Excessive Intraoperative Blood Loss Increases the Risk of Postoperative Complications After Liver Transplantation: A Retrospective Multicenter Study

Introduction

As a key treatment for end-stage liver disease and hepatocellular carcinoma, liver transplantation (LT) is both complex and associated with postoperative complications, similar to other major surgical procedures [1,2]. Postoperative complications, such as excessive blood loss, biliary tract complications, vascular complications, rejection, infection, and graft dysfunction, directly influence patient survival [3–5]. Prevention and management of postoperative complications is a shared priority for surgeons and transplant recipients; it remains essential for improving clinical outcomes.

LT is a complex and time-consuming procedure that often requires large incisions, increasing the risk of excessive intraoperative blood loss (EIBL), particularly in patients with concurrent coagulation dysfunction [6]. Pandey et al [4] and Alhamar et al [7] reported that excessive blood loss during LT is associated with procedural difficulty, liver disease characteristics, abnormal coagulation function, previous major abdominal surgery (PMAS), and prolonged operative duration. EIBL substantially affects postoperative quality of life. It also requires blood transfusion, and the infusion of large fluid volumes can elevate citric acid levels. Citrate toxicity may result in hypothermia and hypocalcemia, further worsening coagulation dysfunction and increasing the risks of continued intraoperative and postoperative blood loss [4,8]. Thus, EIBL and the subsequent need for transfusion can initiate a cascading effect. Previous studies have shown associations of EIBL with prolonged hospital stays, acute rejection, infection, graft dysfunction, biliary complications, and postoperative survival outcomes [5,8–10].

The identification of risk factors for EIBL and implementation of appropriate preoperative planning and surgical strategies can reduce EIBL risk, thereby decreasing postoperative complications and improving survival rates. However, few studies have examined EIBL risk factors and its impact on postoperative complication incidence, especially among patients undergoing deceased donor liver transplantation (DDLT). A large multicenter study of living donor LT indicated that EIBL was associated with biliary stenosis, bile leakage, prolonged hospitalization, and reduced graft survival [6,10]. In contrast, studies focused on DDLT remain limited and have yielded inconsistent findings [5,11]. Therefore, a large multicenter study of DDLT is warranted.

With the aim of addressing gaps in the current literature, we conducted a retrospective cohort study of 822 adult patients who underwent DDLT between January 2015 and June 2024. By comprehensively evaluating the effect of EIBL on postoperative complication incidence, we aimed to provide recommendations for preoperative assessment and to optimize surgical practice to reduce intraoperative blood loss and complication risk, thus improving patient survival.

Material and Methods

STUDY DESIGN:

This retrospective study included data from 5 centers: the First Affiliated Hospital of Army Military Medical University, Sichuan Provincial People’s Hospital, Guizhou Provincial People’s Hospital, Liaocheng People’s Hospital in Shandong Province, and Ruijin Hospital in Shanghai. In total, 1122 patients who underwent DDLT between January 2015 and June 2024 were identified; each center contributed data from at least 20 patients annually. Patients were excluded if they died perioperatively, were younger than 18 years, underwent split LT or living donor LT, or were lost to follow-up. Ultimately, 822 patients were included in the statistical analysis.

ETHICAL CONSIDERATIONS:

This study was conducted in accordance with the Declaration of Helsinki (as amended in 2013). Approval was granted by the Ethics Committee of the First Affiliated Hospital of Army Medical University, PLA (Southwest Hospital) (approval No.: (B)KY2021013), and the study was implemented in strict accordance with the approved protocol (Version V1.0; April 21, 2021). Due to the retrospective nature of the study, the requirement for informed consent was waived. Information regarding organ donation and allocation is available through the Chinese Liver Transplant Registry at http://cltr.cotr.cn. No organs from executed prisoners were used in this study.

DATA COLLECTION:

All patients underwent regular follow-up evaluations, and data were obtained from the electronic medical records of the 5 transplant centers. When uncertainties or missing records were identified, patients were contacted by telephone to obtain updated information. Preoperative data included sex, age, body mass index, blood type, hypertension, Child-Pugh score, pre-LT intensive care unit (ICU) stay (yes/no), mechanical ventilation status, portal hypertension, refractory ascites, Model for End-Stage Liver Disease (MELD) score (calculated as 3.78×ln[bilirubin] + 11.2×ln[international normalized ratio] + 9.57×ln[creatinine] + 6.43×[0 for cholestatic/alcoholic, 1 for others]), PMAS, preoperative endotracheal intubation (Preop. ETI), and laboratory results (routine blood testing, biochemistry, and coagulation studies within 7 days before LT), as well as disease etiology (tumor vs benign disease). Benign diseases mainly included viral hepatitis and autoimmune conditions. Intraoperative data included cold ischemia time, total operating time, intraoperative red blood cell transfusion (yes/no), biliary reconstruction method (defined as choledochojejunostomy or T-tube drainage), and intraoperative blood loss volume. Postoperative follow-up data covered the period from the end of transplantation until June 30, 2024, or until death. Postoperative complications—including early allograft dysfunction (EAD), hepatic vascular thrombosis, biliary complications (stricture, leakage, or stones) [12,13], small-for-size graft syndrome, and early reoperation for bleeding (defined as reoperation within 7 days for active bleeding) – and complication grading according to the Clavien-Dindo classification system were documented in medical records and research software. Survival time was calculated in months [13,14].

Patients with intraoperative blood loss of at least 1000 mL were included in the EIBL group, and those with blood loss below 1000 mL were included in the non-EIBL group [9].

Because duration data were not available in retrospective records, pre-LT ICU stay was analyzed as a binary variable (yes/no) rather than as a continuous variable. Complications were graded using the Clavien-Dindo classification system [14]. Complications classified as Clavien-Dindo grade 3 or higher were considered severe; those classified as grade 2 or lower were considered less severe [15]. All patients received an intraoperative bolus of methylprednisolone 500 to 1000 mg. Postoperatively, maintenance immunosuppression consisted of tacrolimus or sirolimus combined with mycophenolate mofetil. Additional methylprednisolone 80 to 120 mg was administered according to individual clinical indications. For recipients with autoimmune liver disease, methylprednisolone was prescribed at 80 to 120 mg every 12 h and subsequently tapered to 4 mg once daily. Target trough whole-blood tacrolimus concentrations were maintained between 5 and 10 ng/mL.

STATISTICAL ANALYSIS:

Baseline characteristics, including categorical and continuous variables and perioperative variables, are expressed as frequencies, percentages, or medians and interquartile ranges (IQRs), as appropriate. The t-test or Mann-Whitney U test was used for intergroup comparisons of continuous variables, and the Pearson chi-square test was used for intergroup comparisons of categorical variables. Spearman correlation analysis was performed to assess relationships of EIBL with postoperative complications. After univariate logistic regression analysis, only variables with P values <0.1 were included in multivariate analysis. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Kaplan-Meier curves were used to evaluate postoperative survival in the EIBL and non-EIBL groups. All statistical tests were 2-sided, and P values <0.05 indicated statistical significance. Patient survival was defined as the interval from the date of transplantation to the date of death or last follow-up. All statistical and graphical analyses were performed using SPSS 27.0 software.

Results

PREOPERATIVE PATIENT CHARACTERISTICS:

Follow-up was conducted quarterly from January 2015 to June 2024 and included blood tests, assessments of liver and kidney function, coagulation testing, abdominal ultrasound, and computed tomography scans. Any abnormalities were documented in the electronic medical record system. Among the 822 patients, 420 (51.1%) had intraoperative blood loss of at least 1000 mL, and 402 (48.9%) had blood loss below 1000 mL. The median donor age was 48 years (IQR: 39–56), and the median recipient age was 50 years (IQR: 42–57). Statistical analysis revealed no significant differences in donor or recipient age between the EIBL and non-EIBL groups (P=0.640). Among the 101 patients (12.3%) who required pre-LT ICU admission, the rate was higher in the EIBL group (71/420, 16.9%) than in the non-EIBL group (30/402, 7.5%; P<0.001). Portal hypertension, defined as the presence of esophageal varices, was identified in 469 patients (57.1%). Details are provided in Table 1.

RISK FACTORS FOR EIBL:

Univariate logistic regression analysis incorporating clinical and pathological variables showed that benign disease, PMAS, preoperative albumin below 38 g/L, Preop. ETI, pre-LT ICU stay, and a MELD score of 15 or higher were associated with EIBL. After multivariate adjustment, benign disease (P<0.001), PMAS (P<0.001), preoperative albumin below 38 g/L (P<0.001), and pre-LT ICU stay (P<0.001) remained independent risk factors for EIBL. ORs for benign disease, PMAS, preoperative albumin below 38 g/L, and pre-LT ICU stay were 1.660 (95% CI 1.231–2.240), 2.365 (95% CI 1.433–3.900), 1.938 (95% CI 1.425–2.635), and 2.255 (95% CI 1.382–3.678), respectively. Although a MELD score of 15 or higher was significant according to univariate analysis, its predictive value was not confirmed in the multivariate model. Results are summarized in Table 2.

RELATIONSHIP BETWEEN EIBL AND POSTOPERATIVE COMPLICATIONS:

To assess the effect of EIBL on postoperative complication incidence, the Pearson chi-square test was used to compare EIBL and non-EIBL groups. The overall incidence of postoperative biliary complications was 14.8%. In the EIBL group, 74 patients (17.6%) developed biliary complications, whereas 48 patients (11.9%) in the non-EIBL group were affected (P=0.022). Additionally, 9.9% of patients experienced Clavien-Dindo grade 3 or higher complications: 51 patients (12.1%) in the EIBL group and 30 patients (7.5%) in the non-EIBL group (P=0.024). Details are provided in Table 3.

UNIVARIATE AND MULTIVARIATE ANALYSES OF RISK FACTORS FOR POSTOPERATIVE BILIARY COMPLICATIONS:

Univariate logistic regression analysis of clinical, pathological, intraoperative, and postoperative variables showed that recipient age 60 years or older, MELD score of 15 or greater, EIBL, EAD, donation after circulatory death (DCD), and biliary reconstruction were associated with biliary complications. After multivariate adjustment, recipient age 60 years or older (P=0.027), EIBL (P=0.027), and DCD (P=0.005) remained independent risk factors. Corresponding ORs were 1.685 (95% CI 1.061–2.675), 1.571 (95% CI 1.054–2.341), and 2.165 (95% CI 1.270–3.690), respectively. Although significant according to univariate analysis, biliary reconstruction, EAD, and MELD score of 15 or greater did not retain predictive value in the multivariate model. Results are summarized in Table 4.

UNIVARIATE AND MULTIVARIATE ANALYSES OF CLAVIEN-DINDO GRADE 3 OR HIGHER COMPLICATIONS:

Multivariate logistic regression analysis demonstrated that Clavien-Dindo grade 3 or higher complications were significantly associated with an ASA score of 3 or greater (P=0.033, OR 2.080, 95% CI 1.061–4.077), EIBL (P=0.042, OR 1.790, 95% CI 1.021–3.138), and Preop. ETI (P=0.013, OR 4.204, 95% CI 1.349–13.105). Although cold ischemia time of 8 h, Child-Pugh score, intraoperative red blood cell transfusion, EAD, and MELD score of 15 or greater were significant according to univariate analysis, they did not demonstrate predictive value in the multivariate model. Donor age 60 years or older, included in the multivariate model based on a previous report [16], also did not demonstrate predictive value. Results are shown in Table 5.

KAPLAN-MEIER ANALYSIS OF 5-YEAR SURVIVAL RATE:

The Kaplan-Meier method was used to analyze the 5-year postoperative survival rate, comparing EIBL and non-EIBL groups. The EIBL group demonstrated a significantly lower survival rate than the non-EIBL group (P=0.010). Details are presented in Figure 1.

Discussion

Despite considerable advances in surgical technique and anesthetic management, EIBL during LT remains a major concern. In our multicenter cohort, EIBL (≥1000 mL) was associated with increased risks of biliary complications and Clavien-Dindo grade 3 or higher complications. These associations persisted after multivariate adjustment, underscoring the clinical importance of minimizing blood loss during LT [17].

In the present study, patients were divided into malignant and benign disease groups on the basis of etiology; intraoperative blood loss was significantly greater in the benign disease group than in the malignant disease group. This finding is consistent with the results of Tripodi and Mannucci, who reported that benign disease is associated with higher bleeding risk [18]. Consistent with the baseline differences we observed, recipients with benign disease demonstrated features of more advanced hepatic dysfunction (lower values for hemoglobin, platelets, and albumin; higher values for international normalized ratio, total bilirubin, and MELD score). Although a MELD score of 15 or greater was associated with EIBL according to univariate analysis, it did not remain an independent predictor after adjustment, suggesting that the MELD score primarily reflects global disease severity rather than serving as a direct determinant of bleeding in this cohort. Serum albumin is a biomarker commonly referenced when predicting adverse outcomes after liver resection, vascular surgery, and trauma [19]. Our results showed that a preoperative albumin concentration below 38 g/L was an independent risk factor for EIBL. This finding aligns with evidence that hypoalbuminemia impairs platelet function and endothelial stability, thereby increasing the risk of surgical bleeding [20]. Accordingly, preoperative optimization – nutritional support, control of infection or inflammation, and targeted correction of hypoalbuminemia within patient blood management pathways – may help reduce EIBL risk [17,21]. PMAS can result in adhesions between intra-abdominal organs and vessels, increasing operative difficulty and duration and contributing to EIBL risk. Multiple studies have indicated that prior abdominal surgery is a risk factor for major blood loss during LT [4,22]. In our multivariate model, PMAS independently predicted EIBL, consistent with reports that adhesions prolong dissection and increase bleeding risk. Pre-LT ICU stay also remained independently associated with EIBL, reflecting a higher systemic illness burden [23].

EIBL during LT has been significantly associated with an increased risk of postoperative complications [5,8]. Our comparison of postoperative complication rates between the EIBL and non-EIBL groups revealed that the incidence of biliary complications was significantly higher in the EIBL group (17.6%) than in the non-EIBL group (11.9%). Multivariate analysis confirmed EIBL as an independent risk factor for biliary complications. This may be attributable to reduced effective circulating blood volume and vascular collapse during EIBL, which can impair oxygen and nutrient delivery to the sinusoids via distal branches of the hepatic artery [24], resulting in ischemic injury to the bile ducts. This mechanistic pathway aligns with ischemia–reperfusion injury to the peribiliary plexus and the known vulnerability of the biliary tree. Compared with younger individuals, older recipients exhibit slower recovery and a higher likelihood of postoperative complications due to age-related functional decline [25]. Our study demonstrated that recipient age 60 years or older and DCD are independent risk factors for biliary complications. Although biliary reconstruction technique did not demonstrate statistical significance in the multivariate model, previous studies have indicated that arterial complications and suture or material factors may influence anastomotic outcomes [26]. Postoperative thrombosis, inflammation, impaired immunity, surgical technique, and direct bile duct injury also represent important risk factors for biliary complications [24,27,28]. Collectively, these findings support prioritizing EIBL reduction and the meticulous protection of arterial and biliary perfusion, particularly in older recipients and those receiving DCD grafts.

This study revealed a significant association between EIBL and the risk of severe postoperative complications (Clavien-Dindo grade ≥3) after LT. The incidence of severe complications was higher in the EIBL group (12.1%) than in the non-EIBL group (7.5%). On multivariate analysis, EIBL remained an independent predictor of Clavien-Dindo grade 3 or higher complications (P=0.042). An ASA score of 3 or greater (P=0.033) and Preop. ETI (P=0.013) were also independent risk factors, whereas cold ischemia time of 8 h or longer (P=0.056) and DCD (P=0.071) showed borderline associations. These findings indicate that both preoperative condition and intraoperative hemodynamic stability influence postoperative outcomes [8]. Hypotension secondary to massive intraoperative blood loss may impair perfusion to all organs, leading to microcirculatory disturbances and increasing the risk of postoperative complications [5,8]. These observations are consistent with patient blood management–oriented evidence that intraoperative coagulopathy, hypoperfusion, and non-targeted transfusion can worsen outcomes, reinforcing the importance of restrictive transfusion strategies and real-time coagulation monitoring [21,29]. Kaplan-Meier analysis also showed significantly lower 5-year survival in the EIBL group, emphasizing the long-term prognostic relevance of controlling intraoperative bleeding [30]. Although DCD and prolonged cold ischemia time showed trends toward increased complications, neither displayed statistical significance in the multivariate model (Table 5).

In summary, the present study highlights the importance of EIBL as a predictor of postoperative complications after LT and demonstrates its adverse impact on patient outcomes. Clinically, these findings support early identification and optimization of modifiable risk factors (eg, hypoalbuminemia, prior abdominal surgery, pre-LT ICU status) and the use of standardized patient blood management bundles to reduce EIBL and improve short- and long-term outcomes [17,21]. Future research should explore strategies to prevent EIBL and develop methods to reduce biliary and other severe complications, with the aim of improving long-term outcomes in LT recipients.

Several limitations of this study should be noted. Due to the retrospective nature of this multicenter analysis and potential biases in data inclusion, the findings may be subject to selection bias. Additionally, several critical perioperative variables – such as detailed use of vasoactive agents, renal replacement therapy, and comprehensive information regarding biliary reconstruction – could not be fully standardized across centers. We reported the extent of data availability transparently and adjusted for key covariates when feasible. Variations in intraoperative coagulation management, surgical technique, and immunosuppressive protocols may also have influenced the outcomes; however, these factors could not be fully controlled in the present dataset. Consequently, the mechanistic association between EIBL and postoperative complications should be viewed as hypothesis-generating; it warrants validation in prospective, patient blood management–oriented studies.

Conclusions

The present results underscore the importance of EIBL in predicting postoperative complications after LT and demonstrate its adverse impact on recipient outcomes. Future studies should focus on strategies to reduce EIBL, as well as methods to prevent and manage biliary and other severe complications, with the goal of improving long-term outcomes for LT recipients.

Data Availability

The data supporting the findings of this study are not publicly available due to privacy and ethical restrictions. Requests for access to these data should be directed to the corresponding author ([email protected]) and will be considered on a case-by-case basis.