Pediatric living donor liver transplantation (LDLT): Short- and long-term outcomes during sixteen years period at a single centre- A retrospective cohort study

Background and objectives Pediatric living donor liver transplantation (LDLT) is an effective tool for managing pediatric patients with end-stage liver disease (ESLD) with good long-term graft and patient survival, especially after improvement in peri-operative care, surgical tools and techniques; however, the morbidity and mortality after such a procedure are still a challenging matter. The study aimed to analyze short-and long-term outcomes after pediatric LDLT in a single centre. Methods We retrospectively analyzed 67 pediatric patients who underwent LDLT in the period from April 2003 to July 2018. The overall male/female ratio was 40/27. Results Forty-one (61.2%) of patients had ≥1 early and/or late morbidities; the early (less than 3months) and late (≥3months) ones affected 36(53.7%) and 12(17.9%) of them respectively. The 16-year graft and patient survivals were 35(52.2%) while early and late mortalities were 23(34.3%) and 9(13.4%) respectively. Sepsis and chronic rejection were the most frequent causes of early and late mortalities respectively. Moreover, more packed RBCs transfusion units, bacterial infections, and pulmonary complications were independent predictors of poor patient survival. Conclusions More packed RBCs transfusion units intra-operatively, and post-liver transplant (LT) bacterial infection, sepsis, chronic rejection, as well as pulmonary complications had a negative insult on our patients' outcomes, so proper management of them is mandatory for improving outcomes after pediatric LDLT.


Introduction
Living donor liver transplantation (LDLT) has become the gold standard treatment option for paediatrics with end-stage liver disease (ESLD), especially after improved patient selection, increased experience, advancement in (pediatric anaesthesia, surgical techniques, graft preservation, peri-operative and intensive care, medical management, antimicrobial medications as well as immunosuppressive agents) [1][2][3][4]. However, the complication rate after such a pediatric procedure is still high with a negative insult on transplanted grafts, pediatric recipient morbidities and mortalities [5][6][7].
To our knowledge; the short-and long-term outcomes after pediatric LDLT is few in literature studies, so we analyzed this issue in a single tertiary Egyptian centre for 16 years period.

Pediatric recipients and methods
We did this cohort study that analyzed short-and long-term outcomes after pediatric LDLT after being approved by our institutional review board and after obtaining written informed consent regarding surgeries and research from both the recipients' parents/Guardians and the donors. It was performed in the department of hepato-pancreatobiliary surgery, National liver institute, University of Menoufiya, Menoufiya, Egypt during the period from April 2003 to June 2019(the liver transplantation (LT) operations were done between April 2003 and July 2018 and the follow-up started from POD1 until June 2019 or until patient loss(median: 18 months; range(0.03-194 months))).
Our series involved 67 pediatric recipients (less than 18years) after exclusion of adults, recipients with data loss, and cases who refused research. Our work was registered in the research registry with registration NO of researchregistry4593 (www.researchregistry.com) and it was reported in line with the STROCSS criteria [27].
All donors were ≥19.5 years old and their assessment included clinical assessment, psychological assessment, lab studies (liver function tests (LFT), virology, etc), abdominal ultrasound (US), computed tomography (CT angiography and CT volumetric studies), magnetic resonance cholangeopancreatography(MRCP), liver biopsy, etc. In late cases; we did CT with hepatic protocol and 3d imaging reconstruction for determining liver graft volume and vascular variations, moreover; we did our best to avoid cases with an estimated graft recipient weight ratio (GRWR) less than 0.8, as well as GRWR>4 to avoid small for size graft (SFSG), and LFSG respectively for being away from their bad sequels Figure  The study parameters were collected from a prospectively maintained database in our LT unit and were analyzed retrospectively. Those parameters included pediatric recipients' pre-and intra-operative variables, their donors' variables, primary liver diseases, and postoperative measures.
The details of donors' and recipients' surgical techniques including recipients' vascular and biliary reconstructions have been described previously [28][29][30]. In short; in the donor surgery; the graft type was chosen concerning the estimated GRWR, and the ratio of the graft volume to the recipient's standard liver volume (GV/SLV), furthermore, the hepatectomy was done using Cavitron ultrasonic surgical aspirator (CUSA) device. The donor biliary anatomy was determined according to both the pre-operative MRCP, and the intra-operative cholangiography (IOC), while the vascular anatomy depended upon pre-operative CT angiography ± intra-operative Doppler US. On the other hand, in the recipient surgery, the total hepatectomy phase was done with meticulous dissection and good hemostasis especially in cases with PHN to decrease blood loss, moreover; the hilar portal structures dissection was performed near the liver for obtaining the maximum length of those structures for better future reconstruction, also, the inferior vena cava (IVC) was carefully preserved with temporary portocaval shunts in some cases. On the other hand; on the back table, Hydroxyl tryptophan ketoglutarate solution was used for graft preservation with vascular manipulations of its hepatic veins(HV)/portal veins(PV) in some cases;  3. Then in the implantation phase; HV and PV anastomoses were performed with the aid of surgical loupes using continuous 5/0 and 6/ 0 prolene sutures respectively; Fig. 4: A, B; moreover, PV anastomosis was done with a growth factor. Then, the hepatic artery (HA) anastomosis was achieved with the help of surgical loupes or microscopy using interrupted 8/0 prolene stitches; Fig. 4:C. The biliary anastomoses were done with the aid of surgical loupes using interrupted 6-0 prolene/ Polydioxanone(PDS) stitches; Fig. 5. Doppler US was done routinely after vascular reconstruction and after abdominal closure to determine the pattern and velocity of blood flow. Finally, all our recipients' abdomens were closed primarily without the occurrence of any abdominal compartment syndrome (ACS). The post-operative measures have been described previously [28][29][30]. In brief; they included: 1-Immunosuppression therapy and protocol; it consisted of tacrolimus(FK506) and prednisolone, however, some cases were given cyclosporine when side effects (i.e. neurotoxicity or nephrotoxicity) developed with tacrolimus. Mycophenolate mofetil (MMF) was given for multiple episodes of acute rejection, chronic rejection, and for decreasing tacrolimus dose to prevent or treat renal impairment. On the other hand, sirolimus and/or everolimus were given to some patients to replace tacrolimus if side effects developed and to treat chronic rejection. Lastly, an interleukin-2 receptor blocker was given in late cases at POD 0 and 4 for minimizing the tacrolimus dose.

List of abbreviations
2-To prevent infection; antibacterial (pre-operative 3rd generation cephalosporine, then intra-/post-operative Imepanem + metronidazole until culture result), antifungal (fluconazole), and antiviral (acyclovir) were given. 3-For prophylaxis of vascular thromboses; Heparin infusion was given (dose; 180-200units/kg/day) adjusted according to activated partial thromboplastin time (APTT) (target levels; 50-70 s), then acetylsalicylate and dipyridamole were given at POD8 at doses of 2 mg/ kg/d and 4 mg/kg/d respectively for 3 months. 4-The follow-up of pediatric recipients (i.e. by transplant surgeons, pediatric hepatologists, pediatric endoscopists, and pediatric intervention radiologists) was done daily until hospital discharge, then weekly until the end of the 1st 3 months then monthly until the end of the 1st year, then yearly until the end of the follow-up period to detect: a-Early (short-term; less than 3months), and late (Long-term; ≥3months) morbidities(i.e. infection, pulmonary, vascular, biliary, renal, rejection, etc); they were graded according to Clavien grading. b-Early (less than 3months), and late (≥3months) mortalities, as well as mortality causes. c-Graft and patient survival outcomes.
The statistical analysis was done by SPSS 21 software (SSPS Inc, Chicago, IL, USA). Nominal variables were expressed in frequencies and percentages and analyzed using Fisher exact or Chi-square tests. Continuous variables were expressed as medians (ranges) or means±SDs and were compared using the t-or Mann-Whitney U tests. Univariate and then multivariate analyses were performed to detect predictors of early and/or late morbidities as well as predictors of patients' survival. The Kaplan-Meier method was applied for analysis of the survival of recipients and was compared using log-rank tests. In all tests, a P-value of <0.05 was significant.
Early pulmonary complications affected 13(19.4%) of our recipients and were categorized into chest infection (11(16.4%)), pulmonary  embolism (1(1.5%)) and hemothorax(1(1.5%)). As regards Clavien's grading; grades II, III and V affected 1, 1 and 11 of them respectively. They were managed as follows: Antibiotics for the chest infection, anticoagulants for pulmonary embolism and a chest tube for hemothorax with successful treatment in 2 of them only. Table 4.
Early acute rejection affected 10(14.9%) of our paediatrics. Clavien grades II and V involved 9 and 1 of them respectively. They were managed by pulse steroids where 4, 3, 1 and 2 of patients were given 1, 2, 3 and 4 boluses respectively with good outcomes in 9 of them. Table 4.
The incidence of early biliary complications was 7(10.5%) in the form of biliary leak + biloma, biliary leak and cholangitis in 2(3%), 3 (4.5%) and 2(3%) of patients respectively. Regarding Clavien grading;   grades II, III and V involved 2, 2 and 3 of them respectively. The cases with biliary leak ± biloma were managed conservatively, by percutaneous drainage, endoscopic retrograde cholangeopancreatography (ERCP) and/or surgery (Open drainage, and/or external biliary diversion) under antibiotic coverage; however cholangitis cases were managed by antibiotics with a good result in 4 of the 7 patients. Table 4. Early wound complications affected 7(10.5%) of patients in the form of wound infection and wound infection + burst abdomen in 6(9%) and 1(1.5%) of patients respectively. They were managed medically by antibiotics for the infection or surgically for burst abdomen. They were in the category of II and III regarding Clavien grades with excellent results in all of them. Table 4. Five (7.5%) of our patients had early renal impairment where Clavien grade II and V affected 3 and 2 of them respectively; they were managed by renal supportive treatment with good outcomes in 3 of them. Table 4.
Early gastrointestinal (GIT) complications in the form of haematemesis, colonic perforation and hepatic encephalopathy affected 2 (3%), 1(1.5%) and 1(1.5%) of our transplanted children respectively. They were managed by endoscopy, surgically and by medical treatment respectively with a successful outcome in 3 of them. Table 4.
Finally; early neurological complications affected 1(1.5%) of our patients who underwent neurological support with a good outcome, lastly, we had a case with recurrent Budd Chiari syndrome (BCS) (IVC stenosis) and another case with early graft failure, they were managed by angiographic dilatation, and liver support respectively but unfortunately; both cases died. Table 4.
As regards late morbidities; they affected 12(17.9%) of our patients. Late bacterial infections involved 4(6%) of our smart recipients; those infections were classified into chest, and biliary infections that affected 3 (4.5%) and 1(1.5%) of patients respectively, furthermore, they were sorted regarding Clavien grading into grades II, III and V in 1, 1 and 2 of them respectively. They were managed by antibiotics; moreover, the cholangitis case was managed surgically. The outcome was successful in 2 of the 4 cases. Table 5.
The incidence of chronic rejection was 4(6%) that occurred in the 11th, 12th, 16th and 18th post-transplant months. It was diagnosed histologically according to updated Banff criteria [31]. Patients were given MMF beside FK for its management, furthermore, when FK toxicity occurred they were shifted to Sirolimus or Everolimus, however, the 4 patients, unfortunately, died (Clavien grade V). Table 5.
In our series, the late pulmonary complications involved 4(6%) of recipients, they were divided into chest infections (3(4.5%) and pleural effusions (1(1.5%), moreover, they were categorized into Clavien grades II, III and V in 1, 1 and 2 of them respectively; they were managed by antibiotics for infection and chest tube for effusion with 2 mortalities; one from sepsis and the other from acute respiratory distress syndrome (ARDS). Table 5.  Four (6%) of our pediatric patients had late biliary complications in the form of HJ stricture + cholangitis, HJ stricture and D-D stricture that involved 1(1.5%), 1(1.5%) and 2(3%) of them respectively; they were all Clavien grade III as they were managed by ERCP, percutaneous transhepatic drainage (PTD) and/or surgical reconstruction under antibiotic coverage with final improvement in all of them. Table 5.
As regards late vascular complications; they were 3 cases (4.5%) and were classified into HAT and HV stenosis in 1(1.5%) and 2(3%) of them respectively. Clavien grades II, III and V involved 1, 1 and 1 of them; they were managed by anticoagulants and fibrinolytic for HAT as well as by angiographic dilatation and stenting for HV stenosis with successful results in 2 of the 3 cases. Table 5.
Lastly, late acute rejection, renal impairment and recurrent BCS affected 2(3%), 2(3%) and 1(1.5%) of patients respectively; they were managed by pulse steroids for rejection, renal supportive treatment for renal impairment and angiographic dilatation and stenting for BCS with a successful outcome in 2, 1 and 1 of them respectively. Table 5.

Predictors of early and/or late morbidity
On univariate analysis, CTP class C, higher PELD/MELD scores, biliary stents, more intra-operative packed RBCs transfusion units and longer duration of operation were predictors of early and/or late morbidities, however, on multivariate analysis, there was no independent predictor of those morbidities. Table 6. The early (less than 3months) mortality reached 34.3% mostly due to sepsis, renal impairment and LFSGs; however, the late (≥3months) mortality was 13.4% mostly from chronic rejection and late sepsis. Table 7; Fig. 6.

Pre-and intra-operative parameters as predictors of patient survival outcome
On univariate analysis, CTP class C, Pre LT PHN, biliary stents, more   intraoperative Packed RBCs transfusion units and longer duration of operation were predictors of poor patient survival, however, on multivariate analysis, GRWR≥3 had a trend towards independent correlation with patient mortality; moreover, more units of transfused Packed RBCs had an independent association with patient mortality. Table 8; Fig. 7.

Early and/or late morbidities as predictors of patient survival outcomes
On univariate analysis, the overall early and/or late morbidities, bacterial infections, pulmonary complications, acute rejection, chronic rejection, vascular complications, and renal impairment were significant risks of patient mortality, on the other hand, on multivariate analysis, bacterial infections and pulmonary complications were independent predictors of poor patient outcome. Table 9; Fig. 7.

Discussion
Despite being a challenging procedure; pediatric LDLT is a life-saving option for paediatrics with ESLD, and other catastrophic liver conditions like tumours especially in countries like Egypt that don't have a deceased donor liver transplantation (DDLT); however, the complications after it are still a big problem with devastating effects despite the recent improvement in such a field of pediatric LDLT [32].   Our early and/or late morbidities that affected 41(61.2%) of our patients lie within the literature range of post pediatric LDLT morbidities (15%-89.9%) [2,8,[33][34][35].
Despite advanced infection control policies and antibacterial prophylaxis; bacterial infections that cause remarkable morbidities and mortalities in the early and late periods after pediatric LT are still common due to poor patient general condition, being ultra major operation, using immunosuppressants, etc [16,39]. They ranged from 39.1% to 67% in the previous literature [39,42,and47]], however, they were less in our series (34.3%); and this is due to our improved infection control policies, especially in our later cases. Higher PELD/MELD scores were associated with morbidities in our work, in the same line; they were significant or independent predictors of morbidities in Kitajima et al., 2017 [2], Raices, et al., 2019 [8], and Chung et al., 2020 [48] studies, moreover, they were independent predictors of early re-laparotomy due to morbidities in Okada et al., 2019 [49] study.
Longer operative time was a significant predictor of morbidity in our present work, also, it was associated with re-laparotomies due to morbidities in Yoeli et al., 2018 [6] and Okada et al., 2019 [49] studies.
Increased amounts of intra-operative packed RBCs transfusion was a predictor of morbidity in the present series, in similar, it was correlated with relaparotomy from morbidities in Yoeli et al., 2018 [6] work.
Higher PELD/MELD scores had a trend towards significant correlation with poor patient survival in our work, also, a higher PELD score was an independent predictor of poor patient survival in Pan et al., 2020 [22], Lu, et al., 2020 [25] and Kehar et al., 2019 [32] studies, moreover, it was a significant predictor of patient loss in Oh et al., 2010 [42]; study, in contrast, it did not affect survival in Kitajima

studies.
Post LDLT bleeding that comes from technical issues, collaterals, and bleeding tendency has a negative insult on survival outcomes [7]. Similarly, increased intraoperative blood loss was associated with patients' mortalities in Pan et al., 2020 [22] and Lu et al., 2020 [25] studies. Furthermore, More RBC units' transfusion was correlated with patient mortality in our and Boillot et al., 2021 [24] studies. Conversely; increased intra-operative blood loss was not associated with patient survival in Shehata et al., 2012 [58] study.
Longer operative time was a significant predictor of patient mortality in our study, also, it was an independent predictor of poor patient survival in Pan et al., 2020 [22]  The overall post-transplant complications were associated with patient mortality in the recent study, also, complications were significant predictors of mortality in Ho et al., 2004 [7] study, and early relaparotomy from early morbidities was significantly associated with poor patient survival in Okada et al., 2019 [49] study.
LFSG affects post LT outcomes by reducing oxygen and blood supplies of the liver graft, increasing vascular complications rate, inducing allograft dysfunction and/or loss and/or necrosis, inducing renal dysfunctions, as well as the occurrence of ACS and large for size syndrome [8,23]. In a similar line in our work; The 2 cases with LFSGs died from their sequels, and GRWR≥3 had a trend towards independent correlation with patient mortality, also; LFSG was an independent predictor of poor patient survival in Lu  Post pediatric LDLT vascular complications are common causes of morbidity and mortality [39]. Also, they were predictors of patients' mortality in ours, Steinbrück, et al., 2011 [61] and Sieders et al., 2000 [62] studies. But they did not affect survival in Shehata et al., 2012 [58] study.
Post LT acute rejection is a known cause of graft dysfunction [63]. It was correlated with patient loss in our series; also, it was a major cause of death in Kitajima et al., 2018 [34] study. However, it did not affect graft or patient survival in Yilmaz et al., 2006 [40] or Shehata et al., 2012 [58] studies.
Despite advanced immunosuppression after LT; chronic rejection remains a major reason for graft and/or patient loss [42,43,59,64]. Also, in our work; it was the major cause of late mortality and a significant predictor of patient loss, similarly, it was an independent predictor of patient loss in Oh et al., 2010 [42] study, In contrast, it did not affect graft or patient survival in Yilmaz et al., 2006 [40] study.
We found a significant correlation between post LT renal impairment and patient mortality, similarly; Post LT hemodialysis was an independent predictor of poor patient survival in Boillot et al., 2021 [24] study. Regarding pulmonary complications; they were associated with patient mortality in Alam et al., 2017 [45] study. Also, they were independently associated with mortality in our work.
Bacterial infection was associated with patient mortality in Pouladfar et al., 2019 [47] and Shepherd et al., 2008 [65] studies. Also, it was an independent predictor of patient loss in our study and in a similar line; sepsis was the major cause of early mortality, and the 2nd most common cause of late mortality in our study, similarly; it was the major cause of mortality in Kitajima et al., 2017 [2], Kehar, et al., 2019 [32], Kitajima, et al., 2018 [34], Mohan, et al., 2017 [43] and Tanaka et al., 2010 [66] pediatric LDLT studies.
Lastly, to our knowledge; this is one of the unique pediatric LDLT studies mentioning the independent association between both pulmonary complications and bacterial infection and patient mortality, and this is due to sepsis which has led to those catastrophic mortalities. In conclusion; more packed RBCs transfusion units intra-operatively, and post LT bacterial infection, sepsis, chronic rejection, as well as pulmonary complications had a negative insult on our patients' outcomes, so proper management of them is mandatory for improving outcomes after pediatric LDLT.

Ethical approval
The approval by National liver institute (IRB), Menoufia University that was done retrospectively.

Sources of funding
No source of funding for this research.

Author contributions
Emad Hamdy Gad: Surgical procedures, study design, data collection, writing, analysis and publication.
Ahmed Nabil Sallam: Surgical procedures, data collection, and analysis.
Tarek Ibrahim: Surgical procedures, and analysis. Tahany Abdel Hameed Salem: Study design, data collection, and writing.
Mohammed Abdel-Hafez Ali: Study design, data collection, and writing.
Mohammed abdelsamee: Data collection and writing.Islam Ayoub: Study design, surgical procedures, data collection and analysis.

Guarantor
All the authors of this paper accept full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.

Provenance and peer review
Not commissioned, externally peer-reviewed.

Declaration of competing interest
No conflict of interest to declare.

Acknowledgement: We thank our LT team for their hard work
The main limitation of the study is the small NO of patients and being a retrospective one as well as the comparisons of our results were done with previous heterogonous literature studies (prospective or retrospective, small or large volumes, short or long duration studies) despite being pediatric living donor liver transplantation studies; so, we recommend doing further large prospective studies of pediatric LDLT and making the comparison with similar large prospective pediatric LDLT studies.

Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.amsu.2022.103938. E.H. Gad et al.