Ann Thorac Surg 1995;59:626-631
© 1995 The Society of Thoracic Surgeons
Bacterial Wound Colonization After Broad-Spectrum Versus Narrow-Spectrum Antibiotics
Darwin L. Palmer, MD,
Stuart B. Pett, MD,
Bechara F. Akl, MD
New Mexico Regional Federal Medical Center and University of New Mexico School of Medicine, Albuquerque, New Mexico
Accepted for publication November 10, 1994.
 |
Abstract
|
|---|
Broad-spectrum versus narrow-spectrum antibiotic prophylaxis for patients who undergo cardiac operations is variously advocated to reduce the incidence of all infections or, conversely, to prevent resistant superinfections. Previous studies of prophylaxis have shown a reduction in the incidence of staphylococcal infections with some increased resistance. We studied preoperative and postoperative wound colonization as a surrogate for infection. Among 78 patients undergoing cardiac procedures, the type of prophylaxis was allocated as follows: narrow-spectrum (nafcillin), 24 patients; midspectrum (cephapirin), 26 patients; and broad-spectrum (ceftriaxone), 28 patients. Seventeen patients who underwent other procedures received no antibiotics and served as controls. Cultures of the operative site were done preoperatively, and 3 and 6 days postoperatively. The incidence of preoperative skin colonization with staphylococci was identical (95%) in all groups. Postoperatively, more patients receiving nafcillin (48%) were culture-negative for all organisms than were either of the other groups receiving antibiotics (27% and 22%) (p < 0.05). Gram-negative bacilli were infrequent colonizers and neither did the incidence of infection with these organisms increase nor did resistance develop in any group. The infection rates were not different among the treatment groups. Thus, a narrow-spectrum antistaphylococcal penicillin may offer an advantage in terms of both prophylaxis for cardiac operations and hospital costs.
|
Introduction
|
|---|
See also page 631.
Prophylactic antibiotics are routinely given perioperatively in patients undergoing cardiac procedures [1], although few clinical trials have demonstrated their efficacy [2–4]. The authors of a critical review concluded that the value of antibiotic prophylaxis was not proved [5]. Moreover, antibiotic prophylaxis has resulted in increased bacterial colonization and the emergence of resistant organisms. Sallam and associates [6] showed that the use of prophylactic antibiotics resulted in a greater incidence of all infections in patients undergoing closed heart operations, as well as an increased rate of gram-negative bacillary infections. Both gram-negative bacilli [7, 8] and coagulase-negative staphylococci [2–11] have demonstrated increased resistance after antibiotic prophylaxis. Palmer and colleagues [12] have shown an increased incidence of Enterobacter mediastinal infections among postoperative cardiac patients when prolonged antibiotics have been given. Although staphylococci (both coagulase-positive and coagulase-negative) continue to be the major pathogens in cardiac surgical patients [1], an increasing number of postoperative infections appear to be caused by gram-negative bacilli [1, 8, 12, 13]. Increasingly, broad-spectrum cephalosporins (as opposed to antistaphylococcal penicillins) are used as prophylactic agents in such patients [1, 5, 13, 14]. Unfortunately, the results of clinical trials comparing narrow-spectrum and broader-spectrum agents have not resolved the problem. Certain studies have shown fewer postoperative infections resulting from the use of cephalosporins than from the use of antistaphylococcal penicillins [14, 15]; others have shown more Staphylococcus aureus infections associated with the use of first-generation than with second-generation cephalosporins [16]. Two studies have shown no differences in the infection rates associated with the use of narrow-spectrum versus broad-spectrum agents [17, 18], but more gram-negative infections associated with the use of broad-spectrum antibiotics [13]. Variable choice of antibiotics with widely differing pharmacokinetics makes these study findings difficult to compare.
At issue is whether broad-spectrum antibiotics, either by suppressing some bacterial flora and allowing superinfection or the overgrowth of more resistant organisms [19, 20], or by directly inducing resistance among Enterobacteriaceae [21, 22], increase the risk of postoperative infection. Superinfection with nosocomial gram-negative bacilli appears to be particularly likely when prolonged-use, high-dosage or broad-spectrum antibiotics are prescribed [7–9, 12, 13, 19, 21].
Our study was designed to investigate the hypothesis that broad-spectrum and narrow-spectrum antibiotics might result in different rates of skin colonization at sternal wound sites in patients undergoing cardiovascular operations. We chose to sample bacterial colonization at the skin wound site as a surrogate marker for potential infections, as colonization is a necessary preliminary step in infection [19] and changes in rates may more rapidly be detected in this way. To evaluate the effect of narrow-spectrum versus broad-spectrum antibiotics on high-risk cardiac surgical patients, the rates, quantity, and bacterial types of skin-colonizing organisms were determined before and after operation in patients who were randomly allocated to receive nafcillin (narrow-spectrum), cephapirin (midspectrum), or ceftriaxone (broad-spectrum) as antibiotic prophylaxis. Non–cardiac surgical patients not given antibiotic prophylaxis were also studied concurrently to control for the effects of the presence of a surgical wound alone on wound colonization.
|
Material and Methods
|
|---|
All patients undergoing elective cardiac operations (both valve procedures and coronary artery bypass grafting) at the Albuquerque Veterans Affairs Medical Center between January 1985 and September 1988 were eligible for enrollment. Written informed consent was obtained from all patients before random selection of the antibiotic regimen. Excluded from the study were patients who had received antibiotics within the previous two weeks, who were allergic to either cephalosporins or penicillins, who had a serum creatinine level greater than 2.0 mg/dL, or who had recently undergone another operation (in the past month). The control surgical patients, who did not receive antibiotics, were those undergoing knee operations or inguinal hernia repair.
The allocation of cardiac surgical patients to receive nafcillin (2 g intravenously every 6 hours), cephapirin (2 g intravenously every 4 hours), or ceftriaxone (1 g intravenously every 12 hours) was done by random assignment, but was not investigator blinded. The antibiotic therapy was started immediately preoperatively and continued for 72 hours. The serum levels of the antibiotics in the first 10 or 12 patients in each treatment group were measured by bioassay at four interoperative points to ensure that appropriate concentrations had been reached. These points were before the start of cardiopulmonary bypass, after the start of bypass, 4 hours into the procedure, and on arrival in the surgical intensive care unit. Bioassay was done essentially as described elsewhere [23] using S aureus as a test organism and measuring inhibition by the pour plate method. All sera were tested in duplicate, and all determinations were accompanied by a control determination using known concentrations of identical antibiotics in the broth.
Cultures were obtained by swab from the posterior pharynx, skin operative site (sternum), and leg (donor site) 24 hours or less preoperatively, and on postoperative days 3 and 7 (4 days after discontinuing antibiotic prophylaxis). Colony counts of all major species were done for both skin sites. Species were identified using routine clinical laboratory technology. Additional information gathered on each patient consisted of the duration of the preoperative hospital stay; preexisting illness; the duration and type of operation; the number of drains, catheters, and lines and duration of use; operative complications; and postoperative infections.
Patients were dropped from evaluation if their antibiotic regimen was mismanaged (eg, wrong dose or timing); if other antibiotics were given during the 72-hour prophylactic interval (as in the event of a suspected or proven postoperative infection); or if adverse reactions occurred necessitating a switch in the antibiotic agent.
Statistical analysis was done by Stat-pack, using chi-square analysis for the comparison of multiple categories and Fisher's exact test (two-tailed) for two-by-two tables.
|
Results
|
|---|
A total of 91 patients were enrolled in the study after giving written, informed consent; 5 patients refused to be enrolled. Of the enrollees, 8 were not evaluable, leaving 78 study patients. Drop-outs were evenly distributed among the antibiotic groups and did not differ by age or underlying disease. More drop-outs had a valve procedure or combined coronary artery bypass grafting and valve procedure (mostly due to the addition of antibiotics) than grafting alone. As shown in Table 1
, 24 patients were randomly allocated to receive nafcillin, 26 to receive cephapirin, and 28 to receive ceftriaxone; 17 patients served as the surgical controls. All patients were male, the average age was 58.6 years, and the preoperative risk factors such as the duration of hospital stay and the underlying illness were identical. As shown in Table 2, the operative risks among the antibiotic prophylaxis groups were identical (data for control group not shown). Thus, the type of procedure, duration of operation, and duration of urinary catheter drainage, chest tube drainage, postoperative ventilatory support, and central venous line support were the same. Sixty-seven patients underwent coronary artery bypass grafting; 8, a valve operation; and 3, both. The mean duration of operation was 4.9 hours. As shown in Figure 1, the antibiotic levels, as measured in the first 10 or 12 patients in each treatment group, exceeded the 90% minimum inhibitory concentration for all staphylococci and common Enterobacteriaceae throughout the operation and immediately postoperatively.
The colony count quantification of skin cultures proved unreliable because of the extreme range encountered; no trends were apparent. No differences were found either preoperatively or postoperatively in the throat culture findings (not shown); pathogenic streptococci were found only twice and Staphylococcus (either epidermidis or aureus) was found only once. The leg vein graft donor site culture findings largely paralleled those of the sternal cultures; both sites showed a preponderance of S epidermidis and both sites showed a significant decrease after 3 days of antibiotic therapy (not shown). Gram-negative bacilli and other organisms were seen in less than 25% of all patients and the incidence did not vary significantly among the antibiotic groups.
The results of sternal cultures are shown in Table 3. Preoperatively, S epidermidis was detected on 97% of the cultures; S aureus was found in only 5 patients (5%), with 1 or 2 in each treatment group. Because S aureus was so infrequently found, the number of cases were added to those of S epidermidis and reported as the total rate of staphylococcal colonization. In all treatment groups, the preoperative detection of gram-negative rods was rare (only in 1 patient, a control); this increased to 4 patients (3 in the nafcillin group and 1 in the control group) during treatment and decreased again to 1 (in the cephapirin group) after antibiotic therapy was stopped. Neisseria, diphtheroid, and Bacillus species were skin ``contaminant'' flora found equally in all groups, with the incidence declining from 21% to 9%, then rising back to 13%, in all groups in the pretherapy, therapy, and posttherapy periods. Sternal cultures demonstrated significant differences between the antibiotic prophylaxis groups during the postoperative period (not shown). Of all the groups, those patients receiving nafcillin showed the greatest decrease in the total staphylococci rate of colonization of from 100% to 52%. The decrease in the cephapirin group was from 100% to 73%; that in the ceftriaxone group was from 100% to 71%; and that in the control group was from 94% to 72%. The difference between the nafcillin group and the other therapy groups in terms of this reduction approached statistical significance (p = 0.08) by Fisher's exact test. The number of patients with gram-negative rods increased to 3 (1 case of Enterobacter and 2 of Escherichia coli) in the nafcillin group, versus 1 patient each in the control and cephapirin groups, an insignificant change. In the nafcillin group, 10 patients (42%) became culture-negative (for staphylococci or gram-negative bacilli) while on therapy, a difference that was significantly lower than that in the other groups (p < 0.05). The rates of colonization were not statistically associated with the duration of operation.
Postoperative infections occurred in a total of 16 of the 78 patients (21%) (Table 4); in terms of the total number of infected sites, there were 23 infections in the 78 patients (29%). The latter included both operative wound and non–operative wound infections. These were highest in the ceftriaxone group (25% of the patients) and lowest for the cephapirin group (15% of the patients), but the difference between the two groups was not statistically significant. The sternal site was infected in 4 patients (3 receiving ceftriaxone and 1, nafcillin); three of these infections were superficial and one was deep. Leg infections were seen in 7 patients (4 receiving nafcillin; 2, ceftriaxone; and 1, cephapirin). The remaining infections were at distant sites: respiratory, 5; urinary tract, 3; blood (bacteremia), 2; and other infection and other wound in 1 each. The majority of the infections were staphylococcal (n = 8); gram-negative rods (n = 6) were the next most frequent infecting organism, and the cause of infection was unknown (no or negative culture) in 7 cases. Thirteen of the 23 infections occurred during the acute postoperative period and 10 occurred (or were detected) after discharge. All four of the sternal wound infections were late occurrences; one was due to S epidermidis and the others to unknown organisms (culture-negative). Although not related to the antibiotic choice, the surgical time was more likely to be prolonged in the 16 patients with infections (mean, 5.4 hours) than it was in the 62 patients without infection (mean, 4.8 hours; p < 0.04 by t test). The increased surgical time was associated with increased surgical complexity; those undergoing coronary artery bypass grafting in conjunction with a valve procedure as opposed to coronary artery bypass grafting alone.
|
Comment
|
|---|
Antibiotic prophylaxis during cardiac operations appears warranted because of the devastating nature of postoperative infection, despite the paucity of carefully conducted random allocation studies clearly indicating their value [1–3, 7]. The selection of antibiotic has been perplexing, although most surgical centers now use a cephalosporin. With this as with other surgical sites, antibiotic use has been associated both with increased bacterial colonization [8, 9, 11, 13] and with infection with resistant organisms [9–11]. Antibiotic prophylaxis may result in a greater incidence of all infections, an increased rate of gram-negative bacillary infections [7, 8, 12], and resistance to the perioperative antibiotic used [21, 22]. This problem has become so acute that some centers have had to shift from ß-lactam antibiotic prophylaxis to the use of vancomycin because of the increase in methicillin-resistant S epidermidis and S aureus [10, 11].
The choice of broad-spectrum versus narrow-spectrum antibiotics for perioperative prophylaxis is controversial. A narrow-spectrum penicillinase-resistant penicillin (directed solely at staphylococci) may reduce the incidence of colonization with gram-negative bacilli by least disturbing the endogenous flora [19–21].
Proponents of broad-spectrum antibiotic (eg, third-generation cephalosporins) prophylaxis claim a lesser incidence of gram-negative infections [1, 7, 16], particularly those causing non–wound infections. Both classes of antibiotics share the risk of increasing the carriage of methicillin-resistant staphylococci [10, 11]. Superinfection with gram-positive or gram-negative organisms appears related to prolonged antibiotic use, high dosage, or the use of a broad-spectrum agent [9, 12, 20, 21, 24].
The main thrust of this study was to investigate the differences in the induction of gram-negative bacterial wound colonization through the use of a narrow-spectrum (nafcillin), midspectrum (first-generation cephalosporin), or broad-spectrum (third-generation cephalosporin) antibiotic prophylaxis. Because colonization with pathogenic organisms seems to be a necessary precondition to subsequent infection of a wound [19, 20], this rather than infection was measured. Patients in all three groups were similar in terms of age, the severity of their underlying disease, the types of operations performed, drop-outs from the groups due to inappropriate usage, and basic demographics. Study of the initial 10 or 12 patients in each group indicated that the antibiotic levels reached in all were adequate to attain effective antimicrobic tissue levels.
All groups showed a significant decline in staphylococcal colonization between the preoperative determination and postoperative day 3. These organisms were clearly the predominant bacteriologic risk, as S epidermidis was detected preoperatively in 97% of all patients and S aureus was found in 5%. Of all groups, the patients receiving nafcillin showed the greatest decrease in the total rate of colonization with staphylococci of from 100% to 52%. The decline in patients receiving cephapirin (27% decrease) was identical to that in the those receiving ceftriaxone (29%), and the difference between the nafcillin and the other therapy groups approached, but did not reach, statistical significance by Fisher's exact test (p = 0.08). Gram-negative bacilli did not increase significantly in terms of number of cases or antibiotic resistance in any of the groups, although the sample size was small. One Enterobacter and two E coli colonizations occurred in the nafcillin group, and one gram-negative isolate was found in the control and the cephapirin group each. Ten patients (42%) in the nafcillin group became culture-negative while on therapy. This difference was significantly lower than that of other groups (p < 0.05). The rates of colonization did not change with the duration of operation. The non–cardiac surgical control group, which received no antibiotics, showed an increase in the postoperative colonization rate (a finding that was expected because of the incisional operations performed in these patients), but no infections. Finally, during the course of this study, no specific nosocomial pathogen causing excess or problem wound infections was noted.
In summary, this study was designed to use changes in the colonization of the operative wound site after antibiotic prophylaxis as a surrogate marker for potential pathogenic bacteria capable of causing infection. In the treatment groups, the patients receiving narrow-spectrum antistaphylococcal penicillin therapy showed a significantly greater reduction in staphylococcal carriage at the operative wound site in the posttherapeutic, postoperative period than did those receiving midspectrum or broad-spectrum agents. No change in gram-negative colonization or antibiotic sensitivity was seen among the treatment groups. We therefore suggest the use of such narrow-spectrum antibiotic prophylaxis unless a preponderance of preexisting resistant nosocomial bacteria mandates other therapy. This may result in an equivalent or better reduction in wound staphylococci, less chance of the development of antibiotic resistance, and lower patient costs.
|
Footnotes
|
|---|
Address reprint requests to Dr Palmer, Infectious Disease Division, 2100 Ridgecrest Dr SE, Albuquerque, NM 87108.
|
References
|
|---|
- Kaiser AB. Antimicrobial prophylaxis in surgery. N Engl J Med 1986;315:1129–38.[Medline]
- Fong IW, Baker CB, McKee DC. The value of prophylactic antibiotics in aorta-coronary bypass operations: a double-blind randomized trial. J Thorac Cardiovasc Surg 1979;78:908–13.[Abstract]
- Penketh ARL, Wansbrough-Jones MH, Wright E, et al. Antibiotic prophylaxis for coronary artery bypass graft surgery [Letter]. Lancet 1985;1:1500.
- Kaiser AB, Clayson KR, Mulhern JL Jr, et al. Antibiotic prophylaxis in vascular surgery. Ann Surg 1978;188:283–9.[Medline]
- Hirschmann JV, Inui TS. Antimicrobial prophylaxis: a critique of recent trials. Rev Infect Dis 1980;2:1–23.[Medline]
- Sallam IA, Sammon A, McGeachie J, Bain WH. Prophylactic antibiotics in closed heart surgery. Chest 1971;60:252–5.[Abstract/Free Full Text]
- Bor DH, Rose RM, Modilin JF, et al. Mediastinitis after cardiovascular surgery. Rev Infect Dis 1983;5:885–97.[Medline]
- Freeman R. Short-term adverse effects of antibiotic prophylaxis for open heart surgery. Thorax 1980;35:941–4.[Abstract/Free Full Text]
- Hansen BG. Per-operative antibiotic treatment in cardiovascular surgery: the influence of methicillin vs cephalothin on post-operative infections and bacterial colonization. J Hosp Infect 1986;8:184–92.[Medline]
- Archer GL, Armstrong BC. Alteration of staphylococcal flora in cardiac surgery patients receiving antibiotic prophylaxis. J Infect Dis 1983;147:642–9.[Medline]
- Kernodle DS, Barg NL, Kaiser AB. Low-level colonization of hospitalized patients with methicillin-resistant coagulase-negative staphylococci and emergence of the organisms during surgical antimicrobial prophylaxis. Antimicrob Agents Chemother 1988;32:202–8.[Abstract/Free Full Text]
- Palmer DL, Kuritsky JN, Lapham SC, et al. Enterobacter mediastinitis following cardiac surgery. Infect Control 1985;6:115–9.[Medline]
- Wilson APR, Treasure T, Gruneberg RN, et al. Antibiotic prophylaxis in cardiac surgery: a prospective comparison of two dosage regimens of teicoplanin with a combination of flucloxacillin and tobramycin. J Antimicrob Chemother 1988;21:213–24.[Abstract/Free Full Text]
- Myerowitz PD, Caswell K, Lindsay WG, Nicoloff DM. Antibiotic prophylaxis for open-heart surgery. J Thorac Cardiovasc Surg 1977;73:625–9.[Abstract]
- Ghoneim ATM, Tandon AP, Ionescu MI. Comparative study of cefamandole versus ampicillin plus cloxacillin: prophylactic antibiotics in cardiac surgery. Ann Thorac Surg 1982;33:152–8.[Abstract]
- Kaiser AB, Petracek MR, Lea JW IV, et al. Efficacy of cefazolin, cefamandole, and gentamicin as prophylactic agents in cardiac surgery: results of a prospective, randomized double-blind trial in 1030 patients. Ann Surg 1987;206:791–7.[Medline]
- Slama TG, Sklar SJ, Misinski J, Fess SW. Randomized comparison of cefamandole, cefazolin and cefuroxime prophylaxis in open heart surgery. Antimicrob Agents Chemother 1986;29:744–7.[Abstract/Free Full Text]
- Brook I. Microbiology of post-thoracotomy sternal wound infection. J Clin Microbiol 1989;27:806–7.[Abstract/Free Full Text]
- Johanson WG, Pierce AK, Sanford JP, Thomas GD. Nosocomial respiratory infections with gram-negative bacilli. The significance of colonization of the respiratory tract. Ann Intern Med 1972;77:701.
- Weinstein L. Superinfection: a complication of antimicrobial therapy and prophylaxis. Am J Surg 1964;107:704–9.
- Saunders CC, Saunders WE Jr. Microbial resistance to newer generation beta-lactam antibiotics: clinical and laboratory implications. J Infect Dis 1985;151:399–406.[Medline]
- Chow JW, Fine MJ, Shlaes DM, et al. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med 1991;115:585–90.
- Reeves DS, Wise R. Antibiotic assays in clinical microbiology. In: Reeves DS, Phillips I, Wise R, Williams JD, eds. Laboratory methods in antimicrobial chemotherapy. Edinburgh: Churchill Livingston 1978:137–43.
- Rosendorf LL, Daicoff G, Baer H. Sources of gram-negative infection after open heart surgery. J Thorac Cardiovasc Surg 1974;67:195-201.[Medline]
Related Article
-
Noninvasive Imaging of Congenital Heart Disease: Before and After Surgical Reconstruction
Ann. Thorac. Surg. 1995 59: 631.
[Extract]
[Full Text]
Top
Abstract
Introduction
