We retrospectively reviewed the medical records of 924 adult recipients of LDLT at Kaohsiung Chang Gung Memorial Hospital between July 2008 and September 2017. The study was approved by the institutional review board (number:202200107B0).

HRS was defined following the diagnostic criteria published by the ICA: (1) cirrhosis and ascites; (2) diagnosis of AKI according to the ICA criteria; (3) no response after two consecutive days of diuretic withdrawal and plasma volume expansion with albumin; (4) absence of shock; (5) no current or recent use of nephrotoxic drugs; and (6) no macroscopic signs of structural kidney injury, which is indicated by proteinuria, microhematuria, or abnormal findings on renal ultrasonography (9). The initial criteria were revised in 2007 and 2015 and subsequently updated in 2019 (1). All patients with HRS included in this study fit the diagnostic criteria described by the ICA; the definition of type-1 HRS is now referred to as HRS-AKI and be defined based on serum creatinine ≥0.3 mg/dl within 48 h or ≥50% from baseline value according to ICA consensus document and/or urinary output ≤0.5 ml/kg body weight ≥6 h. Whereas type-2 HRS is now known as HRS-CKD and be defined as eGFR <60 ml/min/1.73 m² for ≥3 months in the absence of other causes (2).. (Table 1) According to the criteria, the patients were divided into the HRS1 group (11 patients) and the HRS2 group (19 patients). In addition to the HRS1 and HRS2 groups, we included two control groups: the CKD group (43 patients) which consisted of patients diagnosed with CKD stage ≥ III (determined by the estimated glomerular infiltration rate (eGFR) using the Modification of Diet in Renal Disease (MDRD) equation before LDLT) (10) and the normal renal function group (67 patients) which consisted of patients with normal pre-LDLT renal function (determined using propensity score matching based on sex and age). CKD stages ≥ III were defined as eGFR <60 ml/min/1.73 m² 3 months or kidney damage 3months per the Kidney Disease Outcomes Quality Initiative guidelines (11). To distinguish between HRS2 and CKD need to rule out structural kidney injury, proteinuria and microhematuria were routinely checked in our patients. However, for patients with advanced cirrhosis or hepatorenal syndrome, the risk of renal biopsy is relatively high. For such reason, we did not obtain renal biopsy in our recipients. The differential diagnosis was mainly done by the timing of renal deterioration (clinical course) and the duration of renal dysfunction.

Patients with liver cirrhosis who were diagnosed with HRS and developed more than one organ failure were admitted to the liver intensive care unit before LDLT. The first aim was to evaluate and treat the reversible etiology of renal injury such as gastrointestinal bleeding, sepsis, and spontaneous bacterial peritonitis (3). Additionally, exposure to nephrotoxic agents was avoided. Examples of these include nonsteroidal anti-inflammatory medications, antibiotics with renal toxicity, and radiocontrast. In patients with HRS, vasoconstrictors (e.g., terlipressin) and albumin are used, and precise fluid management is evaluated by continued central venous pressure and urine output monitoring (7). Renal replacement therapy (RRT) is indicated when pulmonary edema, severe hyperkalemia, metabolic acidosis, and/or complications of uremia develop (2). There was no RRT for HRS patients before LDLT in this study. The preoperative eGFR was routinely measured within one week before transplant, and the preoperative eGFR value was used as the baseline eGFR for further comparison.

All LDLTs in this study were ABO-compatible, and partial liver grafts were obtained from living relatives following the Organ Transplant Act of Taiwan. All operations were approved by the Institutional Ethics Committee. Both recipient and donor surgeries followed the protocol used in our liver transplant center, as described elsewhere (12). The graft-to-recipient weight ratio (GRWR) threshold was > 0.8% and the graft-to-recipient standard liver volume ratio was > 40% (12).

Post-LDLT management followed the protocol established in our center (13). Briefly, basiliximab was administered 6 h after portal vein reperfusion (day 0) and on postoperative day 4. Prostaglandin E1 and heparin were started once bleeding tendency was at an acceptable point for 2 weeks. Prophylactic antibiotics including antibacterial and antifungal agents are routinely administered. Hyperimmune cytomegalovirus (CMV) globulin was routinely administered once a week for 1 month to prevent CMV infection. Hepatitis B immunoglobulin (HBIG) and lamivudine were administered to prevent hepatitis B recurrence or de novo hepatitis B infection. Nephrotoxic agents such as aminoglycosides were avoided. The details of this process are described later in this text (6, 14). The maintenance immunosuppressive drugs were tacrolimus and mycophenolate mofetil, mainly used in stable recipients (15). Given the significant nephrotoxic effects of calcineurin inhibitors, renal-sparing regimens were used to preserve renal function in patients with HRS (3). Tacrolimus administration was delayed until renal function improved, as evidenced by adequate urine output (1 ml/kg/hour) and decreased serum creatinine levels (14). A mammalian target of rapamycin (mTOR) inhibitor was added or replaced with tacrolimus in patients with impaired renal function (16). Liver ultrasonography, blood biochemistry, including eGFR, and immunosuppression serum levels were routinely followed up every 2–3 months (6). Additionally, regular follow-up by the nephrologist is essential for patients with significant CKD risk factors, such as diabetes and hypertension (14, 17). The ESRD was defined as eGFR <15 ml/min/1.73 m² and the time of RRT was determined by the nephrologist. The last follow-up in this study ended in August 2022, and the minimum follow-up period was 5 years.

Primary outcomes were eGFRs measured using the MDRD equation (10) preoperatively and at 1 week, 1 month, 3 months, 6 months, 1 year, 2 years, and 5 years post-LT. Compared to other formulae, the MDRD equation is a more precise and accessible method for measuring renal function in LT recipients (18). Secondary outcomes were the incidence of CKD (defined as eGFR falling below 60 ml/min/1.73 m² according to the KDIGO guideline) at 5 years post-LDLT (11).

Data were collected and analyzed using IBM SPSS version 20 (IBM Corporation, Armonk, NY, United States). Categorical variables between the groups were compared using the Chi-square test. Numerical variables are expressed as medians with interquartile ranges (IQR) and compared using the Kruskal-Wallis test. The change in eGFR (mean ± 2 standard deviations) was calculated using the generalized estimating equation. Univariate and multivariate logistic regression analyses were used to determine the risk factors for post-LT permanent CKD in patients with HRS. Variables with P values less than 0.2 were considered significant in the univariate analysis and were included in the multivariate analysis. The preoperative cutoff values of eGFR to determine the prognostic value of post-LT CKD in patients with HRS were calculated using a receiver operating characteristic curve. Kaplan-Meier survival curves and log-rank tests were used to evaluate the 5-year survival. Statistical significance was set at p < 0.05.

A total of 115 (82.1%) males and 25 (17.9%) females were included in this study. The median age was 53–58 years in each group (Table 2). According to the diagnostic criteria (Table 1), 11 patients were classified into the HRS1 group, 19 into the HRS2 group, and 43 into the CKD group. Sixty-seven patients with preoperative normal renal function were identified by propensity score 1:1 matching method and then defined as the reference group. None of the patients were lost to follow-up in this study.

The prevalence of underlying liver diseases was as follows: hepatitis B (48.9%), hepatitis C (24.8%), alcoholic liver disease (17.0%), and other etiologies (9.3%). There were significantly fewer patients with HCC in HRS1 (n = 0, 0%) and HRS2 (n = 3, 15.8%) than in the CKD (n = 16, 37.2%) and reference groups (n = 36, 53.7%) (P = 0.018 and 0.001, respectively). There were no differences in age, sex, primary liver disease, or pre-existing comorbidities, including hypertension and diabetes, among the four groups (P > 0.05). However, the severity of liver dysfunction, in terms of Child-Pugh score, MELD score, and serum total bilirubin level, was significantly higher in the HRS1 and HRS2 groups than in the CKD and reference groups (P < 0.05).

All three groups (HRS1, HRS2, and CKD) had significantly lower eGFR and higher serum creatinine levels than those of the reference group (Table 2). In addition, the HRS1 group had worse renal function than the HRS2 group did. Nonetheless, none of the patients with HRS required RRT before LDLT. Interval time from onset of HRS1 to transplantation was showed in Table 3.

The perioperative parameters, including cold/warm ischemia time, GRWR, and intraoperative blood loss (7000 vs. 4,650 vs. 5,000 vs. 2300 ml; P = 0.891), did not show statistical differences among the HRS1, HRS2, CKD, and reference groups. The ascites amount at the time of LT showed significant differences among 4 groups (P < 0.01) but no statistical differences among the HRS1, HRS2, CKD (P = 0.159) (Table 2). The risk of postoperative Clavien-Dindo complications ≥ grade IIIb (59% vs. 26.3% vs. 58.1% vs. 41.7%; P = 0.073) was similar in the HRS1, HRS2, CKD, and reference groups. No 30-day postoperative mortalities were observed.

More than 90% of the patients were administered tacrolimus. As expected, mTOR was more frequently prescribed to patients with renal insufficiency (91.6% for HRS1, 78.9% for HRS2, 86.0% for CKD, and 38.8% for the reference group, P < 0.001; Table 2). The details between each groups were present in Supplementary Material Table S1.

The eGFR of the reference group declined gradually 1 month after LT and stabilized 1 year after LDLT (Table 4). We further compared the eGFR between the HRS1, HRS2, and CKD groups. One week after LT, the eGFR (58, 82, and 51 ml/min/1.73 m², respectively) markedly improved comparing to the preoperative values in all three groups. One month after transplantation, the eGFR in HRS1 continued to improve, whereas in HRS2 and CKD, the eGFR declined (Table 4). However, at the third month of transplantation, the eGFR (54, 62, and 41 ml/min/1.73 m², respectively) deteriorated in all three groups. Between the 3rd and 6th months after LT, the eGFR of HRS2 kept decreasing whereas the eGFR of HRS1 and CKD groups maintained at relatively levels. One year after transplantation, eGFR stabilized. There was no significant difference in eGFR in the HRS1, HRS2, and CKD groups at the first year post-LT (55, 47, and 39 ml/min/1.73 m², respectively; P = 0.098), second year (56, 49, and 41 ml/min/1.73 m², respectively; P = 0.082), and fifth year (56, 52, and 48 ml/min/1.73 m², respectively; P = 0.149). The pre-LT and early post-LT eGFR among the three groups were significantly different (all P < 0.05), but became similar 1 year after LT (all P > 0.05). Interestingly, despite HRS1 started with the worst pretransplant eGFR among all the groups, the eGFR could be maintained at around 55 ml/min/1.73 m² from the 3rd month after LT throughout the 5-year follow-up (Figure 1).

CKD stage ≥ III developed in 8 (72.7%), 15 (78.9%), 39 (90.7%), and 5 (7.5%) patients in the HRS1, HRS2, CKD, and reference groups (respectively) after 5 years of LDLT (P = 0.001). Moreover, 12 (27.9%) patients in the CKD group eventually progressed to the ESRD status, while one (8.3%) patient in the HRS1 group, and two (10.5%) patients in the HRS2 group required RRT (P = 0.248). Finally, one patient (8.3%) in the HRS1 group and three patients (6.9%) in the CKD group underwent kidney transplantation at the end of follow-up (P = 0.293) (Table 5).

The 5-year survival rates after LDLT were 89.6%, 91.7% for HRS1, 94.7% for HRS2, and 81.4% for CKD group, respectively. There was no significant difference between the groups (P = 0.399) (Figure 2). Infection was the most common cause of mortality in the CKD group (n = 5). HCC progression was the most common cause of mortality in the reference group (n = 5) (Table 5).

In a univariate analysis (Table 6), the variables significantly associated with CKD stage ≥ III after 5 years of LDLT included pre-LT eGFR (odds ratio [OR] = 0.96, 95% confidence interval [CI] = 0.920–0.995, P = 0.028) and postoperative complications ≥ Gr.IIIb (OR = 0.16, 95% CI = 0.025–0.999, P = 0.050). MELD score, pre-LT creatinine, pre-LT eGFR, GRWR, post-LDLT complication ≥ Gr. IIIb, and use of terlipressin were entered into multivariate logistic regression. The significant variable in the multivariate analysis was the pre-LT eGFR (OR = 0.96, 95% CI = 0.920–0.995, P = 0.028). When the cutoff value of pre-LT eGFR was set at <46.4 ml/min/1.73 m² (AUC = 0.807, 95% CI = 0.617–0.997, P = 0.011), the sensitivity and specificity for predicting CKD stage ≥ III after 5 years of LDLT were 68.2% and 87.5%, respectively (Figure 3).