valvular endocarditis. One survey of an institutionalized

Transcription

1 Valve Replacement for Left-Sided Endocarditis in Drug Addicts Robert B. Mammana, M.D., Sidney Levitsky, M.D., David Sernaque, M.D., Charles B. Beckman, M.D., and Norman A. Silverman, M.D. ABSTRACT Eighteen drug addicts with left-sided valvular endocarditis requiring operation are reviewed. Gram-positive bacteria were the most common organisms cultured (61% 1, with Staphylococcus aureus present in 7 of 11 patients. Gram-negative bacteria, exclusively Pseudomonas aeruginosa, were cultured in the remaining 39%. Indications for operation included sepsis (61% 1, heart failure (78%), and systemic emboli (22%). Abscesses formed in 6 of 11 patients with gram-positive endocarditis, while only one abscess was present with gram-negative endocarditis. Normal valves were infected in 17 of 18 patients (94%). Early surgical mortality (less than 30 days) was 11%. There were major complications in 79% of these patients, including persistent sepsis (50%), valvular dehiscence, prosthetic endocarditis or perivalvular leakage (37%), and mycotic aneurysms (22%). These complications were directly related to a late mortality of 44%, yielding an overall mortality of 50% in the first nine months after operation. Contrary to previous reports of acceptable surgical survival for valvular endocarditis, these data suggest that endocarditis involving the aortic or mitral valve in a drug addict is a highly lethal disease due to the virulence of the organisms, the severity of the complications encountered, and the predisposition to continued addiction. Despite the large body of recent literature on infective endocarditis [l-61, there are few reports that specifically detail surgical results and complications in drug addicts with left-sided From the University of Illinois Medical Center, Westside Veterans Administration and Cook County Hospitals, Chicago, IL. Presented in part at the American Heart Association Meeting, Nov 16-19, 1981, Dallas, TX. Accepted for publication Apr 13, Address reprint requests to Dr. Mammana, Assistant Professor, Department of Surgery, Division of Cardiothoracic Surgery, University of Arizona Health Sciences Center, 1501 N Campbell Ave, Tucson, AZ valvular endocarditis. One survey of an institutionalized addict population showed infective endocarditis to be the cause of death in 14% (18/128) of postmortem examinations performed [7]. Since the exact number of addicts in the United States is not known, the incidence of potential endocarditis, though theoretically great, cannot be established. Many surgical reports suggest that operative intervention during active infection can result in acceptable survival [2, 5, 8, 91, although the presence of major congestive heart failure may limit survival in both medically and surgically treated patients [l]. Most of these reports, however, concern nonaddicts with underlying heart disease who develop endocarditis. In the reports describing endocarditis in drug addicts, right-sided and left-sided valvular endocarditis are considered as a single entity, even though they are different diseases. Furthermore, several reports [lo, 111 have suggested an increase in the incidence of gram-negative organisms causing endocarditis in drug addicts; predominantly gram-positive organisms had been cultured in the past. The purpose of this review is to describe the surgical results and complications of left-sided valvular endocarditis in a population of drug addicts. Materials and Methods The cases of all individuals undergoing valve replacement at the University of Illinois Hospital, Cook County Hospital, and Westside Veterans Administration Medical Center in Chicago, IL, from January 1, 1976, to December 31, 1979, were reviewed. Eighteen patients who had undergone left-sided valvular procedures and who were known narcotic addicts were studied in detail. Follow-up was accomplished by examination of each patient, review of clinic records, and telephone interviews when possible. Because of the transient nature of the addict 436

2 437 Mammana et al: Valve Replacement for Endocarditis in Addicts Table 1. Characteristics of Study Population (N = 18) Duration Patient Age Organism of Symptoms Valve Antibiotic Complications Postoperative No. (yr) Cultured (days) Procedure(s) Therapy at Operation Outcome S. aureus 30 S. aureus 17 S. aureus 40 S. aureus 42 S. aureus 37 S. aureus 29 S. aureus 20 S. viridans 26 N. gonorrhoeae 24 N. gonorrhoeae 27 Enterococcus 39 P. aeruginosa AVR K, 0, CI for 6 wkpreop + 2 wk AVR G, 0 for 2 wk preop + 5 wk MVR N for 6 wk AVR AVR P. aeruginosa 49 MVR P. aeruginosa 47 P. aeruginosa P. aeruginosa 14 AVR P. aeruginosa 14 MVR P. aeruginosa 30 AVR P for 2 wk preop + 6 wk P, G for 6 wk G for 5 wk preop + C for 4 wk G for 5 wk preop + C, Ti, To for 4 wk MVR C, G, Ti, To for 10 wk AVR + MVR To, C for 12 wk C, G for 6 wk C for 2 wk preop + G for 8 wk C for 2 wk preop + G for 4 mo Signed out of hospital AMA 2 wk ; Transient low Alive & well output Low output + Infected hepatic renal failure arterv aneurvsm requiring ligation; AVR + MVR 0 for 6 wk died 6 mo of PVE AVR 0 for 31 days died on POD 29 of ruptured mycotic aortic arch aneurysm AVR MVR 0 for 6 wk preop and N for 6 wk preop Transient renal failure; Alive & well ' and AVR P for 6 wk Alive; has perivalvular leak & refuses further operations MVR P for 6 wk Late death at 6 mo of S. aureus PVE MVR Dense preop left hemiplegia; Bleeding dyscrasia requiring 40 units of blood transfusion Late death at 8 mo of PVE died 1 mo of valvular dehiscence signed out AMA 4 wk ; returned 2 wk later with valvular dehiscence & died Craniotomy for brain abscess at 6 wk ; late death at 3 mo of dehiscence of both valves Alive & well; nonaddicted; splenic abscess Late death at 2 mo of persistent sepsis late death at 4 mo of infected aneurysm of aortic suture line AVR = aortic valve replacement; MVR = mitral valve replacement; K = Keflin (cephalothin); 0 = oxacillin; CI = Cleocin hydrochloride (clindamycin); G = gentamicin; N = nafcillin; P = penicillin; C = carbenicillin; Ti = ticarcillin; To = tobramycin; AMA = against medical advice; POD = erative day; PVE = prosthetic valve endocarditis.

3 438 The Annals of Thoracic Surgery Vol 35 No 4 April 1983 population, 2 of the 9 surviving patients could not be contacted. However, family members confirmed that both were alive and apparently in good health (Patients 1 and 2). Table 1 outlines the clinical characteristics of this population. Results The study group consisted of 17 men and 1 woman, ranging in age from 17 to 52 years (mean, 33.2 years). Despite a greater mean age for patients with gram-negative endocarditis (41.2 years) compared with that for patients with gram-positive endocarditis (28.5 years), no statistically significant difference was noted. The duration of symptoms ranged from 12 days to one year for the gram-positive group and 12 to 75 days for the gram-negative group. Eleven of 18 patients (61%) had a history of infection (fever, chills, anorexia, night sweats) for 30 days or more prior to hospitalization. Four patients (Nos. 3, 4, 9, and 14) claimed to have had symptoms for periods ranging from 2Y2 months to one year. Bacteriology GRAM-POSITIVE INFECTION. Eleven of 18 patients (61%) had gram-positive endocarditis, with Staphylococcus aureus the predominant organism in 7 of 11 patients (64%). Among the remaining patients, 1 was infected with Streptococcus viridans, 1 with Streptococcus faecalis (enterococcus), and 2 with Neisseria gonorrhoeae. Patient 9 originally had gonococcal endocarditis but subsequently died of S. aureus prosthetic valve endocarditis. For this report he was considered to have had gonococcal endocarditis and has not been included in the S. aureus group. GRAM-NEGATIVE INFECTION. All patients with gram-negative endocarditis were infected by Pseudomonas aeruginosa (7/18) (39%). No patients in this series had Candida albicuns or any other type of fungal endocarditis. There was no significant difference between the incidence of gram-positive and gram-negative endocarditis. Operative lndications The indications for operation overlapped, with sepsis the most common indication in 11 of 18 patients (61%). Sepsis was present in 6 of the 11 patients in the gram-positive group (55%), and in 6 of the 7 patients in the gram-negative group (86%). Cardiovascular collapse was the primary indication for operation in only 2 patients, although signs of cardiac decompensation (rales, gallop rhythm, dyspnea, elevated left ventricular end-diastolic pressure) were present in 13 patients (72%): 9 (82%) with gram-positive endocarditis and 4 (57%) with gram-negative endocarditis. Systemic emboli were the least common indication for operation, occurring in 4 patients (22%): 3 (27%) with gram-positive endocarditis and 1 (14%) with gram-negative endocarditis. There was no statistically significant difference between the two groups in terms of indications for operation. Valve Replacement Seventeen of 18 of the valves excised (94%) were previously normal. Patient 8 had a bicuspid aortic valve at surgical removal. Periannular or myocardial abscess was present in 7 patients (39%): more so in gram-positive infections (6/11; 55%), than in gram-negative infections (1/7; 14%). Aortic valve replacement (AVR) was performed in 9 of 18 patients (50%), 6 of whom were in the gram-positive group and 3 of whom were in the gram-negative group. Mitral valve replacement (MVR) was performed in 7 patients (39%), 4 with gram-positive endocarditis and 3 with gram-negative endocarditis. Combined AVR and MVR was required in 2 patients (ll%), 1 with gram-positive endocarditis and 1 with gram-negative endocarditis. Early and Late Mortality Early surgical mortality (less than 30 days) was 11% (2/18); late mortality was 44% (7/16). Overall mortality was 50% (9/18), with gram-positive endocarditis fatal in 4 patients (36%) and gramnegative endocarditis fatal in 5 others (71%). All deaths occurred within the first nine months after surgery. No statistical significance could be established as to relationship between valve replacement and mortality: AVR, 4/9 (44%); MVR, 317 (43%); combined replacement, 2/2 (100%). A course of antibiotic treatment was com-

4 439 Mammana et al: Valve Replacement for Endocarditis in Addicts pleted in all but 2 patients (Nos. 1 and 13, who left the hospital against medical advice). In the remaining 16 patients, antibiotics were continued for a minimum of 6 weeks unless death occurred while on treatment. Patient 5 died on erative day 29 (POD 29), and Patient 12 died on POD 30 (see Table 1). The remaining 14 patients completed their antibiotic treatments. No correlation could be established between the duration of antibiotic therapy and survival. Cornplica tions Major complications were encountered. Persistent sepsis was present in 9 of 18 patients (50%) and fatal in 7 (77%). Infected aneurysms were found in 4 of 18 patients (22%), 2 of whom died (50%). Valvular dehiscence, prosthetic endocarditis, or perivalvular leakage occurred in 7 of 18 patients (39%), 6 of whom died (86%). One patient (No. 8) had a perivalvular leak but refused further operative intervention. Patient 14 had an abscess of the brain requiring craniotomy, but this was nonfatal. Five of 9 survivors (56%) admitted to continued drug addiction. Comment Despite the advances in valve substitutes and antibiotics, valvular endocarditis still remains a great technical challenge to cardiovascular surgeons. Reports over the past decade have demonstrated that surgery can be performed safely in the presence of congestive heart failure [l] and active infection [5, 9, 111. Little, however, is known about left-sided valvular replacement for endocarditis in a population of narcotic addicts. A review of the literature that deals selectively with aortic and mitral valvular endocarditis (Table 2) shows a widespread variation in the outcome of endocarditis in addicts. Most reports consider right-sided and left-sided endocarditis collectively; however, these are distinct entities with differing mortalities, ranging from 0 to 25% for right-sided involvement and 23 to 100% for left-sided involvement. When data presented by Banks and colleagues [12] were analyzed, right-sided endocarditis was found in 38 of 50 addicts (76%); however, leftsided endocarditis was present in only 12 of 50 addicts (24%). The reported mortality for rightsided endocarditis was 15.8%, compared with a 57% mortality for left-sided endocarditis. Those addicts who underwent operation had a mortality rate of 100%. In that series, as in ours, s. aureus was present in over 50% of the patients. However, only 1 patient had gram-negative infection. Other series have confirmed that s. aureus is the most common gram-positive organism, with incidences ranging from 25% [13] to 68% [14]. S. viridans, the next most common gram-positive organism encountered, ranges in incidence from 30 to 37% [13], although we observed it only once. In our series normal valves were involved in 17 of 18 (95%) episodes of endocarditis. This compares favorably with other series, in which native valves were involved 63% [l], 85% [15], and 86% [9] of the time. The aortic valve was involved 50% of the time in our series, a figure corresponding to other reported series in which the aortic valve was involved in 51% [16] and 61% [9] of patients. The mitral valve was involved less frequently (37%), and combined valvular involvement was the least common (11%). Periannular and myocardial abscess formation is common with gram-positive endocarditis and was present in 57% of the patients reported by Hiratzka and associates [9]. Similarly, in our series periannular or myocardial abscess was present in 45% of patients with gram-positive endocarditis and accounted for death in 50%- similar to a mortality of 45% reported by Richardson when a myocardial abscess was present [16]. Abscess formation is uncommon with gram-negative endocarditis and was seen in only one instance in our series. The surgical mortality for left-sided grampositive endocarditis is 0 [13], 23% [15], 33% [3], 66% [4], and 100% [ll, 12, 141, depending on the series reviewed. In our 18 patients the mortality for left-sided gram-positive endocarditis was 36% and was similar to the results of previous studies. Despite the prevalence of gram-positive endocarditis in addicts and nonaddicts, the emergence of gram-negative endocarditis in addicts is causing great concern in metropolitan centers [lo, 111 due to the high associated mortality, inability to identify a common source, and inadequate knowledge of proper surgical and medical therapy. With the exception of the

5 440 The Annals of Thoracic Surgery Vol 35 No 4 April 1983 Table 2. Selective Review of the Literature on Endocarditis in Addicts No. of Addicts Type of Treatment with Involve- Operative ment of AV, Medical No. of Medical/ No. of Mortality Source MV, or Both Only Deaths Surgical Deaths (%) Comment Ramsey et a High incidence ( ) [13] of S. viridans Chicago, IL (37%); S. nureus present in 25% Stimmel et a S. aureus present ( ) [15] in 69%; no New York, NY gram-negative Dreyer et a S. aureus present ( ) [14] in 68%; 1 gram- New York, NY negative infection with 1 death; no P. aeruginosa present Banks et a S. aureus present ( ) [12] in 56%; 1 P. Washington, DC aeruginosa resulting in death Boyd et a S. aureus in 48%; ( ) [3] P. aeruginosa New York, NY in 3 patients with 2 deaths Reyes et a All patients had ( ) [ll] P. aeruginosa Detroit, MI endocarditis Pelletier et a S. aureus in 56%; ( ) [4] no P. aeruginosa Seattle, WA Hiratzka et a Four deaths due to ( ) [9] surgery; 3 due Los Angeles, CA to drug overdose; 1 due to aortic aneurysm (not infected) Mammana et a S. aureus present (this report) in 61%; P. aeruginosa in 39% aoverau mortality is taken from the combined deaths for patients having medical treatment only and those who had medicalsurgical treatment. AV = aortic valve; MV = mitral valve. series reported by Archer [lo] and Reyes Ill] and their co-workers, Pseudomonas endocarditis is an uncommon entity, noted primarily in case reports [3, 12, 171. Right-sided Pseudomonas endocarditis, though frequently lethal, can be cured using a combination of medical and surgical therapy [ 111. Left-sided Pseudomonas endocarditis, however, has a reported mortality of 87.5% [ll] and 100% [lo]. In our series Pseudomonas endocarditis was found in 39% of pa- tients, accounting for a mortality of 71% and due primarily to persistent uncontrollable sepsis, although a large number of other lethal complications occurred. Despite the apparent severity of Pseudomonas endocarditis, we were unable to prove statistically that any difference exists between gram-positive and gram-negative endocarditis. The short-term operative mortality was 11% (2/18), and the late operative mortality was 44%

6 441 Mammana et al: Valve Replacement for Endocarditis in Addicts (7/16), giving an overall mortality in the first nine months of 50% (9/18). The complications contributing to late mortality include continued drug addiction, persistent sepsis, valvular dehiscence, prosthetic endocarditis or periannular leakage, and infected aneurysms. We conclude from these data that bacterial endocarditis secondary to intravenous drug abuse is markedly more virulent than endocarditis in individuals who do not use drugs. Several features of the disease can be cited for these distinctions. First, the patients themselves may be less reliable and more inclined to procrastinate in obtaining medical care, thereby increasing the possibility of metastatic abscess formation and continued erative sepsis. Moreover, the nature of the bacteriological flora encountered, with the high incidence of gramnegative organisms frequently resistant to currently available antibiotics, would seem to preclude successful surgical cure. This is borne out by the low perioperative mortality reported as compared with the excessive patient attrition over the first year; mortality correlated primarily with persistent infection (prosthetic endocarditis and dehiscence as well as mycotic aneurysms). Others have noted similar long-term mortality and complications [18]. Contrary to previous reports of acceptable surgical survival for valvular endocarditis, our study suggests that left-sided endocarditis in drug addicts is a highly lethal disease by virtue of the organisms encountered, the presence of myocardial abscess formation, the severity of the complications encountered, and the prevalence of continued refractory sepsis. References 1. Utley JR, Mills J, Hutchinson JC, et al: Valve replacement for bacterial and fungal endocarditis. Circulation 47:Suppl3:42, Stinson EB, Griepp RB, Vosti K, et al: Operative treatment of active endocarditis. J Thorac Cardiovasc Surg 71:659, Boyd AD, Spencer FC, Isom OW, et al: Infective endocarditis: an analysis of 54 surgically treated patients. J Thorac Cardiovasc Surg 73:23, Pelletier LL, Petersdorf RG: Infective endocarditis: a review of 125 cases from the University of Washington Hospitals, Medicine 56: 287, Wilson WR, Danielson GK, Giuliani ER, et al: Valve replacement in patients with active infective endocarditis. Circulation 58:585, Richardson JV, Karp RB, Kirklin JW, Dismukes WE: Treatment of infective endocarditis: a 10- year comparative analysis. Circulation 58:589, Sapira JD, Ball JC, Penn H: Causes of death among institutionalized narcotic addicts. J Chron Dis 22:733, Wallace AG, Young WG Jr, Osterhoust S: Treat- ment of acute bacterial endocarditis by valve excision and replacement. Circulation 31:450, Hiratzka LF, Nelson RJ, Oliver CB, Jengo JA: Operative experience with infective endocarditis: drug users compared with non-drug users. J Thorac Cardiovasc Surg 77:335, Archer G, Fekety FR, Supina R: Pseudornonas aeruginosu endocarditis in drug addicts. Am Heart J 88:570, Reyes MP, Palutke WA, Wylin RF, Lerner AM: Pseudornonns endocarditis in the Detroit Medical Center, Medicine 52173, Banks T, Fletcher R, Ali N: Infective endocarditis in heroin addicts. Am J Med 55:444, Ramsey RG, Gunnar RM, Tobin SR Jr: Endocarditis in the drug addict. Am J Cardiol 25:608, Dreyer NP, Fields BN: Heroin-associated infective endocarditis: a report of 28 cases. Ann Intern Med 78:699, Stimmel B, Donoso E, Dack S: Comparison of infective endocarditis in drug addicts and non-drug users. Am J Cardiol 32924, Factor S, Frishman W: Sudden death in a narcotic addict four months following aortic valve replacement. Am Heart J 98233, 1979 Graham DY, Reul GJ, Martin R, et al: Infective endocarditis in drug addicts. Circulation 47 Suppl337, 1973 Hubbell G, Cheitlin MD, Rappaport E: Presentation, management, and follow-up evaluation of infective endocarditis in drug addicts. Am Heart J 102:85, 1981