Clinical prediction score for identifying patients with pulmonary veno-occlusive disease/pulmonary capillary hemangiomatosis

Background: Pulmonary veno-occlusive disease (PVOD) and pulmonary capillary hemangiomatosis (PCH) are rare causes of pulmonary hypertension. Although diagnosis is based on pathological findings, an early diagnosis is crucial because of poor prognosis compared to other types of pulmonary hypertension. Furthermore, vasodilators may cause fatal pulmonary edema in patients with PVOD/PCH. This study aimed to identify specific characteristics for patients with PVOD/PCH to clinically diagnose PVOD/PCH. Methods: Clinical data were obtained at baseline and were compared between 19 patients with PVOD/PCH and 55 patients with idiopathic/heritable pulmonary arterial hypertension. Receiver operating characteristic analyses were used to determine characteristics specific for patients with PVOD/PCH and a scoring system to diagnose PVOD/PCH was developed. Results: Patients with PVOD/PCH had a smoking history and were predominantly male. Six-minute walk distance was significantly lower and oxygen desaturation was severe during the walk. Diffusion capacity of carbon monoxide was significantly low. Radiological findings included ground glass opacity on chest high-resolution computed tomography (CT) in all patients with PVOD/PCH, and thickened septal lines in 90% of the patients. Lung perfusion scintigraphy showed defect in >70% of the patients. Pulmonary edema after initiation of vasodilation therapy was frequently observed in PVOD/PCH patients. Based on these results, we identified nine important clinical characteristics and a novel scoring system was designed to clinically diagnose PVOD/PCH: male sex, smoking history, 6-minute walk distance < 285 m, minimum SpO(2) < 92% during the 6-minute walk test, %DLco < 34%, ground glass opacity and thickening of the interlobular septa in high-resolution CT, defects in the perfusion lung scan, and pulmonary edema due to vasodilators. Score >= 5 points had 95% sensitivity and 98% specificity to predict PVOD/PCH (area under the curve: 0.991; 95% CI: 0.976-1.000). Conclusions: Our novel prediction rule for diagnosing PVOD/PCH may offer an early clinical diagnosis of these diseases. (C) 2018 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.