Physical activity end-points in trials of chronic respiratory diseases: summary of evidence

Search strategy and overview of the outcomes

To evaluate physical activity measures with a focus on patient relevance, the variety of end-points was first assessed using Citeline's database Trialtrove (a database, constantly updated, covering the entire public domain using major – and over 40 000 unique information sources – i.e. trial registries, portals, PubMed; figure 1). We searched for the terms “Trial Title contains Physical Activity” OR “Trial Title contains exercise” OR “End-point is Daily Physical Activity” OR “Primary End-point contains Physical Activity”. It is noteworthy that the inclusion criteria included “Trial title contains Exercise”: frequently studies claim to measure exercise when they record daily physical activity levels through end-points such as steps per day. To prevent inclusion of studies truly measuring exercise and endurance, any exercise intervention or exercise challenge studies were then manually excluded from the search results (figure 1). For each indication, the top two to four most frequently used end-points were cross-compared using resources from Trialtrove, supplemented with primary literature found through end-point-specific Google Scholar and PubMed searches. The choice of end-point comparison was owing to the availability of evidence relating to each end-point, with both time and resource limitation taken into consideration. The end-points were compared across multiple factors, including their construct and content validity. Construct validity refers to ensuring the end-point measures what it claims to be measuring, whereas content validity ensures the end-point is measuring appropriate content.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Search strategy diagram. Workflow to assess landscape of end-points which assess physical activity in respiratory trials; details of inclusion/exclusion criteria used for Trialtrove searches; and end-point evaluation criteria. IPF: or idiopathic pulmonary fibrosis.

The search criteria returned 15 studies in patients with COPD, 6 in patients with asthma and 2 in patients with IPF (table 1). The COPD studies identified were published between 2007 and 2018, whereas the asthma and IPF studies were published more recently, between 2017 and 2018 and between 2018 and 2019, respectively. Step count was used most frequently across all the indications, appearing in 16 out of 23 clinical trial results. This was followed by time spent in moderate-to-vigorous physical activity (MVPA) in 10 out of 23 clinical trials, the 6-min walk distance (6MWD) in 9 studies and activity-related energy expenditure (AEE) in 5 studies. Two novel instruments have been used specifically in patients with COPD: the daily-patient-reported outcome (PRO)active and clinical visit-PROactive physical activity (D-PPAC and C-PPAC). D-PPAC and C-PPAC are hybrid instruments which combine a patient-reported outcome (PRO) with accelerometer-derived data, to capture the amount of physical activity and patient experience during activity [10]. Both instruments were developed and validated specifically within the COPD patient population. Other end-points used included time spent in light physical activity, sedentary time, active time, “healthy lifestyle”, duration of exercise, intensity of exercise and time spent in degrees of activity as determined by the metabolic equivalent of task (METs). The variety and inconsistency of end-points used reflects the unmet need for relevant and validated physical activity measures for use in clinical trials.

Assessing physical activity in patients with COPD

The European Medicines Agency (EMA) qualified both D-PPAC and C-PPAC as suitable instruments to capture physical activity experience in patients with COPD and are supported as end-point use in clinical trials [11]. Both instruments demonstrate strong construct validity, content validity, with scores reflecting COPD status, positively affected by bronchodilator therapy, and negatively affected by exacerbations (table 2) [10]. Additionally, out of 1595 patients who participated in 7 validation studies, 83% of patients produced sufficient data from both accelerometer (8 h wearing time on at least 3 days across 1 week recording) and respective questionnaires, confirming sufficient acceptability from a patient perspective, across multiple nations, disease severities and languages [10].

In comparison, step count displays some aspects of construct validity such as correlations to dyspnoea and exercise capacity [12, 13], but an inconsistent relationship with health-related quality of life questionnaires (HRQoL) [12, 14]. Content validity is poor, as it cannot portray intensity of activity nor patient experience during daily activities. Step count does however show sensitivity to severity of COPD [12] and improvement following pharmacological intervention [13, 15, 16]. The 6MWD has commonly been used as a surrogate end-point for physical activity prior to introduction of pedometers and commercialisation of activity monitors. Despite the fact the 6MWD captures functional capacity and not daily activity (demonstrating poor content validity), it remains a popular end-point within respiratory research to indicate patient activity levels. 6MWD shows some correlation with levels of dyspnoea [17] and HRQoL [18], inverse correlation with disease severity [18] and sensitivity to therapeutic intervention [15, 17].

Assessing physical activity in patients with asthma

In contrast to the COPD population, less work has thus far been conducted in patients with asthma with respect to physical activity. The majority of studies have utilised the end-points step count and time spent in MVPA (table 3). Of the two end-points, step count offers more advantages: it is an intuitive end-point, easily understood by patients, and easily assessed by wearable user-friendly gadgets. Fundamentally, step count is an important measure owing to the fact that patients with asthma do not complete the recommended 10 000 daily steps per day and the consistent (but limited) findings that step count declines with worsening disease status [8, 21, 22]. Step count is responsive to intervention, specifically improving with anti-5 therapy [23]. Furthermore, step count associates with dyspnoea and exercise capacity [8], two end-points which significantly impact quality of life. Conversely, step count can be viewed as a crude representation of physical activity which is impacted by occupation and does not at first glance reflect patient experience. However, recent evidence may suggest otherwise; in 2020, Neale and colleagues [24] showed that step count in patients with asthma is inversely correlated with HRQoL.

The concept of measuring time spent in MVPA by patients is meaningful and perhaps has potential to be a clinically useful physical activity end-point. Increasing the time spent in MVPA has endless physical, mental and social benefits for patients, and largely this end-point is not impacted by occupation. Unfortunately, the initial studies present inconsistent findings. Firstly, the raw values of time spent in MVPA by both asthma and healthy populations varies substantially between studies [8, 22, 24]. Secondly, time spent in MVPA is not significantly different between asthma and healthy populations in the studies describing it, once adjusting for confounding factors, such as in a study by Bahmer and colleagues in 2017 [22]. Finally, there are no data available looking at the effect of treatment on time spent in MVPA in patients with asthma.

Assessing physical activity in patients with IPF

There are few clinical trials that have investigated physical activity in patients with IPF (table 4). Patients are reportedly completing 2728±2475 steps per day on average, a variation in steps almost as large as the step count itself [26]. A study by Nakayama and colleagues in 2015 [27] found that patients averaged 6520 steps per day, and that during the monitoring period over 1 month, there was no significant day-to-day variation. Both studies show that IPF patients complete less steps than the recommended 7000 per day for older adults, and other studies show initial indications of associations between step count and clinically important end-points such as serum Krebs von den Lungen (KL)-6 [27], dyspnoea [28], lung function measures [28], HRQoL [28] and 6MWD [26–28].

Time spent in MVPA was recently approved by the Food and Drug Administration (FDA) as a Phase III primary end-point in Bellerophon Therapeutics Inc's study investigating the inhaled nitric oxide treatment of pulmonary hypertension associated with interstitial lung disease, confirming the validity of MVPA as a clinically meaningful end-point [29]. This was due to the positive results reported from cohort 1 of their ongoing Phase 2b/3 study, where patients on active treatment demonstrated a 34% placebo-adjusted improvement in MVPA after 8 weeks [29]. It is noteworthy that this Phase 2b/3 study also measured step count, but the largest difference in activity between treatment arms was shown through MVPA. A study by Hur and colleagues in 2018 estimated that an increase of MVPA by 26 min a week is a realistic but beneficial goal for patients with fibrotic interstitial lung disease; this study included a subset cohort of IPF patients. Time in MVPA has proven to correlate with exercise capacity in patients with IPF [30] and HRQoL [31].

AEE is the relative energy expended to perform a task above resting metabolism [26]. AEE recorded in IPF patients is significantly less than expended in healthy controls, averaging 133±127 kcal per day in IPF patients compared to 201±111 kcal per day in healthy controls [26]. AEE has been shown to be correlated with 6MWD and survival of IPF patients [26], dyspnoea and serum KL-6 [27]. Two papers measuring 6MWD in IPF patients have shown the end-point's unsuitability as a surrogate marker for physical activity, as it accounts for a low percentage of the variance observed in step count [28, 32]. There is initial data suggesting that 6MWD associates with dyspnoea and quality of life [28], and predicts mortality [32]. However, there are no published data to suggest 6MWD reflects IPF severity or treatment response.