Background/Purpose: The 2022 ESC/ERS Guidelines recommend comprehensive risk stratification at diagnosis of pulmonary arterial hypertension (PAH) to guide optimized management.¹Several risk stratification tools have been developed with data derived mainly from patients with idiopathic PAH. However, patients with systemic sclerosis (SSc)-associated PAH have a worse prognosis. We aimed to assess the performance of current risk stratification tools to predict mortality in SSc-PAH by adding variables specific to SSc.

Methods: We included SSc patients from the EUSTAR database who were diagnosed with PAH by right heart catheterization (RHC) between 2001-2021 (Project Number: CP122). PAH was defined as a mean pulmonary arterial pressure >20 mmHg, pulmonary artery wedge pressure ≤15 mmHg and pulmonary vascular resistance >2 WU. We excluded patients with previous PAH-specific treatment and meaningful interstitial lung disease (ILD), defined as an extent of ILD >20% on HRCT or FVC < 70% in patients with missing quantification. We applied four different approaches:

(I) 2022 ESC/ERS 3-strata: Three risk groups based on the mean of up to 17 risk parameters from the guidelines graded 1-3 representing low-high risk

(II) 2022 ESC/ERS 4-strata: Equals no. (I), but divides the intermediate-risk group into two groups

(III) COMPERA 2.0: Four risk groups based on the mean of WHO-functional class (FC) and/or six-minute walk distance (6MWD) and NT-proBNP graded 1-4 representing low-high risk.²

(IV) REVEAL Lite 2:Three risk groups based on six weighted variables (WHO-FC, systolic blood pressure, heart rate, 6MWD, NT-pro-BNP, and eGFR)³

We performed Cox regression adjusted for general and SSc-specific factors associated with worse outcome based on expert opinion and published literature (age, male sex, anti-centromere antibodies, digital ulcers, DLCO and limited ILD). Harrell’s C-index and ROC analysis with area under the curve (AUC) were applied to compare the performance and discriminating ability of the models with >0.7 defined as acceptable.

Results: Of 890 patients who had RHC, 367 were eligible. Among these, 87% were females, mean age was 66 years, and 83% had limited cutaneous SSc.

In univariable analysis, only COMPERA 2.0 and REVEAL Lite 2 had acceptable predictive value with a C-index >0.7 (table). Adjusted for general and SSc-specific variables, all models were acceptable (table, fig. 1).Numerically, COMPERA 2.0 and REVEAL Lite 2 were the most accurate to predict mortality (table, fig. 1). Hazard ratios increased with higher risk scores (fig. 2). However, COMPERA 2.0 and REVEAL Lite 2 did not significantly discriminate the lower risk groups (fig. 2).

Conclusion: In SSc-PAH, we suggest that risk stratification at time of PAH diagnosis should take general and SSc-specific variables into account. The COMPERA 2.0 and REVEAL Lite 2 risk stratification models perform best when used as stand-alone tools. Neither requires invasive measurements, making them easily applicable in clinical practice. Importantly, these models identify patients in the high-risk group where aggressive upfront treatment is recommended.

References:
¹Humbert, Eur Heart J, 2022
²Hoeper, Eur Respir J, 2022
³Benza, Chest, 2021

« Back to ACR Convergence 2023

ACR Meeting Abstracts - https://acrabstracts.org/abstract/comparison-of-four-risk-stratification-models-for-prediction-of-mortality-in-systemic-sclerosis-associated-pulmonary-arterial-hypertension-in-the-eustar-cohort/