Cardiopulmonary exercise testing and pulmonary function testing for predicting the severity of CTEPH

The study protocol was reviewed and approved by the Ethics Committee of Shanghai Pulmonary Hospital. Written informed consent was obtained from each patient for inclusion into the study prior to the performance of any study-related procedures.

All methods including CPET, PFT, RHC and blood test were carried out in accordance with relevant guidelines and regulations.

This study retrospectively enrolled 86 inpatients with CTEPH who were referred to Shanghai Pulmonary Hospital from November 2015 to December 2019. All patients were definitely diagnosed by RHC. Patients with mean pulmonary artery pressure (mPAP) ≥ 25 (mm Hg) and pulmonary arterial wedge pressure (PAWP) ≤ 15 (mm Hg) were considered to be diagnosed with CTEPH. They also should had thromboembolic disease performance which can be detected by ventilation/perfusion scan or pulmonary angiogram. Patients were excluded from study if they had any evidence of the following: right-to-left cardiac shunt, coexisting lung diseases (identified clinically or on CT scan), FEV1/FVC% < 65%, history of treatment with BPA and PEA. Enrolled patients’ data including demographics, medication, NT-pro BNP, hemodynamics, PFT and CPET were collected. Ethical approval by the medical ethics committee of Shanghai Pulmonary Hospital was obtained.

CPET was performed on an electromagnetically braked cycle ergometer (Master Screen CPX, Jaeger crop, Hoechberg, Germany) to record gas exchange data over 10-s intervals by using a breath-by-breath system. The protocol was consisted of the rest phase of 3 min, the unloaded phase of 3 min, the incremental phase, and the recovery phase of 5 min. The patients were instructed to pedal at 55–60 revolutions/min in the unloaded and the incremental phase, and once they reached their limit, entered the recovery phase. Patients could quit at any time if they developed fatigue, dyspnea, chest tightness or any other discomfort during the process. There were three ramp increments models that we used in the incremental phase: 10 W/min, 15 W/min and 20 W/min. We would choose an appropriate model according to the patient's clinical condition and PFT result. Each subject’s exercise time includes the unloaded phase of 3 min and the incremental phase. Some basic information in our research is as follows. In “Mild” group, 8 subjects used 15 W/min ramp increments model, and 2 subjects used 20 W/min ramp increments model. In “Moderate” group, all 10 subjects used 15 W/min ramp increments model. In “Severe” group, 1 subject used 10 W/min ramp increments model, 29 subjects used 15 W/min ramp increments model and 5 subjects used 20 W/min ramp increments model. Each group’s exercise time is as follows: “Mild” group (8.6 ± 1.4 min), “Moderate” group (7.6 ± 1.2 min) and “Severe” group (6.9 ± 1.3 min).

Measurements including load, minute ventilation (VE), carbon dioxide output (VCO₂), oxygen uptake (VO₂), oxygen pulse (VO₂/HR), end-tidal partial pressure for carbon dioxide (PETCO₂), end-tidal partial pressure for oxygen (PETO₂), heart rate (HR), breathing reserve (BR), respiratory exchange ratio (RER) and breathing frequency (BF) were recorded and calculated. Anaerobic threshold (AT) which represents the beginning of anaerobic metabolism was determined by the V-slope method and was independently defined by two experienced investigators who have been engaged in clinical and scientific research on CPET for several years. VE/VCO₂ slope was obtained by linear regression analysis of the relation between VE and VCO₂. Oxygen uptake efficiency slope (OUES) was computed by linear square regression from the oxygen uptake on the logarithm of the minute ventilation according to the following equation: VO₂ = a*lgVE + b. Constant “a” is called the OUES. Oxygen uptake efficiency plateau (OUEP) was at 90 s of the highest consecutive values for VO₂ (mL/min)/VE (L/min).

Spirometry and body plethysmography were performed on each patient using standard equipment (Masterscreen-PFT, Jaeger crop, Hoechberg, Germany; Masterscreen-plethysmography, Jaeger crop, Hoechberg, Germany). Forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), residual volume (RV), total lung capacity (TLC) and diffusing capacity for carbon monoxide (DLCO) were determined by standard procedures [12, 13]. For each patient, data were presented in absolute terms and normalized to percentage of normal predicted (% Pred). All measurements were calculated using accepted equations for Chinese adults [14].

55 patients with CTEPH were involved in the test set. They were divided into “Mild”, “Moderate” and “Severe” group according to PAP (mm Hg). Mild: 35 > PAP (mm Hg) ≥ 25; Moderate: 45 > PAP (mm Hg) ≥ 35; Severe: PAP (mm Hg) ≥ 45.

The characteristics of all groups were summarized in Table 1. The “Severe” group had the highest value of NT-proBNP (pg/mL) (1082 (642.0, 2674)) when compared with the “Mild” (143.0 (63.8, 286.7)) and “Moderate” (648.5 (266.3, 2049) group. The values of PAP (mm Hg) of the “Mild”, “Moderate” and “Severe” groups were statistically different (28.6 ± 3.3 vs. 39.5 ± 3.9 vs. 56.4 ± 8.0 mm Hg; P < 0.001). Additionally, the values of PVR (wood u) and RAP (mm Hg) of the three groups were also statistically different. PVR (wood u) values of the “Mild”, “Moderate” and “Severe” group were listed (3.7 ± 1.3 vs. 7.6 ± 2.4 vs. 10.9 ± 3.8 wood u; P = 0.011). And RAP (mm Hg) values of the “Mild” (0.5 (0, 4.5)), “Moderate” (1.0 (0, 1.3)) and “Severe” (4.0 (2.0, 7.0)) group were also listed.

Phosphodiesterase type 5 (PDE-5) inhibitors, endothelial receptor antagonists (ERAs), prostacyclin analogs and soluble guanylate cyclase (sGC) activators were therapeutic agents used by the CTEPH patients, which included sildenafil, tadalafil, vadenafil, ambrisentan, bosentan, beraprost and iloprost. Details were listed in Table 1.

FEV1/FVC (%) was the only parameter that was statistically different between the “Severe”, “Mild” and “Moderate” group (78.8 ± 9.3 vs. 81.7 ± 8.0 vs. 73.4 ± 8.0%; P = 0.041). Details were listed in Table 1.

16 parameters were found to be statistically different among the “Mild”, “Moderate” and “Severe” group. They were listed as followed: Load @ Peak (W), VO₂ @ Rest (mL/kg/min), VO₂ @ Peak (mL/kg/min), VE @ AT (L/min), BR @ Rest (%), BR @ AT (%), VE/VCO₂ @ AT, VE/VCO₂ @ Peak, VE/VO₂ @ AT, PETCO₂ @ Rest (mm Hg), PETCO₂ @ AT (mm Hg), PETCO₂ @ Peak (mm Hg), PETO₂ @ AT (mm Hg), PETO₂ @ Peak (mm Hg), VE/VCO₂ slope and LOWEST VE/VCO₂. Details were listed in Tables 2 and 3.

55 patients with CTEPH in the test set were re-grouped into “Mild-Moderate” and “Severe” group to analyze predictors for severe CTEPH. All the CPET and PFT parameters indicated were analyzed with the univariate analysis for the severe CTEPH, and 20 parameters were found to have a P < 0.05. They were listed in Additional file 1: Table S1. Considering the sample size, 4 parameters with the minimum P value were fitted into the multivariate analysis, including VE @ AT (L/min) (OR 1.169, P = 0.004), PETCO₂ @ AT (mm Hg) (OR 0.809, P = 0.005), VO₂ @ peak (mL/kg/min) (OR 0.627, P = 0.005) and LOWEST VE/VCO₂ (OR 1.129, P = 0.006). By using multivariate logistic regression analysis, it was found that VE @ AT (L/min) (OR 1.162, P = 0.024) and VO₂ @ peak (mL/kg/min) (OR 0.633, P = 0.026) were independent predictors for the severe CTEPH. Details were listed in Table 4 and Fig. 1.

Multivariate logistic regression was used to establish a prediction model for predicting severe CTEPH: Logit(P) = Log(P/1 − P) = 1.753 + 0.168 * VE @ AT (L/min) − 0.505 * VO₂ @ peak (mL/kg/min). To evaluate the ability of the VE @ AT (L/min), VO₂ @ peak (mL/kg/min) and the prediction equation to discriminate severe CTEPH, ROC curves and calibration curves analysis were performed. Details were listed in Table 5 and Fig. 2. It should be noted that the AUC of the prediction equation was better than that for each parameter, indicating that the equation based on two parameters could improve the prediction performance for severe CTEPH.

Additionally, the optimum cut-off value of Logit(P) ≥ 0.716 to predict severe CTEPH was determined by using ROC analysis (AUC = 0.843, 95% CI = 0.732 to 0.954, Youden index = 0.586). We calculated the value of Logit(P) of each subject in the validation datasets to validate the results. Only 4 of 31 patients with CTEPH of the validation set were ambiguous, and the accuracy was 87.10%. Details were listed in Table 6.