Antibiotic prophylaxis in acute cholecystectomy revisited: results of a double-blind randomised controlled trial

The standard treatment for mild to moderate acute cholecystitis is early laparoscopic cholecystectomy [1‐4]. Given that the overall complication rate after cholecystectomy is approximately 11% (Swedish GallRiks register) [5], the risk for postoperative infectious complications (PIC) after gallstone surgery is small but not negligible. The risk appears to be enhanced by the severity of the ongoing cholecystitis, and PIC has been shown to occur in 17% or more of patients with grades I and II acute cholecystitis [6]. The pathogenesis of PIC in connection with elective and acute operations for gallstone disease includes contamination with infected bile. Under normal conditions, bile in the gallbladder and the biliary tree is sterile [7]. Obstruction of the outflow of bile by gallstone(s) initiates an inflammatory process that eventually leads to bacterial colonisation. In acute cholecystitis, the bile becomes colonised in 35–60% of the cases [8], the most common agents being gram-negative organisms (Escherichia coli 31–44%, Klebsiella 9–20%) and Enterococci (3–34%) [9, 10]. The mechanisms behind PIC are complex, and the details of the dynamics of colonisation are lacking. A recent retrospective study in Sweden showed that the only significant risk factor for PIC (approximately 6%) was a positive bile culture. Gender, age, ongoing or previous cholecystitis and preoperative antibiotic prophylaxis (PAP) had no impact on PIC [11]. However, despite the fact that surgery for acute cholecystitis is one of the most frequently performed procedures by general surgeons throughout the world, the use of PAP varies between hospitals and between surgeons, as recently shown in a nationwide study from Sweden [12]. Current guidelines do not support the routine use of PAP for elective cholecystectomy for uncomplicated gallstone disease [13‐15], but surprisingly adherence to international guidelines is generally low in Europe [16, 17]. Lack of evidence-based guidelines could be a significant factor behind the disparities seen worldwide. Liang et al., in a systemic review and meta-analysis of low-risk patients undergoing elective laparoscopic cholecystectomy, concluded that PAP is safe and effective in reducing surgical site infections, global infections and postoperative length of hospital stay [18]. On the other hand, the 2013 Tokyo Guidelines (TG13) recommended PAP for interventions in acute cholangitis and cholecystitis only. Implementation of TG13 has led to uniform antibiotic treatment without increasing the risk for PIC [19]. However, the effectiveness of PAP in acute cholecystitis needs to be confirmed.

The aim of this study was to determine the effect of PAP with piperacillin/ tazobactam on PIC following acute cholecystectomy for mild to moderate cholecystitis (grades I and II according to TG18).

The study was undertaken as a prospective randomised double-blinded parallel group study at Karolinska University Hospital in Huddinge, with treatment allocation 1:1 between PAP and placebo.

Altogether 106 patients were invited to participate in the study. Of these, 16 were excluded because of protocol violation (3 did not fulfil eligibility criteria, 1 withdrew consent, 8 due to postponed operation because of OR overload and 4 because the allocation envelope was missing). The remaining 90 patients were allocated to PAP or placebo (Fig. 1). The first patient was included on December 14, 2009, and the last to be included was followed until April 4, 2017. There were no differences between the groups regarding demographic and disease-specific characteristics (Table 1). Fifteen patients were given antibiotic treatment in the immediate postoperative period due to severe contamination of the abdomen, regardless of allocation. These were classified as PIC even though the decision to give antibiotic treatment was taken perioperatively whether prophylaxis had been given or not. Seventeen patients were lost to follow-up (Fig. 1). Four patients were included based on primary intent to perform laparoscopic cholecystectomy, but the surgeon responsible for the procedure decided to do an open procedure for technical reasons. These patients remained in the study.

The rate of PIC was 24% (9% in the PAP arm and 16% in the placebo arm in the intention-to-treat and 5% in each group in the PP analysis). There was no significant difference in PIC between the groups regardless of the analytical approach (p = 0.193, Table 2).

Bile cultures were obtained in 48/90 (53%) cases (evenly distributed between the study groups), and cultures were positive in18/48 (38%) (13 in the antibiotic group and 5 in the placebo group, p = 0.076). Gram-negative aerobes predominated in the cultures (N = 11), followed by gram-positive aerobes (N = 10), anaerobes (N = 3) and fungus (N = 1). PIC was numerically more common in those with a positive culture, but this did not reach statistical significance (p = 0.054, Table 3). There was no statistically significant association between the period of PAP and a positive bile culture.

CRP levels were significantly higher in patients with PIC, on both the day of randomisation (p = 0.008) and the day of surgery (p = 0.004). PIC was also more frequently seen in cases converted to an open procedure (p = 0.017), whereas patient comorbidity was not associated with PIC (p = 0.075, Table 2).

In the present study, we were unable to detect any benefit of administrating PAP to reduce the risk for PIC after emergency cholecystectomy for grades I and II acute cholecystitis. Although we did not reach sufficient statistical power to detect a minor reduction in PIC rate, the absence of any significant impact speaks against the routine use of PAP. We observed a PIC rate of less than 20%: 9% in the PAP arm and 16% in the placebo arm in the ITT analysis and 5% in each group in the PP analysis. These figures compare well with the expected PIC incidences seen after emergency cholecystectomy for cholecystitis [5]. It can be argued that the preconditions applied in the sample size calculation lack clinical relevance. On the other hand, there are no definited criteria for defining clinically important differences when assessing the effectiveness of PAP. The widespread overuse of antibiotics and increasing resistance figures emphasises the importance of assessing the evidence basis of all indications for use of antibiotics. A recent randomised clinical trial, comparing single-dose antibiotic prophylaxis with single-dose AP plus an extra dose postoperatively, in patients listed for acute cholecystectomy for mild (grade I) calculus cholecystitis, demonstrated no difference in PIC rate [21]. In three other randomised trials, continuation of prophylactic antibiotic after cholecystectomy had no significant impact on PIC rate in grades I and II cholecystitis [6, 22, 23]. The results of the present study further question the use of PAP in procedures for cholecystitis.

The use of PAP in planned cholecystectomy varies between units and regions. There is still a lack of evidence supporting the use of PAP in cases of acute cholecystectomy, but even so it is obviously a generally held view that PAP reduces PIC; otherwise, the use of PAP should not be practised to such a great extent [12].

Guidelines on antibiotic use are important for all clinicians, not only to ensure proper use of antibiotics but also to minimise its overuse. Reported proportions of ESBL-producing E. coli range between 31.2% in two university hospitals in Germany [24] and 70.0% in a Korean university medical centre [25]. In Europe, the rate of fully susceptible E.coli is only 41.7% and 63.4% for Klebsiella [26]. The increasing problem of drug resistance makes evidence-based guidelines even more important. In the present study, we tested piperacillin/ tazobactam as prophylaxis. The use of piperacillin/tazobactam was motivated by the drug susceptibility pattern found in a previous Swedish study [11] and was one of the several drugs suggested in an update of the Tokyo guidelines on antimicrobial therapy for acute cholangitis and cholecystitis [10]. It could be argued, however, that its spectrum is too broad to be used for prophylaxis. Antibiotics with a spectrum more appropriate for prophylaxis, e.g. sulfamethoxazole/trimethoprim, could be a better alternative. Nevertheless, there are few other drugs that can be expected to be more effective as prophylaxis than piperacillin/tazobactam. Concerns other than just effectiveness must be considered when choosing an antibiotic for prophylaxis, and there is no reason to believe that the outcome of the present study would have been different if another antibiotic had been used.

As the timing of the procedure varied, the interval between study inclusions ranged from 0 to 72 h. This may have affected the impact of AP in the PAP group, although the time from the last administration of the infusion to start of surgery never exceeded 8 h since the antibiotic was administrated from the time of inclusion until surgery. There was a significant correlation between CRP at baseline and PIC rate, which may have had an impact on the outcome.

It is obvious that after procedures for acute cholecystitis, the PIC rate is significant. The mechanisms behind the development of PIC are complex including factors such as bile leakage, haemorrhage and tissue damage related to surgical trauma. Intraoperative contamination with infected bile is probably the main risk factor for PIC, which is why we specifically followed contamination and bile cultures [8‐10]. We found a weak association between bacterial counts in the aspirated bile and risk for PIC which was in contrast to the risk associated with conversion to an open procedure. The rate of bile culture sampling was low, which may have affected the results of this study. The higher rate of positive cultures in the placebo group is probably due to a type I error.

We defined PIC as any infectious complication requiring drainage, surgical intervention or antibiotic treatment. This definition corresponds to grade ≥ II complications according to the Clavien-Dindo classification, but does not include surgical site infections not requiring antibiotics. This may have led to an underestimation of the PIC rate.

PAP did not seem to affect the risk for PIC in patients with grades I and II acute cholecystitis. The major factors determining PIC in these patients must be better defined, in particular the role of bacteriobilia. The present study did not have sufficient statistical power to reveal minor reductions in PIC rate. PAP may also be motivated in subgroups with high risk for PIC. Further studies are needed to evaluate the effectiveness of PAP in patients undergoing surgery for acute cholecystitis.