Relative Hyperlactatemia in the Emergency Department: A Retrospective Cohort Study

Ralphe Bou Chebl American University of Beirut Medical Center Sarah Jamali American University of Beirut Medical Center Nancy Mikati American University of Beirut Medical Center Reem Al Assaad American University of Beirut Medical Center Karim Abdel Daem American University of Beirut Medical Center Nadim Kattouf American University of Beirut Medical Center Rawan Safa American University of Beirut Medical Center Maha Makki American University of Beirut Medical Center Hani Tamim American University of Beirut Medical Center Gilbert Abou Dagher (  ga66@aub.edu.lb ) https://orcid.org/0000-0002-9147-1515


Conclusion
The optimal cutoff of initial serum lactate that discriminates between survivors and non-survivors is in the ED 1.33 mmol/L. Relative hyperlactatemia is associated with increased mortality in emergency department patients, and this interaction seems to be more important in healthy patients.

Background
The breakdown of pyruvate via the enzyme lactate dehydrogenase leads to the formation of lactate.
Healthy individuals produce basal lactate levels of 1.0 ± 0.5 mmol/L 1, 2 . Normal lactate levels in the blood usually refer to levels < 2 mmol/L 3, 4 . Current theories relate hyperlactatemia to decreased oxygen delivery and tissue malperfusion, or to impaired oxygen utilization and adrenergic stress, and both paradigm mechanisms may be compounded by impaired elimination [5][6][7][8][9][10] . There have been signi cant advances in our understanding of the physiology of lactate, and it has since become a mainstay biomarker, heavily integrated into clinical decision-making of septic patients in the emergency department 11,12 . Furthermore, hyperlactatemia (> 2.00 mmol/L) has been associated with poor outcomes and independently predicts mortality in diverse patient populations presenting to the emergency department (ED) 13 . In its most recent guidelines, the Surviving Sepsis Campaign recommend using lactate levels ≥ 4 mmol/L to initiate IV uid resuscitation and recommend re-measuring lactate levels if they are > 2 mmol/L to monitor the response to the resuscitation 8 . Lactate levels within the reference range (≤ 2 mmol/L) have a less clear clinical interpretation and may result in less attention given to these patients in the ED. Furthermore, there is a considerable group of patients who present in shock and have elevated in-hospital mortality rates despite having lactate levels within the normal range 14 . It has been proposed that this subgroup of patients with septic shock possess distinctive clinical and physiological pro les, and may have unique treatment parameter considerations [15][16][17] . Moreover, there is emerging evidence that suggests that relative hyperlactatemia (i.e. lactate above an identi ed threshold) has a more appropriate consideration in certain subgroups of patients, such as those with sepsis 18 , septic shock 14,19 or cancer 20 . The clinical interpretation of lactate ≤ 2.00 mmol/L in ED patients is not well characterized. This allows us to question what the best prognostic cutoff value is for lactate. One study proposed that a cutoff of 1.35 mmol/L best discriminates between survivors and non-survivors in the intensive care unit (ICU) 21 . Nonetheless, there is a paucity of data on this issue, particularly in the ED. This study aims to evaluate the optimal cutoff threshold for lactate that distinguishes between survivors and non-survivors and predicts in-hospital mortality among patients presenting to the ED with an initial serum lactate levels within the reference range (0.01 to 2.00 mmol/L).

Design and Setting
This was a retrospective cohort study of adult patients presenting to the academic ED of a tertiary care center between the dates of January 1, 2014 and June 30, 2019. All patients aged 18 years of age or older who presented to the ED and had a serum lactate level drawn had their charts queried. All patients who had an initial serum lactate level drawn and within the reference range (0.01-2.00 mmol/L) were included in the study. Exclusion criteria consisted of patients with serum lactate > 2.00 mmol/L, patients who were pregnant, patients who presented with cardiac arrest and patients who had been admitted less than 10 days prior to presentation. The data collection protocol was standardized, and information was extracted from the patient's electronic medical records and anonymized. The variables collected included patient demographics and characteristics, vital signs and initial laboratory tests upon presentation to the ED, diagnosis, presence of sepsis on admission, interventions performed (renal replacement therapy, mechanical ventilation, vasopressor use, steroid use, antibiotic administration, intravenous uid administration), disposition, length of stay, readmission rates, in-hospital mortality and 30-day mortality rates. In this study, sepsis was de ned as the presence of an infection with signs of organ dysfunction, as represented by the Sequential [Sepsis-related] Organ Failure Assessment (SOFA) score of 2 points or greater according to the Third International Consensus De nitions for Sepsis and Septic Shock (Sepsis-3) guidelines 22 . Patient who did not meet this de nition were labeled as having an infection. The study was approved by the hospital's Institutional Review Board (IRB).

Outcomes
The primary outcome was in-hospital mortality. Secondary outcomes included mechanical ventilation, vasopressor use, steroid use, intravenous uid administration, lengths of stay (ED, ICU and total).

Statistical analysis
Statistical analysis was conducted with IBM SPSS Statistics for Windows, version 24 (IBM Corp., Armonk, N.Y., USA). Continuous variables are presented as mean ± standard deviation and categorical variables are presented as frequency with valid percent. Patients were strati ed into survivors and non-survivors. A Youden's index was used to determine the optimal threshold that predicts in-hospital mortality, and patients above the threshold were reclassi ed as having relative hyperlactatemia. A multivariate logistic regression was performed to determine the association of relative hyperlactatemia and in-hospital mortality. All variables with statistical and clinical signi cance were included in the analysis. The variables included were age, gender comorbidities (hypertension, diabetes, dyslipidemia, chronic obstructive pulmonary disease, heart failure, immunocompromised), diagnosis category, sepsis, lymphocyte count, and white blood cells (WBC) count. We looked at the in-hospital mortality among patients with and without relative hyperlactatemia, strati ed by selected subgroups. These subgroups included the following: male versus female patients; age younger than 50 years versus age older to equal to 50 years; diabetes versus no diabetes; hypertension versus no hypertension; dyslipidemia versus no dyslipidemia; coronary artery disease versus no coronary artery disease, chronic obstructive pulmonary disease (COPD) versus no COPD, congestive heart failure versus no congestive heart failure, sepsis versus no sepsis and vasopressor versus no vasopressor use.

Results
During the study period, a total of 2692 patients were identi ed with lactate levels within the reference range (0.01 to 2.00 mmol/L). Of these, 868 were excluded because their initial lactate value upon ED presentation was greater than 2.00 mmol/L, 163 were excluded because they were under 18 years of age, 4 were excluded because they were pregnant, 9 were excluded because they presented with cardiac arrest and 10 were excluded because they had been recently admitted to the hospital less than 10 days prior to presentation. A total of 1638 patients were included in the study, and their characteristics are summarized in Table 1. ROC curve Figure 2 demonstrates the receiver operating curve for lactate upon presentation and in-hospital mortality. The optimal cutoff that differentiates between survivors and non-survivors was found to be 1.33, and the area under the curve for that value was 0.545 (95% CI 0.477-0.614).

Laboratory and vital signs
Youden's index was used to nd the threshold that best discriminates between survivors and nonsurvivors and it was found to be 1.33 mmol/L. Table 2 summarizes the characteristics of patients with initial lactate levels below 1.33 mmol/L and of those with initial lactate equal to or above 1.33 mmol/L. Patients with relative hyperlactatemia (equal to or above 1.33 mmol/L) were older (  Patients with relative hyperlactatemia were also more likely to have a diagnosis of sepsis (56.2% vs. 34.1%, p < 0.001),. Furthermore. they were more likely to receive vasopressors (5.0% vs. 2.2%, p = 0.015), steroids (13.0% vs. 7.5%, p < 0.001), and higher volumes of IV uids in the rst 24 hours (2.1 ± 1.2, p < 0.001) compared to patients with lactate levels below 1.33 mmol/L. In-hospital mortality was higher in patients with relative hyperlactatemia (4.4% vs. 1.9%, p = 0.029) compared to patients with lactate levels below 1.33 mmol/L. The outcomes are summarized in Table 3.

Multivariate logistic regression
After adjusting for the multiple confounding variables such as age, gender, laboratory results and comorbidities, we found that patients with relative hyperlactatemia had a 1.78 greater odds of in-hospital mortality (95% CI 1.18-4.03; p -0.02) than patients without.

Subgroup analysis
The association between relative hyperlactatemia and mortality in various subgroups is demonstrated in Table 5. Relative hyperlactatemia was associated with increased hospital mortality consistently across the different subgroups, however the difference was only statistically signi cant in patients without hypertension (4.7% versus 1.1%; p 0.008), patients without diabetes (4.2% versus 1.0%; p 0.01), patients without dyslipidemia (5.4% versus 1.5%; p 0.008) and patients without COPD (4.3% versus 1.8%; p 0.04).

Discussion
The results of this study have shown that lactate levels of 1.33 mmol/L were found to have the optimal threshold to discriminate between survivors and non-survivors. Furthermore, relative hyperlactatemia (1.33 mmol/L to 2.00 mmol/L) patients had 1.78 times greater odds of dying than patients with lactate < 1.33 mmol/L. The overall hospital mortality in our population was 3.8, with the relative hyperlactatemia subgroup having a higher mortality rate of 4.4% in patients with initial lactate ≥ 1.33 mmol/L, compared to 1.9% in patients with initial lactate < 1.33 mmol/L. Similar results were demonstrated in a study by Rishu et al. who looked at a discriminatory level of lactate in an intensive care unit and found that a cutoff value of 1.35 mmol/L adequately discriminated between survivors and non-survivors 21 . In addition, they also found that mild hyperlactatemia was an independent predictor of hospital mortality (OR 1.60; 95% CI 1.29-1.98). De ning hyperlactatemia as a serum lactate level greater than or equal to 2.00 mmol/L insinuates that lactate values between 1.00 and 2.00 mmol/L can be interpreted as normal 22,23 . Historically, authors have used the cutoff of 1.3 mmol/L to de ne hyperlactatemia 2, 24 . Over the years, the reference ranges for hyperlactatemia have varied from lactate > 1.5 mmol/L to lactate > 2.5 mmol/L 15,19,25 . Following this, a number of studies found an increased mortality risk when using the cutoff lactate of > 2.0 mmol/L 26, 27 . This led to the gradual adoption of 2.0 mmol/L as the cutoff that de nes the reference range in contemporary literature and in the latest national and international guidelines 1, 22 .
In our study, when we looked at the mortality among septic patients, relative hyperlactatemia patients had a higher mortality (4.4% versus 1.9%). They were also more likely to receive more antibiotics, the reference range that has the greatest prognostic value 21 . Using the Youden index, they found that the optimal cutoff was 1.35 mmol/L, and that relative hyperlactatemia above that cutoff was associated with increased hospital mortality (OR 1.60). This study adds stronger evidence that a lactate within the normal range should be interpreted cautiously as it may still be associated with an increased mortality.
Finally, it is also interesting that amongst the subgroup analysis, relative hyperlactatemia patients without any comorbidities were found to have a statistically signi cant increased mortality when compared to "normal" lactate levels. This is an important nding as it illustrates the important prognostic value of relative hyperlactatemia in healthy individuals versus patients with comorbidities. A possible explanation could be due to the fact that patients with comorbidities are sicker at baseline and tend to raise their lactate more easily in the setting of an acute illness.

Limitations
To the best of our knowledge, this investigation is the rst to explore mild hyperlactatemia in a population of adult ED patients. It is important to note that our results have shown that relative hyperlactatemia patients were older and had higher rates of comorbidities, laboratory derangements, vasopressor use, steroid use, in-hospital mortality and IV uid use. Our strengths include the large sample size and the standardized lactate measurement in the same laboratory for all patients. This study is limited by its retrospective nature and is thus prone to selection bias. Second, our inclusion criteria were restricted to patients on whom there was a clinical decision to draw a serum lactate level, and this also might have introduced a selection bias. Third, since the data pertains to a single center there may be limitations to the generalizability of the ndings. In an attempt to minimize information bias, the authors held multiple meetings to ensure correct patient identi cation and underwent training to standardize the rigorous data abstraction protocol. The data collected is insu cient to de ne the most likely etiology of the mild hyperlactatemia or the effect of early interventions in patients with mild hyperlactatemia on clinical outcome although this was beyond the scope of this study.

Conclusion
The optimal cutoff of initial serum lactate within the reference range that discriminates between survivors and non-survivors is in the ED 1.33 mmol/L. Relative hyperlactatemia is associated with increased inhospital mortality in patients presenting to the ED regardless of diagnosis or co-morbidities. Further studies are required to determine the optimal management to patients with mild hyperlactatemia.

Declarations
Ethics approval and consent to participate: Approved by the Institutional Review Board and carried out in accordance with the recommendations provided. The research was performed according to ethical principles and in compliance with all prevailing and applicable laws, rules and regulations and policies regarding the protection of human subjects and research conduct as outlined by the declaration of Helsinki. Subject privacy and data con dentiality were of paramount concern at all times, and every effort was made to protect subjects' rights and welfare.
Contribution: GAD, RBC, SJ have made substantial contributions to conception and design of the study .
NM, RAA, KAD, NK, RS, and SJ were involved in acquisition of data, data entry, and data cleaning. MK and HT were involved in analysis and interpretation of data. RBC, GAD, SJ, and RS and have been involved in drafting the manuscript. RBC, GAD and HT were involved in revising manuscript critically for important intellectual content. All authors contributed substantially to its revision. GAD and RBC take responsibility for the paper as a whole. Receiver operating curve and area under the curve.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. STROBEChecklistv4MSWordScand.doc