Risk Factors for Multidrug-Resistant Bacterial Infections After Liver Transplantation

Introduction

Recently, the total number of liver transplants performed in China has exceeded 6896 [1]. Liver transplantation is internationally recognized as the most effective treatment for end-stage liver diseases [1,2]. Although the number of liver transplants and the postoperative quality of life have improved in recent years, bacterial infection remains a major global challenge, and postoperative infection has become the leading cause of liver transplantation failure [3]. Studies have shown that the postoperative infection rate after liver transplantation is high, with reported incidences ranging from 30% to 70% [4–6]. Infection is the primary cause of death among liver transplant recipients and prolongs hospital stays and increases medical costs [5]. Infections can occur at any time following liver transplantation [5]. Kalpoe et al [7] reported that liver transplant recipients who do not develop infections have a 1-year survival rate of 90%, compared with only 67% in those who do. Notably, the 1-year survival rate after surgery for patients with carbapenem-resistant Klebsiella pneumoniae infections is only 29% [7].

Given the frequent use of invasive medical devices in liver transplant patients and the misuse of antibiotics in recent years, bacterial drug resistance has been increasing. Multidrug-resistant bacteria (MDRB) have emerged as major pathogens after liver transplantation [8,9], accounting for more than 50% of bacterial infections. Clinically, these include methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococci, extended-spectrum β-lactamase-producing Enterobacteriaceae (eg, Escherichia coli and K. pneumoniae), carbapenem-resistant Enterobacteriaceae, multidrug-resistant Pseudomonas aeruginosa, and multidrug-resistant Acinetobacter baumannii [10], as well as multidrug-resistant gram-negative bacteria [11].

MDRB infections after liver transplantation include the following as risk factors: age, preoperative use of antibacterial drugs, liver reserve function, operation time, blood transfusion volume, blood loss volume, and anhepatic phase, as well as catheter-related factors (eg, gastric tube, endotracheal intubation, nasobiliary tube, urinary catheter), nutritional factors, intensive care unit (ICU) length of stay, and postoperative biliary complications [12–15]. Therefore, analyzing and understanding the risk factors for MDRB infections after liver transplantation and, on that basis, formulating targeted prevention and control measures are of great importance for reducing the incidence and spread of MDRB infections and lowering infection-related complications. We performed a retrospective analysis of 4 years of clinical data on MDRB infections after liver transplantation at Beijing Youan Hospital, and we discuss the relevant risk factors to provide a basis for early infection control strategies and guiding treatment.

Material and Methods

GENERAL DATA:

This study was retrospective, and a total of 350 patients who underwent liver transplantation at the Liver Transplantation Centre of Beijing Youan Hospital between January 2019 and March 2023 were included. The inclusion criteria were: 1) complete medical records such as laboratory examinations, surgical records, anaesthesia records, special care records, and progress notes; 2) postoperative survival longer than 15 days; 3) aged over 18 years; 4) underwent only liver transplantation, not combined liver and kidney transplantation; and 5) no infection or bacterial colonization before surgery (diagnosed by bacterial culture). Data on survival outcomes and time of death within the 1-year follow-up were collected. The exclusion criteria were: 1) undergoing a second liver transplant; 2) received long-term immunosuppressive therapy before surgery; 3) underwent solid organ or bone marrow transplantation; and 4) received antibiotics before the operation. A flowchart of the study subject screening is shown in Figure 1.

Donor livers were obtained from organ donation after death, including 293 male donors and 57 female donors, with a mean age of 43±18 years. The causes of death included brain death resulting from traffic accidents in 244 cases, stroke in 75 cases, and other causes in 31 cases. All legal immediate family members of the donors signed the Chinese Confirmation Registration Form for Human Organ Donation, witnessed by organ coordinators and personnel from the Red Cross Society. The donor brain death assessment, organ acquisition, and distribution processes and recipient transplantation complied with relevant national laws, regulations, and medical ethics requirements. All donor livers were allocated to appropriate recipients on the liver transplantation waiting list at our center through the Chinese Human Organ Distribution and Sharing Computer System. Routine medical evaluation prior to organ donation was conducted to exclude contraindications, including assessment of cardio-cerebral injury, medical history, primary disease, immune status, biochemical indicators, pathogenic microorganisms, and solid organ imaging. If fatty liver or malignant tumours were suspected, rapid frozen section pathology was performed during acquisition to determine organ suitability.

The specimens included blood, urine, sputum, drainage fluid, and surgical incision secretions. All infectious samples were identified using the BD Phoenix 100 NMIC/ID4 and VITEK-2 GP-67 systems. The KB disc diffusion method was employed to assess the antibiotic sensitivity of pathogens, and the double-disc method was used to detect extended-spectrum β-lactamase-producing strains. Methicillin-resistant strains were identified using the cefoxitin disk diffusion method. The instrument used was the VITEK-2 automated bacterial analyzer (purchased from bioMérieux, France), and the quality control strains were obtained from the National Centre for Clinical Laboratories, Ministry of Health.

Diagnostic criteria for MDRB infections: The same strain tested positive in 2 or more consecutive tests and showed resistance to 3 or more classes of commonly used antibiotics (anti-Pseudomonas cephalosporins, anti-Pseudomonas carbapenems, compound preparations containing β-lactamase inhibitors, fluoroquinolones, and aminoglycoside antibiotics) [16]. Based on the Diagnostic Standard of Nosocomial Infection issued by the former Ministry of Health in 2001 [17], recipients were divided into an MDRB infection group (50 cases) and a non-MDRB infection group (300 cases). This study was approved by the Ethics Committee of Beijing Youan Hospital.

Extubation procedure: After liver transplantation, extubation was performed in the operating room or ICU according to the extubation criteria used in the operating room. The criteria for extubation in the operating room were being awake, spontaneous eye-opening, the ability to respond correctly to verbal instructions (nodding or eye-opening), hemodynamic stability, normal body temperature, spontaneous breathing with a tidal volume >6 mL/kg, a respiratory rate of 10–18 breaths/min, PETCO₂ of 35–45 mmHg, and SpO₂ >95% (FiO₂ <40%). All patients were admitted to the ICU for continuous monitoring and treatment after surgery.

Antimicrobial management protocol: Initial empirical antimicrobial prophylaxis consisted of third-generation cephalosporins (cefoperazone-sulbactam) or carbapenems (imipenem or meropenem) during surgery. Upon detection of MDRB infection, antimicrobial therapy was adjusted based on the culture and susceptibility results. Treatment typically involved combination therapy with 2 or more antibiotics, most commonly including polymyxin B or colistin plus a carbapenem for carbapenem-resistant organisms, or vancomycin plus linezolid for resistant gram-positive organisms. The median duration of the targeted antimicrobial therapy was 14 days (range 10–21 days).

Postoperative immunosuppressive therapy: Tacrolimus, mycophenolate mofetil, and corticosteroids were used to prevent rejection after liver transplantation, and the Tac dose was adjusted according to the monitored plasma concentration. The target range for Tac trough levels in the cohort was 5–15 ng/mL.

Stopping protocol for bacterial culture: Bacterial culture was discontinued after 2 consecutive negative results, provided the patient’s clinical symptoms had improved, such as no swelling or pain at the surgical incision, no purulent discharge, no fever, and normal laboratory test indicators.

Definitions of time intervals included postoperative hospitalization duration, calculated from the end of liver transplantation surgery to discharge or death, excluding preoperative waiting time; ICU stay, defined as total calendar days spent in the ICU after surgery, including multiple admissions; and intubation time, which was the cumulative duration of endotracheal tube placement from initial intubation to final extubation, measured in hours.

DATA COLLECTION:

Recipients’ clinical data were collected by reviewing hospital records, medical examination reports, and patient follow-up. The data included patients’ conditions on admission, postoperative conditions, and post-discharge review information. By reviewing the literature and combining it with clinical data, the risk factors included in the analysis were: preoperative data included sex, age, body mass index, drinking history, smoking history, disease type, preoperative comorbidities, the model for end-stage liver disease (MELD) score, the Child-Pugh classification, and the presence of infection; intraoperative data included operation time, anhepatic phase, blood loss and transfusion; and postoperative data included tracheal intubation post-transplantation time, length of stay in the ICU, whether reoperation occurred (any surgical reintervention within 30 days postoperatively), biliary complications, Tac blood concentration peak point (postoperative day 7 following liver transplantation), and laboratory examination indicators (hemoglobin, albumin and serum creatinine [Scr]). Smoking history was defined as smoking for 6 months or more, and alcohol history as consumption of 30 g per week for 1 year or more [18]. The anhepatic phase was defined as the period from cutting off the portal vein and the superior and inferior hepatic vena cava of the diseased liver to completing and reopening the blood flow after anastomosis of the new liver’s portal vein and the superior and inferior hepatic vena cava. Co-infection referred to infection with 2 or more bacteria. Preoperative comorbidities referred to pre-existing underlying diseases such as diabetes and hypertension. The cut-off point was hospital discharge.

Child-Pugh classification criteria [19]: The main indicators were serum bilirubin, plasma albumin, prothrombin time extension, ascites, and hepatic encephalopathy. Each indicator was scored 1, 2, or 3, and the scores of the 5 indicators were added to give a total score ranging from 5 to 15. Liver reserve function was then divided into 3 levels: A, B, and C, based on the total score, indicating different severities of liver damage (the higher the score, the worse the liver reserve function).

Model for end-stage liver disease scoring criteria: Cultures were performed when infection was suspected. Based on Scr, total bilirubin (TBil), the international normalized ratio (INR), and the type of primary disease, a recipient was calculated using the MELD score [1,20].

Timeframe and criteria for biliary complications: Biliary complications were recorded within 1 year after liver transplantation. Diagnostic criteria were based on clinical presentation, imaging (such as MRCP or ultrasound), and laboratory findings, consistent with definitions in the transplant literature.

The first postoperative phase (within the first 24 hours after surgery): Monitor the patient’s physical condition. Re-examinations were performed once a week during the first 3 months after surgery, once a month after 3 months, once every 3 months within the first year, once every 6 months within the second year, and once every 6–12 months thereafter. Subsequent follow-up visits were scheduled based on the patient’s condition and the physician’s recommendations. Patients were monitored continuously during their ICU and hospital stays.

STATISTICAL ANALYSIS:

Statistical analysis was conducted using IBM SPSS version 22.0. Normality was assessed using the Shapiro-Wilk test. Data following a normal distribution were expressed as mean±standard deviation (X±S), and independent sample t-tests were used to compare the 2 groups. For non-normally distributed continuous variables, the Mann-Whitney U test was applied. Categorical variables were compared using Fisher’s exact test. Categorical data were expressed as frequency (n) and percentage (%), and comparisons between groups were performed using the chi-square test. Univariate analysis was performed to identify factors associated with multidrug-resistant organism infections after liver transplantation. A P-value <0.05 was considered statistically significant. For non-normally distributed continuous variables, the Mann-Whitney U test was used for comparisons between 2 groups. Chi-square tests assume adequate expected frequencies (≥5 per cell). When this assumption fails for categorical variables, Fisher’s exact test provides valid inference. In the multivariate analysis, clinically relevant variables and potential confounders were included in the logistic regression model, regardless of their significance in the univariate analysis, to ensure a more comprehensive assessment of the independent association between risk factors and outcomes.

Results

COMPARISON OF BASELINE DATA BETWEEN THE 2 GROUPS:

A total of 350 patients were included after excluding 1 secondary liver transplant, 6 patients with incomplete medical records, and 21 patients who died within 15 days. Of these, 50 were in the MDRB-infected group and 300 were in the non-MDRB-infected group. During the follow-up period, the overall mortality rate was significantly higher in the MDRB infection group than in the non-MDRB infection group (28.0% vs 12.3%, P<0.001). Among the 14 deaths in the MDRB infection group, 9 (64.3%) were directly attributable to MDRB infection-related complications, primarily severe sepsis and multiple organ failure. The median time from MDRB infection diagnosis to death was 18 days (range: 7–45 days). One-year survival rates differed significantly between groups (MDRB infection group: 72.0% vs non-MDRB infection group: 87.7%, P<0.001). Kaplan-Meier analysis showed that MDRB infection was associated with significantly worse survival (log-rank test, P<0.001, see Figure 2).

In the MDRB group, infections occurred 15–26 days after surgery (mean 20.34±4.56 days). In the non-MDRB group, infections occurred 14–25 days after surgery (mean 21.23±5.68 days). The multidrug-resistant pathogens detected in the MDRB infection group included A. baumannii in 14 cases; K. pneumoniae in 14 cases; Enterococcus faecium in 8 cases; and E. coli, S. aureus, and Acinetobacter yoelii in 2 cases each. Moreover, A. baumannii and K. pneumoniae were co-detected in 6 cases, and E. faecium and P. aeruginosa were co-detected in 2 cases each. Regarding infection sites, there were 18 cases of pulmonary infection, 16 abdominal infections, 10 incision infections, 8 bloodstream infections, 6 biliary tract infections, 2 urinary tract infections, and 1 central nervous system infection.

The comparison of baseline data between the 2 groups showed no statistically significant differences in indicators such as sex distribution, age, smoking history, drinking history, body mass index, history of hypertension, history of diabetes, or distribution of disease types (P>0.05), indicating good comparability between the groups, as shown in Table 1. Among the 350 patients, 50 (14.29%) developed MDRB infection, whereas 300 (85.71%) did not.

UNIVARIATE ANALYSIS OF MULTIDRUG-RESISTANT BACTERIAL INFECTIONS:

Univariate analysis identified 6 potential risk factors for MDRB infection after liver transplantation: endotracheal intubation ≥48 h after transplantation, reoperation, Tac plasma concentration, hospital stay ≥30 days, Child-Pugh classification, and ICU stay ≥72 h. These factors were significantly more common in the MDRB group (P<0.05; see Table 2). Moreover, the Child-Pugh classification indicated that the survival rate of patients without MDRB infection was considerably higher than that of patients with MDRB infection.

MULTIVARIATE LOGISTIC REGRESSION OF MULTIDRUG-RESISTANT BACTERIAL INFECTIONS:

The logistic model used MDRB infection occurrence as the dependent variable (1=occurrence, 0=non-occurrence). Independent variables included statistically significant factors in univariate analysis and clinical factors. The results show that tracheal intubation ≥48 h after liver transplantation, reoperation, and the peak point of Tac blood concentration peak point ≥15 ng/mL on postoperative day 7 following liver transplantation are independent risk factors for MDRB infections after liver transplantation, as shown in Table 3.

Discussion

Liver transplantation has advanced considerably in recent decades, with alloimmune rejection becoming a manageable challenge due to improved immunosuppressive strategies. The 1- and 5-year survival rates have reached approximately 90% and 70%, respectively, making liver transplantation the most effective treatment for end-stage liver diseases. However, infections have emerged as a major early complication due to surgical trauma and immunosuppressive therapy [21]. Among these, MDRB infections substantially increase recipient mortality. Multidrug-resistant bacterial organisms resist antibiotics through multiple mechanisms, including the overexpression of efflux pumps, β-lactamase production, and chromosomal mutations leading to reduced porin expression [9,22–24]. In this study, we identified postoperative endotracheal intubation ≥48 h after transplantation, reoperation, and Tac blood concentration peak point ≥15 ng/mL as independent risk factors for MDRB infections. These findings underscore the need for early identification of high-risk patients and timely intervention strategies to reduce MDRB-associated morbidity and mortality.

The risk of MDRB infection after liver transplantation was 2.656 times higher in patients with postoperative endotracheal intubation ≥48 h after transplantation than in those with intubation <48 h, making it an independent risk factor for MDRB infection. However, whether the trachea can be extubated promptly after surgery is closely related to the patient’s physical condition. For patients experiencing respiratory difficulties or those who have undergone liver transplantation, timely extubation may not be possible. We should closely monitor the physical condition of patients with prolonged postoperative intubation and consider how medication affects MDRB infections. Relevant studies show that long-term ventilator support is a risk factor for gram-negative MDRB and fungal infections after liver transplantation [21,22]. This may be attributable to damage to the oral and upper-respiratory mucosa by intraoperative orotracheal intubation, which inhibits the normal cough reflex and compromises the mucosal immune barrier, weakening the respiratory system’s natural defences. In addition, recipients with indwelling endotracheal tubes often have longer ICU stays, where confined space, use of broad-spectrum antibiotics, and invasive procedures more easily lead to MDRB infections [25–27]. Zhong et al [28] found that prolonged postoperative tracheal intubation (>48 h) was an independent risk factor for multidrug-resistant gram-negative infections after liver transplantation, consistent with the findings of this study.

Reoperation increases the risk of MDRB infections, possibly by causing secondary trauma to the patient’s body, leading to a stronger stress response and greater injury, which weakens the immune system and increases vulnerability to bacterial invasion. Studies have found that the incidence of infection, including surgical site infections, is higher after retransplantation than after primary transplantation [29], and data suggest that the relative risk of infection following retransplantation is as high as 6.417 [30]. Some patients may still require additional surgeries after liver transplantation, which inevitably affects their condition. It has been reported that patients undergoing reoperation have a 10-fold increased risk of infection [31,32]. Reoperation after liver transplantation not only causes additional trauma but also considerably increases the risk of infection.

Tacrolimus remains a cornerstone of immunosuppressive therapy following liver transplantation, with standard target blood concentrations often based in protocols from kidney transplantation [33,34]. However, growing evidence suggests that high Tac concentrations, particularly in the early postoperative period, increase the risk of infection. Our study found that a Tac blood concentration peak point ≥15 ng/mL was independently associated with MDRB infections. This may be due to excessive immunosuppression, as high Tac levels strongly inhibit T-cell activation and impair host immune defences. Previous studies have also shown that reducing Tac exposure can lower infection rates and improve long-term renal outcomes [35]. Therefore, careful monitoring and individualized adjustment of Tac dosing are critical to balancing rejection prevention with infection risk.

Factors associated with infection following liver transplantation are multifaceted. For liver transplant patients with the above risk factors, clinicians should conduct thorough preoperative evaluations, continuously develop and refine surgical techniques, remain vigilant for postoperative infections, and provide targeted treatment or preventive measures based on the combination of different risk factors. This approach can help reduce infection rates after liver transplantation and improve patient outcomes. In addition, liver transplant surgeons should work closely with relevant physicians and infectious disease specialists, using multidisciplinary approaches to comprehensively assess and manage the various risk factors for post-transplant infections.

Although previous international studies have reported a high MELD score as an independent risk factor for post-liver transplantation infections, our study did not observe a statistically significant association between MELD score and postoperative infections in either the univariate or multivariate analyses. Several factors may have contributed to this discrepancy. First, the sample size of our cohort was relatively limited, which may have reduced the statistical power to detect subtle associations. Second, differences in patient populations, baseline characteristics, perioperative management strategies, and infection control protocols across institutions and regions may also have influenced outcomes. Third, in our center, patients with high MELD scores may receive more intensive monitoring and individualized perioperative care, potentially mitigating the risk of infection. These factors should be considered when interpreting our results, and further studies with larger, multicenter cohorts are warranted to validate our findings.

This study has several limitations. First, as a single-center retrospective analysis with a limited sample size, baseline consistency could not be fully ensured, and the statistical power was relatively low. Variability in postoperative management and incomplete documentation (eg, time to infection onset, infection site, and pre-transplant antibiotic use) can introduce bias and limit the accuracy of infection risk assessment. Additionally, the absence of detailed pre-transplant status (eg, ICU admission) and inconsistent records of postoperative indicators (eg, creatinine levels) restricted further analysis. The retrospective design also limits the ability to determine the temporal sequence between risk factors and MDRB infection. Importantly, some known risk factors reported in previous studies were not validated, possibly due to sample size constraints. Although patients with preoperative MDRB infections were excluded, in clinical practice such cases may still proceed to transplantation after risk assessment. In future research, we aim to include multicenter prospective cohorts and more comprehensive perioperative variables, such as the MELD score, intraoperative blood loss, and postoperative lactate levels, to better understand the dynamics and predictors of MDRB infections.

Conclusions

In summary, MDRB infections are a considerable threat to patients’ prognosis following liver transplantation, especially in the context of prolonged intubation, reoperation, and high Tac blood concentrations. Our findings highlight the importance of early risk identification and individualized perioperative management. Targeted strategies, such as timely extubation when feasible, close monitoring, and adjustment of Tac levels, and minimizing the need for reoperation, may help reduce the incidence of MDRB infections and improve clinical outcomes. Given the study’s retrospective nature and limited sample size, future multicenter prospective studies are warranted to validate these findings and further refine prevention strategies.