- Email: [email protected]
Pediatric infective endocarditis in the modern era
Pediatric infective endocarditis in the modern era Lisa Saiman, MD, Alice Prince, MD, and Welton M. Gersony, MD From the Department of Pediatrics, Divisions of Pediatric Infectious Diseaseand Pediatric Cardiology, Columbia Presbyterian Medical Center, New York, New York
Sixty-two cases of endocarditis occurring in children between January 1977and February 1992 were reviewed and compared with series from the 1970s and early 1980s.Changes in risk factors, pathogens, diagnostic modalities, and outcome were determined. Complex congenital heart disease (22 cases) and unrepalred ventricular septal defect (9 cases) were the most common underlying lesions. A total of 19 children with normal anatomy had endocarditis; 6 had communltyacquired infection and 13 had hospital-acquired endocarditis (11 of these 13 children had central venous catheters in place, including 7 premature infants). Echocardiograms revealed vegetations In 25 of 49 patients; 24 of these patients had positive echocardiographic findings on the first study. Echocardlographic findings were most otten negative In children with complex cyanotic heart disease. Staphylococcus aureus (39%) was the most common pathogen isolated and was associated with a higher Incidence of central nervous system complications (p <0.0015) and a greater need for surgical Intervention (p = 0.01) than were other pathogens. Methicillin-resistant S. aureus (eight cases) and coagulase-negative staphylococci (three cases) emerged as important pathogens but were not associated with increased morbidity or mortality rates. Fungal endocarditis (six cases) had a 67% mortality rate. Overall the mortality rate was 11%. Endocarditis remained undiagnosed in seven seriously ill patients until postmortem examination. This study indicates that, during the past decade, important changes in risk factors, pathogens, and the susceptible population have altered the presentation and management of endocarditis in children. (J PEDJATR 1993;122:847-53) During the past two decades an increasing number of risk factors for endocarditis have been recognized in pediatric patients!": (1) advances in medical and surgical management of children with complex congenital heart disease have led to increased life expectancy; (2) routine use of central venous catheters has increased the risk of endocarditis; (3) mitral valve prolapse is being recognized with increasing frequency, but the importance of this abnormality as a cause of pediatric endocarditis has been controversial; and Presented at the American Heart Association meeting, New Orleans, La., Nov. 15-19, 1992. Submitted for publication Sept. 8, 1992; accepted Jan. 26, 1993. Reprint requests: Lisa Saiman, MD, Columbia Presbyterian Medical Center, Department of Pediatrics, Division of Pediatric Infectious Disease, 650 W. 168th St., BB4-427, New York, NY 10032. Copyright @ 1993 by Mosby-Year Book, Inc. 0022-3476/93/$1.00 + .10 9/20/45928
(4) the importance of drug addiction as a risk factor in adolescents has yet to be determined. Few new data have appeared describing the recent epidemiologic and clinical spectrum of endocarditis in children, despite the everincreasing population at risk. The spectrum of organisms MIC MVP VSD
Minimum inhibitory concentration Mitral valve prolapse Ventricularseptal defect
causing endocarditis has changed during the past few decades. During the 1970s Staphylococcus aureus became increasingly important in comparison with streptococci. However, the importance of hospital-acquired (nosocomial) pathogens and the effect of antimicrobial resistance on the treatment of endocarditis need to be examined. Finally, the results of medical and surgical management of endocarditis during the past decade have not been reported.
847
848
Saiman, Prince, and Gersony
We reviewed the occurrence of endocarditis at our pediatric center from January 1977 through February 1992 to determine whether the current epidemiologic factors and changes in the spectrum of pathogens have had an effect on morbidity and mortality rates. We compared this more recent pediatric population with those treated for endocarditis during the 1970s and early 1980s. METHODS Patient selection. Cases of endocarditis treated at Babies Hospital, Columbia Presbyterian Medical Center, from January 1977 through February 1992 were reviewed. Cases were identified from the discharge diagnoses of all pediatric patients, the pathology reports, and the records of the infectious diseases and cardiology divisions. To be included in the study, patients had to meet one of the following case criteria 1-7: 1. Blood cultures positive for pathogens, with a vegetation demonstrated by echocardiography in a patient with or without structural heart disease 2. Blood cultures positive for pathogens, without a vegetation but with a history of fever, fatigue, joint pain or swelling, and recent dental manipulations; supporting laboratory data, including an increased leukocyte count, an elevated erythrocyte sedimentation rate, hematuria, and positive findings on a ventilation-perfusion scan; and/or physical findings of fever, a murmur or a changing murmur, hepatosplenomegaly, and embolic phenomena 3. One or more blood cultures positive for pathogens, with fever and congenital heart disease, and no other apparent source of infection 4. Fever and congenital heart disease with positive findings on an echocardiogram, embolic phenomena, or both, but no blood cultures positive for pathogens In addition, patients with histopathologic evidence of endocarditis at autopsy were analyzed. Microbiology. The susceptibilities of clinical isolates were determined by using microtiter plates (Microscan; Baxter Healthcare Corp., Microscan Division, West Sacramento, Calif.). Streptococci were defined as susceptible to penicillin ifthe minimum inhibitory concentration was <0.1 ,ug/ml and as relatively resistant if the MIC was >0.1 ,ug/ ml.s Methicillin-resistant strains of staphylococci were initiallyidentified as having an MIC to oxacillin >4 ,ug/ml and confirmed by growth of the organism on agar containing oxacillin, 6 /Lg/ml, at 37° C.9 RESULTS Patient population. Sixty-two patients with endocarditis were included in this review. Twenty-three patients had blood cultures positive for pathogens and vegetations on an
The Journal of Pediatrics June 1993
echocardiogram and therefore fulfilled criterion 1. Twentyfive patients fulfilled criterion 2: blood cultures positive for pathogens, no demonstrable vegetation, but with evidence of endocarditis by history, laboratory data, or physical examination. Twenty-three of these patients had structural heart disease and two had normal hearts. Three patients fulfilled criterion 3, and four children had culture-negative endocarditis, criterion 4. Endocarditis remained undiagnosed until autopsy in seven additional patients. Thirty-nine patients were male (63%). The ages ranged from 1 month to 19 years (median age 8.2 years); 20 patients were younger than 2 years of age. One third of the patients were transferred from other hospitals. The distribution of cases during the study period, including cases of endocarditis that remained undiagnosed until autopsy and cases noted in premature infants, is shown in Fig. 1. There were more cases during the later years of the study, which in part is accounted for by an increase in the number of premature infants with endocarditis. Prior medical conditions. The majority of the children (70%) in whom endocarditis developed had structural heart disease (Fig. 2). Congenital heart disease. Complex cyanotic heart disease with palliative shunts, conduits, or prosthetic valves accounted for about half of the congenital lesions (22 cases); ventricular septal defects (9 cases) predominated among the acyanotic lesions. Mitral valve prolapse. Four adolescents had mitral valve prolapse by echocardiography when endocarditis was diagnosed. Three had murmurs consistent with mitral regurgitation, but none had previously diagnosed mitral valve disease. The fourth patient had been told that she had a "click." Rheumatic heart disease. Rheumatic heart disease was an uncommon risk factor in this series. Three children had a history of rheumatic fever with residual valve involvement, and one had a prosthetic mitral valve. There were no cases of endocarditis associated with acute rheumatic fever. Normal cardiac anatomy. Nineteen children with normal cardiac anatomy had endocarditis. Six had contracted endocarditis in the community, and five of these children were previously healthy. None was an intravenous drug abuser. Four of these six had infection of the mitral valve. At the time of onset all appeared to have sepsis; their clinical courses were complicated by embolic phenomena with the need for surgical intervention, but all survived. Nosocomially acquired endocarditis occurred in 13 children with normal cardiac anatomy; 11 had central venous catheters, including 7 premature infants. All these patients had rightsided endocarditis. Surgical status associated with endocarditis. Of the children with congenital heart disease, 50% had had previous
The Journal of Pediatrics Volume 122, N umber 6
Salm an, Prince, and Gersony
849
12
B fZf
10 1/1
•
41
II)
ClI
8
0
6
-... u
Cases Premature Infants Autopsy diagnosed
41
.c
E
4
:::I
Z
2 0 777879808 1828 38485868 788899 09 192 Years FIg. 1. Yearly distrib ution of cases of endocarditis. Distribution of cases during study period (Jan uary 1977 through February 1992) is shown, including cases diagnosed in premature infants and at autopsy.
•
Comp ex cyanolJc heart disease
•
Ventricular septal defect
•
Other acyanotlc loolons
o
Rheumatic heart disease
m Mitral valve prolapl4l IZ] Normal anatomy
Fig. 2. Cardiac anatomy associate d with endocardit is. Cardiac anatomy of 62 children who had endocarditis is shown graphically, and number of children with each lesion is indicated.
cardiac surgery. Four children with congenital heart disease ha d early postoperative endocarditis (within 2 months of surg ery). Late postoperative endocar ditis (more than 2 months afte r surgery) occurred in 14 children with cyanotic lesions. Seven had palliativ e shunts and seven had had complex intraca rdiac reparative procedures. Eighteen children with congenital heart disease wit hout previous surgery had endocarditis. Unrepaired VSD was the most common congenita l lesion and accoun ted for nine cases. Antecedent antibiotics and prophylaxis. Before the diagnosis of endocar ditis, 21 patients received oral antibiotics for a variety of illnesses, including pharyngitis, gastroen teritis, "i nfluenza ," bronchitis, and urinar y tract infection . Despite the oral administration of antibiotics, cultu res of blood specimens from 18 of 21 patien ts grew pathogens; 3
of the patients with endocarditis had negative culture results. Only 1 of 41 children not pretreated with antibiotics had negative blood cultu re results (odds ratio 6.7; confidence int erval 0.8 to 53.7; Fisher exact two-sided test p = 0.2). Thus endocarditis in association with negative cu lture results was not significantly associated with antecedent treatment with antibiotics or with increased an mortality rate. Ap propriate antimicrobial prophylaxis did not always prevent endocarditis; two pat ients received penicillin before dental work, and one patient received penicillin for rheumat ic fever prophylaxis. These three children had endocarditis caused by penicillin-susceptible streptococci. Diagnostic use of echocardiograms. The results of echocar diography were available for 49 patients. (Echocardio g-
850
TheJournal of Pediatrics JUlie 1993
Salman, Prince, and Gersony
N=15
N= 10
N=8
N= 16
100 (II
E
...
ClI
Cl
0 "C III
80
...
o 0 ..c
60 (-) echocardiograms
(,)
-Gl
0
C 41
...Gl
(,)
(+) echocardiograms
40
20
c.. 0
NI. anatomy Valvular
VSDs
Cyanotic
Les io ns Fig. 3. Cardiac anatomy correlated with echocardiography. Percentage of echocardiograms demonstrating vegetations in four groups of patients, including those children with normal anatomy, isolated valvular lesions, unrepaired VSDs, and complex cyanotic heart disease. is shown. The number of children in each group is indicated. Nl, Normal.
raphy was not performed for 6 of 7 patients in whom endocarditis remained undiagnosed until autopsy.) Of these 49 patients, 25 (51 %) had demonstrable vegetations. In all but one patient (24/25), vegetations were present on the initial study, Children with normal anatomy, including prematur.e infants, or with isolated valvular abnormalities were most likely to have vegetations noted (Fig. 3). Most often, echocardiograms showedno abnormalities in patients with complex cyanotic lesions; 13 of these 16 children did not have demonstrable vegetations. Pathogens. The most common pathogen in this series was S. aureus; it caused 39%of all episodesof endocarditis. Half of the patients with congenital heart disease without previous surgery were infected with S. aureus (9/18), as were previously healthy patients with newly diagnosed MVP (3/ 4), children with normal anatomy (3/4), and premature infants with central venous catheters in place (4/7). Coagulase-negative staphylococci caused endocarditis in seven patients (I 1%), including three patients during the early postoperative period and one during the late postoperative period. Cefazolin has been used routinely as perioperative prophylaxis against staphylococci for cardiac surgery, but three of the patients with early postoperative endocarditis were infected with methicillin-resistant strains. Streptococci were the second most common organisms causing endocarditis and accounted for 22% of the cases. The species isolated included Streptococcus viridans, Streptococcus mitis, Streptococcus pneumoniae, Streptococcus sanguis, Streptococcus bovis, and group B strepto-
cocci; one third were relatively resistant to penicillin (MIC >0.1 ,ug/ml). Streptococci infected 4 of 13 patients with late postoperative infection, 5 of 9 with unrepaired VSDs, and 2 of 3 with rheumatic heart disease. Community-acquired endocarditis of unrepaired congenital lesions was caused equally by S. aureus or streptococci. Ten patients had endocarditis caused by more unusual pathogens, including eight who acquired endocarditis in the hospital. Of the 10 cases, 7 remained undiagnosed until autopsy. The pathogens isolated were typical of nosocomial organisms but were not usually associated with endocarditis; they included Candida species, Aspergillus, and gramnegative organisms. Medical management. Older children with uncomplicated streptococcal endocarditis were managed with both intravenous and oral regimens, including completion of parenteral therapy at home. Two children with S. viridans infection (penicillin MIC <0.1 ,ug/ml) were treated for 2 weeks with intravenously administered penicillin and gentamicin and for 2 additional weeks with orally administered amoxicillin; children witb relatively resistant streptococci were treated for 4 weeks with penicillin .and gentamicin. Children with methicillin-sensitive S. aureus were treated for 4 to 6 weeks with oxacillin; gentamicin, rifampin, or both were added for 2 to 4 weeks for synergy. Premature infants with tricuspid valve endocarditis caused by infection with methicillin-resistant S. aureus received intravenously administered vancomycin for 4 to 6 weeks with similar regimens for synergy.
Morbidity. Eighteen children had uncomplicated courses and responded to conventional medical therapy. These patients included those with congenital heart disease and streptococcal infection, and five premature infants with anatomically normal hearts and tricuspid valve endocarditis. Serious morbidity occurred in 37 of 55 patients and included valvular insufficiency and pulmonary emboli; central nervous system complications (Table) were significantly more frequent in patients infected with S. aureus than in patients infected with other pathogens (p <0.001; chisquare value = 10.1). Of 18 patients with central nervous system complications, 12 were infected with S. aureus. Morbidity was not increased in children infected with methicillin-susceptible S. aureus (16 cases), in comparison with those infected with methicillin-resistant S. aureus (8 cases) (p >0.5). Surgical intervention. Of 55 children, 12 (22%) bad surgery either while being treated for endocarditis or immediately after microbiologic cure. The surgical procedures included three vegetectomies for blood cultures that were persistently positive for pathogens, evacuation of an intracerebral and a retroperitoneal hematoma, five systemic atrioventricular valve replacements, placement of an aortic graft, and replacement of a ventriculoatrial shunt. Surgery was more likely to be required for S. aureus infection (9 of 12 cases) than for infection with other pathogens (p = 0.012; chi-square value = 6.5). Of 12 children who underwent surgery, 11 were cured and survived, although I child had serious neurologic sequelae. Mortality rate. Thirteen children (21%) in this series died of endocarditis. Diagnosis before death. Six children (II %) died of endocarditis while being treated with appropriate antimicrobial agents. Each had known congenital heart disease, and four were infected with S. aureus. The fifth child had been in the intensive care unit because of progressive congestive heart failure; subsequently endocarditis with Candida albicans developed, and the child died despite treatment with amphotericin B. The sixth child, infected with Enterococcus faecalis, died of an unsuspected mycotic aneurysm after 6 weeks of antibiotic therapy and apparent clinical cure. Diagnosis after death. Endocarditis remained undiagnosed in seven seriously ill children until postmortem examination. In six instances endocarditis developed during the treatment of another illness, including suspected immunodeficiency (two cases), Goodpasture syndrome, autoimmune neutropenia, metabolic disease, and familial erythrophagocytosis. All seven patients were being treated with broad-spectrum antibiotics for presumed bacterial sepsis. Three children infected with gram-negative organisms had blood cultures positive for the organisms, but the
851
Saiman, Prince, and Gersony
The Journal of Pediatrics Volume 122, Number 6
Table. Morbidity associated with endocarditis Cases (n
=
55)'
Morbidity
No.
%
Cardiac complications Valvular insufficiency Intraventricular abscess Pulmonary complications Emboli CNS complications Emboli Seizures Meningitis
37 10 9 I 9 9 18 9 6 3
67 18
16 33
'Cases diagnosed before death.
three patients with fungal endocarditis did not. None had echocardiography performed. The seventh child, known to have pulmonary atresia and two Blalock shunts, initially had a fever of unknown origin, but methicillin-resistant S. aureus was recovered at autopsy from his heart valves and multiple embolic sites. His echocardiograms had shown no abnormalities. All seven children had multisystem emboli at postmortem examination.
DISCUSSION When this series of pediatric patients with endocarditis is compared with series reported during the 1970s and early 1980s, both similarities and significant differences are found."? Patients with cyanotic congenital heart disease continued to represent the largest group at risk, but the most common cyanotic lesion from the earlier era was tetralogy of Fallot, whereas patients in our series had more complex anatomy with palliative shunts and valved conduits. The reason for this difference is presumably that children with tetralogy of Fallot are at lower risk after repair, and children with more complex lesions are surviving longer as a result of surgical palliation. Early postoperative endocarditis has become a rare event because infection control during open heart surgery has been excellent. In this series no cases of endocarditis were recognized in patients with surgically repaired VSDs. A previous multicenter report has also indicated a lower incidence of endocarditis in patients with successfully corrected YSDs.IO Rheumatic heart disease had traditionally been an important risk factor for endocarditis.t but as the incidence of rheumatic fever decreased throughout the 1970s and 1980s, it became a less important risk factor'. From OUr center, 66 cases of acute rheumatic fever were reported from 1977 to 1988, and a small resurgence was noted in 1985and 1986. 11 Two patients in that series did have endocarditis, 7 and 16 years, respectively, after the initial episodes of rheumatic fever.
852
Saiman, Prince, and Gersony
Cen tral venous catheters appeared to be an important risk factor in this series. Premature infants were especially at riskp-14 However, unlike patients in other reports with high mortality rates, all these infants survived . Thi s may be attributed to improved medical care of premature infants, earlier recognition of endocard itis, and prompt removal of the catheter. Previous reports generally have not shown that central venous catheters predispose older patients with cancer to endocarditis.P despite frequent episodes of bacteremia. 16. 17 However, several older children with other underlying diseases in this series did have endocarditis associated with the use of centr al venous catheters and appear to represent a unique population of seriously ill children at significant risk . The diagnosis of endocarditis has been greatly augmented by echocardiography, but demonstrable vegetations are not always present. In two previous series of twodimens ional echoc ardiography in children, 59% and 82% of patients with congenital heart disease had vegetations noted. 18, 19 Bricker et a1.18 found that only one of nine children with uncorrected cyanotic heart disease had positive findings on an echocardiogram. In our series, echocardiograms were leas t useful in patients with complex congenital hear t disease and were most effective in children and premature infants with normal anatomy. This finding may reflect characteristics of both the hosts and the pathogens. Vegetations are difficult to image within conduits and shunts utilized to manage complex cyanotic lesions. Vegetation size may also vary with specific pathogens; S. aureus, the most common organism isolated from specimens of patients with normal anatomy, tends to form large vegetations. During the 1980s nosocomial pathogens were frequently associat ed with endocarditis. S. aureus was already an important pathogen during the 1970s, but methicillin-resistant S. aureus and coagulase-negative staphylococci have become increasingly prevalent. Methicillin-resistant S. aureus did not appear to be a more virulent pathogen than methicillin-sensitive strains. There has also been an increased incidence of fungal endocarditis in recent yea rs, re. Iiecting more pr olonged courses of broader-spectrum antibacterial drugs in seriously ill patients. In the previou sly published series combined.ls fungi had caused endocarditis in only 3 (2%) of 192 cases, whereas in our series 10% were infected with fung i. Antibiotic prophylaxis to prevent endocarditis is not always effective.20The necessity for antibiotic prophylaxi s for patients with MVP has been debated because increasing numbers of cases of endocarditis are associated with this lesion.21,22 The most recent American Heart Association recommendations suggest prophylaxis for MVP only when valvular regurgitation is present. 20 S. aureus was the most common pathogen isolated in this series, and it is impossible to develop prophylactic strategies such as those devel-
The Journal of Pediatrics June 1993
oped for infection with streptococci and dental work. However, pre treatment of patients with orally administered antibiotics did not interfere significantly with recovery of the bacterial pathogens causing endocarditis. Medical management of endocarditis in children is derived largely from published recommendations for adults.f Despite the increased number of available antibiotics, there has been a limited role for newer agents such as ureidopenicillins, monobactams, and quinolones except in the treatment of infection with rare gram-negative pathogens. Among children with recognized endocarditis, the morta lity rate of II % in th is series was slightly less than the mortality rates of 14% to 22% previously reported. J-5 During the 1970s, children younger than 2 years of age gene rally had mortality rates> 30%, whereas the mortality rate in this series was 20%. Thus younger patients continued to have the worst prognos is, perhaps because of the severity of their cardiac lesions and underlying diseases . Seven children died of endocarditis that remained unsuspected until postmortem examin at ion. These cases reflect unfortunate outcomes associated with modern pediatric intensive care units. The pat ients had rem ark abl y similar clinical courses that masked the presence of endocarditis. Despite blood cultures that were persistently positive for gram-negative organisms, other foci of infection were sought. Endocarditis was not suspected in the children with fungal endocarditis because blood culture results remained negative. A previous report from this institution noted th at endocarditis caused by infection with Aspergillus developed after open heart surgery.P However, the two cases of autopsy-proved Aspergillus endocarditis described in the present series occurred in children with normal hearts and central venous catheters. Thus endocarditis should be considered in seriously ill patients with normal hearts who (I) receive prolonged courses of antibiotics , (2) have central venous catheters in place , and (3) have continued fevers of uncertain cause. We conclude that infective endocarditis presents an increasingl y difficult challenge in diagnosis and management. The patient with a simple cardiac defect and the classic findings of suba cute bacterial endocarditis caused by infection with S. vir idans has become the exception in the presen t era. We thank Dr. S. Griffiths, Dr. F. Bierman, Dr. P. Ursell, Dr. J. Garvin, R. Hawks, Dr. J. Driscoll, and K. Loizzo for sharing their records and insights. REFERENCES l. Johnson CM, Rhodes KH . Pediatric endocarditis. Mayo Clin Proc 1982;57:86-94. 2. Stanton BF, Baltimore RS, Clemens JD. Changing spectrum
Salman, Prince, and Gersony
The Journal of Pediatrics Volume 122, Number 6
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
of infective endocarditis in children: analysis of 26 cases, 1970-1979. Am J Dis Child 1984;138:720-5. Kramer JJ, Bourgeois M, Liersch R, et a!. Current clinical aspects of bacterial endocarditis in infancy, childhood, and adolescence. Eur J Pediatr 1983;140:253-9. Van Hare GF, Ben-Shachar G, Liebman J, Boxerbaurn B, Riemenschneider TA. Infective endocarditis in infants and children during the past 10 years: a decade of change. Am Heart J 1984;138:1235-40. Yokochi K, Sakamoto H, Mikajima T, et a1. Infective endocarditis in children: a current diagnostic trend and embolic complications. Jpn Circ J 1986;50:1294-7. Walterspiel IN, Kaplan SL. Incidence and clinical characteristics of "culture-negative" infective endocarditis in a pediatric population. Ped Infect Dis 1986;5:328-32. Johnson DH, Rosenthal A, Nadas AS. A forty-year review of bacterial endocarditis in infancy and children. Circulation 1975;51:581-8. Bisno AL, Dismukes WE, Durack DT, et a1. Antimicrobial treatment of infective endocarditis due to viridans streptococci, enterococci, and staphylococci. JAMA 1989;261:1471-7. I-lindler W, Warner NL. Effect of source of Mueller-Hinton agar on detection of oxacillin resistance in Staphylococcus aureus using a screening methodology. J Clin Microbiol1987; 25:734-5. Gersony WM, Hayes CJ. Bacterial endocarditis in patients with pulmonary stenosis, aortic stenosis, or ventricular septal defect. Circulation 1977;56(suppl): 184-7. Griffiths SP, Gersony WM. Acute rheumatic fever in New York City (1969 to 1988): a comparative study of two decades. J PEDIATR 1990;116:882-7. Edwards K, Ingall D, Czapek E, Davis AT. Bacterial endocarditis in 4 young infants. Clin Pediatr 1977;16:607-9.
853
13. Symchych PS, Krauss AN, Winchester P. Endocarditis following intracardiac placement of umbilical venous catheters in neonates. J PEDIATR 1977;90:287-9. 14. Oelberg DG, Fisher DJ, Gross DM, Denson SE, Adcock EW. Endocarditis in high-risk neonates. Pediatrics 1983;71:392-7. 15. Liepman MK, Jones PG, Kauffman CA. Endocarditis as a complication of indwelling right atrial catheters in leukemic patients. Cancer 1984;64:804-7. 16. Shapiro ED, Wald ER, Nelson KA. Broviac catheter-related bacteremia in oncology patients. AmJ Dis Child 1982;136:67981. 17. Hiemenz J, Skelton J, Pizzo PA. Perspective on the management of catheter-related infections in cancer patients. Pediatr Infect Dis 1986;5:6-11. 18. Bricker JT, Latson LA, Huhta JC, Gutgesell HP. Echocardiographicevaluation of infective endocarditis in children. Clin Pediatr 1985;24:312-7. 19. Kavet JEW, Frank DM, Byrum CJ, Blackman MS, Sondheimer HM, Bove EL. Two-dimensional echocardiographic assessment of infective endocarditis in children. Am J Dis Child 1983;137:851-6. 20. Dajani AS, Bisno AL, Chung KJ, et a1. Prevention of bacterial endocarditis: recommendations by the American Heart Association. JAMA 1990;264:2919-22. 21. Clemens JD, Horwitz RI, JalTeCC, Feinstein AR, Stanton BF. Controlled evaluation of the risk of bacterial endocarditis in persons with mitral valve prolapse. N EnglJ Med 1982;307:77681. 22. Danchin N, Briancon S, Mathieu P, et al. Mitral valve prolapse as a risk factor for infective endocarditis. Lancet 1989;2:743-5. 23. Barst R, Prince A, Neu HC. Aspergillus endocarditis in children: case report and review of the literature. Pediatrics 1981;68:73-8.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1993 issues of THE JOURNAL OF PEDIATRtCS are available to subscribers (only) from the Publisher, at a cost of $63.50 for domestic, $83.96 for Canadian, and $79.50 for international subscribers, for Vol. 122 (January-June) and Vol. 123 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA/800-453-4351, or 314-453-435l.
Subscriptions mustbe in force to qualify. Bound volumes are notavailable in place of a regular Journal subscription.
Sixty-two cases of endocarditis occurring in children between January 1977and February 1992 were reviewed and compared with series from the 1970s and early 1980s.Changes in risk factors, pathogens, diagnostic modalities, and outcome were determined. Complex congenital heart disease (22 cases) and unrepalred ventricular septal defect (9 cases) were the most common underlying lesions. A total of 19 children with normal anatomy had endocarditis; 6 had communltyacquired infection and 13 had hospital-acquired endocarditis (11 of these 13 children had central venous catheters in place, including 7 premature infants). Echocardiograms revealed vegetations In 25 of 49 patients; 24 of these patients had positive echocardiographic findings on the first study. Echocardlographic findings were most otten negative In children with complex cyanotic heart disease. Staphylococcus aureus (39%) was the most common pathogen isolated and was associated with a higher Incidence of central nervous system complications (p <0.0015) and a greater need for surgical Intervention (p = 0.01) than were other pathogens. Methicillin-resistant S. aureus (eight cases) and coagulase-negative staphylococci (three cases) emerged as important pathogens but were not associated with increased morbidity or mortality rates. Fungal endocarditis (six cases) had a 67% mortality rate. Overall the mortality rate was 11%. Endocarditis remained undiagnosed in seven seriously ill patients until postmortem examination. This study indicates that, during the past decade, important changes in risk factors, pathogens, and the susceptible population have altered the presentation and management of endocarditis in children. (J PEDJATR 1993;122:847-53) During the past two decades an increasing number of risk factors for endocarditis have been recognized in pediatric patients!": (1) advances in medical and surgical management of children with complex congenital heart disease have led to increased life expectancy; (2) routine use of central venous catheters has increased the risk of endocarditis; (3) mitral valve prolapse is being recognized with increasing frequency, but the importance of this abnormality as a cause of pediatric endocarditis has been controversial; and Presented at the American Heart Association meeting, New Orleans, La., Nov. 15-19, 1992. Submitted for publication Sept. 8, 1992; accepted Jan. 26, 1993. Reprint requests: Lisa Saiman, MD, Columbia Presbyterian Medical Center, Department of Pediatrics, Division of Pediatric Infectious Disease, 650 W. 168th St., BB4-427, New York, NY 10032. Copyright @ 1993 by Mosby-Year Book, Inc. 0022-3476/93/$1.00 + .10 9/20/45928
(4) the importance of drug addiction as a risk factor in adolescents has yet to be determined. Few new data have appeared describing the recent epidemiologic and clinical spectrum of endocarditis in children, despite the everincreasing population at risk. The spectrum of organisms MIC MVP VSD
Minimum inhibitory concentration Mitral valve prolapse Ventricularseptal defect
causing endocarditis has changed during the past few decades. During the 1970s Staphylococcus aureus became increasingly important in comparison with streptococci. However, the importance of hospital-acquired (nosocomial) pathogens and the effect of antimicrobial resistance on the treatment of endocarditis need to be examined. Finally, the results of medical and surgical management of endocarditis during the past decade have not been reported.
847
848
Saiman, Prince, and Gersony
We reviewed the occurrence of endocarditis at our pediatric center from January 1977 through February 1992 to determine whether the current epidemiologic factors and changes in the spectrum of pathogens have had an effect on morbidity and mortality rates. We compared this more recent pediatric population with those treated for endocarditis during the 1970s and early 1980s. METHODS Patient selection. Cases of endocarditis treated at Babies Hospital, Columbia Presbyterian Medical Center, from January 1977 through February 1992 were reviewed. Cases were identified from the discharge diagnoses of all pediatric patients, the pathology reports, and the records of the infectious diseases and cardiology divisions. To be included in the study, patients had to meet one of the following case criteria 1-7: 1. Blood cultures positive for pathogens, with a vegetation demonstrated by echocardiography in a patient with or without structural heart disease 2. Blood cultures positive for pathogens, without a vegetation but with a history of fever, fatigue, joint pain or swelling, and recent dental manipulations; supporting laboratory data, including an increased leukocyte count, an elevated erythrocyte sedimentation rate, hematuria, and positive findings on a ventilation-perfusion scan; and/or physical findings of fever, a murmur or a changing murmur, hepatosplenomegaly, and embolic phenomena 3. One or more blood cultures positive for pathogens, with fever and congenital heart disease, and no other apparent source of infection 4. Fever and congenital heart disease with positive findings on an echocardiogram, embolic phenomena, or both, but no blood cultures positive for pathogens In addition, patients with histopathologic evidence of endocarditis at autopsy were analyzed. Microbiology. The susceptibilities of clinical isolates were determined by using microtiter plates (Microscan; Baxter Healthcare Corp., Microscan Division, West Sacramento, Calif.). Streptococci were defined as susceptible to penicillin ifthe minimum inhibitory concentration was <0.1 ,ug/ml and as relatively resistant if the MIC was >0.1 ,ug/ ml.s Methicillin-resistant strains of staphylococci were initiallyidentified as having an MIC to oxacillin >4 ,ug/ml and confirmed by growth of the organism on agar containing oxacillin, 6 /Lg/ml, at 37° C.9 RESULTS Patient population. Sixty-two patients with endocarditis were included in this review. Twenty-three patients had blood cultures positive for pathogens and vegetations on an
The Journal of Pediatrics June 1993
echocardiogram and therefore fulfilled criterion 1. Twentyfive patients fulfilled criterion 2: blood cultures positive for pathogens, no demonstrable vegetation, but with evidence of endocarditis by history, laboratory data, or physical examination. Twenty-three of these patients had structural heart disease and two had normal hearts. Three patients fulfilled criterion 3, and four children had culture-negative endocarditis, criterion 4. Endocarditis remained undiagnosed until autopsy in seven additional patients. Thirty-nine patients were male (63%). The ages ranged from 1 month to 19 years (median age 8.2 years); 20 patients were younger than 2 years of age. One third of the patients were transferred from other hospitals. The distribution of cases during the study period, including cases of endocarditis that remained undiagnosed until autopsy and cases noted in premature infants, is shown in Fig. 1. There were more cases during the later years of the study, which in part is accounted for by an increase in the number of premature infants with endocarditis. Prior medical conditions. The majority of the children (70%) in whom endocarditis developed had structural heart disease (Fig. 2). Congenital heart disease. Complex cyanotic heart disease with palliative shunts, conduits, or prosthetic valves accounted for about half of the congenital lesions (22 cases); ventricular septal defects (9 cases) predominated among the acyanotic lesions. Mitral valve prolapse. Four adolescents had mitral valve prolapse by echocardiography when endocarditis was diagnosed. Three had murmurs consistent with mitral regurgitation, but none had previously diagnosed mitral valve disease. The fourth patient had been told that she had a "click." Rheumatic heart disease. Rheumatic heart disease was an uncommon risk factor in this series. Three children had a history of rheumatic fever with residual valve involvement, and one had a prosthetic mitral valve. There were no cases of endocarditis associated with acute rheumatic fever. Normal cardiac anatomy. Nineteen children with normal cardiac anatomy had endocarditis. Six had contracted endocarditis in the community, and five of these children were previously healthy. None was an intravenous drug abuser. Four of these six had infection of the mitral valve. At the time of onset all appeared to have sepsis; their clinical courses were complicated by embolic phenomena with the need for surgical intervention, but all survived. Nosocomially acquired endocarditis occurred in 13 children with normal cardiac anatomy; 11 had central venous catheters, including 7 premature infants. All these patients had rightsided endocarditis. Surgical status associated with endocarditis. Of the children with congenital heart disease, 50% had had previous
The Journal of Pediatrics Volume 122, N umber 6
Salm an, Prince, and Gersony
849
12
B fZf
10 1/1
•
41
II)
ClI
8
0
6
-... u
Cases Premature Infants Autopsy diagnosed
41
.c
E
4
:::I
Z
2 0 777879808 1828 38485868 788899 09 192 Years FIg. 1. Yearly distrib ution of cases of endocarditis. Distribution of cases during study period (Jan uary 1977 through February 1992) is shown, including cases diagnosed in premature infants and at autopsy.
•
Comp ex cyanolJc heart disease
•
Ventricular septal defect
•
Other acyanotlc loolons
o
Rheumatic heart disease
m Mitral valve prolapl4l IZ] Normal anatomy
Fig. 2. Cardiac anatomy associate d with endocardit is. Cardiac anatomy of 62 children who had endocarditis is shown graphically, and number of children with each lesion is indicated.
cardiac surgery. Four children with congenital heart disease ha d early postoperative endocarditis (within 2 months of surg ery). Late postoperative endocar ditis (more than 2 months afte r surgery) occurred in 14 children with cyanotic lesions. Seven had palliativ e shunts and seven had had complex intraca rdiac reparative procedures. Eighteen children with congenital heart disease wit hout previous surgery had endocarditis. Unrepaired VSD was the most common congenita l lesion and accoun ted for nine cases. Antecedent antibiotics and prophylaxis. Before the diagnosis of endocar ditis, 21 patients received oral antibiotics for a variety of illnesses, including pharyngitis, gastroen teritis, "i nfluenza ," bronchitis, and urinar y tract infection . Despite the oral administration of antibiotics, cultu res of blood specimens from 18 of 21 patien ts grew pathogens; 3
of the patients with endocarditis had negative culture results. Only 1 of 41 children not pretreated with antibiotics had negative blood cultu re results (odds ratio 6.7; confidence int erval 0.8 to 53.7; Fisher exact two-sided test p = 0.2). Thus endocarditis in association with negative cu lture results was not significantly associated with antecedent treatment with antibiotics or with increased an mortality rate. Ap propriate antimicrobial prophylaxis did not always prevent endocarditis; two pat ients received penicillin before dental work, and one patient received penicillin for rheumat ic fever prophylaxis. These three children had endocarditis caused by penicillin-susceptible streptococci. Diagnostic use of echocardiograms. The results of echocar diography were available for 49 patients. (Echocardio g-
850
TheJournal of Pediatrics JUlie 1993
Salman, Prince, and Gersony
N=15
N= 10
N=8
N= 16
100 (II
E
...
ClI
Cl
0 "C III
80
...
o 0 ..c
60 (-) echocardiograms
(,)
-Gl
0
C 41
...Gl
(,)
(+) echocardiograms
40
20
c.. 0
NI. anatomy Valvular
VSDs
Cyanotic
Les io ns Fig. 3. Cardiac anatomy correlated with echocardiography. Percentage of echocardiograms demonstrating vegetations in four groups of patients, including those children with normal anatomy, isolated valvular lesions, unrepaired VSDs, and complex cyanotic heart disease. is shown. The number of children in each group is indicated. Nl, Normal.
raphy was not performed for 6 of 7 patients in whom endocarditis remained undiagnosed until autopsy.) Of these 49 patients, 25 (51 %) had demonstrable vegetations. In all but one patient (24/25), vegetations were present on the initial study, Children with normal anatomy, including prematur.e infants, or with isolated valvular abnormalities were most likely to have vegetations noted (Fig. 3). Most often, echocardiograms showedno abnormalities in patients with complex cyanotic lesions; 13 of these 16 children did not have demonstrable vegetations. Pathogens. The most common pathogen in this series was S. aureus; it caused 39%of all episodesof endocarditis. Half of the patients with congenital heart disease without previous surgery were infected with S. aureus (9/18), as were previously healthy patients with newly diagnosed MVP (3/ 4), children with normal anatomy (3/4), and premature infants with central venous catheters in place (4/7). Coagulase-negative staphylococci caused endocarditis in seven patients (I 1%), including three patients during the early postoperative period and one during the late postoperative period. Cefazolin has been used routinely as perioperative prophylaxis against staphylococci for cardiac surgery, but three of the patients with early postoperative endocarditis were infected with methicillin-resistant strains. Streptococci were the second most common organisms causing endocarditis and accounted for 22% of the cases. The species isolated included Streptococcus viridans, Streptococcus mitis, Streptococcus pneumoniae, Streptococcus sanguis, Streptococcus bovis, and group B strepto-
cocci; one third were relatively resistant to penicillin (MIC >0.1 ,ug/ml). Streptococci infected 4 of 13 patients with late postoperative infection, 5 of 9 with unrepaired VSDs, and 2 of 3 with rheumatic heart disease. Community-acquired endocarditis of unrepaired congenital lesions was caused equally by S. aureus or streptococci. Ten patients had endocarditis caused by more unusual pathogens, including eight who acquired endocarditis in the hospital. Of the 10 cases, 7 remained undiagnosed until autopsy. The pathogens isolated were typical of nosocomial organisms but were not usually associated with endocarditis; they included Candida species, Aspergillus, and gramnegative organisms. Medical management. Older children with uncomplicated streptococcal endocarditis were managed with both intravenous and oral regimens, including completion of parenteral therapy at home. Two children with S. viridans infection (penicillin MIC <0.1 ,ug/ml) were treated for 2 weeks with intravenously administered penicillin and gentamicin and for 2 additional weeks with orally administered amoxicillin; children witb relatively resistant streptococci were treated for 4 weeks with penicillin .and gentamicin. Children with methicillin-sensitive S. aureus were treated for 4 to 6 weeks with oxacillin; gentamicin, rifampin, or both were added for 2 to 4 weeks for synergy. Premature infants with tricuspid valve endocarditis caused by infection with methicillin-resistant S. aureus received intravenously administered vancomycin for 4 to 6 weeks with similar regimens for synergy.
Morbidity. Eighteen children had uncomplicated courses and responded to conventional medical therapy. These patients included those with congenital heart disease and streptococcal infection, and five premature infants with anatomically normal hearts and tricuspid valve endocarditis. Serious morbidity occurred in 37 of 55 patients and included valvular insufficiency and pulmonary emboli; central nervous system complications (Table) were significantly more frequent in patients infected with S. aureus than in patients infected with other pathogens (p <0.001; chisquare value = 10.1). Of 18 patients with central nervous system complications, 12 were infected with S. aureus. Morbidity was not increased in children infected with methicillin-susceptible S. aureus (16 cases), in comparison with those infected with methicillin-resistant S. aureus (8 cases) (p >0.5). Surgical intervention. Of 55 children, 12 (22%) bad surgery either while being treated for endocarditis or immediately after microbiologic cure. The surgical procedures included three vegetectomies for blood cultures that were persistently positive for pathogens, evacuation of an intracerebral and a retroperitoneal hematoma, five systemic atrioventricular valve replacements, placement of an aortic graft, and replacement of a ventriculoatrial shunt. Surgery was more likely to be required for S. aureus infection (9 of 12 cases) than for infection with other pathogens (p = 0.012; chi-square value = 6.5). Of 12 children who underwent surgery, 11 were cured and survived, although I child had serious neurologic sequelae. Mortality rate. Thirteen children (21%) in this series died of endocarditis. Diagnosis before death. Six children (II %) died of endocarditis while being treated with appropriate antimicrobial agents. Each had known congenital heart disease, and four were infected with S. aureus. The fifth child had been in the intensive care unit because of progressive congestive heart failure; subsequently endocarditis with Candida albicans developed, and the child died despite treatment with amphotericin B. The sixth child, infected with Enterococcus faecalis, died of an unsuspected mycotic aneurysm after 6 weeks of antibiotic therapy and apparent clinical cure. Diagnosis after death. Endocarditis remained undiagnosed in seven seriously ill children until postmortem examination. In six instances endocarditis developed during the treatment of another illness, including suspected immunodeficiency (two cases), Goodpasture syndrome, autoimmune neutropenia, metabolic disease, and familial erythrophagocytosis. All seven patients were being treated with broad-spectrum antibiotics for presumed bacterial sepsis. Three children infected with gram-negative organisms had blood cultures positive for the organisms, but the
851
Saiman, Prince, and Gersony
The Journal of Pediatrics Volume 122, Number 6
Table. Morbidity associated with endocarditis Cases (n
=
55)'
Morbidity
No.
%
Cardiac complications Valvular insufficiency Intraventricular abscess Pulmonary complications Emboli CNS complications Emboli Seizures Meningitis
37 10 9 I 9 9 18 9 6 3
67 18
16 33
'Cases diagnosed before death.
three patients with fungal endocarditis did not. None had echocardiography performed. The seventh child, known to have pulmonary atresia and two Blalock shunts, initially had a fever of unknown origin, but methicillin-resistant S. aureus was recovered at autopsy from his heart valves and multiple embolic sites. His echocardiograms had shown no abnormalities. All seven children had multisystem emboli at postmortem examination.
DISCUSSION When this series of pediatric patients with endocarditis is compared with series reported during the 1970s and early 1980s, both similarities and significant differences are found."? Patients with cyanotic congenital heart disease continued to represent the largest group at risk, but the most common cyanotic lesion from the earlier era was tetralogy of Fallot, whereas patients in our series had more complex anatomy with palliative shunts and valved conduits. The reason for this difference is presumably that children with tetralogy of Fallot are at lower risk after repair, and children with more complex lesions are surviving longer as a result of surgical palliation. Early postoperative endocarditis has become a rare event because infection control during open heart surgery has been excellent. In this series no cases of endocarditis were recognized in patients with surgically repaired VSDs. A previous multicenter report has also indicated a lower incidence of endocarditis in patients with successfully corrected YSDs.IO Rheumatic heart disease had traditionally been an important risk factor for endocarditis.t but as the incidence of rheumatic fever decreased throughout the 1970s and 1980s, it became a less important risk factor'. From OUr center, 66 cases of acute rheumatic fever were reported from 1977 to 1988, and a small resurgence was noted in 1985and 1986. 11 Two patients in that series did have endocarditis, 7 and 16 years, respectively, after the initial episodes of rheumatic fever.
852
Saiman, Prince, and Gersony
Cen tral venous catheters appeared to be an important risk factor in this series. Premature infants were especially at riskp-14 However, unlike patients in other reports with high mortality rates, all these infants survived . Thi s may be attributed to improved medical care of premature infants, earlier recognition of endocard itis, and prompt removal of the catheter. Previous reports generally have not shown that central venous catheters predispose older patients with cancer to endocarditis.P despite frequent episodes of bacteremia. 16. 17 However, several older children with other underlying diseases in this series did have endocarditis associated with the use of centr al venous catheters and appear to represent a unique population of seriously ill children at significant risk . The diagnosis of endocarditis has been greatly augmented by echocardiography, but demonstrable vegetations are not always present. In two previous series of twodimens ional echoc ardiography in children, 59% and 82% of patients with congenital heart disease had vegetations noted. 18, 19 Bricker et a1.18 found that only one of nine children with uncorrected cyanotic heart disease had positive findings on an echocardiogram. In our series, echocardiograms were leas t useful in patients with complex congenital hear t disease and were most effective in children and premature infants with normal anatomy. This finding may reflect characteristics of both the hosts and the pathogens. Vegetations are difficult to image within conduits and shunts utilized to manage complex cyanotic lesions. Vegetation size may also vary with specific pathogens; S. aureus, the most common organism isolated from specimens of patients with normal anatomy, tends to form large vegetations. During the 1980s nosocomial pathogens were frequently associat ed with endocarditis. S. aureus was already an important pathogen during the 1970s, but methicillin-resistant S. aureus and coagulase-negative staphylococci have become increasingly prevalent. Methicillin-resistant S. aureus did not appear to be a more virulent pathogen than methicillin-sensitive strains. There has also been an increased incidence of fungal endocarditis in recent yea rs, re. Iiecting more pr olonged courses of broader-spectrum antibacterial drugs in seriously ill patients. In the previou sly published series combined.ls fungi had caused endocarditis in only 3 (2%) of 192 cases, whereas in our series 10% were infected with fung i. Antibiotic prophylaxis to prevent endocarditis is not always effective.20The necessity for antibiotic prophylaxi s for patients with MVP has been debated because increasing numbers of cases of endocarditis are associated with this lesion.21,22 The most recent American Heart Association recommendations suggest prophylaxis for MVP only when valvular regurgitation is present. 20 S. aureus was the most common pathogen isolated in this series, and it is impossible to develop prophylactic strategies such as those devel-
The Journal of Pediatrics June 1993
oped for infection with streptococci and dental work. However, pre treatment of patients with orally administered antibiotics did not interfere significantly with recovery of the bacterial pathogens causing endocarditis. Medical management of endocarditis in children is derived largely from published recommendations for adults.f Despite the increased number of available antibiotics, there has been a limited role for newer agents such as ureidopenicillins, monobactams, and quinolones except in the treatment of infection with rare gram-negative pathogens. Among children with recognized endocarditis, the morta lity rate of II % in th is series was slightly less than the mortality rates of 14% to 22% previously reported. J-5 During the 1970s, children younger than 2 years of age gene rally had mortality rates> 30%, whereas the mortality rate in this series was 20%. Thus younger patients continued to have the worst prognos is, perhaps because of the severity of their cardiac lesions and underlying diseases . Seven children died of endocarditis that remained unsuspected until postmortem examin at ion. These cases reflect unfortunate outcomes associated with modern pediatric intensive care units. The pat ients had rem ark abl y similar clinical courses that masked the presence of endocarditis. Despite blood cultures that were persistently positive for gram-negative organisms, other foci of infection were sought. Endocarditis was not suspected in the children with fungal endocarditis because blood culture results remained negative. A previous report from this institution noted th at endocarditis caused by infection with Aspergillus developed after open heart surgery.P However, the two cases of autopsy-proved Aspergillus endocarditis described in the present series occurred in children with normal hearts and central venous catheters. Thus endocarditis should be considered in seriously ill patients with normal hearts who (I) receive prolonged courses of antibiotics , (2) have central venous catheters in place , and (3) have continued fevers of uncertain cause. We conclude that infective endocarditis presents an increasingl y difficult challenge in diagnosis and management. The patient with a simple cardiac defect and the classic findings of suba cute bacterial endocarditis caused by infection with S. vir idans has become the exception in the presen t era. We thank Dr. S. Griffiths, Dr. F. Bierman, Dr. P. Ursell, Dr. J. Garvin, R. Hawks, Dr. J. Driscoll, and K. Loizzo for sharing their records and insights. REFERENCES l. Johnson CM, Rhodes KH . Pediatric endocarditis. Mayo Clin Proc 1982;57:86-94. 2. Stanton BF, Baltimore RS, Clemens JD. Changing spectrum
Salman, Prince, and Gersony
The Journal of Pediatrics Volume 122, Number 6
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
of infective endocarditis in children: analysis of 26 cases, 1970-1979. Am J Dis Child 1984;138:720-5. Kramer JJ, Bourgeois M, Liersch R, et a!. Current clinical aspects of bacterial endocarditis in infancy, childhood, and adolescence. Eur J Pediatr 1983;140:253-9. Van Hare GF, Ben-Shachar G, Liebman J, Boxerbaurn B, Riemenschneider TA. Infective endocarditis in infants and children during the past 10 years: a decade of change. Am Heart J 1984;138:1235-40. Yokochi K, Sakamoto H, Mikajima T, et a1. Infective endocarditis in children: a current diagnostic trend and embolic complications. Jpn Circ J 1986;50:1294-7. Walterspiel IN, Kaplan SL. Incidence and clinical characteristics of "culture-negative" infective endocarditis in a pediatric population. Ped Infect Dis 1986;5:328-32. Johnson DH, Rosenthal A, Nadas AS. A forty-year review of bacterial endocarditis in infancy and children. Circulation 1975;51:581-8. Bisno AL, Dismukes WE, Durack DT, et a1. Antimicrobial treatment of infective endocarditis due to viridans streptococci, enterococci, and staphylococci. JAMA 1989;261:1471-7. I-lindler W, Warner NL. Effect of source of Mueller-Hinton agar on detection of oxacillin resistance in Staphylococcus aureus using a screening methodology. J Clin Microbiol1987; 25:734-5. Gersony WM, Hayes CJ. Bacterial endocarditis in patients with pulmonary stenosis, aortic stenosis, or ventricular septal defect. Circulation 1977;56(suppl): 184-7. Griffiths SP, Gersony WM. Acute rheumatic fever in New York City (1969 to 1988): a comparative study of two decades. J PEDIATR 1990;116:882-7. Edwards K, Ingall D, Czapek E, Davis AT. Bacterial endocarditis in 4 young infants. Clin Pediatr 1977;16:607-9.
853
13. Symchych PS, Krauss AN, Winchester P. Endocarditis following intracardiac placement of umbilical venous catheters in neonates. J PEDIATR 1977;90:287-9. 14. Oelberg DG, Fisher DJ, Gross DM, Denson SE, Adcock EW. Endocarditis in high-risk neonates. Pediatrics 1983;71:392-7. 15. Liepman MK, Jones PG, Kauffman CA. Endocarditis as a complication of indwelling right atrial catheters in leukemic patients. Cancer 1984;64:804-7. 16. Shapiro ED, Wald ER, Nelson KA. Broviac catheter-related bacteremia in oncology patients. AmJ Dis Child 1982;136:67981. 17. Hiemenz J, Skelton J, Pizzo PA. Perspective on the management of catheter-related infections in cancer patients. Pediatr Infect Dis 1986;5:6-11. 18. Bricker JT, Latson LA, Huhta JC, Gutgesell HP. Echocardiographicevaluation of infective endocarditis in children. Clin Pediatr 1985;24:312-7. 19. Kavet JEW, Frank DM, Byrum CJ, Blackman MS, Sondheimer HM, Bove EL. Two-dimensional echocardiographic assessment of infective endocarditis in children. Am J Dis Child 1983;137:851-6. 20. Dajani AS, Bisno AL, Chung KJ, et a1. Prevention of bacterial endocarditis: recommendations by the American Heart Association. JAMA 1990;264:2919-22. 21. Clemens JD, Horwitz RI, JalTeCC, Feinstein AR, Stanton BF. Controlled evaluation of the risk of bacterial endocarditis in persons with mitral valve prolapse. N EnglJ Med 1982;307:77681. 22. Danchin N, Briancon S, Mathieu P, et al. Mitral valve prolapse as a risk factor for infective endocarditis. Lancet 1989;2:743-5. 23. Barst R, Prince A, Neu HC. Aspergillus endocarditis in children: case report and review of the literature. Pediatrics 1981;68:73-8.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1993 issues of THE JOURNAL OF PEDIATRtCS are available to subscribers (only) from the Publisher, at a cost of $63.50 for domestic, $83.96 for Canadian, and $79.50 for international subscribers, for Vol. 122 (January-June) and Vol. 123 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA/800-453-4351, or 314-453-435l.
Subscriptions mustbe in force to qualify. Bound volumes are notavailable in place of a regular Journal subscription.