Between January 2012 and October 2016, we retrospectively studied the records of 237 LDLT patients using data collected among 241 cases of consecutive adult LDLT performed at our institution. The medical records were reviewed retrospectively in terms of patient demographics and intra- and postoperative findings, including patient age, disease etiology, laboratory data, operative outcomes, postoperative complications, and long-term outcomes. For standardization purposes, 4 patients in the period primarily using a surgical loupe were excluded due to only availability of operative microscope for the following cases: right posterior sector grafts with too small caliber and short stump of donor HA or right lobe grafts with double donor HA stumps which have both a smaller caliber and short length. The study was approved by the Institutional Review Board of our institution. The patients were divided into two groups, depending on whether an operative microscope or surgical loupe was used during HA reconstruction. In group I (n=136), HA reconstruction was performed under a surgical microscope by an experienced microsurgeon until September 2014. Subsequently, a transplant surgeon who had no experience in arterial anastomosis during LDLT performed the HA reconstruction using a surgical 5×loupe (group II, n=101). He only had experience in vascular anastomosis in large vessels such as the hepatic vein or portal vein during LDLT and received no prior microsurgical training. The mean follow-up period was 50.9±17.1 months in group I and 23.5±9.1 months in group II (p=0.000). However, this difference was due to the historical nature of the two groups.

The detailed surgical procedure used for HA reconstruction is described elsewhere and the anastomosis procedures were similar in the two groups.1011 The recipient right and left hepatic arteries can be dissected out higher in the hilum to maximize artery length and mobility for the application of double microvascular clamps. After adequate debridement, the quality of the recipient HA was inspected carefully, including the interior of the vessel and the strength of the pulsatile arterial flow. A double micro-clamp is first applied to the donor HA and then to the recipient HA after aligning both terms of length and rotation. An end-to-end anastomosis is performed using interrupted 8-0 Nylon sutures with the aid of an operating microscope or surgical loupe. We didn't perform continuous suturing in all cases. First, both the dorsal and ventral ends were anastomosed and four or five sutures were placed while first two staying sutures were stretched apart gently by first assistant. Back wall stiches were placed in similar fashion after the micro-clamp were rotated 180° keeping two corner sutures stretched like conventional twist technique. The back wall-first anastomosis technique was employed in cases which couldn't flip the artery due to inadequate vessel length.

Immediately after completion of the vascular reconstruction, intraoperative Doppler ultrasound (DUS) is performed routinely to check adequacy of the HA and portal vein inflow and they were addressed immediately if the intraoperative DUS findings were abnormal. If there was no HA flow during intraoperative assessment, after a brief observation, we tried to do HA anastomosis repeatedly to establish the arterial flow. If the cause of low HA flow is arterial spasm, we advocate a brief waiting period for the spasm to relieve and to rule out any anatomical factors such as kinking, angulation or thrombosis. However, if this is not related to HA spasm, we performed immediate revision of the HA anastomosis.

DUS was performed on post-transplant days 1 and 3, and dynamic computed tomography (CT) scans were performed on days 7 and 14 to assess the inflow and outflow graft patency in the early posttransplant period. If hepatic arterial or portal vein complications were suspected during DUS with elevated liver enzymes, those complications were further confirmed by a CT angiography.

The immunosuppression regimen consisted of calcineurin inhibitors, mycophenolate mofetil and low-dose steroids. Fresh-frozen plasma was transfused only when massive bleeding and oozing were observed. Prostaglandin E1 was administrated intravenously immediately following graft reperfusion and continued for seven days. All patients were not heparinized in the postoperative period as prophylaxis. Acetylsalicylic acid 100 mg/day was started when the patient tolerated oral intake.

All numerical data are reported as the mean and standard deviation. Student's t-test or the Mann–Whitney U test was used to compare continuous variables according to their distributions. The chi-square test or Fisher's exact test was used to compare categorical variables after assumptions were verified. Survival and patency rates were determined with the Kaplan–Meier method and were compared with the log-rank test. Multivariate analysis was applied to assess the statistical association and risk factors for HAT and intraoperative HA revision. Multivariate analyses with logistic regression were performed to quantify the adjusted odds ratio (OR) to control for established risk factors and to assess the independent impact of the use of surgical loupe on postoperative outcome. All analyses were performed using SPSS Statistics ver. 19.0 (IBM, Armonk, NY, USA). A value of p<0.05 was taken to indicate statistical significance.

Table 1 summarizes the recipient and donor demographic data. No significant difference was observed between the two groups, except for recipient age and incidence of ABO-incompatible LDLT. The difference in the incidence of ABO-incompatible LDLT was only due to the chronological nature of the two groups. The principal etiology of liver disease was viral hepatitis-related liver cirrhosis in 153 patients (101 of whom also had hepatocellular carcinoma). The mean recipient and donor ages were 51.7±9.1 and 31.5±11.4 years old, respectively. The mean score for model for end-stage liver disease (MELD) was 17.0±10.1. Thirty-seven recipients (15.6%) underwent ABO-incompatible LDLT. A single HA anastomosis was performed in 223 patients (94.1%), while double anastomoses were performed in 14 cases (5.9%) and the mean donor HA diameter also did not differ between the two groups. The incidence of intraoperative HA revision was not different between the two groups, and the most common cause was thrombosis due to a poor HA condition and intimal dissection. The HA reconstruction took longer in group I than in group II, although the HA reconstruction was performed by a more experienced microsurgeon in group I (p=0.001). No other perioperative variables differed significantly between the two groups (Table 2).

In multivariate logistic regression analyses, only significant hilar fibrosis (OR, 22.033; 95% confidence interval [CI], 1.179–411.899; p=0.038) and a smaller HA diameter (defined as less than 1.8 mm) (OR, 0.002; 95% CI, 0.000–0.166; p=0.007) were associated with HA thrombosis (HAT) but not related to the use of surgical loupe (OR, 0.969; 95% CI, 0.0076–12.351; p=0.981) when adjusted for the established risk factors of HAT such as recipient age, sex, previous hepatectomy, pretransplant locoregional therapy including transarterial chemoembolization, portal vein thrombosis or stenosis, use of surgical loupe, significant hilar fibrosis which cause difficult dissection, multiple graft HA, small-caliber HA and ABO incompatible LDLT. Moreover, a small-caliber HA was only associated with intraoperative HA revision (OR, 0.051; 95% CI, 0.008–0.322; p=0.002) but was also not related the use of surgical loupe (OR, 0.831; 95% CI, 0.224–3.086; p=0.782) when adjusted for the same factors.

Table 3 shows details of the main postoperative complications; no significant difference was found between the groups. There were four cases (1.7%) of HA thrombosis: two cases in group I and two cases in group II. All thromboses were detected within 2 days after LDLT, and inappropriate manipulation for intimal detachment due to poor arterial conditions was regarded as the main cause. All thromboses were resolved with prompt re-exploration and HA revision. The resistive index and peak velocity of the HA on serial postoperative DUS were not different between the groups. Moreover, mild HA stenosis (defined as anastomotic narrowing of less than 50% with no clinical sign)10 was diagnosed in 27 patients (11.2%) by dynamic CT scan on postoperative day 14 but did not differ between the groups. No significant differences in biliary stricture were observed between the two groups. The overall patient and graft survival also did not differ significantly between the groups by Kaplan–Meier log-rank analysis (p>0.05).

We divided group II into subgroup era I (the first 50 cases) and subgroup era II according to evaluate the effect of learning curve phenomenon on LDLT outcomes. According to subgroup analysis, HA reconstruction time in era I have significantly higher than that in era II (25.8±8.3 vs. 18.1±4.4 minutes, p=0.031) and the increased HA occlusion rates including HAT and HA revision in era I than that in era II were shown (12.0% vs. 3.9%) although there was no significance. As compared with microscopy group, these rates related to HA occlusion in subgroup era I have slightly higher than that in microscopy group (5.9% vs. 12.0%, p=0.161).