Isolated tricuspid valve surgery: impact of aetiology and clinical presentation on outcomes

Abstract

Aims

The aim of this study was to identify determinants of in-hospital and mid-term outcomes after isolated tricuspid valve surgery (ITVS) and more specifically the impact of tricuspid regurgitation (TR) mechanism and clinical presentation.

Methods and results

Among 5661 consecutive adult patients who underwent a tricuspid valve (TV) surgery at 12 French tertiary centres in 2007–2017 collected from a mandatory administrative database, we identified 466 patients (8% of all tricuspid surgeries) who underwent an ITVS. Most patients presented with advanced disease [47% in New York Heart Association (NYHA) III/IV, 57% with right-sided heart failure (HF) signs]. Tricuspid regurgitation was functional in 49% (22% with prior left-sided heart valve surgery and 27% isolated) and organic in 51% (infective endocarditis in 31% and other causes in 20%). In-hospital mortality and major complications rates were 10% and 31%, respectively. Rates of survival and survival free of HF readmission were 75% and 62% at 5 years. Patients with functional TR incurred a worse in-hospital mortality than those with organic TR (14% vs. 6%, P = 0.004), but presentation was more severe. Independent determinants of outcomes were NYHA Class III/IV [odd ratios (OR) = 2.7 (1.2–6.1), P = 0.01], moderate/severe right ventricular dysfunction [OR = 2.6 (1.2–5.8), P = 0.02], lower prothrombin time [OR = 0.98 (0.96–0.99), P = 0.008], and with borderline statistical significance, right-sided HF signs [OR = 2.4 (0.9–6.5), P = 0.06] while TR mechanism was not [OR = 0.7 (0.3–1.8), P = 0.88].

Conclusion

Isolated TV surgery was associated with high mortality and morbidity, both in hospital and during follow-up, predicted by the severity of the presentation but not by TR mechanism. Our results suggest that TV intervention should be performed earlier in the course of the disease.

Open in new tabDownload slide

Open in new tabDownload slide

See page 4318 for the editorial comment on this article (doi: 10.1093/eurheartj/ehaa740)

Introduction

Tricuspid regurgitation (TR) is a frequent condition that has been overlooked compared to left-sided valve diseases. Epidemiological studies suggest that moderate or severe TR affects over 1.6 million individuals in the USA.^{1  ,  2} Multiple studies, in various settings such as heart failure (HF) with reduced left ventricular ejection fraction (LVEF), tricuspid flail, or isolated functional TR, have shown that the presence and severity of TR are associated with an increased risk of mortality and of cardiovascular (CV) events.^3–8 In a recent meta-analysis, moderate/severe TR was associated with more than a two-fold increased cardiac mortality independently of pulmonary pressures, LVEF, and right ventricular (RV) dysfunction.⁹ Despite its high prevalence, TR is often left untreated and the number of tricuspid valve (TV) surgeries, mostly performed at the time of mitral valve surgery, remains remarkably low.¹⁰

Isolated tricuspid valve surgery (ITVS) is reputed to be a high-risk surgery with a 10% in-hospital mortality rate that had remained constant across studies and time despite improvement in surgical techniques and post-operative management.^10–12 However, most studies relied on small single-centre studies or on administrative databases with limited clinical, biological, and echocardiographic information and limited follow-up. Thus, determinants of outcome after ITVS have remained so far elusive. In a previous study, using the French Programme de Médicalisation des Systèmes d’Information (PMSI), a mandatory administrative database, congestive heart failure (CHF) at presentation was associated with mortality and major complications rates.¹² However, granularity was limited and we were not able to account for the severity of the clinical presentation and echocardiographic features. In addition, TR mechanism may be either functional or organic and encompass multiple aetiologies. Consequently, patients referred for ITVS represent a heterogeneous population. Impact of TR aetiology on outcome remained unclear and data are very limited.

In the present study, relying on in depth characterization of patients’ medical history, functional status, laboratory, and echocardiography results, we aimed to evaluate the determinants of in-hospital and mid-term outcomes after ITVS according to TR mechanism and aetiology and more specifically the impact of the severity of the clinical presentation.

Methods

Study design

Using the local PMSI database,¹³ we collected all consecutive cases of TV surgery performed in adult patients at 12 French tertiary centres between 1 January 2007 and 31 December 2017 (Supplementary material online, Figure S1). Tricuspid valve interventions were identified using Common Classification of Medical Procedures classification (DBKA004, DBMA008, or DBMA012).¹⁴ All healthcare institutions in France are mandated to transfer information regarding their activity into the PMSI database and the PMSI database is therefore considered exhaustive. Among all TV surgery cases performed in these centres, we excluded congenital TRs and patients who underwent a combined TV surgery, meaning a TV surgery concomitant with other cardiac surgery such as mitral valve surgery, coronary artery bypass graft, or atrial septal defect closure. Finally, we excluded all redo TV surgery either on native or prosthetic valves and our study population consisted of all patients who underwent an isolated non-congenital TV surgery performed on native valve. Clinical presentation, laboratory results, echocardiography, and immediate and mid-term outcome were retrospectively collected locally at each centre. The study was approved by each local institutional review board.

Patients’ characteristics

Information regarding clinical presentation, past medical history, and laboratory results were retrospectively collected from patients’ charts. Glomerular filtration rate was calculated using the Cockcroft–Gault formula according to sex. Surgical risk scores [logistic EuroSCORE, EuroSCORE II, and Society of Thoracic Surgeons (STS)] were calculated; mitral valve repair intervention was used for the STS as there is no dedicated score for ITVS.^{15  ,  16} The Charlson index was used to assess for patients’ comorbidities.

All echocardiographic reports before intervention were reviewed and when needed measurements were retrospectively performed on digitally stored images. Tricuspid regurgitation aetiology was first classified as functional (no intrinsic tricuspid disease) or organic (due to abnormalities of the tricuspid leaflets or chordae). Functional TR was then subdivided into functional TR after left-sided valve surgery or isolated functional TR, and organic into endocarditis and other aetiologies including carcinoid diseases or traumatic. Left ventricular ejection fraction was measured visually or using the biplane method of discs (modified Simpson’s rule). Right ventricular size was visually assessed as normal, mildly, moderately or severely enlarged. The RV basal diameter and the tricuspid annulus (TA) diameter (TA) were measured in diastole in the apical four-chamber view.¹⁷ Right ventricular function was semi-quantitatively assessed using an integrative approach based on visual assessment, tricuspid annulus plane systolic excursion (TAPSE), and Doppler tissue imaging peak systolic annular velocity S′ as recommended.^{18  ,  19} Pulmonary artery systolic pressure was measured using the peak systolic trans-tricuspid pressure gradient and estimated right atrial pressure according to the diameter of the inferior vena cava and its variation with respiration, applying the Bernoulli equation.

Outcome

In-hospital mortality was defined as death occurring between the intervention and hospital discharge during the same hospital stay. Major post-operative complications were defined as death, shock, tamponade, and acute renal failure requiring dialysis or prolonged (>72 h) mechanical ventilation. Lengths of stay (total and intensive care unit) were calculated as the time duration between the admission and hospital discharge. During follow-up, all-cause death and CV readmissions, defined as due to CHF, atrial fibrillation (AF), infective endocarditis, stroke, acute coronary syndrome, re-intervention (tamponade, sternum infection), or pacemaker related, were also collected based on hospital chart reviews and telephone interviews with patients or family members.

Statistics

Variables were expressed as mean ± standard deviation or median (percentile 25 and 75) for continuous variables and number of patients (percentage) for categorical variables. Trends in the number and type of TV surgery were estimated by the Mann–Kendall trend test. Comparisons between groups of TR mechanism/aetiologies were performed using ANOVA, χ  ², Student’s t-test or non-parametric Wilcoxon test as appropriate. Complication rates were expressed as percentage and 95% confidence intervals. Logistic regressions in univariate and multivariate analyses (including variables with P < 0.10 in univariate analysis) were used to assess the determinants of in-hospital death and major post-operative complications. Variables independently associated with mortality in each category (clinical, laboratory, and echocardiographic) were entered into the final model. Multiple comparisons for continuous variables were tested with the Bonferroni correction. Odd ratios (OR) along with 95% confidence intervals were reported. We assessed the risks of death and CV readmission overall and according to TR mechanism/aetiology using Kaplan–Meier estimates and Cox proportional hazard models. Data were analysed using JMP version 9.0 (SAS Institute) or XLSTAT (Microsoft, Redmond, WA, USA). All tests were two-sided and a P-value of <0.05 was considered statistically significant.

Results

Trends in tricuspid valve surgery

A total of 5661 adult patients underwent a TV surgery (Supplementary material online, Figure S2) at the 12 participating French tertiary centres between 1 January 2007 and 31 December 2017. Among them, 466 patients (8%) had a non-congenital isolated TV surgery on native valve (Figure 1). Trends in TV surgery are presented in Figure 2. The number of TV surgery increased from 280 in 2007 to 774 in 2017 (P for trend < 0.0001). The increase in number of TV surgery performed was mainly related to a marked increase in the number of combined TV surgery (P for trend < 0.0001) although isolated TV modestly but significantly increased during the study period (P for trend = 0.0001).

Figure 1

Flow chart of study population. A total of 5661 adult patients underwent a tricuspid valve surgery at 12 French tertiary centres between 1 January 2007 and 31 December 2017. Among them, 466 had an isolated non-congenital surgery, 229 for functional tricuspid regurgitation, and 237 for organic tricuspid regurgitation. Functional tricuspid regurgitations were then subdivided into functional tricuspid regurgitation after left-sided surgery or isolated functional tricuspid regurgitation and organic into endocarditis aetiology and other aetiologies including carcinoid diseases or traumatic.

Open in new tabDownload slide

Figure 2

Trends in tricuspid valve surgery between 1 January 2007 and 31 December 2017 in the 12 French tertiary centres. (A) All tricuspid valve surgery, combined tricuspid valve surgery, and isolated tricuspid valve surgery on native valve. (B) Tricuspid valve repair and tricuspid valve replacement—bioprosthetic and mechanical valves. TV, tricuspid valve.

Open in new tabDownload slide

Clinical presentation and aetiologies of isolated tricuspid valve surgery

Characteristics of the 466 patients with a non-congenital isolated TV surgery on native valve are presented in Table 1. Overall mean age was 60 ± 16 years and 229 patients (49%) were female. One hundred and eleven patients (24%) had a previous left-sided heart surgery. Close to half of the patients were severely symptomatic [47% in New York Heart Association (NYHA) functional Class III/IV] and 163 (35%) were admitted for CHF within the year preceding the surgery. Right-sided HF signs were present in 264 patients (57%) in whom 39 (8%) had ascites. Chronic kidney disease was present in 154 patients (33%) and chronic liver disease in 54 patients (12%). Mean LVEF was 58 ± 9%. A moderate/severe RV dilatation was present in 54% of patients and a moderate/severe RV dysfunction in 17%. The mean systolic pulmonary artery pressure was 40 ± 11 mmHg (≥50 mmHg in 19%). The median Charlson comorbidity index was high [3.0 (1–4)], and 324 patients (70%) had a Charlson comorbidity index of ≥2.

Tricuspid regurgitation mechanism was functional in 229 patients (49%) and organic in 237 patients (51%). Among patients with functional TR, 101 patients (22%) had a prior left-sided heart valve surgery and 128 (27%) patients had an isolated functional TR (Figure 1). Among the 237 patients with organic disease, TR was due to infective endocarditis in 142 patients (31%) and to other causes in 95 patients (20%). Patients with functional TR were older, more frequently female and presented more often with CV risk factors and comorbidities than those with organic TR. Importantly clinical presentation was also more severe. Patients with functional TR presented more advanced signs of HF signs, worse laboratory results, and more often moderate/severe RV enlargement and dysfunction than patients with organic TR. The Charlson index was twice as high in patients with functional TR than in patients with organic TR. As expected, TV infective endocarditis affected a younger mainly male population with infrequent CV risk factors. Clinical and echocardiographic RV parameters also tended to be less severe than those of the three other groups. Comparisons of baseline characteristics between functional and organic TR and between TR aetiologies are presented in Table 1.

In-hospital outcomes of isolated tricuspid valve surgery

A TV repair was performed in 193 patients (41%) and a replacement in 273 patients (59%) (Supplementary material online, Figure S2). A TV replacement was more often performed when the TV surgery was performed isolated than concomitant with other cardiac interventions (59% vs. 4%, P < 0.0001). In most cases of ITVS, a bioprosthetic valve was used (n = 264, 57%) while a mechanical valve was implanted in only nine patients (2%). Surgery was urgent in 102 patients (22%). Overall, in-hospital mortality rate was 10% (n = 48) and major post-operative complications rate was 31% (n = 145) (Table 2 and Figure 3). In the post-operative period, 87 patients (19%) presented a cardiogenic shock, 36 patients (8%) a tamponade, and 123 (26%) acute renal failure, of whom 49 (11%) requiring dialysis, 40 (9%) a prolonged mechanical ventilation, and 84 (18%) needed a definitive pacemaker. The median intensive care unit and total lengths of stay were 4 days (2–8) and 18 days (12–29), respectively. Residual TR was rare and 98% of patients alive at discharge presented with less than moderate TR.

Figure 3

In-hospital mortality and major post-operative complications rates overall and according to the aetiology of the tricuspid regurgitation.

Open in new tabDownload slide

Factors associated with in-hospital death in univariate analysis are presented in Table 3. In-hospital mortality was associated with older age and more severe presentation based on clinical, laboratory, and echocardiographic assessment such as history of CHF, severe symptoms (NYHA Class III/IV), presence of ascites, cirrhosis, dose of loop diuretics, kidney impairment, comorbidity index, and surgical risk scores and degree of RV dilatation and dysfunction. Functional TR was also associated with higher in-hospital mortality rate than organic aetiology (14% vs. 6%, P = 0.004) while history of previous left-sided heart valve surgery was not (P = 0.10). Mortality rates according to TR aetiology were 16% for functional TR after left-sided valve surgery, 13% for isolated functional TR, 5% for infective endocarditis, and 8% for organic TR of other aetiology (P = 0.02) (Figure 3). In multivariate analysis, independent determinants of in-hospital death were NYHA Class III/IV [OR = 2.7 (1.2–6.1), P = 0.01], moderate/severe RV dysfunction [OR = 2.6 (1.2–5.8), P = 0.02], lower prothrombin time [OR = 0.98 (0.96–0.99), P = 0.008], and with borderline statistical significance, presence of right CHF signs [OR = 2.4 (0.9–6.5), P = 0.06] while mechanism was not [OR = 0.7 (0.3–1.8), P = 0.88]. Systolic pulmonary artery pressure ≥50 mmHg was also of borderline statistical significance [OR = 2.5 (0.9–7.0), P = 0.08]. Independent determinants of major in-hospital complications in multivariate analysis were similar, with moderate/severe RV dysfunction [OR = 2.2 (1.2–4.1], P = 0.01], lower glomerular filtration rate [OR = 0.98 (0.96–0.99), P = 0.006], lower prothrombin time [OR = 0.98 (0.97–0.99), P = 0.02], and with borderline statistical significance, NYHA Class III/IV [OR = 1.6 (0.95–1.1), P = 0.07], AF [OR = 1.7 (0.9–3.1), P = 0.08] and moderate/severe RV dilatation [OR = 1.6 (0.95–2.6), P = 0.08]. Excluding patients with permanent pacemaker at the time of the surgery did not change our results. Our results also remained similar after the exclusion of patients with infective endocarditis or if the prothrombin time was not included into the model.

Mid-term outcomes of isolated tricuspid valve surgery

Follow-up was 90% complete. During an overall mean follow-up duration of 3.1 ± 2.7 years [median 2.6 years (0.8–4.9), up to 11.8 years], there were 40 additional deaths (all-cause) and 138 patients were readmitted for CV reasons including 64 patients for CHF. Overall rates of survival, survival free of CV readmission, and survival free of HF readmission were 86%, 73%, and 79% at 1 year and 75%, 44%, and 62% at 5 years, respectively (Figure 4A, Take home figure). Event rates among the 371 patients alive at 30 days and with available follow-up were 98%, 83%, and 90% at 1 year and 85%, 50%, and 70% at 5 years (Figure 4B). Survival, survival free of CV readmission, and survival free of HF readmission curves according to TR aetiology are presented in Figure 5 and were significantly worse for functional than for organic TR (all P < 0.01). Mortality and CV events according to TR aetiologies were also significantly different when subdivided into functional TR after left-sided valve surgery, isolated functional TR, organic TR due to infective endocarditis, and organic TR due to other aetiologies (all P < 0.05) (Supplementary material online, Figure S3). However, outcome of both functional groups was not different (all P > 0.40).

Figure 4

Overall event rates (A) and a landmark analysis for patients alive at 30 days (B). The figure displays Kaplan–Meier curves of all-cause mortality (in grey), all-cause mortality or readmission for cardiovascular reasons (in yellow), and all-cause mortality or readmission for heart failure (in brown).

Open in new tabDownload slide

Figure 5

Event rates according to the aetiology (functional or organic) of the tricuspid regurgitation. The figure displays Kaplan–Meier curves of (A) survival (from all-cause death), (B) survival (from all-cause death) free of cardiovascular readmission, and (C) survival (from all-cause death) free of heart failure readmission for functional (in red) and organic (in purple) aetiology.

Open in new tabDownload slide

Take home figure

Isolated tricuspid valve surgery is associated with high mortality and morbidity, both in hospital and during follow-up, predicted by the severity of the pre-operative clinical, biological, and echocardiographic presentation but not by tricuspid regurgitation mechanism. NYHA, New York Heart Association.

Open in new tabDownload slide

In the multivariate analysis, independent determinants of all-cause mortality were NYHA Class III/IV [HR = 1.7 (1.1–2.8), P = 0.04], presence of right CHF signs [OR = 1.8 (1.01–3.3), P < 0.05], lower prothrombin time [OR = 0.98 (0.97–0.99), P = 0.04], and with borderline statistical significance, functional aetiology [OR = 0.6 (0.3–1.1), P = 0.07], lower glomerular filtration rate [OR = 0.99 (0.98–1.01), P = 0.09], and moderate/severe RV dilatation [OR = 1.6 (0.9–2.7), P = 0.08]. Similar determinants were observed for all-cause mortality or CV readmissions {NYHA Class III/IV [OR = 1.4 (1.0–2.0), P = 0.05], presence of right CHF signs [OR = 1.8 (1.2–2.6), P = 0.002], lower prothrombin time [OR = 0.99 (0.99–0.99), P = 0.03], and moderate/severe RV dilatation [OR = 1.6 (1.1–2.3), P = 0.006]} and for all-cause mortality or HF readmissions {NYHA Class III/IV [OR = 1.7 (1.1–2.3), P = 0.02], presence of right CHF signs [OR = 1.9 (1.2–3.1), P = 0.006], and moderate/severe RV dilatation [OR = 2.0 (1.3–3.1), P = 0.002]}. Functional aetiology was not an independent determinant of mortality or (CV or HF) readmissions [OR = 0.9 (0.6–1.4), P = 0.67 and OR = 0.9 (0.5–1.3), P = 0.37, respectively].

Discussion

In this retrospective series, we collected all cases of TV surgery performed at 12 French tertiary centres between 1 January 2007 and 31 December 2017 and assessed the clinical, biological, and echocardiographic presentation and immediate and mid-term outcomes of ITVS. Main results of the present study can be summarized as follows: (i) although total numbers had increased over time, TV surgery remained seldom performed; (ii) TV surgery was mainly concomitant with another surgical procedure and rarely isolated; (iii) ITVS was performed in approximately half of cases for functional TR (after left-sided valve surgery in 46% and for isolated functional TR in 54%) and in the other half for organic TR (due to infective endocarditis in 60%); (iv) patients were often referred late in the course of the disease with severe symptoms, frequent impairment of liver and kidney functions, and advanced RV consequences; (v) in-hospital mortality and CV events rates were high with long in-hospital lengths of stay; (vi) although in-hospital mortality and morbidity were significantly worse for functional TR than for organic TR, outcome was more related to the severity of the clinical presentation observed in this subgroup than to TR mechanism or aetiology, which portended limited independent prognostic information; and (vii) mid-term outcomes showed substantial rates of death and readmissions also predicted by the severity of the presentation.

Using the PMSI, a mandatory French administrative database, we were able to identify all TV surgeries performed at 12 French tertiary centres over the past 10 years. Based on our previous nationwide study, these 12 centres accounted for 40% of all TV surgery performed in France.¹² The number of TV surgeries increased over time, from less to 300 in 2007 to close to 800 in 2017, but remained far from the number of aortic valve replacements (>10 000/year) or mitral valve interventions (>6000/year) currently performed in France.²⁰ The vast majority of TV interventions consisted of a TV repair performed concomitant with another procedure, mostly with left-sided valve surgery. Both the American College of Cardiology/American Heart Association and the European Society of Cardiology guidelines recommend performance of a combined TV surgery during left-sided valve surgery in case of moderate–severe TR or TV annulus diameter >40 mm.^{21  ,  22} Assuming that the number of mitral valve surgeries has remained stable during the study period, the present study suggests a wider adoption of these recommendations but also suggests that a substantial number of patients with qualifying indication for combined TV interventions remain neglected.

Recent epidemiological studies have shown that TR prevalence was similar to that of aortic stenosis and mitral regurgitation.^{2  ,  23} In the community, prevalence of significant—moderate or severe—TR was estimated to be as high as 0.55% and up to 3% after 75 years.¹ Despite its high prevalence and its strong association with poor outcomes,^{3–5  ,  9  ,  24  ,  25} TR is often left untreated. Intervention (surgery or transcatheter) is the only curative treatment for severe TR and medical therapy should be considered as only palliative. The magnitude of the gap between the estimated TR prevalence and the observed rates of interventions for TR (∼40/years ITVS in the present study, 100–120/year at the nationwide level) is thus striking. Management of isolated severe TR often relies on two contradictory beliefs: one that severe TR is a benign condition and the other that ITVS is a high-risk surgery. In clinical practice, patients are often initially not referred for an intervention arguing that it is too early until they are deemed at too high risk and denied surgery due to a too advanced disease. The literature is fairly consistent showing that in-hospital mortality of ITVS is ∼10% and has remained stable over time.^{10  ,  12} The present study confirms these findings with a similar 10% in-hospital mortality rate. In addition, we also showed that mid-term mortality and CV event rates were substantial. At 5 years, one quarter of patients are dead, more than half are either dead or admitted for CV reasons, and one-third dead or admitted for HF.

Importantly, mortality rates are highly variable according to patients’ characteristics. In contrast to previous studies including ours, we were able to obtain much more details regarding TR aetiology and clinical presentation including laboratory and echocardiographic results and to evaluate the determinants of outcomes after ITVS. Most patients presented with advanced disease, severe CHF, kidney, liver, or RV dilatation/dysfunction. It is worth noting that the Charlson index in the present study was more than twice as high as the Charlson index observed in patients referred for transcatheter aortic valve implantation in the early days.²⁰ Functional TR accounted for half of ITVS performed and was associated with mortality and rates twice as high as organic TR. However, the observed worse outcome in patients with functional TR was mainly due to a more severe clinical, biological, and echocardiographic pre-operative presentation. Thus, NYHA class, presence of right CHF signs, prothrombin time, or moderate/severe RV dilatation or dysfunction were important determinants of outcomes whereas TR mechanism or aetiology was not when adjusted on patients’ characteristics including infective endocarditis as these patients presented milder TR consequences. Similarly, previous left-sided heart valve surgery and TV replacement (vs. repair) were markers of advanced disease but were not independent determinants of the outcome. It is worth noting that repair rates for ITVS were similar than in the USA, but use of mechanical valves was much lower (one-third of TV replacements in the USA compared to only 3% in the present study).¹⁰ Thus, in-hospital mortality of patients with ascites and RV dysfunction were 26% and 22%, respectively. Importance of RV size and function and of their combination has also been recently highlighted in a large series of patients with functional TR.²⁶ Results of ITVS are even more striking considering that the population was relatively young and that TV surgery is relatively simple. Interestingly, there is no dedicated surgical risk scoring for ITVS and those for mitral valve surgery markedly underestimated the observed mortality rates. Altogether, our results suggest that the paradigm should be shifted: the high mortality and morbidity observed after ITVS are more related to patients’ late referral with advanced symptoms, kidney and liver dysfunction, and RV consequences than to the surgical intervention per se.

Strengths and limitations of the present study need to be discussed. First, we retrospectively collected TV surgery cases performed at 12 French centres. However, identification of cases at these centres was performed using the PMSI, a mandatory database considered exhaustive, and we thus captured all cases done at these centres. In addition, based on our previous study, we were able to estimate that our 12 centres accounted for 40% of all ITVS performed in France insuring generalizability of our findings. Second, in contrast to previously published studies using administrative databases, which lack granularity, and with limited follow-up, we were able to provide in-depth assessment of patients’ presentation including laboratories and echocardiographic results. Such level of details, as well as our sample size, unique in the literature, enabled us to identify the mechanism of TR and determinants of in-hospital and mid-term outcomes. Third, echocardiographic evaluation relied on centres’ evaluation and not on a centralized laboratory. Assessment of RV dilatation or dysfunction was based on an integrative approach performed by each centre, and we could not provide specific thresholds for RV dilatation or dysfunction that could have guided surgical indications. Nevertheless, as an internal validation, RV basal diameter was markedly different between patients classified as presenting a moderate/severe RV dilatation compared to those who did not (54 ± 8 vs. 42 ± 10 mm, P < 0.0001), and TAPSE and Doppler tissue imaging peak systolic annular velocity S′ were markedly different between patients who presented with moderate/severe RV dysfunction compared to those who did not (13 ± 4 vs. 21 ± 6 mm and 7.6 ± 1.5 vs. 12.4 ± 3.2 cm/s, respectively, both P < 0.0001). Fourth, as a retrospective study, not all laboratory and echocardiographic parameters could be collected for all patients. However, main parameters of interest were available in the vast majority of our population. In addition, we did not collect information regarding right catheterization performance and results. It is probable that patients with high systolic pulmonary pressure would be those with the highest event rates and those who might benefit the less from the intervention. Systolic pulmonary artery pressure measured using echocardiography was of borderline statistical significance but was only available in 274 patients (59%) and may be underestimated in patients with severe/laminar TR. Fifth, TA diameter was not predictive of outcome in the overall population and in the subgroup of patients with functional TR. These findings should be interpreted in light of the baseline characteristics of our population as most patients presented with significant TA enlargement while prognostic value of TA diameter has been shown in less selected patients.²⁶ Sixth, assessment of TR aetiology and mechanism may be challenging in few patients in particular in the presence of permanent pacemaker. However, our results remained unchanged after exclusion of the 104 patients with pacemaker. Finally, we could only infer that the high mortality and morbidity observed in the present study are related to a late surgical referral and that an earlier surgery would have improved the outcome. However, the limited impact of TR aetiology on in-hospital and mid-term outcomes compared to the weight of the clinical presentation including laboratory and echocardiographic TR consequences strongly supports this hypothesis. Future randomized controlled trials will evaluate the impact of TR correction on outcome and identify patients who will benefit the most of an intervention.

With regard to clinical implications, the present study highlights the need to prevent the occurrence of late TR after left-sided surgery and the importance of timely referral for TV interventions. As mentioned above, although debated, a more aggressive approach on TV intervention at the time of the left-sided surgery should be advocated to prevent the occurrence of late TR and the need for a repeat intervention seldom performed in real life. The high mortality and morbidity rates observed during ITVS for functional TR after left-sided valve surgery should be regarded as a strong incentive for the performance of concomitant TV surgery when indicated. On the other hand, timely referral for the intervention of patients with severe TR is also critical. Current guidelines^{21  ,  22} are quite vague with no Class I recommendation for intervention, but as for other valvular diseases, patients with severe TR should be referred earlier in the course of the disease avoiding the performance of TV intervention in patients with advanced disease at a desperate stage. In this regard, evaluation of RV size and function, often challenging especially in the setting of severe TR, should be refined. It is worth noting that we have illuminated only part of the problem as many patients with severe TR are often conservatively managed and denied any intervention. Thus, mortality and morbidity related to TR are likely underestimated. In a near future, the development of transcatheter therapies may represent an alternative to surgery and will further push for an early intervention if proved to be safe and efficient. As for functional mitral regurgitation,^{27  ,  28} patients’ selection will be critical and ‘early’ intervention before the occurrence of irreversible consequences will be key. Impact of TR correction on outcome has been recently disputed.^{29  ,  30} A later referral in surgical than transcatheter series may explain the contradictory results reported in the literature. Compared to our population, patients who underwent a transcatheter intervention were older (77 years), the vast majority presented with functional TR in AF (90% and 83%, respectively) with concomitant 3+/4+ mitral regurgitation while pulmonary pressure was similar. Patients referred for surgery were of similar age, similarly symptomatic than our functional TR subset but presented with higher systolic pulmonary pressures.

Conclusion

In this large multicentre series, we collected all cases of TV surgery performed at 12 French tertiary centres between 1 January 2007 and 31 December 2017 and assessed the clinical presentation and immediate and mid-term outcomes of ITVS. Compared to aortic or mitral valve surgery, tricuspid surgery remained uncommon and mainly performed at the time of left-sided surgery. Isolated TV surgery was rare, associated with high mortality and morbidity, both in hospital and during follow-up, predicted by the severity of the clinical presentation, while TR mechanism/aetiology had limited impact. Our results suggest adopting a more aggressive approach for TV intervention, either at the time of left-sided surgery to prevent the occurrence of late TR, or intervening earlier in the course of the disease for patients with severe isolated TR, whatever the mechanism, before the occurrence of advanced/irreversible consequences and intractable HF.

Data sharing

Data sharing with qualified researchers may be considered after submission of a proposal to Docteur Dreyfus.

Supplementary material

Supplementary material is available at European Heart Journal online.

Conflict of interest: B.I. received consultant fees from Edwards Lifesciences. J.-F.O. received consultant fees for Abbott Vascular, Carmat, Delacroix-Chevalier, Landanger, Medtronic, and Sorin. D.M.-Z. received consultant fees from Edwards Lifesciences. All other authors declared no conflict of interest.

References