Advertisement

Improving outcome of valve replacement for carcinoid heart disease

Open ArchivePublished:October 03, 2018DOI:https://doi.org/10.1016/j.jtcvs.2018.09.025

      Abstract

      Objective

      Carcinoid heart disease is characterized by tricuspid valve regurgitation and varying degrees of pulmonary valve regurgitation or stenosis. Valve replacement procedures may be complicated by systemic effects of carcinoid syndrome, as well as hepatic dysfunction and right heart failure. This study was performed to identify factors that might be associated with improving early mortality rates and late outcomes.

      Methods

      Between November 1985 and January 2018, 240 adult patients underwent surgery for carcinoid heart disease at the Mayo Clinic. We analyzed the association of multiple clinical and echocardiographic variables on early mortality and late survival.

      Results

      The median (interquartile range) age of patients was 63 years (55-69), and 117 patients (49%) were male. Before operation, 157 patients (70%) had New York Heart Association class III or IV limitation. Somatostatin analogs were used in 221 patients (92%), and long-acting somatostatins were used in 130 patients (54%). Loop diuretic therapy was used preoperatively in 125 patients (52%). Early mortality rate was 29% (9/22) between 1985 and 1994, but decreased to 7% (6/81) during 1995 to 2004, and to 5% (7/128) from 2005 onward. Overall survival estimates at 1, 3, and 5 years were 69%, 48%, and 34%, respectively. Older age, advanced New York Heart Association class, and a nonlinear effect of creatinine were independently associated with overall mortality.

      Conclusions

      Valve replacement for carcinoid heart disease has acceptable short-term mortality, and early risk has decreased in the current era. Earlier intervention may improve overall survival.

      Key Words

      Abbreviations and Acronyms:

      CI (confidence interval), NYHA (New York Heart Association), OR (odds ratio)
      Valve surgery for patients with carcinoid heart disease has acceptable operative mortality. Early operation on patients with lower NYHA class may improve short- and long-term outcomes.
      In patients with carcinoid heart disease and symptoms of right-sided heart failure, valve surgery improves functional outcomes. Early operative mortality has decreased significantly over the last decades, but late survival is limited by carcinoid tumor progression. Reintervention due to carcinoid involvement of bioprostheses is rare.
      See Commentary on page 108.
      Carcinoid tumors are malignancies of the neuroendocrine system. These tumors secrete high levels of vasoactive substances, including serotonin, and typically arise from the terminal ileum or appendix. Carcinoid syndrome is characterized by facial flushing, bronchoconstriction, and secretory diarrhea and may occur in patients with invasive carcinoid tumors or metastases. Cardiac involvement (carcinoid heart disease) occurs in approximately 20% to 50% of patients with carcinoid syndrome. Carcinoid heart disease leads to fibrosis and endocardial thickening of the heart valves and is thought to occur as the result of chronic exposure of valve tissue to high levels of serotonin.
      • Pellikka P.A.
      • Tajik A.J.
      • Khandheria B.K.
      • Seward J.B.
      • Callahan J.A.
      • Pitot H.C.
      • et al.
      Carcinoid heart disease. Clinical and echocardiographic spectrum in 74 patients.
      • Lundin L.
      • Norheim I.
      • Landelius J.
      • Oberg K.
      • Theodorsson-Norheim E.
      Carcinoid heart disease: relationship of circulating vasoactive substances to ultrasound-detectable cardiac abnormalities.
      • Bhattacharyya S.
      • Toumpanakis C.
      • Caplin M.E.
      • Davar J.
      Analysis of 150 patients with carcinoid syndrome seen in a single year at one institution in the first decade of the twenty-first century.
      • Davar J.
      • Connolly H.M.
      • Caplin M.E.
      • Pavel M.
      • Zacks J.
      • Bhattacharyya S.
      • et al.
      Diagnosing and managing carcinoid heart disease in patients with neuroendocrine tumors: an expert statement.
      Patients with carcinoid heart disease typically present with right heart failure and associated right heart enlargement (Figure 1, A) caused by tricuspid and pulmonary valve regurgitation; the tricuspid valve leaflets are thickened and retracted with shortening of subvalvular attachments (Figure 1, B). Carcinoid involvement of the pulmonary valve leads to thickening and retraction of the cusps, and there may be additional scarring and constriction of the pulmonary valve annulus leading to mild stenosis.
      • Fox D.J.
      • Khattar R.S.
      Carcinoid heart disease: presentation, diagnosis, and management.
      In addition, the left-sided heart valves can be affected (<10% of patients).
      • Connolly H.M.
      • Schaff H.V.
      • Mullany C.J.
      • Rubin J.
      • Abel M.D.
      • Pellikka P.A.
      Surgical management of left-sided carcinoid heart disease.
      • Arghami A.
      • Connolly H.M.
      • Abel M.D.
      • Schaff H.V.
      Quadruple valve replacement in patients with carcinoid heart disease.
      Cardiac involvement dramatically reduces survival of patients with carcinoid syndrome and metastatic disease, and heart failure contributes to poor quality of life in these patients who have limited life expectancy.
      Figure thumbnail gr1
      Figure 1Intraoperative photographs of carcinoid heart disease. A, Enlargement of the right atrium. B, Carcinoid fibrosis of the tricuspid valve. C and D, Carcinoid metastases.
      In patients with severe tricuspid and pulmonary valve regurgitation and symptoms of heart failure, valve replacement can improve functional capacity and may allow more aggressive treatment of hepatic metastases.
      • Connolly H.M.
      • Schaff H.V.
      • Abel M.D.
      • Rubin J.
      • Askew J.W.
      • Li Z.
      • et al.
      Early and late outcomes of surgical treatment in carcinoid heart disease.
      • Lillegard J.B.
      • Fisher J.E.
      • Mckenzie T.J.
      • Que F.G.
      • Farnell M.B.
      • Kendrick M.L.
      • et al.
      Hepatic resection for the carcinoid syndrome in patients with severe carcinoid heart disease: does valve replacement permit safe hepatic resection?.
      However, operation in patients with carcinoid heart disease may be challenging because of systemic effects of carcinoid syndrome, especially those with impaired hepatic function. Careful preoperative planning and intraoperative and postoperative management are particularly important in these patients.
      • Weingarten T.N.
      • Abel M.D.
      • Connolly H.M.
      • Schroeder D.R.
      • Schaff H.V.
      Intraoperative management of patients with carcinoid heart disease having valvular surgery: a review of one hundred consecutive cases.
      Despite these clinical challenges, previous reports from our institution have highlighted the feasibility of surgical valve replacement in patients with carcinoid heart disease.
      • Connolly H.M.
      • Schaff H.V.
      • Abel M.D.
      • Rubin J.
      • Askew J.W.
      • Li Z.
      • et al.
      Early and late outcomes of surgical treatment in carcinoid heart disease.
      • Møller J.E.
      • Pellikka P.A.
      • Bernheim A.M.
      • Schaff H.V.
      • Rubin J.
      • Connolly H.M.
      Prognosis of carcinoid heart disease: analysis of 200 cases over two decades.
      In the present study, we describe our surgical experience with management of carcinoid heart disease and analyzed preoperative factors associated with early (<30 day) mortality. Furthermore, we aimed to describe predictors of long-term mortality in patients with carcinoid heart disease.

      Materials and Methods

       Study Population

      This study was approved by the Mayo Clinic Institutional Review Board. Between November 1985 and January 2018, 240 patients underwent cardiac surgery for carcinoid heart disease. We collected patient data from a prospectively maintained database, which records information on all patients undergoing cardiac surgery at the Mayo Clinic and supplemented this with data obtained by chart review. Follow-up information was collected through questionnaires that are sent to all cardiac surgery patients at 1, 3, 5, 10, 15, and 20 years after operation. Vital status was determined using Accurint (LexisNexis, New York City, NY).

       Preoperative Assessment and Operative Planning

      All patients included in this study were cared for by a multidisciplinary team with special interest in carcinoid heart disease. Preoperative assessment included transthoracic echocardiography, and cardiac computed tomography or cardiac magnetic resonance imaging was performed as necessary. All patients underwent assessment of renal and liver function, and coagulation testing before surgery.
      Perioperative management of patients with carcinoid heart disease is important to avoid triggering a carcinoid crisis, which is a life-threatening complication that may be exacerbated by general anesthesia and trauma of surgery. A carcinoid crisis is characterized by sudden onset of hypotension (or labile blood pressure), tachycardia, flushing, and bronchoconstriction. To prevent this complication, all patients receive somatostatin analogs preoperatively, which, in current practice, include the long-acting formulation.
      • Strosberg J.
      • Weber J.
      • Feldman M.
      • Goldman J.
      • Almhanna K.
      • Kvols L.
      Above-label doses of octreotide-LAR in patients with metastatic small intestinal carcinoid tumors.
      • Ramage J.K.
      • Ahmed A.
      • Ardill J.
      • Bax N.
      • Breen D.J.
      • Caplin M.E.
      • et al.
      Guidelines for the management of gastroenteropancreatic neuroendocrine (including carcinoid) tumours (NETs).
      Patients with very active disease despite outpatient somatostatin therapy are hospitalized preoperatively for intravenous infusion of octreotide (100-200 μg/h, 12 hours preoperatively). The infusion is continued during and after the procedure. If the patient experiences signs and symptoms suggestive of carcinoid crisis, a 1000 μg bolus of octreotide is administered with intravenous fluids.
      • Davar J.
      • Connolly H.M.
      • Caplin M.E.
      • Pavel M.
      • Zacks J.
      • Bhattacharyya S.
      • et al.
      Diagnosing and managing carcinoid heart disease in patients with neuroendocrine tumors: an expert statement.

       Operative Methods

      Almost all patients with carcinoid heart disease have some degree of tricuspid valve regurgitation, and tricuspid valve replacement is performed in the majority of patients.
      • Killu A.M.
      • Newman D.B.
      • Miranda W.R.
      • Maleszewski J.J.
      • Pellikka P.
      • Schaff H.V.
      • et al.
      Carcinoid heart disease without severe tricuspid valve involvement.
      In our earlier experience, the pulmonary valve was frequently excised at the time of tricuspid valve replacement. Although pulmonary valve regurgitation was tolerated well early postoperatively, longitudinal follow-up of patients who had pulmonary valvectomy showed increased risk of progressive right ventricular enlargement.
      • Connolly H.M.
      • Schaff H.V.
      • Mullany C.J.
      • Abel M.D.
      • Pellikka P.A.
      Carcinoid heart disease: impact of pulmonary valve replacement in right ventricular function and remodeling.
      Therefore, it is our practice to replace the pulmonary valve in patients with pulmonary valve regurgitation.
      After median sternotomy, we use bicaval cannulation with direct cannulation of the superior vena cava for cardiopulmonary bypass (Video 1). The heart is arrested using cold blood cardioplegia. After snaring the caval cannulas to isolate the right atrium, we make a longitudinal atriotomy and inspect the septum for a defect or patent foramen ovale, which is then closed. Next, the anterior and posterior tricuspid valve leaflets are excised, and the septal leaflet and chordal attachments are left in place. After calibrating the annulus, valve sutures are placed through leaflet tissue remnants in the area of the penetrating bundle of His, and the crossclamp is removed and sinus rhythm restored to allow the heart to beat during the remainder of the procedure.
      Figure thumbnail fx2
      Video 1Intraoperative video of a patient with carcinoid heart disease who underwent tricuspid and pulmonary valve replacement. Video available at: https://www.jtcvs.org/article/S0022-5223(18)32526-1/fulltext.
      For pulmonary valve replacement, an incision in the pulmonary artery begins 3 to 4 cm distal to the valve annulus and is carried across the valve onto the right ventricular outflow tract. Remnants of pulmonary valve cusps are excised, and a large oval autologous pericardial patch is sewn to the pulmonary arteriotomy with 4-0 polypropylene suture. The augmented valve annulus is calibrated, and the appropriate size prosthesis is secured with running 3-0 polypropylene suture that anchors the valve to the native annulus posteriorly and to the pericardial patch anteriorly. The remainder of the pericardial patch is then used to close the ventriculotomy.

       Postoperative Management

      The intravenous octreotide infusion is continued in the early postoperative phase until the patient is stable. Anticoagulation with warfarin is instituted in all patients with a target international normalized ratio of 2.5 (range, 2-3)
      • Whitlock R.P.
      • Sun J.C.
      • Fremes S.E.
      • Rubens F.D.
      • Teoh K.H.
      Antithrombotic and thrombolytic therapy for valvular disease: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest physicians evidence-based clinical practice guidelines.
      and is continued for 3 to 6 months for those with biological tissue valves and indefinitely for those receiving mechanical prostheses. Patients should undergo continued surveillance by an experienced cardiology team, with frequent echocardiographic assessment every 3 to 12 months after operation.
      • Davar J.
      • Connolly H.M.
      • Caplin M.E.
      • Pavel M.
      • Zacks J.
      • Bhattacharyya S.
      • et al.
      Diagnosing and managing carcinoid heart disease in patients with neuroendocrine tumors: an expert statement.

       Statistical Analysis

      Categorical data are presented as frequencies and percentages, and continuous variables are expressed as medians and interquartile range. Unadjusted survival time was estimated by Kaplan–Meier analysis, with differences determined using log-rank test. In the subset of patients with follow-up of 30 days or more, cumulative incidence of reintervention was estimated in a competing risk analysis in which death before reoperation was treated as a competing event. Prespecified preoperative variables were assessed as potential predictors of long-term mortality with multivariable Cox proportional hazards regression. Logistic regression analysis was used to analyze the association of the same factors with early mortality, although the number of 30-day deaths was too small to support a multivariable analysis. In all models, continuous predictors were expanded into multiple terms using restricted cubic splines to allow for nonlinear effects. All statistical analyses were carried out using SAS software (Version 9.4; SAS Institute Inc, Cary, NC).

      Results

       Baseline Characteristics and Preoperative Echocardiography

      The baseline characteristics of the study population are shown in Table 1. For these 240 patients, the median (interquartile range) age was 63 years (55-69), and 117 (49%) were male. New York Heart Association (NYHA) class III or IV dyspnea was present in 157 patients (70%). Loop diuretic therapy to control edema was used in 125 patients (52%). Somatostatin analog therapy was used in 221 patients (92%); 108 patients (45%) received short-acting somatostatins, and from 1999 onward, when long-acting somatostatins became available, 130 (72%) of 180 patients had long-acting somatostatin therapy. Markers of renal and hepatic function were normal in most patients.
      Table 1Baseline characteristics
      VariableMedian or nIQR or percentage
      Age (y)63(55-69)
      Male11749%
      BMI (kg/m2)24(22-27)
      BSA (m2)1.8(1.7-2.0)
      Smoker9038%
      Diabetes2410%
      NYHA class
       I73%
       II5926%
       III12255%
       IV3516%
      Chemotherapy6125%
      Loop diuretics12552%
       Dose of loop diuretics (mg)40(20-60)
      Somatostatin therapy22192%
       Short-acting somatostatins10845%
       Long-acting somatostatins130/18072%
      Creatinine (mg/dL)1.2(1.0-1.4)
      AST (IU/L)30(23-38)
      ALP (IU/L)209(134-355)
      Total bilirubin (mg/dL)1.0(0.7-1.5)
      INR1.2(1.1-1.3)
      Year of surgery2005(1999-2011)
      IQR, Interquartile range; BMI, body mass index; BSA, body surface area; NYHA, New York Heart Association; AST, aspartate transaminase; ALP, alkaline phosphatase; INR, international normalized ratio.
      Preoperative left ventricular ejection fraction was 61% (56-66). Tricuspid valve regurgitation was present in almost all patients, and 231 patients (97%) had severe valve regurgitation on preoperative echocardiography. Severe pulmonary valve regurgitation was seen in 157 patients (66%). Left-sided valves were less commonly affected, and severe regurgitation of the mitral or aortic valve was present in 15 patients (6%) and 17 patients (7%), respectively.

       Operative Details and Postoperative Complications

      As shown in Table 2, tricuspid valve replacement was performed in 237 patients (99%) and bioprosthetic valves were used in 201 patients (85%). Pulmonary valve operation was performed in 202 patients (84%). Operation on left-sided valves was performed in 38 patients (16%), and in 30 of these patients (79%), the mitral or aortic valve dysfunction was due to carcinoid plaque. In these 38 patients with left-sided valve surgery, the atrial septum was intact in 25 (66%). Median crossclamp time was 11 (7-34) minutes.
      Table 2Operative details and postoperative complications
      Variablen (%)
      Tricuspid valve replacement237 (99)
       Mechanical valve36 (15)
       Tissue valve201 (85)
      Pulmonary valve operation202 (84)
       Excision59 (29)
       Valve replacement143 (71)
      Mitral valve operation26 (11)
       Repair9 (35)
       Replacement17 (65)
      Aortic valve operation25 (10)
       Repair6 (24)
       Replacement19 (76)
      Left-sided operation (mitral or aortic valve)38 (16)
      No. of valves operated on
       Single32 (13)
       Double175 (73)
       Triple24 (10)
       Quadruple9 (4)
      ASD/PFO closure57 (24)
      Tumor excision6 (3)
      Reoperation27 (11)
       Reoperation due to bleeding18 (8)
      Postoperative permanent pacemaker implantation14 (6)
      Early (30-d) mortality22 (9)
      ASD, Atrial septal defect; PFO, patent foramen ovale.
      Among the 26 patients (11%) who underwent concomitant mitral valve operations, prosthetic replacement was performed in 17 patients (65%) who had carcinoid involvement of leaflet tissue. Mitral regurgitation was thought to be due to degenerative etiologies in 9 patients (35%) who had repair. Likewise, 25 patients (10%) required aortic valve intervention, of whom 19 (76%) had valve replacements. Most patients (n = 175; 73%) required double valve surgery, but triple and quadruple valve surgery were performed as necessary in 24 patients (10%) and 9 patients (4%), respectively. At the time of surgery, an atrial septal defect or patent foramen ovale was closed in 57 patients (24%). Cardiac metastases were excised in 6 patients (3%) (Figure 1, C and D).

       Early Mortality and Morbidity

      Early mortality, defined as death within 30 days of operation, was 9% (22 patients) over the course of the study. Early mortality decreased as time progressed and as operative experience increased (Figure 2). For operations performed in our earlier experience (1985 to 1994), early mortality was 29% (9/22). However, the rate decreased to 7% (6/81) over the next 10 years (1995 to 2004) and to 5% (7/128) from 2005 onward. A more recent year of surgery correlated with older patient age (Spearman's correlation rs = 0.20, P = .002) and with lower baseline levels of total bilirubin (rs = −0.19, P = .004) and creatinine (rs = −0.17, P = .01).
      Figure thumbnail gr2
      Figure 2Trend in early operative mortality over study period.
      On univariate analysis (Table E1), older age and higher NYHA class were associated with early mortality (odds ratio [OR], 2.31, 95% confidence interval [CI], 1.30-4.11, P = .008 and OR, 2.67, 95% CI, 1.31-5.45, P = .007, respectively). Use of long-acting somatostatins appeared to have some protective effects on early mortality, although this result was not statistically significant (OR, 0.45; 95% CI, 0.18-1.12; P = .085). The number of valves involved and presence of left-sided valve disease did not influence early mortality.
      Eighteen patients (8%) had to be taken back to the operating room because of postoperative bleeding. Fourteen patients (6%) required permanent pacemaker insertion for complete heart block (n = 12) or sustained sinus bradycardia (n = 2).

       Late Survival and Functional Outcomes

      Figure 3 shows survival over time for the entire study population, with survival estimates at 1, 3, and 5 years of 69%, 48%, and 34%, respectively. Survival was significantly better in patients with lower preoperative NYHA class (P < .001). On univariate analysis, older age, higher NYHA class, and a nonlinear effect of creatinine were associated with increased overall mortality. In a multivariable model, these factors remained independently predictive of mortality (Table 3). Use of short- or long-acting somatostatins was not associated with overall mortality. Among the late deaths in which cause was known, 80% died of noncardiac conditions, primarily tumor progression.
      Figure thumbnail gr3
      Figure 3A, Overall survival curve. B, Survival stratified by preoperative NYHA class. NYHA, New York Heart Association.
      Table 3Predictors of overall mortality on univariate and multivariable analysis
      VariableComparison
      Effects of predictor variables were estimated with unadjusted and adjusted hazard ratios (and 95% confidence limits) from univariable and multivariable models, respectively. Unless noted otherwise, continuous variables are reported as interquartile range hazard ratios (comparing the 75th percentile with the 25th percentile of the predictor). Because these variables are modeled nonlinearly, confidence limits of the hazard ratio are not necessarily consistent with general tests of association. That is, inclusion of 1.0 in the interval does not imply nonsignificance, and conversely, it is possible to have intervals excluding 1.0 without having evidence of an overall association.
      Univariate analysisMultivariable analysis
      HR (95% CI)P valueAdjusted HR (95% CI)
      A multivariable Cox regression model was fitted for all-cause mortality on n = 214 who had complete data for predictor variables shown in the table.
      P value
      Age
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      69:551.37 (1.10-1.69).0021.55 (1.22-1.97)<.001
      SexMale:Female1.22 (0.92-1.63).1711.12 (0.77-1.64).548
      BMI
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      26.5:21.70.87 (0.72-1.05).3050.90 (0.71-1.15).556
      Short-acting somatostatinYes:No1.04 (0.78-1.40).7741.57 (1.00-2.49).052
      Long-acting somatostatinYes:No1.17 (0.87-1.58).3051.57 (0.97-2.56).069
      Loop diureticsYes:No1.22 (0.91-1.63).1771.13 (0.80-1.57).489
      No. of valves operated on
      Because of data discreteness, a linear relationship was assumed for numeric variable (1 d.f. test).
      4:11.60 (0.77-3.33).2101.06 (0.34-3.38).916
      Left-sided operationYes:No1.29 (0.88-1.90).1951.52 (0.81-2.85).190
      NYHA class
      Because of data discreteness, a linear relationship was assumed for numeric variable (1 d.f. test).
      III:II1.48 (1.18-1.85).0011.53 (1.15-2.03).004
      AST
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      37.5:230.98 (0.75-1.28).1851.15 (0.84-1.55).196
      Total bilirubin
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      1.5:0.71.09 (0.82-1.45).8331.05 (0.76-1.46).892
      Creatinine
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      1.4:1.00.98 (0.79-1.22)<.0010.81 (0.63-1.06).013
      HR, Hazard ratio; CI, confidence interval; BMI, body mass index; NYHA, New York Heart Association; AST, aspartate transaminase.
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects and assessed for general association based on a 2 d.f. test.
      Because of data discreteness, a linear relationship was assumed for numeric variable (1 d.f. test).
      Effects of predictor variables were estimated with unadjusted and adjusted hazard ratios (and 95% confidence limits) from univariable and multivariable models, respectively. Unless noted otherwise, continuous variables are reported as interquartile range hazard ratios (comparing the 75th percentile with the 25th percentile of the predictor). Because these variables are modeled nonlinearly, confidence limits of the hazard ratio are not necessarily consistent with general tests of association. That is, inclusion of 1.0 in the interval does not imply nonsignificance, and conversely, it is possible to have intervals excluding 1.0 without having evidence of an overall association.
      § A multivariable Cox regression model was fitted for all-cause mortality on n = 214 who had complete data for predictor variables shown in the table.
      Early functional outcomes were assessed at a median follow-up of 0.5 (0.3, 0.9) years after surgery in 107 patients. Subanalysis of these patients showed that preoperative NYHA class III or IV symptoms were present in 64%. Postoperatively, only 15% had NYHA class III limitation with no patients reporting NYHA class IV dyspnea (Figure E1).
      During early follow-up, hepatic operation was performed in 21 patients (9%). Eighteen patients had resection of hepatic metastases, and 3 patients had liver transplant.

       Cardiac Reintervention

      Cardiac reintervention was necessary in 20 patients (8%) (Table E2). Figure 4 shows the cumulative incidence of cardiac valve reintervention in the presence of the competing risk of death. Reasons for reintervention were native or prosthetic valve dysfunction in 16 patients and pericardiectomy for constrictive pericarditis in 4 patients. Of the 16 late valve interventions, 3 were percutaneous procedures, 1 was balloon dilatation, 1 was insertion of a Melody valve (Medtronic, Minneapolis, Minn) in the pulmonary bioprosthesis, and 1 was insertion of Melody valves in dysfunctional tricuspid and pulmonary valves. Among the patients who underwent late valve operations, 2 developed carcinoid involvement of native mitral or aortic valves. Late prosthetic valve dysfunction was due to carcinoid involvement of the bioprosthesis in 3 patients. All patients with valve thrombosis had tissue bioprostheses implanted at the time of initial surgery.
      Figure thumbnail gr4
      Figure 4Cumulative risk of cardiac valve reintervention.

      Discussion

      This study updates and expands our surgical experience with carcinoid heart disease over a 33-year period.
      • Connolly H.M.
      • Schaff H.V.
      • Abel M.D.
      • Rubin J.
      • Askew J.W.
      • Li Z.
      • et al.
      Early and late outcomes of surgical treatment in carcinoid heart disease.
      Early mortality declined in the 240 patients who underwent surgical valve replacement during this time frame, and risk of perioperative death was associated with older age and more advanced heart failure (higher NYHA class); although not significant, long-acting somatostatins may have some protective effect on early mortality. The study adds important information on durability of valve prostheses in patients with carcinoid heart disease.
      In previous reports, operative mortality for valve replacement in patients with carcinoid heart disease has ranged from 18% to 63%, but most of these series were small with fewer than 50 patients.
      • Castillo J.G.
      • Milla F.
      • Adams D.H.
      Surgical management of carcinoid heart valve disease.
      In the present series between 1985 and 1994, early mortality was 29%; this risk decreased to 7% between 1995 and 2004, and has been 5% in the 128 patients undergoing operation since 2005 (Figure 2). Recent studies have reported similar early risks, 5% to 10%.
      • Kuntze T.
      • Owais T.
      • Secknus M.-A.
      • Kaemmerer D.
      • Baum R.
      • Girdauskas E.
      Results of contemporary valve surgery in patients with carcinoid heart disease.
      • Mokhles P.
      • van Herwerden L.A.
      • de Jong P.L.
      • de Herder W.W.
      • Siregar S.
      • Constantinescu A.A.
      • et al.
      Carcinoid heart disease: outcomes after surgical valve replacement.
      Indeed, in current practice, multiple valve replacement for carcinoid heart disease has an operative risk similar to that for multiple valve disease of other etiologies (8%-13%).
      • Rankin J.S.
      • He X.
      • O'Brien S.M.
      • Jacobs J.P.
      • Welke K.F.
      • Filardo G.
      • et al.
      The Society of Thoracic Surgeons risk model for operative mortality after multiple valve surgery.
      It is notable that the number of valves replaced was not predictive of mortality in this surgical cohort.
      Decreasing early mortality may be due to multiple factors, including refinement in patient selection and improved overall surgical care. These measures, especially improved perioperative management, are difficult to quantify and were not captured in our database.
      However, increased early mortality was associated with older age and higher NYHA class at presentation (Table E1), suggesting that patients with carcinoid heart disease may benefit from earlier surgical intervention. Indeed, in our more recent experience, patients with carcinoid heart disease had less severe symptoms, and this accounts, in part, for better outcome. Some clinicians argue that operation should be performed before onset of significant functional impairment, and Bhattacharyya and colleagues
      • Bhattacharyya S.
      • Toumpanakis C.
      • Caplin M.E.
      • Davar J.
      Analysis of 150 patients with carcinoid syndrome seen in a single year at one institution in the first decade of the twenty-first century.
      • Bhattacharyya S.
      • Raja S.G.
      • Toumpanakis C.
      • Caplin M.E.
      • Dreyfus G.D.
      • Davar J.
      Outcomes, risks and complications of cardiac surgery for carcinoid heart disease.
      reported NYHA class I or II symptoms in 55% of 22 patients with carcinoid heart disease who underwent cardiac surgery at their institution. Nevertheless, optimal timing of surgery may be difficult to determine in these complex patients whose symptoms occur in combination with gastrointestinal/hepatic impairment. For example, it is often challenging to determine to what degree right heart failure is contributing to functional decline and ascites in a patient with carcinoid syndrome and extensive hepatic metastases.
      In our cohort, survival at 3 years was 48%, but was significantly better in patients with NYHA class I or II symptoms preoperatively (Figure 3). Overall survival of our surgical patients compares favorably to medically managed patients with carcinoid heart disease, in whom 3-year survival estimates of 31% have been reported.
      • Pellikka P.A.
      • Tajik A.J.
      • Khandheria B.K.
      • Seward J.B.
      • Callahan J.A.
      • Pitot H.C.
      • et al.
      Carcinoid heart disease. Clinical and echocardiographic spectrum in 74 patients.
      In most patients, late survival after valve replacement for carcinoid heart disease is limited by tumor progression. Despite limitations of overall survival in patients undergoing valve surgery for carcinoid heart disease, early functional outcomes were excellent (Figure E1). Symptoms from right-sided heart failure were greatly improved in the majority of patients, and enhanced quality of life should be considered in these patients with limited life expectancy.
      An important consideration for patients undergoing valve replacement is the choice of valve prosthesis. At the Mayo Clinic, we have generally preferred tissue valves for patients with carcinoid heart disease; bioprostheses were implanted in more than 80% of patients. Tissue valves are appropriate for most patients with carcinoid heart disease who have limited life expectancy. Exceptions to this practice include patients who require anticoagulation with warfarin for other reasons or patients with ovarian carcinoid tumors who have a favorable outlook after valve surgery and oophorectomy.
      • Chaowalit N.
      • Connolly H.M.
      • Schaff H.V.
      • Webb M.J.
      • Pellikka P.A.
      Carcinoid heart disease associated with primary ovarian carcinoid tumor.
      However, as seen in the present review, tissue bioprostheses were not free of risk from thrombosis; 3 patients required surgical reintervention because of thrombosis of a bioprosthetic valve. Thus, anticoagulation is advised for patients receiving tissue bioprostheses, and we suggest using warfarin for 3 to 6 months for patients in sinus rhythm and indefinitely for those with atrial fibrillation.
      • Davar J.
      • Connolly H.M.
      • Caplin M.E.
      • Pavel M.
      • Zacks J.
      • Bhattacharyya S.
      • et al.
      Diagnosing and managing carcinoid heart disease in patients with neuroendocrine tumors: an expert statement.
      • Egbe A.C.
      • Connolly H.M.
      • Pellikka P.A.
      • Schaff H.V.
      • Hanna R.
      • Maleszewski J.J.
      • et al.
      Outcomes of warfarin therapy for bioprosthetic valve thrombosis of surgically implanted valves: a prospective study.
      • Egbe A.
      • Pislaru S.V.
      • Ali M.A.
      • Khan A.R.
      • Boler A.N.
      • Schaff H.V.
      • et al.
      Early prosthetic valve dysfunction due to bioprosthetic valve thrombosis: the role of echocardiography.
      Recurrent carcinoid fibrosis of bioprostheses is another concern for patients undergoing valve replacement with tissue valves. In our series, this was documented by pathologic examination of valves from 3 patients (Table E2). Carcinoid recurrence seems to be a rare complication, but has been reported by van Rijswijk and colleagues.
      • van Rijswijk J.W.
      • Vink A.
      • van Herwerden L.A.
      • Kluin J.
      Recurrent carcinoid involvement of a tricuspid bioprosthesis.
      In the present series, carcinoid involvement of bioprostheses occurred between 3 and 8 years after initial surgery, but cause(s) of recurrence could not be ascertained from our data. Two of these 3 patients were treated with somatostatins preoperatively and remained on this medication after surgery; 24-hour urinary 5-hydroxyindoleacetic acid levels were elevated before reoperation in 1 of these patients. The other patient had no systemic symptoms from his metastatic carcinoid tumor and did not receive somatostatin therapy preoperatively or postoperatively. It is important to note that valve thrombosis and carcinoid recurrence occurred only in patients who received tissue bioprostheses, and thus mechanical heart valves may be appropriate for some patients, such as young patients in whom further surgical treatment of primary or metastatic tumor is not planned.
      • Lillegard J.B.
      • Fisher J.E.
      • Mckenzie T.J.
      • Que F.G.
      • Farnell M.B.
      • Kendrick M.L.
      • et al.
      Hepatic resection for the carcinoid syndrome in patients with severe carcinoid heart disease: does valve replacement permit safe hepatic resection?.
      In the present series, left-sided heart valves were replaced or repaired in 16% of patients. However, presence of left-sided disease did not appear to be associated with intracardiac shunts because 66% of these patients had an intact atrial septum. This result is consistent with those reported by Connolly and colleagues,
      • Connolly H.M.
      • Schaff H.V.
      • Mullany C.J.
      • Rubin J.
      • Abel M.D.
      • Pellikka P.A.
      Surgical management of left-sided carcinoid heart disease.
      who observed left-sided heart disease without shunts in 55% of their study population. Furthermore, presence of left-sided heart disease was not associated with early or overall mortality in our cohort. In our experience, surgery in patients with associated aortic or mitral valve regurgitation is safe and should be performed if clinically indicated.

       Study Limitations

      This is a retrospective study from a single center and may be influenced by referral bias. These surgical results also reflect the importance of an experienced team specializing in carcinoid heart disease and may not be replicable at other institutions.

      Conclusions

      Operation on patients with carcinoid valvular disease leads to excellent improvement in symptoms during early follow-up and has acceptable short-term mortality and complication rates. Early mortality has improved in recent experience, but overall survival is limited by tumor progression. Bioprostheses are appropriate for most patients because structural valve deterioration due to carcinoid involvement is rare.

       Conflict of Interest Statement

      Authors have nothing to disclose with regard to commercial support.

      Appendix

      Figure thumbnail fx3
      Figure E1Functional outcomes of 107 patients at median follow-up of 0.5 (0.3-0.9) years. NYHA, New York Heart Association.
      Table E1Predictors of early mortality on univariate analysis
      VariableComparison
      Effects are estimated as unadjusted ORs (and 95% CIs) from univariable logistic regression models. Unless noted otherwise, continuous variables are reported as IQR ORs (comparing the 75th percentile with the 25th percentile of the predictor). Because these variables are modeled nonlinearly, CIs of the ORs will not be entirely consistent with general tests of association, unlike the case with linear effects. That is, inclusion of 1.0 in the interval does not imply nonsignificance, and conversely, it is possible to have intervals excluding 1.0 without having evidence of a significant association.
      OR (95% CI)P value
      Age
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      69:252.31 (1.30-4.11).008
      SexMale:Female0.86 (0.36-2.08).746
      BMI
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      26.5:21.71.04 (0.60-1.78).826
      Short-acting somatostatinsYes:No1.87 (0.77-4.56).169
      Long-acting somatostatinsYes:No0.45 (0.18-1.12).085
      Loop diureticsYes:No1.69 (0.68-4.18).259
      No. of valves operated on
      Because of data discreteness, a linear relationship was assumed for the numeric variable (1 d.f. test).
      1-level increase1.01 (0.50-2.05).976
      Left-sided operationYes:No0.83 (0.23-2.94).767
      NYHA class
      Because of data discreteness, a linear relationship was assumed for the numeric variable (1 d.f. test).
      1-level increase2.67 (1.31-5.45).007
      AST
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      38:231.87 (1.06-3.30).067
      Total bilirubin
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      1.5:0.71.15 (0.47-2.80).559
      Creatinine
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      1.4:1.01.14 (0.58-2.25).906
      OR, Odds ratio; CI, confidence interval; BMI, body mass index; NYHA, New York Heart Association; AST, aspartate transaminase.
      Continuous variables were expanded into multiple terms using restricted cubic splines (using 3 knots) to allow for nonlinear effects, and assessed for general association based on a 2 d.f. test.
      Because of data discreteness, a linear relationship was assumed for the numeric variable (1 d.f. test).
      Effects are estimated as unadjusted ORs (and 95% CIs) from univariable logistic regression models. Unless noted otherwise, continuous variables are reported as IQR ORs (comparing the 75th percentile with the 25th percentile of the predictor). Because these variables are modeled nonlinearly, CIs of the ORs will not be entirely consistent with general tests of association, unlike the case with linear effects. That is, inclusion of 1.0 in the interval does not imply nonsignificance, and conversely, it is possible to have intervals excluding 1.0 without having evidence of a significant association.
      Table E2Details of cardiac valve reintervention
      PatientYear of initial surgeryTime to reintervention (y)Initial procedure on the valve requiring reinterventionReintervention procedureAdditional details
      119920.9Pulmonary homograftPercutaneous balloon dilatation of pulmonary valvePulmonary artery rupture after percutaneous procedure
      219942.6Tricuspid bioprosthesisTricuspid valve replacementCarcinoid recurrence of tricuspid bioprosthesis
      319950.5Tricuspid and pulmonary bioprosthesesTricuspid and pulmonary valve replacementThrombosis of both bioprostheses
      419960.9Tricuspid bioprosthesis, pulmonary homograftTricuspid and pulmonary valve replacementTricuspid valve thrombosis, pulmonary homograft stenosis
      5199714.5Native mitral valve, tricuspid mechanical prosthesisMitral and tricuspid valve replacementMitral valve stenosis (late carcinoid involvement)
      6199712.7Pulmonary valvectomyPulmonary valve replacementSevere pulmonary valve regurgitation (s/p valvectomy)
      719975.0Native aortic valve, pulmonary valvectomyAortic and pulmonary valve replacementAortic valve regurgitation (late carcinoid involvement), severe pulmonary valve regurgitation (s/p valvectomy)
      8199713.6Tricuspid bioprosthesis, pulmonary valvectomyTricuspid and pulmonary valve replacementThrombosis of tricuspid bioprosthesis, severe pulmonary valve regurgitation (s/p valvectomy)
      920018.6Tricuspid and pulmonary bioprosthesesTricuspid and pulmonary valve replacementCarcinoid recurrence of both bioprostheses
      1020017.8Tricuspid and pulmonary bioprosthesesTricuspid and pulmonary valve replacementCarcinoid recurrence of both bioprostheses
      11200210.2Tricuspid and pulmonary bioprosthesesTricuspid and pulmonary valve replacementUnknown etiology, reintervention at outside hospital
      1220048.2Pulmonary bioprosthesisPercutaneous pulmonary valve replacementSevere pulmonary valve regurgitation
      1320044.0Tricuspid bioprosthesisTricuspid valve replacementThrombosis of tricuspid bioprosthesis
      1420075.7Tricuspid and pulmonary bioprosthesesPercutaneous tricuspid and pulmonary valve replacementSevere tricuspid and pulmonary valve regurgitation
      1520091.9Tricuspid bioprosthesisTricuspid valve replacementPannus overgrowth of tricuspid bioprosthesis cusps
      1620152.0Tricuspid and pulmonary bioprosthesesTricuspid and pulmonary valve replacementStructural valve deterioration due to infective endocarditis

      Supplementary Data

      References

        • Pellikka P.A.
        • Tajik A.J.
        • Khandheria B.K.
        • Seward J.B.
        • Callahan J.A.
        • Pitot H.C.
        • et al.
        Carcinoid heart disease. Clinical and echocardiographic spectrum in 74 patients.
        Circulation. 1993; 87: 1188-1196
        • Lundin L.
        • Norheim I.
        • Landelius J.
        • Oberg K.
        • Theodorsson-Norheim E.
        Carcinoid heart disease: relationship of circulating vasoactive substances to ultrasound-detectable cardiac abnormalities.
        Circulation. 1988; 77: 264-269
        • Bhattacharyya S.
        • Toumpanakis C.
        • Caplin M.E.
        • Davar J.
        Analysis of 150 patients with carcinoid syndrome seen in a single year at one institution in the first decade of the twenty-first century.
        Am J Cardiol. 2008; 101: 378-381
        • Davar J.
        • Connolly H.M.
        • Caplin M.E.
        • Pavel M.
        • Zacks J.
        • Bhattacharyya S.
        • et al.
        Diagnosing and managing carcinoid heart disease in patients with neuroendocrine tumors: an expert statement.
        J Am Coll Cardiol. 2017; 69: 1288-1304
        • Fox D.J.
        • Khattar R.S.
        Carcinoid heart disease: presentation, diagnosis, and management.
        Heart. 2004; 90: 1224-1228
        • Connolly H.M.
        • Schaff H.V.
        • Mullany C.J.
        • Rubin J.
        • Abel M.D.
        • Pellikka P.A.
        Surgical management of left-sided carcinoid heart disease.
        Circulation. 2001; 104: I36-I40
        • Arghami A.
        • Connolly H.M.
        • Abel M.D.
        • Schaff H.V.
        Quadruple valve replacement in patients with carcinoid heart disease.
        J Thorac Cardiovasc Surg. 2010; 140: 1432-1434
        • Connolly H.M.
        • Schaff H.V.
        • Abel M.D.
        • Rubin J.
        • Askew J.W.
        • Li Z.
        • et al.
        Early and late outcomes of surgical treatment in carcinoid heart disease.
        J Am Coll Cardiol. 2015; 66: 2189-2196
        • Lillegard J.B.
        • Fisher J.E.
        • Mckenzie T.J.
        • Que F.G.
        • Farnell M.B.
        • Kendrick M.L.
        • et al.
        Hepatic resection for the carcinoid syndrome in patients with severe carcinoid heart disease: does valve replacement permit safe hepatic resection?.
        J Am Coll Surg. 2011; 213: 130-138
        • Weingarten T.N.
        • Abel M.D.
        • Connolly H.M.
        • Schroeder D.R.
        • Schaff H.V.
        Intraoperative management of patients with carcinoid heart disease having valvular surgery: a review of one hundred consecutive cases.
        Anesth Analg. 2007; 105: 1192-1199
        • Møller J.E.
        • Pellikka P.A.
        • Bernheim A.M.
        • Schaff H.V.
        • Rubin J.
        • Connolly H.M.
        Prognosis of carcinoid heart disease: analysis of 200 cases over two decades.
        Circulation. 2005; 112: 3320-3327
        • Strosberg J.
        • Weber J.
        • Feldman M.
        • Goldman J.
        • Almhanna K.
        • Kvols L.
        Above-label doses of octreotide-LAR in patients with metastatic small intestinal carcinoid tumors.
        Gastrointest Cancer Res. 2013; 6: 81-85
        • Ramage J.K.
        • Ahmed A.
        • Ardill J.
        • Bax N.
        • Breen D.J.
        • Caplin M.E.
        • et al.
        Guidelines for the management of gastroenteropancreatic neuroendocrine (including carcinoid) tumours (NETs).
        Gut. 2012; 61: 6-32
        • Killu A.M.
        • Newman D.B.
        • Miranda W.R.
        • Maleszewski J.J.
        • Pellikka P.
        • Schaff H.V.
        • et al.
        Carcinoid heart disease without severe tricuspid valve involvement.
        Cardiology. 2016; 133: 217-222
        • Connolly H.M.
        • Schaff H.V.
        • Mullany C.J.
        • Abel M.D.
        • Pellikka P.A.
        Carcinoid heart disease: impact of pulmonary valve replacement in right ventricular function and remodeling.
        Circulation. 2002; 106: I51-I56
        • Whitlock R.P.
        • Sun J.C.
        • Fremes S.E.
        • Rubens F.D.
        • Teoh K.H.
        Antithrombotic and thrombolytic therapy for valvular disease: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest physicians evidence-based clinical practice guidelines.
        Chest. 2012; 141: e576S-e600S
        • Castillo J.G.
        • Milla F.
        • Adams D.H.
        Surgical management of carcinoid heart valve disease.
        Semin Thorac Cardiovasc Surg. 2012; 24: 254-260
        • Kuntze T.
        • Owais T.
        • Secknus M.-A.
        • Kaemmerer D.
        • Baum R.
        • Girdauskas E.
        Results of contemporary valve surgery in patients with carcinoid heart disease.
        J Heart Valve Dis. 2016; 25: 356-363
        • Mokhles P.
        • van Herwerden L.A.
        • de Jong P.L.
        • de Herder W.W.
        • Siregar S.
        • Constantinescu A.A.
        • et al.
        Carcinoid heart disease: outcomes after surgical valve replacement.
        Eur J Cardiothorac Surg. 2012; 41: 1278-1283
        • Rankin J.S.
        • He X.
        • O'Brien S.M.
        • Jacobs J.P.
        • Welke K.F.
        • Filardo G.
        • et al.
        The Society of Thoracic Surgeons risk model for operative mortality after multiple valve surgery.
        Ann Thorac Surg. 2013; 95: 1484-1490
        • Bhattacharyya S.
        • Raja S.G.
        • Toumpanakis C.
        • Caplin M.E.
        • Dreyfus G.D.
        • Davar J.
        Outcomes, risks and complications of cardiac surgery for carcinoid heart disease.
        Eur J Cardiothorac Surg. 2011; 40: 168-172
        • Chaowalit N.
        • Connolly H.M.
        • Schaff H.V.
        • Webb M.J.
        • Pellikka P.A.
        Carcinoid heart disease associated with primary ovarian carcinoid tumor.
        Am J Cardiol. 2004; 93: 1314-1315
        • Egbe A.C.
        • Connolly H.M.
        • Pellikka P.A.
        • Schaff H.V.
        • Hanna R.
        • Maleszewski J.J.
        • et al.
        Outcomes of warfarin therapy for bioprosthetic valve thrombosis of surgically implanted valves: a prospective study.
        JACC Cardiovasc Interv. 2017; 10: 379-387
        • Egbe A.
        • Pislaru S.V.
        • Ali M.A.
        • Khan A.R.
        • Boler A.N.
        • Schaff H.V.
        • et al.
        Early prosthetic valve dysfunction due to bioprosthetic valve thrombosis: the role of echocardiography.
        JACC Cardiovasc Imaging. 2018; 11: 951-958
        • van Rijswijk J.W.
        • Vink A.
        • van Herwerden L.A.
        • Kluin J.
        Recurrent carcinoid involvement of a tricuspid bioprosthesis.
        Eur J Cardiothorac Surg. 2017; 51: 1015