INTRODUCTION
Serum lactate level is determined by lactate production and elimination, both of which are equivalent under homeostatic conditions [1]. Hyperlactatemia can occur in conditions of excess lactate production or diminished lactate clearance. Increased production of lactate may be caused by anaerobic metabolism due to tissue hypoxia, aerobic glycolysis from sepsis, trauma, or hemorrhagic shock, enzymatic alterations in pyruvate dehydrogenase, mitochondrial dysfunction, and pharmacologic effects such as metformin, propofol, and linezolid [1-5]. Hyperlactatemia has been reported to be associated with mortality in patients with septic shock and patients who underwent major abdominal surgery or cardiac surgery [6-10].
The liver plays a crucial role in lactate metabolism. Lactic acid produced by skeletal muscle and red blood cells is released into the bloodstream and then metabolized by the liver and kidneys, of which 70% is metabolized by the liver [11,12]. Lactate is taken up into hepatocytes via the monocarboxylate transporter and converted to pyruvate by lactate dehydrogenase, which then produces glucose and energy through the Cori cycle and the Krebs cycle [13]. However, the liver can also become a producer of lactate in cases of hepatic parenchymal hypoxia [14]. Therefore, patients with acute liver failure or chronic liver disease can cause excessive production and reduced elimination of lactic acid, resulting in the elevation of lactate levels.
Liver transplantation (LT) surgery is characteristically associated with elevated serum lactate levels. It is known that lactate concentration increases from the beginning of surgery through the anhepatic phase, peaks after reperfusion, and then gradually declines [15]. Several studies have identified the association between postoperative hyperlactatemia and poor outcomes including primary graft function and mortality [16-18]. Although one study investigates the relationship between intraoperative lactate levels and short-term mortality in LT [19], there are no studies that focus on the relationship between intraoperative hyperlactatemia and long-term mortality. Similarly, lactate clearance after reperfusion has been correlated with post-LT graft function and short-term mortality [15,20,21], but the association of post-reperfusion lactate clearance and long-term mortality has yet to be explored. Thus, the primary objective of this study is to investigate the relationship between intraoperative lactate levels and long-term mortality after LT. We additionally evaluated whether lactate clearance can be a prognostic factor for long-term mortality.
MATERIALS AND METHODS
Study design and patient populations
We conducted a single-center retrospective observational study. All patients aged ≥ 18-years who underwent either living donor or deceased donor LT at University of Pittsburgh Medical Center between January 2012 and June 2023 were included in this study. All liver transplant surgeries during the study period were performed by 8 experienced attending transplant surgeons following standardized surgical techniques at our center. Patients who underwent simultaneous transplantation, died intraoperatively, had chronic kidney disease dependent on hemodialysis, or had missing data were excluded from this study. This study was approved by the Institutional Review Board (IRB) of University of Pittsburgh (STUDY20050148) and conducted in accordance with the Declaration of Helsinki. The requirement for written informed consent was waived by the IRB owing to the retrospective design of the study. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement of reporting observational studies [22].
The following data were collected for all subjects: demographics including age, sex, and body mass index (BMI); presence of hypertension or diabetes mellitus (DM); etiology of liver failure, preoperative laboratory findings within 1 week prior to LT; and model for end-stage liver disease (MELD) score. Data for the duration of procedure, volume of infusion, and blood loss were recorded as intraoperative factors. We also investigated allograft information including donor age, cold ischemic time, and warm ischemic time. Intraoperative lactate values were collected, analyzed, and reported from the arterial blood gas analysis performed intraoperatively. Intraoperative lactate measurements were used to calculate the time-weighted average lactate (TWAL) level for each patient. TWAL was calculated as the area under the curve (AUC) divided by the time interval between the first and the last measurement (Fig. 1), using a trapezoidal rule as previously described in studies of critically ill patients [23,24]. This approach reflects lactate dynamics over time rather than relying on a single measurement [23,24]. Lactate clearance in this study was defined as the absolute difference between lactate at reperfusion and lactate at the end of surgery (Lactate [reperfusion]-Lactate [end of surgery]; mM). We applied receiver operating characteristics (ROC) curve analysis for 1-year mortality and selected the optimal cut-off TWAL value using the Youden index. The patients were divided into two groups based on the optimal cut-off TWAL value; lower lactate (LL) group included patients who had lower TWAL than the optimal cut-off TWAL value, and higher lactate (HL) group included patients who had higher TWAL.
Outcome
The primary outcome was mortality at 30-days, 1-year, and 3-years after LT. Secondary outcomes included ventilation time, intensive care unit (ICU) and hospital length of stay (LOS), early allograft dysfunction (EAD), and acute kidney injury (AKI). AKI was defined according to the Kidney Disease Improving Global Outcomes definition [25]: an increase in serum creatinine of 0.3 mg/dl within 48 h or a rise 1.5 times baseline or more within 7 days. We set the baseline creatinine as the most recent serum creatinine level within 7 days prior to surgery. EAD was defined as the presence of one or more of the following variables: bilirubin ≥ 10 mg/dl on postoperative day (POD) 7, or International Normalized Ratio ≥ 1.6 on POD 7 according to the Olthoff criteria [26].
Statistical analysis
Continuous variables are reported as median with interquartile range (IQR) and were compared using the Student t-test or the Mann-Whitney U test, as appropriate. Categorical data are reported as numbers (%) and were compared using the chi-square test. ROC curves were constructed to evaluate TWAL as a predictor of 1-year mortality. The optimal cut-off point was identified using the Youden index. Kaplan-Meier analyses were applied for mortality at 30-days, 1-year, and 3-years, and the Log rank test was evaluated to compare between LL group and HL group. Multivariable Cox hazard models were used to estimate the adjusted association between TWAL and mortality at 30-days, 1-year, and 3-years. Multivariable analysis was performed with known risk factors for post-LT mortality, including age, sex, BMI, DM, MELD score, and intraoperative blood loss [27-32]. Furthermore, lactate clearance was compared between survivors and non-survivors for each mortality timepoint using the Mann-Whitney U test. To identify the adjusted association between lactate clearance and each mortality, multivariable Cox hazard models were analyzed.
The results of the Cox proportional hazard model are expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). All P values were two-sided, and values < 0.05 were considered statistically significant. All analyses were conducted using StataSE version.17.0 (StataCorp LLC) and EZR (Saitama Medical Center, Jichi Medical University), a graphical user interface for R version 4.4.1 (R Foundation for Statistical Computing).
RESULTS
During the study period, 1,067 of the 1,145 patients who underwent LT were eligible for analysis (Supplementary Fig. 1). The median number of arterial blood gas analyses measured during the procedure was 10, and TWAL was calculated for each patient using all the collected data.
Determination of high and low lactate groups
Median TWAL was 2.9 (IQR 2.2, 4.0) mM, and intraoperative lactate levels increased from the beginning of surgery (1.2 [0.9, 1.6] mM), peaked at the time of reperfusion (4.3 [3.3, 5.8] mM), and then slowly decreased toward the end of surgery (3.1 [2.0, 5.4] mM). ROC curve analysis identified the optimal TWAL cut-off value for predicting 1-year mortality was 3.1 mM, with a sensitivity of 65.8%, specificity of 58.2%, and AUC of 0.64 (95% CI 0.57-0.71). According to the optimal cut-off TWAL value, patients were divided into LL group (≤ 3.1 mM, n = 603) and HL group (> 3.1 mM, n = 464).
Patient characteristics and intraoperative variables were summarized in Table 1. Recipient age and MELD scores were similar between the groups. Notably, the HL group had longer duration of procedure and greater intraoperative crystalloid administration compared to the LL group (P < 0.001 and P < 0.001, respectively).
Primary outcome
In our cohort, mortality at 30-days, 1-year, and 3-years was 2.3%, 7.1%, and 11.8%, respectively. Kaplan-Meier survival analysis for mortality at 30-days, 1-year, and 3-years demonstrated significantly worse survival in patients in the HL group compared to the LL group (P = 0.012, P < 0.001, and P < 0.001, respectively) (Fig. 2). Multivariable Cox hazard models identified that TWAL was independently associated with post-LT mortality at 30-days, 1-year, and 3-years (HR 1.41, 95% CI 1.19-1.67; P < 0.001, HR 1.29, 95% CI 1.17-1.42; P < 0.001, and HR 1.20, 95% CI 1.10-1.30; P < 0.001, respectively) (Fig. 3).
Fig. 4 shows the comparison of lactate clearance between survivors and non-survivors in each mortality time frame. Non-survivors had reduced lactate clearance compared to survivors at 30-days, 1-year, and 3-years (non-survivors vs. survivors: 0.2 [-2.2, 1.1] mM vs. 1.1 [0, 1.9] mM; P = 0.003, 0.5 [-0.9, 1.7] mM vs. 1.1 [0, 1.9] mM; P = 0.003, and 0.6 [-0.7, 1.7] mM vs. 1.2 [0.1, 1.9] mM; P = 0.002, respectively). In multivariable Cox hazard models, lactate clearance was also an independent predictor for mortality at 30-days, 1-year, and 3-years (HR 0.69, 95% CI 0.60-0.80; P < 0.001, HR 0.80, 95% CI 0.73-0.88; P < 0.001, and HR 0.83, 95% CI 0.77-0.90; P < 0.001, respectively) (Fig. 5).
Secondary outcomes
Secondary outcomes are shown in Table 2. The HL group showed significantly longer ventilation time as well as ICU and hospital LOS (HL vs. LL: 11 [3, 22] h vs. 8 [0, 15] h; P < 0.001, 3 [2, 4] days vs. 2 [2, 4] days; P = 0.019, and 10 [7, 21] days vs. 9 [7, 16] days; P = 0.048, respectively). The incidence of post-LT AKI was also significantly higher in the HL group compared to the LL group (HL vs. LL: 44.2% vs. 37.7%; P = 0.031), while the incidence of EAD was similar between the two groups (HL vs. LL: 19.5% vs. 16.1%; P = 0.150). Although TWAL was similar between patients with and without EAD (3.0 [2.3, 4.3] mM vs. 2.8 [2.2, 3.9] mM, P = 0.076), lactate clearance was worse in the patients with EAD compared to patients without EAD (0.8 [-0.5, 1.7] mM vs. 1.2 [0.1, 1.9] mM, P = 0.003).
DISCUSSION
In this study, we found that intraoperative lactate level was independently associated with post-LT short- and long-term mortality. Furthermore, we also found that post-reperfusion lactate clearance was a significant prognostic factor for post-LT mortality. To the best of our knowledge, the association of intraoperative lactate level with long-term mortality has not been previously reported. In addition, this was the first study to demonstrate the relationship between intraoperative lactate clearance and mortality in LT.
Since Broder and Weil [33] demonstrated the correlation between high lactate level and adverse outcomes in patients with septic shock in 1964, multiple studies have been performed to examine the association between lactate levels and poor outcomes in septic shock patients [10,34,35]. Accordingly, the first edition of the Surviving Sepsis Campaign guidelines included lactate measurement as a severity and prognostic assessment of therapeutic endpoints [36]. Afterwards, many studies have identified the correlation between lactate level and postoperative poor outcomes in various surgical fields including major abdominal surgery, cardiac surgery, and liver surgery [6-9]. In LT, Kim et al. [19] identified that intraoperative hyperlactatemia was associated with significantly higher post-LT mortality at 30-day and 90-day. However, they only measured lactate levels at 3 time points (preanhepatic, anhepatic, and neohepatic periods) during the LT surgery, and only the peak lactate level was used for analysis. Thus, there was no standardized timing of lactate measurement. Because transient hyperlactatemia does not necessarily predict a poor clinical outcome, repeat lactate measurements are suggested to evaluate prognosis [37]. In the present study, we recorded a median of 10 lactate levels per patient during the procedure and evaluated the intraoperative lactate level by means of the time-weighted average method, which is less affected by transient hyperlactatemia. Furthermore, we identified the association of lactate level with long-term mortality as well as short-term mortality.
In our cohort, a high lactate level was associated with longer surgical duration and higher infused crystalloid volume. Although prolonged surgical duration may suggest an extended anhepatic phase, detailed records of anhepatic time were not available in our dataset, and this association could not be directly verified. Prolonged anhepatic time is known to contribute to elevated lactate levels because of impaired lactate metabolism during this phase. The longer surgical duration may also be attributed to more complex procedures, which can result in unstable hemodynamics. It is possible that the higher amount of crystalloid administration is a reflection of intraoperative hypotension. Unstable hemodynamics can lead to tissue hypoxia, resulting in anaerobic metabolism and increased production of lactic acid [38]. Intraoperative hyperlactatemia is thought to be produced as a combined result of these factors and led to a robust association with postoperative mortality.
Lactate clearance calculated intraoperatively was also an independent predictor of post-LT mortality in the current study. Lactate clearance has been studied as a potential prognostic factor in the setting of LT [15,39,40]. The definition of lactate clearance varies among studies, and all studies examining the relationship between lactate clearance and post-LT mortality have been calculated using lactate level at reperfusion and lactate level measured at 6 h or later postoperatively [20,21]. However, we have shown an independent association between lactate clearance and mortality that can be calculated intraoperatively using lactate level at reperfusion and lactate level at the end of surgery. This may facilitate earlier prognostication and could inform therapeutic decision-making, such as fluid administration, although standardized intervention protocols specifically targeting lactate dynamics in LT have not been established.
Since baseline lactate levels can widely vary among individuals with cirrhosis, absolute lactate levels do not necessarily reflect postoperative graft function. However, lactate clearance after reperfusion reflects the function of the implanted graft [20]. In fact, in the present study, while absolute lactate levels were similar between patients with or without EAD, lactate clearance was diminished in the patients with EAD compared to patients without EAD.
Despite our key findings, this study has several limitations. First, this was a retrospective, single-center study with the associated risks of unintended bias. Second, we did not evaluate several important confounders that may influence lactate dynamics, including graft-related variables such as donation after circulatory death grafts and donor risk index, as well as intraoperative factors like hemodynamic parameters, postreperfusion syndrome, and vasopressor use. Furthermore, although LT surgeries were performed by 8 experienced attending transplant surgeons using standardized surgical techniques, the potential influence of surgeon variability cannot be completely excluded. These factors should be considered when interpreting our findings. Third, because the focus of this study was on intraoperative lactate, we did not evaluate postoperative lactate levels. Therefore, we could not compare intraoperative lactate levels with postoperative lactate levels as a prognostic tool. Fourth, we did not evaluate graft size in this study because records of graft weight were inconsistently documented. It has been reported that graft volume would affect lactate levels, especially during the neohepatic phase [15]. Fifth, this study did not include precise records of the anhepatic phase duration, which is a potential factor influencing intraoperative lactate levels. Therefore, we were unable to directly analyze its impact. Future prospective studies incorporating detailed intraoperative time logs are needed to verify this association. Finally, although high lactate levels and reduced lactate clearance were associated with poor prognosis in the current study, it is unclear whether therapeutic intervention against them will improve prognosis, particularly as there is limited evidence for lactate-guided resuscitation [41,42].
In conclusion, the present study demonstrates that intraoperative lactate level was independently associated with short- and long-term mortality after LT. Furthermore, lactate clearance calculated intraoperatively was an independent prognostic factor for post-LT mortality.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
