Significance of lactate clearance in septic shock patients with high bilirubin levels

Lactate clearance is affected by hepatic function. However, it is unclear whether the association between hepatic dysfunction and lactate clearance can act as a prognostic marker of clinical outcomes in patients with septic shock. We aimed to evaluate the association between lactate clearance and mortality in two cohorts of septic shock patient who had hepatic dysfunction based on their total serum bilirubin levels (TBIL). Lactate clearance at 24 h after the onset of septic shock was analyzed using two cohorts, sub-categorized into two groups based on TBIL: < 2 mg/dL and ≥ 2 mg/dL. In the derivation cohort, lactate clearance was lower in non-survivors than in survivors with TBIL ≥ 2 mg/dL, while there was no significant difference in lactate clearance between non-survivors and survivors with TBIL < 2 mg/dL. Multivariate logistic regression analysis revealed that increased lactate clearance was significantly associated with decreased 28-day mortality in the TBIL ≥ 2 mg/dL group (10% lactate clearance, adjusted odds ratio [OR]: 0.88, 95% confidence interval (CI): 0.80–0.97, P = 0.0075), Creatinine level ≥ 2 mg/dL group (adjusted OR: 0.88, 95% CI: 0.81–0.95, P = 0.00069) and APACHE II score ≥ 35 group (adjusted OR: 0.93, 95% CI: 0.87–0.98, P = 0.013). In the validation cohort, lactate clearance was lower in non-survivors than in survivors with TBIL ≥ 2 mg/dL, while no significant difference in lactate clearance was observed between non-survivors and survivors with TBIL < 2 mg/dL. Increased lactate clearance was significantly associated with decreased 28-day mortality in the TBIL ≥ 2 mg/dL group (10% lactate clearance, adjusted OR: 0.89, 95% CI: 0.83–0.96, P = 0.0038) and the association was just about significant in APACHE II score ≥ 35 group (adjusted OR: 0.86, 95% CI: 0.74–1.00, P = 0.051). In conclusion, increased lactate clearance in septic shock patients with hepatic dysfunction (TBIL ≥ 2 mg/dL) or high severity (APACHE II score ≥ 35) was associated with decreased 28-day mortality.

Blood lactate levels potentially reflect the imbalance between oxygen delivery and consumption during global tissue hypoxia, which reduces the availability of pyruvate for the tricarboxylic acid (TCA) cycle and accelerates aerobic glycolysis as a result of excess beta-adrenergic stimulation 1,2 . Blood lactate level is the cornerstone of diagnosis and management of patients with septic shock [3][4][5][6] . Septic shock patients with increased blood lactate levels have poor clinical outcomes; therefore, persistently elevated lactate due to decreased lactate clearance may be a danger signal for the management of septic shock [7][8][9][10] .
Blood lactate level indicates the balance between lactate production and lactate clearance 11,12 . Hepatic function is key to eliminating lactate, while hepatic dysfunction potentially decreases lactate clearance 13,14 . Lactate clearance has been extensively investigated in septic shock [7][8][9]15,16 ; however, it remains unknown whether the association between lactate clearance and clinical outcomes in patients with septic shock differs between patients with and without hepatic dysfunction.
Thus, we hypothesized that hepatic function would influence the relationship between lactate clearance and mortality in septic shock. We investigated the association between lactate clearance and 28-day mortality in two cohorts of septic shock patients who were sub-divided into two groups based on blood total bilirubin levels (TBIL) indicating hepatic dysfunction. We also determined changes in the anion gap and base excess, which directly reflect the amount of nonvolatile acid produced during tissue hypoxia and may be affected by hepatic function.

Materials and methods
Study design, definition, and patients. This was a retrospective observational study. Septic shock was characterized by the presence of systemic inflammatory response syndrome caused by infection 17 , at least one new organ dysfunction according to the Brussels criteria, and persistent hypotension despite adequate fluid resuscitation 18 . This study utilized the previous definition of septic shock (Sepsis-2) because the current one (Sepsis-3) includes hepatic dysfunction assessed by the Sequential Organ Failure Assessment (SOFA) score, and one of the cohorts we used for validation was assessed using the Sepsis-2 definition. According to these criteria, hepatic dysfunction was defined as TBIL ≥ 2 mg/dL 19 . We selected TBIL for the assessment of hepatic dysfunction as opposed to albumin levels and coagulation tests because they can be affected by nutrition and coagulation disorders in patients with septic shock. The need to obtain informed consent from the patients was explicitly waived by the institutional review board which approved this study's protocols.
Derivation cohort: CHIBA cohort. We screened total 9290 patients who were admitted to the intensive care unit (ICU) at Chiba University Hospital (CHIBA), Chiba, Japan between October 2012 and September 2018. The ICU had 22 beds and was utilized by patients following emergency room admission and after elective surgery, accounting for approximately 80% of ICU beds. Of 9290 patients, 230 were diagnosed with septic shock on ICU admission and had blood lactate clearance, delta anion gap, and delta base excess data available; these patients were subsequently included in the analysis. This study was approved by the Institutional Review Board of Chiba University Graduate School of Medicine and was performed in accordance with the committee's guidelines; the requirement to obtain informed consent was waived. Measurements and lactate clearance. Blood lactate level, base excess, and anion gap were measured on the day patients presented with septic shock (day 1) as well as 24 h later (day 2). Blood lactate clearance (%) was calculated using the following formula 7-9 : where Lactate day1 = lactate level at the diagnosis of septic shock and Lactate day2 = lactate level 24 h after the diagnosis of septic shock.
The anion gap was calculated using the concentrations of sodium, chloride, and bicarbonate ions and was available only in the derivation cohort. Delta base excess and anion gap were calculated by subtracting day 2 values from day 1 values. Statistical analyses. The primary outcome variable was lactate clearance. The secondary outcome variables were delta base excess and delta anion gap. Patients were categorized into two groups based on their TBIL on day 1: < 2 mg/dL and ≥ 2 mg/dL.
We used logistic regression analysis to test the association between lactate clearance and 28-day mortality in patients with TBIL < 2 mg/dL or ≥ 2 mg/dL. Multivariate analysis was performed to adjust for potential baseline imbalances (age, sex, and Acute Physiology and Chronic Health Evaluation II (APACHE II) score), as reported previously 9,20-22 . Creatinine (Cre) level (< and ≥ 2 mg/dL), platelet count (< and ≥ 80 × 10 3 /mm 3 ), and APACHE II score (< and ≥ 35) were also analyzed to evaluate the association between severity and organ dysfunction. Cutoff values were derived according to the Brussels criteria and using the Youden's index. The correlation between lactate clearance and delta anion gap was analyzed using the Pearson correlation coefficient. Different TBIL cutoff values of 1 mg/dL, 3 mg/dL, and 4 mg/dL were tested using multivariate analysis to determine the appropriate threshold. The cutoff value of lactate clearance was determined to be 10% using the derivation cohort as per the Youden's index.
Univariate analysis was performed using the Mann-Whitney U test. A two-tailed P value of < 0.05 was considered statistically significant. Data are presented as medians with interquartile range. All analyses were performed using R version 3.3.3 (R Foundation for Statistical Computing, Vienna, Austria, http://www.R-rojec t.org/) and PRISM version 8 (GraphPad Software Inc, California, USA).

Ethics approval and consent to participate. This study was approved by the Institutional Review
Board of Chiba University Graduate School of Medicine, the Institutional Review Board of St. Paul's Hospital and the University of British Columbia. The requirement to obtain written informed consent was waived because of the study design.
Multivariate logistic regression analysis with adjustments for baseline imbalances revealed that patients with increased lactate clearance had significantly decreased 28-day mortality in the TBIL ≥ 2 mg/dL group (10% lactate clearance, adjusted odds ratio [OR]: 0.88, 95% confidence interval [CI]: 0.80-0.97, P = 0.0075, absolute risk difference: 25%, 95% CI: 4-46%). However, lactate clearance was not significantly associated with altered 28-day mortality in the TBIL < 2 mg/dL group in the derivation cohort (10% lactate clearance, adjusted OR: 0.99, 95% CI: 0.93-1.05, P = 0.69) (Table 3A). Similar to lactate clearance, patients with an increased delta anion gap had significantly decreased 28-day mortality in the TBIL ≥ 2 mg/dL group (delta anion gap, adjusted OR: 1.09, 95% CI: 1.00-1.19, P = 0.045). However, delta anion gap was not significantly associated with altered 28-day mortality in the TBIL < 2 mg/dL group (delta anion gap, adjusted OR: 0.96, 95% CI: 0.83-1.11, P = 0.58) (Table 3B). Table 1. Baseline characteristics, vital signs, blood gas analysis and laboratory data in the derivation cohort (CHIBA cohort, n = 230). Median (inter quartile range). P values were calculated using Pearson's chi-square test and Mann-Whitney U test. a APACHE, acute physiology and chronic health evaluation. b Delta lactate was calculated by subtracting lactate level on day2 from day1. c Lactate clearance was calculated by delta lactate divided by day1 value. d Delta base excess was calculated by subtracting base excess on day1 from day2. e Delta anion gap was calculated by subtracting anion gap on day1 from day2. www.nature.com/scientificreports/ Delta base excess was also not significantly associated with altered 28-day mortality in the multivariate logistic regression analysis following the same adjustments for baseline imbalances (delta base excess, TBIL < 2 mg/dL, In the Cre ≥ 2 mg/dL group, non-survivors had significantly lower lactate clearance compared to survivors (P = 0.0006). However, there was no significant difference in lactate clearance between survivors and non-survivors in the Cre < 2 mg/dL group (P = 0.71) (eFigure 2A, Supplemental Digital Content). Non-survivors had significantly lower lactate clearance compared to survivors (P = 0.03), in the APACHE II score ≥ 35 group, while there was no significant difference in lactate clearance between survivors and non-survivors in the APACHE II < 35 group (P = 0.84) (eFigure 2B, Supplemental Digital Content).
We then analyzed lactate clearance in the TBIL ≥ 2 mg/dL group using the validation cohort (Table 2). In terms of baseline characteristics, patients in the TBIL ≥ 2 mg/dL group were younger and had an increased probability of chronic hepatic failure compared to those in the TBIL < 2 mg/dL group. Similar to the derivation cohort, non-survivors had significantly lower lactate clearance compared to survivors in the TBIL ≥ 2 mg/dL group (P = 0.0006), while no significant difference in lactate clearance was observed between non-survivors and survivors in the TBIL < 2 mg/dL group (P = 0.46) (Fig. 2). Furthermore, patients with increased lactate clearance had significantly decreased 28-day mortality in the TBIL ≥ 2 mg/dL group (10% lactate clearance, adjusted OR: 0.89, 95% CI: 0.83-0.96, P = 0.0038, absolute risk difference: 20%, 95% CI: 2-37%). However, in the TBIL < 2 mg/ dL group, lactate clearance was not significantly associated with altered 28-day mortality in the validation cohort (Table 3C).
Non-survivors had significantly lower lactate clearance compared to survivors in the Cre ≥ 2 mg/dL group (P = 0.049), while no significant difference in lactate clearance was observed in the Cre < 2 mg/dL group (P = 0.12) (eFigure 3A, Supplemental Digital Content). In the APACHE II score ≥ 35 group, non-survivors had significantly lower lactate clearance compared to survivors (P = 0.02). However, there was no significant difference in lactate In both the Cre ≥ 2 mg/dL and APACHE II score ≥ 35 groups, lactate clearance was not significantly associated with altered 28-day mortality in the validation cohort (eTable 3, Supplemental Digital Content).

Discussion
In this study, we investigated lactate clearance in septic shock patients, stratified according to hepatic dysfunction. The results of this study showed that increased lactate clearance during the initial 24 h was significantly associated with decreased mortality in septic shock patients with TBIL ≥ 2 mg/dL. In contrast, lactate clearance was not associated with altered mortality in patients with TBIL < 2 mg/dL. Lactate clearance has previously been investigated as a potential prognostic marker for sepsis/septic shock 7,9,15,20 . However, increased lactate clearance is not always associated with decreased mortality in sepsis/ septic shock. In this study including two septic shock cohorts, the association between increased lactate clearance and decreased 28-day mortality was significant in the TBIL ≥ 2 mg/dL group but not in the TBIL group < 2 mg/ dL; this may explain the inconsistent conclusions of previous studies in mixed patient populations concerning hepatic dysfunction. The present study highlights the importance of hepatic dysfunction when considering lactate clearance as a prognostic marker.
Several studies have investigated lactate kinetics [23][24][25][26][27] , which is a complicated topic, and blood lactate level depends on the rate of production and removal of lactate from the blood as well as cell metabolism. Since sustained hyperlactatemia or even a rise in lactate levels could potentially reflect decreased lactate clearance rather than an increase in lactate production in septic shock patients, hepatic function could be important for the elimination of lactate 28 . A study using intravenous infusion of sodium lactate revealed that the clearance of lactate from the blood in normal subjects in the postprandial state was 17.9 mL/kg/min 23 ; it decreased to 14.5 mL/kg/ min in patients with hepatic dysfunction 24 . Additionally, due to impaired hepatic lactate uptake, lactate half-life was significantly longer in patients with than that in patients without liver cirrhosis (18.8 min vs. 14.7 min) 25 . Similarly, in severe sepsis/septic shock, hepatic dysfunction was significantly associated with impaired lactate clearance and normalization during early quantitative resuscitations 13 . Thus, hepatic dysfunction decreased lactate elimination. In cases of insufficient lactate elimination, lactate generation has a large effect on blood lactate level, potentially explaining the significance of lactate clearance in hepatic dysfunction.
In this study, delta base excess, which reflects the bicarbonate level at the end of lactate metabolism, was not associated with 28-day mortality in both the TBIL ≥ 2 mg/dL and TBIL < 2 mg/dL groups. High base excess reflects increased levels of non-volatile acids, including lactate, in sepsis 29 , whereas low base excess predicts increased lactate levels 30,31 . However, base excess is affected by other abnormalities such as metabolic acidosis (ketoacidosis, renal tubular acidosis, and uremia), acute respiratory acidosis, and changes in hemoglobin levels, all of which can occur in critically ill patients 32 . The application of the change in base excess as a prognostic marker for septic shock may be affected by these complex acidosis mechanisms.
Conversely, delta anion gap was significantly associated with altered mortality and lactate clearance. The anion gap, which simply reveals ion balance, reflects the level of nonvolatile acid more directly compared to the base excess, calculated using pH and bicarbonate level 33,34 . This explains the significant correlation between lactate clearance and delta anion gap, which in turn indicates the significant association between delta anion gap and altered 28-day mortality because lactate clearance is significant associated with 28-day mortality.
In the analysis of the derivation cohort, increased lactate clearance was also associated with decreased 28-day mortality in both Cre ≥ 2 mg/d group and APACHE II score ≥ 35 group. In the validation cohort, the association was just about significant in APACHE II score ≥ 35 group. Furthermore, the point estimate for 28-day mortality in lactate clearance in APACHE II score ≥ 35 group (0.86) was lower than in the TBIL ≥ 2 mg/dL group (0.89). It may indicate the lactate clearance being beneficial also in patients with severe disease, while no previous study showed benefits of lactate clearance as a prognostic value for mortality in patients with higher severity of sepsis or septic shock. Further studies of larger scale will be required to focus on the association between severities and lactate clearance in septic shock patients.
Our study has few limitations. First, it was a retrospective study; however, hepatic dysfunction had the same effect on both septic shock cohorts. Second, we defined hepatic dysfunction as TBIL ≥ 2 mg/dL based on the Brussels criteria; however, there are different published criteria or cutoff values for defining hepatic dysfunction. Finally, lactate clearance is associated with survival in not only patients with increased bilirubin levels but also in those with high APACHE II score and increased creatinine levels in the derivation cohort, which shows that increased lactate clearance is beneficial in all patients with increased severity and evidence of organ dysfunction.

Conclusions
In septic shock patients with hepatic dysfunction (TBIL ≥ 2 mg/dL) or high APACHE II score (≥ 35), increased lactate clearance was correlated with decreased 28-day mortality. The findings of this study highlight the importance of hepatic dysfunction and severity in lactate clearance when considering lactate clearance as a clinical parameter.

Data availability
The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.