Abstract
Cardiopulmonary exercise testing (CPET) is a comprehensive and invaluable assessment used to identify the mechanisms that limit exercise capacity. However, its interpretation remains poorly standardised. This scoping review aims to investigate which limitations to exercise are differentiated by the use of incremental CPET in literature and which criteria are used to identify them. We performed a systematic, electronic literature search of PubMed, Embase, Cochrane CENTRAL, Web of Science and Scopus. All types of publications that reported identification criteria for at least one limitation to exercise based on clinical parameters and CPET variables were eligible for inclusion. 86 publications were included, of which 57 were primary literature and 29 were secondary literature. In general, at the level of the cardiovascular system, a distinction was often made between a normal physiological limitation and a pathological one. Within the respiratory system, ventilatory limitation, commonly identified by a low breathing reserve, and gas exchange limitation, mostly identified by a high minute ventilation/carbon dioxide production slope and/or oxygen desaturation, were often described. Multiple terms were used to describe a limitation in the peripheral muscle, but all variables used to identify this limitation lacked specificity. Deconditioning was a frequently mentioned exercise limiting factor, but there was no consensus on how to identify it through CPET. There is large heterogeneity in the terminology, the classification and the identification criteria of limitations to exercise that are distinguished using incremental CPET. Standardising the interpretation of CPET is essential to establish an objective and consistent framework.
Shareable abstract
There is large heterogeneity in the interpretive strategies to identify limitations to exercise with CPET. To provide a framework for objective and consistent CPET interpretation, there is a need for standardisation of CPET interpretation. https://bit.ly/3Vs3cHE
Introduction
Cardiopulmonary exercise testing (CPET) is a type of stress test predominantly performed on a cycle ergometer or treadmill during which pulmonary gas exchange, heart rate, blood pressure, electrocardiogram and oxygen saturation are recorded at rest and while exercising, along with symptoms. Sometimes more invasive measurements, such as blood lactate levels or arterial blood gasses, are also collected. CPET allows the holistic evaluation of the respiratory, cardiovascular, neurosensory and skeletal muscle system at the time of physical exertion [1]. The test is usually conducted using an incremental exercise protocol, wherein the work rate is progressively increased in a continuous or stepwise manner over time until maximal exercise is reached. Constant work rate tests are also used, especially for research purposes and to evaluate the impact of therapeutic interventions [2].
CPET is performed for a variety of reasons in a clinical setting. It is used to determine a person's cardiorespiratory fitness level, to estimate prognosis in patients with known cardiac or respiratory disease, to assess perioperative risk and to evaluate the impact of therapeutic interventions [3, 4]. However, one of its main indications is to identify the mechanisms that limit exercise tolerance and to determine the aetiology of exercise-induced symptoms [5]. This is helpful for diagnostic purposes, especially in people with unexplained dyspnoea on exertion, because they will often show a normal workup at rest [6]. Although CPET rarely pinpoints a specific diagnosis, it helps to narrow the differential diagnosis and guide further investigations [7]. CPET results within normal limits can also limit further, unnecessary testing. In patients with known cardiac or pulmonary disease, identification of the cause(s) of exercise limitation can reveal potential therapeutic targets and guide therapy or rehabilitation strategies, especially when multiple comorbidities are present [8–10].
The interpretation of CPET is dominated by the principle that the body's oxygen consumption (VʹO2) is the primary measure of exercise capacity [10]. During exercise, skeletal muscles need more energy in the form of ATP in order to sustain contractions. The increased ATP demand is accompanied by a greater requirement for oxygen because ATP production primarily relies on oxidative metabolism. To meet this increased oxygen demand, the different steps of the oxygen cascade, which describes the transport of oxygen from the atmosphere to the mitochondria, have to adjust accordingly [11]. This cascade roughly consists of the uptake of oxygen by the pulmonary system into the bloodstream, where it mainly binds to haemoglobin in the red blood cells, the delivery of oxygen to the exercising muscle tissue by the cardiovascular system and the diffusion of oxygen into this muscle tissue, where it is utilised in the mitochondria. Physiologically, exercise is considered maximal when one or more steps of the oxygen cascade are loaded to their limit [12]. In other words, exercise capacity is limited by the “weakest” step(s) of this cascade. In general, cardiac output is regarded as the principal determinant of maximal VʹO2 in healthy subjects, but recently there has been debate on whether oxygen diffusion into the muscle tissue is a (contributing) limiting step of the cascade [13, 14]. Additionally, even though historically the pulmonary system was considered overbuilt for exercise in healthy individuals, it is now known that well-trained subjects can reach or even surpass their predicted pulmonary ventilation, as they have superior cardiovascular function [15, 16]. By measuring key variables that reflect the response to exercise of the involved organ systems, CPET can help to evaluate which systems showed an abnormal response to exercise and/or reached its limits.
The interpretation of CPET remains poorly standardised. As a result, a uniform classification of limitations to exercise (also called exercise limitations, causes of exercise limitation or exercise limiting factors) that can be distinguished using incremental CPET is lacking and there is no consensus regarding the criteria to identify these limitations. Therefore, the aim of this scoping review is to explore the existing literature to gain insight into which limitations to exercise are differentiated using incremental CPET and which criteria should be used to identify these limitations when interpreting CPET results. The scoping review will also highlight areas or uncertainty in need of further investigations.
Methodology
This scoping review was performed following the updated methodological guidance for the conduct of scoping reviews by Peters et al. [17] and reported according to the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA-ScR) checklist guidelines [18].
Databases and search strategy
An electronic literature search of PubMed, Embase, Cochrane CENTRAL, Web of Science and Scopus was performed twice, the last one was done on 5 September 2023. The search strategy was developed together with an experienced librarian and was as follows: (“Exercise Test”[Mesh] OR “exercise test*”[tiab] OR CPET[tiab]) AND (“limitation*”[tiab] OR “limiting factor*”[tiab] OR “limiting cause*”[tiab]) for PubMed; (“cardiopulmonary exercise test”/exp OR “exercise test*”:ti,ab,kw OR CPET:ti,ab,kw) AND (“limitation*”:ti,ab,kw OR “limiting factor*”:ti,ab,kw OR “limiting cause*”:ti,ab,kw) for Embase; ([mh “Exercise Test”] OR ((exercise NEXT test*) OR CPET):ti,ab,kw) AND ((limitation* OR (limiting NEXT factor*) OR (limiting NEXT cause*)):ti,ab,kw) for Cochrane Central; TS=(“exercise test*” OR “CPET”) AND TS=(“limitation*” OR “limiting factor*” OR “limiting cause*”) for Web of Science; and TITLE-ABS(“exercise test*” OR “CPET”) OR AUTHKEY(“exercise test*” OR “CPET”) AND TITLE-ABS(“limitation*” OR “limiting factor*” OR “limiting cause*”) OR AUTHKEY(“limitation*” OR “limiting factor*” OR “limiting cause*”) for Scopus.
The results of the initial search were imported into EndNote 20. Duplicates were identified and removed. One researcher (M.S.) shortlisted the search results based on conservative title and abstract screening, excluding studies that did not fulfil the selection criteria. Consequent full-text screening was done independently by two researchers (M.S. and I.G.). When discrepancies were present, a consensus-based decision was reached after discussion. Backward citation searching by reviewing reference lists of included publications was done subsequently.
Selection criteria
Inclusion criteria encompassed 1) any type of publication, 2) that provided clear identification criteria for at least one limitation to exercise 3) based on variables obtained during incremental, maximal CPET (also including blood sampling) and clinical parameters, and 4) published in English, French, Dutch or Spanish. In addition, studies that validated the presence of limitations to exercise with methods other than conventional CPET were considered for inclusion if they also reported corresponding CPET variables. This was done to be able to uncover the scientific evidence behind the use of certain thresholds. Studies were only included if they used incremental CPET performed on cycle ergometer in upright position or on a treadmill. Exclusion criteria were 1) publications including paediatric populations and 2) publications that only relied on invasive haemodynamic CPET parameters, since these invasive haemodynamic measurements are not routinely performed in clinical practice.
Data extraction and reporting
M.S. extracted all necessary data into Microsoft Excel. Information on publication type, author, publication year, limitation(s) to exercise and their identification criteria were extracted. If mentioned, the source of the used criteria to identify limitations to exercise was collected, along with the criteria to define the maximality of a CPET. In the case of research studies, information about the study design, study population and method of conducting the CPET was also gathered. If studies compared different identification criteria for limitations to exercise, results on the found association were included.
Results
Search results
The search identified 10 052 records. After removing 5076 duplicates, 4976 unique records were screened by title and abstract. 74 publications met the selection criteria. An additional 12 publications were found through backwards reference searching, resulting in a total of 86 publications included in this scoping review. The PRISMA flowchart with visualisation of the screening process is shown in figure 1.
PRISMA flowchart of the included publications. CPET: cardiopulmonary exercise testing.
Characteristics of included publications
General characteristics of the included publications are shown in table 1. All included publications were published between 1981 and 2023. 57 publications were primary literature [19–75] and the remaining 29 consisted of secondary literature [76–104]. Four studies compared the use of different criteria to identify limitation(s) to exercise [56–59] and two studies validated the presence of a limitation to exercise with another method than conventional CPET [62, 63]. The main findings of these studies are summarised in tables 2 and 3. All extracted data of other included publications can be found in supplementary tables S1 and S2.
Characteristics of the included publications (total, n=86)
Studies comparing the use of different identification criteria to identify limitation(s) to exercise
Studies using additional methods to assess the presence of a certain limitation to exercise
Maximality criteria for CPET
38 publications mentioned criteria to assess the maximality (or submaximality) of CPET. Most of these publications only required the achievement of one criterium to identify CPET as maximal.
By far the most commonly used variable to evaluate CPET maximality was a high respiratory exchange ratio (RER). However, there was no consensus on an optimal cut-off value, as illustrated by the different values proposed by various statements included in this scoping review, namely >1.15 by the 2003 American Thoracic Society (ATS)/American College of Chest Physicians (ACCP) statement [78], >1.10 by the 2007 American Heart Association (AHA) statement [97], >1.00 by the 2012 Heart Failure Association (HFA)/European Association of Preventive Cardiology (EAPC)–European Society of Cardiology (ESC) statement [100] and >1.05 by the 2019 European Respiratory Society (ERS) statement [86].
The attainment of a certain heart rate [24, 28, 39, 41, 42, 70, 78, 84, 86, 88, 92–94, 103], a plateau in VʹO2 [28, 39, 41, 70, 78, 84, 86, 88, 94, 97, 103], i.e. no further increase in VʹO2 despite an increase in work rate, high Borg score(s) rating exertional symptoms [28, 41, 42, 65, 70, 72, 78, 84, 93, 94] and the presence of ventilatory limitation [41, 70, 84, 86, 88, 92, 94] were other frequently used maximality criteria. Less often used criteria were the attainment of a certain peak VʹO2 (VʹO2peak) [24, 41, 70, 78, 86, 88, 92] or work rate [41, 78] and high lactate levels (defined as >8 mmol·L–1 [70, 86, 102], >5 mmol·L–1 [60] or above the predicted value [69]).
While Kersten et al. [60] claimed that oxygen desaturation indicates maximal exhaustion, the 2003 ATS/ACCP statement [78] and Datta et al. [85] explicitly stated that oxygen desaturation is not a marker of a maximal CPET.
All maximality criteria that were used in included publications can be found in supplementary table S3.
Classifying limitations to exercise with CPET
The most commonly reported limitations to exercise, categorised by site of exercise limitation, are provided in figure 2. 11 publications only identified the limitation(s) to exercise if exercise capacity was reduced, defined as VʹO2peak <80% [19, 37, 46, 71], <83% [21, 25] or <85% [45, 60, 86, 92, 95] of the predicted VʹO2peak. In contrast, the other publications identified the limitation(s) to exercise independently of VʹO2peak.
Bar chart of most commonly reported limitations to exercise. Limitations to exercise at the level of the respiratory and cardiovascular system that were reported twice or less were included in the category “other”, while for limitations at the level of the peripheral muscle and other limitations only limitations mentioned once were included in this category. #: Pulmonary vascular limitation was categorised as a limitation at the level of the respiratory system, but it also involves the cardiovascular system. EIB: exercise-induced bronchoconstriction.
Identifying limitations to exercise with CPET
Identifying limitations at the level of the cardiovascular system
The most often used terms to describe exercise limitation at this level were cardiovascular limitation, cardiac limitation and circulatory limitation. These terms were interchangeable, aside from one study that distinguished cardiac from circulatory (and also ischaemic cardiac) limitation [20]. Two categories could be broadly distinguished: a physiological cardiovascular limitation and a pathological cardiovascular limitation. This distinction arises from the fact that a physiological cardiovascular limitation is typically considered the default finding in healthy people, as mentioned in the introduction [84].
The 2003 ATS/ACCP statement [78] proposed to identify a physiological cardiovascular limitation by reaching >90% of the predicted heart rate and having a heart rate reserve of <15 age beats·min–1. Eight additional publications mentioned identification criteria that implied the reaching of a physiological cardiovascular limitation. They relied on peak heart rate to identify this limitation, with reaching >80% [36, 53], >85% [75] and >90% [47, 67] of the predicted maximal heart rate used as thresholds. Different age-dependent formulas to estimate peak heart rate were used between publications. These formulas were 220−age beats·min–1 [53, 75], 210–(0.65×age) beats·min–1 [28] and 202–(0.72×age) beats·min–1 [45]. Criner et al. [67] used 220−age beats·min–1 for men and 226−age beats·min–1 for women.
The other 27 publications that mentioned cardiovascular, cardiac, circulatory or cardiocirculatory limitation implied the presence of a pathological limitation. The most frequently used identifier for this limitation was a low or early plateau in oxygen pulse (n=17), calculated as VʹO2/heart rate, specified as <70% pred [88, 102], <80% pred [28, 32, 72], <81% pred [73] or <12 mL·beat–1 in men and <10 mL·beat–1 in women [43]. Other frequently used criteria were ECG changes (n=10), blood pressure abnormalities (n=8), a low VʹO2/work rate relationship (n=8), chronotropic incompetence (n=7), arrhythmias (n=4) and angina pectoris (n=2). An abnormal blood pressure response was defined as exercise systolic blood pressure <120 mmHg [103], a fall in systolic blood pressure >10 mmHg [22, 28], values of >220/120 mmHg [28] or >250/120 mmHg [41] and a lower blood pressure at peak exercise compared with blood pressure at rest [45]. 10 publications included a decreased anaerobic threshold (AT) as indicative of cardiovascular limitation, usually described as AT at <40% of predicted normal VʹO2peak [20, 21, 43, 73, 86, 88, 92, 99, 102, 103].
Identifying limitations at the level of the respiratory system
Identifying ventilatory limitation
Ventilatory limitation was by far the most frequently identified limitation to exercise and breathing reserve (BR) was typically looked at for its identification. BR is the difference between the maximal voluntary ventilation (MVV) and the maximum exercise ventilation, which can be expressed in absolute terms in litres per minute or in relative terms as a fraction of the MVV. All but two publications [19, 81] that identified ventilatory limitation included low BR as a criterium and a majority of studies that formulated fixed identification criteria even used it as only criterium (n=27/37). The 2003 ATS/ACCP statement [78], the two included ERS statements [79, 86] and the 2018 ERS monograph on clinical exercise testing [103] proposed a cutoff value of a BR less than 15% to identify ventilatory limitation, based on the lower limit of the 95% confidence interval in healthy populations. This threshold was adopted by most reviews (n=12/17) [80, 82, 84, 85, 87, 89–91, 94, 95, 99, 104] and studies using fixed criteria (n=23/37) [27, 31, 37, 38, 41, 44–48, 50–52, 64–67, 69, 70, 72–75]. However, other cut-off values such as BR less than 40% [77], 30% [20, 24, 29, 34, 39, 68, 96, 98] and 20% [21, 25, 33, 36, 42, 53, 55, 61] were also used. In absolute terms, a BR of less than 11 L·min–1 was most commonly used [26, 47, 66, 69, 72, 89, 91, 95], but thresholds of less than 8 [67], 12 [55] or 15 L·min–1 [52, 103] were also utilised. The influential book Wasserman & Whipp's Principles of Exercise Testing and Interpretation by Sietsema et al. [101], hereafter referred to as the Wasserman textbook, states that ventilatory limitation is likely when BR is less than 10% or 11 L·min–1.
There was no consensus on how to estimate MVV. MVV can be measured directly based on a 12–15 s manoeuvre during which an individual performs rapid and deep breathing or it can be calculated indirectly based on forced expiratory volume in 1 s (FEV1). Each method has its own shortcomings [78]. Multiple formulas for indirect MVV measurement were reported, such as multiplying FEV1 by 41 [20], 40 [23, 26, 27, 32, 46, 52, 58, 66, 81, 82, 86, 93, 97, 99, 101, 103], 37.5 [36, 57], 35 [34, 40, 41, 43, 47, 75, 85, 93] and (FEV1×20.1)+15.4 [40]. The 2019 ERS statement [86] and the Wasserman textbook [101] advise using indirect MVV, preferably calculated as FEV1 multiplied by 40, rather than direct MVV, unless there are compelling circumstances such as inspiratory obstruction or severe neuromuscular disease.
Another frequently used method to identify ventilatory limitation was the assessment of dynamic respiratory mechanics, based on serial inspiratory capacity (IC) manoeuvres at rest and during exercise. This way, operating lung volumes throughout exercise can be measured and exercise tidal flow–volume loop (FVL) can be compared with the maximum FVL. 18 publications mentioned the measurement of dynamic respiratory mechanics, but only 13 of them proposed criteria derived from IC manoeuvres to identify ventilatory limitation. The generally used criteria reflected the presence of significant expiratory flow limitation (measured as the ratio of overlap between the tidal breathing and maximum expiratory FVL, with 25% [88, 103], 30% [84], 40% [59, 86, 94] and 50% [100] overlap used as cutoff values), the occurrence of dynamic hyperinflation (identified by a decrease in IC with more than 100–200 mL [59, 86, 88, 94, 100, 102, 103] or increase in end-expiratory lung volume [84, 100]) or the inability to further expand tidal volume (VT) (identified by an end inspiratory lung volume greater than 90% of total lung capacity [58, 88–90, 94, 102, 103], decrease of inspiratory reserve volume below 0.5–1.0 L [57, 59, 90, 94, 103] or a ratio of tidal breathing (VT)/IC of more than 60%, 70% or 85% [81, 88, 90, 94, 102, 103]). The assessment of dynamic respiratory mechanics to identify ventilatory limitation can be supported by the findings of a study by Chin et al. [63] (table 3). By adding dead space, they found that a cohort of patients with COPD and healthy controls showed evidence of having reached ventilatory limits during CPET, even in the presence of considerable BR. However, they observed significant mechanical constraints at peak exercise in the studied individuals, suggesting that these mechanical factors limited further increases in ventilation. Three included studies that compared breathing reserve against different markers of dynamic respiratory mechanics to identify ventilatory limitation to exercise found large discordance between the two methods in the studied populations (table 2) [57–59].
Extreme tachypnoea, with respiratory rate above 50 or 55 breaths·min–1, was less frequently used as criterium to identify ventilatory limitation [60, 81, 88, 101, 102].
Identifying pulmonary gas exchange limitation
All eight publications that outlined criteria for pulmonary gas exchange limitation based its identification on a decrease in oxygen saturation and/or high ventilatory equivalent for carbon dioxide (VʹE/VʹCO2) slope. The most often mentioned thresholds were ≥34 for VʹE/VʹCO2 slope [65, 69, 72] and an absolute drop below 90% [53] or 88% [88, 102] or a relative drop ≥5% from baseline [45, 88, 102] for oxygen saturation. Three publications also added criteria based on invasive measurements, such as a high dead space to tidal volume ratio (VD/VT), an increased alveolar–arterial gradient (PA–aO2), a decreased partial pressure of oxygen and an increased partial pressure of carbon dioxide [86, 88, 102]. Eight additional publications used the terms “gas exchange abnormality” and “gas exchange impairment” as cause of exercise limitation instead of the term “limitation”, but used the same parameters with (almost) identical thresholds for its identification [19, 20, 41, 48, 70, 73, 79, 82].
Various publications outlined criteria to identify gas exchange abnormalities without explicitly attributing them as a cause of exercise limitation. The 2003 ATS/ACCP statement [78] stated that PA–aO2 of >35 mmHg indicates possible gas exchange abnormality. It also calls a fall in arterial oxygen saturation ≥4%, below ≤88% or alveolar oxygen tension ≤55 mmHg significant. Similar criteria, based on a drop in oxygen saturation and/or an increase in PA–aO2, were found in a review by Stickland et al. [94] and the 2022 HFA/EAPC-ESC statement [100].
Identifying other limitations at the level of the respiratory system
15 publications described pulmonary limitation to exercise. This term was sometimes used as a synonym for ventilatory limitation [35, 54, 81, 83, 87], mainly identified by a low BR. However, more often, it was used as an umbrella term for a ventilatory and gas exchange limitation, identified by a low BR and/or oxygen desaturation [23, 25, 60, 71, 84, 92, 93, 97, 99]. Three publications even specified ventilatory limitation as a separate criterium within the criteria for pulmonary limitation [25, 84, 99]. The same ambiguity was seen for the term respiratory limitation. Four out of eight publications used it as an umbrella term for ventilatory and gas exchange limitation [28, 32, 70, 86], while the other half used it as synonym for ventilatory limitation [26, 40, 48, 49].
Criteria to identify pulmonary vascular limitation were given by six publications, of which three mentioned an elevated VʹE/VʹCO2 slope for its recognition [60, 92, 100]. Other used identification criteria were a drop in oxygen saturation [92, 100] and a reduction partial pressure of end-tidal CO2 [60, 92]. Laveneziana et al. [88, 102] used the same criteria for cardiocirculatory limitation and pulmonary vascular limitation in two included publications. One study by Markowitz et al. [62] supported the use of the combination of reaching <58% of the predicted VʹO2peak, VʹO2 at AT <38% of predicted VʹO2peak, BR >8 L·min–1 and VʹE/VʹCO2 at AT >34 had an accuracy of 85% to identify pulmonary vascular limitation, identified by invasive haemodynamic measurements. This algorithm had a sensitivity of 79% and a specificity of 88%.
Identifying limitations at the level of the peripheral muscle
The terms peripheral limitation, muscle limitation and peripheral muscle limitation were used as synonyms in literature. Six publications referred to a decreased AT [8, 28, 32, 78, 86, 88] in cases of peripheral muscle limitation. Other identification criteria were high fatigue Borg score ≥7/10 (n=2) [28, 32], a low RER (n=2) [54, 78] and blood lactate anomalies (n=2) [54, 88]. Most often, however, peripheral muscle limitation was described after exclusion of cardiopulmonary limitations, especially in the presence of a decreased VʹO2peak [42, 45, 54, 71, 78, 86].
Identifying other limitations to exercise
All but two publications that mentioned deconditioning as limitation to exercise included low VʹO2peak as a compulsory criterium, using <80% [19, 37, 46], <83% [25] and <85% [60, 69] of predicted VʹO2 as threshold value. A low oxygen pulse was considered a sign of deconditioning according to four publications [25, 60, 91, 92], while two publications indicated a normal oxygen pulse in deconditioned individuals [19, 96]. Another contradiction was seen regarding peak heart rate, with some publications describing reaching predicted peak heart rate in deconditioning, while Glaab et al. [92] mentioned the opposite. Multiple publications required the absence of pulmonary or cardiovascular limitations to identify deconditioning [37, 60, 69, 72, 91].
Dysfunctional breathing was identified as high variability in tidal volume and/or BR, with Ingul et al. [46] adding reaching VʹO2 <80% pred and Frésard et al. [70] adding the absence of limitations in the respiratory and circulatory system.
Sources of used identification criteria
The 2003 ATS/ACCP statement [78] and the Wasserman textbook [101], both also included in this scoping review, were by far the most commonly referred to sources of the used identification criteria. They were cited 19 and 14 times, respectively. However, two included studies referred to the interpretation algorithm of the Wasserman textbook [101] to identify circulatory limitation, but this algorithm leads to clinical entities rather than limitations to exercise [37, 46]. 25 publications (including eight conference abstracts) did not mention the source of the identification criteria they employed. Five studies used arbitrary identification criteria. The whole citation network diagram of included publications is provided in figure 3.
Citation network diagram illustrating the sources of the used identification criteria for limitation(s) to exercise of included publications. Blue dots represent publications included in this scoping review, red dots are publications that were referred to but were not included, because they did not meet the selection criteria or were not retrieved. The size of the dot is relative to the amount of times the publication was referenced to. Publications that did not reference any source for the identification criteria they used nor were referenced to by other publications were left out this diagram.
Discussion
The aim of this scoping review was to explore which limitations to exercise are differentiated using conventional CPET and which criteria are used to identify them. In general, we found large heterogeneity in the terminology, the classification and the identification criteria of limitations to exercise that are distinguished by CPET.
This variability was also observed in the criteria employed to detect a maximal CPET. While a plateau in VʹO2 is often considered the primary indicator of attainment of an individual's maximal VʹO2, it is not the most commonly used maximality criterium. This could possibly be due to the fact that this VʹO2 plateau is not often attained despite good effort or not well interpreted by physicians [41, 88]. Other frequently used maximality criteria such as a high RER, reaching a certain heart rate, high Borg score and ventilatory limitation all exhibit a lack of consensus on an optimal cutoff.
In current literature, two approaches to identify limitations to exercise can be distinguished. The first approach reserves the use of the term “limitation” only for individuals with a diminished exercise capacity, as people with a normal exercise capacity are not considered “limited”. This way, the term only refers to a “pathological limitation”. However, this approach risks categorising a CPET as normal solely based on a normal exercise capacity, potentially resulting in not detecting abnormal exercise responses [13]. The second approach uses the term “limitation” regardless of exercise capacity to refer to an individual's inherent exercise limiting factor(s), even if it is physiological in nature. Therefore, according to this view, a physiological cardiovascular limitation, the default finding in most healthy subjects, should be distinguished from a pathological one. A physiological cardiovascular limitation is identified by approaching or achieving maximal heart rate in the absence of signs indicating an abnormal cardiovascular response. The main caveat here is the inability of the various equations used to estimate maximal heart rate to accurately predict maximal heart rate on an individual level, due to large inter-individual variation [113, 114].
Findings specific for a pathological cardiovascular limitation are angina, ECG changes, the development of cardiac arrhythmias, the presence of chronotropic incompetence or an abnormal blood pressure response, although exact definitions vary or are lacking. A low or early flattening oxygen pulse trajectory, a low VʹO2/work rate relationship and early AT are commonly used criteria for a cardiovascular limitation, but they are not specific. The oxygen pulse is used as a surrogate for changes in stroke volume, but also depends on the arteriovenous oxygen difference and thus can be reduced in case of low peripheral oxygen extraction too [115]. This also holds true for the VʹO2/work rate relationship. Finally, a decreased AT reflects an early switch to anaerobic metabolism. This could be due to impaired cardiovascular function, but also due to impaired oxygen uptake or utilisation by the muscle and even early oxygen desaturation.
On the level of the respiratory system, ventilation and gas exchange are often differentiated as causes of exercise limitation. Ventilatory limitation is most commonly identified by a BR of <15%, with MVV calculated as FEV1×40. The use of calculated MVV based on FEV1 is generally favoured over directly measured MVV. This is because the breathing strategy adopted when measuring MVV differs vastly from the breathing pattern observed during exercise and cannot be sustained for more than 15–20 s, making its relevance for exercise longer than this time questionable [78]. However, using calculated MVV also has shortcomings, as it is not appropriate in people with inspiratory obstruction or severe neuromuscular disease, because of the risk of overestimating true breathing capacity [101]. Reliance on only BR to identify ventilatory limitation, however, has been criticised, because BR only reflects an imbalance between ventilatory capacity and ventilatory demand and provides little information about the specific factors limiting the ventilatory response [78, 90]. Therefore, the use of serial IC manoeuvres during exercise to assess dynamic respiratory mechanics has been encouraged, as it allows a comprehensive evaluation of ventilatory abnormalities during exercise and their contribution to exercise limitation in the individual person [90, 110]. Various studies found that the measurement of dynamic respiratory mechanics is complementary to the use of BR to identify ventilatory limitation, but no consensus exists on which markers of dynamic respiratory mechanics should be used [57–59]. Limitations of the measurement of dynamic respiratory mechanics are the inability to account for the thoracic gas compression artifact or exercise induced bronchodilation/bronchoconstriction when comparing exercise tidal and maximal FVLs and the risk of incorrect alignment of the tidal breathing curve with the maximal FVL. Serial IC manoeuvres also require good cooperation and some concern exists about their potential impact on other CPET measurements [59, 84, 90].
Although pulmonary gas exchange limitation is a frequently encountered term, many publications refer to “gas exchange abnormality” or “gas exchange impairment”, with or without explicitly categorising it as cause of exercise limitation. This observed hesitancy towards the use of the term gas exchange limitation could be attributed to the uncertainty if gas exchange really is an independent limitation to exercise or merely a contributor to higher ventilatory requirement. It has been shown that in patients with COPD with gas exchange abnormalities arterial blood gas homeostasis is adequately maintained during exercise as a result of compensatory increases in minute ventilation, but at the expense of earlier occurrence of dynamic respiratory mechanical constraints and greater respiratory discomfort [116]. Two noninvasive CPET variables that are often used as surrogates for gas exchange are oxygen saturation, measured by pulse oximeter, and VʹE/VʹCO2 slope, but both have their shortcomings. While oxygen desaturation can be a sign of impaired gas exchange, it can also be a result of an inadequate ventilatory response during exercise [84]. Furthermore, pulse oximeters are good to monitor trends but not reliable to determine the absolute magnitude of change [78]. The 95% confidence limit of oxygen saturation measured by pulse oximetry values relative to direct measurement of arterial oxygen saturation has been shown to be approximately 4–5% [82]. The presence of a high VʹE/VʹCO2 slope, on the other hand, does not necessarily prove the presence of gas exchange abnormalities either, because its value can be influenced by other factors causing a lower arterial carbon dioxide tension set-point, such as hyperventilation [117]. Additionally, the relationship between VʹE/VʹCO2– and ventilation perfusion mismatch becomes less reliable after the respiratory compensation point, when glycolytic pathways become the dominant driver of ventilation instead of CO2 production, rendering the VʹE/VʹCO2 slope not useful for the evaluation of gas exchange abnormalities at peak exercise. While both oxygen desaturation and the VʹE/VʹCO2 relationship have their shortcomings in evaluation pulmonary gas exchange, clustering both variables may be useful to estimate the presence of gas exchange abnormalities in a noninvasive way. However, only blood gas assessment with calculation of PA–aO2 and VD/VT can confirm and quantify gas exchange abnormalities [101].
The terms “pulmonary limitation” and “respiratory limitation” are sometimes used either as a synonym for ventilatory limitation or as an umbrella term combining criteria for ventilatory and gas exchange limitation, but their use seems of no added value and could cause confusion. Another frequently described limitation to exercise within the respiratory system is pulmonary vascular limitation, referring to the pulmonary vessels as locus of exercise limitation. Abnormalities in the pulmonary vasculature can influence both gas exchange and circulation. The use of this term can be potentially confusing given this overlap with other limitations to exercise, which is also reflected in the identification criteria that are used for its detection.
Although a limitation in the peripheral muscle is frequently described in literature, no single parameter measured during conventional CPET specifically captures the exercise response at this level. The inherent limitation of conventional CPET to discriminate between an impaired oxygen delivery by the cardiovascular system and impaired oxygen uptake or utilisation at the level of the muscle (in the absence of specific pathological cardiovascular findings) is often overlooked in current practice. In line with this, deconditioning represents a broad adaptation within the body impacting various organ systems, including a lower cardiac stroke volume and impaired peripheral muscle function, rather than constituting a distinct physiological limitation [16]. Despite its frequent mention as an exercise limiting factor, it is important to note that no specific CPET parameter exclusively characterises the presence of deconditioning.
The heterogeneity we found in CPET interpretation can be attributed to the lack of its standardisation. In 2019, an important step was taken by the publication of the ERS statement on standardisation of cardiopulmonary exercise testing in chronic lung diseases [86], but its main focus lay on the standardisation of CPET protocols and not its interpretation. Currently, no gold standard method exists for interpreting CPET. An integrative, “patterns-based” approach has been advocated, in which exercise response patterns, rather than clear cut-off values, are combined with clinical information for the interpretation of CPET [78]. This approach, however, is very interpreter-dependent and makes CPET interpretation arbitrary and prone to confirmation bias.
To provide a framework to ensure objective and consistent CPET interpretation, standardisation is necessary. This standardisation should include the CPET maximality criteria as well as a classification of physiological limitations to exercise and criteria to identify them. Hereby, nevertheless, it is essential to acknowledge the earlier described, inherent discriminative limitations of (noninvasive) CPET, to avoid overinterpretation of CPET findings and oversimplification of exercise limitation. This vision is in analogy to the latest 2021 ATS/ERS technical standard for the interpretation of pulmonary function tests (PFTs) [118], which says that the main emphasis when using PFTs should be on classifying physiology and not making a clinical diagnosis per se. The standardisation of CPET interpretation would enhance clinical applicability of CPET, facilitate comparability of CPET results in different settings and help with education, which are all not feasible at the moment. Furthermore, it would facilitate the development of evidence-based guidelines for clinical decision-making and tailoring of therapy. It also does not contradict an integrative approach to CPET interpretation. After objectively identifying exercise limitations by standardised criteria, which should be done with flexibility, this information can be integrated with other clinical data to help with clinical diagnosis and decision-making.
Currently, little evidence exists to support the use of specific thresholds to identify limitations to exercise and most recommendations on CPET interpretation are based on limited evidence [119]. Further research is needed to validate which criteria determine the attainment of physiological limits by CPET, by correlation of findings to more invasive measurements and imaging techniques [43]. Nevertheless, it appears unlikely that ideal criteria to identify all limitations to exercise will be deduced for everyone based on experimental research, given that the large variability within the population makes extrapolation difficult. An alternative and more pragmatic approach is a statistical-driven approach where measured values are compared to a healthy population, which is also the approach used for the interpretation of PFTs [118]. However, there is lack of good reference values for many CPET variables and the use of different reference equations can considerably affect the interpretation of CPET results [119]. A recent systematic review by Takken et al. [120] found that although the number of studies reporting reference values for CPET variables is rising, all studies showed methodological flaws. In accordance with the interpretation of PFTs, they encouraged the use of Z-scores, instead of expressing obtained values as percentage of the predicted value. This way, the influence of variability in the reference population can be accounted for. Attempting to solve the lack of good and unified reference values for CPET, the ERS founded a Global Lung Initiative task force to develop reference equations for cardiopulmonary exercise testing, including Z-scores. These reference equations should also take into account the exercise modality, given the differences between CPET performed on a cycle ergometer versus a treadmill. It is recognised that CPET performed on a treadmill yields higher VʹO2peak values and can lead to greater oxygen desaturation and increased breathlessness scores compared to cycle ergometry [86]. Consensus will be necessary to establish a lower and upper limit of normal.
To our knowledge, this is the first review to summarise the classification and identification of limitations to exercise by incremental CPET. Although we used systematic methodological and reporting techniques, this paper holds some limitations inherent to scoping reviews. While we utilised well-defined eligibility criteria and adhered to the latest guidelines during the study selection process, the potential for selection bias needs to be acknowledged. This stems from the subtle boundary between the use of descriptive patterns and precise criteria to identify limitations to exercise, and the fact that identification criteria were sometimes rather implied than explicitly reported. Many publications also used different terms interchangeably or even changed identification criteria throughout the text. Furthermore, reported findings are only based on quantification of literature as quality appraisal of the included publications was not performed.
Conclusion
There is large heterogeneity in the terminology, the classification and the identification criteria of limitations to exercise that are distinguished using incremental CPET. Without denying the complexity of exercise limitation, there is a need for standardisation of CPET interpretation to provide a framework for objective and consistent CPET interpretation. This way, the true potential of this highly valuable test, which is currently underutilised in clinical practice, can be unlocked. Hereby, however, the inherent discriminative limitations of CPET need to be acknowledged.
Points for clinical practice
Given the large heterogeneity in the interpretive strategies of CPET to identify limitations to exercise, it is important to mention which criteria were used for the interpretation of CPET results.
While CPET provides valuable information on the mechanisms of exercise limitation, caution should be exerted not to overinterpret CPET findings, given the inherent discriminative limitations of the test.
Questions for future research
Validation of CPET criteria correlating best with the attainment of physiological limits is needed.
There is a need for better reference values for many CPET variables, including Z-scores, with the establishment of lower and upper limits of normal.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary table S1 ERR-0010-2024.SUPPLEMENT
Supplementary table S2 ERR-0010-2024.SUPPLEMENT2
Supplementary table S3 ERR-0010-2024.SUPPLEMENT3
Acknowledgements
The authors wish to thank Thomas Vandendriessche, Chayenne Van Meel, Norin Hamouda and Krizia Tuand, the biomedical reference librarians of the KU Leuven Libraries – 2Bergen – learning Centre Désiré Collen (Leuven, Belgium), for their help in conducting the systematic literature search.
Footnotes
Provenance: Submitted article, peer reviewed.
Conflict of interest: I. Gyselinck reports grants from Research Foundation Flanders, and support for attending meetings from AstraZeneca. W. Janssens reports grants from AstraZeneca and Chiesi, consultation fees from AstraZeneca, Chiesi, GSK and Sanofi, payment or honoraria for lectures, presentations, manuscript writing or educational events from AstraZeneca, Chiesi and GSK, support for attending meetings from AstraZeneca and Chiesi, and the following financial (or non-financial) interests: co-founder and chairholder of ARTIQ, a spin-off company of KULeuven. M. Staes, K. Goetschalckx and T. Troosters have nothing to disclose.
- Received January 20, 2024.
- Accepted May 28, 2024.
- Copyright ©The authors 2024
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
