Download files
Abstract
Background: Pulmonary hypertension (PH) complicating interstitial lung disease (ILD) may be associated with significant morbidity and mortality. Transthoracic echocardiography (TTE), the most widely used screening test for PH, may be inaccurate in up to 50% of ILD patients. We evaluate the utility of TTE, in isolation and in combination with several non-invasive screening tests, for predicting the presence and severity of ILD-PH. Novel techniques, including non-invasive assessment of pulmonary blood flow and dual energy CT pulmonary angiography (DE-CTPA), were also evaluated for identifying pulmonary vasospasm in systemic sclerosis (SSc).
Methods: In 265 consecutive patients with ILD and suspected PH, the predictive utility of the updated European Society of Cardiology/European Respiratory Society (ESC/ERS) TTE PH screening recommendations was evaluated against RHC testing. In a subgroup of 114 patients, combined non-invasive investigations (including TTE, respiratory function tests, brain natriuretic peptide (BNP), vascular dimensions on CT, and 6 minute walk test) was assessed for predicting PH. Equations predicting the presence and severity of PH were then tested prospectively in a validation cohort.
Results: In patients classified as ‘high probability’ of PH by ESC/ERS TTE criteria, PH was confirmed on RHC in 86%. However, 40% of patients were misclassified as ‘low probability’ of PH (tricuspid regurgitation velocity <2.8 m/s), when PH was confirmed on subsequent RHC. At regression analysis, TTE derived right ventricular systolic pressure (RVSP) (R2=0.37; p<0.01), BNP (R2=0.25; p<0.01), and main pulmonary artery diameter (mPA) on CT scan (R2=0.23; p<0.01) correlated most strongly with RHC measured mean pulmonary artery pressure (mPAP), and a ‘best fit’ equation for predicting mPAP (tested in the validation cohort) was more strongly predictive of PH (with a receiver operating characteristic AUC of 0.93) than any single non-invasive test in isolation.
Conclusion: Current TTE screening recommendations are associated with a significant misclassification rate (40%) in excluding ILD-PH. Combining RVSP, mPA diameter and BNP more accurately predicted mPAP, the presence of PH and the presence of severe PH, compared to any single test in isolation. While pulmonary vasospasm was detectable in a subgroup of SSc patients, extensive post-processing of DE-CTPA data precluded this as a clinically useful test at this time.