Hierarchical Capability in Distinguishing Severities of Sepsis via Serum Lactate: A Network Meta-Analysis

Background: Blood lactate is a potentially useful biomarker to predict the mortality and severity of sepsis. The purpose of this study is to systematically review the ability of lactate to predict hierarchical sepsis clinical outcomes and distinguish sepsis, severe sepsis and septic shock. Methods: We conducted an exhaustive search of the PubMed, Embase and Cochrane Library databases for studies published before 1 October 2022. Inclusion criteria mandated the presence of case–control, cohort studies and randomized controlled trials that established the association between before-treatment blood lactate levels and the mortality of individuals with sepsis, severe sepsis or septic shock. Data was analyzed using STATA Version 16.0. Results: A total of 127 studies, encompassing 107,445 patients, were ultimately incorporated into our analysis. Meta-analysis of blood lactate levels at varying thresholds revealed a statistically significant elevation in blood lactate levels predicting mortality (OR = 1.57, 95% CI 1.48–1.65, I2 = 92.8%, p < 0.00001). Blood lactate levels were significantly higher in non-survivors compared to survivors in sepsis patients (SMD = 0.77, 95% CI 0.74–0.79, I2 = 83.7%, p = 0.000). The prognostic utility of blood lactate in sepsis mortality was validated through hierarchical summary receiver operating characteristic curve (HSROC) analysis, yielding an area under the curve (AUC) of 0.72 (95% CI 0.68–0.76), accompanied by a summary sensitivity of 0.65 (95% CI 0.59–0.7) and a summary specificity of 0.7 (95% CI 0.64–0.75). Unfortunately, the network meta-analysis could not identify any significant differences in average blood lactate values’ assessments among sepsis, severe sepsis and septic shock patients. Conclusions: This meta-analysis demonstrated that high-level blood lactate was associated with a higher risk of sepsis mortality. Lactate has a relatively accurate predictive ability for the mortality risk of sepsis. However, the network analysis found that the levels of blood lactate were not effective in distinguishing between patients with sepsis, severe sepsis and septic shock.


Introduction
Sepsis, characterized by life-threatening organ dysfunction resulting from a dysregulated host response to infection [1], represents a significant global health concern, impacting millions of individuals worldwide.Several significant infectious conditions and inappropriate treatment may culminate in the progression to severe sepsis or septic shock [2].To enhance clinical outcomes in these patients, early identification of individuals at risk of mortality and timely optimization of clinical decision-making are of paramount importance [3].
Clinical scoring systems, such as Sequential Organ Failure Assessment (SOFA) (Supplementary Table S1) and systemic inflammatory response syndrome (SIRS) (Supplementary Table S2), have been proposed for predicting sepsis-related outcomes, including mortality [4].Nevertheless, SOFA entails a time-consuming process that necessitates multiple laboratory and clinical data inputs, while SIRS exhibits limited sensitivity in predicting mortality [5].In contrast, blood lactate levels offer a rapid and easily obtainable measurement, serving as a surrogate for tissue hypoperfusion in critically ill patients [6].Combining point-of-care lactate assessment with the qSOFA (Supplementary Table S3) score in rapid bedside assessments more accurately identifies the risk of sepsis-related mortality compared to using the qSOFA score alone [7].
Elevated blood lactate levels may result from tissue hypoxia and anaerobic metabolism; they can also develop in other ways (Figure 1) [8].Studies consistently demonstrate that the duration and severity of hyperlactatemia are directly correlated with mortality in septic shock patients.Elevated blood lactate levels serve as a valuable marker for assessing the severity of sepsis and predicting patient outcomes [9][10][11].This is attributed to their ability to accurately and promptly reflect the perfusion status of peripheral tissues in the body and their sensitivity in indicating the presence of cellular hypoxia or tissue hypoperfusion [12][13][14].However, there exists inconsistency in the reference criteria for predicting the prognosis of sepsis in cases of elevated lactate levels.While certain studies emphasized an initial lactate level greater than 2.0 mmol/L as indicative of elevated lactate levels [15,16], others suggested that elevated lactate levels are defined by an initial lactate level exceeding 4.0 mmol/L [17][18][19].Furthermore, there is no relevant quantitative metaanalysis focusing on different blood lactate level to predict different outcomes in sepsis, such as severe sepsis and septic shock.
Biomedicines 2024, 12, x FOR PEER REVIEW 2 of 32 mortality [5].In contrast, blood lactate levels offer a rapid and easily obtainable measurement, serving as a surrogate for tissue hypoperfusion in critically ill patients [6].Combining point-of-care lactate assessment with the qSOFA (Supplementary Table S3) score in rapid bedside assessments more accurately identifies the risk of sepsis-related mortality compared to using the qSOFA score alone [7].Elevated blood lactate levels may result from tissue hypoxia and anaerobic metabolism; they can also develop in other ways (Figure 1) [8].Studies consistently demonstrate that the duration and severity of hyperlactatemia are directly correlated with mortality in septic shock patients.Elevated blood lactate levels serve as a valuable marker for assessing the severity of sepsis and predicting patient outcomes [9][10][11].This is a ributed to their ability to accurately and promptly reflect the perfusion status of peripheral tissues in the body and their sensitivity in indicating the presence of cellular hypoxia or tissue hypoperfusion [12][13][14].However, there exists inconsistency in the reference criteria for predicting the prognosis of sepsis in cases of elevated lactate levels.While certain studies emphasized an initial lactate level greater than 2.0 mmol/L as indicative of elevated lactate levels [15,16], others suggested that elevated lactate levels are defined by an initial lactate level exceeding 4.0 mmol/L [17][18][19].Furthermore, there is no relevant quantitative metaanalysis focusing on different blood lactate level to predict different outcomes in sepsis, such as severe sepsis and septic shock.For this purpose, we conducted a systematic review and diagnostic systematic review to rigorously and quantitatively assess the accuracy of blood lactate levels in predicting sepsis mortality.Additionally, we performed a network meta-analysis to evaluate whether blood lactate can differentiate between sepsis, severe sepsis and septic shock.For this purpose, we conducted a systematic review and diagnostic systematic review to rigorously and quantitatively assess the accuracy of blood lactate levels in predicting sepsis mortality.Additionally, we performed a network meta-analysis to evaluate whether blood lactate can differentiate between sepsis, severe sepsis and septic shock.

Study Protocol
This analysis was conducted in accordance with a predetermined protocol following the recommendations of a guideline for systematic reviews of prognostic factor studies.The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) Extension Statement for Reporting of Systematic Reviews Incorporating Network Meta-analyses of Health Care Interventions was followed for data abstractions.Our study protocol was registered on PROSPERO (CRD449572).

Study Selection
Titles and abstracts of search results were screened independently (YL).The full texts of the remaining results were assessed independently by another two of us (BZ, RZ) for inclusion based on predetermined criteria.Any discrepancies were resolved through discussion, potentially with a third reviewer.We manually searched the reference lists of included studies and existing systematic reviews as well as all articles citing the included studies on Google Scholar.
In accordance with the objectives of our meta-analysis, we developed a 'Population, Index prognostic factor, Comparator prognostic factor, Outcome, Timing, Settings' (PICOTS) framework adapted from the guideline proposed by Riley et al. [20].Our study inclusion criteria were as follows according to PICOTS framework: (1) population: sepsis patients with a well-defined diagnostic reference standard for sepsis; (2) index prognostic factor: before-treatment blood lactate levels measured; (3) outcome: nonsepsis, sepsis, severe sepsis, sepsis shock and death; (4) if studies were based on overlapping patients, the most completed one was chosen; and (5) studies were restricted to English publications.We used the following criteria for study exclusion: (1) studies lacking relevant outcomes or lactate levels; and (2) conference abstracts, reviews, case reports, and experiment studies.

Data Collection and Assessment of Study Quality
The relevant articles and eligible data were assessed and extracted by two authors (BZ, RZ), respectively.If a disagreement occurred, it was discussed and the consensus with a third author was reached.The quality of evidence was assessed by the modified Grading of Recommendations Assessment, Development, and Evaluation system (GRADE) by consensus among the authors [21,22].
The following data were collected from each study: first author name, area, publication date, the type of studied design, number of patients, timing of lactate measurements and primary outcome (nonsepsis, sepsis, severe sepsis, sepsis shock and death).When an included study reported different cut-off values, we chose one which made both sensitivity and specificity more than 50% as possible.When an included study reported the same outcome at different follow-up timepoint (e.g., 7-day mortality and 30-day mortality), we chose the earliest one.If the included studies did not report the mean and standard deviation, estimates for these parameters were derived from the sample size, median and quartiles [23,24].Table 1 presents the baseline characteristics of included studies.Two independent reviewers (BZ, RZ) performed quality assessments of selected studies using the QUADAS-2 criteria [25].This checklist consists of four key domains: patient selection, index test, reference standard, and flow and timing.Within each study, the domains are assessed in terms of risk of bias and the first three of these domains are assessed in terms of concerns about applicability.Signaling questions as specified in the QUADAS-2 tool enable the reviewer to give each domain a rating of high, low or unclear.If the answers to all signaling questions for a domain are "yes", then risk of bias can be judged low.If any signaling question is answered "no", potential for bias exists.The "unclear" category should be used only when insufficient data are reported to permit a judgment.When there was a disagreement, the third author made the final decision based on the criteria.Details of the QUADAS-2 criteria are elaborated on Table 2.  Using the Newcastle-Ottawa Scale (NOS) [26] for cohort studies, the risk of bias was assessed for each outcome in all included studies.According to the selection of cohort (up to four points), the comparability of cohort design and analysis (up to two points) and the adequacy of result measurement (up to three points), a maximum of nine points will be obtained.Seven to nine points are considered high quality (low risk of bias) (Table 3).
Each ⋆ indicates one score received in the according section.Superscript numbers are used to distinguish between two different studies by the same author.

Statistical Analysis
The meta-analysis used the combined effects of each result.We calculated odds ratios (ORs) and 95% confidence intervals (CIs) for each result using a random effects model, the between-study heterogeneity was evaluated by the χ 2 -based Q statistics and I 2 test, and a significant heterogeneity was as p (value of Q test) < 0.1 or I 2 > 50%.When significant heterogeneity was observed, we would apply the random effects models for analysis.Otherwise, we would apply the fixed effects models.We applied funnel plots as well as Egger's test [27] to assess publication bias.A two-sided p value of 0.05 was deemed as statistical significance.Based on different classifications of blood lactate levels and the categorization of children and adults, we conducted subgroup analyses.
We performed meta-analysis by using the hierarchical summary receiver operating characteristic (HSROC) model to estimate and compare SROC curves [28].Sensitivity and specificity were calculated by true positives, false positives, true negatives, and false negatives.In order to further quantify the lactate level of various outcome, we calculated the frequentist analogue of the surface under the cumulative ranking curve (SUCRA) for each outcome [29].
Data was analyzed using STATA Version 16.0 [30].The network was evaluated using frequentist multivariate meta-analysis (commands network meta and mvmeta) in Stata 16.0.Additionally, publication bias and sensitivity analysis were also conducted by STATA version 16.0.

Assessment of Methodological Quality
The QUADAS-2 tool was employed to evaluate the quality of the included studies.The majority of these studies met a significant portion of the criteria outlined in the QUADAS list.Detailed results of the QUADAS assessments are provided in Table 2.Among the enrolled studies, a total of 124 studies were included, with 110 of them achieving a NOS score equal to or greater than seven points, indicating their classification as high-quality studies.Please refer to Table 3 for a comprehensive overview of these high-quality studies.

Higher Blood Lactate Value Was Associated with Mortality of Sepsis
A total of 78 articles were included in the comparative analysis, assessing the impact of various blood lactate levels on sepsis prognosis.Heterogeneity testing revealed significant heterogeneity across the research findings (I 2 = 92.8%,p = 0.000).To account for this heterogeneity, a random effects model was applied for the meta-analysis, confirming the prognostic significance of elevated blood lactate levels in sepsis mortality [OR = 1.57, 95% CI 1.48-1.65](Figure 3).This finding indicates that higher blood lactate levels are associated with sepsis mortality.Given the substantial heterogeneity, we conducted a subgroup analysis based on different blood lactate cutoff values.Specifically, 13 studies utilized a blood lactate cutoff of ≥2 mmol/L (OR = 2.49, 95% CI 2.00-3.10,I 2 = 67.4%,p = 0.000), while 17 studies employed a cutoff of ≥4 mmol/L (OR = 3.48, 95% CI 2.79-4.34,I 2 = 55.5%, p = 0.003) (Figure 3).Despite the presence of considerable heterogeneity, the pooled effect sizes remained robust, as confirmed by sensitivity analysis (Supplementary Figure S1).Such data demonstrated the blood lactate value was positive associated with sepsis mortality.

Blood Lactate Levels Failed to Distinguish Sepsis, Severe Sepsis and Septic Shock Based on Network Meta-Analysis
As the blood lactate level could indicate the prognosis of sepsis, we attempt to underline whether the value of blood lactate was correlated different outcomes of sepsis (sepsis, severe sepsis and septic shock).For that, network meta-analysis was used to compare the average blood lactate values among non-septic patients, septic patients, severe septic patients and septic shock patients with frequentist statistics (Supplementary Figure S4).In the network meta-analysis, there were significant differences between nonsepsis and sepsis, severe sepsis, septic shock (sepsis vs. nonsepsis, mean 2.03, 95% CI 0.76, 3.30, p < 0.005; severe sepsis vs. nonsepsis, mean 3.03, 95% CI 1.34, 4.72, p < 0.001; septic shock vs. nonsepsis, mean 3.07, 95% CI 1.70, 4.44, p < 0.001), but there were no significant differences among sepsis, severe sepsis and septic shock (severe sepsis vs. sepsis, mean 1.00, 95% CI −0.51, 2.51; septic shock vs. sepsis, mean 1.04, 95% CI −0.05, 2.13; septic shock vs. severe sepsis, mean 0.04, 95% CI −1.58, 1.67) (Figure 6).However, in SUCRA statistical analysis, it still showed the ranked blood lactate levels among patients with sepsis, severe sepsis and septic shock, with blood lactate levels in septic shock patients ranked first, followed by severe sepsis patients and sepsis patients (Figure 7).Funnel plots suggested no publication bias based on its symmetry (Supplementary Figure S4).6).However, in SUCRA statistical analysis, it still showed the ranked blood lactate levels among patients with sepsis, severe sepsis and septic shock, with blood lactate levels in septic shock patients ranked first, followed by severe sepsis patients and sepsis patients (Figure 7).Funnel plots suggested no publication bias based on its symmetry (Supplementary Figure S4).

Discussion
Our study conclusively establishes an association between elevated blood lactate levels and increased mortality, providing updated statistical conclusions and risk values.Importantly, this association holds true across diverse demographic factors such as age, gender, race, geographic region, and the specific assay method employed to measure blood lactate.Furthermore, diagnostic systematic review results demonstrate the significant predictive ability of lactate in sepsis mortality.Higher cutoff values for lactate are associated with increased mortality risk.However, this network analysis did not underline any significant differences in average blood lactate levels among individuals with sepsis, severe sepsis and septic shock.In clinical practice, we cannot overly rely on lactate to determine the severity of sepsis.Mildly elevated lactate levels may also progress to severe sepsis or septic shock, and excessively high lactate levels may indicate non-septic infections.These complex clinical scenarios require a more flexible and nuanced approach to assessment.
Several recent studies have highlighted the potential prognostic markers of outcome in severe sepsis, including the central venous minus arterial carbon dioxide pressure to arterial minus central venous oxygen content ratio (Pcv-aCO2/Ca-cvO2 [141,142], the CRP/albumin ratio [86,143], and lactate clearance [65,113,116,144].The diagnostic performance of the Pcv-aCO2/Ca-cvO2 ratio > 1.696 at 24 h was analyzed using the ROC curve, and it was found to have an AUC of 0.82 (95% CI, 0.661-0.979).The lactate > 1.6 mmol/L had an AUC of 0.853 (95% CI 0.712-0.915)for predicting 28-day mortality [141].The CRP/albumin ratio had an AUC of 0.621 for predicting mortality [86].Other biomarkers, such as first urine liver-type fa y acid binding protein (L-FABP), plasma mtDNA also had a predictive value for sepsis mortality [138,145].The AUC of urine L-FABP and plasma mtDNA for mortality were 0.647 and 0.726, respectively.The AUC was 0.864 for lactate [138].Lactate had the greatest association with mortality.However, these markers necessitate multiple laboratory measurements, which can be unfavorable for early and accurate

Discussion
Our study conclusively establishes an association between elevated blood lactate levels and increased mortality, providing updated statistical conclusions and risk values.Importantly, this association holds true across diverse demographic factors such as age, gender, race, geographic region, and the specific assay method employed to measure blood lactate.Furthermore, diagnostic systematic review results demonstrate the significant predictive ability of lactate in sepsis mortality.Higher cutoff values for lactate are associated with increased mortality risk.However, this network analysis did not underline any significant differences in average blood lactate levels among individuals with sepsis, severe sepsis and septic shock.In clinical practice, we cannot overly rely on lactate to determine the severity of sepsis.Mildly elevated lactate levels may also progress to severe sepsis or septic shock, and excessively high lactate levels may indicate non-septic infections.These complex clinical scenarios require a more flexible and nuanced approach to assessment.
Several recent studies have highlighted the potential prognostic markers of outcome in severe sepsis, including the central venous minus arterial carbon dioxide pressure to arterial minus central venous oxygen content ratio (Pcv-aCO2/Ca-cvO2 [141,142], the CRP/albumin ratio [86,143], and lactate clearance [65,113,116,144].The diagnostic performance of the Pcv-aCO2/Ca-cvO2 ratio > 1.696 at 24 h was analyzed using the ROC curve, and it was found to have an AUC of 0.82 (95% CI, 0.661-0.979).The lactate > 1.6 mmol/L had an AUC of 0.853 (95% CI 0.712-0.915)for predicting 28-day mortality [141].The CRP/albumin ratio had an AUC of 0.621 for predicting mortality [86].Other biomarkers, such as first urine liver-type fatty acid binding protein (L-FABP), plasma mtDNA also had a predictive value for sepsis mortality [138,145].The AUC of urine L-FABP and plasma mtDNA for mortality were 0.647 and 0.726, respectively.The AUC was 0.864 for lactate [138].Lactate had the greatest association with mortality.However, these markers necessitate multiple laboratory measurements, which can be unfavorable for early and accurate assessment.Blood lactate levels are considered sensitive markers for sepsis and septic shock, reflecting cellular metabolism [146].In our analysis, we included a larger number of studies, enhancing the comprehensiveness and reliability of our results.
In our analysis, we observed heterogeneity among the included studies.The levels of blood lactate varied significantly across different studies, resulting in substantial unmanageable heterogeneity in the pooled effects.This heterogeneity can be attributed, in part, to the diverse sources of arterial and venous blood lactate, variations in measurement equipment and assays, as well as differences in methodologies for lactate measurements and the use of various lactate cut-off values.Moreover, even after conducting subgroup analysis, high heterogeneity persisted due to significant disparities in diagnostic criteria, primary diseases, disease severity, and treatment status, necessitating further validation and analysis of clinical trial results.It is important to note that blood lactate levels in the body are influenced by multiple processes, including lactate generation, transformation, clearance, and recycling [8].The dynamics of blood lactate levels can provide valuable insights into identifying individuals at high risk for poor clinical outcomes.However, our study has several limitations.It underscores that a single measurement of blood lactate levels in clinical practice may not fully capture the dynamic state of the body.Therefore, dynamic monitoring of blood lactate levels is recommended for more effectively guiding sepsis treatment and assessing prognosis.Additionally, a model incorporating more variables may offer improved predictive capability compared to a single lactate variable.Recently, a study suggested that compared to lactate and albumin alone, the predictor value of the lactate and albumin ratio was outstanding in predicting death and hospital stay (discharge) among sepsis participants, with a sensitivity of 100% and a specificity of 88% [147].As most of the included studies are observational, they cannot infer causation.Further research, ideally through controlled trials or experimental studies, is needed to establish causal relationships between blood lactate levels and sepsis outcomes.While lactate plays a significant role in predicting the risk of sepsis mortality, in clinical practice, it is essential to consider other clinical indicators and the overall condition of the patient.This comprehensive approach allows for a more thorough evaluation of the patient's condition and the development of appropriate treatment plans.

Conclusions
Based on the meta-analysis, blood lactate revealed the capability to predict the multiple sepsis mortality.This study demonstrated the high-level blood lactate was associated with an elevated risk of death and predicted higher risk of mortality.However, the levels of blood lactate failed to distinguish sepsis, severe sepsis and septic shock.Large-scale multicenter randomized clinical trials are needed to provide high-level evidence and confirm the optimal cut-off for prognosis of sepsis.In clinical practice, we cannot overly rely on lactate to determine the severity of sepsis; the dynamic monitoring of blood lactate levels is recommended for more effectively guiding sepsis treatment and assessing prognosis.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biomedicines12020447/s1, Figure S1.Sensitivity analysis of the individual trials on the results for blood lactate level associated with sepsis mortality; Figure S2.Sensitivity analysis of the individual trials on the results for blood lactate level associated survivors and non-survivors of sepsis; Figure S3.Funnel plot with Egger's test for association between blood lactate levels and mortality; Figure S4.The network meta-analysis of available comparisons of blood lactate levels of patients with various outcomes; Figure S5.Comparison-adjusted funnel plot for blood lactate levels of patients with various clinical outcomes.A: nonsepsis; B: sepsis; C: severe sepsis; D: septic shock; Table S1.Original Sequential Organ Failure Assessment (SOFA) score; Table S2.Systemic Inflammatory Response Syndrome (SIRS) Criteria; Table S3.qSOFA (Quick SOFA) Criteria.

Figure 1 .
Figure 1.The pathogenesis of sepsis and the relationship with increased lactate.

Figure 1 .
Figure 1.The pathogenesis of sepsis and the relationship with increased lactate.

Figure 2 .
Figure 2. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram for study identification and selection.

Figure 2 .
Figure 2. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram for study identification and selection.

Figure 6 .
Figure 6.Forest plot of network meta-analysis of lactate and sepsis, severe sepsis and septic shock.

Figure 6 .
Figure 6.Forest plot of network meta-analysis of lactate and sepsis, severe sepsis and septic shock.

Figure 7 .
Figure 7. Results of network rank test and the surface under the cumulative ranking curve (SUCRA).

Figure 7 .
Figure 7. Results of network rank test and the surface under the cumulative ranking curve (SUCRA).

Table 1 .
Basic characteristics of included studies.

Table 1 .
Cont.Surviving Sepsis Campaign; MIMIC: Medical Information Mart for Intensive Care; SFAR-SRLF *: recommendations of the French Intensive Care Societies; APACHE: Acute Physiology and Chronic Health Evaluation; SCCM/ESICM/ACCP/ATS/SIS: Society of Critical Care Medicine/European Society of Critical Care Medicine/American College of Chest Physicians/American Thoracic Society, Surgical Infection Society.Assessment *: a: network analysis; b: mortality; c: different type of sepsis; d: diagnostic analysis.Superscript numbers are used to distinguish between two different studies by the same author.

Table 3 .
Quality assessment of included studies by Newcastle-Ottawa Scale (NOS).