Cardiopulmonary Exercise Testing for Prognostication in Advanced Cystic Fibrosis Lung Disease and Beyond

Cardiopulmonary exercise testing (CPET) has not only been established as a useful diagnostic tool in the evaluation of dyspnea and exercise limitation but has shown utility in prognostication in multiple cardiopulmonary disorders (1–4). The use of CPET for prognostication in cystic fibrosis (CF)–related lung disease has been evaluated among the overall CF population, including people with all degrees of lung function (5). Despite drastically improved clinical outcomes among people with CF (PwCF) after the approval of CFTR (CF transmembrane conductance regulator) modulator therapies, PwCF continue to have structural lung disease, with 6.5% of participants in the Cystic Fibrosis Foundation Patient Registry living with advanced CF lung disease (ACFLD) (6). Approximately 10% of the overall CF population in the United States is ineligible for CFTR modulator therapies because of their mutation and/or age, with a higher percentage ineligible among those with ACFLD (6). Although established clinical markers such as forced expiratory volume in 1 second (FEV₁), 6-minute-walk distance, hypoxemia, hypercarbia, and pulmonary hypertension are used in consideration of lung transplantation (LTx) referral, CPET has not previously been evaluated as a prognostic tool among patients with ACFLD, specifically those with FEV₁ of 40% or lower (7). This is in contrast to systolic heart failure and heart transplantation, for example, for which the prognostic value of CPET was established, and CPET is now routinely used in pre–heart transplantation evaluations (3, 8). In the context of an evolving landscape in CF care, unknown trajectory of lung disease in those using CFTR modulator therapy, and a fraction of the population who will continue to have ACFLD and require LTx, additional prognostic tools are necessary to estimate mortality risk.

In this issue of AnnalsATS, Radtke and colleagues (pp. 411–420) present a large, international, retrospective study of CPET for people with ACFLD (9). The authors retrospectively analyzed data from 2008 to 2017 across 20 CF centers in Asia, Australia, Europe, and North America. Data were collected on 174 PwCF aged 10 years and older with FEV₁ ⩽ 40% predicted. Follow-up data from subjects were available for 2 years after CPET, with an endpoint of death or LTx within 2 years. Using mixed Cox proportional-hazards regression adjusting for clinical factors such as age, sex, and FEV₁, the investigators found that for every increase of 10% predicted peak oxygen uptake ($\dot{\text{V}}$o_2peak) or peak work rate (W_peak), risk of death or LTx within 2 years was reduced by 40%. The authors used tree structured regression models including 12 prognostic factors for survival and found that W_peak was most strongly associated with 2-year risk of death or LTx. They used receiver operating characteristic curves to provide possible cutoffs for $\dot{\text{V}}$o_2peak and W_peak values, 40% and 50% predicted, respectively. Compared with prior studies, the investigators used a random intercept for study center to at least partially account for differences in protocols among centers.

This is the first international study of CPET as a prognostication tool among specifically those with ACFLD. The study includes a range of FEV₁ values among people with ACFLD across a broad span of age groups. Importantly, these findings can be applied to an important portion of the CF population: the more than 10% of PwCF who remain ineligible for CFTR modulator therapies in the United States and in whom the timing of LTx remains an important consideration (6). This population also includes those who are not able to tolerate CFTR modulator therapy because of side effects and PwCF in low- and middle-income countries that do not have access to modulator therapy (10, 11). Although receiver operating characteristic cutoff points for $\dot{\text{V}}$o_2peak and W_peak values require prospective validation, these findings could suggest a clinically applicable threshold for transplantation referral and/or listing considerations. W_peak is measured using a cycle ergometer and does not require the use of a face mask to measure gas exchange. As the authors highlight, this is time and cost efficient, may be performed with people using supplemental oxygen, and reduces the need for specialized equipment or expertise in using complex systems. Using W_peak is therefore more plausible in resource- or staff-limited settings, and in a population of patients who are more likely to use supplemental oxygen and therefore have more difficulty using face masks.

As acknowledged by the authors, there are a few limitations affecting the interpretation of this study. First, the period of interest, between 2008 and 2017, saw rapidly advancing CF care, and data were collected before the broad use of highly effective modulator therapies (HEMTs). Few patients from this cohort were on CFTR modulators, an important variable in the consideration of survival in CF, even among those with ACFLD.

Evolving evidence shows that death and LTx are less common among PwCF now being treated with HEMTs, making the present study’s prognostication less applicable in this population specifically (12–14). There are important considerations regarding the decision and indication to perform CPET in this retrospective study. Although the random effect of CF center was considered, indications for CPET varied among CF centers, and the centers included in this study were more likely to perform CPET than CF centers more broadly. Although this study included supplemental oxygen use as a prognostic factor in modeling, CPET was not performed with patients using supplemental oxygen, therefore possibly excluding those individuals dependent on continuous oxygen during exercise. Importantly, oxygen use is a consistent predictor of death and has been incorporated as an indication for transplantation referral (15, 16). In prior survival analyses, approximately 35% of patients with FEV₁ less than 30% used oxygen continuously or at night compared with the less than 9% of patients in the present study (15). Finally, only a small proportion of patients (approximately 18%) had 6-minute-walk test data available, limiting the ability to directly compare the prognostic value of CPET with that of 6-minute-walk distance in this study.

CPET has been evaluated in mortality risk assessment in small studies in idiopathic pulmonary fibrosis and pulmonary arterial hypertension, as well as comparison of those with interstitial lung disease with and without pulmonary hypertension (2, 4, 17). Similarly, cutoff points for $\dot{\text{V}}$o_2peak and W_peak were shown to correspond to mortality in a small study of 34 subjects with idiopathic pulmonary fibrosis (2). Given similar findings of the prognostic value of CPET in other lung diseases, there may be a role for the broader study and use of CPET among LTx candidates with all diagnoses. Radtke and colleagues (9) demonstrate that CPET is valuable in estimating the risk of mortality or transplantation in people with ACFLD without HEMTs. This study highlights the need for prospective validation of CPET as a tool in the evaluation of persons with ACFLD. Use of CPET should be considered in the risk stratification algorithm of individuals with ACFLD and potentially for a more general population of individuals with advanced lung disease. Finally, it will be important to understand whether interventions such as pulmonary rehabilitation and exercise programs can improve CPET performance and clinical outcomes for people with ACFLD.