Postoperative outcome after reoperative isolated tricuspid valve surgery—is there a predictor for survival?

Abstract

 

Open in new tabDownload slide

OBJECTIVES

Reoperative tricuspid valve (TV) surgery is considered high risk even in the absence of additional concomitant cardiac procedures. The purpose of this study was to evaluate preoperative clinical parameters as predictors for survival after isolated reoperative TV surgery.

METHODS

From January 2005 to January 2019, 85 patients (mean age: 66.7 ± 10.3 years, 34 male) with severe isolated TV regurgitation and prior cardiac surgery were referred to our centre for elective or urgent TV repair/replacement; patients with endocarditis were excluded. We retrospectively analysed preoperative hepatorenal function [reflected by widely used clinical and laboratory parameters and the Model of End-stage-Liver Disease excluding International Normalized Ratio (MELD-XI) score] as a predictor for postoperative survival.

RESULTS

At hospital admission, the patients’ average preoperative New York Heart Association class was 2.9 ± 0.6, left ventricular ejection fraction 52.5 ± 10.6%, mean pulmonary artery pressure 24.7 ± 8.0 mmHg, creatinine 115.4 ± 66.6 μmol/l, bilirubin 20.0 ± 19.6 μmol/l and the mean MELD-XI score was 13.3 ± 4.0 μmol/l. The mean follow-up was 5.4 ± 4.2 years. Thirty-day mortality was 5%, 5-year survival was 60.6 ± 5.4% and 10-year survival was 42.9 ± 6.5%. The multivariable Cox regression analysis evaluated the MELD-XI score [hazard ratio (HR 1.144, confidence interval 95% 1.0–1.3, P = 0.005] and diabetes mellitus (HR 2.27, confidence interval 95% 1.0–5.0, P = 0.04) as significant predictors for excess mortality while age and mean pulmonary artery pressure did not reliably predict clinical outcome.

CONCLUSIONS

Hepatorenal dysfunction was one main factor accounting for limited postoperative survival in our patient cohort. The MELD-XI score is easy to calculate and seems to reliably predict the perioperative risk in patients with prior cardiac surgery and indication for TV surgery.

INTRODUCTION

Patients with indication for isolated tricuspid valve (TV) surgery after previous cardiac surgery are frequently suffering from secondary, functional TV regurgitation. TV regurgitation—remaining asymptomatic for a long period of time—is associated with a reduced life expectancy, as shown by several expert groups [1–5].

Consequently, the ‘European Guidelines for the management of heart valve disease’ state that TV surgery ‘should be carried out sufficiently early to avoid irreversible RV dysfunction’, but also to avoid organ failure [6]. In particular, isolated reoperative TV repair is considered remarkably high-risk surgery, despite few reports on favourable postoperative outcomes [7, 8].

While modern treatment options include reoperative TV surgery, various interventional methods, as well as watchful conservative treatment in these patients, the individual decision-making process oftentimes is complex. This is particularly true, when the consequences of long-lasting TV regurgitation are present—renal failure, ascites, anasarca, cardiac related liver cirrhosis and oesophageal varices.

The aim of our retrospective clinical study was to assess the outcome of patients with reoperative isolated TV surgery focusing on preoperative predictors for postoperative survival.

PATIENTS AND METHODS

The local Ethics Committee approved this research project and patient consent was obtained (ID 069/19-ek; 26 February 2019). In a 14-year period (January 2005 to January 2019), a total of 85 consecutive patients underwent reoperative TV surgery for severe isolated TV regurgitation after various previous cardiac procedures at our institution. All operations involving other concomitant procedures than isolated TV surgery (i.e. MAZE procedures, closure of an atrial septum defect/persistent foramen ovale with contact to the left side of the heart or patients with active endocarditis) were excluded from this study.

Pre-, intra- and postoperative parameters of all 85 patients were prospectively entered into our institutional patient data management system and retrospectively analysed. In addition, in-depth chart review was undertaken to gather further information from preoperative and predischarge echocardiographic reports for subsequent analysis. TV regurgitation was uniformly calculated by the assessment and quantification of the Vena Contracta in a four-chamber view [9]. To identify predictors for postoperative survival, we comprehensively analysed all available preoperative laboratory parameters, transoesophageal echocardiography data and right heart catheter reports. Pulmonary hypertension was defined as an increase in mean pulmonary artery pressure (PAP) ≥25mmHg at rest, measured by right heart catheterization [10]. In addition, we used the modified Model of End-stage-Liver Disease excluding International Normalized Ratio (MELD-XI) score [11] for patients with anticoagulation to calculate the preoperative hepatorenal dysfunction. To calculate the prognostic relevance for our patient cohort, we have chosen the cut-off value of MELD-XI >12—as verified before by Wernly et al. [12].

TV surgery was performed in institutional standard technique as described elsewhere, including bicaval cannulation and mild hypothermic cardiopulmonary bypass. TV repair or replacement was performed depending on the surgeon’s choice.

Follow-up

Individual patient follow-up included post-discharge questionnaires—inquiring into postoperative complications and quality of life—sent via post mail to all patients. Patients who did not submit a written response were personally contacted and a detailed phone interview conducted. If no further information about a patient could be obtained, the family physician and/or registration office was contacted. Thereby, follow-up was 94% complete with a mean follow-up time of 5.4 ± 4.2 years (range: 1 day to 14.3 years).

Statistical evaluation

Results are displayed in the standard format with continuous variables expressed as mean ± standard deviation and categorical data as proportions. Univariable analysis was performed using the chi square and Fisher’s exact tests. The appropriateness of variable transformations was determined by means of univariable analysis. Variables with univariable P < 0.05 were submitted to the multivariable models. Cox regression analysis was performed to evaluate the predictors for postoperative mortality and calculated with 95% confidence interval. The assumptions of the Cox Proportional Hazard model were analysed with the Schoenfeld residuals test. If there was not seen independence between residuals and time, a time-dependent Cox-Regression analysis was performed. Cumulative survival was calculated by the Kaplan–Meier method for follow-up and was calculated with standard error; for the comparison of curves, the log-rank test was utilized.

All statistical analyses were performed using SPSS statistical package 20.0 (IBM Corp. Released 2011. IBM SPSS Statistics for Macintosh, Version 20.0; IBM Corp., Armonk, NY).

RESULTS

Demographic parameters of the entire patient cohort are shown in Table 1. The mean New York Heart Association class of patients undergoing isolated reoperative TV repair/replacement was 2.9 ± 0.6, surgery was performed at a mean age of 66.7 ± 10.3 years and more than half of patients were female. Left and right ventricular function was mildly reduced with a mean left ventricular ejection fraction of 52.5 ± 10.6%, and a right ventricular function—measured as tricuspid annular plane systolic excursion—of a mean of 17.2 ± 4.7 mm. The mean PAP, as assessed by right heart catheterization, was elevated with a mean of 24.7 ± 8.0 mmHg. Half of the patients presented with atrial fibrillation and the majority had systemic arterial hypertension. A permanent pacemaker/automatic implantable cardioverter defibrillator was present in 36 patients (42%).

Surgical access for the procedure was in more than half of the patients (n = 52, 61%) through a right anterolateral minithoracotomy versus a median sternotomy in the remaining 33 patients (39% of the cohort).

The majority of patients underwent beating heart reoperative TV surgery (n = 79, 81%), and the remaining patients were operated in cardioplegic cardiac arrest with a mean cross-clamp time of 49.5 ± 24.7 min. Overall mean cardiopulmonary bypass time was 96.0 ± 38.3 min, and overall operation time in mean 173.4 ± 50.4 min.

Intraoperatively, the TV leaflets were described as unremarkable in 34 patients (40%), pathologies of the leaflets were observed in the remaining patients—with isolated restrictive movement in 21 patients (25%), restrictive movement due to a pacemaker lead in 6 patients (7%), sclerotic or thickened leaflets in 10 patients (14%), retraction in 8 patients (9%) and prolapsing, myxomatous degeneration or otherwise destructed leaflets in a total of 6 patients (7%).

TV repair was performed in 41 patients (48%) versus TV replacement in 44 patients (52%). Postoperative rethoracotomy due to bleeding in first 24/36 hours after surgery occurred in 6 respectively 8 cases (7%/9%). Haemorrhagic complications, which were followed by rethoracotomy, were seen in altogether 17 patients, affecting 10/33 patients (30%) after median sternotomy and 7/52 patients (13%) after anterolateral minithoracotomy.

Within 30 days after surgery, 4 patients (5%) died: 3 from acute cardiac failure and 1 patient due to a fatal cerebral haemorrhage. In each 2 of the patients, surgery was performed electively and urgent; in each 2 patients, the operation was undertaken via an anterolateral minithoracotomy and a median sternotomy. The Kaplan–Meier survival graphs according to ‘MELD-XI ≤ 12’ versus ‘MELD-XI > 12’ showed a 5-year survival of 73.8 ± 6.8% vs 44.2 ± 9.3% (log rank: P = 0.001), respectively (central image). Five-year survival for all patients was 60.6 ± 5.4%, and 10-year survival 42.9 ± 6.5%.

In the univariable Cox regression analysis, were detected the clinical parameters age (P = 0.034), MELD-XI (P = 0.001), diabetes mellitus (P = 0.001) and mean PAP (P = 0.006) as significant predictors for postoperative mortality (Table 2). The Schoenfeld residuals test showed for all variables that the slopes were not 0 (age: −0.348, P < 0.01; PAP mean: −0.377; P < 0.01; MELD-XI: −0.324, P < 0.01; diabetes mellitus: −0.339, P < 0.01). Due to that, we applied multivariable Cox regression analysis with the time-dependent covariable MELD-XI (categorized) × (time ≤ 1 year) as well as age, PAP mean and diabetes mellitus with variable selection stepwise backward (P_in = 0.15, P_out = 0.2) regarding risk factors for postoperative mortality: the procedure excluded age at surgery as well as the time-dependent covariatble. The final model estimated the following hazard ratios (HRs): MELD-XI: HR 1.14 [95% confidence interval (CI): 1.04–1.25; P = 0.005]; diabetes mellitus: HR 2.27 (95% CI: 1.02–5.02; P = 0.044); and PAP mean: 1.04 (95% CI: 0.99–1.10; P = 0.094) That is, that diabetes mellitus (HR 2.27) and an MELD-XI score by 5 point higher is associated with nearly double risk of postoperative mortality (HR = 1.96).

Altogether 45 patients died during follow-up. Among the 40 survivors, complete independence with regard to ‘carrying out daily tasks’ was reported in 20 patients (24% of the entire cohort of 85 patients), ‘occasionally in need of help’ was reported by 10 patients (12%), 3 patients were reported to be ‘immobile and in need of care’ (4%), while the status of the remaining 7 surviving patients (8%) was unknown.

DISCUSSION

Historically, reoperative isolated TV surgery is considered to be associated with a significant operative risk and a poor long-term prognosis. Bernal et al. [13] reported on a hospital mortality of over 35% and a 5-year survival of only 45% in a mixed cohort of 74 patients who underwent reoperations after TV repair between 1976 and 2002. While Casselman et al. [14] reported an operative mortality as low as 3.8% and a 4-year survival of 85.6+/−6.4% in patients operated at a high volume centre between 1997 and 2006. In an even more favourable report, Lee et al. [15] claimed an early mortality of only 2.1%; however, a 5-year survival of 65% between 1997 and 2008—both reoperative patient cohorts underwent minimally invasive TV surgery. We reported our overall results of 82 patients after reoperative isolated TV surgery (patients with acute TV endocarditis, MAZE procedures and persistent foramen ovale closing were included in contrast to the present investigation) between 1997 and 2010 with a 30-day mortality of overall 14.6% and 4.0% in patients who underwent elective surgery, 2 years survival was 63.0 ± 5.5% altogether [16]. The 30-day mortality in our current patient cohort—isolated reoperative TV surgery and operated in a period between 2005 and 2019—was 5%, independent from surgical access chosen or urgency of operation. Five-year survival was 60.6 ± 5.4%, revealing comparable longevity in these patients as reported by Bernal et al. [13] and Lee et al. [15]. Long-term survival is sobering—particularly with regard to ‘quality of life’. With completion of the follow-up, it became evident that only one-third of all patients carried out their daily tasks independently or with only occasional support.

Adequate patient selection requires reliable risk stratification.

In this predominately elective study cohort, survival was not associated with cardiac related factors as left ventricular ejection fraction or tricuspid annular plane systolic excursion. However, diabetes mellitus (HR 2.27) and an MELD-XI score by 5 point higher were associated with nearly double risk of postoperative mortality (HR = 1.96).

The Model of End-stage-Liver Disease (MELD) score is commonly calculated to assess the perioperative risk for hepatorenal dysfunction [17]. The MELD score is based on widely available laboratory parameters, i.e. bilirubin, albumin, creatinine, sodium blood levels and international normalized ratio (INR). In our cohort, half of the patients suffered from atrial fibrillation. These patients are commonly under anticoagulation, affecting the INR (and the MELD score) independently from hepatorenal dysfunction—also affecting the iNR.

In 2007, Heuman et al. [11] addressed this issue presenting the modified MELD-XI score accounting for the influence of dialysis on creatinine, excluding sodium and—most importantly in our patient cohort—excluding the INR. Consequently, in the past decade, the MELD-XI score has been progressively utilized to predict the mortality risk in critically ill patients, e.g. prior to heart transplantation [18, 19] or TV surgery [20]. Wernly et al. [12] verified the MELD-XI score >12 as a cut-off regarding prognostic relevance of the MELD-XI score in a retrospective study with 4381 medical patients admitted to an intermediate care unit (ICU) between 2004 and 2009—admission diagnoses were, e.g. myocardial infarction (n = 2034), sepsis (n = 694) and heart failure (n = 688).

While modern treatment options include reoperative TV surgery and increasingly also interventional methods, as well as watchful conservative treatment in patients with tricuspid regurgitation and previous surgery—the individual decision-making process for the best therapeutic option is oftentimes complex. This is particularly true, when the consequences of long-lasting TV regurgitation are present, i.e. renal failure, ascites, anasarca, cardiac related liver cirrhosis and oesophageal varices. The MELD-XI score excluding INR seems to be a parameter of prognostic relevance in this special patient cohort, especially due to the fact that many patients of these are anticoagulated.

Limitations

Our study is mainly limited through its retrospective nature and the small patient cohort, which is also very heterogeneous. Therefore, the statistical value of the Kaplan–Meier estimation regarding postoperative survival and the MELD-XI score >12 vs ≤12 are somewhat limited.

The small sample size and missing values limited the model building. Thus, a sensible model with time-dependent covariates was not possible. However, exploratory analysis supports the hypothesis that high MELD-XI values are associated with high mortality risk.

CONCLUSION

In conclusion, however, hepatorenal dysfunction was the main factor for limited postoperative survival in our patient cohort. The MELD-XI score is easy to calculate and offers reliable perioperative risk prediction in patients with prior cardiac surgery and indication for TV surgery. Eventually, our clinical data support our hypothesis—the best predictor for survival after reoperative TV surgery is early surgery in patients with TV regurgitation.

Conflict of interest: none declared.

Author contributions

Bettina Pfannmueller: Conceptualization; Methodology; Project administration; Writing—original draft. Luca-Marie Budde: Data curation; Formal analysis; Investigation. Christian D. Etz: Writing—review & editing. Thilo Noack: Data curation. Mateo Marin Cuartas: Data curation. Martin Misfeld: Data curation. Michael A. Borger: Supervision; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Shiv Kumar Choudhary, Arnaud Van Linden, Yoshihisa Tanoue and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• CI

  Confidence interval

 
• INR

  International normalized ratio

 
• MELD

  Model of End-Stage-Liver Disease

 
• MELD-XI

  Model of End-Stage-Liver Disease excluding International Normalized Ratio

 
• TV

  Tricuspid valve