The original and two new derivative versions of the COMPERA 2.0 risk assessment model: useful tools for guiding balloon pulmonary angioplasty

BACKGROUND: The COMPERA 2.0 4-stratum (4-S) risk score has been demonstrated superior over the 3-stratum (3-S) one in patients with pulmonary arterial hypertension and medically managed patients with chronic thromboembolic pulmonary hypertension (CTEPH). We aimed to determine the prognostic value of the original 4-S and 3-S COMPERA 2.0 risk score and two new derivative versions in CTEPH patients who underwent balloon pulmonary angioplasty (BPA). METHODS: We retrospectively enrolled 175 BPA-treated patients with CTEPH. We assessed the risk stratification before and after each BPA session of CTEPH patients by the original 4-S and 3-S COMPERA 2.0 risk score (by rounding decimal to the nearest integer) and two new proposed derivative versions: the modified version (by rounding decimal to the next integer) and a hybrid version that fuses the original and modified versions. The primary endpoint was clinical worsening events. The secondary outcomes were achieving low-risk profile and mean pulmonary arterial pressure (mPAP) < 30 mmHg at follow-up. We used the Kaplan–Meier curve analysis to assess the survival differences between stratified patients. The comparative model’s performance was evaluated in terms of discrimination by Harrell’s C-index. RESULTS: All versions of COMPERA 2.0 4-S model outperformed the 3-S one in discriminating the differences in echocardiographic and hemodynamic parameters and clinical worsening-free survival rates. The original and hybrid 4-S model could independently predict the primary and secondary endpoints, and the hybrid version seemed to perform better. The first BPA session could significantly improve risk profiles, and these changes were associated with the likelihood of experiencing clinical worsening events, achieving a low-risk profile and mPAP < 30 mmHg at follow-up. The number of BPA sessions required to achieve low risk/mPAP < 30 mmHg increased as the baseline risk score escalated. CONCLUSIONS: The COMPERA 2.0 4-S model outperformed the 3-S one in BPA-treated patients with CTEPH. The 4-S model, especially its hybrid version, could be used to predict clinical outcome before the initiation of BPA and monitor treatment response. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12931-022-02232-1.