ERS statement on standardisation of cardiopulmonary exercise testing in chronic lung diseases

Pre-test procedures

Subjects are usually asked to refrain from eating for at least 2 h before testing [8] and to avoid strenuous exercise for at least 24 h before CPET. Subjects are advised to avoid caffeine on the day of the test [8] and smoking for at least 8 h prior to the test [11] due to the deleterious impact of carboxyhaemoglobin on the oxygen transport capacity. Subjects are asked to wear comfortable clothing and appropriate shoes suitable for exercise. Upon arrival in the laboratory, subject's medical history (with a complete recording of all medications in use), along with physical examination and recording of vital signs (heart rate (HR), BP, S_pO2 and body temperature) is performed to determine any contraindication to exercise testing. In subjects with diabetes mellitus, it is usually advised to measure blood glucose before exercise to reduce the risk of hypoglycaemia [50]. In case of severe hypoglycaemia, defined as a blood glucose ≤2.8 mmol·L⁻¹ or 50 mg·dL⁻¹, or a hyperglycaemic event in the previous 24 h requiring assistance, exercise testing is contraindicated [50]. Subjects currently taking bronchodilators are asked to administer their usual bronchodilator treatment at least 10 min prior to CPET [51]. Attention is paid to the temperature in the exercise laboratory (20–22°C) [52], as cold air may foster bronchospasm in susceptible individuals.

It is important to make clear to the subjects that sensations during the test such as shortness of breath and a heightened sensation of leg effort are to be expected and their quantification and differentiation will be important for test interpretation. Subjects are informed that the CPET will be terminated at any time they choose to stop. However, it is emphasised that data collected throughout the test will be highly valuable if the participant works to near maximal capacity by the end of the CPET. Subjects are then introduced to a perceived exertion scale and carefully “anchored” to the extremities of the scale as pertaining to both “shortness of breath” and “leg effort”. The 0–10 modified Borg scale is the most widely used in clinical practice wherein “0” means no dyspnoea at all and “10” the worst shortness of breath ever felt [53, 54]. The following is an example of general instructions, used by the Task Force members: “During every stage of exercise, we are going to ask you about the intensity of your breathlessness and leg discomfort at that point in time. As you should avoid speaking during the test, you will use a finger of your hand to point which number between 0 and 10 best reflects the intensity of each of these sensations”. It is good practice for the technician or physician to say the number out loud to confirm the selected number.

Most commercially available CPET systems now allow measurements of exertional inspiratory capacity (IC) to track the behaviour of operational lung volumes (especially relevant in obstructive lung diseases) [55]. For the identification of mechanisms of exertional dyspnoea, measurements of IC are necessary to quantify the magnitude of exercise-induced dynamic hyperinflation. If the measurement of IC is planned during the test, it is important to explain the correct execution of the manoeuvre to the individual before the test with special emphasis on the importance of fully inflating their lungs. An example for general instructions is given in the online supplement [56].

Spirometry should be performed prior to CPET [57]. If direct measurement of maximal voluntary ventilation (MVV) is to be performed, this also should be performed prior to CPET testing [57]. Whilst direct measurement of MVV may be the gold standard, it is highly effort dependent and can be unpleasant to perform with the participant feeling light-headed or dizzy, and in some subjects performance of MVV may induce fatigue. Furthermore, patients often hyperinflate during a voluntary hyperpnoea manoeuvre which can increase their work of breathing relative to the same ventilation during the exercise test [58, 59]. Finally, after an MVV manoeuvre, in the experience of Task Force members, 15 min are required to recover to baseline with respect to ventilation and gas exchange.

Calibration of the flow sensor and the gas analysers of the metabolic cart are usually performed just prior to starting the CPET (see below for more information).

Resting phase

A brief resting period before initiation of exercise allows the patient to familiarise with the testing apparatus (facemask or mouthpiece, ergometer, ECG probes, etc.) and perform baseline measurements of S_pO2, arterial blood sampling (in selected cases only), ECG, BP, V′_E and gas exchange variables. Resting measurements, especially respiratory exchange ratio (RER), can be helpful in identifying patients hyperventilating before exercise and can be compared with data obtained during maximal effort. The literature review showed that the duration of a resting phase was only reported in 42.0% and 14.3% for cycle and treadmill tests, respectively. Of publications selected during the review process, the most common resting phase duration being reported was 3 min (table 2). This duration is in line with previous guidelines and statements [8, 11]. Most Task Force members keep the duration of the resting phase up to 3 min to check that the system is working properly or as long as it is required to obtain accurate and stable resting measurements (e.g. RER <1.0, V′_O2 approximately 3.5 mL·kg⁻¹·min⁻¹) [52]. This is in agreement with the results from the literature review (table 2). It is possible not to take measurements during the entire resting phase but only during the final minute or as long as it takes to reach baseline values. When tests are conducted with oxygen supplementation, a longer resting period may be needed to allow for equilibration of higher fraction of inspired oxygen (F_IO2) in the alveolar space. Patients can be asked to take a few deep breaths to expedite this process. If the measurement of IC is planned during the test, at least three reproducible IC manoeuvres are performed during the resting phase [60].

Unloaded phase

An unloaded phase follows the resting phase. This phase is important in assessing baseline V′_O2 and ventilation associated with vertical leg movements and is necessary for the patient to familiarise themselves with exercise on the treadmill/cycle ergometer and to warm-up adequately. Most commercially available cycle ergometers cannot provide a true minimum load of 0 W. Preferably, the minimum work rate on the cycle ergometer does not exceed 10 W during the “unloaded” phase, and this work rate should be stated in the test report. A three-minute duration for the unloaded phase has previously been recommended [8, 9, 11] and is in line with the literature review results (table 2). In patients with severe lung disease, the unloaded phase may be shorter (but not <2 min) in order to perform the incremental exercise phase before inducing significant amount of fatigue. If a treadmill is used, the lowest speed may be chosen for baseline measurements, e.g. 1.0–1.6 km.h⁻¹ [11].

Incremental exercise phase

After resting and unloaded phases, the incremental exercise phase starts. An incremental exercise protocol adapted to the patient's characteristics is used in order to allow adequate evaluation and follow-up. Ramp and uniform minute-by-minute incremental protocols are both acceptable and provide similar outcomes, when total duration of the incremental phase is the same [61, 62]. According to our literature review, minute-by-minute protocols were used more frequently than ramp protocols for both cycle ergometer (272 versus 216 studies) and treadmill (38 versus 7 studies) exercise testing with 58 (cycle) and 4 (treadmill) studies not reporting the protocol specifications (table 1). While there was a clear preference for minute-by-minute cycle ergometer testing protocols in studies on COPD/emphysema and CF, a protocol preference was less clear for other respiratory conditions. Previous work suggests no differences in physiological parameters (e.g. maximal HR, V′_O2peak and V′_Epeak) between minute-by-minute and ramp exercise testing protocols, in both health and disease [11, 63, 64].

The selected work rate increment allows the duration of the incremental exercise phase to be approximately 10 min, typically ranging between 8 and 12 min [8]. Choice of the rate of increment depends on underlying disease, age, sex, body size, pulmonary function, physical activity, fitness and possible other factors such as genetics and medication [65–67]. The incremental phase is designed to last about 10 min, somewhat shorter or longer tests may also provide useful information. However, the work rate–V′_O2 relationship may not be linear in tests lasting <6 min, and may not be useful to assess submaximal data such as AT [68]. Most importantly, on repeated testing, the same work rate increment is used. If the measurement of IC is planned during the test, the IC manoeuvres are preferably performed every 2 min and, if possible, at the end of exercise [60].

Cycle ergometry protocols

Some studies have provided work rate increment estimation for cycle ergometry based on different variables for patients with COPD [65], CF [66], asbestosis, silicosis and asthma [67]. Other, more general recommendations for the selection of work rate, increments are based on estimated V′_O2peak and V′_O2 during unloaded pedalling [8]. An approach to establish work rate increments similar to that of Wasserman et al. [12] was adopted by our Task Force. Table 1 and table S11 provide information on the work rate increments used and the median peak work rate and V′_O2peak measured, as well as the reported protocol specifications (i.e. median time of each phase, etc.) in the studies included in the literature review which can contribute to the determination of work rate increments in individual patients. Very few studies reported test–retest reliability of V′_O2peak and peak work rate (table S12).

Based on data available in the literature, we estimated median V′_O2peak per kg body weight for the populations of each pulmonary disease/disease condition (table S13). The patient's V′_O2 at unloaded pedalling (in mL·min⁻¹) can be estimated from the patient's body weight according to the equation (150+(6×bodyweight, kg)) [12]. Assuming an average V′_O2 increase of 10 mL·W⁻¹·min⁻¹ and an assumption regarding the kinetic delay between increase in exercise intensity and V′_O2 response [11], work rate increments (in W), in order to obtain a ∼10-min incremental exercise phase duration, are calculated as follows:

Cycling work rate increment (Watts per min)=(estimated V′_O2peak−estimated V′_O2 unloaded (mL·min⁻¹))/92.5

For example, given a patient with CF, 172 cm tall and weighing 60 kg, his estimated V′_O2 during unloaded pedalling is (150+(6×60))=510 mL·min⁻¹. According to table S13 estimated V′_O2peak per kg body weight for this patient should be 34.4 mL·kg⁻¹·min⁻¹, therefore estimated V′_O2peak for this patient should be 34.4 mL·kg⁻¹·min⁻¹ × 60 kg=2064 mL·min⁻¹. Hence, for this patient cycling work rate increment can be calculated as follows:

Cycling work rate increment (per min)=(2064 mL·min⁻¹−510 mL·min⁻¹)/92.5 mL·W⁻¹=16.80 W·min⁻¹

Tables S8 and S9 provide information on weighted median V′_O2peak per kg body weight for multiple lung diseases indicated for cycle ergometer and treadmill testing. Estimated work rate increments have proved difficult to establish due to the heterogeneity of reported patient data sets (>300) and lung diseases (table S13). Depending on a subject's regular physical activity, fitness, reported dyspnoea in daily life (e.g. Medical Research Council dyspnoea scale) and resting lung function impairment (e.g. forced expiratory volume in 1 s (FEV₁)), work rate increments for an individual test will need to be adjusted. Based on the current practice of the Task Force members, adjustments of ±5 W·min⁻¹ are usually sufficient. Nevertheless, validated, easy to use, and disease-specific equations are required and most likely will improve estimation of work rate increment.

Task Force members would only consider repeating a test with a different work rate increment if data from the test are insufficient to answer the key clinical or research questions that were posed when indicating the test.

Treadmill protocols

Various well-established treadmill protocols exist for CPET testing including those of Bruce et al. [69], Nagle et al. [70] and the modified protocol of Naughton et al. [71], during which treadmill speed and/or slope is increased over time. However, none of these protocols, including their modifications, allows a linear increase of work rate as a ramp or minute-by-minute increment. However, ramp protocols with linear increase in work rate have been developed for treadmill testing in healthy adults [43, 60]. Porszasz et al. [43] developed an algorithm using body weight to produce a linear increase in work rate on the treadmill that more closely mimics the linear increase in metabolic rate seen with cycle ergometry. The algorithm of Porszasz et al. [43] has been adapted to assess patients with diseases such as COPD [47, 72]. For other conditions, treadmill work rate increments have not been published. Due to the low number of studies using treadmill protocols identified in this review, disease-specific estimation of work rate increments was felt to be inadequate.

Recovery phase

After the end of the incremental exercise phase (i.e. after reaching the patient's limit of tolerance), all Task Force members employ a recovery period of at least 2–3 min with unloaded pedalling at a reduced cadence of ∼30 revolutions·min⁻¹ or very slow walking. A prolonged recovery period may be indicated for individual patients. The recovery phase is useful for safety reasons and may also be of interest for patients to demonstrate improvements in exercise recovery time after pulmonary rehabilitation and exercise training programmes. During this period, cardiovascular monitoring (ECG, BP, S_pO2, etc.) is typically performed. Since delayed post-exercise HR recovery is frequently observed in lung disease patients and associated with dynamic hyperinflation [73] and poor prognosis [74, 75], HR recordings may offer additional useful information. Furthermore, assessment of perceived exertion, dyspnoea and leg fatigue is done using standardised scales suggested by Borg et al. [76, 77] and Gift [78]. Gas exchange measurements can be terminated and the mask/mouthpiece removed in order to make the patient more comfortable (unless these measurements are required for specific evaluations during the recovery phase), as was done in patients suffering from hyperventilation syndrome [79]. In addition, IC manoeuvres may be performed to assess the rate of recovery of dynamic hyperinflation, but this is mainly done for research purposes.

A flow chart describing the usual practice by Task Force members on the different test protocol phases is given in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Flow chart describing the usual practice by Task Force members on different protocol phases and specifications during cardiopulmonary exercise testing. It is not intended as a recommendation for clinical practice.

V′_O2peak verification phase

A supramaximal verification test is a high-intensity, constant load test performed after CPET [80, 81]. Work rate selected for this task is >100% peak power output achieved during CPET. Details on the methodology of supramaximal verification of V′_O2 are provided in the online supplement. Due to paucity of data in lung disease patients [80–82], the majority of Task Force members abstain from a supramaximal verification test as part of routine clinical CPET. The test may, however, be useful for research questions and in cases with suspected invalid peak data during CPET.

Cycle ergometer

Mechanically braked cycling requires maintenance of a specific cycling rate to achieve the specified work rate [35]. In contrast, performed work rates during electronically braked cycling are directly quantified and can be computer controlled, leading to work rate being maintained over a wide range of pedalling rates (40–70 revolutions per minute) [11]. This allows work rate to be incremented automatically and continuously (e.g. ramp protocol) [11]. According to the American Heart Association (AHA), electronically braked ergometers have become the standard in clinical testing; but are more expensive and less portable [52]. Cycle ergometers used in patient testing are medical devices and need to fulfil all respective requirements. A cycle ergometer needs to be capable of adjusting work rate in increments either automatically or manually [52]. For the most debilitated patients with lung diseases, it is important to use a cycle ergometer that allows very low (near zero) work rates during unloaded pedalling, so that patients do not have to pedal against flywheel inertia [11]. In addition, the cycle ergometer must include adjustable handlebars and seat (and crank arms in smaller children), allowing the knee to be slightly flexed when the pedal is at its lowest position [83]. Meters, dials or digital displays should be appropriately sized and placed for easy reading [52]. It is advised that adaptable pedal grips (i.e. to fix the patients' shoes on the pedal) are employed to optimise performance and increase patients safety [52]. Regular calibration of the equipment is necessary following the manufacturer's specifications. This usually includes measurements of crank arm torque at various speeds using a certified instrument. If not required more often by local regulations or manufacturer, verification of calibration should be done annually and after each service/repair of the ergometer [11]. Further information on calibration procedures and quality control can be found elsewhere [84].

Treadmill

Treadmills for CPET should be electrically driven, are medical devices, and therefore must meet all safety standards [52]. Optimally, these allow the use of a wide range of individualised incremental and/or ramp protocols. Specifically, in patients with lung diseases, it is essential that the treadmill offers very slow walking speed (e.g. 1 km.h⁻¹) and allows inclination of choice. The AHA proposes that a suitable treadmill should minimally be 127 cm in length and 41 cm in width and should accommodate body weights up to 150 kg [52]. It is advised that a padded front rail and at least one side rail are present to optimise patient safety, but minimal or (preferably) no handrail support needs to be encouraged as it influences the metabolic cost of locomotion [68]. An emergency stop button is always visible and directly accessible to staff and patient. As in cycle ergometry, verification of speed and inclination readings of the treadmill should be done at least annually [11].

Mouthpiece versus facemask

Subjects can be interfaced to the metabolic cart system via a bite-block mouthpiece, which is combined with a nose clip and often stabilised by headgear or via a facemask assembly made of pliable rubber that covers the area of the nose and mouth. Although the mouthpiece/nose clip combination is the gold standard for measuring respiratory parameters during an incremental exercise test, it can be uncomfortable for some participants as it makes swallowing difficult, and promotes dry mouth, and/or throat irritations [85]. In contrast, a facemask assembly provides an alternative to mouthpiece/nose clip that allows oral and often nasal breathing, as well as swallowing [86]. However, a major concern is that, with a facemask, it may not be possible to get a tight fit to the face, associated with relevant leakage particularly at high exercise intensities, thus yielding inaccurate respiratory measurements. Furthermore, the dead space (V_D) is usually larger when using a facemask compared to a mouth piece/nose clip. Past studies in athletes and lung disease patients have reported lower values of V′_O2 (ranging from 5.8% to 24%) during facemask breathing as compared to mouthpiece/nose clip [87–90]. In addition, parameters such as V′_E, tidal volume (V_T), carbon dioxide output (V′_CO2) and respiratory rate have been reported to be changed when a facemask was used [85, 86, 89, 90]. Nevertheless, current facemasks are lighter in weight, have improved contact surface area to prevent leakage, and create a more comfortable fit as compared to previous models. Indeed, a recent study in healthy individuals did not report any differences either in gas exchange or breathing patterns between the two different breathing apparatus during both sub-maximal and maximal levels of exercise [91]. In any case, V_D of the system used needs to be fed into the metabolic cart/analysing software.

From a patient's perspective, the survey respondents showed no clear preference between mouthpiece/nose clip and facemask assembly use among 34 (14%) participants who had experience of using both (figures S2–S4). Of these, 12 participants preferred a mouthpiece, 10 preferred a facemask and 12 had no preference. Based on the evidence and patient input, both approaches are, therefore, acceptable. Some additional concerns outlined by respondents who had used both assemblies to those outlined above were that use of the mouthpiece gave pain in the jaw and palette, a sensation of choking and saliva dripping from the mouth can be embarrassing for patients. Additional concerns from using a facemask were that it gave a feeling of tightness/pressure, panic, fear of suffocation and pain from wearing the mask. It would seem that use of each assembly is likely to be a personal preference and therefore, where possible, a choice should be offered to patients following discussion of the advantages and disadvantages of each.

Pulse oximetry

Pulse oximeters rely on the differential transmission of certain light wavelengths by reduced and oxygenated haemoglobin [92] to estimate arterial oxyhaemoglobin saturation. Pulse oximetry is used to monitor S_pO2 during CPET, which gives a general oxygenation estimate providing a trend during the test [11, 68]. However, it is reported as an outcome in only around 30% of published studies using CPET as part of patient assessment (table S10).

A ≥5% fall in S_pO2 during CPET suggests exercise-induced hypoxaemia [11]. In some laboratories, falls of S_pO2 below 85% or 80% constitute an indication to stop the test. Pulse oximeters' accuracy rates during exercise are reasonably good (95% CI 4–5%) when compared with direct measurement in arterial blood oxygenation [93]. Wider confidence limits, however, have been reported, particularly when S_pO2 falls below 88% [94] with even further decrease in reliability in populations with dark skin [95, 96]. Most common inaccuracies in the measurements are due to poor capillary perfusion during exercise (mainly in cardiovascular diseases) and motion artefacts [97] providing inaccurately lower levels of S_pO2. An inaccurate pulse rate reading from the oximeter can sometimes identify these inaccuracies, but this is not always the case. Some authors have reported overestimation of S_pO2 measured with pulse oximetry [98–100].

Moreover, sensor placement (i.e. finger, earlobe, nose, forehead) may also affect the stability and reliability of S_pO2 recordings. For example, in healthy people and those with cardiovascular disease, forehead sensors have better precision compared to finger sensors under various testing conditions (i.e. normoxia, hypoxia and hyperoxia) [101]. Hand motion has been shown to reduce signal quality during finger S_pO2 recordings [102] suggesting use of ear or forehead sensors during cycling or treadmill exercise to be more appropriate. In patients with digital clubbing, for example in CF, finger sensors underestimate S_pO2 readings [103]. Forehead sensors have the advantage that they are less affected by motion and can possibly reduce the measurement error [101].

It should be recognised that pulse oximeters do not detect methaemoglobin or carboxyhaemoglobin and will overestimate oxyhaemoglobin saturation in circumstances where these are elevated [104]. In addition, due to the haemoglobin dissociation curve shape, S_pO2 measurements are relatively insensitive to changes in arterial oxygen tension (P_aO2) at levels corresponding to the plateau of the curve [105] (i.e. S_pO2 will remain above 93% and hardly change at P_aO2 values >70 mmHg).

Arterial blood gas measurements

The measurement of arterial blood gas requires the placement of an arterial catheter prior to exercise, preferably into the radial artery [11]. The clinical decision for arterial blood gas measurements during CPET depends on the purpose of the test; e.g. if pulmonary gas exchange abnormalities are to be expected and noninvasive estimation of oxygenation is considered inadequate [11]. Blood samples can be taken at rest, after the unloaded phase, every other minute during the incremental phase and after 2 min of recovery [11].

In general, measurement of P_aO2 allows the calculation of gas exchange indices such as the alveolar–arterial oxygen tension difference. In this case, arterial blood sampling needs to be done using arterialised capillary blood or from an indwelling arterial catheter at the end of exercise [106]. There is evidence showing that a puncture performed within 20 s after exercise may capture all episodes of significant hypoxaemia (i.e. P_aO2 <60 mmHg) [107]. Moreover, arterial or arterialised capillary blood gas analyses have the potential to provide relevant information on arterial carbon dioxide tension, pH, bicarbonate and lactate levels. The measurement of arterial carbon dioxide tension allows the calculation of V_D/V_T, a measure of the efficiency of CO₂ exchange. P_aO2 levels obtained from capillary blood are less accurate and should be corrected for the systematic bias between adequately detected disturbances in pulmonary gas exchanges with a high sensitivity and specificity as previously described [108]. Of note, arterialised earlobe blood is not a reliable measurement of P_aO2 [109].

Heart rate/blood pressure

Metabolic carts

Accurate recording of respiratory volumes and airflow parameters, as well as inspired and expired gases, requires accurate equipment. With the readily available computers and by utilising specific algorithms, it has become practical to compute respired flow, volumes and gas concentrations breath-by-breath [114]. Pneumotachographs [115], mass flow sensor transducers [116] and turbines, among other devices, are widely used in commercially available metabolic carts for sensing volumes and flows at rest and during laboratory-based exercise testing [117]. These types of transducers must comply with the standards that have been established by the ATS and ERS for flow and volume measurements [57]. These standards specify that the devices must have low V_D, low resistance to breathing flows and be immune to temperature changes, and to the water vapor or pools of saliva that may accumulate during exercise [57]. For example, pneumotachographs are sensitive to gas composition variation and temperature changes [118] as compared to turbine flow meters, that are lightweight and have low V_D. However, pneumotachographs have been considered as more accurate than mass flow sensor transducers [117] With regard to gas analysers, several systems exist, namely gas collection systems, mixing chamber systems, gas-exchange measurement at elevated inspired O₂ concentrations and more highly developed devices without employing the Haldane transformation [119]. The advantages and disadvantages of each system have been previously described in detail [84]. A breath-by-breath system employing a discrete O₂ analyser using paramagnetic or electro-chemical cell technology and a CO₂ analyser using non-dispersive infrared sensor technology or thermal conductivity, as well as mass spectrometry, are suitable for the demands of CPET [11].

For ensuring valid measurements, the examiner has to follow manufacturer-suggested calibration procedures for flow and gas analysers. Both flow sensors and gas analysers tend to drift, therefore calibration procedures immediately preceding each test are recommended. For flow analysers, this includes calibration using standard volume (typically using a certified 3 L syringe) in at least three flow ranges, i.e. 2, 4–6 and 8 L·s⁻¹ to assure linearity. Agreement in calculated volumes within ±3% signifies acceptable performance [84]. The minimum requirements of gas analysers sensors' performance for measurements during room-air breathing are: 1) for O₂ sensors a detection range between 0–100% and for CO₂ sensors a detection range between 0–10%; 2) sensors' accuracy of 1%; 3) sensors' response time of <13 ms; and 4) two calibration points (for each sensor): one in the range of ambient concentration and the other in the range of exhaled concentration [11, 120]. Preferably, two gases with known gas concentrations should be used for calibration rather than only one gas and room air [11]. With the help of a metabolic simulator [121] or testing the same subject (healthy and fit) in at least two or three different submaximal exercise levels (physiological calibration) [122] it is possible to regularly perform a global functioning check for assuring correct ventilatory and gas exchange calculations [84]. The pick-up lead for sampling inspiratory and expiratory gas needs to be regularly replaced by a new one every 3–6 months. Between tests, it is also necessary to change the lead to avoid cross infection.

Test with high oxygen concentration

CPET in the condition of hyperoxia (supplemental oxygen) has several applications such as detection of desaturation during exercise despite oxygen supplementation in patients with intrapulmonary or cardiac right-to left shunt, and response to supplemental oxygen, particularly in patients with COPD [123–125] and, although with limited literature support, in CF and interstitial lung diseases [10]. Hyperoxia, however, can interfere with the performance of the equipment sensors (i.e. oxygen and flow sensors and gas exchange calculation algorithms). Similarly, at high altitude (hypoxic conditions), reduced humidity and barometric pressure as well as extreme temperatures can impose a challenge to the equipment. Supplemental oxygen can alter viscosity of the respired gas (i.e. 11% greater for F_IO2=1 in comparison to F_IO2=0.21). In regard to flow sensors, pneumotachographs are more sensitive to changes in air viscosity, temperature and humidity than turbine volume transducers [11, 126, 127]. In addition, density also influences Pitot tube flow meters [11] and mass flow meters (that measure mass flow in contrast to flow of a volume of air) [116]. As for gas concentration sensors, not all of the available methods to measure concentrations of gases, particularly O₂, are accurate in hyperoxic conditions. Paramagnetic oxygen analysers' ranges might be limited to F_IO2 between 0 and 0.25 [128]. Thus, the technician or physician may check whether the metabolic cart allows recordings of gas exchange variables when the inspired air contains a high fraction of oxygen.

Another important issue to consider is the kinetics response of the analysers that can be lengthened by the increased viscosity of the air in hyperoxic conditions [129].

When intended to use hyperoxic (or hypoxic) conditions, it is mandatory for the equipment to be adequately validated and calibrated for these conditions. Furthermore, calibration gas mixtures should be selected to reflect the concentrations of inspired and expired air expected [84].

Infection control considerations

In a busy cardiopulmonary exercise-testing laboratory with a large heterogeneity of patients and disease entities examined, cross-infection can occur. Therefore, a number of precautions are usually taken and standards applied.

Cross-infection can occur either via direct contact between patients, indirect contact with contaminated surfaces, equipment or healthcare personnel, and inhalation of aerosolised particles or droplets via airborne route, tubing or mouthpieces/facemasks [130]. Direct contact between patients with lung diseases should be minimised in all cases [131, 132]. Appropriate hand hygiene by healthcare workers should be applied according to established standards [133] and supervised occasionally in order to improve compliance. Sufficient time intervals between tests should be used for surface decontamination and room aeration according to local infection control guidelines [134]. Parts of CPET apparatus that might be contaminated (e.g. tubing and hand grips) should be disinfected according to manufacturer's recommendations [135]. In case of contamination that cannot be dealt with by available means of disinfection, all contaminated parts are replaced.

In patients known to be infectious, additional precautions may be considered. For example, to prevent inhalation of microorganisms, barrier filters may be added in case of exercise testing with mouthpieces [136]. However, the additional V_D of the filter should be considered in the setup of the metabolic cart. Moreover, calibration of the flow–volume sensor with the bacterial filter in place is required to accommodate for the additional resistance. Each patient should be provided with a new or disinfected nose clip. Regardless of whether CPET is performed with the use of facemask or mouthpiece, after every test facemasks and multi-use mouthpieces are disinfected according to manufacturer's and the laboratory standards [135].

Basic information on the patient

The report includes name, date of birth, date of test, patient's characteristics (ECG and pulmonary function report may be added), and indication(s) for the test.

Technical report

Secondly, there is a technical report, defining protocol and work rate increment used, along with reason(s) for test termination.

Response to exercise

The report contains the reason for exercise termination, shortness of breath, leg discomfort or both. In addition, Borg scale scores may be provided at the limit of tolerance. Then, the report describes exercise responses at peak exercise and, if reliably assessable, also at AT, as well as slopes, namely the following (table S14).

1) Aerobic/anaerobic response: V′_O2peak (% predicted) and peak work rate (% predicted), V′_O2 and work rate at AT, and V′_O2/work rate slope.

2) Cardiovascular response: HR_max (% predicted) and V′_O2/HR (oxygen pulse) at peak exercise and at AT, along with BP and ECG responses to exercise.

3) Ventilatory response: V′_Epeak and V′_Epeak related to MVV. Breathing reserve and/or ventilatory limitation at maximal exercise is discussed. If available, assessment of dynamic hyperinflation by serial IC measurements is provided.

4) Gas exchange response: S_pO2, as well as ventilatory equivalents (V′_E/V′_O2, V′_E/V′_CO2) and end-tidal O₂ and CO₂ measures at peak exercise. Report V′_E/V′_CO2 slope. If available, arterial blood gas analysis and V_D/V_T (calculated from arterial carbon dioxide tension) are also added.

5) Metabolic response: RER values at rest, AT and peak exercise.

Report summary

Finally, a report summary puts the technical and exercise response information into the context of the exercising subject; e.g. What does this mean for the patient? What disease processes can be diagnosed?

Equipment

The type of ergometer (cycle ergometer or treadmill), metabolic cart, equipment used for ECG and S_pO2 monitoring and pulmonary gas exchange measurements (mouthpiece or facemask).

The calibration procedures if not standard (e.g. calibration gases for testing with supplemental oxygen).

Exercise protocol and measurements

The CPET protocol including duration of resting, unloaded phase (warm-up), incremental and recovery periods.

The work rate used during the unloaded phase.

The choice of work rate increments (i.e. ramp or incremental) and how the increment was determined for individual patients, if different for different patients (i.e. those with severe lung function impairment).

The actual duration of the incremental phase (mean and standard deviation).

How the criteria were defined for reaching the limit of exercise tolerance and/or a maximal effort and how many patients actually met these criteria.

Outcomes

The following variables to describe patient characteristics (separately for different disease entities, if applicable): age, sex, height, weight, body mass index, indices of pulmonary function, V′_O2peak, peak work rate, peak HR, S_pO2 at peak exercise, V′_E/V′_CO2 slope, and V′_E/MVV at peak exercise.

Additional CPET outcomes, if relevant, to describe the patient's or study outcomes.

Report the interpretation of the reason for limited exercise tolerance (i.e. cardiovascular, ventilatory, muscular and/or detraining).