Beyond Pretto and Towards Improved CPET in Clinical Practice

The cardiopulmonary exercise test (CPET) is used to evaluate exercise intolerance and associated symptoms (breathlessness, fatigue), as well as operative risk, disease trajectory, and treatment response. CPET is a sophisticated test that is challenging to implement and interpret. In practice, the test is initiated at a low intensity, which is progressively increased until the patient reaches their exercise limit. A challenge to the clinician or scientist is to set an incremental work rate (Ẇinc) that enables patients to achieve this limit within the recommended 8–12 min [1-3]. A persistent problem is the short test (< 8 min), which yields insufficient data that causes uncertainty about test outcomes [2], requires repeated testing, and increases the burden to the patient and the healthcare system. There is no consensus regarding an effective method of predicting Ẇinc that consistently yields test times of the required duration in healthy individuals. This problem is heightened when testing patients with diverse, complex pathologies. CPET is performed using a treadmill or cycle ergometer. According to our recent (unpublished) survey of 55 sites in Australia and New Zealand, CPET is conducted mainly in respiratory departments, cycling is the predominant mode of testing, and 13 different methods of determining Ẇinc are used. This diversity of Ẇinc methods suggests a lack of methodological consistency with potential implications for the effectiveness of CPET when used in clinical settings, but it is surprising given the long-standing presence of an Australian-based method—the 'Pretto equation'—that accurately estimates Ẇinc in ~80% of respiratory cases [4]. There are several types of Ẇinc method, including use of a fixed incremental work rate. Use of a fixed Ẇinc is characteristic of rigid protocols that fail to account for differences in patient fitness or disease pathology, resulting in highly variable test durations. In testing respiratory patients, Pretto et al. showed that use of a fixed Ẇinc (15 W min−1) resulted in just 28% of tests lasting 8–12 min, with most tests (59%) lasting less than 8 min [4]. Thus, continued use of a fixed Ẇinc is likely to yield too many short tests and increase diagnostic uncertainty. An alternative to a fixed Ẇinc is use of a formula to predict peak power output (Ẇpeak) from which Ẇinc can be calculated. Since the first edition of their pre-eminent textbook, Wasserman and colleagues have advocated use of a formula to estimate Ẇpeak but which they adjust for patients with low FEV1 [2, 3]. This formula, designed for cycle ergometry, is based on the relationship between V̇O2 and power output and estimates of V̇O2 predicted from non-exercise measurements (Figure 1A). Unfortunately, evidence of its efficacy is lacking. To address this, we analysed their case studies [2] which correspond to common cardiopulmonary pathologies referred to our CPET centre. For patients considered 'normal' by the authors, CPET times were generally within the recommended timeframe (mean ± SD: 10.6 ± 2.8 min, N = 14) but significantly lower for patients with 'common pulmonary diseases' (6.8 ± 1.9 min, N = 16), 'common cardiovascular disorders' (7.7 ± 1.5 min, N = 10) and 'primary or secondary pulmonary vascular diseases' (6.56 ± 3.30 min, N = 9). This analysis suggests that the 'Wasserman formula' tends to overestimate Ẇinc and result in shorter duration tests for patients with common cardiopulmonary pathologies. Similar formulae for cycle ergometry have been proposed by others, along with supporting evidence. A study of 108 healthy men showed close agreement between predicted and measured Ẇpeak (< 1% difference, limits of agreement = −16% to +17%) [8]. A smaller study (Figure 1B) involving healthy men and women also showed close agreement between predicted and measured Ẇpeak but with greater precision (LoA = −10 to +12%) [6]. These formulae included 'non-exercise' predictive equations for V̇O2peak developed from earlier studies of healthy individuals [7, 9-15] and shown to have good predictive accuracy (multiple R2 = 0.66–0.77). Whereas these studies were limited to healthy individuals, Pretto et al. [4] studied patients with respiratory diseases and proposed an empirical equation to predict Ẇpeak based on routine clinical measurements. They analysed 180 consecutive tests performed at a fixed Ẇinc and showed that four measurements (FEV1, TLCO, age, sex) accounted for 72% of the variance in Ẇpeak. In subsequent testing (N = 231), an empirical equation based on these measurements was used to predict Ẇpeak and calculate Ẇinc for a 10-min test. Compared with tests at a fixed Ẇinc, this empirical approach yielded a higher percentage of tests (62% vs. 28%) that lasted 8–12 min, although its effectiveness was limited by the necessity to round Ẇinc estimates to the nearest 5 W min−1. There is clearly room for improvement. Determining an appropriate Ẇinc for clinical use remains a critical yet unresolved challenge in CPET. While fixed and formula-based methods are widely used, they often fail to deliver tests of optimal duration, reducing the diagnostic utility of CPET. Fixed Ẇinc methods are not appropriate for CPET, whereas there is a need to improve formulaic approaches, establish their validity for females, and assess their predictive accuracy in patients. The 'Pretto equation' [4] is specific to the testing of patients with respiratory diseases, but its relevance to patients with other diseases is unclear. This equation requires a measurement of FEV1 that is often lacking in patients referred from non-respiratory specialists. However, spirometry measurements can be obtained in conjunction with CPET, and their incorporation into formulae might improve prediction of Ẇpeak beyond what current 'non-exercise' formulae achieve. Finally, little consideration has been given to the potential of measurements obtained during brief and mild exercise used to familiarise patients to CPET. New research and innovative approaches to the 'Ẇinc problem' are needed to improve the precision, consistency, and clinical value of CPET. We thank Evan McAuley for assistance with the analysis of the Wasserman case series. Open access publishing facilitated by Western Sydney University, as part of the Wiley - Western Sydney University agreement via the Council of Australian University Librarians. The authors declare no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.