CT signs of pulmonary hypertension before TAVI: long-term prognosis in a cohort of 469 patients

Abstract

Abstract Introduction Patients undergoing TAVI often present with comorbidities, such as pulmonary hypertension (PH), which has been shown to be predictive of poor outcome after TAVI, using echocardiography (TTE). CT scan is systematically performed before TAVI, and PH signs on CT may be used to assess long-term survival after TAVI, but data are scarce on this matter. Purpose The main objective of this study was to analyze the association between signs suggestive of PH on CT performed before TAVI and 4-years mortality. Methods Between 2013 and 2018, 469 consecutive patients undergoing TAVI at our institution were included. We analyzed all pre-procedural CT for signs of PH: pulmonary artery (PA) diameter, PA to aorta (PA/A) ratio, right ventricle (RV) basal diameter, and inferior vena cava (IVC) diameter, with double reading, blinded to clinical and echocardiographic data. Results Our population of elderly patients (mean age 85±6) having a severe aortic stenosis was highly symptomatic (NYHA III–IV in 71% of cases) with moderate PH on TTE (mean sPAP of 43±14 mmHg), and intermediate operative risk (STS-score 6.2±4.5%). After a median follow-up of 4 years after TAVI, multivariate analysis identified IVC and RV axial diameters as CT PH parameters associated with long term mortality (adjusted HR and 95% CI of 1.08 [1.05–1.10] (p<0.0001) and 1.03 [1.01–1,05] (p=0.002), respectively). An IVC diameter ≥20 mm and a RV axial diameter ≥37 mm on CT were independently associated with late mortality, with areas under the curve (AUC) of 0.75; IC95% (0.70–0.80) and 0.67 (0.62–0.72), better than any significant clinical or biological data and than echocardiographic sPAP (Figure 1). 4-year survival was 65±4% if IVC <20 mm vs 44±3% if IVC ≥20 mm (p<0.0001) (Figure 2), and 35±4% if RV <37 mm vs 54±5% if RV ≥37 mm (p<0.0001). Interestingly, TTE IVC diameters were not significantly associated with long-term mortality and could not be assessed in 66 patients (14%) due to poor abdominal echogenicity. On the 403 patients with IVC diameters, assessed both by TTE and CT scan, correlation was poor although significant (R2= 0.1), with significant differences between the 2 methods using a paired T test (17.3±5.4 mm versus 20.3±6.0; P<0.001). Moreover, on the 403 patients, the discrimination was far better with CT scan IVC diameters than with TTE diameters: 0.75 (0.70–0.80) vs 0.54 (0.48–0.59); p<0.001. Regarding RV diameter, too many data were missing on TTE to compare their prognostic impact with CT diameters. Conclusion This study is the first to outline IVC and RV diameters on preprocedural CT as being associated to long-term mortality after TAVI. They outperformed other clinical, biological and TTE prognostic factors as well as the usual operative risk scores in terms of discrimination. These simple measurements, which can be performed on any pre-TAVI CT, could therefore be included in routine practice to improve the selection of patients undergoing TAVI. Funding Acknowledgement Type of funding sources: None.