Background: Pulmonary hypertension (pH) due to left heart disease (pH-LHD) is the most common form of pH in clinical practice.
Objectives: The purpose of the study is to develop a diagnostic nomogram predictive model combining conventional noninvasive examination and detection indicators.
Methods: Our study retrospectively included 361 patients with left heart disease (LHD) who underwent right heart catheterization between 2013 and 2020. All patients were randomly divided into a training cohort (253, 70 %) and a validation cohort (108, 30 %). pH was defined as resting mean pulmonary arterial pressure (mPAP) ≥25 mmHg measured by RHC examination. Data dimension reduction and feature selection were used by Lasso regression model. The nomogram was constructed based on multivariable logistic regression.
Results: A total of 175 patients with LHD were diagnosed with pH during their hospitalization, representing 48.5 % of the cohort. The mean age of the overall group was 55.6 years, with 76.7 % being male patients. Excessive resting heart rate, elevated New York Heart Association functional class, increased red blood cell distribution width, right ventricular end-diastolic diameter, and pulmonary artery systolic pressure measured by echocardiography were independently associated with the prevalence of pH-LHD. The inclusion of these 5 variables in the nomogram showed good discrimination (AUC = 0.866 [95 % CI, 0.820-0.911]) and optimal calibration (Hosmer-Lemeshow test, P = 0.791) for the validation cohort.
Conclusions: The noninvasive nomogram of pH-LHD developed in this study has excellent diagnostic value and clinical applicability, and can more accurately evaluate the presence risk of pH in patients with LHD.
Keywords: Nomogram; Noninvasive diagnosis; Pulmonary hypertension due to left heart disease.
Copyright © 2024 Elsevier Inc. All rights reserved.