Tricuspid Repair for Infective Endocarditis

Abstract

We report our retrospective experience in the treatment of infective tricuspid endocarditis with valve repair. From January 1981 through January 1999, 238 cases of infective endocarditis were seen at our institution, with tricuspid involvement in 19 cases. Tricuspid valve repair was performed in 9 patients whose valves had infective lesions involving a single leaflet. One goal of the repair was to avoid implanting any prosthetic material.

At surgery, the posterior leaflet was completely excised and annuloplasty was performed in 4 patients. Wide quadrangular resection of the anterior leaflet and De Vega annuloplasty were performed in the other 5 patients. All patients had a good postoperative recovery. Postoperative echocardiography showed no tricuspid regurgitation in 4 patients, mild regurgitation in 3, and moderate in 2. Follow-up ranged from 21 to 155 months (mean, 47.56 ± 50 [SD] months). Two late deaths occurred: one, 2 months postoperatively (sudden death), and the other, 108 months postoperatively (lung carcinoma). Late postoperative echocardiography showed no tricuspid regurgitation in 4 patients, mild in 2, and moderate in 2. No recurrent infection was observed.

Tricuspid valve repair rather than valvulectomy or replacement is indicated in cases of right-sided endocarditis with single-leaflet involvement. Tricuspid repair enables eradication of the infection without implantation of prosthetic material.

Reports of tricuspid endocarditis have increased in frequency during the last 2 decades. ^1,2 The growing number of patients addicted to intravenous (IV) drugs and those with long-term IV catheters or with antiarrhythmic devices, such as implantable defibrillators and pacemakers, have considerably increased the number of patients at risk of right-sided endocarditis.

Although the 1st surgical treatment of right-sided endocarditis was reported in 1961, the treatment for tricuspid valve endocarditis is still a matter of debate. ³ Surgical options for patients with tricuspid endocarditis vary from complete excision of the valve to tricuspid valve replacement with a prosthesis. ^2,4,5 Valve excision results in massive tricuspid regurgitation, limiting the quality of life and leading, in more than 20% of patients, to reoperation in order to reduce the degree of regurgitation with a prosthesis. ^6,7

Valve replacement with either a mechanical or a biological prosthesis ^4,8 exposes patients to valve-related complications and to the risk of recurrent endocarditis —especially patients who are addicted to IV drugs. ⁴ Furthermore, valve replacement in patients with pacemakers may preclude the possibility of endocardial pacing, which is still the most effective method of cardiac stimulation. ⁹

Many authors have published their experiences with tricuspid repair in patients with infective endocarditis, reporting high rates of surgical cure, ¹⁰ good hemodynamic results, ^11,12 and good long-term survival. ^13–16 Frequently, the techniques require the use of prosthetic materials such as glutaraldehyde-tanned pericardium, polytetrafluoroethylene (PTFE) chordae, and annuloplasty rings. 11,12,14–16

To avoid the use of foreign materials, thus reducing the incidence of recurrent infection and obviating the need for long-term anticoagulation, we performed a repair consisting of leaflet resection and annuloplasty without any prosthetic material, either biologic or synthetic. We report herein the early and late results of our retrospective experience with tricuspid repair for infective endocarditis.

Patients and Methods

From January 1981 through January 1999, 238 patients with valvular endocarditis underwent surgical treatment at our institution. Tricuspid valve involvement was observed in 19 patients (8%), 9 of whom underwent tricuspid valve repair (for preoperative data, see Table I). Our procedure was contraindicated in the other 10 patients, because 2 leaflets or the entire valve was infected.

TABLE I. Preoperative Clinical Features, Echocardiographic Findings, and Bacteriologic Data

Of the 9 patients who underwent tricuspid valve repair, 7 were men and 2 were women, with ages ranging from 22 to 68 years (mean, 44.2 ± 14.3 [SD] years). Causes of endocarditis were IV drug abuse in 3, pacemaker leads in 2, aortic valve endocarditis in 2, congenital lesion in 1 (ostium primum atrial septal defect), and unknown in 1. Factors predisposing 3 of the patients to endocardial infections were mitral prosthetic endocarditis in 1, dental infections in 1, and lower limb ulcers due to rachischisis in 1.

Blood cultures showed Staphylococcus aureus in 3 patients and S. epidermidis in 1. In 5 patients, there was no bacterial growth, the absence of which was most likely due to the long-term antibiotic therapy administered in the referring hospital before the patients' admission to our institution.

Preoperatively, 7 patients were in New York Heart Association (NYHA) functional class III and 2 were in class II. Indications for surgery were recurrent pulmonary embolism in 4 patients, cardiac failure in 4, and persistence of fever despite appropriate antibiotic therapy in 1.

Standard preoperative transthoracic echocardiography (TTE) was performed in each patient. Diagnosis of tricuspid infective endocarditis was made according to criteria established by the Duke Endocarditis Service, ¹⁷ and the amount of regurgitant flow was evaluated from backflow extension within the inferior vena cava on color-flow Doppler echocardiography in 8 patients (Fig. 1).

Fig. 1 Top: Preoperative color-flow Doppler echocardiogram, left parasternal view, showing a case of severe tricuspid regurgitation due to anterior leaflet prolapse. Bottom: Post-operative echocardiogram, left parasternal view, showing a residual, mild degree of tricuspid regurgitation after anterior leaflet resection and De Vega annuloplasty.

LA = left atrium; LV = left ventricle; RA = right atrium; RV = right ventricle

All patients were found to have moderate-to-severe tricuspid regurgitation, which was due to anterior leaflet prolapse in 4 cases (Fig. 1), posterior leaflet prolapse in 2 (Fig. 2), flail posterior leaflet in 2, and flail anterior leaflet in 1. Additionally, annular dilatation was present in 3 cases. In all patients, vegetations were seen, localized on the anterior leaflet in 5 patients and on the posterior leaflet in 4. Evidence of right ventricular volume overload was found in each case. Right ventricular end-diastolic diameter was measured in the parasternal long-axis view. Right ventricular dilatation was considered mild for end–diastolic diameters up to 3.0 cm, moderate for diameters from 3.0 to 3.5 cm, and severe for diameters measuring more than 3.5 cm. Right ventricular dilatation was mild in 1 patient, moderate in 4, and severe in 4.

Fig. 2 Top: Preoperative echocardiogram, left parasternal view, showing posterior leaflet prolapse in a case of tricuspid endocarditis. Bottom: Postoperative echocardiogram, 4-chamber view, showing the absence of tricuspid prolapse after posterior leaflet resection and Kay annuloplasty.

LV = left ventricle; PL = posterior leaflet; RA = right atrium; RV = right ventricle; TV = tricuspid valve

The operative approach was through a standard median sternotomy incision. All patients underwent valve repair with aortic cross-clamping and hypothermic arrest with use of cold crystalloid cardioplegic solution. In 2 patients, aortic valve replacement with a bioprosthesis was performed. Both patients had aortic vegetations and a tear in the noncoronary aortic cusp. Another patient with a mechanical prosthetic valve in the mitral position had a septic mitral peri-valvular leak requiring prosthesis excision and replacement with another bileaflet valve.

Surgical inspection of the tricuspid valve confirmed preoperative echocardiographic findings in all cases. Patients with vegetations and other signs of infection localized on a single leaflet were considered candidates for valve repair; multileaflet involvement was considered a contraindication to reparative procedures. Our policy was to perform wide-margin resection of the vegetation area and to restore tricuspid valve competence without any prosthetic tissue, either autologous or artificial. To achieve leaflet coaptation and stability of the repair over time, we performed annuloplasty in all cases. When vegetations involved the posterior portion of the tricuspid valve, complete leaflet excision and Kay annuloplasty ¹⁸ were performed. If the veg–etations were localized on the anterior leaflet, wide quadrangular resection, including the vegetations, was performed. In 2 patients, more than three quarters of the anterior leaflet was resected and the leaflet margins were approximated with simple interrupted 5-0 Prolene sutures. In all cases of anterior leaflet resection, a De Vega annuloplasty ¹⁹ was performed with a 3-0 Prolene suture and autologous pericardium pledgets at each end. Resected valvular tissue was sent to the laboratory for histologic examination and culture. Intraoperative testing of tricuspid valve competence was performed by injecting cold saline solution into the right ventricle.

Parenteral antibiotic therapy was continued for 4 weeks postoperatively, with use of specific antibiotics in those patients who had positive blood or valve cultures, and broad-spectrum empirical therapy in other cases.

All patients returned to the outpatient clinic for follow-up: functional recovery was evaluated according to NYHA classification every 6 months, and TTE was performed yearly (Figs. 1 and 2). Right ventricular end-diastolic diameter and degree of tricuspid regurgitation were recorded and compared with preoperative data.

Results

Postoperative and follow-up results are presented in Table II.

TABLE II. Surgical Procedures, Outcome, and Follow-Up Data

Cultures of excised tricuspid leaflets showed bacterial growth in only 2 of the patients with positive blood cultures. Histologic examination confirmed the clinical diagnosis of infective endocarditis in all cases.

All patients survived the operation and were discharged from the hospital within 10 days postoperatively. None of the patients required postoperative therapy with diuretics. Predischarge postoperative echocardiography showed mild tricuspid regurgitation in 3 patients and moderate in 2. No tricuspid regurgitation was seen in the other 4 patients.

Follow-up ranged from 21 to 155 months (mean, 47.56 ± 50 months). One patient died suddenly, 2 months postoperatively, of unknown causes. He was in a rehabilitation program for drug addiction when he died. No postmortem examination was performed. Another patient died of small-cell lung carcinoma 9 years after the operation. The other 7 patients were still alive with good postoperative recovery at the end of the study.

One patient who underwent mitral valve replacement and tricuspid repair had a mitral perivalvular leak requiring reoperation 5 months after the 1st operation. Although no signs of endocarditis were noted, and the blood and valve cultures were sterile, periprosthetic leakage recurred 2 years later and required replacement with another mitral prosthesis. After the last operation, the patient remained asymptomatic: no recurrence of perivalvular leakage and only moderate tricuspid regurgitation were observed on the last echocardiogram, obtained nearly 13 years (155 months) after the initial operation.

The last postoperative echocardiographic studies showed no tricuspid regurgitation in 3 patients, mild in 2, and moderate in 3. No clinical or echocar–diographic signs of recurrent tricuspid endocarditis were observed in any patient. Right ventricular end–diastolic diameter measurements showed mild dilatation in 3 patients and moderate dilatation in 2 (Table II).

Discussion

The incidence of tricuspid valve endocarditis has risen dramatically during the last 2 decades. ^1,2 A review of the literature shows 3 reasons for this increase. 1) Successful cardiac surgery and interventional electrophysiology have considerably prolonged the survival of patients with congenital or acquired heart valve disease, as well as those with heart block or malignant tachyarrhythmias. ^1,2,11 2) Many patients with malignancies are treated with chemotherapy by means of long-term central venous catheters. 3) Intravenous heroin abuse predisposes IV drug users to tricuspid valve endocarditis. ^11,20,21

The surgical treatment of right-sided endocarditis remains an item of debate, but fortunately, most of these infections are caused by organisms that can be treated successfully with antibiotics. Tricuspid valve lesions can be treated medically, because the heart is less susceptible to injury from tricuspid regurgitation and pulmonary embolization than from lesions associated with left-sided endocarditis. ^7,22 For these reasons, associated right-sided endocarditis is probably more common than is recognized and is therefore under diagnosed. ²³ Nevertheless, routine echocardio-graphic investigation in patients with long-term right-heart catheters who have recurrent fever enables diagnosis of an increasing number of cases of tri–cuspid endocarditis. Finally, improved sensitivity of echocardiography in detecting intracardiac vegetations can lead to diagnosis of asymptomatic tricuspid involvement in patients who have left-sided endocarditis. ^24,25

The successful treatment of tricuspid endocarditis should include the excision of all infected tissue and restoration of valvular competence. Arbulu and associates ²¹ proposed tricuspid valve excision without replacement. Such a procedure allows complete removal of infected tissues; moreover, the absence of a tricuspid valve can be well tolerated by patients with normal pulmonary artery pressure. The procedure is indicated especially in patients addicted to IV drugs, in whom complete absence of a tricuspid valve can avoid recurrence of infection. However, about 25% of patients cannot tolerate tricuspid regurgitation and require a 2nd operation for tricuspid valve replacement. ^1,6,7 Further, in this group of patients, pulmonary hypertension may be present immediately, may develop secondary to multiple pulmonary emboli, or may occur later. Pulmonary hypertension may also lead to progressive right-heart failure in patients without a tricuspid valve. ⁷ Tricuspid valve excision without replacement is contraindicated in patients with left-sided endocarditis and high left atrial pressure.

Tricuspid replacement may be another therapeutic option. Although early reports of tricuspid valve replacement indicated a high incidence of valve-related complications including reinfection, heart block, prosthetic thrombosis, and poor hemodynamic performance, ²⁶ the development of bileaflet valves and low-profile porcine valves has greatly improved the prognosis of patients after tricuspid valve replacement. Transvalvular gradients are low when prostheses larger than 31 mm are implanted; the incidence of prosthetic thrombosis is low, even if mechanical valves are used; and calcific degeneration of the tricuspid bioprosthesis develops more slowly than it does in the mitral position. ⁴

The incidence of spontaneous recurrence of prosthetic endocarditis is very low in valves in the tricuspid position. ¹¹ This finding can be explained by the propensity of the infection to involve the free margin of the valve most often, and the annular region rarely. In addition, local factors, such as differences in the blood supply to the annular area and the surrounding myocardium, may account for the different manifestations of endocarditis among the mitral, aortic, and tricuspid valves.

Nevertheless, patients with tricuspid valve endocarditis have low compliance rates with regard to long-term anticoagulation, often because they are elderly and have pacemakers or because they are addicted to IV drugs. ^10,11 Moreover, these 2 groups of patients have a high tendency to become reinfected, due either to the presence of a transvenous lead or to the resumption of intravenous drug abuse. ^4,26

The best immediate and short-term solution to the problem of tricuspid valve endocarditis may be generous excision of valvular tissue when it has vegetations and reconstruction of the native valve without the use of prosthetic material. In our series, every attempt was made to repair the valve using existing tissue alone, rather than using a pericardial patch, ^10,11 which is nonviable tissue, or Gore-Tex® sutures (W.L. Gore & Associates, Inc.; Flagstaff, Ariz) for artificial chordae, ^12,15,16 which are foreign material.

Valve reconstruction after excision of one half to three quarters of the anterior tricuspid leaflet can be considerably more complex than reconstruction after excision of redundant leaflet tissue in mitral valve repair procedures. ¹¹ Indeed, the main drawback of this procedure may be the postoperative persistence of some degree of tricuspid valve regurgitation.

The posterior leaflet can be resected (as shown by Kay's group ¹⁸) and maintain good valve competence; however, when large amounts of anterior leaflet are excised, it becomes essential to retain a full posterior leaflet in order to provide sufficient leaflet surface area. De Vega annuloplasty achieves leaflet coaptation while avoiding excessive tension on the repaired valve; thus it is a useful adjunct to tricuspid valve repair. ¹⁹

The initial moderate insufficiency after tricuspid repair is well tolerated by the patients, and valvular competence generally improves over time. ¹¹ Residual regurgitation subsides with the resolution of right ventricular volume overload. However, tricuspid valve repair provides a somewhat lesser degree of regurgitation than does tricuspid valve excision. In some of our patients, the use of a rigid modeled ring might have achieved better valve competence but would have added a risk of reinfection.

In conclusion, resection of all infected tricuspid tissue and simple repair without prosthetic material achieves a high rate of success in patients who have infective endocarditis. Complete excision of the posterior leaflet or resection of almost all of the anterior leaflet can be accomplished while maintaining good valvular function. Cases of endocarditis involving 2 tricuspid leaflets or the whole valve are best treated with tricuspid valve replacement.