One Small Reference Set Publication, One Giant Leap for CPET Interpretation

Aaron B. Holley, MD


February 18, 2021

For cardiopulmonary exercise testing (CPET), patients exercise on either a stationary bike or a treadmill. Cardiac, respiratory, and metabolic function are measured during warm-up and then during progressively more intensive exercise culminating with exhaustion. As I tell my patients, "if you want to pinpoint and quantify the reasons for your exercise symptoms, this is the best test that medicine has to offer." Several major organizations concur.

CPET is not without its detractors, however, and the test is far from perfect. Interpretation remains challenging for many reasons, one of which is the absence of a good reference set to establish normative values. The 2003 CHEST/ATS guidelines on CPET discuss reference equations and recommend using sets published by Jones or Hansen and colleagues. These are the reference sets programmed into the software that we use at our lab. But both were published in the 1980s and sampled small, rather homogeneous populations. In short, they're helpful but limited.

In a recent CHEST journal issue, investigators from the CanCOLD Collaborative Research Group and the Canadian Respiratory Research Network published new CPET reference equations. There's a lot to like about their data. Screening for baseline disease and abnormalities was rigorous, data collection was comprehensive, their sample size was relatively large for a CPET study (N = 173), they documented baseline activity levels, and they included a validation cohort to test their derived equations. They also compared their findings with those predicted by Jones and Hansen.

The CanCOLD Group makes a strong case for the importance of their data in the discussion section of the paper. They're right — the data are really helpful. Unfortunately, it's unlikely to silence the CPET detractors, and CPET interpretation will be a continued challenge for pulmonologists and other non–exercise physiologists.

For one, the distance between the lower and upper fifth percentiles for most values is quite wide, which means a broad range of responses will be considered normal for many parameters. This is probably a function of their small sample size (while big for a CPET study, it's still too small to establish tight reference ranges) and the physiologic variability inherent to each measure. As a result, lots of patients with exercise complaints will probably be labeled "normal" following their test. This finding is unhelpful to patients looking for an explanation and a path forward. Jones' and Hansen's equations suffer from the same problem, and this is probably why the CanCOLD investigators found that estimates using these equations predominantly fell within their predefined (very wide) equivalence margin.

Second, although several diagnostic algorithms have been proposed for CPET interpretation, they're all too reductionist to adequately describe the complex physiologic response to exercise. Any CPET reference study will suffer from the same issue. For example, reference ranges for respiratory variables will depend heavily on the patient's underlying fitness level and the patient's effort on the test. One person's mechanical respiratory limitations to exercise are another's excellent cardiopulmonary fitness and vigorous effort. Providing average and fifth percentile cut-offs is helpful and is the only way to approach reference data. Still, these cut-offs should not be considered a hard line between normal and abnormal, the way they can be for spirometry.

The CanCOLD group should be commended for their work, and it does move us forward. Their reference ranges should be used to get a feel for where individual patients fall on the spectrum of physiologic responses to exercise. Integration of symptoms, patient medical history, reason for the test, and medications with actual CPET variables remains critical. Physiologic response for any given variable must be viewed in the proper context and must factor in patient effort and baseline activity levels. (Despite the fact that baseline activity apparently fell out of their reference equation models, it still has an enormous effect on most CPET parameters.) In the end, no matter the dataset, it still takes a well-trained, competent physician to interpret a CPET.

Aaron B. Holley, MD, is an associate professor of medicine at Uniformed Services University and program director of pulmonary and critical care medicine at Walter Reed National Military Medical Center. He covers a wide range of topics in pulmonary, critical care, and sleep medicine.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.