Background: Follow-up risk assessment in patients with pulmonary arterial hypertension (PAH) is based on three non-invasive variables according to the 2022 ESC/ERS guidelines. We aimed to systematically assess routinely measured non-invasive clinical variables to identify residual risk utilizing an unbiased feature selection algorithm on real-world data.

Methods: We retrospectively studied 298 incident PAH patients from a German centre diagnosed between 2009 and 2019. LASSO-regularized Cox regression models with mortality or lung transplant after first follow-up assessment as the primary composite endpoint were applied to early follow-up assessments (exercise testing, echocardiography, electrocardiography, lung function testing and biochemistry). 

Results: In a training sub-cohort (n=208), 12 variables were associated with outcomes after correction for multiple testing and entered LASSO modelling. A model combining seven variables (WHO functional class [WHO-FC], 6-minute walking distance [6MWD], tricuspid annular plane systolic excursion to systolic pulmonary arterial pressure ratio [TAPSE/sPAP], right atrial area index, lung diffusing capacity for carbon monoxide, total lung capacity, and aspartate transaminase to alanine transaminase ratio) discriminated 1-year outcome status (area under the curve [AUC] 0.83) and improved discrimination in comparison to the guideline approach in a replication sub-cohort (n=90; DeLongŽs test p=0.002). WHO-FC, 6MWD and TAPSE/sPAP were sensitive to treatment initiation.

Conclusion: Residual risk might be captured by non-invasive procedures (e.g., echocardiography) during routine follow-up assessments in PAH.