Staphylococcus aureus is the leading cause of bacterial infection in liver transplant recipients. Preoperative nasal carriage of methicillin-resistant S. aureus (MRSA) is associated with a high risk of infection. We conducted a retrospective cohort study in order to identify independent risk factors for early-onset S. aureus infection after liver transplantation. Patients were screened preoperatively for methicillin-susceptible S. aureus (MSSA) and MRSA nasal carriage. Risk factor analysis was performed by univariate analysis followed by stepwise logistic regression. Of the 323 patients included, 63 (19.5%) patients developed S. aureus infection (36 MRSA, 27 MSSA) within 1 month of surgery. Variables significantly associated with infection in the univariate analysis were MRSA and MSSA nasal carriage, alcoholic cirrhosis, absence of hepatocellular carcinoma, decreased prothrombin ratio, and presence of ascites. In the multivariate analysis, MRSA carriage (odds ratio [OR]: 20.9, P < 0.0001), MSSA carriage (OR: 3.4, P = 0.0004), alcoholic cirrhosis (OR: 2.4, P = 0.01) and decreased prothrombin ratio (OR: 1.2, P = 0.01) were independent predictors of infection. Molecular typing showed that the infecting isolate was identical to the isolate from the nose in most patients. In conclusion, preoperative nasal carriage of MRSA and MSSA is an independent risk factor for S. aureus infection in liver transplant recipients. The infection is most often of endogenous origin. Alcoholic cirrhosis and the severity of liver failure are also associated with a high risk of infection. (Liver Transpl 2005;11:1093–1099.)
Staphylococcus aureus has emerged as the leading cause of bacterial infection in liver transplant recipients.1, 2 These infections are mostly caused by strains that are resistant to methicillin, have an early postoperative onset, usually within 1 month of surgery, and are associated with high mortality rate.1, 3, 4 Previous studies have shown that preoperative methicillin-resistant S. aureus (MRSA) nasal carriage is associated with a very high risk of S. aureus infection after orthotopic liver transplantation (OLT).3, 5, 6 However, MRSA carriers usually have more advanced liver disease than noncarriers.6, 7 Thus, the relationship between MRSA nasal carriage and infection may indicate that colonization of the nares plays a direct role in the pathogenesis of S. aureus infection or that host characteristics associated with nasal carriage also predispose to infection. To establish that nasal carriage is an independent risk factor for infection, a multiple logistic regression model is required to control for the effects of confounding variables such as the type and severity of underlying disease. Previous studies addressing risk factors associated with S. aureus infection in these patients did not perform multivariate analysis.1, 3, 6 Moreover, these studies did not investigate the risk associated with methicillin-susceptible S. aureus (MSSA) nasal carriage6 or failed to demonstrate a significant association with infection because of a relatively small number of patients.3, 5
The aim of the present study was to determine independent risk factors for early S. aureus infection (i.e., within 1 month of surgery) in liver transplant recipients using a multiple logistic regression model. For this purpose, we conducted a retrospective cohort study including all consecutive liver transplant recipients at Beaujon Hospital over a 75-month period. All patients were screened preoperatively for both MSSA and MRSA nasal carriage.
From January 1, 1997, to March 31, 2003, 363 patients underwent OLT in our hospital. Three patients whose medical records could not be located and 37 patients who died within 1 month of surgery without S. aureus infection were excluded from the study. The 323 remaining patients constituted the study population. When retransplantation occurred within 10 days of the initial transplant, it was considered a single transplant episode. The characteristics of the patients are shown in Table 1.
Table 1. Characteristics of the 323 Liver Transplant Recipients
NOTE: Data are numbers (%) of patients, unless otherwise indicated.
During the study period, standard perioperative antibiotic prophylaxis consisted of cefoxitin, as previously described.3 Most patients received tacrolimus and corticosteroids as primary immunosuppressive therapy. Doses of tacrolimus were adjusted to achieve plasma levels of 10 to 15 ng/mL in the month following transplantation. Corticosteroid therapy consisted of an intravenous methylprednisolone taper from 5 mg/kg/day to 0.3 mg/kg/day on postoperative days 1 to 8 followed by oral prednisone (20 mg/day). From January 1997 to July 1998, ciclosporin was used alternatively to tacrolimus. During the period 1997–2000, patients received azathioprine (2 mg/kg/day) in addition to tacrolimus and corticosteroids. Patients with renal failure at the time of surgery received antilymphocyte serum as induction therapy instead of tacrolimus or ciclosporin.
Isolation procedures were applied to patients with MRSA nasal carriage during their hospitalization.
The following preoperative and perioperative variables were collected from the medical records and analyzed in relation to S. aureus infection. Pretransplantation data included age, sex, underlying liver disease, presence of an hepatocellular carcinoma, history of ascites, spontaneous bacterial peritonitis, encephalopathy or gastrointestinal hemorrhage, prior abdominal surgery, dialysis, prothrombin ratio and serum levels of albumin, creatinine and total bilirubin at the time of enrollment on the waiting list, cytomegalovirus serostatus, S. aureus nasal carrier state, and duration of time on the waiting list. Perioperative variables were duration of surgery, type of transplant (whole or reduced-size organ), type of donor (deceased or living), number of transplants (liver only or multiple organs), and requirement for blood (number of units of packed red blood cells transfused).
Definition of S. Aureus Infection
The medical and microbiological records of the patients were reviewed for S. aureus infection in the month following OLT.
Bacteremia was defined as the isolation of S. aureus from at least 1 set of blood cultures in the presence of clinical signs of infection. It was considered as primary bacteremia if it was of unknown origin or related to an intravascular catheter, and secondary bacteremia when it was related to a documented site of infection. Bacteremia was considered to be catheter-related when quantitative culture of the catheter tip yielded S. aureus with a count of >103 colony-forming units/mL in the absence of another site of infection.
Pneumonia was diagnosed when a new pulmonary infiltrate developed in conjunction with clinical symptoms and S. aureus was isolated from a protected brush specimen (>103 colony-forming units/mL) or a bronchial aspirate (>106 colony-forming units/mL).
An intra-abdominal abscess was defined as a collection of fluid, drained surgically or under ultrasound guidance, containing numerous polymorphonuclear cells and yielding a heavy growth of S. aureus. Peritonitis was diagnosed if the ascitic fluid obtained intraoperatively or by percutaneous aspiration contained ≥250 polymorphonuclear cells/mm3 and S. aureus was cultured, in the absence of an intra-abdominal abscess.
Wound infection was considered to be present when S. aureus was isolated from a purulent fluid drained from the wound.
Sinusitis was diagnosed when the fluid drained from the maxillary sinus contained numerous polymorphonuclear cells and S. aureus was isolated in heavy growth.
Statistical analysis was performed using the SAS version 8.2 program (SAS Institute, Carry, NC). Univariate analysis was used to identify associations between each of the variables recorded and S. aureus infection. The chi-square test and, for small numbers, Fisher exact test were used for comparison of categorical data. Continuous variables were compared using the Student t test. A P value ≤0.05 was considered statistically significant. For multivariate analysis, only variables that demonstrate a P value < 0.2 in the univariate analysis were entered into a stepwise logistic regression model. The interactions for all the couples of the variables that remained in the final model were also tested.
Microbiological Studies and Molecular Typing
All patients were screened preoperatively for nasal carriage of S. aureus. A nasal swab was taken before surgery and plated onto mannitol-salt agar. After incubation at 37°C for 24 to 48 hours, S. aureus was identified by microscopic and growth characteristics, the coagulase test, and DNA hydrolysis. Methicillin resistance was determined by the disk diffusion method on Mueller-Hinton agar incubated for 24 to 48 hours at 30°C with disks of oxacillin 5 μg (Bio-Rad, Marnes-la-Coquette, France).
For all patients who carried MRSA and subsequently developed MRSA infection, the isolate from the nose and the isolate from the infected site were compared by molecular typing. For MSSA carriers with infection, only isolates from blood cultures were available for typing and were compared to the nasal isolates. Typing was done by pulsed-field gel electrophoresis of restriction fragments after DNA digestion with Sma I, as previously described.8
S. Aureus Infection
Overall, 63 (19.5%) of the 323 patients developed S. aureus infection within 1 month of surgery. The infection was caused by MRSA in 36 (11.1%) patients and by MSSA in 27 (8.4%) patients. The mean time to onset of infection was 10.2 days (range, 2–30 days). The types of infection were lower respiratory tract infection (27 patients), primary bacteremia (14 patients), surgical wound infection (13 patients), peritonitis (11 patients), intra-abdominal abscesses (8 patients), and sinusitis (1 patient). Of the 63 infected patients, 24 (38.1%) had bacteremia, including 14 primary (catheter related) bacteremia and 10 secondary bacteremia. The sites of infection did not differ significantly for MRSA and MSSA (data not shown).
Thirty-five patients were treated with a glycopeptide (32 vancomycin, 3 teicoplanin), 22 patients received a beta-lactam agent (11 amoxicillin-clavulanic acid, 5 oxacillin, 4 piperacillin-tazobactam, 2 cefotaxime), in association with a fluoroquinolone in 4 patients and with gentamicin in 2 patients. The remaining 6 patients, who had only a superficial wound infection, were treated by debridement and did not receive antibiotics. The mortality rate among infected patients was 6.3% (4/63) at 30 days and 12.7% (8/63) at 6 months.
Risk Factors Analysis
The characteristics of S. aureus-infected patients and noninfected patients are compared in Table 2. Variables significantly associated with infection in the univariate analysis were the type of underlying liver disease (P = 0.012), absence of hepatocellular carcinoma (P = 0.034), presence or history of ascites (P = 0.01), decreased prothrombin ratio (P = 0.003), MSSA nasal carriage (P = 0.004), and MRSA nasal carriage (P < 0.001). Further analysis of the association between underlying disease and infection showed that S. aureus infection was significantly more frequent in patients with alcoholic cirrhosis than in other patients (23/67 [34.3%] vs. 40/256 [15.6%], P = 0.0006). In contrast, the rate of infection did not differ significantly in patients with hepatic cirrhosis (13.6%), primary biliary cirrhosis (13.8%), fulminant hepatitis (16.7%), and other diseases (19.4%). The presence or absence of alcoholic cirrhosis was therefore included in the multivariate analysis.
Table 2. Comparison of Patients With and Without S. Aureus Infection by Univariate Analysis
Infected Patients (n = 63)
Patients Without Infection (n = 260)
NOTE: Data are numbers (%) of patients unless indicated otherwise.
In the multivariate analysis, 14 risk factors with P < 0.2 were tested. In the final model, only MRSA nasal carriage (odds ratio [OR]: 20.9), MSSA nasal carriage (OR: 3.4), alcoholic cirrhosis (OR: 2.4), and decreased prothrombin ratio (OR: 1.2 for a decrease of 10% in the prothrombin ratio) were found as independent predictors of infection (Table 3). No significant interactions between the covariates kept in the model were found. The differences in infection rates according to the S. aureus nasal carrier status, the alcohol-related origin of cirrhosis, and the prothrombin ratio divided in 2 classes (< or ≥ 40%) are illustrated in Figure 1.
Table 3. Multivariate Analysis of Risk Factors for S. Aureus Infection in Liver Transplant Recipients
OR (95% Confidence Interval)
MRSA nasal carriage
MSSA nasal carriage
Decreased prothrombin ratio
Association Between S. Aureus Nasal Carriage and Infection
Of the 323 patients included in the study, 82 (25.4%) were S. aureus nasal carriers. Of these 82 carriers, 63 harbored MSSA, 18 harbored MRSA, and 1 harbored both MSSA and MRSA. S. aureus infection developed in 15/19 (79%) MRSA carriers, 20/64 (31.2%) MSSA carriers, and 28/241 (11.6%) noncarriers. The mean time of onset was 6.7 days for MRSA carriers, 9.3 days for MSSA carriers, and 12.9 days for noncarriers. The infection occurred within 10 days of surgery in 14/15 (93.3%) MRSA carriers, 13/20 (65%) MSSA carriers, and 11/28 (39.3%) noncarriers (P < 0.01).
In the 15 MRSA carriers who subsequently developed S. aureus infection, the infection was caused by MRSA in all patients. The pairs of isolates were compared by molecular typing. In 13 (86.7%) of the 15 patients, the isolate from the infected site was identical or closely related (≤3 band difference) to the nasal isolate. The 2 isolates had distinct patterns (>3 band difference) in the 2 remaining patients. Figure 2 shows the patterns of selected isolates from patients with MRSA bacteremia.
Of the 20 MSSA carriers in whom S. aureus infection occurred, 18 had MSSA infection and 2 had MRSA infection. Molecular typing showed that the isolate from the nose was identical to that from the blood in the 10 carriers who developed MSSA bacteremia (data not shown). Of the 28 noncarriers who became infected, 19 had MRSA infection and 9 had MSSA infection.
The anterior nares are the main reservoir of S. aureus and may serve as an endogenous source for the development of infection in colonized patients. Thus, nasal carriage has been associated with an increased risk of infection in patients receiving hemodialysis, intensive care unit patients, surgical patients, liver transplant recipients, HIV-infected patients, and patients in long-term care facilities.3, 5, 6, 9–15 The risk may be higher in patients colonized with MRSA than in those colonized with MSSA.3, 5, 12, 13 However, MRSA carriers often have a poorer clinical condition than other patients and, therefore, differences in infection rates could also be related to differences in host characteristics.
In the present study, a multiple logistic regression model was used to identify risk factors for S. aureus infection in liver transplant recipients. Our results indicate that both MRSA nasal carriage and MSSA nasal carriage are independent predictors of infection in these patients. The rate of infection was 79% in MRSA carriers and 31% in MSSA carriers, vs. only 11.6% in noncarriers. Infection occurred within 10 days of OLT in most nasal carriers, indicating that it arose from events during surgery or the immediate postoperative period. As previously reported in other patients,14 molecular typing showed that the infecting isolate was identical to that from the nasal swab in the large majority of patients, thereby confirming the endogenous origin of infection. In the noncarriers who became infected, nasal screening may have failed to detect S. aureus carriage in patients with rectal colonization only, and the strain may have been acquired after surgery in a few patients as a result of horizontal spread despite isolation procedures.
Our results suggest that preoperative nasal carriage plays a key role in the pathogenesis of S. aureus infection in liver transplant recipients. Candidates for OLT should systematically be screened for both MRSA and MSSA nasal carriage. It is noteworthy that although the effects of host characteristics were controlled by multivariate analysis, colonization with MRSA was still associated with a greater risk of infection (OR: 20.9) than colonization with MSSA (OR: 3.4). Although perioperative prophylaxis with cefoxitin probably contributed to decreased rates of MSSA surgical site infection, it cannot entirely account for this finding. Differences in infection rates could also be related to differences in intrinsic virulence of strains. However, whether MRSA is more virulent than MSSA remains a controversial issue.16, 17
In addition to preoperative nasal carriage, the type of underlying liver disease appeared to have an impact on the risk of S. aureus infection, and multivariate analysis identified alcoholic cirrhosis as an independent predictor of infection. Alcohol consumption is associated with an impairment of the humoral and cell-mediated immune response, which explains that chronic alcoholics are particularly susceptible to community-acquired infections.18–21 A recent study showed that heavy drinking also increases the risk of nosocomial infection following general surgical procedures.22 Farges et al.23 found that patients who undergo OLT for chronic alcoholism have a lower incidence of acute rejection and a higher incidence of bacterial infection than other liver transplant recipients. Our study indicates that alcoholic cirrhosis is independently associated with a higher risk of S. aureus infection after OLT. Since the reversal of the immune dysfunction is progressive after ethanol withdrawal,19 the immunosuppressive effect may have persisted in our patients despite several months of sobriety before surgery and subsequently increased the risk of S. aureus infection in the early postoperative period.
The severity of liver failure, as reflected by a decreased prothrombin ratio, was also associated with S. aureus infection in the multivariate analysis. This finding indicates that patients with more advanced liver disease prior to OLT are more likely to develop infection after surgery. This may account for the substantial rate of infection (16.7%) observed in patients with fulminant hepatitis who usually do not have a history of previous hospitalization and, therefore, rarely carry MRSA but present with dramatic liver failure.
Our results have obvious implications for the prevention and treatment of S. aureus infection in liver transplant recipients. The identification of high-risk patients is useful when choosing perioperative antibiotic prophylaxis and empirical therapy for suspected early infections. Thus, the empiric use of glycopeptides may be justified not only in patients with MRSA nasal carriage but also in patients with alcoholic cirrhosis and/or severe preoperative liver failure. In our study, patients were screened in the immediate preoperative period, and because of the delay required for incubation, culture results were available only 24 to 48 hours after surgery. The perioperative and early postoperative management of these patients may be improved by the use of a rapid real-time polymerase chain reaction aimed at detecting S. aureus from nasal swabs.24
Eradication of S. aureus nasal colonization in candidates for OLT is a potential strategy for decreasing the rate of subsequent infection. A recent evidence-based review suggested that intranasal mupirocin is effective at reducing S. aureus nasal carriage and that although it should not be routinely used to prevent infection it may be useful in selected populations of patients with an acute period of risk for infection.25 However, cirrhosis is known to be a risk factor for S. aureus colonization, and candidates for OLT are at high risk for recolonization during the waiting period because of the frequent hospitalizations imposed by severe liver disease.26 In a recent study,27 mupirocin failed to prevent S. aureus infection in a liver transplant unit, since many successfully decolonized patients subsequently became recolonized. Moreover, many patients with nasal carriage are also rectal carriers.28 Thus, the optimal strategy for eradication of S. aureus colonization in candidates for OLT remains to be determined.
In conclusion, preoperative nasal carriage of both MSSA and MRSA is an independent risk factor for S. aureus infection in liver transplant recipients. The infection is most often of endogenous origin. Alcoholic cirrhosis and the severity of liver failure are also associated with a high risk of infection.