DLco, Kco and FVC/DLco in suspected systemic sclerosis-associated pulmonary hypertension: insights from the ASPIRE registry

Objectives There are limited data comparing parameters reflecting gas transfer used to assess the likelihood of pulmonary hypertension (PH) in patients with systemic sclerosis (SSc) and regarding the impact of transitioning to Global Lung Initiative (GLI)-predicted values. Methods 632 patients with suspected SSc associated PH were identified from the ASPIRE registry. Spirometry and CT reports were reviewed to identify significant lung disease. ROC curve analysis and correlations of the 3 markers of gas transfer with pulmonary arterial pressure were performed. Results Correlations of GLI-derived values with mean pulmonary arterial pressure were: DLco% r=−0.45, Kco% r=−0.42 and FVC%/DLco% r=0.37. Correlations in patients without lung disease were: DLco% r=−0.51, Kco% r=−0.44, FVC%/DLco% r=0.38, compared to patients with lung disease: DLco% r=−0.41, Kco% r=−0.39, FVC%/DLco% r=0.39. Area under the curve for the presence of PH in the overall study cohort was significantly superior for DLco% at 0.84 (optimal threshold 53%), compared with Kco% 0.74 (60%) and FVC%/DLco% was 0.74 (1.91), p both <0.001. Compared with European Coal and Steel Community-derived data, GLI-derived percent-predicted lung volumes were lower, DLco% and Kco% were higher and consequently FVC%/DLco% lower (p all <0.001). Conclusion DLco performed as least as strongly as Kco or FVC%/DLco% in terms of correlations with mPAP and diagnostic utility, regardless of the presence or absence of lung disease. Transitioning to GLI equations led to lower predicted spirometric volumes and higher DLco%. This should be considered when interpreting changes in values over time and when using screening algorithms.