Systematic review and meta-analysis of the role of aorto-hepatic conduits in liver transplant: Known knowns and known unknowns

Abstract

BACKGROUND

Aorto-hepatic conduits (AHCs) are an effective revascularization method for liver allografts when the native hepatic artery is unusable. Various studies have confirmed that outcomes with AHCs are inferior to those with native hepatic artery inflow.

AIM

To investigate the published evidence on the outcomes according to different inflow site for AHCs.

METHODS

A systematic search was conducted for studies reporting on AHCs in liver transplantation over the last 10 years (January 2014 onwards). Two independent reviewers selected articles, assessed quality, and evaluated bias in the included systematic reviews. The methodological quality of the included studies was assessed using the Newcastle-Ottawa Scale. The protocol was registered with PROSPERO (CRD42024545810). Review was conducted using the Preferred Reporting Items for Systematic Review and Meta-Analysis statement standards.

RESULTS

Fourteen studies identified a total of 32486 deceased donor liver transplants, of which 1136 (3.5%) required AHCs. The most frequent indications for AHC use included poor arterial flow, intimal dissections, and hepatic artery thrombosis. Among all AHCs, 207 (18.2%) were supra-coeliac (SC) AHCs, 738 (65.0%) infra-renal (IR) AHCs, 25 (2.2%) iliac artery conduits, and 166 (14.6%) had unspecified origins. Pooled analysis revealed comparable demographic characteristics. The median follow-up duration ranged from 18 to 52 months. There were no significant differences in early occlusions of AHCs [odds ratio (OR) = 0.94 (0.48, 1.84); P = 0.86], late occlusions of AHCs [OR = 0.46 (0.16, 1.32); P = 0.15], early allograft dysfunction [OR = 0.82 (0.46, 1.47); P = 0.51], biliary complications [OR = 1.10 (0.69, 1.76); P = 0.68], post-transplant renal replacement therapy (RRT) requirement [OR = 1.12 (0.72, 1.72); P = 0.62], and major surgical complications (Clavien-Dindo > 3b) [OR = 1.06 (0.70, 1.61); P = 0.79]. The median duration for graft occlusion was approximately 142 days, ranging from 13 to 3313 days. One-year graft and patient survival rates for SC conduits were 77% to 81.1% and 80% to 85.1%, respectively. For IR conduits, one-year graft and patient survival rates were 66% to 79.1% and 73% to 88.3%, respectively. Five-year graft and patient survival rates for SC conduits were 53.9% to 67% and 67.8% to 74%, respectively. For IR conduits, five-year graft and patient survival rates were 50% to 56% and 56% to 64.9%, respectively.

CONCLUSION

Considering these findings, there is no significant difference in early and late outcomes between SC and IR AHCs, although there is a discernible tendency towards higher late occlusion rates in the IR group.

Key Words: Liver transplant; Aorto-hepatic conduits; aortic conduits; hepatic artery; thrombosis

Core Tip: This meta-analysis of fourteen studies identified aorta-hepatic conduits utilized in 3.5% of deceased donor liver transplantation, with the majority being infra-renal conduits, however, 5-year patency was better with supra-coeliac aortic conduits.

INTRODUCTION

Effective vascularization of the liver allograft is crucial for the success of a liver transplant (LT). Hepatic artery-related complications, such as poor flow, dissections, thrombosis, and stenosis, are disastrous complications that can lead to graft failure and the need for retransplantation[1-5]. In some instances, it may not be possible to anastomose the native recipient hepatic artery to the donor hepatic artery. Alternatively, the splenic artery, gastroduodenal artery, and coeliac artery have been described as inflow sources[6-10]. However, these vessels often share a common origin from the aorta, which might also be affected by the same pathology. In such scenarios, aorto-hepatic conduits (AHC) can serve as a salvage technique.

AHC inflow sites are typically the supra-coeliac (SC) aorta, infra-renal (IR) aorta, or, in rare cases, the iliac arteries. Most studies report no difference in short- and long-term occlusion rates based on inflow sites[11,12]. However, each site has distinct advantages and disadvantages. SC inflow closely mimics natural anatomical inflow and requires a shorter conduit. However, accessing the SC aorta can be challenging in patients with portal hypertension and large varices. Additionally, aortic clamping may theoretically cause renal insufficiency and spinal neurological effects[11]. In contrast, the IR aorta provides easier access, although it can impact the renal arteries. This region of the aorta may also be prone to severe atherosclerosis, posing significant risks during or after clamping, including dissection, uncontrolled bleeding, and risk of graft or limb ischaemia[13]. Previous abdominal surgeries may complicate access to the IR aorta. The reversal of blood flow in the IR conduit is highly influenced by systemic hemodynamic changes, and over time, this may result in graft ischemia and thrombosis[12,13].

Different materials have been utilized in constructing AHCs depending on the setting and availability. The donor iliac artery is the main conduit material in the deceased donor setting, typically retrieved at the end of the procedure[14]. However, this artery can be compromised by atherosclerosis and aneurysmal dilatation, making it unsuitable as a conduit. In such cases, donor superior mesenteric artery, carotid arteries, or axillary arteries have been used as alternatives. In the living donor setting, the donor external jugular or greater saphenous vein (GSV) has been commonly utilized as a conduit[15-19].

While some aspects of their efficacy and safety are well-documented, uncertainties remain regarding their optimal utilization and impact on patient short- and long-term outcomes. This systematic review and meta-analysis aim to elucidate the established facts and unsolved questions surrounding the role of AHCs in LT. Specifically, it seeks to consolidate existing evidence on the selection of inflow sites (SC vs IR), post-operative complications related to conduit, and overall graft/patient survival.

MATERIALS AND METHODS

Study design

The study design of systematic review and meta-analysis was chosen to define the role of published evidence on different AHCs. The study followed the preferred reporting items for systematic review and meta-analysis (PRISMA) statement standards[20,21]. Our review was registered at the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD42024545810).

Search strategy

A search strategy in line with the Meta-analysis Of Observational Studies in Epidemiology guidelines and previous recommendations for the conduction of systematic reviews of prognostic variables was developed. An electronic search of MEDLINE, EMBASE, PubMed, Cochrane Library, CINAHL, and Google Scholar was conducted independently by authors BU and SG. The databases were searched to identify studies reporting on AHCs in LT in the last 10 years (January 2014 onwards). Full literature search strategy in MEDLINE (OVID) was: (Liver transplantation OR hepatic graft OR ("Liver Transplantation/adverse effects"[Mesh] OR "Liver Transplantation/classification"[Mesh]) AND (y_10[Filter])) AND (Hepatic artery OR ("Blood Vessel Prosthesis/adverse effects"[Mesh] OR "Blood Vessel Prosthesis/classification"[Mesh] OR "Blood Vessel Prosthesis/history"[Mesh] OR "Blood Vessel Prosthesis/supply and distribution"[Mesh]) AND (y_10[Filter])).

Only studies involving human participants were included. No language restrictions were applied. Additional relevant studies were identified through manual screening of reference lists of included articles and by using the “related articles” feature in PubMed. Studies available solely in abstract form or unpublished reports were excluded. The final literature search was conducted on 19 May 2024. The PRISMA chart depicts the search strategy (Figure 1).

Figure 1 PRISMA Flowchart depicting the search strategy and selection of articles for the systematic review and meta-analysis.

Inclusion criteria

We included studies that analyzed the outcome of AHCs. Observational and comparative studies enrolling more than three patients were considered for inclusion. All studies were rigorously assessed for potential duplication or overlapping data. When multiple studies originated from the same institution, the study of superior methodological quality—judged by the primary outcomes assessed—or the most recent publication was selected for inclusion. If a study described different aspects of the overlapping populations, both studies were included. However, only one study, which provided the largest sample size, was counted in the total number calculation.

Exclusion criteria

Studies were excluded if they consisted of case reports or if they involved overlapping patient cohorts or institutional data already presented in higher-quality publications. Review articles, studies involving animal models, and experimental (non-clinical) models were also excluded. Studies with extra-anatomic inflow without a conduit were excluded.

Study selection

Two authors (Appuhamilage BU and Gupta S) independently conducted the literature search and screened all identified abstracts for eligibility based on predefined inclusion criteria. In cases where abstracts were unavailable or insufficiently informative, full-text articles were retrieved and assessed. Discrepancies in study selection between the two reviewers were resolved through consensus with a third (Hakeem AR) and fourth author (Parente A), both of whom independently reviewed all retrieved articles to ensure the robustness and consistency of the selection process.

Quality assessment

The methodological quality of the included studies was assessed using the Newcastle-Ottawa Scale (NOS). Two authors (Hakeem AR and Appuhamilage BU) independently evaluated the level of evidence based on three core domains: Selection of study groups, comparability between groups, and ascertainment of exposure or outcome. For cohort studies reporting solely on AHC outcomes, or those comparing SC-AHCs with IR-AHCs, each domain comprised one to four items, with individual items assigned a maximum score of one or two. The total NOS score was used to categorize the overall quality of each study, as summarized in Table 1.

Table 1 Risk of bias assessment (Newcastle-Ottawa Scale).

Ref.	Year	Design	Selection	Comparability	Reliability	Total
Hummel et al[11]	2014	Retrospective	4	1	2	7
Livingston et al[18]	2019	Retrospective	4	2	3	9
Vivarelli et al[19]	2020	Retrospective	4	2	3	9
Oberkofler et al[21]	2021	Retrospective	4	2	2	8

Statistical analysis

The pooled mean across included studies was calculated by the following formula: Pooled mean of included studies = (N1 × M₁ + N2 × M₂ + N3 × M₃)/(N1 + N2 + N3), where M₁, M₂ and M₃ represent the means of individual studies, and N1, N2, and N3 denote the corresponding sample sizes. For studies that did not report mean and variance, these values were estimated from the reported median, range, and sample size, following methods described in previously published literature[22,23].

Categorical variables were analysed by calculating odds ratios (ORs), while continuous variables were assessed using standardized mean differences (MD). A random-effects meta-analysis was conducted using the DerSimonian-Laird method to account for between-study variability in outcome measures. Funnel plots were used to visually evaluate potential publication bias among the included studies. Study heterogeneity was quantified using the I² to identify the extent of variation beyond chance. I² values were interpreted as indicating low (< 25%), moderate (25%-75%), or high (> 75%) heterogeneity. The Egger test was used to assess funnel plot asymmetry. Statistical significance was defined as a P-value less than 0.05. All statistical analyses were conducted using Review Manager (RevMan) version 5.4.1 (The Cochrane Collaboration, Copenhagen, 2020).

RESULTS

Study characteristics

Excluding duplicates, the search strategy identified a total of 1749 publications from 2014 to 19^th May 2024. After title screening, 106 articles were included for the abstract review. From this, 30 full papers were assessed for eligibility. Ultimately, 14 papers were included in the systematic review. All 14 papers were retrospective, observational studies[8,11-14,18,24-30]. Three papers focused on living donor LT (LDLT)[15,19,30], while the remainder addressed deceased donor LT (DDLT). Four manuscripts, which analyzed SC and IR conduit groups, were included in the meta-analysis[11,13,14,25]. Of the remaining 10 papers, two compared AHC with other extra-anatomical inflow sites[8,12], and two compared AHC characteristics with non-conduits[15,24]. Bhatti et al[17,22] had two publications, from 2019 and 2023. Given the infrequent utilization of AHCs in LDLT and the evolution of techniques over the last three years of the study period, both papers were included in the systematic review[17,22]. Oberkofler et al[21] reported their outcome data in 2018 and 2023, with the latter being a multicenter study. For the calculation of the total number of transplants and total AHCs, only the multicenter study was considered[13]. Among the four papers included in the meta-analysis, two provided data from single-center studies[11,25], while the remaining two reported multicenter data[13,14].

Risk of bias assessment

The risk of bias for all included studies was evaluated using the NOS. The outcomes of this quality assessment are presented in Table 1 and Figure 2, indicating that all studies received NOS scores exceeding 7 (high quality). Consequently, these studies were classified as either fair or good quality.

Figure 2 Risk of bias figure.

Type of conduit

The studies included data from the period between 1990 to 2020. Table 2 summarizes the demographic characteristics of the included studies. Pooled data revealed a total of 32486 DDLTs, of which 1136 (3.5%) required AHCs. Among these, 207 (18.2%) were SC AHC, 738 (65.0%) were IR AHC, and 25 (2.2%) utilized iliac artery inflow conduits. The remaining 166 (14.6%) DDLTs did not specify the type of conduit used.

Table 2 Demographic characteristics of included studies.

Ref.	Year	Country	Study period	Study design	Conduit material	Total No. of LT	Aorto-hepatic conduits	Supra-coeliac/infra-renal	Median follow up (months)
Hummel et al[11]	2014	Germany	2005-2008	Retrospective	IA	114	15	8/7	NA
Denecke et al[12]	2016	Austria	1990-2012	Retrospective	IA	947	43	1/42	NA
Li et al[13]	2017	Taiwan	2002-2015	Retrospective	GSV	NA	11	11/0	NA
Kazemi et al[14]	2017	Iran	2011-2016	Retrospective	IA	2135	76	0/76	NA
Jung et al[15]	2018	Korea	2011-2016	Retrospective	IA	1928	25	0/25	3
Oberkofler et al[16]	2018	Switzerland	2007-2016	Retrospective	IA	361	29	2/27	46
Bhatti et al[17]	2019	Pakistan	2012-2017	Retrospective	GSV	452	19	3/16	18
Livingston et al[18]	2019	United States	2000-2016	Retrospective	IA	3125	104	22/82	52.8
Vivarelli et al[19]	2020	Italy (multicentre)	2003-2018	Retrospective	IA/Axillary	4255	120	64/56	SC: 47.5; IR: 62.5
Devcic et al[20]	2020	USA	2000-2016	Retrospective	IA	3125	104	NA	NA
Oberkofler et al[21]	2021	International multicentre (14 centres)	2007-2016	Retrospective	IA/Prosthetic	11133	565	111/42 (Iliac = 14)	36.5
Beaurepaire et al[8]	2022	France	2002-2017	Retrospective	IA/Prosthetic	1677	68	Iliac artery: SC + IR = 50, Iliac = 6. Prosthesis: SC = 1, IR = 6, Iliac = 5	61
Bhatti et al[22]	2023	Pakistan	2017-2020	Retrospective	GSV	498	14	4/10	34.9
Sohrabi et al[23]	2024	Iran	2009-2020	Retrospective	IA	4010	16	0/16	35

GSV: Great saphenous vein; IA: Iliac artery; IR: Infra-renal; LT: Liver transplant; SC: Supra-coeliac.

Recipient demographics

Pooled analysis revealed comparable demographic characteristics. The OR and MD analysis confirmed the findings of no difference in age [MD -0.57 (-2.48, 1.34); P = 0.56], male gender [OR = 0.86 (0.61, 1.22); P = 0.40], recipient BMI [MD -0.92 (-1.90, 0.05); P = 0.06] for SC vs IR conduits. Further, the pooled analysis revealed that the SC group hads a higher MELD [MD 2.34 (1.30, 3.38); P < 0.0001]. Only two studies reported on the use of transarterial chemoembolization before LT, and both were multicenter studies[13,14]. The pooled analysis did not reveal any difference between the groups [OR = 1.20 (0.75, 1.92); P = 0.45] (Supplementary Figures 1-7 and Table 3). Sensitivity analysis (Supplementary Figures 8-12) did not show any other differences.

Table 3 Recipient characteristics of studies included for meta-analysis, n (%).

Ref.	Year	No. of Patients		Recipient age, median (SD)		Recipient gender (male%)		Recipient BMI (kg/m2)		Redo transplant		Indications for Transplant (V-Viral, A- Alcohol, M-MASLD, O-Other, NA-not available)		MELD		Previous TACE
		SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR
Hummel et al[11]	2014	8	7	50.9 (12.2)	46.6 (14.3)	6 (75)	7 (100)	23.5 (3)	27.4 (3.7)	3 (37.5)	3 (42.9)	NA	NA	22.7 (11.7)	16.5(9.5)	NA	NA
Livingston et al[18]	2019	22	82	51.5 (13.3)	52.7 (10.7)	14 (63.6)	57 (69.5)	25.7 (5.5)	26.3 (5.2)	10 (45.5)	45 (54.9)	NA	NA	19.6 (9.7)	19.0 (10.4)	NA	NA
Vivarelli et al[19]	2020	64	56	49.5 (14.4)	49.5 (11.8)	45 (70.3)	41 (73.2)	24.2(3.6)	25.5 (4.9)	16 (25.0)	14 (25.0)	V-32 (50.0%) A-8 (12.5%) M-8 (12.5%) O-3 (4.7%)	V-30 (53.6%) A-6 (10.7%),M-2 (3.6%),O-4 (7.1%)	22.7 (10.7)	18.7 (7.8)	13 (20.3)	12 (21.4)
Oberkofler et al[21]	2021	111	428	52.5 (11.9)	53.2 (4.9)	65 (59.0)	261 (61.0)	25.1(1.4)	25.5 (2.0)	54 (49.0)	147 (34.0)	NA	NA	25.0 (5.2)	22.7 (6.3)	20.0 (20.0)	61 (15.0)

A: Alcohol related liver disease; BMI: Body mass index; IR: Infra-renal; LT: Liver transplant; MASLD: Metabolic-associated steatotic liver disease; MELD: Model for end-stage liver disease; NA: Not available; O: Other aetiology; SC: Supra-coeliac; TACE: Transarterial chemoembolisation; V: Viral aetiology.

Indications for AHCs

Out of 14 included studies, 10 (71.0%) specified the indication for AHCs. The most frequent indications for AHC use included; poor arterial flow, intimal dissections, and hepatic artery thrombosis (HAT). Other notable reasons were insufficient length, size discrepancy of the artery, challenging dissection, and severe atherosclerosis of the hepatic artery[8,12,13,24-26,30]. Sohrabi et al[23] reported on 16 patients who underwent AHCs for HAT, although it is unclear if other indications were also considered. In the context of LDLTs, Bhatti et al[17] indicated that 85% of cases were due to arterial dissections, whereas Li et al[13] identify failure of the primary anastomosis (due to poor caliber) as the reason for AHCs.

Rationale for type of AHCs

Of the 14 papers reviewed, only 7 (50.0%) provided a rationale for selecting a specific inflow site over alternative options. Three papers cited 'surgeon preference' as the determining factor, while three adhered to predetermined inflow sites based on institutional policy[8,11,17]. Notably, two of these studies favored the splenic artery and coeliac artery before considering the aorta for conduit[15,24]. One study uniquely reported that the choice was made intraoperatively, based on the assessment of the aortic vasculature and the patient's history of previous abdominal surgeries[25].

Prosthetic and other biological grafts usage

In DDLTs, 11 studies identify the iliac artery as the primary conduit material[8,11,12,14,17,24-26,38]. Additionally, the axillary artery is mentioned as an alternative conduit option[14]. Prosthetic grafts were rarely utilized[13]. In LDLT settings, the GSV was employed as the conduit material in all three studies[15,19,30]. Two were published by the same author with varying study periods[15,30]. GSV is the conduit of choice in all patients (100%). Bleeding has been the most common complication after GSV, and conduit and it ranges between 9% to 14%. Both series do not demonstrate any early or late occlusion events.

Anticoagulation usage in patients with AHCs

Among the 14 studies, nine reported their anticoagulation protocols[8,11,13,14,15,24,26,30]. Aspirin was used as the preferred antiplatelet agent in 66% of these studies, primarily due to conduit use. Intravenous heparin was included in 33% of the protocols, with durations ranging from 5 days to a maximum of 3 months[15,24]. One study described the concurrent use of both aspirin and heparin[24], while only one center specified lifelong use of aspirin[15].

Complications of AHCs

The median follow-up duration ranged from 18 to 52 months. Complication rates did not significantly differ between the SC and IR groups (Tables 4 and 5). Forest plot analysis confirmed no significant differences in early occlusions of AHCs [OR = 0.94 (0.48, 1.84); P = 0.86] (Figure 3A), and late occlusions of AHCs [OR = 0.46 (0.16, 1.32); P = 0.15) (Figure 3B), early allograft dysfunction [OR = 0.82 (0.46, 1.47); P = 0.51] (Figure 3C), biliary complications [OR = 1.10 (0.69, 1.76); P = 0.68] (Figure 3D), post-transplant renal replacement therapy (RRT) requirement [OR = 1.12 (0.72, 1.72); P = 0.62] (Figure 3E), and major surgical complications (Clavien-Dindo > 3b) [OR = 1.06 (0.70, 1.61); P = 0.79] (Figure 3F and Table 6). The median duration for graft occlusion was approximately 142 days, ranging from 13 to 3313 days (Table 4).

Figure 3 Forest plot. A: For early occlusions comparing supra-coeliac with infra-renal aorto-hepatic conduits; B: For late occlusions comparing supra-coeliac with infra-renal aorto-hepatic conduits; C: For early allograft dysfunction comparing supra-coeliac with infra-renal aorto-hepatic conduits; D: For biliary complications comparing supra-coeliac with infra-renal aorto-hepatic conduits; E: For post-transplant renal replacement therapy comparing supra-coeliac with infra-renal aorto-hepatic conduits; F: For major surgical complications (Clavien-Dindo 3b >) comparing supra-coeliac with infra-renal aorto-hepatic conduits. IR: Infra-renal; SC: Supra-coeliac.

Table 4 Details of aorto-hepatic conduits in the included studies, n (%).

Ref.	Year	Number of patients		Early occlusions		Late occlusion		Bile duct complications		Graft survival 1-year		Graft survival 5-year		Patient survival 1-year		Patient survival 5-year
		SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR
Hummel et al[11]	2014	8	7	0	0	0	0	4 (50.0)	3 (42.9)	NA	NA	NA	NA	57.1%	85.7%	NA	NA
Livingston et al[18]	2019	22	82	0	0	0	5	NA	NA	85.1%	79.1%	53.9%	56.0%	85.1%	88.3%	67.8%	64.9%
Vivarelli et al[19]	2020	64	56	4 (6.2)	6 (10.7)	3 (4.7)	11 (19.6)	20 (31.2)	22 (39.3)	77.0%	66.0%	67.0%	50.0%	80.0%	73.0%	74.0%	56.0%
Oberkofler et al[21]	2021	111	428	9 (8.1)	31 (7.2)	9 (8.0)	41 (9.8)	21 (19.0)	61 (14.0)	NA	NA	NA	NA	NA	NA	NA	NA

IR: Infra-renal; NA: Not available; SC: Supra-coeliac.

Table 5 Complications related to aorto-hepatic conduits in the included studies, n (%).

Ref.	Year	No. of AHCs	Total occlusions	Surgical interventions for oocclusions			Endovascular intervensions for occlusions				Success rate of surgery without retransplantation	Success rate of endovascular
				Re-anastomosis	Re-conduit formation	Retransplants	Stent	PTA	Stent/PTA	CDTL
Hummel et al[11]	2014	15	0	0	0	0	0	0	0	0	NA	NA
Livingston et al[18]	2019	104	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA
Vivarelli et al[19]	2020	120	24 (20.0)	3 (12.5)	NA	8 (33.0)	8 (33.0)		NA	NA	NA	NA
Oberkofler et al[21]	2021	565	96 (17.0)	8 (8.3)	4 (4.2)	22 (23.0)	4 (4.2)	12 (13.0)	7 (7.3)	6 (11.0)	7 (27.0)	18 (82.0)
Sohrabi et al[23]	2024	16	2 (12.5)	0	0	2 (100)	0	0	0	0	NA	NA
Devcic et al[20]	2020	104	7 (6.7.0)	0	0	0	0	7 (100)	0	0	NA	NA

AHCs: Aorto-hepatic conduits.

Table 6 Other complications in patients receiving aorto-hepatic conduits, n (%).

Ref.	Year	Number of patients		Pre-Cr at LT (mg/dL)		Post-Cr maximum (mg/dL)		Post-transplant CRRT		Early allograft dysfunction		Major surgical interventions (Clavien 3b>0)
		SC	IR	SC	IR	SC	IR	SC	IR	SC	IR	SC	IR
Hummel et al[11]	2014	8	7	0.56 (0.33)	0.97 (0.37)	0.8	1.47	NA	NA	0	0	1 (12.5)	1 (14.3)
Livingston et al[18]	2019	22	82	1.15 (0.71)	1.20 (0.72)	1.17 (0.22)	1.2 (0.26)	2 (10.5)	10 (12.2)	4 (18.2)	24 (29.3)	NA	NA
Vivarelli et al[19]	2020	64	56	1.35 (0.79)	1.82 (1.1)	NA	NA	NA	NA	4 (6.2)	6 (10.7)	NA	NA
Oberkofler et al[21]	2021	111	428	NA	NA	NA	NA	35 (32.0)	122 (29.0)	9 (8.1)	31 (7.2)	66 (60.0)	248 (58.0)

IR: Infra-renal; NA: Not available; SC: Supra-coeliac.

Graft and patient survival based on type of AHCs

Pooled data revealed one-year graft and patient survival rates for SC conduits were 77% to 81.1% and 80% to 85.1%, respectively. For IR conduits, one-year graft and patient survival rates were 66% to 79.1% and 73% to 88.3%, respectively. Five-year graft and patient survival rates for SC conduits were 53.9% to 67% and 67.8% to 74%, respectively. For IR conduits, five-year graft and patient survival rates were 50% to 56% and 56% to 64.9%, respectively.

Except for two studies, all others reported doppler ultrasound scan (USS) as the post-transplant imaging patency surveillance method[8,11,14,24-26,30]. However, the frequency of doppler USS application was inconsistent among the studies. Some studies performed daily USS in the first week post-transplant[12,14,24-26]. One study conducted doppler USS every 12 hours, while others performed it two or three times per week[24]. Angiography was the next reported imaging modality, utilized on demand based on suspicion of conduit occlusion.

Interventional radiology to treat AHCs complications

Out of 14 studies, 5 reported interventions for stenosis[13,17,24,25,28]. The systematic review identified four options for managing AHC occlusion: Endovascular dilation with or without stenting, catheter-directed thrombolysis, conduit revision, and retransplantation. Among these, angioplasty and stenting emerged as the preferred treatments, with retransplantation being the last resort. The re-anastomosis rate ranged from 0 to 12.5%, while the re-conduit rate was 0 to 4.2%[13,17].

DISCUSSION

This study primarily focused on the use of AHCs in LT, specifically comparing the outcomes between the two common inflow sites, namely SC and IR. When the native recipient’s hepatic artery is unusable, some series have shown an increasing preference for splenic artery transposition over the formation of AHCs[7,8]. However, in cases where the coeliac axis is involved by the same pathology, AHC becomes the only option for revascularizing the liver graft. This study showed that IR conduit is the most preferred inflow site over SC, however this systematic review identified choice was mostly “predetermined” rather than an intra-operative decision[31]. The absence of notable differences in early outcomes between SC and IR conduits indicates that surgeons often base their choice on personal expertise and familiarity with the technique. These results provide reassurance that variations in surgical preference do not compromise patient outcomes, affirming that both methods are equally effective when performed by experienced hands. This study revealed a utilization rate of only 3.5% for AHCs in DDLT, which is lower than the 9% conduit rate reported in a previous meta-analysis conducted by Reese et al[32]. In this meta-analysis, 12 of the 13 studies included the patient populations from after 2000, whereas Reese et al[32] included patients from 1989 onwards. The decline in the rate of AHC in LT may be attributed to advancements in pre-transplant imaging and surgical planning over the past two decades. Currently, patients with complex arterial issues and associated vascular and metabolic comorbidities are likely excluded due to more stringent and improved patient selection criteria.

This meta-analysis also identified the donor iliac artery as the main AHC in the DDLT setting. The use of prosthetic grafts has been very rare. In LDLT settings, the GSV serves as the main conduit material. Although data are limited and the patient cohort is small (Table 7), outcomes with GSV appear encouraging, demonstrating a low occlusion rate[15,30]. However, this systematic review did not identify any use of GSV in the DDLT setting. Although the GSV shows promising outcomes as an AHC in LDLT, it is well reported that the arterial conduits generally provide the best results in cardiac bypass surgery[33]. In the context of high-pressure aortic flow and the long length of AHC grafts, particularly in IR aortic conduits, the use of the GSV conduit instead of an arterial conduit, such as the radial artery, requires justification.

Table 7 Summary table aorto-hepatic conduits.

Aorto-hepatic conduit details	Number of studies that have reported (n = 14)
Infow sites	14
Conduit material	14
Second party vs third party donor	1
Storage of the conduits	0
Reason for conduit	9
Choice of inflow site	12
Post-operative doppler monitoring	12
Post-operative anticoagulation	9
Early occlusions	8
Late occlusions	8
Endovascular Interventions for occlusions	3
Surgical interventions (revisions/reconduits/retransplants)	4

Conduit occlusion is primarily driven by endothelial injury and thrombosis, with chronic inflammation—such as that associated with rejection or graft arteriopathy—also playing a key role in its development[13]. The rate of conduit occlusion in the included studies ranged from 0% to 20%[13,14]. The meta-analysis shows no difference between SC and IR conduits in relation to short-term occlusion rates. However, there is a discernible tendency towards higher late occlusion in the IR group. These findings are consistent with the multicenter study conducted by Oberkofler et al[21]. The SC group had similar or slightly superior outcomes despite higher MELD scores, suggesting that the SC conduit may give the best results in the long-term compared to the IR group. The increased late occlusion rates observed in IR conduits may be due to their greater length, which is associated with turbulence-induced arteriopathy, and luminal narrowing. In contrast, SC conduits—being shorter and offering more physiological blood flow—may provide a protective effect, contributing to improved long-term patency. The role of anticoagulation and antiplatelet therapy, aimed at preventing thrombotic events, has become crucial, particularly for IR conduits, to reduce the incidence of late occlusions. Of note, Sohrabi et al[23] identified significantly fewer biliary complications in AHCs compared to HA anastomosis revision (47.4% vs 31.2%), in the context of early HAT which is likely to be attributable to multiple re-explorations after the HA revision[24]. Vivarelli et al's multicenter European study supports our findings, indicating no significant differences in early outcomes between SC and IR AHCs[19]. However, while both studies underscore the long-term occlusion risks associated with the IR group, our review places greater emphasis on the clinical implications of anticoagulation protocols and postoperative monitoring, whereas the referenced study is more focused on procedural differences in conduit placement.

In LDLT, Li et al[13] performed all their conduit placements using the SC approach. In contrast, Bhatti et al[17] utilized both the SC and IR aorta, with a predominance of conduits being placed in the IR aorta. According to Bhatti et al[17], the majority of conduits are utilized for managing arterial dissections, which are typically identified after reperfusion. This post-reperfusion identification of dissection can make access to the SC aorta challenging, prompting the preference for the IR aorta.

Another finding of this systematic review reveals a notable trend towards endovascular interventions and a decreased preference for re-anastomosis or re-conduit in cases of conduit occlusion. Given the limited availability of donor organs, it is understandable that most centers are focusing on salvaging the existing graft whenever possible, reserving retransplantation as a last resort. According to Devcic et al[20], angioplasty achieved a 100% technical success rate in the seven patients who required this intervention. Only two patients needed reintervention and there were no procedure-related adverse events[28].

This systematic review has highlighted the variable use of anticoagulation following acute AHC thrombosis. All studies reviewed employed additional pharmacological anticoagulation methods, such as aspirin, heparin, or a combination of both. However, the specific duration of these treatments was not clearly defined. Considering the limited data and the reported median graft occlusion duration of 142 days, it can be suggested that a pharmacological prophylaxis regimen lasting at least six months could provide beneficial outcomes, potentially enhancing results with AHCs[28]. However, this should be further investigated in a large multicenter study. To date, the value of lifelong anticoagulation as a preventive strategy for AHC occlusion remains contentious, although it is commonly practiced in some centers[15]. Oberkofler et al[21] in their multicentre study identified the use of aspirin as an effective preventive method for conduit occlusion[12]. Due to the great variability in Doppler scan surveillance in graft inflow and outflow, the review is unable to evaluate the validity or effectiveness of these surveillance practices[32,33]. The absence of standardized protocols, particularly during the early postoperative period, can have serious consequences, including graft ischaemia and failure. With endovascular intervention on board, it is of utmost importance to detect AHC occlusions early, as a more proactive approach can salvage the conduit and the graft[12,32].

The main limitation of this study is the potential overlap between the two large multicenter studies. Additionally, the limited LDLT data make it challenging to draw firm conclusions. The higher MELD scores in the SC group may introduce bias in inflow site selection due to issues like suboptimal aortic vessels and bowel adhesions from recurrent SBP. There's also considerable variability in patient characteristics, management strategies, and study designs. While subgroup analyses and I² statistics were used to assess variability, these reflect real-world clinical practice and improve generalizability. However, further subgroup analyses or meta-regression are limited by the small number of studies and the potential for unreliable results.

CONCLUSION

The study highlights several key areas for future research related to AHC. These include investigating the use of larger, high-pressure conduits in pediatric LT, exploring the impact of novel perfusion technologies, and examining how preoperative handling of the donor hepatic artery might influence outcomes. The lack of consensus on post-transplant anticoagulation also underscores the need for further studies and the development of guidelines. Addressing these gaps could improve clinical practices. In the meantime, this review provides valuable data that can serve as a basis for future research