Results
A total of 195 patients who underwent isolated tricuspid valve surgery at our institution were reviewed and, based on the postoperative course, divided in regular and non-regular categories. 23/195 patients (12%) were excluded from the study due to lack of sufficient data and therefore inability of correctly assigning them to either category (Figure 1 ). A total of 172 patients were finally considered for the purpose of this study, among whom 97 (56.3%) had a regular postoperative course and 75 patients (43.7%) a complicated one. In the overall population, 43 patients underwent TVr (25%) and 129 (75%) underwent TVR. Among the 43 patients submitted to repair, a ring annuloplasty was performed in 37 of them (86%), with or without concomitant leaflet repair (including edge-to-edge or clover technique), whereas the remaining 6 patients received suture annuloplasty and/or leaflet repair. 7/43 (16.3%) patients with TVr experienced a complicated postoperative course. In case of TVR a bioprosthesis was used in 123 patients (95.4%) and a mechanical prosthesis was used in the remaining 6 patients (4.6%). Among patients submitted to replacement, 68 (55.3%) had a non-regular postoperative course.
Analysing the preoperative characteristics, patients in the NEG group had worse baseline clinical and echocardiographic parameters, as shown in Table 1 . Indeed, patients with an unfavourable postoperative course were more frequently presenting with higher New York Heart Association (NYHA) functional class, previous episodes of right heart failure, ascites and higher diuretics dose. In addition, NEG patients presented with organ damage, such as chronic kidney disease (CKD), low albumin levels, and higher liver enzymes. Interestingly enough, the MELD score (Model for End-stage Liver Disease) was calculated for all patients and was found to be significantly higher in patients with a negative postoperative course. Furthermore, both right ventricular size and function, together with left ventricular dilation and pulmonary hypertension were more marked in patients with a complicated postoperative course (Table 1 ).
Comparing the operative findings of the two groups, both cardiopulmonary bypass times (CPB) (55min [49-77] in NEG group vs. 56.5min [46-71], p=0.771) and cross-clamp times (XCT) (39min [29-46] vs. 36min [28-46] in REG, p=0.205) were similar. The number of beating heart operations were also similar between the two groups. As expected, in-hospital mortality was significantly higher in the complicated group, resulting in 13% vs 0% in REG patients (p<0.001). Moreover, as shown in Figure 2 , patients within the NEG group developed a higher number of postoperative complications, such as acute kidney injury (AKI), surgical revision for bleeding, low cardiac output syndrome (LCOS), need for high dose inotropic support and infection/sepsis (Table 2 ). There was a higher chance of requiring new pacemaker implantation in patients with an unfavourable course (OR=28.1, 95% CI [3.65-217.04], p=0.001). This can be explained by the fact that patients with a complicated postoperative course needed TVR more frequently than patients with a smooth course, which were mainly subjected to TVr (Figure 3 ). Furthermore, a higher MELD score was associated with a higher risk of developing a greater number of postoperative complications, longer ICU and postoperative length-of-stay and was associated with a higher in-hospital mortality (Table 3 ).
The most important predictors of an unfavourable postoperative course (NEG group) at univariate analysis were chronic kidney disease, ascites, previous hospitalizations for right heart failure, RV dysfunction, previous heart surgeries, need of TVR and an elevated MELD score. The other (less powerful) predictors of this event identified at univariate analysis are detailed in Table 4 . At multivariate analysis, liver enzyme values and preoperative dose of diuretics were identified as the only independent predictors of a negative outcome (Table 4 ).