Neurosurg Rev DOI 10.1007/s10143-014-0524-z
REVIEW
Antibiotic prophylaxis in craniotomy: a review Weiming Liu & Ming Ni & Yuewei Zhang & Rob J. M. Groen
Received: 24 November 2012 / Revised: 24 August 2013 / Accepted: 27 October 2013 # Springer-Verlag Berlin Heidelberg 2014
Abstract The effectiveness of antibiotic prophylaxis (AP) in craniotomies has been clarified through the accumulation of evidence and increased antibiotic knowledge. This paper focuses on the use of AP in craniotomies during different historical periods and collects highly relevant evidence on this issue. This review surveys different AP guidelines and explains why cefazolin was selected by most guidelines. Recent prominent topics, including strategies to update and implement guidelines and antibiotic efficacy in postoperative meningitis and surveillance and decolonization therapies for methicillin-resistant Staphylococcus aureus, are discussed. Keywords Antibiotic prophylaxis . Craniotomy . Surgical site infection . Guidelines . Cefazolin
Although the rate of surgical site infection (SSI) after craniotomy is known to be low, the routine use of antibiotic prophylaxis (AP) was adopted by many neurosurgeons because of the potentially devastating consequences of infectious complications. The use of AP has been an ongoing issue throughout different historical periods. Weiming Liu and Ming Ni contributed equally to this paper. W. Liu (*) : M. Ni Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Tiantan Xili 6, 100050 Beijing, China e-mail: [email protected] W. Liu : R. J. M. Groen Department of Neurosurgery, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, Netherlands Y. Zhang Department of Hospital Infection Control, Beijing Tiantan Hospital, Capital Medical University, Tiantan Xili 6, 100050 Beijing, China
Pre-1980s, exploration and debate Early neurosurgeons were aware of postoperative infections. Cushing said, “Certainly infections cannot be attributed to the intervention of the devil but must be laid at the surgeon’s door” [1]. People went to great lengths to avoid postoperative infections. In the 1940s, after the refinement and production of Alexander Fleming’s penicillin mold extract, AP served as a more advanced form of antisepsis, further reducing postoperative infections. The first report on prophylactic antibiotics in neurosurgery was provided by Hugh Cairns in 1947 [2]. Surgeons placed penicillin powder directly onto patients’ brains during World War II in London. The postoperative infection rate was 0.9 % in the penicillin-treated patients, and this appeared to be lower than the historical controls’ infection rate of 5.4 %. Although this method of prophylactic antibiotic usage was not standardized in today’s view, it was a good attempt. With the increasing diversity of antibiotics, neurosurgeons first chose antibiotics or combined antibiotic regimens with broad antibacterial spectra for prophylaxis. They believed that a “sterile” body could prevent postoperative infections. This idea was prevalent for a long time. The general protocol involved a combination of penicillin and streptomycin. Unfortunately, many tests in other surgical fields did not support this method. In their general surgical cases, Sanchez-Ubeda [3] and Rocha [4] found that AP with penicillin and streptomycin showed no benefit and may have been associated with an increased risk of infection. In neurosurgery, Wright [5] tested the administration of penicillin and streptomycin after craniotomies. The postoperative infection rate was 5.8 % in the AP group and 5.5 % in the no-AP group. There was no evidence that AP decreased the risk of infection. There are two reasons for AP failure during this early period. The first involves the timing of AP use, which was not clear until Burke’s experiment [6]. Burke used guinea pigs
Neurosurg Rev
to mimic surgical incisions with Staphylococcus contamination. He found that the effective period begins the moment the bacteria gain access to the tissue, which occurs within 3 h. He suggested the use of systemic antibiotics before the bacteria had gained access to the tissue. Alexander [7] tested penicillin in staphylococcal surgical wound infections, and his results supported Burke’s suggestion. Since then, it has been gradually accepted that AP should be used before surgery. The second reason for AP failure was an ineffective categorization of wounds. In 1964, the National Academy of Sciences— United States National Research Council classified wounds into four groups [8]: clean, clean–contaminated, contaminated, and dirty. This categorization provides a good distinction between prevention in clean and clean–contaminated wounds and treatment in contaminated and dirty wounds. When these two issues were better understood, the subsequent experiments achieved many positive results. In Mount Sinai Hospital, an intraoperative regimen of lincomycin was used, and the infection rate decreased to 2.3 % [9]. Later, a randomized, double-blind trial was performed to test the use of short-term preoperative lincomycin [10]. This welldesigned trial was terminated unexpectedly because of a sudden epidemic of wound infections. This trial showed an infection rate of 1 % in the lincomycin group and 9 % in the placebo group, proving the efficacy of AP. In 1979, they reported amazing results [11]. In the 1,732 consecutive neurosurgical procedures, there was no evidence of postoperative infections under the routine use of a single intraoperative dose of gentamicin (tobramycin), vancomycin, and topical streptomycin irrigating solution. This regimen was based on a review of organisms cultured from infections throughout hospital over the previous 5 years. When repeating this regimen, Quartey did not obtain zero infections but still had a low infection rate of 0.8 % [12]. Although the efficacy of AP in craniotomies has been verified in some studies, some debate remains ongoing. Certain retrospective studies have shown that the use of prophylactic antibiotics is not valid, regardless of which regimen is used [13–15]. Even without AP, there can be a very low rate of infection in craniotomies [16]. During the first randomized control trial [10], the efficacy of prophylaxis was questioned by Haines [17] because the majority of cases in an infection epidemic fell into the no-AP group. Even in their own hospital, they stated after the infection epidemic that “Antibiotic prophylaxis is not recommended. This may mask an ongoing problem” [11]. Another controversy existed over determining whether a clean wound required AP. Most craniotomies are clean wounds, although a few fall into the clean–contaminated (open paranasal sinuses or mastoid) subgroup. Some experts believe that using AP in clean neurosurgery is not appropriate, and there have been concerns about the side effects of widespread antibiotics use, including antibiotic abuse, bacterial
spectrum changes, and increased resistance [18, 19]. These problems remain a concern today. The value of prophylactic antibiotics in craniotomies could not be resolved with the data available before the 1980s. Most neurosurgeons used AP because they feared the lifethreatening consequences of infection. A variety of AP regimens were chosen, generally based on individual experience. Some consensus was achieved during this period, as follows: antibiotics should be used preoperatively; a short-term regimen or even a single dose can achieve good results; and antibiotics should be chosen based on the spectrum of the local infectious organisms. There was no doubt that rigorous clinical trials were needed to validate the efficacy of AP.
1980s and 1990s: trials and guidelines A few randomized controlled trials were performed in the 1980s and the early 1990s. All but one of these trials showed that AP was more effective than a placebo for SSI. Table 1 shows the details of these trials. Some of these trials had imperfections, such as early termination due to the outbreak of infection [20, 21], no placebo given [22], no blinding [23], and a small sample size [20, 21, 24, 25]. Although these trials had flaws, they all met the basic requirements for prospectively controlled trials and have provided the best evidence to date. It is ethically difficult to repeat similar placebo-controlled studies today. All current AP guidelines are drawn from these trials’ results. Among these trials, only Rocca’s trial indicated that AP was ineffective [26]. In this trial, no infections occurred in the cefamandole prophylaxis group or the placebo group. However, due to the small sample size, the result does not carry strong weight. In 1994, Barker used most of the previously discussed trials to verify AP’s efficacy for craniotomies with a meta-analysis [27]. He found 19 infections in 1,014 craniotomies with AP and 93 infections in 1,061 craniotomies in the control group. The analysis reflected an advantage of antibiotics over the placebo at the P
REVIEW
Antibiotic prophylaxis in craniotomy: a review Weiming Liu & Ming Ni & Yuewei Zhang & Rob J. M. Groen
Received: 24 November 2012 / Revised: 24 August 2013 / Accepted: 27 October 2013 # Springer-Verlag Berlin Heidelberg 2014
Abstract The effectiveness of antibiotic prophylaxis (AP) in craniotomies has been clarified through the accumulation of evidence and increased antibiotic knowledge. This paper focuses on the use of AP in craniotomies during different historical periods and collects highly relevant evidence on this issue. This review surveys different AP guidelines and explains why cefazolin was selected by most guidelines. Recent prominent topics, including strategies to update and implement guidelines and antibiotic efficacy in postoperative meningitis and surveillance and decolonization therapies for methicillin-resistant Staphylococcus aureus, are discussed. Keywords Antibiotic prophylaxis . Craniotomy . Surgical site infection . Guidelines . Cefazolin
Although the rate of surgical site infection (SSI) after craniotomy is known to be low, the routine use of antibiotic prophylaxis (AP) was adopted by many neurosurgeons because of the potentially devastating consequences of infectious complications. The use of AP has been an ongoing issue throughout different historical periods. Weiming Liu and Ming Ni contributed equally to this paper. W. Liu (*) : M. Ni Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Tiantan Xili 6, 100050 Beijing, China e-mail: [email protected] W. Liu : R. J. M. Groen Department of Neurosurgery, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, Netherlands Y. Zhang Department of Hospital Infection Control, Beijing Tiantan Hospital, Capital Medical University, Tiantan Xili 6, 100050 Beijing, China
Pre-1980s, exploration and debate Early neurosurgeons were aware of postoperative infections. Cushing said, “Certainly infections cannot be attributed to the intervention of the devil but must be laid at the surgeon’s door” [1]. People went to great lengths to avoid postoperative infections. In the 1940s, after the refinement and production of Alexander Fleming’s penicillin mold extract, AP served as a more advanced form of antisepsis, further reducing postoperative infections. The first report on prophylactic antibiotics in neurosurgery was provided by Hugh Cairns in 1947 [2]. Surgeons placed penicillin powder directly onto patients’ brains during World War II in London. The postoperative infection rate was 0.9 % in the penicillin-treated patients, and this appeared to be lower than the historical controls’ infection rate of 5.4 %. Although this method of prophylactic antibiotic usage was not standardized in today’s view, it was a good attempt. With the increasing diversity of antibiotics, neurosurgeons first chose antibiotics or combined antibiotic regimens with broad antibacterial spectra for prophylaxis. They believed that a “sterile” body could prevent postoperative infections. This idea was prevalent for a long time. The general protocol involved a combination of penicillin and streptomycin. Unfortunately, many tests in other surgical fields did not support this method. In their general surgical cases, Sanchez-Ubeda [3] and Rocha [4] found that AP with penicillin and streptomycin showed no benefit and may have been associated with an increased risk of infection. In neurosurgery, Wright [5] tested the administration of penicillin and streptomycin after craniotomies. The postoperative infection rate was 5.8 % in the AP group and 5.5 % in the no-AP group. There was no evidence that AP decreased the risk of infection. There are two reasons for AP failure during this early period. The first involves the timing of AP use, which was not clear until Burke’s experiment [6]. Burke used guinea pigs
Neurosurg Rev
to mimic surgical incisions with Staphylococcus contamination. He found that the effective period begins the moment the bacteria gain access to the tissue, which occurs within 3 h. He suggested the use of systemic antibiotics before the bacteria had gained access to the tissue. Alexander [7] tested penicillin in staphylococcal surgical wound infections, and his results supported Burke’s suggestion. Since then, it has been gradually accepted that AP should be used before surgery. The second reason for AP failure was an ineffective categorization of wounds. In 1964, the National Academy of Sciences— United States National Research Council classified wounds into four groups [8]: clean, clean–contaminated, contaminated, and dirty. This categorization provides a good distinction between prevention in clean and clean–contaminated wounds and treatment in contaminated and dirty wounds. When these two issues were better understood, the subsequent experiments achieved many positive results. In Mount Sinai Hospital, an intraoperative regimen of lincomycin was used, and the infection rate decreased to 2.3 % [9]. Later, a randomized, double-blind trial was performed to test the use of short-term preoperative lincomycin [10]. This welldesigned trial was terminated unexpectedly because of a sudden epidemic of wound infections. This trial showed an infection rate of 1 % in the lincomycin group and 9 % in the placebo group, proving the efficacy of AP. In 1979, they reported amazing results [11]. In the 1,732 consecutive neurosurgical procedures, there was no evidence of postoperative infections under the routine use of a single intraoperative dose of gentamicin (tobramycin), vancomycin, and topical streptomycin irrigating solution. This regimen was based on a review of organisms cultured from infections throughout hospital over the previous 5 years. When repeating this regimen, Quartey did not obtain zero infections but still had a low infection rate of 0.8 % [12]. Although the efficacy of AP in craniotomies has been verified in some studies, some debate remains ongoing. Certain retrospective studies have shown that the use of prophylactic antibiotics is not valid, regardless of which regimen is used [13–15]. Even without AP, there can be a very low rate of infection in craniotomies [16]. During the first randomized control trial [10], the efficacy of prophylaxis was questioned by Haines [17] because the majority of cases in an infection epidemic fell into the no-AP group. Even in their own hospital, they stated after the infection epidemic that “Antibiotic prophylaxis is not recommended. This may mask an ongoing problem” [11]. Another controversy existed over determining whether a clean wound required AP. Most craniotomies are clean wounds, although a few fall into the clean–contaminated (open paranasal sinuses or mastoid) subgroup. Some experts believe that using AP in clean neurosurgery is not appropriate, and there have been concerns about the side effects of widespread antibiotics use, including antibiotic abuse, bacterial
spectrum changes, and increased resistance [18, 19]. These problems remain a concern today. The value of prophylactic antibiotics in craniotomies could not be resolved with the data available before the 1980s. Most neurosurgeons used AP because they feared the lifethreatening consequences of infection. A variety of AP regimens were chosen, generally based on individual experience. Some consensus was achieved during this period, as follows: antibiotics should be used preoperatively; a short-term regimen or even a single dose can achieve good results; and antibiotics should be chosen based on the spectrum of the local infectious organisms. There was no doubt that rigorous clinical trials were needed to validate the efficacy of AP.
1980s and 1990s: trials and guidelines A few randomized controlled trials were performed in the 1980s and the early 1990s. All but one of these trials showed that AP was more effective than a placebo for SSI. Table 1 shows the details of these trials. Some of these trials had imperfections, such as early termination due to the outbreak of infection [20, 21], no placebo given [22], no blinding [23], and a small sample size [20, 21, 24, 25]. Although these trials had flaws, they all met the basic requirements for prospectively controlled trials and have provided the best evidence to date. It is ethically difficult to repeat similar placebo-controlled studies today. All current AP guidelines are drawn from these trials’ results. Among these trials, only Rocca’s trial indicated that AP was ineffective [26]. In this trial, no infections occurred in the cefamandole prophylaxis group or the placebo group. However, due to the small sample size, the result does not carry strong weight. In 1994, Barker used most of the previously discussed trials to verify AP’s efficacy for craniotomies with a meta-analysis [27]. He found 19 infections in 1,014 craniotomies with AP and 93 infections in 1,061 craniotomies in the control group. The analysis reflected an advantage of antibiotics over the placebo at the P
