) We find it important to note that these previous studies analytically employed a continuous definition of their study outcomes. This is in contrast to our study, where we were interested in categories that represent clinically meaningful restrictions, i.e., functional limitation. Thus, it is possible that sedentary time may have a relationship with physical function, albeit not as strong as than MVPA.

Study limitations.

First, there is a potential for reverse causation between our exposure (MVPA and sedentary behavior) and outcome (functional limitation) given the observational nature of the study. Further, functional limitation can fluctuate over time. Even though we have removed participants with functional limitation at baseline, we cannot ensure that our study outcome, incident functional limitation, was truly the first instance at 4 year follow-up. Regardless, we have taken advantage of standardized follow-up timepoints from the OAI that allow for the systematic investigation of the relationship between MVPA and sedentary behavior with functional limitation. Second, the majority of our analytical sample were white and highly educated. Therefore, we caution generalizing the results of this study to more diverse knee OA population, in terms of race and education. Third, we used a single cut-point to classify whether participants were inactive vs. active and more vs. less sedentary. Categorizing continuous variables may lead to loss of power and bias. However, we used this approach so we could create four meaningful and mutually exclusive groups to enhance clinical interpretation. Lastly, there is a potential for unmeasured confounding in our analyses because we were not able to account for an exhaustive list of potential confounders. For instance, there is a possibility that participation in MVPA and sedentary behavior may be influenced by psychosocial factors, such as fear of movement and catastrophizing. Future studies are needed to explore the role of psychosocial factors on the association between different activities and functional limitation.

Conclusion:

We found that adults with or at risk of knee OA who were inactive had a greater risk of developing functional limitation compared to those who were active, irrespective of their sedentary category. When possible, healthcare professionals should recommend adults with knee OA to avoid inactivity by encouraging them to go for a 10 minutes of continous brisk walk at least once a week. This recommendation may serve as starting point among adults with knee OA who are highly sedentary.