No abstract

Rationale: The European Society of Cardiology (ESC) and European Respiratory Society (ERS) previously published guidelines in 2009 using right ventricular systolic pressure (RVSP) cut offs to detect pulmonary hypertension. The aim of this study was to define echocardiographic parameters to predict the presence of pulmonary hypertension (PH) when compared to right heart catheterization (RHC). Methods: We studied patients evaluated for PH with a RHC and an echocardiogram within 4 weeks of each other between February 1996 and February 2019 enrolled in our PH registry. Logistic regression and the Youden method were used to calculate RVSP values to maximize sensitivity and specificity to detect the presence or absence of PH. We internally validated these results splitting the data into training and testing randomly, with a ratio of 7:3, using logistic regression to build the training model, and then repeating this process 100 times with different seed. We calculated the likelihood ratios for the RVSP cutoffs. We assessed whether the presence of right atrial dilation (RAD) and right ventricular dilation (RVD) could improve the predictive ability RVSP. The analysis was repeated using the previous definition for PH, mPAP >25mmhg. Results: Our study population included 1,250 patients, mean age 57.4±15.8 years, 63.6% male, mean BMI 30.7±8.5. 1171 patients had mPAP > 20 mmHg and 79 had normal pulmonary pressures. RVSP had a moderate correlation with mPAP (Pearson's correlation coefficient =0.58, p<0.001). The cut offs for RVSP to maximize sensitivity and specificity were 34.5 mmHg and 52.5 mmHg, respectively (c-statistic = 0.89 for the whole cohort analysis and the internal validation analysis). RVSP>52.5 had a sensitivity of 76.9%, specificity of 89.9%, and positive likelihood ratio of 7.6 to detect PH. RVSP <34.5 had a sensitivity of 95.1%, specificity of 35.96%, and a negative likelihood ratio of 0.12 to exclude PH. The addition of RAD and RVD improved the accuracy of RVSP in the low (RVSP < 34.5) and intermediate (34.5 < RVSP <52.5) RVSP groups (c-statistic from 0.52 to 0.63, and from 0.83 to 0.88, respectively, p < 0.05). When using the previous definition of PH, the upper and lower RVSP cutoffs were 58.5 and 36.5 mmHg, respectively. Conclusions: Lower RVSP cut offs are suggested to detect newly defined PH, particularly the lower cutoff associated with a low probability of PH (RVSP 34.5 mmHg). RAD and RVD increases the diagnostic accuracy of echocardiography RVSP in patients with low and intermediate values of RVSP.

No abstract

No abstract