Bacterial Infections after Pediatric Liver Transplantation

Background: Liver transplantation (LT) is the only curative therapy for acute and chronic end-stage liver disease in pediatrics. Nevertheless, bacterial infection post-transplant account for one of the most signicant causes of mortality which requires attention, along with appropriate antibacterial treatment. Therefor this study was designed to provide a multifaceted overview of bacterial infections during a one-year experience in the biggest pediatrics LT center in Iran. Methods: In this retrospective cohort study, records from all liver transplant pediatrics in Abu-Ali Sina hospital, Shiraz, Fars, Iran from April 2019 to February 2020 were assessed. Demographic, laboratory, and clinical data were extracted along with the administered therapeutic approach for the patient. Results: an incidence of 67.9% bacterial infection rate was observed, in which more than 64.06% of isolated pathogens were Gram-negative and 35.93% were gram-positive. The most isolated pathogens were E. coli sp. and pseudomonas sp. for gram negative, and enterococci and staph sp. for gram positive. Our study demonstrated that Intra-abdominal Surgical Site Infections (24.24%) was the most common site of infections in post LT pediatrics. According to the results of the sensitivity-resistance pattern of isolated pathogens, 24.39% of isolated gram-negative bacteria were Extensively drug resistant (XDR). Also, 30.43% of gram-positive bacteria were Vancomycin-resistant enterococci (VRE) and 8.69% of them were methicillin-resistant staphylococcus aureus (MRSA). Furthermore, multivariate regression analysis showed that the only risk factor for bacterial infections after LT is the length of stay in the ICU. The mortality rate in our study was 22%, which was signicantly higher among the infection group (P = 0.008). Also, it was found that patients receiving colistin had a signicantly shorter in length of stay in the ICU and hospital than those who did not receive colistin. There is no any signicant difference between piperacillin-tazobactm versus meropenem as empirical treatment regarding clinical outcomes. Conclusion: a high rate of bacterial infection was detected among hospitalized pediatrics in the early


Background
Advances in medical science have increased the life expectancy of patients suffering from organ failure and end-stage disease. Patients with liver failure also fall into this category [1,2]. One of the most effective treatment strategies for the survival of patients with advanced liver failure is liver transplantation (LT) [3,4], especially in pediatrics in increasing their life expectancy. Indications for LT in pediatrics generally include cholestatic diseases, metabolic disorders, acute liver failure due to viral infections or drug use, chronic hepatitis, and malignancies [5].
Infections have obtained one of the most important causes of morbidity and mortality after solid organ transplants even compared to transplant rejection [6,7]. The rate of infection after liver transplant varies in different studies ranging from 21 to 47% among pediatrics [8][9][10].
The most frequent organism of post-transplant infection is usually bacterial (up to 70%), followed by viral and fungal [11,12]. Bacterial infection in these patients may range from less severe infections, such as super cial wound infections, to severe, such as septic shock [8]. Various factors contribute to predisposing pediatrics liver transplant recipients to a greater risk for colonization and infection with multidrug-resistant organisms (MDRs), which include recurrent and prolonged hospitalizations, impairment of the immune system, and multiple uses of antibiotics [13]. Since there are signi cant developmental variations among pediatrics and adults that can considerably in uence the pathophysiology and prognosis, caution is warranted in extrapolating the results of adult studies to pediatrics. Therefore, the evaluation of infection rates among pediatrics receiving LT is justi ed.
Therefore, this retrospective cohort study was designed to evaluate a multifaceted overview of bacterial infections during a one year experience which includes: incidence, type and susceptibility-resistance patterns of isolated pathogens, risk factors related to infections, as well as evaluation of different antibiotic regimens and their outcome in pediatrics candidate for LT in Shiraz organ transplant center, Abu-Ali Sina hospital as the only and biggest pediatrics LT center in Iran.

Study design
In this retrospective cohort study, clinical and laboratory records from all liver transplant pediatrics under 18 years of age in Abu-Ali Sina hospital, Shiraz, Fars, Iran from April 2019 to February 2020 were assessed. This study was approved by the ethics committee of Shiraz University of Medical Sciences. All of the protocols were based on the ethical guidelines of the 1975 Helsinki Declaration. Written informed consent was obtained from all the patients' parents or guardians.

Data Collection
The medical and laboratory records of all the patients were extracted from available electronic inpatient medical record databases in hospitals and reviewed by a pharmacist under the vision of clinical pharmacists. Demographic data such as age, gender, body mass index (BMI), etiology of end-stage liver disease, Pediatric End-Stage Liver Disease (PELD) score or the Model for End-Stage Liver Disease (MELD) score, and previous history of admission and antibiotic usage within one week to three months before of liver transplant were recorded. The technique of LT, the mean operation time, immunosuppressive regimens, and length of ICU and hospital stay, readmission, rejection or re-transplantation episodes, and clinical outcome after transplantation were also evaluated.

Infection Assessment
All patients in our center received antimicrobial prophylaxis consisting of ampicillin-sulbactam (150 mg/kg/day) and ceftizoxime (150 mg/kg/day till 72 hours after LT in all patients, except in those with biloenteric anastomosis (continued for 5 days). All pediatrics received trimethoprim/sulfamethoxazole (TMP/SMX) prophylaxis with 2-3 mg/kg TMP daily for the rst 6 months after LT for prophylaxis of Pneumocystis jiroveci, uconazole (3-5 mg/kg/day) for prophylaxis of fungal infection, and ganciclovir (10 mg/kg/day) or valganciclovir (13 mg/kg/day) for prophylaxis of cytomegalovirus infection that was commenced and continued for at least one month after LT. [14].
Also, as a suppressive regimens, a combination of tacrolimus, prednisolone, and mycophenolate mofetil are prescribed and according to the condition of the transplanted graft, and also based on the plasma levels and the duration since the time of transplant, doses will be adjusted.
All the patients were visited and evaluated daily by a transplant surgeon, pediatric gastroenterohepatologist, and a subspecialist of pediatric infectious diseases after transplantation.
If an infection is suspected, the necessary work-up in this area will be conducted, including chest and abdominal radiography, in ammatory factors such as erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), procalcitonin, complete blood count (CBC) blood culture samples (central and peripheral), urine and sputum culture, and in case of presence of any drains, samples will be obtained and sent for analysis. Then, at the discretion of the pediatric infectious disease specialist and based on the patient's clinical condition, a combination of vancomycin or clindamycin for empirical coverage of gram-positive pathogens with additional antibiotics covering gram-negative bacteria such as carbapenems, betalactambetalactamase inhibitor or Fluoroquinolones and, if necessary, an antifungal is used and after determining the type of pathogen isolated, de nitive therapy is performed.
In this study, the classi cation type of infections was based on the CDC/NHSN surveillance classi cation of healthcare-associated infection and criteria for speci c types of infections in the acute care setting [15]. sepsis diagnosis is based on Surviving sepsis campaign [16] and Intra-abdominal infection includes peritonitis, Cholangitis, abdominal abscess, and Infected hepatic necrosis based on pathological and radiographic ndings.
Following data were recorded for each episode of infection: sign and symptoms, date of the event according to CDC de nition, site of infection, isolated organism and its antimicrobial susceptibility pattern (which performed by using Kirby-Bauer disc diffusion method and according to CLSI guideline), type and length of antibacterial usage and clinical outcome of infections. Multidrug-resistant (MDR) and Extensive Drug Resistance (XDR) pathogens were de ned in accordance with the Consensus Statement of the European Centre for Disease Prevention and Control and the American CDC evidence of infection [14].

Statistical Analysis
All the statistical analyses were performed by statistical package for social sciences (SPSS Inc., Chicago, Illinois, USA) version 26.0. Descriptive statistics were presented as mean ± SD and proportions as appropriate. Factors affecting the difference between the infected and non-infected group was done by using a binary logistic regression model and multivariate analyses. The Chi-square test or Fisher's exact test was used to compare categorical data. The statistical tests used in the univariate analysis were Student's t-test and Mann-Whitney U test. Covariates signi cant (P < 0.05) in the univariate model were included in the multivariable linear regression. The possible association between different demographic, clinical, and paraclinical characteristics of the study population and the development of infection was assessed by the multivariate logistic regression analysis. In the rst step, each independent variable entered into the univariate model. In the following, those with P < 0.05 in the univariate model were selected and entered into the nal multivariable regression model. Comparison of the effect of different empirical antibiotic regimens on white blood cells, body temperature, and CRP levels at consecutive days after starting antibiotic therapy was done by using a one-way analysis of variance (ANOVA) with repeated measures. The statistical signi cance level was set at p < 0.05.

Results
Over the period of this study, 84 patients had undergone LTs in our center in which four were excluded due to incomplete clinical records. Among the 80 enrolled pediatrics, the median age was 60 months (mean: 72.18 ± 48.48; range from 8 to 204 months), with a 52 (65%) male to 28 (35%) female rate. Table 1 demonstrates the demographic and clinical related data of these patients with a comparison between the infected and non-infected group.  Table 2  Furthermore, the most frequent site for isolated XDR pathogens was the abdomen, while the most frequent one for MDR was the urinary tract. Our results, as demonstrated in Table 1, showed that the length of stay in ICU and hospital, length of mechanical ventilation after transplant, re-hospitalization and mortality rate were signi cantly higher in the group of infected than in non-infected pediatrics.
After assessing the quantitative and qualitative variables and their relationship with infection, multivariate regression analysis showed that the only risk factor for bacterial infections after LT was the length of stay in the ICU (Table 3).  Table 4 demonstrates the frequency of antibiotic use in empirical and de nitive conditions along with the average duration of administration during the study period. Evaluation of antibiotics prescribed revealed that 57.1% of empirical regimens included Vancomycin with carbapenems (mainly meropenem), 30% included vancomycin and piperacillin-tazobactam, and 12.9% included Fluoroquinolones (mostly levo oxacin). Regarding clinical outcome indexes, the rate of re-hospitalization was comparable (P = 0.325) between patients with (37.5%) and without polymyxin E (Colistin) (15.38%) as a part of their de nite antibiotic therapy regimen. The mortality rate was also comparable (P = 1) between patients with (22.22%) and without colistin (16.67%). In contrast, the median lengths of both hospital and ICU stay were signi cantly lower (P = 0.015) in patients who did receive colistin (22 days) than those not given colistin as a part of their de nite antibiotic therapy regimen (32 days).

Discussion
In pediatric LT, postoperative bacterial infections occur frequently and are potentially life-threatening. In this one-year, retrospective study, an incidence of 67.9% infection rate was observed, which was higher than a previous study in Iran, reporting an incidence of 54.3% bacterial infection among liver transplant pediatrics. [17]. Our ndings are also in accordance with the limited reports worldwide regarding LT in pediatrics (51.9% in Germany and 70.8% in France). [18][19][20][21][22]. In a 23-year, retrospective single-center study conducted by Kukreti et al, 49 incidents of microbiologically documented infection were observed among 145 pediatrics (34%) in post LT, which 79% had bacterial infections only during their stay in the pediatric ICU. [23]. Numerous factors have contributed to the ampli ed risk of by bacterial etiology in post-LT pediatrics especially during the early periods; including the di culty of surgical procedures; high levels of immune suppression because of rejection; numerous points of access for microorganisms (eg probes, catheters, incisions); and patient's poor health state [10,18,19,23,24].
Based on our results, more than 64.06% of the isolated pathogens were Gram-negative and 35.93% were gram-positive. Our data regarding the proportion of causative pathogens is in accordance with the previous study among post-liver transplant pediatrics in Iran, which showed Gram-negative isolates were comparatively as common as Gram-positive isolates (51% vs 49%), with a predominance of Enterococcus spp (36.1%) and Staphylococcus spp (11.1%) as Gram-positive bacteria and Enterobacteriaceae (21.3%), Acinetobacter spp (16.7%), as Gram-negative ones. Gram negative predominance may be due to various factors such as LT patients' longer stay in the ICU, longer duration of mechanical ventilation, along with the possibility of kidney failure after transplantation and the need for dialysis, as well as a long history of the use of preoperative broad-spectrum antibiotics due to recurrent hospitalizations and also biliary tract manipulation during surgery, all of which are risk factors for gram-negative infections [25,26].
However, some studies have demonstrated a predominance of Gram-positive bacteria; such as in a French study, bacterial isolates showed a dominance of Gram-positive bacteria (78%) which included Staphylococcus aureus (32%) and Staphylococcus epidermis (26%). It is also worth mentioning that all patients in that study were administered gentamicin, polymyxin, and nystatin during the time of their ICU stay for selective intestinal decontamination or until resumption of oral intake in order to reduce not only the predominance of Gram-positive bacteria, but also the Gram-negative aerobic bacteria in the intestinal ora [19,27,28].
Another concern for pathogens causing infection after transplantation is the emergence of resistant species such as MDR, XDR, and VRE. The results of our studies showed that about 25% of gram-negative species isolated from patients were XDR type; other studies in this eld also pointed to the high incidence of MDR and XDR pathogens after solid organ transplantation [29][30][31].The increase in rates of resistant pathogen species not only affects the e ciency of common antibiotic regimes, but also increases the mortality rate [32,33].Inappropriate administration of antibiotics in empirical conditions, longer hospital and ICU stays, frequent prescription of broad-spectrum antibiotics such as carbapenems and uoroquinolones in treatment of spontaneous bacterial peritonitis conditions before LT, as well as the need for hemodialysis or Continuous Renal Replacement Therapies (CRRT) after LT are among the risk factors for emerging resistant pathogens after LT [25,32].
In the case of gram-positive pathogens, the incidence of VRE infections has been reported to be about 31% in our study, while in the study of Pouladfar et al., 82% of Enterococcus species were VRE [17]. The reasons for this difference in the incidence of VRE can be attributed to the establishment and implementation of stewardship antibiotic programs in our hospital under the supervision of infectious disease and clinical pharmacist specialists in recent years, which has led to a dramatic reduction in the use of inappropriate broad-spectrum antibiotics such as vancomycin.
Our study revealed that Intra-abdominal Surgical Site Infections (SSI) were the most common site of infections in post-liver transplant patients. Based on the studies exploring infection in hospitalized pediatrics during the early periods after LT, the two most frequent sites of infection. [19,22,23] were the blood and abdomen. Execution of complex surgical procedures in LT pediatrics and the requirement for insertion of intra-abdominal JP Drain and central vein catheter can have a major impact on the developing infections in the abdomen and blood.
Based on our results, length of stay in ICU and hospital, length of mechanical ventilation after transplant, re-hospitalization and mortality rate were signi cantly higher in the infected group rather than in noninfected pediatrics. Furthermore, multivariate regression analysis showed that the only risk factor for bacterial infections after LT in pediatrics is the length of stay in the ICU. Factors such as a longer hospitalization and longer stay in ICU have been formerly described to be linked with increased infection in LT by other authors [17,34,35]. Although our study revealed that the tacrolimus level was higher in the infected group compared to the non-infected group, no statistical signi cance was observed. Dohna Schwake et al. demonstrated tacrolimus levels above 20 ng/ml were associated with a higher risk for bacterial infection, especially severe sepsis, septic shock, and SSI. However, this may be due to the complete avoidance of steroids in the absence of rejection in their study in which as a consequence, tacrolimus trough level targets might have been higher [36].
One of the goals of this study was to evaluate the success rate of different antibiotic regimens in empirical as well as de nitive therapy. The results of our study showed that receiving carbapenem along with vancomycin, as an empirical regimen, shortens the mechanical ventilation duration compared to beta lactam -beta lactamase inhibitors, although non-signi cant. Nevertheless, neither of the two regimens was superior in reducing the signs and symptoms of infection. Although the use of carbapenems in the empirical treatment may seem to be necessary with the increase of MDR pathogen species amongst LT, the results of our study showed no difference regarding the administration of carbapenems compared to beta lactam-beta lactamase inhibitors, which is also supported by other studies [37][38][39].
Given that the use of carbapenems increases the risk of carbapenem-resistant Enterobacteriaceae pathogens [40], some studies have compared the e cacy of carbapenems with beta lactam-beta lactamase inhibitors, especially in the management of ESBL species. This suggests that the use of piperacillin-tazobactam might be a promising alternative to carbapenems for the management of Enterobacteriaceae bloodstream infections, especially when MIC piperacillin-tazobactam is low (MIC ≤ 0.5/4 µgL) or the source of bloodstream infection is the genitourinary tract or abdomen [37]. Furthermore, some studies have also suggested that piperacillin-tazobactam may be a potential alternative to carbapenems in the treatment of AmpC-producing Enterobacteriaceae [41].However, further studies are essential to provide a more accurate and de nitive assessment of the effectiveness against beta lactambeta lactamase inhibitors versus carbapenems, especially since some studies still consider carbapenems to be superior to others regimes [37,41]. Furthermore, no studies have been conducted in this regard on pediatric LT until now.
Given the increasing trend of gram-negative XDR species, especially in the case of Acinetobacter pathogen, the use of polymyxin E (colistin) is inevitable. In our study, it was found that pediatrics receiving colistin had a signi cantly shorter length of stay in the ICU and hospital than those who did not receive colistin; however, no signi cant difference was observed in other factors such as mortality. Given that the number of patients receiving colistin in our study was low (n = 9), the generalization of their results is impossible. However, the results of other studies showed that in the treatment of XDR Acinetobacter baumannii (XDR-Ab) colistin-free regimen is not very successful, and a combination of colistin with carbapenems is most successful in the treatment of these pathogens [30].
The results of our study should be carefully evaluated because it suffers from some limitations; only bacterial infections were evaluated and other infections such as fungal and viral ones were not investigated. Also, the study is the result of a one-year evaluation, so more accurate results may be obtained by evaluating the patients during a longer period. Furthermore, lack of detailed information about the isolated pathogens, including the study of resistance genotypic patterns and the MIC value for MDR and XDR species, and also unavailability of some antibiotics in our centers, such as tigecycline or ceftazidime-avibactam, were the other limitations of our study.

Conclusion
Pediatric patients in the immediate postoperative period after LT have a high risk of bacterial infections, which induce morbidity and mortality. To sum it up, a high incidence of bacterial infection was observed among hospitalized pediatrics in the early period after LT. These infections were associated with longer hospital stays and higher mortality rates. Aside from longer ICU stay, there was no other risk factor associated with contracting an infection in multivariate analysis. Furthermore, the use of Piperacillintazobactam instead of carbapenems in empirical treatment may be rational.

Declarations
Ethics approval and consent to participate: The present study was approved by the Medical Ethics Committee of Shiraz University of Medical Sciences. The purpose of this study was completely explained to the patient and their parents and were assured that their information will be kept con dential by the researchers. The written consent form was obtained from the patient's parents Consent for publication