1 Introduction

Pediatric liver transplantation is often the treatment of choice for chronic liver diseases like biliary atresia and acute liver failure. Because of improvements in care over the last two decades, 1-year survival rates reach up to 95%. As mortality declined, morbidity especially in the early postoperative phase on the pediatric intensive care unit (PICU) remains a substantial problem [1-9]. Patients' reduced preoperative condition, long surgery time, massive transfusions, immunosuppression, and a disturbed gastrointestinal barrier mainly contribute to the early complications.

Bacterial infections like bloodstream infections, pneumonia, or surgical site infections (SSIs) are among the most frequent and severe complications after transplantation [10-17] and directly or indirectly cause about half of early postoperative deaths. SSIs occur in about a quarter of all transplant recipients with the clinical presentation ranging from superficial wound infections to intra-abdominal abscess, peritonitis, or infected necrosis leading to sepsis and requiring additional surgery. Observational studies and registries identified [10, 12, 18] multi-resistant bacteria (MRB) colonization, high-level immunosuppression, and insertion of an abdominal patch as risk factors for SSIs.

At the same time, little is known about protective measures to prevent SSI. In contrast to many other surgical procedures, only few and partially differing recommendations or guidelines on duration or drug choice for perioperative antibiotic prophylaxis (PAP) after pediatric liver transplantation exist. For example, the adults and pediatrics guidelines from 2013 recommended a third-generation cephalosporin plus ampicillin or piperacillin-tazobactam alone [19]. Guidelines from 2019 by the American Society of Transplantation Infectious Diseases Community of Practice stated that ampicillin/sulbactam could be an alternative for 48 h or less, not distinguishing between adults and children [20].

Current clinical practice differs from the mentioned recommendations. A study on regimes in different pediatric transplant centers found a huge variety: from single shot antibiotic to more than a week, and from narrow spectrum to broad-spectrum antibiotic drug [21]. Currently, no data on the clinical value of PAP in children after liver transplantation are available.

In this retrospective single-center study, we assessed the effect of a change in the antibiotic perioperative prophylaxis protocol in pediatric liver transplantation from cefotaxime plus ampicillin without predefined duration to ampicillin/sulbactam with preplanned duration. Additionally, we examined the influence of the duration of the PAP on the occurrence of SSIs.

2 Methods

3 Results

A total of 125 patients was included (P1: n = 59, P2: n = 66), of which six patients died (P1: n = 1 [2%], P2: N = 5 [8%]) (Table 1). No deceased patient experienced infectious complications. Because all but one patient died within 2 days after transplantation (Table S1), deceased patients were excluded from further analyses. In both periods, survivors showed similar demographics, MDR bacterial colonization, postoperative laboratory results, incidence of abdominal silicone patch for temporary wall closure, underlying disease, and type of transplantation (split versus full organ and deceased versus living related liver transplantation) (Table 1).

The median length of PICU stay declined from 14.5 days in P1 to 11 days in P2, and the median duration of invasive mechanical ventilation declined from 4 to 2 days. In P1, more patients experienced at least one episode of acute rejection than in P2 (Tables 1 and 2).

3.1 Perioperative Prophylaxis

The duration of PAP was 4.5 (IQR 3–7) days in P1 and 4 (IQR 3–6) days in P2. PAP duration complied with the newly implemented SOP in 45% (n = 54) of cases in both periods combined. In P1, 43% (n = 25) of the patients had a treatment duration that complied with the later implemented SOP, compared to 48% (n = 29) after its implementation.

The predominant antibiotic substance or combination for PAP changed from ampicillin + cefotaxime in P1 to ampicillin + sulbactam in P2 (Figure 1). A strong decline was observed for the combination of vancomycin + meropenem, with a complementary increase in meropenem monotherapy and piperacillin + tazobactam. Postoperative vancomycin usage dropped from 14% (n = 8) in P1 to 3% (n = 2) in P2, reflecting and OR for receiving vancomycin as part of PAP of 0.21 (0.04–1.04).

4 Discussion

In this retrospective single-center study, SSIs occurred in about a fifth of patients and were not impacted by the regimen of PAP. The placement of an abdominal patch was identified as a relevant risk factor for SSIs. The effect of PAP on the reduction in SSIs became smaller with longer duration. In our cohort, efficacy of PAP was limited, as 70% of the identified pathogens were resistant against the prophylaxis we had used. One could argue that this result should prompt prophylaxis with broad-spectrum antibiotics. However, a recent analysis of over 2500 pediatric liver transplantations did not find a benefit of lower SSI rate when broader-spectrum antibiotics were used [27].

Our results confirm that SSIs contribute to morbidity during the early postoperative phase after liver transplantation in children. Similar rates have been observed previously [28, 29]. On the other hand, lower infection rates can be achieved, as reported by Banach et al. [10]: 11% of segmental graft recipients developed an SSI, whereas only 4% of whole graft recipients developed SSI [10]. The same result was observed in our study, with recipients of deceased donor split grafts developing SSIs twice as often as recipients of whole or living donor grafts. According to literature, biliary complications are associated with SSIs [10, 28]. Thus, SSIs might be effectively reduced by surgery technique, potentially including prevention bundles rather than single measures. However, robust evidence on optimal (intraoperative) measures to reduce SSIs are lacking. Results from adult trials cannot be extrapolated directly due to differing risk factors and physiological characteristics [30]. Furthermore, studies from nonimmunocompromised children that failed to identify benefit from PAP [31, 32] cannot be directly transferred to children in the postoperative phase after LTx, which is characterized by high levels of immunosuppressants, often prolonged ventilation, use of vasopressors and sometimes associated wound healing disorders in the early postoperative phase. Prospective randomized controlled trials are warranted to identify optimal prevention bundles against SSIs at all stages and reduce antimicrobial exposure as well as reoperations.

An important cornerstone of balancing harm and benefit from antibiotic exposure is the optimal duration of PAP. In our study, no change in actually applied PAP duration was achieved. Compliance with the newly introduced recommendation on PAP duration remained below 50%, suggesting that additional considerations determined the duration of PAP administration regardless of the SOP. A recently published multicenter study across Europe found great heterogeneity regarding the duration of PAP, with a range from single dose PAP to 10 days [19]. In one study, standardization of perioperative antibiotic regimen in pediatric liver transplantation, gram-negative broad-spectrum antibiotic treatment for >48 h declined from 77% to 44% after implementation without an increase in infectious complications [31]. Our study found that prolonged PAP duration did not substantially reduce SSIs. A group of patients that might benefit from prolonged PAP duration are patients with an abdominal patch, which was an important risk factor for SSIs in this study.

Besides the duration, the choice of antibiotic substance for PAP is important. In this study, the change in PAP regimen resulted in a strong shift toward ampicillin/sulbactam instead of ampicillin + cefotaxime. Furthermore, the exposure to a combination of reserve antibiotics, that is, vancomycin + meropenem declined in favor of meropenem monotherapy or piperacillin/tazobactam without an increase in SSIs. Even stronger effects regarding vancomycin treatment with preserved outcomes were achieved in a study with higher baseline vancomycin usage (50%), which dropped to 7% after antibiotic regimen standardization [33]. Careful choice of antibiotics is especially warranted to minimize the exposure to hepato- and ototoxic medication, which is acknowledged for vancomycin but has been also reported for trimethoprim and amoxicillin/clavulanic acid in pediatric liver transplant recipients [34, 35].

While it is well-acknowledged that unwarranted use of broad-spectrum antibiotics contributes to the selection of multidrug-resistant (MDR) bacteria [20], recipients of solid organ transplantations are predisposed for infections with MDR bacteria due to previous exposure to antibiotics of donors and recipients, prolonged hospital stays, and immunosuppression [36]. In transplant patients with sepsis, MDR organisms were identified as causative pathogens in 48% of cases and conferred increased odds for mechanical ventilation and higher degrees of organ failure [26]. Alike, MDR carrier status and high levels of immunosuppression were independent risk factors for infections, SSIs, severe sepsis, and septic shock [11, 12]. The synthesis of these findings suggests that MDR carriers might benefit from individually tailored antibiotic regimens, as discussed in the literature [11, 37] and already practiced in some European liver transplant centers [19]. However, a study tempering enthusiasm for this approach was published by Cardile et al. on adult LTX recipients who were carriers of carbapenem-resistant Klebsiella pneumoniae [38]: In the standard antibiotic treatment group, microbiota richness after liver transplant patients was increased and harmful Enterobacteriaceae and Klebsiella spp. reduced compared with the group receiving targeted MDR antibiogram-based antibiotic treatment. The authors conclude that it might be appropriate to administer standard prophylaxis and reserve targeted regimens for complications [27], a suggestion that is supported by a study where MDR-associated complications occurred much later than viral or culture-negative infections [29]. Use of an abdominal patch was associated with an increased risk of SSI, possibly explained by the need for re-surgery and a dysfunctional barrier to the skin microbiome as in our patients there was no VAC dressing used with abdominal patch.

The main limitation of this study is the retrospective nature, which prevented reliable assessment of underlying reasons for treatment choices and changes in ongoing therapy. Furthermore, due to the single-center design, unobserved center-specific characteristics that influence the occurrence of SSIs and/or treatment outcomes cannot be ruled out or adjusted. Altogether, protocol violations were frequent making it even more difficult to draw substantial conclusions from our data. Reasons for choice of a different antibiotic prophylaxis were not always documented but included MDR carrier status, long hospital stay before transplantation, and recurrent culture-negative cholangitis.

In summary, this study showed that SSIs contribute substantially to postoperative morbidity in pediatric liver transplant recipients, which remained unaffected by a change in PAP regimen. The results imply that overriding considerations determined PAP duration, whereas the standard substances used were effectively modified. Vancomycin usage was further reduced in a setting with already infrequent routine usage and low proportion of MDR carriers. The failure to detect an impact of the substances used for PAP on the occurrence of SSIs add to the growing awareness that evidence-based PAP regimens tailored to this specific patient population or even to individuals' status of organ dysfunction and preoperative MDR carrier status are urgently warranted but not available. Multicentric joint efforts at international level are therefore needed to minimize pharmacotoxicity deriving from antimicrobial treatment and accompanying treatments including sedation and anesthesia during the vulnerable early postoperative phase.