Can we improve cognitive function among adults with osteoarthritis by increasing moderate-to-vigorous physical activity and reducing sedentary behaviour? Secondary analysis of the MONITOR-OA study

This study was a secondary analysis of Monitor-OA, a six month randomized controlled trial (RCT) examining the efficacy of a technology-enabled counselling intervention for increasing MVPA and reducing SB in people with knee OA [36]. The study occurred between November 1st 2015 and June 1st 2017. We used a randomized delayed-control design. In this study design, randomization determined the timing of when the intervention was provided (i.e., immediately vs. a 2-month delay).

We recruited community-dwelling adults from Vancouver, British Columbia who had a physician confirmed diagnosis of knee OA, or passed two criteria for early OA: 1) aged 50+ years; and 2) experienced knee pain during the previous year lasting > 28 separate or consecutive days [37]. Participants were excluded if they 1) had been diagnosed with inflammatory arthritis, connective tissue diseases, fibromyalgia, or gout; 2) used anti-rheumatic drugs or gout medications; 3) had previously underwent knee arthroplasty, or were on the waitlist to receive total knee arthroplasty; 4) had suffered an acute knee injury in the past six months; 5) had a body mass index (BMI) of > 40 kg/m²; 6) had received a steroid injection or a hyaluronate injection in the last 6 months; 7) were using medications which impaired physical activity tolerance (e.g., beta blockers), or had an inappropriate level of risk for increasing their physical activity. Participants were also excluded if they did not have access to a computer in their home, or did not have a personal email address. Potential participants completed the Physical Activity Readiness Questionnaire (PAR-Q) [38]. If a potential risk was identified by the PAR-Q, physician confirmation was required to ensure the participant was able to be physically active without supervision of a health professional.

The CONSORT (Consolidated Standards of Reporting Trials) flowchart in Fig. 1 shows the number of participants in the treatment arms at each stage of the study [39]. The research protocol was approved by the University of British Columbia Behavioural Research Ethics Board (Application number: H14–01762), and was published in ClinicalTrials.​gov (NCT02315664).

Trained staff members administered all testing procedures. We assessed participants at baseline, 2 months, 4 months, and 6 months follow-up. In this paper we report data from baseline, 2 months, and 4 months.

After completing the baseline assessment, eligible participants were randomly assigned to the Immediate Intervention (I-INT) or the Delayed Intervention (i.e. control; D-INT) in a 1:1 allocation ratio. Randomization was performed using computer-generated random numbers in variable block sizes. The D-INT received the same intervention as the I-INT after a 2-month wait.

Details of the intervention have been described previously [36]. Briefly, the intervention consisted of participants attending a 1.5-h session, where they received: 1) a standardized group education session about the benefits of increasing MVPA and reducing SB; 2) a Fitbit® Flex™; and 3) individual activity counselling with a physiotherapist. The education session was delivered in groups of 2–3 participants. The individual activity counselling session followed the Brief Action Planning approach [49]. The physiotherapist guided participants to identify their MVPA goals (e.g., begin resistance training, start cycling, join a walking group, etc.), develop an action plan (i.e., where they plan to perform their MVPA goal, how often, and for how long), identify barriers and solutions, and then rate their confidence in executing the plan. For SB, the physiotherapists began by asking participants to estimate their sitting time in a normal day. Participants were then asked to identify ways to break up their sitting time into shorter bouts.

Following the education session, participants were provided with a Fitbit Flex™ which they were instructed to wear 24 h/day except during water-based activity (i.e., swimming or bathing) or when charging the device. The Fitbit data were wirelessly synchronized with Fitbit’s online Dashboard which could be viewed only by the participants and the study physiotherapist. During the intervention period, the physiotherapist reviewed the individual’s MVPA on the Dashboard and progressively modified the activity goals during 4 biweekly phone calls. During these phone calls, we also monitored participant adherence to SB goals using self-report. Specifically, participants were asked at each biweekly phone call whether they fully met, partially met, or did not meet their SB goal. These goals were then modified as needed.

We conducted all statistical analyses using R version 3.3.1 in the lsmeans 2.26–3, lmerTest 2.0–33, and mice 2.46.0 packages. Descriptive statistics were used to summarize participant demographics. In order to account for missing data at each follow-up time point, we performed multiple imputation in the mice 2.46.0 package using predicted mean matching (5 imputations; 20 iterations each), and visually checked for convergence. All statistical models used pooled estimates from all 5 imputed data sets. Plots and graphs were created using ggplot2 2.2.1. Our statistical code can be found in Additional file 1.

From 2015 to 2016, 278 people indicated an interest to participate, and 64 met the eligibility criteria (Fig. 1). Of those, we recruited 61 participants (I-INT: n = 30; D-INT: n = 31). As described in Table 1, there were no group differences in age (I-INT: 61.73 [SD 9.40] years; D-INT: 62.61 [8.54]), sex (I-INT: 73.33% female; D-INT: 90.32%), BMI (I-INT: 29.16 [5.46] kg/m²; D-INT: 29.24 [4.82), or time spent in MVPA (I-INT: 83.44 [60.80] minutes/day; D-INT: 86.19 [86.19] minutes/day) and SB (I-INT: 681.96 [111.51] minutes/day; D-INT: 703.05 [161.17] minutes/day) at baseline. Participants in I-INT had a lower picture sequence memory score (I-INT: 102.04 [17.22]; D-INT: 112.53 [14.67]; p = 0.02), but there were no differences between groups for list-sorting score (I-INT: 102.05 [17.22]; D-INT: 102.42 [14.64]).

Group differences in cognitive performance—accounting for baseline cognitive performance—are illustrated in Fig. 2. Briefly, there were no statistically significant differences between groups following the intervention. As described in Table 2, we calculated a small improvement of the I-INT compared to D-INT for picture sequence memory (estimated mean difference: 1.27; 95% CI [− 9.27, 11.81]; d = 0.10), and a small improvement of the D-INT compared to the I-INT for list-sorting memory (estimated mean difference: -1.64; 95% CI [− 8.72, 5.44]; d = − 0.19).

The relationship between changes in MVPA and changes in cognitive function are illustrated in Fig. 3. There were no statistically significant relationships between changes in MVPA and changes in cognitive function. Increases in MVPA were correlated with changes in list-sorting memory in the expected direction (B = 0.04; 95% CI [− 0.07, 0.14]), however changes in picture sequence memory appeared to be negatively correlated with increases in MVPA (B = − 0.02; 95% CI [− 0.15, 0.12]).

The relationship between changes in SB and changes in cognitive function are illustrated in Fig. 4. There were no statistically significant relationships between changes in SB and changes in either picture sequence memory (B = − 0.01; 95% CI [− 0.09, 0.07]) or list-sorting memory (B = 0.00; 95% CI [− 0.09, 0.10]).

Although we did not find that our intervention significantly improved cognitive function, there are several potential clinical applications to our study. First, we previously demonstrated that clinicians (i.e., physical therapists) can use consumer-available wearable activity-monitors such as a Fitbit to promote MVPA and reduce SB among their patients with OA [36]. Given the health care system has an untapped capacity for promoting changes in MVPA and SB, which to date has not been fully developed [58], a first step towards harnessing this potential to promote behaviour change is for clinicians to track their patients’ MVPA and SB using wearable activity-monitors. At minimum, clinicians should query about activity during their consultations with OA patients [59].

Secondly, both OA and SB may increase the risk of cognitive impairment and dementia [7‐9, 26, 29, 30]. In contrast, engagement in MVPA reduces dementia risk and promotes overall cognitive and physical health [14‐16, 27, 28]. Adults with OA should therefore be encouraged to adhere to the current public health recommendations of engaging in ≥150 min/week of MVPA and limiting their SB as much as possible [26]. Importantly, 87% of adults with OA are underactive [31] and adults with OA spend an average of 61% of the day in SB [32]. MVPA promotion and SB reduction may thus be particularly important for physical and cognitive health in adults with OA.

This was a secondary outcomes analysis, and thus our findings are a preliminary investigation of whether increasing MVPA and reducing SB can improve cognitive function in adults with OA. While the SenseWear Mini provides valid and reliable estimates of energy expenditure for both younger and older adults [42, 60], which can be used to derive time spent in MVPA and time spent in SB, we cannot determine whether time estimated as SB was actually spent sitting or lying down. Hence, future studies to examine changes in SB should use measures of body posture such as the activPAL [61], which can accurately determine whether a person is sitting, standing, or walking.

We did not collect information on medication use, however our participants were community-dwelling adults who were healthy enough to start a physical activity program at study entry. We also did not exclude participants based on current activity level and hence the results may not be generalizable to most people with OA—who are often sedentary. Given the high activity level of our participants, the effects of increased MVPA and reduced SB on cognitive function may have been attenuated. We therefore suggest future studies should recruit underactive adults with OA, since these individuals are likely to benefit from increased MVPA and reduced SB.

There is growing evidence that the effects of MVPA (and potentially SB) are moderated by age and sex [62, 63], however due to our small sample size, it was not feasible for us to control for age and sex within our analyses. As detailed previously [36], our intervention did not reduce SB which is perhaps due to several shortcomings of the counselling program, which we have rectified. Specifically, the intervention now includes a new SB counselling strategy, and a Fitbit-compatible web app with enhances functionality for setting goals and rewarding behaviours that break up prolonged sitting [64]. This paradigm is currently being tested in a RCT (ClinicalTrial.​gov identifier: NCT02554474) involving people with rheumatoid arthritis and systematic lupus erythematosus [65].