Donor size and early vascular complications in children receiving a whole liver transplant

Abstract

BACKGROUND

Early vascular complications are one of the leading causes of morbidity and graft loss in paediatric whole liver transplant (WLT). It has been previously suggested that the donor size might influence these outcomes, but evidence remains insufficiently defined.

AIM

To evaluate the influence of donor weight on the incidence of early vascular complications after WLT in paediatric recipients.

METHODS

Single-centre, retrospective cohort study on patients under 18 years old, who received a first WLT between 2000 and 2024. Recipients were classified according to the donor’s weight: Subgroups A (≤ 15 kg) and B (> 15 kg). Demographic, anthropometric, surgical and perioperative variables were analysed, focusing on early vascular complications - hepatic artery thrombosis (HAT), portal vein thrombosis (PVT) - and the need for re-transplantation within 30 days postoperatively.

RESULTS

Sixty-four patients met the inclusion criteria. Subgroup A (n = 12) showed higher rates of early complications compared with subgroup B (n = 52): HAT 25% vs 5.8% [relative risks (RR): 4.3, 95%CI: 1.0-18.9, P = 0.074], PVT 16.7% vs 1.9% (RR: 8.7, 95%CI: 0.9-87.9, P = 0.088), and re-transplantation 16.7% vs 0% (RR: 5.4, 95%CI: 0.9-32.0, P = 0.033). Donor and recipient anthropometrics were highly correlated, precluding assessment of donor weight as an independent risk factor.

CONCLUSION

In paediatric WLT, grafts from small donors appear more susceptible to early vascular complications, reflecting the technical and anatomical vulnerability of small-calibre vessels rather than donor weight alone. Awareness of this relationship may assist in graft selection, intraoperative planning and postoperative surveillance.

Key Words: Whole liver transplant; Hepatic artery thrombosis; Portal vein thrombosis; Re-transplantation

Core Tip: This single-centre study evaluates the impact of donor size on early vascular complications after paediatric whole liver transplantation. Grafts from donors ≤ 15 kg showed higher rates of early vascular complications and retransplantation. Although donor weight itself was not an independent risk factor, it reflects the anatomical and technical challenges of small-calibre vessels. Recognising this association may help optimise graft selection, surgical planning, and postoperative monitoring to improve outcomes in small paediatric recipients.

INTRODUCTION

Liver transplantation is an effective therapeutic intervention for children suffering from end-stage liver disease secondary to chronic conditions, such as biliary atresia, intrahepatic cholestasis, metabolic and unresectable neoplastic disorders, amongst others, as well as irreversible cases of acute liver failure[1]. The paediatric population undergoing liver transplantation is heterogeneous, with a clear division between infants and adolescents. According to the European Liver Transplant Registry (2022)[1], biliary atresia is the primary indication for liver transplantation in children aged < 12 years (53.4%), with more than 25.4% of these interventions occurring in children aged < 1 year. Over time, several technical approaches have been developed to mitigate the shortage of livers for these small patients, such as the use of partial grafts from either living or deceased donors. Diversification has become essential to expand access to organs among smaller patients, as the number of young donors (< 18 years) has decreased from 59% to 24% after 2010, and the percentage of children receiving size-matched whole liver transplants (WLT) has moved from 71.3% to 26.7% over the past three decades[1]. However, despite being a scarce but also valuable source of grafts, the risk of vascular complications associated with the use of small whole livers has been described to be higher, primarily due to the vessel diameter and the complexity of vascular reconstructions[2,3].

The present study aims to evaluate the influence of donor weight on the incidence of early vascular complications after WLT in paediatric recipients, as we believe it is an important factor to consider during graft allocation. We hypothesise that grafts from smaller donors have a higher risk of vascular complications compared to those from bigger donors.

MATERIALS AND METHODS

We present a retrospective cohort study of all patients under 18 years of age who underwent a WLT at our centre between 2000 and 2024 (n = 134). Patients undergoing a multi-visceral transplantation, re-transplantation, or those with incomplete clinical records due to historical data limitations (n = 70) were excluded. Recipients were classified into two subgroups according to the donor’s weight: A (≤ 15 kg) and B (> 15 kg).

The donor-weight cut-off of 15 kg was determined a priori, based on prior paediatric series describing a higher incidence of early vascular complications when donor weights fall within the 6-15 kg range[4-7]. This threshold also reflects our centre’s real-world allocation practice, where grafts below 15 kg represent the natural cluster of “small donor” offers. Importantly, this categorisation was intended for pragmatic comparison, not as a biological inflexion point. The analysis was exploratory, recognising that donor and recipient anthropometrics are inherently correlated for allocation purposes, and therefore, the effect of donor weight cannot be fully isolated from recipient size in this dataset.

The following variables were analysed: Recipient and donor’s gender, age and weight, donor-recipient weight ratio, indication for transplantation, elective or super-urgent status, time on the waiting list, type of donor (donation after brain death vs donation after circulatory death) and arterial anatomy, ischaemia times, surgical time and surgical techniques, time of follow-up and patient survival.

Cold ischaemia time was defined as the time between when cold perfusion started and liver out of ice after static cold storage preservation; warm ischaemia time was defined as the time between taking the liver out of ice and portal vein reperfusion, and surgical time was defined as the time between skin incision and the end of surgery. Regarding surgical technique, a microvascular approach was employed with non-absorbable 6-0 to 8-0 polypropylene sutures for both the portal vein and arterial anastomoses. Portal vein reconstruction´s choice was operator-dependent; end-to-end anastomosis was the preferred technique – either using a continuous suture in the posterior wall and interrupted in the anterior wall, or two continuous sutures incorporating a growth factor. Interpositional grafts were reserved for marked donor-recipient size discrepancies or severe portal vein hypoplasia. For the hepatic artery, interrupted end-to-end anastomosis is the standard technique in our institution, with primary aortic conduits reserved for selected metabolic patients or specific re-transplantation scenarios.

Focusing on our aim, we recorded only early vascular complications, including hepatic artery thrombosis (HAT) and portal vein thrombosis (PVT), within the first 30 postoperative days, along with their management (conservative, surgical, or re-transplantation). Our institutional post-liver transplant protocol specifies six-hourly blood tests during the first 48 hours (including full blood count, liver function tests - alanine transaminase, aspartate transaminase, gamma-glutamyl transferase, bilirubin -, clotting screening - PT, aPTT and fibrinogen - and blood gases), followed by eight-hourly blood tests during intensive care unit stay, and daily Doppler surveillance for five days. To prevent vascular complications, the protocol targets a central venous pressure of 8-10 mmHg, neutral fluid balance (initially achieved with 80% of maintenance fluids) and the early initiation of anticoagulation therapy (continuous heparin infusion starting at 2.5 IU/kg/hour and increased to 5 IU/kg/hour on postoperative day 1, then titrated according to PT). Early vascular complications are diagnosed through a combination of biochemical markers (a sudden rise in transaminases after the initial post-transplant peak) and Doppler findings (direct visualisation of thrombus and/or alteration of flow velocity or Doppler signal in the hepatic artery or the main portal vein). If the diagnosis for HAT or PVT is unclear, it is confirmed with computed tomography (CT) angiography. Anticoagulation therapies, postoperative Doppler surveillance, and the use of CT confirmation have remained unchanged since 2000. Although Doppler imaging was initially performed exclusively by consultant radiologists, from 2015, this practice transitioned to trained sonographers working under consultant supervision.

Qualitative variables were expressed as absolute frequencies and percentages, with comparisons performed using Fisher’s exact test. Relative risks (RR) and odds ratios with 95% confidence intervals were calculated to quantify the strength of associations. Quantitative variables were described as median and interquartile range, and analysed using the U-Mann-Whitney test. Statistical significance was defined as a Fisher’s P < 0.05. All analyses were conducted using the SPSS software package (IBM^® version 25).

RESULTS

During the study period, a total of 64 patients met the inclusion criteria [subgroup A (n = 12, 18.75%) and subgroup B (n = 52, 81.25%)]. Both subgroups were comparable in terms of gender, elective vs super-urgent, time on the waiting list, and type of donors. All donors’ and recipients’ demographics are shown in Table 1. In keeping with expectations, donor and recipient sizes were tightly coupled, with minimal overlap between groups. Median donor-recipient-weight-ratio showed no significant difference between groups.

Table 1 Demographics of recipients and donors.

n = 64		Subgroup A (n = 12)	Subgroup B (n = 52)	P value
Recipients	Gender: M/F (%)	6/6 (50/50)	32/20 (61.5/38.5)	0.525
	Age (years)	1.4 (0.89-2.15)	13.31 (10.34-15.5)	0.000
	Weight (kg)	9.45 (8.67-11.65)	45.5 (30.08-58.5)	0.000
	Elective/SU (%)	11/1 (91.7/8.3)	47/8 (90.4/15.4)	1
	Days waiting list	40.5 (15.5-169.25)	72.5 (9-151.25)	0.750
Donors	Type: DBD/DCD (%)	12/0 (100/0)	44/8 (84.6/15.4)	0.334
	Gender: M/F (%)	7/5 (58.3/41.7)	23/29 (44.2/55.8)	0.523
	Age (years)	2.09 (1.31-2.94)	20.22 (9.61-32.98)	0.000
	Weight (kg)	12 (11.25-14.75)	58.75 (30.75-69.5)	0.000
	DRWR	1.21 (1.1-1.5)	1.16 (0.89-1.46)	0.636

DRWR: Donor-recipient-weight-ratio; M/F: Male/female; SU: Super-urgent; DBD: Donation after brain death; DCD: Donation after circulatory death.

Biliary atresia was the main indication for transplantation in subgroup A (33.3%), whereas indications in subgroup B were more heterogeneous (Figure 1). Regarding ischaemia times, surgical time and surgical techniques, we observed no statistically significant differences (Table 2). In subgroup A, no aortic conduits or interpositional portal vein grafts were used. In subgroup B there were two (3.8%) primary aortic conduits (one was a metabolic patient and the other one an acute liver failure. None of them presented with post-transplant HAT) and three (5.8%) portal vein interpositional grafts (one was biliary atresia with severe portal vein hypoplasia; other one presented with grade III PVT and the third had a cavernoma with superior mesenteric vein and splenic vein thrombosis. None of them presented with post-transplant PVT).

Figure 1 Indications for liver transplantation. Subgroup A (orange line), subgroup B (blue line). PSC: Primary sclerosing cholangitis.

Table 2 Timings and surgical techniques, n (%).

n = 64	Subgroup A (n = 12)	Subgroup B (n = 52)	P value
CIT (minutes)	478 (336-688)	439 (378-544)	0.335
WIT (minutes)	42 (27-49)	30 (26-40)	0.139
Surgical time (minutes)	300 (264-385)	300 (264-371)	0.969
IVC technique	Piggy-back = 5 (41.7); end-end = 7 (58.3)	Piggy-back = 31 (59.6); end-end = 17 (32.7); side-side = 4 (7.7)	0.311
Portal vein technique	End-end = 12 (100)	End-end = 49 (94.2); graft = 3 (5.8)	1
Hepatic artery technique	End-end = 12 (100); aortic conduit = 0	End-end = 50 (96.2); aortic conduit = 2 (3.8)	1

CIT: Cold ischaemia time; WIT: Warm ischaemia time.

The rates of HAT, PVT, and need for re-transplantation in the first 30 days after transplant are summarised in Table 3. In subgroup A, 33% (n = 4) of patients developed vascular complications. Two patients presented with HAT alone, and both of them underwent super-urgent re-transplantation. One of these patients was initially managed with thrombectomy and re-anastomosis, but presented again with intra- and extrahepatic arterial thrombosis, requiring eventual re-transplantation. In the same subgroup A, one patient developed PVT alone, necessitating thrombectomy and re-anastomosis. The last patient of this subgroup had simultaneous HAT and PVT, which was successfully managed with thrombectomy and re-anastomosis, without graft loss. In subgroup B, 8% (n = 4) developed vascular complications. Three patients developed HAT, all of whom were successfully managed surgically, two with thrombectomy and re-anastomosis, and one with the creation of an infrarenal aortic conduit. One patient developed a non-occlusive PVT and was managed conservatively with anticoagulation therapy alone. The median follow-up time was 10.7 (1.8-17.6) years in subgroup A, and 2.5 (1.5-5.7) years in subgroup B. In the whole cohort, there were seven deaths unrelated to early vascular complications.

Table 3 Early vascular complications and the need for re-transplantation, n (%).

n = 64	Subgroup A (n = 12)	Subgroup B (n = 52)	Risk ratio (95%CI)	Odds ratio (95%CI)	Fisher P value
HAT	3 (25)	3 (5.8)	4.3 (1.0-18.9)	5.4 (0.9-32.0)	0.074
PVT	2 (16.7)	1 (1.9)	8.7 (0.9-87.9)	10.2 (0.8-123.6)	0.088
Re-Tx	2 (16.7)	0	-	1.2 (0.9-1.6)	0.033

Re-Tx: Re-transplantation; HAT: Hepatic artery thrombosis; PVT: Portal vein thrombosis.

DISCUSSION

The persistent shortage of grafts and the difficulty of finding suitable size-matched grafts for the younger patients remain among the greatest challenges in paediatric liver transplantation. According to the European Liver Transplant Registry, more than 25% of liver recipients are less than one year of age[1]; similarly, data from the SPLIT registry (United States and Canada) indicate that over 30% of recipients are under one year old, and 37% weigh less than 10 kg[8]. This vulnerable population faces a higher mortality risk due to prolonged waiting times and clinical deterioration. In the United States, more than 1/10 children under two years old die while waiting for a liver transplant, and among those who reach transplantation, more than 40% spend over a year on the waiting list[9]. To address these challenges, the use of technical variants has become essential to expand the donor pool for small paediatric recipients. Over the last 15 years, the percentage of grafts from living donors has shifted from 7% (pre-2000) to 39%, while the percentage of reduced grafts from deceased donors has shifted from 22% (pre-2000) to 33%[1]. Although living-donor liver transplantation provides superior patient and graft survival outcomes compared to deceased donation[10], many centres rely exclusively on the latter due to limited resources[11]. Even though reduced grafts from deceased donors can shorten waiting list times[12,13], there is a significant variation in their utilisation among different countries, reflecting different allocation policies and split criteria[9,14]. In Europe, around 40% of paediatric transplants are performed with reduced grafts from deceased donors[14], whereas in the United States, more than 90% of potentially splitable grafts are allocated as whole livers to adult recipients due to logistical and technical barriers[9]. As a result, the use of WLT from small donors continues to play a crucial role in alleviating the organ shortage in children.

WLT from small donors offers advantages in graft integrity and avoidance of biliary cut-surface complications[15]. Nevertheless, evidence regarding the risk of early vascular complications - specifically HAT and PVT - in WLT in children remains inconclusive. Rokop et al[16] described a higher incidence of early (within the first 90 days) HAT in recipients receiving WLT compared to those receiving left lateral segment from deceased donors, irrespective of the recipient’s age or weight. However, bigger paediatric cohorts like the SPLIT registry, performed by McElroy et al[17] in 2022, demonstrated a higher incidence of early HAT in WLT compared with left lateral segment from both living donors or deceased donors (8% vs 4% vs 5%, respectively; P = 0.0085), with a trend towards lower incidence in older recipients, suggesting that smaller arterial calibre in infant grafts predisposes to thrombosis. On the contrary, PVT was more frequent in reduced grafts (5% vs 2.6%, P = 0.021), probably due to donor-recipient venous calibre mismatch. Despite the differences, rates of reoperation and three-year graft or patient survival did not differ significantly. A meta-analysis by Ye et al[18] revealed significantly lower PVT rates in WLT compared to reduced grafts, but no significant difference in HAT incidence. Although the study pooled short- and long-term outcomes and did not stratify by donor or recipient weight.

Irrespective of graft type, early vascular complications remain a major cause of graft loss, long-term graft dysfunction, and patient mortality[2]. Several studies have focused on the recipient as a key determinant of risk for complications and graft loss[7,19,20]. Based on our institutional experience, we hypothesised that donor anthropometrics might be equally relevant. Previous authors have explored similar hypotheses, although the evidence remains inconclusive[5]. In our cohort, WLT from donors ≤ 15 kg demonstrated a trend towards higher incidences of early HAT and PVT, as well as increased graft loss, with a four-fold increase in relative risk for HAT (RR: 4.33; 95%CI: 1.0-18.5), and a nine-fold increase for PVT 8.7 (0.9-87.9). Re-transplantation was observed only in donors ≤ 15 kg (16.7%), precluding calculation of a risk ratio. For HAT and PVT, differences between subgroups did not reach statistical significance, likely reflecting the small number of events. Re-transplantation, however, was significantly more frequent in recipients of ≤ 15 kg donors (P = 0.033). Nevertheless, the point estimates for relative risk consistently suggest a clinically meaningful trend toward higher early post-transplant vascular complications in recipients of smaller grafts. These findings are consistent with the UNOS analysis by Desai et al[7], which reported an increased HAT risk with donor weight below 10 kg, peaking in grafts from donors < 7.8 kg. Interestingly, the donor-recipient-weight-ratio did not achieve statistical significance in this study or in ours, suggesting that the observed early post-transplant complications cannot be attributed to donor-recipient weight mismatch, and the association between donor size and vascular complications should be interpreted as a size-matching phenomenon rather than an independent donor-weight effect. Other publications differ from our results; Li et al[4] advocate the good outcomes of low-weight grafts. Although no statistically significant differences were observed in the incidence of HAT or PVT between grafts weighing < 150 g and those 150-400 g, the overall rate in donors weighing < 15 kg was 13.5%, highlighting the clinically relevant risk associated with small donor grafts.

It is also worth mentioning that among paediatric liver transplant patients, those with metabolic conditions show a higher tendency of HAT compared to other risk groups[21,22]. In our cohort, three patients in subgroup A and 2 in subgroup B presented with metabolic conditions; all of them had normal graft anatomy, four had end-to-end arterial anastomosis, and one had an aortic conduit created; however, none of them presented with HAT. Besides, none of these explanted grafts were utilised for domino transplantation.

Finally, paediatric liver grafts are particularly vulnerable to ischaemia-reperfusion injury, clinically manifesting as post-reperfusion syndrome. Longer ischaemia times have been identified as significant predictors of endothelial injury, early vascular thrombosis and poor graft function[23-25]. Although machine perfusion technologies - such as hypothermic-oxygenated machine perfusion, normothermic regional perfusion, and normothermic ex-vivo machine perfusion - have been widely adopted in adult transplantation, emerging data suggest potential benefits in reducing biochemical injury and preserving microvascular integrity, particularly in small-calibre grafts[26-28]. Future studies will clarify whether these modalities can mitigate vascular risk in small-donor WLT.

We acknowledge the limitations of our study, particularly the inherently small sample size associated with paediatric liver transplantation. While this reflects the rarity of the procedure, it nevertheless warrants caution in the interpretation and external validity of our findings. We describe a single-centre retrospective experience based on institutional protocols. The inclusion of patients transplanted more than two decades ago introduces a potential era-related bias, as perioperative management and imaging modalities may have evolved during this period. However, surgical technique and anticoagulation protocols have remained stable, which we believe to be of the utmost importance in this cohort. To reduce the impact of incomplete postoperative information, cases lacking adequate follow-up were excluded from the analysis; however, this approach may itself have introduced selection bias by preferentially retaining patients with more complete surveillance or more morbid postoperative courses.

Higher incidence of early HAT and PVT is clinically intuitive given the smaller vessel calibre, greater technical complexity, and microvascular fragility characteristic of infant grafts. However, this observation does not establish donor weight as an independent causal factor. Smaller donors were almost exclusively transplanted into smaller and younger recipients, a pattern that mirrors real-world allocation rather than an experimental design. Consequently, the observed association most likely reflects the composite risk of the “small donor-small recipient” pair rather than donor weight alone. We deliberately did not perform case matching or propensity analysis owing to the small cohort and low event frequency, which would yield unstable estimates and spurious precision. The present analysis should therefore be regarded as hypothesis-generating, highlighting the need for heightened vigilance and meticulous surgical technique in extreme size-mismatch scenarios. The 15 kg cut-off, although arbitrary, corresponds to thresholds pragmatically used in prior literature to define “small donors” and represents a natural separation point in our dataset. More granular approaches-such as decile stratification or receiver operating characteristic derived thresholds-would require larger, multicentre cohorts to avoid statistical artefacts and refine the weight limits for acceptable WLT grafts in children.

CONCLUSION

WLT from small donors may be associated with an increased risk of early vascular complications and subsequent graft loss in paediatric recipients. While donor weight itself may not act as an independent risk factor, it serves as a proxy for technical and anatomical vulnerability within this unique population. Recognising this relationship can help refine graft allocation strategies and guide preventive intraoperative and postoperative measures to improve outcomes.

ACKNOWLEDGEMENTS

To all members of the Liver Unit in Birmingham Children’s Hospital for their dedication, and to all the patients and their families for trusting our work.