Risk Factors for Mortality in Sepsis Patients without Lactate Levels Increasing Early

Objectives Our purpose was to investigate the influencing factors for mortality in sepsis patients without lactate levels increasing in the early stage. Methods We conducted a retrospective observational study involving 830 adult sepsis patients admitted to ICU. We calculated time-weighted lactate (LacTW), a dynamic value that incorporates both the magnitude of change and the time interval of such change, to represent lactate levels in the first 24 hours. ROC curve was used to find the cutoff of LacTW for predicting mortality, and the influencing factors for lactate levels and mortality in the low lactate group were further studied. The primary outcome was hospital mortality. Results Among 830 patients, LacTW > 1.975 mmo/L was found to be the cutoff threshold for predicting mortality (AUC = 0.646, P < 0.001). The following indexes related to organ dysfunction influenced LacTW: acute physiology and chronic health evaluation II (APACHE II) score (P < 0.001), activated partial thromboplastin time (APTT) (P = 0.002), total bilirubin (P = 0.012), creatinine (P = 0.037), with hypotension (P < 0.001), chronic kidney disease (P = 0.013), and required continuous renal replacement therapy (CRRT) (P < 0.001). Of the 394 patients in the low lactate group, age (P = 0.002), malignancy (P < 0.001), lactate dehydrogenase (P = 0.006), required treatment such as mechanical ventilation (P < 0.001), CRRT (P < 0.001), vasoactive drugs (P < 0.001), and glucocorticoid (P < 0.001), and failure to reach the target fluid resuscitation of 30 ml/kg within 6 hours (P = 0.003) were independently associated with hospital mortality. Conclusions Due to the lower incidence of early organ dysfunction, lactate levels are not increased or delayed in some septic shock patients in the early stage, thus affecting the alertness of clinicians and the timeliness and adequacy of fluid resuscitation, and finally affects the prognosis.


Introduction
Sepsis and septic shock are essential global emergency and critical care medicine issues. More than one million people sufer from sepsis worldwide each year, causing at least 1/6 to 1/3 of deaths. Improving prevention, identifcation, and treatment of sepsis is a global health priority [1][2][3][4].
Hyperlactatemia is closely related to disease severity and prognosis in patients with severe sepsis or septic shock [5][6][7]. Sepsis patients are complicated with hyperlactatemia due to tissue hypoperfusion, excess lactate production, and decreased lactate clearance caused by hepatic and renal insufciency. However, it does not entirely represent tissue hypoxia. Te concept that lactate is merely a metabolic waste product has now evolved to the point that lactate is seen as an energy shuttle [8]. And hyperlactatemia may indicate adaptive responses to the metabolic processes of severe infections and treatment [9][10][11]. A decrease in lactate level with fuid resuscitation in septic shock may indicate clinical improvement, so lactate clearance can help to evaluate the overall response. However, it responds too slowly to guide acute changes in treatment. Terefore, hyperlactatemia should not be seen as a problem but as a warning of altered cellular function, and therapy should not be guided by a single lactate indicator [12,13].
A secondary analysis compared the prognosis of patients who met the new defnition of septic shock in sepsis3.0 [14] with those who did not meet the new diagnostic criteria but met the previous one. In this study, low lactate level was the main reason patients did not meet the new diagnostic criteria. Although patients with high lactate levels meet the new diagnostic criteria and have a signifcantly higher mortality rate, patients with low lactate levels also have a mortality rate of 14.4%, which is considerably higher than the incidence of serious diseases such as stroke and heart disease [15]. At present, multiple studies focus on patients with hyperlactatemia because elevated lactate level was well established as an independent predictor of mortality in patients with sepsis. But there exist clinical cases of patients diagnosed with sepsis or septic shock without lactate levels increasing in the early stage despite clinical deterioration. Tis study aimed to investigate the risk factors of mortality in patients with sepsis or septic shock without early lactate level elevation.

Study Design.
Our study retrospectively collected data from patients diagnosed with sepsis or septic shock and admitted to the EICU or ICU in the First Afliated Hospital of Anhui Medical University from May 2019 to March 2022, which had been approved by the Clinical Medical Research Ethics Committee of our hospital (Registration Number: PJ-2022- [11][12][13][14][15][16][17][18]. A total of 830 patients were enrolled in the study. Clinical laboratory data included: blood routine: white blood cell count (WBC), neutrophil ratio, hemoglobin (Hb), platelet count (PLT), red blood cell distribution width (RDW), and hematocrit (HCT); anticoagulant function: prothrombin time (PT), activated partial thromboplastin time (APTT), international normalized ratio (INR), and prothrombin time activity (PTA); liver function: total bilirubin (TBIL), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and lactate dehydrogenase (LDH); renal function: urea nitrogen (BUN), creatinine (CRE), and estimated glomerular fltration rate (eGFR); myocardial indexes: creatine kinase (CK), creatine kinase isoenzyme (CK-MB), troponin, and myoglobin; electrolytes: serum potassium, serum sodium, serum calcium, serum phosphorus, and serum bicarbonate concentration (HCO3); infammatory indicators: procalcitonin (PCT) and C-reactive protein (CRP); all lactates within 24 hours after ICU admission. Acute physiology and chronic health evaluation II (APACHE II) score, sepsis-related organ failure assessment (SOFA) score, and Glasgow Coma Scale (GCS) score when entering ICU were used to evaluate the severity of disease and organ dysfunction. In addition, the length of hospital stay and ICU stay were collected, and the clinical outcome was hospital mortality.

Statistical Analysis.
We applied the time-weighted lactate value (LacTW) within 24 hours after ICU admission to represent lactate level in the early stage. Te LacTW is calculated as follows: the average value of lactate at diferent time points multiplied by the time interval, the sum of these lactate weighted values, and then divided by the total time. Tis approach is based on the model established by Finney and colleagues in blood glucose control [17,18]. Unlike in randomized, controlled trials, it is impossible to specify the absolute time interval for lactate measurement in the retrospective clinical study. As a comprehensive parameter that refects the degree and duration of lactic acidosis, LacTW efectively avoids the potential impact of monitoring deviation caused by more frequent blood lactate monitoring in more severe patients. So, it may better represent the actual lactic acid level than the arithmetic means during hospitalization and a single lactate value at admission. SPSS 26.0 statistical software and GraphPad Prism 9 was used to analyze the data. Te Shapiro-Wilk (SW) test was used to test the normality of the distribution. Data conforming to the normal distribution were expressed as mean ± standard deviation (x ± s), and a t-test was used to compare groups. Non-normally distributed data were expressed as median (interquartile range), and a nonparametric test was used for comparison between groups. Frequencies and percentages were used to describe categorical variables such as gender, underlying disease, mechanical ventilation, use of vasoactive drugs, etc., and were analyzed by chi-square tests. Simple correlation analysis (Ttest, nonparametric test, chi-square test, etc.) was performed on the potential risk factors for hospital mortality in patients with sepsis or septic shock, and the statistically signifcant variables were analyzed by binary logistic regression analysis to obtain the risk factors for predicting mortality. Te ROC curve was used to calculate the optimal time-weighted lactate value for predicting mortality. We divided all the participants into a high lactate group and a low lactate group according to the cutof of LacTW, and further analyzed the diferences between the two groups. And the risk factors for lactate levels and mortality in the low lactate group were analyzed by subgroup analysis. Te P value less than 0.05 was considered statistically signifcant.

Results
830 patients with sepsis or septic shock were enrolled in this study, with an average age of 63 years, including 521 males and 309 females. 418 patients died during hospitalization, with an overall mortality rate of 50.36%.
We compared the baseline data and clinical characteristics between the nonsurvivors and the survivors among all patients. Tere were signifcant diferences in age, length of stay in ICU, complicated with malignancy, two or more underlying diseases, scores representing disease severity, and requirement of special treatments such as mechanical ventilation, vasoactive drugs, glucocorticoid, and CRRT ( Table 1). As for clinical indicators, whether the fuid resuscitation volume at 6H reached the target volume was statistically signifcant. Tere were signifcant diferences in Hb, RDW, HCT, PT, INR, PTA, BUN, LDH, CK-MB, eGFR, serum potassium, serum phosphorus, troponin, myoglobin, PCT, initial blood lactate value, and LacTW between the nonsurvivors and the survivors (Supplementary Table 1).
As for LacTW, the ROC curve showed that the highest Youden value corresponded to a LacTW value of 1.975 mmo/L, with a sensitivity of 64.1% and a specifcity of 59.2% (AUC � 0.646, P < 0.001, 95% CI: 0.609-0.683) for predicting mortality in patients with sepsis or septic shock ( Figure 1).
In order to further study the indicators afecting lactate level, we conducted multiple linear regression analysis (Table 3), the following indexes had statistically diferent efects on LacTW: APACHE II score (b � 0.066, t � 3.607, Based on LacTW ≤1.975 mmol/L, patients were divided into the high lactate group (n � 436) and the low lactate group (n � 394). Tere were no signifcant differences in baseline data, including gender, age, weight, history of smoking and drinking, and underlying diseases between the two groups (Table 4). Te fuid resuscitation volume in the high lactate group was signifcantly higher than that in the low lactate group at four time points (3H, 6H, 12H, and 24H) within 24 hours after admission to ICU. With the recommended target fuid resuscitation volume of 30 ml/kg, there were statistically signifcant diferences in whether the fuid resuscitation volume reached the target volume of the frst 6H, 12H, and 24H between the two groups ( Table 4). Among them, the mortality of the high lactate group was signifcantly higher than that of the low lactate group. However, the mortality rate in the low lactate group was still as high as 38.1%. Indicators related to organ dysfunction afecting lactate levels listed in Table 3, were signifcantly diferent between the two groups (Table 5). Terefore, it can be speculated that the low incidence of early organ dysfunction leads to the absence or delayed increase of lactate levels in these patients.
We conducted a subgroup analysis on the factors infuencing the mortality of patients in the low lactate group. Supplementary Table 2 shows the clinical characteristics of hospital survivors and nonsurvivors in the low lactate group. Te signifcant univariate indicators were included in the multivariate logistic regression analysis (Table 6). Te mortality of patients with low lactate increased with age (OR � 0.966, 95% CI: 0.945-0.987, P � 0.002). Te mortality of patients with malignancy was higher, and the diference was statistically signifcant (OR � 0.148, 95% CI: 0.061-0.359, P < 0.001). Required vasoactive drugs (OR � 4.745, 95% CI: 1.876-11.999, P � 0.001), glucocorticoid (OR � 2.579, 95% CI: 1.321-5.034, P � 0.005), and mechanical ventilation (OR � 2.926, 95% CI: 1.151-7.442, P � 0.024) and CRRT (OR � 3.079, 95% CI: 1.435-6.607, P � 0.004) had signifcant efects on the mortality of patients with low lactate in the early stage. In addition, LDH also had an impact on the mortality of patients in the low lactate group, with a statistical diference (OR � 0.999, 95% CI: 0.999-0.9998, P � 0.006). Statistical signifcance was not observed about fuid resuscitation volume in any of the time points mentioned above. But compared with the patients who met the standard fuid resuscitation volume in the frst 6 hours, the mortality was higher in patients who failed to meet the standard, and the diference was statistically signifcant (OR � 22.621, 95% CI: 2.987-171.333, P � 0.003). For all participants, time-weighted lactate level was positively correlated with the amount of fuid resuscitation at each time point within 3H, 6H, 12H, and 24H after ICU admission (Figure 2), that is, the amount of fuid resuscitation increased with the increase of lactate level within 24 hours after ICU admission.

Discussion
Tis study confrmed the association between lactate level and mortality in patients with sepsis and septic shock again. Te cutof value of LacTW was 1.975 mmol/L, which is very close to the critical value of lactate in sepsis 3.0 [14]. However, the initial lactate value in ICU had no signifcant efect on mortality. It has been proved that time-weighted lactate is superior to static indices of lactate concentration and has signifcant independent predictive value of outcome  in critically ill patients [19][20][21]. Our result reinforces the notion that, as a dynamic parameter, time-weighted lactate is a better proxy for early lactate levels than the initial lactate value in sepsis. Te defnition of septic shock includes two indispensable components: elevated lactate levels and the need for vasoactive agents to maintain mean arterial pressure ≥65 mmHg after adequate fuid resuscitation. So, we may miss patients in the early stages of sepsis. In a multicenter study, patients with only refractory hypotension with normal lactate levels had a prevalence of 21.9% and a mortality of 29.6%. Te morbidity and mortality of patients with hyperlactatemia and normal blood pressure were 31.6% and 27.5%, respectively [22,23]. In this study, the mortality rate of patients with early sepsis and septic shock whose LacTW was less than 1.975 mmol/L for the frst 24 hours after admission to ICU was 38.1%. Terefore, the severity of the disease cannot be misjudged by the absence of lactate level increasing, such inappropriate neglect may afect the timeliness and adequacy of specifc treatment interventions, which may ultimately afect patient outcomes. Tat is, lactate is not omnipotent.
Terefore, the causality of lactate levels and mortality remains to be further verifed. Tis is a clinical observational study, and the causality is less convincing than RCTs. Because doctors do not make decisions randomly but based on changes in the patient's condition. So, in this case, the relationship between lactate level and mortality is not a simple causal relationship. High lactate level may cause clinicians to attach great importance to it, and lactate level may decrease after treatment. In contrast, normal lactate level may lead to clinicians' neglect, thus missing the opportunity for treatment. In fact, the inference of causality of the longitudinal data can be further verifed according to the method of Zhang et al. [24].
Why is lactate not elevated or delayed in some patients with septic shock despite circulatory failure in the early stage? We found that lactate levels increased in patients with liver, kidney, and coagulation dysfunction (especially in patients with renal insufciency), hypotension, and high APACHE II score. Lactate remains the product of its production and elimination. Tissue hypoxia can lead to impaired mitochondrial function or microcirculation dysfunction, and lactate is overproduced and underutilized [25,26]. Lactate can also be produced under aerobic conditions. It can be used as an energy source, messenger molecules, and gluconeogenesis precursors by various cells, tissues, organs, and the whole body to regulate body metabolism. Lactic acid is not a mark of hypoxia [8,27,28]. And lactate is mainly cleared in the liver and kidney, among which the liver is responsible for 70% of the systemic clearance rate. Studies have shown that patients with acute liver failure will lead to accelerated liver glycolysis and reduced gluconeogenesis, resulting in increased lactic acid levels [29,30]. But in our study, renal insufciency and creatinine level were positively correlated with lactic acid level. On the one hand, renal insufciency leads to decreased lactate clearance rate; on the other hand, acute kidney injury is also a severe complication of sepsis, and recent studies have shown that lactic acid can induce immunosuppression by inducing lymphocyte apoptosis in septic acute kidney injury [31]. According to the results of this study, patients in the low-lactate group had a lower rate of organ dysfunction, so we can speculate that due to the lower incidence of early organ dysfunction, lactate levels are not increased or delayed in the early stage.
To avoid inappropriate neglect of such patients with sepsis who may have clinical deterioration but do not have elevated lactate levels in the early stage, we further analyzed that age, malignancy, and requirement of special treatments such as mechanical ventilation, CRRT, vasoactive drugs, and glucocorticoid had statistically diferent efects on mortality in the low lactate group. Te fact that these indicators refected disease severity has an impact on mortality is no surprise, which had been well established. But interestingly, we found that whether fuid resuscitation volume reached the standard of 30 ml/kg at 6H was an independent risk factor for predicting mortality in the low lactate group. Tis fnding may suggest that low lactate levels cause clinicians to neglect, leading to inadequate or delayed treatment, and even further clinical deterioration of patients.
Te surviving sepsis campaign guidelines recommend that in the resuscitation from sepsis-induced hypoperfusion (hypotension or lactic acidosis), at least 30 mL/kg of IV crystalloid fuid should be given within the frst 3 hours [16,32,33]. A retrospective study has shown that sepsis patients who do not receive the recommended 30 ml/kg intravenous fuids within 3 hours have an increased risk for in-hospital mortality, delayed hypotension, and increased ICU length of stay. Moreover, elderly, the obese, male, patients with a history of heart failure, and end-stage renal disease are unlikely to achieve the standard in the frst 3 hours [34]. However, few studies paid attention to fuid resuscitation in sepsis patients without increased lactate levels or the efect of lactate levels on fuid resuscitation.
In our study, there was no signifcant diference in baseline data including age, gender, weight, and underlying diseases between the low lactate group and the high lactate group. However, there were statistical diferences in the fuid resuscitation volume at the time point of 3H, 6H, 12H, and 24H within the frst 24 hours after ICU admission, and LacTW is positively correlated with fuid resuscitation volume. Te low-lactate group had fewer patients who achieved the required fuid resuscitation and had a higher mortality rate than the high-lactate group. Terefore, the adequacy of fuid resuscitation should not be ignored because of the low lactic acid level. But more fuid resuscitation volume did not translate into lower mortality. Rivers and colleagues reported early goal-directed therapy (EGDT) to guide fuid resuscitation through continuous monitoring of prespecifed physiological goals in 2001 [35]. However, several multicenter randomized controlled trials subsequently showed that EGDT has no signifcant advantage in improving mortality and organ dysfunction compared with ordinary care [23]. In addition, studies have shown that 67% of patients undergoing resuscitation with EGDT protocol had clinical evidence of fuid overload after 24 hours, and 48% of them had the characteristics of continuous fuid overload on the third day of hospitalization [36]. And positive fuid balance is independently associated with an increased risk of organ dysfunction and death in critically ill patients [37,38]. So, in the early stage of sepsis, adequate fuid resuscitation should be performed as soon as possible, while restricted fuid resuscitation should be advocated in the later stage, rather than the more the better. Although guidelines currently recommend an initial fuid resuscitation volume of 30 ml/kg, the specifc amount of fuid resuscitation required should vary from person to person. Dynamic monitoring of vital signs, microcirculation, fuid responsiveness, and other indicators is needed to adjust the patient's fuid balance dynamically.
As for the timing of early fuid resuscitation, the latest sepsis bundle recommends starting fuid resuscitation within 1 hour, and the volume is required to reach the standard within 3 hours [33,39]. Diferently, in this study, the time point of getting the standard of fuid resuscitation volume afecting mortality was 6 hours, it may be explained by the inappropriate neglect of clinicians caused by low lactate level, which not only afected the adequacy of fuid resuscitation volume but also afected the timeliness.

Limitations and Future
Research. Tis study has several limitations. First, it is a single-center retrospective observational analysis, and all the clinical data were collected in one hospital. Second, since we focus on patients with sepsis or septic shock after entering ICU, the clinical index and treatment in the emergency department were not obtained. Tird, patients with ICU hospitalization of fewer than 24 hours were excluded. Most of these patients died quickly after entering ICU, or their families gave up further treatment, which may lead to selection bias. In addition, this study did not specifcally analyze the impact of the infection sources, pathogen, the type and dose of vasoactive drugs, and fuid type of resuscitation on mortality, which can be further followed up in subsequent studies.

Conclusions
Te lactate level in the early stage of sepsis is related to organ function and prognosis. Compared with the initial lactate value, LacTW showed a more signifcant prognostic efect in patients with sepsis. Mortality in patients with no elevated lactate level in the early stage was associated with advanced age, malignancy, and the need for supportive treatment, such as the use of vasoactive drugs, glucocorticoid, mechanical ventilation, and CRRT, and the mortality in this group was also associated with whether the target fuid resuscitation volume within 6 hours. So, it can be calculated that due to the lower incidence of early organ dysfunction, lactate levels are not increased or delayed in some septic shock patients despite circulatory failure in the early stage, thus afecting the alertness of medical workers and the timeliness and adequacy of fuid resuscitation, and fnally afecting the prognosis of these patients with septic shock. It 8 Emergency Medicine International may serve as a warning to clinicians to avoid ignoring sepsis patients at potential risk for clinical deterioration and whose early lactate levels are not elevated.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.