Acta Med Scand 1986; 219: 275-82

Role of Surgery in Infective Endocarditis

PER ANDERSSON,' WIESLAW DUBIEL, EBBA ENGHOFF, GORAN FRIMAN, ANDERS HAGG, SVEN-OLOF NYSTROM and TORKEL ABERG From the Departments of Internal Medicine, Thoracic Surgery, and Infectious Diseases, University Hospital, Uppsala, Sweden

ABSTRACT. Anderson P, Dubiel W, Enghoff E et al. (Departments of Internal Medicine, Thoracic Surgery, and Infectious Diseases, University Hospital, Uppsala, Sweden.) Role of surgery in infective endocarditis. Acta Med Scand 1986; 219: 275-82.

One-hundred-and-thirteen patients with endocarditis and valvular insufficiency were stud- ied retrospectively with special regard to indications for operation and the optimum time for cardiac valve surgery. Thirty patients (group I) had acute, 63 (group 11) subacute and 20 (group 111) prosthetic valve endocarditis. Group I: Eleven patients underwent surgery in the acute stage, 8 while bacteremic; 5 of the latter died perioperatively. Of the 19 patients treated medically, 16 died. Group 11: All patients underwent operation in a bacteria-free state. The mortality was 5 % . Group 111: Eight patients had early (<60 days postoperative- ly) and 12 late endocarditis. Total mortality was 40% (71 % early and 2 5 % late mortality). Ten patients underwent reoperation, with a mortality of 20%, compared with 60% in the medically treated group. The results support the indication for early operation in acute endocarditis with progressive cardiac failure and renal failure and prosthetic valve endo- carditis. even during bacteremia. Key words: operation, acute endocarditis, prosthetic valve endocarditis, cardiac failure, bacteremia.

With the effective antibiotics of today, most cases of infective endocarditis can be treated successfully by medical means alone. There is, however, a small group of patients with acute endocarditis who develop serious complications (progressive heart failure, renal failure, embolism) and in addition patients with prosthetic valve endocarditis, whose mortality is very high when medical measures are the sole form of treatment and in whom early surgical intervention with valve replacement has been found to be life-saving ( 1 4 , even during the course of bacteremia ( 5 , 6, 7). The indications for surgical treatment of patients with acute infective endocarditis have thus changed in recent years, and in most thoracic surgical units early operation is now favored, especially in cases of progressive, refractory cardiac failure (8). We therefore considered it important to make a retrospective study of patients with acute or subacute endocarditis treated surgically and patients with endocarditis following prosthetic valve operations. In addition, patients with acute endo- carditis who had not undergone surgery were reviewed retrospectively. Special interest was paid to patients with endocarditis and prosthetic valve endocarditis with a fulminant course.

The purpose of this study was to gain a better idea of the indications for surgery and of the optimum time for performing heart valve operations in patients with endocarditis, especially the acute form.

PATIENT MATERIAL AND METHODS A total of 1574 cardiac valve operations (933 aortic valve, 458 mitral valve and 183 combined aortic and mitral valves) were performed in 1966-82 at the Department of Thoracic Surgery, University Hospital, Uppsala (Fig. I ) . In the great majority of cases mechanical valve prostheses were implant-

' Now at the Department of Internal Medicine, Hudiksvall Hospital, Sweden.

216 P . Andersson et ul. Acta Med Scand 1986; 219

130 Sterile on antibiotics w Persistent bacteremia

No of operations

/.qp TOTAL

1966-68 -70 -12 -74 -76 -78 -80 -82

t MVD A t t

Fig. 1 Fig. 2 Fig. 1 . Number of cardiac valve operations during the study period (1%6-82). AVD = aortic valve disease, MVD = mitral valve disease.

Fig. 2. Results of medical and surgical therapy in acute infective endocarditis. NVE = native valve endocarditis, Op = operated on, t = dead.

ed, the Bjork-Shiley prosthesis predominating. Only in a small number of patients (n= 143) were biological valves used-mainly fascia lata. Of the total number of patients who had undergone valve operations, infective endocarditis had led to the operation in 4.7 % and prosthetic valve endocarditis had developed in 0.6%. Data on patients who had had acute or subacute endocarditis preoperatively and those in whom prosthetic valve endocarditis had occurred after surgery were taken from registers of medical case records of the Department of Thoracic Surgery. In addition, through the University Hospital's collected register of medical records, data from the same period of time were obtained for all patients with medically treated acute endocarditis who had received primary care at different clinics. These patients came from a smaller catchment area than the surgically treated patients. The study population comprised 113 patients who were divided into the following three groups.

Group I: Thirty patients with acute endocarditis (fulminant course with septic fever) with involve- ment of the aortic valve in 13, mitral valve in 10 and both the aortic and mitral valve in 4. All affected valves were insufficient. In one patient the tricuspid valve was involved and in another the pulmonary valve; in one patient the location was unknown.

Group 11: Sixty-three patients with subacute endocarditis-3 1 with involvement of the aortic valve, 18 with involvement of the mitral valve and 14 in whom both these valves were affected.

Group 111: Eighteen patients with prosthetic valve endocarditis following operations in 1966-82, of whom two patients had two episodes of endocarditis during this period. In 13 patients an aortic valve prosthesis, in two patients a mitral valve prosthesis and in five patients both an aortic and a mitral valve prosthesis were infected.

The sex and age distributions, the occurrence of previously known organic heart defects, and blood

Table I . Duta on the patients divided into three groups I. Acute endocarditis ( a ) treated surgically, (b ) not treated surgically. 11. Subacute endocarditis treated surgically. 111. Posthetic valve endocarditis (2 patients with recurrence) VHD = valvular heart disease

Previously Age (Y.) known Blood

No. of Sex VHD culture Group pats. (d/?) Mean Range (+I-) (+/-)

Ia I 1 714 44 1348 417 1110 I b 19 10/9 53 21-83 4/15 ' 1112 I1 63 46/17 51 19-74 37/26 41/22 I11 20 I6/4 48 24-62 1911

Acta Med Scand 1986: 219 Surgery in infective endocarditis 277

Table 11. Occurrence of bacterial strains in the three patient groups Division into patient groups as in Table I

Group

l a I h I1 Ill

Staphylococcus aureus albus

Streptococcus viridans Streptococcus pyogenes Streptococcus fecalis Aspergillus fumigatus Other microorganisms Negative culture No culture report available

10 0 2 0 2 0 3 2 0

4

18 0 3 0 4

22 12

culture results in the three groups are presented in Table I. The bacteriological findings are summa- rized in Table 11.

RESULTS

Group I. Acute endocarditis (Fig. 2 ) ( a ) Surgically treated patients (aortic valve 7 , mitral value 2 , aortic and mitral valve 2 ) . Eleven patients with acute endocarditis underwent cardiac valve surgery in the acute stage of the disease. The valvular damage was verified by echocardiography (with obvious vegetations in seven of nine cases) and/or angiography. All I 1 patients had had refractory, progressive cardiac failure preoperatively. Eight of them had not responded to adequate antibiotic therapy and they were operated upon in the presence of bacteremia. Five of these eight patients died-four perioperatively and one 14 days postoperatively in associ- ation with dislodgement of a prosthesis. The other three patients who underwent surgery during the course of infection withstood the operation well.

One of these patients, who had had therapy-resistant gonococcal infection and aortic insufficiency preoperatively, had protracted fever after operation, but otherwise the postoperative period was fairly free from complications. The three patients who were operated upon during a bacteria-free stage had no complications at all postoperatively.

(h ) Patients not treated surgically. Nineteen patients with acute endocarditis received medical treatment alone, and only three of them survived. These three patients, all with positive blood cultures, showed no obvious signs of cardiac failure despite a newly acquired cardiac murmur and have not as yet required any valve operation. In 10 of the 16 patients who died, the diagnosis was not made until autopsy was performed. They had been treated mainly as unclear cases of sepsis and had received adequate antibiotic therapy but had not responded to the treatment and had died either of the infection or of cardiac failure. In the other six patients, ranging in age from 34 to 80 years, the diagnosis was clear or suspected before death and later confirmed at autopsy. Surgical intervention had been considered, but had not been undertaken for various reasons: Two patients died while waiting for operation, which was planned, because of acute, progressive cardiac failure. Three patients with cardiac failure were considered inoperable (two had pro- nounced renal insufficiency and one was in a very poor general condition). One 40-year- old woman died preoperatively of recurrent cerebral embolism.

278 P . Andersson et al. Acta Med Scand 1986; 219

30 -

10

5

Fig. 3 . Ranking of the patients according to years between the onset of subacute infec- tive endocarditis and cardiac valve opera- n , , a m : m Time interval tion.

The frequency of positive blood cultures was high in groups Ia and I b (Table I). Thus, the bacteriological findings were positive in all surgically treated and in all except two of the medically treated patients, i.e. in totally 28 patients.

Staphylococcus aureus was predominant in both groups and occurred in a total of 16 patients. Streptococcus viridans, pyogenes and fecalis were found in seven cases and pneumococci, gonococci, Klebsiella, Escherichia coli and Bacterium coroides in the other five cases, respectively. Of the surgically treated patients, all except one of those with Staphylococcus aureus infection had involvement of the aortic valve with consequent aortic insuficincy, in two cases combined with mitral insufficiency. In 16 patients a probable portal of entry of the infection was demonstrated, most commonly the skin (7 cases). No fewer than four patients had concurrent meningitis. Only just over one fourth of the patients had a previously known valvular heart lesion.

Group I I . Subacute infectiue endocarditis All 63 patients were operated on in a bacteria-free conditon at a point of time when there were clear indications for prosthetic valve surgery according to our current criteria. Forty per cent of the patients underwent surgery during the first year and two thirds within 3 years after their episode of endocarditis (Fig. 3). The early postoperative mortality in this group was 5 % (3 patients).

As shown in Table 11, the predominant cause in group I1 patients with positive blood culture w a s Streptococcus viridans (18 patients), whereas staphylococci occurred in only four patients. Blood culures were negative in one third of the patients and in one-fifth no culture report was available.

Group I I I . Prosthetic valve endocarditis Nine of the patients in whom a prosthetic heart valve was inserted (6 fascia lata and 3 mechanical prostheses) during the first 8-year period (1966-73) developed prosthetic valve endocarditis, one of them with recurrence. Eleven of those who underwent implantation of a mechanical prosthetic valve during the following 9-year period ( 1974-82) developed this condition, including one who had had an earlier episode of endocarditis in 1972. As the number of valve operations per year more than doubled during the later period (Fig. I ) , the annual incidence of prosthetic valve endocarditis decreased from 2.4 % during the former period to 1 .O % in the latter. For the mechanical valves the incidence was the same during both periods, namely 1 .O %, but for the biological valves it was 6.5 % in the first period and 0% in the second. Most episodes of endocarditis (16 of 20) occurred in patients operated upon during the winter half-year between October and March.

Eight cases of prosthetic valve endocarditis occurred at an early stage, within 2 months

Acta Med Scand 1986; 219 Surgery in infective endocarditis 279

No of potlents

k 1 2 3 L

Fig. 4

I 5 6

Sterile on antibiotics’ ‘Persistent bacteremia

7 8 9 10 11 12 25 t t t t Time interval (months) Fig. 5

Fig. 4. Ranking of the patients according to months between cardiac valve operation and the onset of prosthetic valve endocarditis. Fig. 5 . Results of medical and surgical therapy in active prosthetic valve endocarditis. PVE = prosthetic valve endocarditis, Op = operated on, t = dead.

after operation, and 12 at 3-25 months postoperatively, with the highest incidence during the first year (8 patients) (Fig. 4). Aortic prosthetic valve endocarditis was the most commonly occurring type both among the early and among the late cases (6/8 and 7/12 cases, respectively). In five patients (20%) both an aortic and a mitral prosthesis were infected, and only in two patients did the infection involve the mitral valve alone. Staphylococci were the responsible bacterial agents in 16 of the 20 patients with prosthetic valve endocarditis-Staphylococcus aureus in seven and Staphylococcus albus in nine. Among the early cases of prosthetic valve endocarditis, Staphylococcus aureus was the causal agent in four, Staphylococcus albus in three and Aspergillus fumigatus in one. Staphylococcus albus was predominantly responsible for the late cases of prosthetic valve endocarditis (7 out of 12 cases), whereas Staphylococcus aureus occurred in only three of these cases and Streptococcus viridans and fecalis in one case each.

The treatment followed the principles generally adopted for sepsis therapy. Patients with hemodynamically significant prosthetic dysfunction underwent reoperation with valve replacement; this was necessary in 10 cases. If possible, the operations were performed when the patients had been free from bacteria for a few months, but in three patients progressive cardiac failure made it necessary to operate during the course of bacteremia (Fig. 5 ) . Total mortality among the patients with prosthetic valve endocarditis was 40% (71 % in early and 25 % in late cases). Of the ten patients who underwent surgery, two died (20%), and of ten patients treated with medical measures alone, six died (60%). Prosthetic valve endocarditis caused by Staphylococcus aureus carried the poorest prog- nosis and five of these seven patients died, whereas there was only one death among ten patients with Staphylococcus albus. The prognosis was also poorer among patients with infection of biological than of mechanical prosthetic valves, with mortality figures of 66% (4 of 6 patients) and 28% (4 of 141, respectively.

DISCUSSION Early surgical intervention in acute endocarditis, with replacement of an infected native valve by a valvular prosthesis, has become increasingly advocated in recent years (2 .4 , 6, 9, 10). Our results support this approach. Eleven patients with. infective endocarditis were

280 P. Andersson et al. Acta Med Scand 1986; 219

operated upon during an active stage of the disease in 1974-82 at the University Hospital, Uppsala. Eight of them underwent surgery after 1979, with increasingly better results, and this was also true for patients with prosthetic valve endocarditis. Common to those who died perioperatively or in the immediately postoperative phase was that they were in a very poor, hardly operable state, with pronounced cardiac failure and in some cases renal insufficiency. The latter has been claimed to be the main decisive factor for a poor outcome in prosthetic valve endocarditis (1 1 ) . It is also important that the patients come for operation before the cardiac failure has led to irreversible damage. Even though we found that the most favorable results were obtained at operations performed when the bacteremia had been brought under control with antibiotics, the patient's heart condition sometimes necessitates operation even in the presence of bacteremia-a measure which can be successful, as indicated in our study and also demonstrated by other authors ( 5 ) . Even if the blood cultures are negative, valve tissue samples obtained at operation often yield positive cultures. For example, Stinson (2) reported such cultures in 36% of surgically treated patients with negative blood cultures.

In cases of acute aortic insufficiency due to endocarditis, the diagnosis is not seldom rendered difficult by somewhat atypical bedside findings, for instance a diastolic murmur that is difficult to hear on auscultation and, in cases of cardiac failure, absence of the peripheral pulse phenomenon-an experience which is made not least in prosthetic valve endocarditis. In these cases echocardiography is often very helpful, and in prosthetic valve endocarditis fluoroscopy and cineradiography of the valve prosthesis-as long as it is radiopaque-can be valuable for assessing its motional pattern. In sporadic cases with, for example, involvement of two valves, echocardiograms that are difficult to interpret and a degree of regurgitation that is difficult to assess, supplementary angiocardiography may be required.

Staphylococcus aureus was, not unexpectedly, the predominant causal agent in our patients. Concerning antibiotic therapy in surgical treatment of valvular lesions due to active endocarditis, it has recently been claimed that the risk of recurrence and therewith the long-term prognosis are dependent more upon the duration of postoperative antibiotic therapy (adequate period 4-6 weeks) than upon the preoperative treatment ( I , 10). Most authors agree, however, that ongoing bacteremia increases the risk associated with operation ( 1 , 2, 12). Even in this situation, however, surgery is considered to be the superior method of treatment and is best undertaken before severe cardiac insufficiency, renal failure and the more rare recurrent embolization to the major arteries have occurred ( 5 , 6, 7, 13). Our results support this view. Alstrup and Froysaker (14) found good long- term results in 73 % of their patients with acute aortic insufficiency and cardiac failure who underwent surgery in an active stage of endocarditis.

The patients with subacute endocarditis, all of whom fulfilled current indications for heart valve operation, showed good early results ( 5 % mortality), corresponding to those found in general in elective valve surgery. In this group Streptococcus viridans was the most common bacterial agent.

In our patient series, no single factor could be identified that was associated throughout with postoperative endocarditis, except the fact that the initial symptoms appeared in all eight patients with early prosthetic valve endocarditis after discontinuation of postoper- ative antibiotic therapy. In this context it is interesting that Stinson (2), in his series of 31 patients with prosthetic valve endocarditis, found that 39% of the patients who according to their blood culture results had become free from infection preoperatively, nevertheless showed signs of infection in excised valvular tissues at reoperation. Like Lewis et al. (IS) , he found no relationship between the duration of the preoperative antibiotic therapy and the postoperative course, including the operation mortality. The incidence of prosthetic

Acta Med Scand 1986: 219 Surgery in infective endocarditis 281

valve endocarditis in our series (1.3 %) was of approximately the same order of magnitude as in other series (16). The biological valves showed a notably high frequency of infection during the first half of the study period, which can be explained by the inadequate sterility conditions associated with the shaping of the fascia lata grafts (17). When these conditions had been corrected, the infections ceased altogether.

The mortality in prosthetic valve endocarditis is high and rates of 5 ( M o % have been reported (16, 18). In our cases of prosthetic valve endocarditis, of which Staphylococcus albus and aureus were common causes, the poorest results, as in acute endocarditis, were obtained with medical treatment alone. In that group the mortality was 5 5 % , which may be compared with the figure of 22% in the group which underwent valve replacement. There was a clear difference in prognosis between early (within 2 months after operation) and late prosthetic valve endocarditis, the latter being considerably more favorable. This finding is in accordance with literature reports of mortality figures of 70430% in early and 40-50% in late endocarditis (191, although the mortality rate was lower (25%) among our cases of late endocarditis. Staphylococcus aureus predominated among the early cases of prosthetic valve endocarditis and Staphylococcus albus among the late ones. A predomi- nance of streptococci is often reported in the latter type (16, 19).

Our treatment results in prosthetic valve endocarditis, with a substantially better prognosis with operation and replacement of the infected prosthetic valve than with antibiotic therapy alone, are in accordance with the results in most other published series (6, 19, 20). The role of Staphylococcus epidermidis (albus) in infections in implanted foreign material deserves attention, in view of the increasing frequency of this type of operation in different branches of surgery, including cardiac surgery (2). This bacterial strain of comparatively low virulence can be extremely difficult to master with antibiotics alone and the resistance pattern is often highly unfavorable (21). However, one third of our patients with prosthetic valve endocarditis caused by Staphylococcus albus recovered without operation.

Thus, we have found that in valve operations performed 2 months to several years after a clinically cured subacute endocarditis the mortality is not higher than in other elective valve surgery. In acute endocarditis it is important not to delay operative correction of the valvular damage when complications occur, especially in the form of progressive cardiac failure and renal failure, or when bacteremia cannot be brought under control with antibiotics. Consideration must be paid to this development at an early stage. Since a fulminant course can develop very rapidly (sometimes within 24 hours), as is seen particularly in acute aortic valve endocarditis, as many as possible of these patients should be referred to hospitals with a thoracic surgical unit. This is even more important in cases of prosthetic valve endocarditis, in whom hemodynamic evaluation may be difficult. The point of time at which operation should be performed ought to depend more on the patient’s hemodynamic condition than the degree of activity of the infection or the duration of preoperative antibiotic therapy. The presence of large vegetations, evident at echocardiography , have recently been added to the indications for early operation, on account of the risk of embolization.

REFERENCES I . Jung JY, Saab SB. Almond CH. The case for early surgical treatment of left-sided primary

2. Stinson EB. Surgical treatment of infective endocarditis. Progr Cardiovasc Dis 1979; 22: 145-68. 3. Brandenburg RO, Giuliani ER, Wilson WR, Geraci JE. Infective endocarditis-A 25 year

4. Croft Ch H , Woodward W, Elliott A, Commerford PJ, Barnard Ch N , Beck W . Analysis of

infective endocarditis. A collective review. J Thorac Cardiovasc Surg 1975; 70: 509-18.

overview of diagnosis and therapy. J Am Coll Cardiol 1983; I : 28CL91.

282 P . Andersson et al. Acta Med Scand 1986; 219

5 .

6.

7.

8. 9.

10.

1 1 .

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

surgical versus medical therpay in active complicated native valve infective endocarditis. Am J Cardiol 1983; 51: 1650-5. Wilson WR, Danielson GK, Giuliani ER, Washingon JA, Jaumin PM, Geraci JE. Valve replace- ment in patients with active infective endocarditis. Circulation 1978; 58: 585-8. Richardson JV, Karp RB, Kirklin JW, Dismukes WE. Treatment of infective endocarditis: A 10- year comparative analysis. Circulation 1978; 58: 589-97. Young JB, Welton DE, Raizner AE et al. Surgery in active infective endocarditis. Circulation 1979; 60 (Suppl I ) : 77-81. Dinubile MJ. Surgery in active endocarditis. Clinical review. Ann Intern Med 1982; 96: 650-9. Rapaport E. Editorial: The changing role of surgery in the management of infective endocarditis. Circulation 1978; 58: 598-9. Raychaudhury T. Faichney A, Cameron EWJ, Walbaum PR. Surgical management of native valve endocarditis. Thorax 1983; 38: 168-74. Baumgartner WA, Miller DC, Reitz BA et al. Surgical treatment of prosthetic valve endocarditis. Ann Thorac Surg 1983; 35: 87-104. Wilson WR, Danielson GK, Giuliani ER, Washington JA, Jaumin PM. Geraci JE. Cardiac valve replacement in congestive heart failure due to infective endocarditis. Mayo Clin Proc 1979; 54: 223-6. Rudolph W, Kraus F, Mudra H, Dennig D, Dacian S, Froer KL. Erkennung und Beurteilung der infektiosen Endokarditis. Herz 1983; 8: 241-70. Alsturp P, Froysaker T. Immediate and long-term results of emergency aortic valve replacement in acute bacterial endocarditis. Acta Med Scand 1976; 200: 373-7. Lewis BS, Agathangelou NE, Colsen PR, Antunes M, Kinsley RH. Cardiac operation during active infective endocarditis. Results of aortic, mitral and double valve replacement in 94 patients. J Thorac Cardiovasc Surg 1982; 84: 579-84. Wilson WR, Danielson GK, Giuliani ER, Geraci JE. Prosthetic valve endocarditis. Mayo Clin Proc 1982; 57: 1 5 5 4 1 . Dubiel WT, Hallen A, Johansson L. Aortic valve replacement with frame-supported autologous fascia lata grafts. Scand J Thorac Cardiovasc Surg 1975; 9: 94-107. Ivert TSA, Dismukes WE, Cobbs CG, Blackstone EH, Kirklin JW, Bergdahl LAL. Prosthetic valve endocarditis. Circulation 1984; 69: 223-32. Watanakunakorn Ch. Prosthetic valve infective endocarditis. Progr Cardiovasc Dis 1979;

Kay PH, Oldershaw PJ, Lincoln JCR, Lennox SC, Paneth M. The management of prosthetic valve endocarditis. J Cardiovasc Surg 1983; 24: 127-31. Editorial: Coagulase-negative staphylococci. Lancet 1981 ; I : 139-40.

22: 181-92.

Received March 21, 1985.

Correspondence: E. Enghoff, M.D., Department of Internal Medicine, University Hospital, S-7518.5 Uppsala, Sweden.