Safety and feasibility of cardiopulmonary exercise testing in head and neck cancer survivors

Assess safety and feasibility of the cardiopulmonary exercise test (CPET) for evaluating head and neck cancer (HaNC) survivors. Also compare their cardiorespiratory fitness to age and sex‐matched norms and establish current physical activity levels.


| INTRODUCTION
Head and neck cancer (HaNC) and its treatment are often associated with a high symptom burden, including weight loss, fatigue, intolerance to physical activity, shoulder and neck dysfunction, head and neck oedema, dry mouth, mouth sores, trismus, dysphagia, pain, depression, and reduced health-related quality of life (Haisfield-Wolfe et al., 2012;Lokker et al., 2013). Exercise programmes undertaken during and/or post-treatment can have positive effects on HaNC survivors, such as improvements in lean body mass, levels of fatigue, physical functioning, and health-related quality of life (Bye et al., 2020;Lynch et al., 2021). The cardiopulmonary exercise test (CPET) provides important information for determining the appropriateness of referring someone to an exercise programme, informing the content and evaluating the efficacy of that programme (ATS/ ACCP, 2003), giving cancer survivors insight into their physical limitations, and setting a baseline for motivating them to exercise (Knutsen et al., 2006). Other indications relevant to HaNC survivors include assessing fitness for surgery and evaluating exercise intolerance and unexplained dyspnoea (ATS/ACCP, 2003). Despite the considerable potential benefits of conducting CPET with HaNC survivors, currently this is typically not a routine procedure in this population.
A CPET involves progressive increments in work rate where the participant is encouraged to exercise until symptom-limited tolerance, during which minute ventilation, pulmonary gas exchange, blood pressure, electrocardiogram (ECG), and oxyhaemoglobin saturation are measured. The CPET is a non-invasive global assessment of the pulmonary, cardiovascular, hematopoietic, skeletal muscle, and neuropsychological systems while experiencing maximal physiological strain, which makes it a unique clinical procedure (ATS/ ACCP, 2003). Moreover, it is considered the most reliable and objective assessment of functional capacity (Robson et al., 2016) and regarded as the gold standard measure of cardiorespiratory fitness (ATS/ACCP, 2003). Limited research suggests that conducting CPETs with survivors of various cancers, such as breast and lung (Jones et al., 2007), and prostate (Wall et al., 2014) is relatively safe and feasible. A recent study investigated the utility of the CPET for perioperative risk assessment in HaNC survivors and observed that lower cardiorespiratory fitness predicts day 5 cardiopulmonary morbidity (Lalabekyan et al., 2021). To our knowledge, however, no study has investigated the safety and feasibility of CPET in HaNC survivors beyond the perioperative period (which extends from the moment of contemplation of treatment through to recovery at home), or after treatment with (chemo)radiotherapy.
An important issue is that peak physiological values obtained from the CPET are supposed to represent the limits of oxygen transport and any factors that result in early test termination (i.e., not related to cardiorespiratory endurance) can potentially invalidate results (ATS/ACCP, 2003). Many unique post-treatment symptoms are the most cited barriers to exercise by HaNC survivors, such as dry mouth and throat, difficulty swallowing, drainage in the mouth and throat, and shoulder weakness and pain (Midgley et al., 2018). Some symptoms can last over 1-year post-treatment (Wulff-Burchfield et al., 2019) and might prove problematic in regards the feasibility of completing a CPET. The maximal nature of a CPET also means the risk of an adverse event is increased compared to rest or a submaximal exercise test. Therefore, the main aim of the present study was to establish the safety and feasibility of CPET for evaluating a heterogenous group of HaNC survivors tested up to  (Miller et al., 2005) to determine forced vital capacity and forced expiratory volume in 1 s. A CPET was then performed that conformed to the joint procedural guidelines of the American Thoracic Society Participants then completed a questionnaire to establish what they perceived as the major and minor contributory factors that influenced the decision to terminate the CPET. This questionnaire has previously been used with apparently healthy individuals and designed primarily based on the responses from semi-structured interviews (Midgley et al., 2017), but was adapted for HaNC survivors using research that reported physical symptoms and barriers to exercise in this population (Haisfield-Wolfe et al., 2012;Lokker et al., 2013;Midgley et al., 2018). The level of weekly physical activity within the 7 days immediately before testing was established for each participant using a self-reported physical activity questionnaire (Godin & Shephard, 1985). The weekly Leisure-time Activity Score was calculated for each participant using the weekly frequencies of strenuous, moderate, and mild physical activity from the questionnaire responses, as follows: (9 × Strenuous) + (5 × Moderate) + (3 × Mild). A weekly Health Contribution Score also was calculated for each participant as follows: (9 × Strenuous) + (5 × Moderate). Health Contribution Scores were then classified according to expected health-related benefits: ≥24 units = active (substantial benefits); 14-23 units = moderately active (some benefits); and <14 units = insufficiently active (less substantial or low benefits) (Godin, 2011).
Participants were instructed to refrain from exercise on the day of the CPET, abstain from smoking for at least 8 h before the test, ingest a light meal no less than 2 h before the test, and to take any medications as prescribed (ATS/ACCP, 2003). Participants also were advised on appropriate clothing and footwear suitable for exercise participation and to arrive at the laboratory wearing this attire.

| Cardiopulmonary exercise test
All CPETs were performed on an electronically braked cycle ergometer (Ergoselect 200; Ergoline GmbH). Participants first sat stationary on the ergometer for 3 min for baseline physiological measurements, followed by a 3-min warm-up consisting of unloaded cycling, and then small work rate increments every 5 s. An appropriate work rate increment was calculated using the guidelines of Cooper and Storer (2001) and was that which predicted the achievement of peak work rate in 10 min during the ramp- Medicine guidelines were used to categorise cardiorespiratory fitness level according to sex and age (Liguori et al., 2022).
Ventilatory reserve was defined as the difference between the peak minute ventilation and maximal voluntary ventilation (predicted from the forced expiratory volume in 1 s × 35; ATS/ ACCP, 2003), and was expressed as a percentage. Exertional dyspnoea and leg discomfort were evaluated at 1-min intervals during the ramp-incremented portion of the CPET using Borg's Category-Ratio (CR10) Scale (Borg, 1998). Participants were given detailed verbal instructions before the CPET on how to interpret this scale. Participants were observed for at least 15 min after CPET termination and ECG, blood pressure, and pulse oximetry were monitored for the first 5 min 30 s to 7 min 30 s of this observation period. The longer monitoring period in some participants reflects that the blood pressure had not decreased sufficiently at the time of the first measurement, and a further measurement was taken. The change in heart rate at peak work rate during the CPET to 1 min into the recovery period was calculated, since a reduction of ≤12 beats·min −1 has been shown to be a powerful predictor of overall mortality (Cole et al., 1999).
Traditional VȮ 2 max criteria were not applied given their apparent limitations and lack of validity (Martin-Rincon & Calbet, 2020; Midgley et al., 2009;Poole et al., 2008). The term 'peak' has been used when reporting physiological values, as no assumption has been made that the data reflect maximal physiological values (Meyer et al., 2005). The CPET outcome variables are reported in accordance with published clinical oncology guidelines .

| Statistical analyses
All data were analysed using SPSS statistical analysis software The relationship between the percentage of predicted VȮ 2 max attained and the Leisure-time Activity Scores was investigated using the Spearman rank correlation coefficient given the non-linear relationship. Sex differences between the Leisure-time Activity Score and Health Contribution Score were analysed using Mann-Whitney U Tests due to non-normal distributions. All statistical assumptions were checked using standard methods (Grafen & Hails, 2002). Twotailed statistical significance was accepted as p < 0.05. Table 1 shows summary statistics for participant characteristics, clinical information, timing of the CPET, and physical activity levels for the 50 participants. Only two participants were medicated with beta-blockers, and none were medicated with other drugs known to lower heart rate, such as ivabradine and non-dihydropyridine calcium channel blockers.

| Participant characteristics
Notable differences were observed between males and females for TNM staging and treatment. For example, only 14% of females in the sample were TNM stage 4 compared to 38% of males. Also, only 5% of females underwent surgery with post-operative radiotherapy compared to 38% of males. The number of days between treatment and presenting for the CPET were not significantly different between males and females (Z = 0.2, p = 0.82). Twenty six percent of the participants were categorised as active, 8% as moderately active, and 66% as insufficiently active. No significant differences were observed between males and females for the Leisure-time Activity Score (Z = 0.6, p = 0.52) or Health Contribution Score (Z = 0.4, p = 0.67).

| Safety
No major adverse events occurred during or within 15 min after CPET termination. The CPET of one participant with a muscle tremor was terminated early by the test administrator for safety reasons, since the ECG response could not be properly monitored due to tremor-related movement artefacts.

| Feasibility
Data for three participants were not included in the analysis of CPET responses. One participant did not start the test because facial disfiguration meant he could not seal his lips around the mouthpiece, and a naso-oral mask would not fit properly. Surgery had left exposed metal from the cheek and a gap continually present between the lips.
One participant could not complete the CPET as knee pain was elicited as soon as she started cycling. The CPET of the third participant was terminated early due to not being able to properly monitor the ECG, as detailed above in the 'Safety' section. The mean (SD) CPET duration of 7:52 (2:29) min:s for the 47 participants that exercised to their limit of tolerance was significantly lower than the target test duration of 10 min based on the prediction equation of  Table 2 shows the factors that 47 participants perceived as contributing to them volitionally terminating the CPET. There were 21 different factors among the cohort that contributed to test termination. The most common major contributory factor was muscle aches in the legs, cited by 70% of the participants, followed by breathing discomfort/breathlessness (32%). In regards factors specific to HaNC, 13% of participants cited dry mouth and throat and/or drainage in the mouth and throat as major contributory factors. No participants terminated the test due to concerns about their safety, although one participant was concerned she would vomit, but this was only cited as a minor contributory factor. Table 3 shows the mean (SD) resting, peak, and recovery CPET responses. The mean (SD) VȮ 2peak of 18.6 (5.7) ml·kg −1 ·min −1 was significantly lower than the mean (SD) predicted VȮ 2 max of 28.3 (7.6) ml·kg-1 ·min −1 based on age and sex (mean difference = −9.7; 95% CI = −11.9 to −7.5; p < 0.001). Significant effects were observed for sex (F = 5.7, p = 0.022, partial η 2 = 0.12), smoking status (F = 12.0, p = 0.001, partial η 2 = 0.23), and TNM staging (F = 3.7, p = 0.019, partial η 2 = 0.22) for the percentage of predicted VȮ 2 max that was
c Calculated as the heart rate at peak work rate minus the heart rate at 1 min into recovery. 43 (25 male, 18 female) of the 47 participants achieved a reduction in heart rate of >12 beats·min −1 at 1 min of recovery. d Measured at between 3 and 6 min into recovery. Measurements taken close to 6 min represent a second measurement taken during recovery because the blood pressure had not decreased sufficiently when the first measurement was taken.
values and establish HaNC survivors' physical activity levels. The main findings were that the CPET was feasible in all but 3 (6%) of the 50 participants, with 47 (94%) of the participants volitionally terminating the test due to symptom-limited tolerance. No serious adverse events occurred. Other findings were that HaNC survivors typically exhibited considerably lower cardiorespiratory fitness than age and sex-matched norms and had low levels of physical activity.

| Safety
No adverse events were observed during the CPET or the 15 min post-CPET observation period, including no ECG irregularities or abnormal blood pressure responses. This contrasts to the 5.9% incidence of adverse events reported for 85 advanced lung and breast cancer survivors during CPET, including three positive tests for myocardial ischemia, one exercise-induced right bundle branch block, and one patient where the CPET induced hip pain following the test that was later diagnosed as lytic metastasis (Jones et al., 2007).
Another study reported a 3.2% incidence of adverse events for 95 prostate cancer survivors being treated with androgen deprivation therapy, which also were related to exercise-induced ST segment depression and right bundle branch block (Wall et al., 2014). The only notable safety issue that arose in the present study was early test termination by the test administrator due to one participant experiencing muscle tremor resulting in substantial ECG artefacts.
These artefacts meant that the ECG could not be properly monitored in regards accurately establishing any changes in heart rhythm or ST-T changes, which is an absolute indication for early test

| Feasibility
Overall, feasibility was good with 47 of the of the 50 participants completing the CPET. One participant could not undertake the CPET because of a HaNC-related issue, in which the oro-nasal mask and mouthpiece could not be fitted properly due to facial disfigurement.
To save time and resources and avoid unnecessary patient burden it may be prudent to check the feasibility of using a mask or mouthpiece before referring HaNC survivors to CPET. Alternatively, a traditional exercise stress test could be performed if a patient presents with this issue on arrival to the laboratory, since exercise stress tests do not involve assessment of minute ventilation and pulmonary gas exchange. An exercise stress test also might be preferable for patients that have had a laryngectomy, due to the considerable technical issues associated with determining minute ventilation and pulmonary gas exchange in these patients (Overstreet et al., 2015).
There were 21 different factors that participants in the present study perceived as contributing to volitional termination of the CPET.
The most common major contributory factor was muscle aches in the legs, cited by 70% of the participants, followed by breathing discomfort/breathlessness (32%). Myers et al. (1992) reported 73% for general or leg fatigue and only 14% for breathlessness. The difference between studies relating to breathing could be due to random variability, although another plausible explanation is that Myers et al. (1992) used treadmill testing and involved apparently healthy participants that were, on average, 16 years younger than the cohort in the present study. Dry mouth or throat was the most common perceived barrier of 36 barriers to engagement in exercise cited by HaNC survivors, with 40% of participants perceiving them as 'often' or 'very often' being a barrier to exercise (Midgley et al., 2018).
High minute ventilation and the effects of the silicone orofacial mask might predispose HaNC survivors to mouth and throat drying during a CPET, especially considering fluid cannot be ingested during this time. However, only 5 (10%) of the participants in the present study reported dry mouth and throat as a major contributory factor for CPET termination. This could be due to the relatively low CPET duration that might somewhat mitigate the risk of mouth and throat drying when compared to more prolonged bouts or exercise performed as part of a training programme. Shoulder weakness and/or pain was the sixth most common perceived barrier to exercise (Midgley et al., 2018), however, no participants in the current study cited this as a factor influencing test termination. This may be explained by the focus on lower body musculature during a cycling CPET. It is possible that performing the CPET on a treadmill would have altered these results due to the greater involvement of upper body musculature.

| CPET responses
The mean observed VȮ 2peak was 9.7 ml·kg −1 ·min −1 (34%) lower than MET as a standardised value of 3.5 ml‧kg −1 ‧min −1 ). Notably, each 1 MET increase in cardiorespiratory fitness is associated with a 13% and 15% reduction in all-cause mortality and cardiovascular events, respectively, in apparently healthy men and women (Kodama et al., 2009 during the CPET (Meyer et al., 2005). An important issue is that there are currently no robust methods to establish with a sufficient level of confidence that VȮ 2 max has been attained. Traditionally, a plateau in the VȮ 2 response at the end of the CPET has been regarded as the primary criterion (Midgley et al., 2007). The validity of this criterion is questionable, however, since many participants do not exhibit a VȮ 2 plateau despite an apparent maximal effort (e.g., Doherty et al., 2003;Rossiter et al., 2006). One study even reported that more participants exhibited an accelerated VȮ 2 response (27%) than plateau-like behaviour (17%) at the end of a CPET (Day et al., 2003). In the absence of a VȮ 2 plateau, secondary criteria have been used that are based on participants attaining arbitrary thresholds for peak values of the respiratory exchange ratio, heart rate, and post-exercise blood lactate concentration (Midgley et al., 2007). Secondary criteria also have questionable validity since there is a large inherent interindividual variation in the maximal values that individuals can attain for each of these variables and the criterion thresholds often can be satisfied well before VȮ 2peak has been attained (Midgley et al., 2009;Poole et al., 2008). Due to the limitations in establishing whether VȮ 2 max has been attained, it could be argued that the percent predicted VȮ 2 max values reported in the present study have questionable validity. However, it is important to appreciate that the study from which the predicted VȮ 2 max values were derived also was subject to the same limitations in establishing whether 'true' VȮ 2 max values were attained. The mean peak ratings of perceived exertion during the CPET of around 5 was relatively low, which could indicate that some participants gave a poor effort, since patients typically stop exercise at ratings of 5-8 (ATS/ACCP, 2003). This discrepancy might be explained by the fact that the present study used differential perceived exertion (i.e., dyspnoea and leg discomfort) rather than whole body exertion. Notably, the mean peak respiratory exchange ratio of 1.17 indicates substantial lactic acidosis and suggests a good effort when considering the cohort overall (ATS/ ACCP, 2003).

| Physical activity levels
Low levels of physical activity were observed in the present study.
Similarly low levels have been reported in previous studies for HaNC survivors in the United States (Rogers et al., 2006), Netherlands (Douma et al., 2020) and Sweden (Karczewska-Lindinger et al., 2021).
Moreover, the physical activity levels of HaNC survivors have been found to be significantly lower post-treatment compared to pretreatment (Sammut et al., 2016). The Health Contribution Score for each participant in the present study highlighted that 26% of the participants were categorised as active, 8% as moderately active, and 66% as insufficiently active. Based on these data, two-thirds of participants were insufficiently physically active to gain appreciable health benefits (Godin, 2011), although the real figure may be more than this. For logistical reasons, a self-reported physical activity questionnaire was used in the present study and participants tend to over self-report the amount of physical activity they perform compared to when an objective measure such as accelerometery is used to assess physical activity levels (Troiano et al., 2008).