Antibiotic Prophylaxis in Open-Heart Surgery:

Antibiotic Prophylaxis in Open-Heart Surgery: A Comparison of Cefamandole, Cefuroxime, and Cefazolin Layne. Gentry, M.D., Barry J. Zeluff, M.D., and Denton A. Cooley, M.D. ABSTRACT We undertook a prospective randomized evaluation of cefamandole nafate, cefuroxime sodium, and cefazolin sodium as prophylaxis in open-heart operations. A total of 93 patients having an elective procedure were enrolled in the study, and 62 of them were eventually considered evaluable. There were no significant differences between the three study groups. The overall rate of infection and the rate of infection according to demographic variables was not significantly different between the three antibiotics. Serious complications and deaths were also similar between the three agents. The presence of multiple severe underlying conditions was a risk factor for infection, independent of the antibiotic used. We conclude that there are no differences in the efficacies of the three agents in preventing postoperative infections in patients having open-heart operation. Cefuroxime, principally because of its every-12-hour dose, is far less expensive than cefamandole or cefazolin. Antibiotic prophylaxis in open-heart surgery was used for one of the first times in 1963 [l], in an effort to prevent an unacceptably high incidence of prosthetic heart valve endocarditis, later documented in 1965 [2]. Initially, penicillin and streptomycin were used to combat gram-positive skin flora, especially Staphylococcus aureus and S. epidermidis, opportunistic to patients undergoing sternotomy. Between 1964 and 1977, much folklore was accumulated regarding the use of antibiotics in preventing infections. In particular, the benefit of prophylaxis for patients undergoing aorta-coronary artery bypass procedures was still unclear in 1978. In 1979, using rigid criteria for an evaluable study of prophylaxis [3], a prospective double-blind drug-versusplacebo trial proved that preoperative treatment with methicillin sodium prevented infections in patients undergoing aorta-coronary artery bypass procedures [4]; in fact, none of the methicillin-treated patients had a postoperative infection, while 21% of the patients who received the placebo did. There have been reports of low-risk open-heart operations without antibiotic prophylaxis [5]. However, these procedures utilized continuous intrapericardial lavage with an antibiotic-laden irrigant. Although we prefer intravenous administration because of the improved dosage accuracy, continuous From the Infectious Diseases Section, St. Luke s Episcopal Hospital, and the Texas Heart Institute, Houston, TX. Accepted for publication Jan 22, 1988 Address reprint requests to Dr. Layne Gentry, St. Luke s Episcopal Hospital, 672 Bertner 2-233, Houston, TX 773. antibiotic irrigation has been shown to result in adequate serum levels for prophylaxis [6]. In 1985, another drug-versus-placebo trial was reported [7]. This trial was discontinued because of a 24% rate of infection in the placebo group. Perhaps investigators now realize that the need for antibiotic prophylaxis in open-heart surgery is clearly established. Drug treatment and selection have been extensively examined. In 1961, it was demonstrated that preoperative dosing would be required for prophylaxis [8]. Studies published in 1972 indicated that prophylactic antibiotic treatment past the fourth postoperative day would be unnecessary [9]. As for the selection of antimicrobials, it was believed that a penicillin in combination with an aminoglycoside would provide adequate protection. In 1973, it was shown that targeted prevention of gram-positive infections might lead to a significant increase in the incidence of gram-negative infections because of the lack of activity of the antibiotic against gram-negative pathogens [lo]. The cephalosporins, with in vitro activity against both gram-positive and gram-negative organisms, penetration into heart muscle, and low toxicity, were obvious choices for prophylaxis [ll]. At our institution, cephalothin sodium, 1. gm 3 minutes preoperatively and every 6 hours for five postoperative days, was used as antibiotic prophylaxis in open-heart operations from 1964 to 1976, when we switched to cefamandole nafate, at the same dose, because of the enhanced in vitro activity of cefamandole against many pathogens [12]. In 1985, after preliminary reports that cefuroxime sodium and cefazolin sodium demonstrated in vivo activity comparable to that of cefamandole and at a much reduced cost [13], we prospectively evaluated these three drugs in a randomized study. Material and Methods Between 1985 and 1987, all patients undergoing elective open-heart operations were considered for the study. By elective, we mean that the patient was admitted to the hospital for an open-heart procedure. Patients were excluded from the study if they were less than 18 years old, allergic to penicillin, or experiencing an infection that might require concurrent antimicrobial therapy. A total of 93 patients were initially entered into the study. Those patients who were included in the study received cefamandole nafate (Mandol; Eli Lilly and Company), cefuroxime sodium (Zinacef; Glaxo Inc.), or cefazolin sodium (Kefzol; Eli Lilly and Company) in a random manner according to the month of operation. 167 Ann Thorac Surg 46167-171, Aug 1988. Copyright 1988 by The Society of Thoracic Surgeons

168 The Annals of Thoracic Surgery Vol 46 No 2 August 1988 All surgical procedures were performed by the same surgical team at the Texas Heart Institute, and inpatient care was in the postoperative recovery unit, intensive care centers, and inpatient rooms at St. Luke's Episcopal Hospital. The study protocol, approved by the institutional review board at St. Luke's Hospital, called for intravenous antibiotic administration 3 minutes prior to operation and at regular intervals for 48 hours after the procedure. The schedule for cefamandole was 2. gm preoperatively and every 6 hours postoperatively; for cefuroxime, 1.5 gm preoperatively and every 12 hours postoperatively; and for cefazolin, 2. gm preoperatively and every 8 hours postoperatively. Doses were increased to elevate serum levels to the mean inhibitory concentration (MIC) of staphylococci isolated in our institution. In 1984, the 48-hour MIC9 of S. aureus was 16 pg/ml for cefamandole, 16 pg/ml for cefuroxime, and 32 pg/ml for cefazolin, and the 48-hour MIC9 of S. epidermidis was 16 pg/ml for cefamandole, 32 pg/ml for cefuroxime, and 32 pg/ml for cefazolin. Intraoperative serum levels of antibiotics were randomly measured, and laboratory data were used to monitor postoperative liver and kidney function. A complete review of medical records assured compliance with the protocol. An "infection" was defined as clinical signs and symptoms of infection, with confirmation by the isolation in culture of a potential pathogen. Each infection was classified as "surgery related' or "nosocomial," according to the site and the pathogen isolated. Surgeryrelated infections were wound infections, endocarditis, or catheter-associated bacteremias; nosocomial infections were urinary tract infections or respiratory infections. An "infectious death" was defined as a death due in part to a culture-proven infection. A total of 283 patients entered into the study were later deemed inevaluable. Recognizing that length of preoperative stay is a significant risk factor for gram-negative nosocomial infectious complications, we excluded 159 patients whose operation occurred more than 6 hours after admission. Others excluded were 77 patients who received other antimicrobials concurrent with the prophylaxis regimen, 31 patients for whom the prophylaxis regimen was erroneously extended by one or two doses, and 16 patients for whom the prophylaxis was prematurely discontinued because of evidence of liver or kidney failure. The study concluded with a total of 62 evaluable patients-more than 2 evaluable patients for each of the three antibiotics. Although it would have been necessary to evaluate more than twice this many patients to rule out significant type I1 error for infection rates less than 5%, a study that large was beyond our practical ability to perform. Results Shown in Table 1 are the patient populations and antibiotic levels for the three study groups. Using a chi-square test of statistical significance, there were no significant differences in patient demographics, predisposing diagnosis, or procedures between the three groups (p >.5). Because of a concurrent study that would later validate the 2.-gm dose of cefamandole [14], more of the cefamandole patients were treated in 1985. A "predisposing diagnosis" was defined as the presence of multiple severe underlying conditions, such as diabetes mellitus, myocardial infarction, congestive heart failure, or immunodeficiency. The intraoperative serum antibiotic concentration levels verified that the preoperative dose of each drug was appropriately administered, and the levels we measured were compatible with the serum half-lives of the three antibiotics. Detailed in Table 2 are the infectious complications. Of the 62 patients, a postoperative infection developed in 45 (7%). By classification, 22 infections were surgery related and 23, nosocomial. Of the 22 surgery-related infections, there were 13 wound infections, 2 instances of endocarditis, and 7 instances of bacteremia. Using another chi-square test, there were no differences in the overall infection rates between patients with cefamandole, cefuroxime, or cefazolin. Also, there were no differences between the three groups in the rates of surgery-related or nosocomial infections, or in the types of pathogens isolated. In other words, no data were collected to indicate any significant difference in the prophylactic benefits of the three agents. There were no adverse reactions to the antibiotics requiring a discontinuation of the regimen, and there were no deaths directly related to the operation. The 2 deaths in the cefamandole patients occurred in 1985, and were due to nosocomial infections caused by Pseudomonus aeruginosa and Citrobacter diversus. The 1 death among the cefazolin patients occurred in 1986, and was due to a nosocomial infection caused by Citrobacter diversus. Despite in vitro sensitivity data to the contrary, none of the three drugs appeared clinically active against methicillin-resistant S. epidermidis, a pathogen of increasing concern in our institution. We also estimated the average increase in the length of stay in the hospital attributable to a postoperative infection, and found that a postoperative infection adds, on average, five days to the patient's stay. Presented in Table 3 are the rates of infection by age, sex, weight, diagnosis, procedure, and pump time. There were no significant differences in the infection rates between the three agents for any of the variables listed (p >.5). However, the rates for those patients more than 6 years old or for those on bypass for longer than 6 minutes were markedly elevated. With repetitive univariate analysis, the only significant risk factor, independent of prophylaxis, for our patients was the diagnosis of multiple severe underlying conditions. Shown in Table 4 are the costs of administering the drugs according to the approved protocol, the average

169 Gentry et al: Antibiotic Prophylaxis Comparison Table 1. Populatiolr Profiles and Antibiotic Levelsa Variable Cefamandole Cefuroxime (N = 217) (N = 23) Cefazolin (N = 2) Overall (N = 62) Age (yr) >6 <6 Sex Male Female Weight (kg) >SO <SO Diagnosis >I uc <1 uc Procedure ACAB Valve Both Other Pump time (min) >6 <6 Dosage Serum levels (kg/ml) 118 (54) 99 (46) 175 (81) 42 (19) 11 (51) 17 (49) 86 (4) 131 (6) 17 (78) 25 (12) 9 (4) 13 (6) 27 (12) 19 (88) 2. gm q. 6h. 13 95 (47) 18 (53) 151 (74) 52 (26) 96 (47) 17 (53) 1 (49) 13 (51) 161 (79) 25 (12) 11 (5) 6 (3) 79 (39) 124 (61) 1.5 gm q. 12h. 125 154 (77) 46 (23) 17 (53.5) 93 (46.5) 99 (49.5) 11 (5.5) 149 (74.5) 29 (14.5) 11 (5.5) 11 (5.5) 2. gm q. 8h. 15 313 (51) 37 (5) 48 (77) 14 (23) 313 (51) 37 (5) 285 (46) 335 (54) 48 (77) 79 (13) 31 (5) 3 (5) 26 (33) 414 (67) NA NA "Numbers in parentheses are percentages. UC = severe underlying condition, such as diabetes mellitus, myocardial infarction, or congestive heart failure; ACAB = aorta-coronary artery bypass; NA = not applicable. Table 2. Infectious Complications" Variable (N = 217) (N = 23) (N = 2) (N = 62) All infections 16 (7) 14 (7) 15 (7.5) 45 (7) Operation relatedb 8 (4) 6 (3) 8 (4) 22 (4) Staphylococcus species 5 (2) 4 (2) 7 (3.5) 16 (3) Pseudomonas species 2 (.9) l(.5) 3 (.5) Enterobucter species Citrobacter species l(.5) 1 (.5) 1 (.5) 2 (.3) 1 (.2) Nosocomiap 8 (4) 8 (4) 7 (3.5) 23 (4) Escherichia species 3 (1) 3 (2) l(.5) 7 (1) Klebsiella species 4 (2) l(.5) 5 (.8) Enterococci 3 (1) 2 (1) 5 (.8) Pseudomonas species l(.5) l(.5) 2 (1) 4 (.6) Citrobacter species l(.5) l(.2) Proteus species 1 (.5) 1 (.2) Infectious deaths 2 (Ps, C) 1 (C) 3 (.5) Increase in LOS (d) 4.7 5.1 5.3 5. "Numbers in parentheses are percentages. bthirteen had wound infection, 2 had endocarditis, and 7 had bacteremia 'These were urinary tract or respiratory infections. LOS = length of stay (estimated average increase in length of stay because of a postoperative infection); Ps = Pseudomonus; C = Citrobucter.

17 The Annals of Thoracic Surgery Vol 46 No 2 August 1988 Table 3. Infection Rates and Risk Factorsa Variable (N = 217) (N = 23) (N = 2) (N = 62) Age (yr) >6 9 (4, 5) 1 (5, 5) 12 (7, 5) 1 (5, 5) < 6 4 (3, 1) 4 (1, 3) 3 (1, 2) 4 (2, 2) Sex Male 7 (4, 3) 6 (3, 3) 2 (1, 1) 5 (3, 2) Female 9 (2, 7) 1 (4, 6) 6 (4, 2) 8 (3, 5) Weight (kg) > 8 9 (5, 4) 5 (3, 2) 7 (3, 4) 7 (4, 3) < 8 6 (2, 4) 9 (3, 6) 8 (5, 3) 8 (3, 5) Diagnosis >1 UCb 11 (6, 5) 12 (5, 7) 14 (7, 7) 12 (6, 6) <1 uc 5 (2, 3) 2 (1, 1) 1 (1, ) 2 (1, 1) Procedure ACAB 7 (4, 3) 6 (3, 3) 8 (5, 3) 7 (4, 3) Valve 16 (4, 12) 6 (3, 3) 7 (2, 5) Both 36 (9, 27) 9 (, 9) 15 (3, 12) Other 9 (, 9) 3 (, 3) Pump time (min) > 6 15 (4, 11) 1 (4, 6) 8 (4, 4) 11 (4, 7) < 6 7 (4, 3) 4 (2, 2) 7 (4, 3) 6 (3, 3) All data are shown as percentages. The first number is the total percentage in the specific subcategory; the two numbers in parentheses are the percentage with an operation-related infection and a nosocomial infection, in that order. bthis was a significant risk factor. UC = severe underlying condition; ACAB = aorta-coronary artery bypass. Table 4. Economic Considerations Variable (N = 217) (N = 23) (N = 2) (N = 62) Treatment 2. gm, 9 doses 1.5 gm, 5 doses 2. gm, 7 doses NA Cost of drug (incl. admin.) $29 $145 $225 $222 Infection rate wio drug (%) 2 2 2 2 Increase in LOS (per patient) (d)b 5 5 5 5 Cost of care (per patienty $5 $5 $5 $5 Cost savings (per patient) $21 $355 $275 $278 This is based on the minimum estimated infection rate last reported without prophylaxis [5]. bthis is the average estimated increase in LOS because of postoperative infection. The average for the entire study group was 5 days. This is based on the minimum estimated daily charge at our institution of $5 per day. NA = not applicable; LOS = length of stay. increased length of hospital stay associated with a postoperative infection, and the computed economic impact of antibiotic prophylaxis. At our institution, antibiotic prophylaxis for open-heart procedures is not only clinically mandated, but cost-effective as well. Comment The scientific evidence clearly indicates that open-heart operations without antibiotic prophylaxis would be ill advised. Our data reveal no significant differences in the efficacies of cefamandole, cefuroxime, and cefazolin in the prevention of postoperative infections for patients having an open-heart procedure. This conclusion essentially substantiates recent observations of other investigators [15,16]. While our research was unable to rule out type I1 error, our data nonetheless contribute to the conclusion that there is no in vivo evidence whatsoever that any of the three drugs offers markedly superior performance as antibiotic prophylaxis in open-heart surgery * Since cefamandole, cefuroxime, and cefazolin are equally safe and effective, the surgeon, in committee with the infectious disease consultant and the pharmacist, may make a decision on which antibiotic prophylaxis to use based on economic considerations. At our

171 Gentry et a1 Antibiotic Prophylaxis Comparison institution, we now use cefuroxime as the antibiotic prophylaxis of choice in open-heart surgery. Future Requirements Each year in our institution, more than 1,2 organisms are isolated from more than 1, blood cultures, and of these organisms, 28% are S. epidermidis. Between 1985 and 1987, our surveillance of the susceptibility of pathogens revealed that the percentage of isolates of s. epidermidis resistant to methicillin rose from 16 to 61% in our institution. It is now apparent that any new agents to be evaluated for prophylactic use must show enhanced activity against this troublesome organism, to improve on the performance of the cephalosporins. Vancomycin is the current drug of choice against methicillin-resistant staphylococci, yet it is hoped that newer broad-spectrum, less-toxic antimicrobials will provide increased protection to the patient undergoing an openheart operation. This research was supported by grants from Eli Lilly and Company, Indianapolis, IN, and Glaxo Inc., Research Triangle Park, NC. We thank Clayton Vernon for the preparation of the manuscript. References Slonim R, Litwak RS, Gadboys HL, Ehrenkranz NJ: Antibiotic prophylaxis of infections complicating open-heart operations. Antimicrob Agents Chemother 3:731, 1963 Herr R, Starr A, McCord CW, Wood JA: Special problems following valve replacement. Ann Thorac Surg 1:43, 1965 Chodak GW, Plaut ME: Use of systemic antibiotics for prophylaxis in surgery. Arch Surg 112:326, 1977 4. 5. 6. 7. 8. 9. 1. 11. 12. 13. 14. 15. 16. Fong IW, Baker CB, McKee DC: The value of prophylactic antibiotics in aorta-coronary bypass operations. J Thorac Cardiovasc Surg 78:98, 1979 Sutherland RD, Martinez HE, Guynes WA: Antibiotics and coronary artery bypass operations: an 11-year experience. Infect Surg 4:585, 1985 Raine PA, Courtis GR, Fowler R, et al: The intraperitoneal use of cefazolin. J Surg Res 25232, 1978 Penketh AR, Wansbrough-Jones MH, Wright E, et al: Antibiotic prophylaxis for coronary artery bypass graft surgery. Lancet 1:15, 1985 Burke JF: The effective period of preventive antibiotic action in experimental incisions and dermal lesions. Surgery 1:161, 1961 Conte JE, Cohen SN, Roe BB, Elashoff RM: Antibiotic prophylaxis and cardiac surgery. Ann Intern Med 76:943, 1972 Slaughter L, Morris JE, Starr A A prosthetic valvular endocarditis. Circulation 47:1319, 1973 Quintilliani R, Klimek JJ, Nightingale C: The penetration characteristics of cephapirin and cephalothin into the right atrial appendage and pericardial fluid in patients undergoing open-heart surgery. J Infect Dis 139:348, 1979 Griffith RS, Black HR, Brier GL, Walny JD: Cefamandole in vitro and clinical pharmacokinetics. Antimicrob Agents Chemother 1:814, 1976 Bain WH, Nasr A, Towler C: Use of cefuroxime sodium as a prophylactic antibiotic in patients undergoing coronary artery bypass surgery. In Wood C, Rue Y (eds): Cefuroxime Update. Royal Society of Medicine International Congress and Symposium Series No. 38. London, Academic, 1981 Gentry LO, Zeluff BJ: Cefamandole prophylaxis for cardiovascular surgery: a dosage comparison. Tex Heart Inst J 14: 374, 1987 Slama TG, Sklar SJ, Misinski J, Fess SW: Randomized comparison of cefamandole, cefazolin and cefuroxime prophylaxis in open-heart surgery. Antimicrob Agents Chemother 29:744, 1986 Peterson CD, Lake KD, Arom KV, Love KR. Antibiotic prophylaxis in open-heart surgery patients: comparison of cefamandole and cefuroxime. Drug Intelligence Clin Pharm 21:728, 1987