Surgical Intervention for Infective Endocarditis in Infancy and Childhood

Transcription

1 Surgical Intervention for Infective Endocarditis in Infancy and Childhood Fumikazu Nomura, MD, Daniel J. Penny, MD, Samuel Menahem, MD, Ash Pawade, FRCS, and Tom R. Karl, MD Victorian Paediatric Cardiac Surgery Unit and Department of Cardiology, Royal Children's Hospital, Melbourne, Australia Background. Infective endocarditis is an uncommon but serious disease in children. Optimal treatment strategy, especially surgical indications, continues to evolve. Methods. Retrospective review of 98 patients treated for infective endocarditis during the past 13 years at the Royal Children's Hospital, including medically and surgically treated patients. Results. Thirty of 98 patients had surgical intervention with 6.7% hospital mortality, and 76% survival probabil- ity at 45 months. The remaining patients were treated medically, with 10% hospital mortality and 52% 5-year survival probability. The incidence of structural heart disease, congestive heart failure, and spectrum of organisms was similar in the two groups. Conclusions. Despite advances in antibiotic therapy, early surgical intervention is required in a significant subset. Concurrent intracardiac repair may be appropriate. (Ann Thorac Surg 1995;60:90-5) I nfective endocarditis is an uncommon but serious disease in children, accounting for approximately I in 4,500 admissions to pediatric hospitals [1]. The association between infective endocarditis and structural congenital heart disease has been well described [2-4]. Although advances in antibiotic therapy and in microbiological isolation techniques have increased the likelihood of successful medical management, there remains a group of patients in whom acute surgical intervention is required. The indications for surgical intervention have been well documented in the literature pertaining to adults [5, 6], but less information is available regarding pediatric patients. The aim of this review of 98 children who underwent treatment for endocarditis over a period of 13 years is to present an analysis of our experience, including the indications and outcome for surgical intervention in our institution. Material and Methods The hospital records, operative notes, echocardiographic reports, and cineangiograms of all children who were treated for infective endocarditis at the Royal Children's Hospital, Melbourne, between January 1971 and May 1994, were reviewed. The diagnosis of infective endocarditis was based on the presence of one of the following: (1) persistent fever, with vegetations seen on echocardiography; (2) persistent fever, with positive blood cultures in a child with congenital heart disease, in whom no other source of infection could be identified; and (3) a classic clinical picture (e.g., fever, splenomegaly, embolic phenomena), positive blood cultures, and a newly ac- Presented at the Thirty-first Annual Meeting of The Society of Thoracic Surgeons, Palm Springs, CA, Jan 30-Feb 1, Address reprint requests to Dr Karl, Victorian Paediatric Cardiac Surgery Unit, Royal Children's Hospital, Flemington Rd, Parkville, Victoria 3052, Australia. quired murmur in a child without previous structural heart disease. Results One hundred seven episodes of endocarditis were documented in 98 patients. The demographic and clinical features of the patient group are presented in Table 1. The median age at the time of diagnosis was 94.5 months (range, 1 to 257 months), and 24 patients were less than 2 years old. Ten episodes of neonatal endocarditis were encountered. The overall mortality for the patient group was 24.5% (95% confidence limits, 16% to 39%), and the actuarial survival at 10 years was 79.3% (95% confidence limits, 66% to 93%). Seventy-six patients had underlying structural heart disease, of which 39 had previously undergone some type of palliative or corrective procedure (Table 2). Two patients without structural heart disease had a central venous line in situ at the time of diagnosis. Staphylococcus and Streptococcus species, isolated in 38 and 26 patients, respectively, were the most common organisms. Blood cultures were repeatedly sterile in 14 patients, despite otherwise classic clinical findings (Table 3). The likelihood of structural heart disease and embolism was similar in patients in whom Staphylococcus was identified, as compared with the overall group (Table 4). Although there were four early deaths in the 32 patients in whom Staphylococcus aureus was identified, compared with 5 of the 66 remaining patients, the differences did not reach statistical significance (Table 4). Vegetations were identified in 43 patients (Table 5). The incidence of embolization was higher in patients in whom a vegetation was identified, but the risk of congestive heart failure and death were similar in patients with and without vegetations (Table 5) by The Society of Thoracic Surgeons /95/$ (95)00350-T

2 Ann Thorac Surg NOMURA ET AL ;60:90-5 INFECTIVE ENDOCARDITIS IN CHILDREN Table 1. Clinical Characteristics of the 98 Patients Who Presented to The Royal Children's Hospital With Endocarditis Characteristic Medical p Value Therapy Surgical (surgical Overall Alone Intervention versus (n = 98) (n = 68) (n= 30) medical) Age ~2 years Findings at presentation Fever NS Positive blood NS cultures Site of vegetation Tricuspid valve Pulmonary valve Mitral valve Aortic valve Other Embolism NS Congestive heart NS failure Recurrence NS Early death NS Late death NS Survival NS NS - not significant. Sixty-eight patients had medical therapy alone, and in 30 patients, antibiotic therapy was combined with operation. Surgical Patients Thirty-three surgical interventions were required in 30 patients. Median age at operation was 22.5 months Table 3. Microbiological Isolates in the Patient Group Medical p Value Therapy Surgical (surgical Microbiological Overall Alone Intervention versus Isolate (n = 38) (n = 28) (n - 10) medical) Staphylococcus spp NS S aureus 23 9 S epidermidis 5 1 Streptococcus spp NS S viridans 11 5 S pneumonia hemolytic 1 1 others 3 1 Haemophilus influenzae NS Other gram-positive NS organism Gram-negative NS organism Klebsiella 1 1 Serratia 2 2 Other 3 3 Candida albicans NS No isolate NS NS = not significant. (range, 1 month to 15 years). Twenty-two of the surgical patients had underlying structural congenital heart disease, of whom 9 had undergone palliative and 6 had undergone corrective procedures (Table 2). One patient had an acquired cardiomyopathy. Staphylococcus aureus was the organism most commonly isolated in the surgical group (Table 3). The indications for surgical intervention and the procedures performed are presented in Table 6. Table 2. Cardiac Lesions in the 76 Patients With Underlying Heart Disease ~ Medical Surgical p Value Overall Therapy Alone Intervention (surgical versus Expected Congenital Heart Disease (n = 76) (n = 54) (n = 22) medical) number overall VSD/AVSD +_ coarctation NS PDA + coarctation NS 4-9 Coarctation NS 3-7 Tetralogy of Fallot/PA VSD NS 3-6 Double outlet right ventricle NS Truncus arteriosus NS 0-2 PA IVS NS Pulmonary stenosis NS 2-10 Aortic stenosis NS 3-6 Tricuspid atresia/univentricular heart NS Mitral valve prolapse NS Ebstein's anomaly NS EFE/cardiomyopathy NS Hypoplastic left heart NS Transposition + VSD -+ LVOTO NS 3-9 a For selected lesions, the number of patients (combined medical and surgical) that would have been expected from population studies of all children with congenital heart disease (data from references 7-11 also is presented. AVSD = atrioventricular septal defect; EFE = endocardial fibroelastosis; IVS = intact ventricular septum; LVOTO = left ventricular outflow tract obstruction; NS = not significant; PA = pulmonary atresia; PDA = patent ductus arteriosus; VSD = ventricular septal defect.

3 92 NOMURA ET AL Ann Thorac Surg INFECTIVE ENDOCARDITIS IN CHILDREN 1995;60:90-5 Table 4. Comparison of Patients With and Without Evidence of Staphylococcus aureus Infection Staphylococcus Staphylococcus aureus aufeus Diagnosis Positive Negative p Value Total Congenital heart disease NS Embolism 7 11 NS Operation 9 21 NS Early death 4 5 NS NS not significant. Common indications included left-sided vegetations (in 1 patient with evidence of embolic phenomena), evidence of persisting infection (despite antibiotic therapy), and significant hemodynamic compromise (usually related to valvular dysfunction). Infection related to artificial tissue was an indication in 3 patients with pulmonary artery bands (leading to band erosion) and in 2 patients with systemic-to-pulmonary artery shunts. Although removal of a vegetation was often the only procedure performed, some patients required extensive debridements, associated with either valve repair or replacement, exclusion of an abscess, closure of a false aneurysm, or removal of prosthetic material (Table 6). Operative specimens were Table 5. Comparison of Patients With and Without Vegetations No p Value Vegetation Vegetation (vegetation Diagnosis (n = 43) (n - 55) versus none) Underlying heart disease Organism Staphylococcus NS Streptococcus NS Other NS Embolus 16 3 < Congestive heart 11 8 NS failure Operation 23 7 NS Death NS NS not significant. cultured in all patients. Evidence of recurring infection developed in 4 patients, and 3 of them required a second operation. Recurrent infection developed in 2 of 6 (33%) patients with positive tissue cultures (obtained at operation) and 2 of 24 (8%) patients with negative cultures. These differences did not reach statistical significance. There were two early deaths in the surgical group Table 6. Indications for Operation and Operative Procedures Performed for Infective Endocarditis in the 30 Surgical Patients Primary Indication for Operation Left-sided vegetation and embolic phenomena (3) Vegetation and continued sepsis (2) + aortic abscess (1) Significant aortic incompetence (3) Significant mitral incompetence (3) Infected systemic-to-pulmonary artery shunt (2) Infected band with band rupture (2) + pseudoaneurysm of main pulmonary artery (1) Infected Hancock conduit + pulmonary embolism (1) Infection (mycotic aneurysm) related to Gore-Tex angioplasty for coarctation (2) Infected central venous line + large vegetation (3) Large obstructive vegetation in RVOT (1) Large obstructive vegetation in SVC and RA (1) Septal abscess and aortic incompetence (1) Mycotic aneurysm in ascending aorta (1) + cerebral embolism (1) Calcified vegetation on tricuspid valve chorda, with tricuspid incompetence + VSD (1) Vegetation in ductus arteriosus and left pulmonary artery (1) Procedure Resection of vegetation (3) Resection of vegetation (2) + debridement of aortic abscess and aortic valve repair (1) a Prosthetic valve replacement (2) Debridement and repair of aortic valve (1) b Mitral valve repair + resection of vegetation (2) Replacement of shunt (1) Removal of shunt, VSD closure, RV-PA conduit Removal of band (2) Removal of band + resection of aneurysm + RV-PA homograft conduit (1) Homograft conduit replacement (1) Resection of infected aortic patch subclavian flap repair (1) + replacement of patch (1) Removal of line and vegetation (3) Resection of vegetation (1) Resection of vegetation (1) Homograft replacement of aortic root (1) bc Removal of aneurysm (2) Resection of vegetation, VSD closure, chordaplasty of tricuspid valve (1) Transpulmonary removal of vegetation, division ductus arteriosus (1) ~,b,c Patients who required second procedures: ~ subsequently underwent repair of VSD and aortic root replacement for aortic incompetence and continuing sepsis, b subsequently underwent aortic valve replacement for aortic incompetence, c subsequently underwent repeat aortic root replacement because of a false aneurysm. RA - right atrium; RV-PA = right ventricular-to-pulmonary artery; RVOT right ventricular outflow tract; SVC = superior vena cava; VSD = ventricular septal defect.

4 Ann Thorac Surg NOMURA ET AL ;60:90-5 INFECTIVE ENDOCARDITIS IN CHILDREN 1.0 ;. Kaplan-Meler Probability (95% Confidence Intervals) 0.6 ~ i _ ~ i 0.2 Therapy -- Non Surgical - - Surgical 0.0,i... i... i... i... i Months Fig 1. Actuarial survival curves (with 95% con~dence intervals)for the 98 patients with infective endocarditis, categorized into nonsurgical (n = 68) and surgical groups (n = 30). (6.7%) (95% confidence limits, 0.8% to 22%). Two patients required a left ventricular assist device and 1 patient required extracorporeal membrane oxygenation in the early postoperative period. There were four late deaths, three related to recurrent endocarditis. The fourth occurred suddenly during a flight 10 months after operation. The actuarial survival probability for the surgical patients was 76% (95% confidence limits, 53% to 89%) at 45 months, with no further deaths occurring to a maximum follow-up time of 120 months (Fig 1). Medical Patients Sixty-eight patients had medical therapy alone. The incidence of structural heart disease, congestive heart failure, and the spectrum of infective organisms were similar in the medical and surgical groups. However, only 9 of the 68 medical patients were aged less than 2 years, compared with 15 of the 30 surgical patients (p = ) (Table 1). The incidence of vegetations was significantly lower in the medical patients. There were 7 early and 11 late deaths in the medically treated group; therefore, the 5-year actuarial survival probability was 52% (95% confidence limits, 37% to 65%) (Fig 1). This was not significantly different from that in the surgical group (p > 0.05). Comment Infective endocarditis is a rare disease in children. This account of 107 episodes of endocarditis in 98 children is one of the largest reported pediatric series. Nonetheless, considering early and late outcome, endocarditis accounts for a significant morbidity and mortality in the pediatric population. The close association between structural heart disease and endocarditis seen in other series [2-4] is supported by our data. Underlying structural heart disease was seen in 79 of our 98 patients. There is little agreement in the published literature with regard to the relative risk of endocarditis with different cardiac lesions. In our experience, as in that of Van Hare and co-workers [2], left-toright shunting through a ventricular or atrioventricular septal defect forms the underlying substrate for infective endocarditis in a significant proportion of patients. However, although some series suggest that the risk of endocarditis in patients with right-to-left shunts may be low [2], our data do not support this impression. Of the 79 patients with congenital heart disease in our patient group, 10 had tetralogy of Fallot, with or without pulmonary atresia. It may not be possible to make statistically valid statements regarding the relative risk of endocarditis for different congenital heart lesions by comparing our patient group with the published population-based series for congenital heart disease. Nonetheless, in Table 2 we present the proportional distribution of some of the more common lesions seen in our endocarditis patients. Also shown are data from some of the larger population-based series, which assessed the distribution for the same lesions in all live-born infants with congenital heart disease [7-11]. These data suggest that the relative risk of endocarditis for lesions such as atrial septal defect, pulmonary stenosis, and coarctation of the aorta is low, whereas tetralogy of Fallot appears to be overrepresented in our series. As in other centers, the most common organisms isolated in our patients were Staphylococcus and Streptococcus [2, 12, 13]. The proportion of patients with staphylococcal infection was similar in those with and without congenital heart disease. Although it has been suggested that staphylococcal infection associated with left-sided vegetations may be a prima facie indication for operation [14], we operated on only 4 of 11 patients with this clinical picture. Six patients were cured with antibiotic treatment alone. Indeed, in our experience, unlike that of others [13], staphylococcal endocarditis was not associated with a greater likelihood of surgical intervention. Embolization occurred with similar frequency in patients with Staphylococcus compared with those with other organisms. There were four early deaths in the 32 patients with staphylococcal infection compared with 5 of the 66 patients with other types of endocarditis, although the differences did not reach statistical significance. Our experience regarding the clinical presentation of patients with Staphylococcus aureus infection is similar to that of others [15]. Many of these patients present with an acute septic crisis, rather than with the classic syndrome of "subacute" endocarditis. Our series includes children who progressed from being vaguely unwell, without any cardiovascular signs, to septic shock, severe cardiac failure, and valve destruction within less than 24 hours. In our institution, the medical management of patients with endocarditis is based on extensive microbiological investigations before the institution of antibiotics, the use of combinations of synergistic bactericidal agents, and measurement of their minimal bactericidal concentrations. Although we recognize that there has been much recent interest in the potential for short antibiotic courses [16], or predominantly oral therapy, we continue to recommend the proven 4- to 6-week course of intravenous antibiotics. All patients with endocarditis undergo regular echocardiograms during treatment, principally

5 94 NOMURA ET AL Ann Thorac Surg INFECTIVE ENDOCARDITIS IN CHILDREN 1995;60:90-5 for the detection of valvular dysfunction, abscess formation, and deteriorating ventricular function. It is our impression, however, that serial echocardiographic estimations of vegetation size are of little value for the most part, because of variations in technique and interpretation, and because a vegetation may show no significant reduction in size despite adequate antibiotic therapy [17]. Nonetheless, we would consider that continued growth of a vegetation, in the presence of continuing bacteremia, may indicate treatment failure and precipitate surgical intervention. Data from adult patients have suggested that the identification of a vegetation identifies a group of patients who are at high risk of life-threatening complications [14, 18]. In our patients, although the presence of a vegetation was associated with an increased risk of embolism, the risk of death and congestive heart failure was similar in patients with and without vegetations. It is likely, however, that this may reflect, in part, that most of our patients who did not have identifiable vegetations had underlying structural heart disease. The likelihood of surgical intervention for infective endocarditis increased over the period of this study. Patients in the most recent decade were twice as likely to have an operation during the course of treatment (p ). This may relate more closely to the increased availability and quality of pediatric open heart surgery in Melbourne during this time course, rather than a specific change in the virulence of the disease. The latter possibility cannot be discounted completely, however, as the growing population of survivors of corrective or palliative operations for congenital heart disease may have been more difficult to cure with antibiotics alone than their unoperated counterparts. The surgical procedures used during this study period constitute a wide technical spectrum, but with uniform goals. Removal of as much infected tissue as is reasonably possible and restoration of a favorable hemodynamic state were the priorities. When possible, concurrent complete repair of congenital cardiac defects was considered highly desirable. The technical details of operation for infective endocarditis have been well described, and are outside the scope of the present discussion. Improvements in aortic valvuloplasty techniques may preclude the need for valve or root replacement in many cases of antibiotic failure. When required, however, homograft rather than prosthetic replacement has been considered preferable. It is likely, however, that the pulmonary autograft will prove to be the best option in suitable patients. Infective endocarditis after a palliative operation may present the surgeon with a difficult decision regarding surgical options. We have encountered this situation with infected polytetrafluoroethylene modified Blalock shunts (2 children) and pulmonary artery (PA) bands (3 children). In 2 patients we were able to perform a complete repair, and although implantation of further prosthetic material was required, antibiotic therapy resulted in a cure. In 3 patients with infection at PA band sites, loss of banding effect was encountered when the band eroded the wall of the main PA. One of these patients had been banded in preparation for conversion from Senning to arterial switch operation for late right ventricular failure [19]. Because left ventricular "'retraining" was still in an early stage, left ventricular-to-pulmonary artery continuity was restored using a homograft tube tailored to an hourglass configuration. The patient was cured with antibiotic therapy and later had successful switch conversion. Indications for operation in infective endocarditis remain debatable and challengeable in children, and each patient must be considered on its own merits. This is especially true when dealing with the aortic root. Abnormalities, such as cavity and sinus tract formation and leaflet abnormalities, may or may not constitute indications for immediate operation. The key issues are control of systemic sepsis and hemodynamic status. Patients who are satisfactory in both regards can usually wait. Although many will eventually require operation, it is vastly preferable to complete antibiotic therapy beforehand. This is especially true in lesions of the aortic root, in which a radical reconstruction will often be required. In our series, 3 patients with essentially untreated infective endocarditis who were operated on for acute hemodynamic deterioration or multiple emboli required postoperative extracorporeal support. Two were supported with a left ventricular assist device and one with extracorporeal membrane oxygenation [20, 21]. All 3 patients survived and were cured of infection after antibiotic therapy. These cases illustrate that severe sepsis is not an absolute contraindication to extracorporeal life support in postoperative patients. In conclusion, infective endocarditis remains a complex and virulent disease in children. Staphylococcus aureus is the most common organism isolated, commonly leading to an acute septic crisis, rather than the classic syndrome of subacute bacterial endocarditis. Despite advances in antibiotic therapy, early surgical intervention is necessary in a subgroup of patients. All children with infective endocarditis should be treated during the acute phase of the infection in a cardiac surgical center, where close surveillance is possible. Despite advances in antibiotic therapy, a significant proportion of patients will need early surgical intervention. Complete repair of the congenital defect may be appropriate as part of the treatment of endocarditis in selected patients. Postoperative extracorporeal life support, which may be a useful adjunct, is not absolutely contraindicated in patients with severe sepsis. References 1. Zakrzewski T, Keith JD. Bacterial endocarditis in infants and children. J Pediatr 1965;67: Van Hare GF, Ben-Shachar G, Liebman J, Boxerbaum B, Reirnenschneider TA. Infective endocarditis in infants and children during the past 10 years: a decade of change. Am Heart J 1984;107: Karl T, Wensley D, Stark J, de Leval M, Rees P, Taylor JFN.

6 Ann Thorac Surg NOMURA ET AL ;60:90-5 INFECTIVE ENDOCARDITIS IN CHILDREN Infective endocarditis in children with congenital heart disease: comparison of selected features in patients with surgical correction or palliation and those without. Br Heart J 1987;58: Geva T, Frand M. Infective endocarditis in children with congenital heart disease. Eur Heart J 1988;9: Manhas DR, Mohri H, Hessel EA, Merendino KA. Experience with surgical management of primary infective endocarditis: a collective review of 139 patients. Am Heart J 1972;84: Cukingnan RA, Carey JS, Wittig JH, Cimochowski GE. Early valve replacement in active infective endocarditis. J Thorac Cardiovasc Surg 1983;85: Carlgren LE. The incidence of congenital heart disease in children born in Gothenburg Br Heart J 1959;21: Rose V, Boyd ARJ, Ashton TE. Incidence of heart disease in the city of Toronto. Can Med Assoc J 1964;91: Laursen HB. Some epidemiological aspects of congenital heart disease in Denmark. Acta Paediatr Scand 1980;69: Bound JP, Logan WFWE. Incidence of congenital heart disease in Blackpool Br Heart J 1977;39: Hoffman lie, Christianson R. Congenital heart disease in a cohort of births with long term followup. Am J Cardiol 1987;42: Citak M, Rees A, Mavroudis C. Surgical management of infective endocarditis. Ann Thorac Surg 1992;54: Salman L, Prince A, Gersony WM. Pediatric infective endocarditis in the modern era. J Pediatr 1993;122: Richardson JV, Karp RB, Kirklin JW. Treatment of infective endocarditis: a 10-year comparative analysis. Circulation 1978;58: Bayer AS. Infective endocarditis. Clin Infect Dis 1993;17: Chambers HF, Miller T, Newman MD. Right-sided Staphylococcus aureus endocarditis in intravenous drug abuser: two week combination chemotherapy. Ann Intern Med 1991;115: Stewart JA, Silimperi D, Harris P, Wise NK, Fraker TD, Kisslo IA. Echocardiographic documentation of vegetative lesions in infective endocarditis: clinical implications. Circulation 1980;61: Buda AJ, Zotz RJ, LeMire MS, Bach DS. Prognostic significance of vegetations detected by two-dimensional echocardiography in infective endocarditis. Am Heart J 1986;112; Cochrane AD, Karl TR, Mee RBB. Arterial switch for late failure of the systemic right ventricle. Ann Thorac Surg 1993;56: Karl TR, Horton SB, Sano S, Mee RBB. Centrifugal pump left heart assist in pediatric cardiac surgery: indications, technique and results. J Thorac Cardiovasc Surg 1991;102: Karl TR, Pennington GD. Extracorporeal circulatory support in infants and children. Semin Thorac Cardiovasc Surg 1994;6: Bound volumes available to subscribers Bound volumes of the 1994 issues of The Annals of Thoracic Surgery are available only to subscribers from the Publisher. The cost is $99.00 (outside US add $25.00 for postage) for volumes 57 and 58. Each bound volume contains a subject and author index, and all advertising is removed. The binding is durable buckram with the name of the journal, volume number, and year stamped on the spine. Payment must accompany all orders. Contact Elsevier Science Inc, 655 Avenue of the Americas, New York, NY 10010; or telephone (212) (facsimile: (212) ).