Is high-intensity interval cycling feasible and more beneficial than continuous cycling for knee osteoarthritic patients? Results of a randomised control feasibility trial

Introduction

Osteoarthritis (OA) is a highly prevalent degenerative joint disease affecting many middle-aged and older adults, with recent global data indicating OA of the hip and knee was ranked as the 11th highest contributor to global disability and 38th highest in disability adjusted life years of the 291 health conditions analysed (Cross et al., 2014). The knee is one of the most affected osteoarthritic joints, resulting in a range of symptoms including pain and tenderness that typically limit the individuals’ physical function and mobility (Flores & Hochberg, 1988; Hootman et al., 2003). Such changes typically result in considerable physical inactivity, with a recent accelerometer cross-sectional study involving 1,111 participants reporting that only 13% of men and 8% of women with OA were meeting the recommended levels of aerobic physical activity and that an additional 40% of men and 57% of women could be classified as inactive, participating in no bouts of moderate to vigorous physical activity for more than 10 min at a time (Dunlop et al., 2011). These knee OA-related reductions in physical function and physical activity may accelerate the age-related loss of muscle mass, muscle strength and function (especially around the knee joint), referred to as sarcopenia (Cruz-Jentoft et al., 2010; Senior et al., 2015); and contribute to an increased risk of obesity and poor cardiovascular health (Roubenoff, 2000). Collectively, the symptoms and adverse events associated with the development of knee OA may further compromise health and well-being and contribute to a downward spiral into greater disability, poor health-related quality of life and further health complications (Yoshimura et al., 2012).

Several recent meta-analyses indicate that a variety of forms of exercise significantly improve pain, function and health-related quality of life in individuals with knee OA (Fransen et al., 2015; Tanaka et al., 2015; Uthman et al., 2013). While a number of different modes of exercise have been examined in the studies included in these meta- analyses, cycling and aquatic exercise are often the most commonly recommended (Arthritis Foundation) and performed (Hootman et al., 2003) exercise options for individuals with knee OA in the community. Such popularity may reflect the benefits and tolerability of these forms of exercise, with aquatic exercise and cycling imposing lower knee joint loads compared to walking, stair climbing and common resistance training exercises such as the leg extension and squat (Hall et al., 2013; Heino Brechter & Powers, 2002; Kutzner et al., 2010, 2012; Powers et al., 2014).

The benefits of cycling for middle-aged and older adults with knee OA are also well documented. These include significant improvements in health-related quality of life, as assessed by the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire (Alkatan et al., 2016; Salacinski et al., 2012); physical function, as assessed by gait speed (Alkatan et al., 2016; Salacinski et al., 2012) and Sit to Stand (STS) performance (Mangione et al., 1999) and body composition, as assessed by body mass, waist and hip circumference and visceral adipose tissue (Alkatan et al., 2016). As the majority of these cycling studies utilised moderate-intensity continuous training (MICT) protocols, exercise prescriptions involving higher intensity muscle contractions may better improve sarcopenic outcomes including muscle mass, strength and function (Landi et al., 2014).

We propose that the utilisation of high-intensity interval training (HIIT) may be an exercise approach that simultaneously improves sarcopenic and cardiovascular outcomes in the OA patients, while at the same time having a similar degree of feasibility and tolerability as MICT. HIIT typically requires participants to alternate short periods (∼8–60 s) of high-intensity activity with longer (∼20–90 s) recovery periods of lower intensity activity (Boutcher et al., 2013; MacInnis & Gibala, 2017; Shiraev & Barclay, 2012). Based on data obtained from other clinical populations, HIIT cycling may provide a better stimulus for overcoming the knee OA related sarcopenic outcomes (i.e. loss in leg muscle mass, strength and function (Alvarez et al., 2017; Jia et al., 2018), while still providing a range of cardiometabolic benefits than is currently achieved by MICT exercise mode such as cycling or aquatic exercise (Francois & Little, 2015; Liou et al., 2016; Ramos et al., 2015; Shiraev & Barclay, 2012).

Unfortunately, almost no research has assessed the feasibility and potential benefits of HIIT in musculoskeletal conditions like knee OA. This may reflect the concern that some health professionals may have regarding the potential musculoskeletal and cardiovascular health risks associated with HIIT. While HIIT appears well tolerated by older and middle-aged individuals with a variety of cardiovascular diseases, even when performed at home (Aamot et al., 2014; Rognmo et al., 2012), only one study appears to have assessed the safety of HIIT in arthritic populations. This recent pilot study involving young to middle-aged adults with rheumatoid arthritis (n = 7) and juvenile idiopathic arthritis (n = 11), reported no significant change in disease activity or pain during the 10 week HIIT and MICT cycling exercise program (Sandstad et al., 2015).

Therefore, this randomised control pilot trial was conducted to compare two unsupervised home-based (HIIT and MICT) cycling exercise programs in terms of their relative feasibility and ability to improve health-related quality of life, physical function and body composition in middle-aged and older knee OA patients. It was hypothesised that both forms of home-based cycling would be feasible and produce a range of benefits for the participants.

Methods

Results

The demographic characteristics of the participants who completed the HIIT and MCIT cycle programs are described in Table 1. There were no significant differences between the groups for any of the outcome measures at baseline (p = 0.09–0.94).

A summary of the key feasibility outcomes are presented in Fig. 1 and Table 2. As seen in Fig. 1, a total of 27 individuals from the 50 (54% enrolment rate) who were invited/expressed interest to participate were enrolled in this study. As only 17 of the 27 participants completed the trial, the trial had a withdrawal rate of ∼37%, with the most common reasons provided for withdrawal reflecting unrelated illness, unrelated injury or family related issues.

As summarised in Table 2, a high level of exercise adherence was observed for the HIIT (94%) and MICT (88%) groups. Three participants (one in the MICT group and two in the HIIT group) reported adverse events during the eight week study, with the one individual in the MICT group reporting two adverse events relating to discomfort and pain they felt was caused by the bicycle seat. In contrast, 26 adverse events were reported for the HIIT group. Of these 26 adverse events, 24 were reported by one individual who stated that the cycling program aggravated their Bakers cyst behind their knee. Interestingly, this individual still completed 25 of the requested 32 exercise sessions.

The potential efficacy of the two home-based cycling programs with respect to improving the participants’ health-related quality of life, physical performance (function) and body composition is summarised in Table 3. The HIIT group demonstrated significant pre–post-test improvements in WOMAC, TUG and STS performance over the course of the training program; with the MICT demonstrating a significant pre–post-test improvement in the WOMAC only. Based on effect size calculations, all of these significant improvements could be described as moderate effects, with the exception of the WOMAC change for the HIIT group which would be considered a large effect. When comparing the changes between the two groups, the only significant between-group difference was observed for the TUG, whereby the HIIT group improved to a significantly greater extent than the MICT group. No significant improvements were observed in the Lequesne index or any of the body composition outcomes for either of the two groups.

Discussion

The primary aim of this pilot study was to examine and compare the feasibility and safety of unsupervised, home-based HIIT and MICT cycling in middle-aged and older adults with knee OA; with the secondary aim being to gain some insight into the relative efficacy of these forms of cycling for improving health-related quality of life, physical performance and body composition.

The results of the current study are heterogeneous with regards to the feasibility of the two home-based cycling programs for middle-aged and older individuals with knee OA. For example, the research team took 1.5 years to identify the 50 potential middle-aged and older individuals with knee OA. The relatively slow recruitment to the study may reflect a variety of factors, perhaps including concerns from potential participants, medical practitioners and physiotherapists that the HIIT exercise program may aggravate knee OA symptoms or perhaps that relatively few individuals with knee OA had access to a suitable stationary bicycle. Qualitative and/or quantitative studies examining these potential barriers, as well as the facilitators and motives to home-based HIIT cycling may need to be conducted as has been done recently for individuals with knee OA performing group-based aquatic exercise programs (Fisken et al., 2014, 2015), to further improve these recruitment rates.

It must also be acknowledged that only 54% of the potentially interested 50 individuals enrolled in this exercise study. While this enrolment rate is relatively low it should be noted that a total of 15 potential participants (30% of those were initially interested) were deemed ineligible, with the majority of these individuals being located too far away to come in for baseline and post-tests or due to a lack of access to a stationary bike. Furthermore, the 54% enrolment rate of the current study was slightly higher than the 44% and 46%, respectively reported by Mangione et al. (1999) and Rewald et al. (2015), who also performed cycling interventions for middle-aged and older adults with knee OA. Such relatively low enrolment rates further support the need for more research needs to investigate the barriers to enrolment in such cycling studies.

Our results also indicated that only 17 of the 27 (63%) of the original participants completed the cycling program, with most of the reasons for withdrawal not related to the exercise program. This withdrawal rate was similar to Salacinski et al. (2012) who reported 68% of their original participants completed the project; but substantially lower than other cycling studies that observed 83–95% completion rates (Alkatan et al., 2016; Mangione et al., 1999; Rewald et al., 2015). For those participants who continued in our two cycling programs, a very high level of exercise adherence was reported for both the MICT (88%) and HIIT (92%) groups. Such values appear to be consistent (Alkatan et al., 2016; Mangione et al., 1999; Salacinski et al., 2012) or greater (Rewald et al., 2015) than reported in previous cycling studies involving individuals with knee OA.

Analysis of the adverse event data indicated that only three participants (one in the MICT and two in the HIIT group) reported any adverse events, defined as ‘an exercise induced change that worsens an aspect of your condition that is greater than expected day-to-day variation’ over the course of the intervention. Two of these individuals each reported two adverse events, with one individual in the HIIT group unfortunately reporting 24 adverse events over the course of the 25 exercise sessions they performed. While the proportion of participants reporting adverse events (HIIT, 2 out of 9, 22%; MICT, 1 out of 8, 13%) and a total of 24 adverse events for just one participant in the HIIT group appears very high, it needs to be remembered that this is an unsupervised exercise trial involving middle-aged and older adults with knee OA. Further, it could be argued that the severity of these adverse events was quite low as the even the individual who reported 24 further adverse events continued to perform the majority of their requested exercise sessions even though such cycling aggravated their Bakers cyst. While such adverse events may therefore not truly reflect the individual’s knee OA diagnosis, some caution may need to be made when prescribing cycling, especially of a HIIT nature to individuals with Bakers cysts; and to perhaps include Bakers cysts as an exclusion criteria for future cycling studies.

Comparison of our adverse data to that of literature was also difficult as some cycling studies did not report such data (Alkatan et al., 2016; Mangione et al., 1999) and even those studies that did, no clear definition of adverse events was often given (Rewald et al., 2015; Salacinski et al., 2012). It was also observed that some studies stated that no adverse events occurred but then indicated that a number of individuals withdrew from the study due to knee pain (Alkatan et al., 2016; Salacinski et al., 2012), although it was not explicitly stated whether such knee pain was considered an adverse event or whether it was related to the exercise program.

According to a recent review by Wellsandt & Golightly (2018), adherence and the level of improvement resulting from an exercise program for individuals with knee OA may be influenced by a variety of factors including the participant’s preference for the degree of supervision and mode of exercise. This suggests that the true feasibility of the home-based HIIT and MICT cycling observed in the present study could be higher in individuals with knee OA who are able to select their preferred exercise activity. Such a finding would suggest that clinicians and exercise professionals should endeavour to find out the exercise preferences of their patients and match these where possible to the efficacy evidence reported in randomised controlled trials (RCTs) and meta-analyses.

Regarding the efficacy outcomes, both cycling programs demonstrated some significant benefits for the participants. Both groups significantly improved their health-related quality of life as measured by the WOMAC, but not the Lequesne index. The significant improvements in their WOMAC score (HIIT: 14.9 points; MICT: 11.9 points) were comparable or greater than the 10.6–11.6 point improvements reported in the literature for other cycling exercise studies involving knee OA participants (Alkatan et al., 2016; Salacinski et al., 2012). The significant increases in the WOMAC score in the current study appear clinically significant as they exceed the nine point improvement identified by Tubach et al. (2005) as being a clinically relevant change. While the improvements in the Lequesne index were non-significant (HIIT: p = 0.102; MICT: p = 0.081), both groups reported small effect size improvements the HIIT group had a reduction of ∼2 points on the scale that resulted in their overall group being classified as ‘intermediate disability’ at post-test compared to their baseline classification of ‘serious disability’ (Lequesne, 1991, 1997).

The tendencies for the participants to report improved health-related quality of life, including reduced pain, stiffness and disability were consistent with some of the functional changes reported for the HIIT group. Specifically, the HIIT group significantly improved their TUG and STS performance, with the change in TUG performance significantly greater than the MICT group. The magnitude of these changes for the HIIT group (TUG: −1.1 s; STS: +2 repetitions) would appear clinically significant based on the minimally detectable change of 1.1 s for the TUG (Alghadir, Anwer & Brismée, 2015) and 1.64 repetitions for the STS (Gill & McBurney, 2008) reported in previous studies for individuals with knee and/or hip OA. It must however be acknowledged that neither exercise group significantly improved their habitual gait speed over the 3.66 m course. Such a lack of change in gait speed was consistent with the low (40% of heart rate reserve) and high (70% of heart rate reserve) intensity MICT cycle training groups in the study of Mangione et al. (1999), but inconsistent with two cycling studies for individuals with knee OA who reported increases of between 0.08 and 0.20 m/s over distances of 3.66–6 m (Alkatan et al., 2016; Salacinski et al., 2012).

The lack of any significant changes in body composition for either of the two groups in the current study appears relatively inconsistent with the very limited number of cycling studies assessing these outcomes for individuals with knee OA. Specifically, Alkatan et al. (2016) reported that the cycling group significantly improved body mass (−1.5 kg), waist circumference (−3 cm), hip circumference (−2 cm) and visceral adipose tissue (−0.1 kg); although no significant changes were observed for BMI, body fat percentage or lean tissue mass. The discrepancy between the results of the current study and Alkatan et al. (2016) may reflect two primary factors. The first is that since Alkatan et al. (2016) used dual-energy X-ray absorptiometry (DEXA) compared to our study’s BIA, there is the potential that the increased sensitivity of the DEXA was required to observe such body composition changes over relatively small training periods. Further, the duration of the cycling program for Alkatan et al. (2016) was 1.5 times longer (12 vs 8 weeks) than the current study. The potential for cycling, particularly the novel HIIT cycling program examined in this study to improve body composition may also be influenced by the exercise preferences of the individuals with knee OA. In a recent review, Wellsandt & Golightly (2018) found that exercise preference (e.g. degree of supervision and exercise mode) contributes to exercise adherence and to the magnitude of change in body composition outcomes with exercise in individuals with OA. Specifically, studies which randomly allocated participants to an exercise program reported no significant change in body composition, whereas the one study that allowed participants to choose their exercise program (Loew et al., 2017), reported significant reductions in body mass; with those adhering to the walking program also reporting significantly greater reductions in waist circumference.

There were however several limitations associated with the current study. The first limitation reflects the self-report nature of much of the data including the WOMAC and Lequesne index as well as the training and adverse events diary. Even though the participants typically completed their training and adverse event diaries, the actual exercise dose performed is always difficult to quantify during home-based exercise, as no measure of external (power output) or internal (heart rate) workload was collected from the participants. It is also acknowledged that the home-based nature of the intervention may not have optimised the potential benefits of the HIIT program as a trained exercise professional supervising the exercise sessions would typically monitor and progressively increase workloads over the course of the training program. The sample size was also relatively small, which may affect our ability to detect within and between group significant differences and the potential generalisability of the trial findings. Nevertheless, the sample size in the study was greater (Rewald et al., 2015) than or somewhat comparable (Mangione et al., 1999; Salacinski et al., 2012) to some other cycling studies involving middle-aged and older adults with knee OA.

Conclusion

Our results are somewhat supportive of the feasibility of home-based HIIT and MICT cycling programs for middle-aged and older adults with knee OA, with the enrolment rate and adherence rate for both the HIIT and MICT groups comparable to other cycling studies involving similar populations (Alkatan et al., 2016; Rewald et al., 2015; Salacinski et al., 2012). The number of adverse events (HIIT: two of nine participants, 26 adverse events; MICT: one of eight participants, two adverse events) appeared very high, although 24 of these adverse events were reported by one HIIT individual with a Bakers cyst on their knee, who still completed 25 of the requested 32 exercise sessions. Efficacy data demonstrated significant benefits in health-related quality of life (WOMAC) for both groups, with the HIIT group also reporting significant increases in functional performance as assessed by TUG and STS. The major limitation of this study is the small sample size, which may limit our statistical power and increase our chance of type II error. This suggests that larger scale RCTs should further investigate the feasibility and efficacy of HIIT compared to MICT cycling for a variety of knee OA sub-populations; including studies comparing standard supervised and home-based exercise programs.

Supplemental Information