Objectives: Infections are major causes of morbidity and mortality in the early postoperative period after liver transplant. We observed a high rate of enterococcal infections at our center. Therefore, we added an intraoperative single shot of vancomycin to the standard regimen of meropenem given over 5 days. The aim of this study was to determine the prevalence of both Enterococcus faecium and Enterococcus faecalis infections during the first 28 days after surgery depending on the type of antibiotic prophylaxis and their implications on mortality and morbidity.
Materials and Methods: Our retrospective cohort analysis included 179 patients: 93 patients received meropenem only and 86 patients were treated with meropenem plus vancomycin.
Results: During the first 28 days after transplant, microbiological tests showed that 51 patients (28.5%) were positive for Enterococcus faecium and 25 patients (14.0%) were positive for Enterococcus faecalis. Enterococcus faecium infections appeared significantly more often in patients without vancomycin (P = .013). In the second week after transplant, there was a significant reduction in Enterococcus faecium infections in the meropenem plus vancomycin group (P = .015). Enterococcus faecalis infections occurred more often in the patients receiving meropenem alone, but results were not statistically significant (P = .194). There was a trend toward more frequent renal replacement therapy in the meropenem plus vancomycin group. We found no differences between the groups regarding survival after 1 and 2 years, length of hospital stay, or duration in the intensive care unit. Overall 1-year survival was 78.8% (141/179 patients).
Conclusions: Although postoperative Enterococcus species infections can be reduced after liver transplant by adding vancomycin to the intraoperative antibiotic regimen, it does not improve the long-term outcomes.
Key words : Antibiotic prophylaxis, Enterococcal infections, Meropenem
Postoperative infections after liver transplant are major causes of morbidity and mortality.1,2 There are many different regimens of antibiotic prophylaxis that can be used during the perioperative period; however, evidence is scare regarding the optimal strategy.3 Our standard perioperative antibiotic regimen is meropenem, but high rates of infection with Enterococcus faecalis and Enterococcus faecium were observed in the early postoperative period.2,4,5 Enterococci are emerging nosocomial pathogens, which can lead to endocarditis and bloodstream and surgical site infections (SSI).6,7 Colonization of the biliary tract with Enterococcus species is common after liver transplant8,9 and in patients with chronic liver disease.10 There are contradictory data concerning the virulence of Enterococcus faecalis compared with Enterococcus faecium.11,12 However, Enterococcus faecalis is much more common in isolates that tested positive for Enterococcus species (60%-95% vs 5%-40%).12
Susceptibility of Enterococcus faecalis against meropenem is low. The MYSTIC study revealed a European resistance rate of 53.3% and up to 68.3% in the United States.13 In regard to our local resistance data, Enterococcus faecalis has shown resistance against ciprofloxacin in 43.7% and against erythromycin in 44.6% of isolates.14 The statistical records of resistance data for Enterococcus faecium at the University Hospital of Munich have shown local resistance rates of more than 90% for ampicillin, amoxicillin, meropenem, or erythromycin from 2004 to 2010. Resistance against ciprofloxacin has been greater than 87%, whereas resistance against vancomycin has been less than 9%.
As a consequence, an intraoperative single-shot application of vancomycin was added to our standard therapy with meropenem. The aim of this study was to determine the effects of this intraoperative single shot of vancomycin on postoperative Enterococcus faecalis and Enterococcus faecium infections.
Materials and Methods
This study is a retrospective cohort analysis. Individual informed consent was waived by the local ethics committee. Patients who received transplants between February 2004 and February 2011 were included. Until December 2007, the perioperative antibiotic regimen consisted of 1 g meropenem intravenously 30 minutes before skin incision (repeated after 8 h or blood loss > 5000 mL, with subsequent doses thereafter adapted to renal function) and was continued until the fifth postoperative day. In all patients, selective oral decontamination was done with polymyxin B, gentamicin, vancomycin, and amphotericin B postoperatively until extubation. After December 2007, an intraoperative single-shot dose of intravenous vancomycin (1 g intravenously 30 min before skin incision, no repetition) was added to the existing regimen. Immunosuppression was achieved with cyclosporine or tacrolimus and prednisolone, as reported elsewhere.15
For this study, 200 consecutive patients were identified, with half of the patients treated before and half after introduction of vancomycin to the standard protocol. Clinical data, laboratory values, and history of infection were extracted from comprehensive clinical documentation. Laboratory Model for End-Stage Liver Disease (MELD) scores were calculated retrospectively for patients transplanted before the introduction of a MELD-based allocation mode in the Eurotransplant zone in 2006.
Screening smear tests from all surgical drains, intraabdominal smears during transplant or consecutive operations (intraabdominal infections), surgical wounds (SSIs), urine (urinary tract infections), and sputum or secretion gained from endotracheal suction (respiratory tract infections) were performed at least once per week. In addition, a smear test was performed if an infection was suspected. Cultures were routinely taken from all removed central venous lines (catheter-related infections). Blood cultures were routinely taken during the change of central venous lines or arterial catheters and additionally at the discretion of the attending physician (bloodstream infections).
A patient was considered positive for an infection if Enterococcus species were detected in at least one of the microbiological cultures from the listed primary sterile regions (in urinary tract, > 105 colony-forming units/mL) in combination with any increase of serum levels of interleukin 6. The first postoperative positive microbiological culture for Enterococcus species was identified for each patient. Patients who repeatedly had positive smears for Enterococcus species were counted once only.
According to the existing literature, the following risk factors for early postoperative infections were included in the analysis: preoperative MELD score > 20, intraoperative transfusion of > 6 units of red blood cells (RBC), intraoperative transfusion of > 12 units of fresh frozen plasma (FFP), and type of biliary reconstruction.
Basic demographic and clinical data were collected from the original electronic data records of each patient. Statistical analyses were performed with Sigma Stat 12.0 (Systat Software, Chicago, IL, USA). Normal distribution was checked using the Lilliefors test. Clinical and microbiological data were assessed using either the t test or Mann-Whitney rank sum test as well as the Pearson chi-square test for evaluation of statistical significance. P < .05 was considered statistically significant. Subgroup analysis was performed using Fisher exact test and a logistic regression model.
Medical records from 200 patients were screened (Figure 1): 100 patients before (meropenem group) and 100 after (meropenem + vancomycin group) implementation of a single-shot vancomycin application to the standard protocol with meropenem. Eleven patients were excluded due to inconsistent data.
Of the remaining 189 patients, 10 patients in the meropenem-vancomycin group were excluded because of ongoing treatment with either linezolid (n = 1), tigecycline (n = 1), or vancomycin (n = 8) on the day of transplant. None of the patients in the meropenem group received an antibiotic agent with activity against Enterococcus species before surgery.
Pretransplant patient characteristics are shown in Table 1. In brief, there were no significant differences between the groups with regard to pretransplant status. Of total patients, 67.6% were male (n = 121). The predominant reasons for liver transplant were hepatitis C cirrhosis (32%), alcoholic cirrhosis (27%), cryptogenic cirrhosis and cirrhosis of other causes (10%), hepatitis B cirrhosis (9%), primary sclerosing cholangitis (8%), and acute hepatic failure (8%).
During the first 28 days after transplant, microbiological tests showed that 51 patients (28.5%) were positive for Enterococcus faecium and 25 patients (14.0%) were positive for Enterococcus faecalis (Table 2). Altogether, 131 microbiological tests were positive for Enterococcus species: 96 probes (73.3%) detected Enterococcus faecium and 35 (26.7%) were positive for Enterococcus faecalis. Table 3 and Table 4 show the exact distribution of positive microbiological cultures.
Enterococcus faecium infections appeared significantly more often in patients receiving meropenem only (P = .013). In the second week after transplant, there was a significant reduction in Enterococcus faecium infections in the meropenem plus vancomycin group (P = .015), as shown in Table 2.
Enterococcus faecalis infections occurred more often in the meropenem group (16 vs 9 patients), but results were not statistically significant (P = .194). There was a peak of new diagnosed infections in the second week, with no statistical difference between the 2 groups (P = .068).
There was a trend toward a more frequent renal replacement therapy in the meropenem plus vancomycin group, although it was not a significant effect (P = .07; Table 5). The overall 1-year survival rate was 78.8% (141/179) with no differences shown between the 2 groups (P = .651). Moreover, survival after 2 years, length of hospital stay, and duration of treatment in the intensive care unit did not differ significantly (Table 6).
For patients with prolonged postoperative treatment, defined as a hospital stay of longer than 28 days, we performed a subgroup analysis (109 patients fulfilled these criteria). Of these, 57 had received meropenem only and 52 patients had received meropenem plus vancomycin. The potency of vancomycin to prevent an enterococcal infection in the second postoperative week was measured via the area under a receiver operator characteristics curve (AUC). An AUC of 0.5 indicates no predictive effect, whereas 1.0 indicates perfect separation. With a measured AUC of 0.62, there is sufficient predictive effect. The Kaplan-Meier curve for infections with Enterococcus species in this subgroup is shown in Figure 2. There was a tendency toward a longer median posttransplant stay in the intensive care unit in the meropenem plus vancomycin group (13 days [range, 3-124 days] vs 7 days [range, 2-110 days]), although it was not statistically significant (P = .089). The 1-year and 2-year survival rates were comparable (P = .96 and P = .16) in the meropenem group (74.6% and 67.3%) and the meropenem plus vancomycin group (75.0% and 53.9%).
In univariate analysis, only the biliary reconstruction with a biliodigestive anastomosis was a predictor for a higher rate of Enterococcus species infections in the study population (P = .012). There was no difference regarding high or low MELD score values (P = 0.135). A transfusion of more than 6 units of RBC or more than 12 units of FFP had no significant effect (P = .349 and P = .581). The effect of these variables on the risk of an infection with Enterococcus species were similar in a bivariate analysis with the addition of vancomycin to the standard protocol (meropenem plus vancomycin), with a significant effect shown for the variables MELD score > 20 (P = .073) and biliodigestive anastomosis (P = .017). Again, there was no effect of transfusion of > 6 units of RBCs (P = .140) or > 12 units of FFP (P = .506).
This study is the first to provide data on the influence of an intraoperative single-shot application of vancomycin on postoperative enterococcal infections after liver transplant. We found a significant reduction of infections with Enterococcus faecium in the early postoperative course if vancomycin was added to the intraoperative antibiotic regimen. Nevertheless, there was no influence on the long-term results.
Postoperative bacterial infections after liver transplant are a common problem. Posttransplant bacteremia occurs in 33% to 68% of patients,1,2 with 8.8% to 18% having SSIs, including peritonitis.16,17 Enterococcus species are the predominant pathogens in liver transplant patient, with a higher proportion of Enterococcus faecium than Enterococcus faecalis, as we found in our population.16,18 In an analysis of bile colonization and intestinal flora of liver transplant recipients, the most common gram-positive isolates were Enterococci and predominantly Enterococcus faecium, which might contribute to the dominance of enterococcal infections.8
The reason for the dominance of infections with Enterococcus faecium after liver transplant might be because proven or suspected spontaneous bacterial peritonitis is often treated with quinolones or third-generation cephalosporins.19 Both groups of antibiotics have limited effects on Enterococcus species.20
In the literature, the median postoperative time to first infection is between 12 and 15 days after transplant, and most infections (77%) occur in the first postoperative month.2,18 This is in line with our findings. In patients treated with meropenem only, and therefore lacking an effective therapy especially against enterococcal infections, we found a peak of enterococcal infections during the second postoperative week. The second postoperative week might be a phase of special vulnerability for infections due to the strong immunosuppression, which reaches its peak in plasma and tissue concentrations.
There is only limited information on the efficacy of diverse perioperative antibiotic regimens to prevent bacterial infections after liver transplant.1 A recent Cochrane analysis failed to show any benefit of antibiotic prophylactic regimens for SSIs after liver transplant.1 Asensio and associates described the influence of different perioperative antibiotic regimens on the frequency of surgical site and intraabdominal infections within the first 30 days after liver transplant.16 The different perioperative antibiotic regimens were analyzed with respect to their influence on postoperative infections. Although no association was found between antibiotic regimen and risk of SSI after risk adjustment for center and Child-Pugh classification, the lowest risk for SSI was seen when glycopeptides (ie, vancomycin) and aztreonam were combined. In the study of Asensio and associates, Enterococcus faecium and feacalis were among the 3 most common gram-positive pathogens, which may explain the low risk of infections if glycopeptides were used.16
Our primary hypothesis was that it is possible to reduce the incidence of early postoperative infections with Enterococcus species by adding a single shot of vancomycin to the perioperative meropenem regimen. However, it remains unclear whether this is beneficial for the patient. No differences in length of hospital stay or in mortality after 1 or 2 years were shown in our study. In contrast to the expected positive effects, we observed a tendency to a higher incidence of renal replacement therapy in patients exposed to vancomycin intraoperatively. Alternative antibiotics could avoid this complication. In principle, daptomycin or tigecycline could be candidates in patients with reduced renal function. However, high-dose daptomycin could cause rhabdomyolysis and thrombocytopenia and tigecycline interacts with coagulation tests, making its perioperative use unfavorable in liver transplant patients.
If necessary at all, vancomycin should be used only in areas with a low prevalence of vancomycin-resistant Enterococcus species.
Although we focused on the effects of vancomycin on postoperative infections, we also observed other known risk factors for postoperative infections after liver transplant. In line with the findings of Asensio and associates,16 we found choledochojejunostomy to be associated with a higher risk of postoperative infections versus choledochocholedochostomy.
The limitations of our study are due to its retrospective nature. To give a definitive answer to the question regarding whether patients will benefit from an perioperative antibiotic regimen that includes agents against Enterococcus species in general and especially Enterococcus faecium, larger prospective studies are needed because the results presented here may only generate hypotheses and do not provide definite evidence.
Although a reduction of postoperative infections with Enterococcus species after liver transplant can be achieved by adding vancomycin to the intraoperative antibiotic regimen, it does not improve long-term outcomes.
Volume : 16
Issue : 6
Pages : 701 - 707
DOI : 10.6002/ect.2017.0153
From the 1Department of Anaesthesiology, Ludwig-Maximilians University, and the
2Department of General, Visceral and Transplantation Surgery, Ludwig-Maximilians
University, Munich, Germany
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Christian Siebers, Department of Anaesthesiology, Ludwig-Maximilians-University (LMU), Marchioninistraße 15, 81377 Munich, Germany
Phone: +49 89 4400711169
Table 1. Patient Characteristics
Table 2. Time Course of Infections with Enterococcus Species
Table 3. Site of Infection With Enterococcus faecium After Liver Transplant
Table 4. Site of Infection With Enterococcus faecalis After Liver Transplant
Table 5. Morbidity After Liver Transplant
Table 6. Mortality After Liver Transplant
Figure 1. Patients Included in the Study
Figure 2. Kaplan-Meier Curve for Infections With Enterococcus faecium and Enterococcus faecalis in Patients With Prolonged Hospital Stay (> 28 d) After Liver Transplant