Association of Plasma Lactate Level With 28-day Mortality in Non-elderly and Elderly Sepsis Patients: A Retrospective Cohort Study Based on the Mimic-III Database

Purpose: The objective of this study is to assess the clinical usefulness of lactate as a predictor of 28-day mortality and the relationship between lactate and 28-day mortality in non-elderly (<65 years) and elderly ( ≥ 65 years) sepsis patients who were admitted to an intensive care unit (ICU). Methods: This retrospective study used the Medical Information Mart for Intensive Care (cid:0) , a publicly available database of ICUs. Prognosis was evaluated using receiver operating characteristic (ROC) analysis. Univariate and multivariable binary logistic regression models were used to identify the association lactate with 28-day mortality. We converted continuous variable lactate into a categorical variable based on tri-segment quantile to explore segmentation effects. Results: The average age of 2848 patients was 68.01 years old, and about 55.40% of them were male. The overall 28-day mortality was 30.41%, and the rate in elderly patients was 65.82%. Among non-survivors, the lactate level was signicantly greater for the non-elderly than the elderly. Lactate level was positively associated with risk of 28-day mortality of the non-elderly sepsis patients (p for trend < 0.001), but there was no signicant association between lactate level and 28-day mortality in the elderly group (p for trend = 0.830). The association between lactate and 28-day mortality for sepsis patients without liver cirrhosis was stronger than for sepsis patients with liver cirrhosis (OR 1.28 vs. OR 1.10, P =0.027). Conclusion: Increased lactate level is associated with higher 28-day mortality in the non-elderly sepsis patients, but there is no signicant association between the lactate level and 28-day mortality in the elderly group. suggested as a biomarker for diagnosis and prognostic evaluation of sepsis [2]. An elevated level of plasma lactate during sepsis is associated with more severe disease and poor prognosis [3]. Previous studies revealed a relationship between lactate and mortality of patients with sepsis [4,5] and indicated that patients who presented with lactate>4mmol/L are signicantly associated with in-hospital mortality[4]. However, they only focus on the overall population and did not specically address the relationship between different lactate level and mortality in different age groups. Therefore, this study aims to investigate whether lactate is independently associated with 28-day mortality and the relationship between lactate and 28-day mortality in non-elderly (<65 years) and elderly ( ≥ 65 years) sepsis patients who were admitted to ICU.


Methods of analysis
Continuous variables are expressed as medians and inter-quartile range, and the Mann-Whitney U test was used for comparisons. Categorical variables are expressed as numbers and percentages, and were compared using the chi-square test or Fisher exact test. Prognosis was evaluated using receiver operating characteristic (ROC) analysis, and the ability of lactate to predict 28-day mortality was assessed by calculating the area under the ROC (AUROC). Youden's index was used to assess the performance of the diagnostic test, and the maximum point of Youden's index was used as the cut-off point (sensitivity + speci city -1).
Univariate and multivariate binary logistic regression were used to identify the association between lactate and 28-day mortality of sepsis patients. The Hosmer-Lemeshow test was used to evaluate the suitability of the model. The results were shown as ORs (95% CIs). To investigate the possibility of linearity, we converted continuous variable lactate into a categorical variable based on tri-segment quantile and calculated P values. Data were analyzed with the use of Empower (R) (www.empowerstats.com; X&Y solutions inc.) and the statistical packages R (The R Foundation; http://www.r-project.org; version 3.4.3). For all analyses, a P value below 0.05 was considered signi cant.

Results
Baseline characteristics There were 46,476 patients who were rst admitted to the ICU, and 3512 had diagnoses of sepsis (ICD 995.91), severe sepsis (ICD 995.92), or septic shock (ICD 785.52; Figure 1). After exclusion of 5 patients who were younger than 18 years-old, 309 patients who were discharged from the ICU within 24 h, 4 patients who did not have chart event data, and 346 patients whose initial lactate levels were not measured, there were 2848 patients. Among these 2848 patients, 1249 were younger than 65 years and 1599 were 65 years or older.
The average age of 2848 patients was 68.01 years old, and about 55.40% of them were male. The overall 28-day mortality rate was 30.4%, and the rate in elderly patients was 65.82%. Most patients in the non-elderly and elderly groups were male. The non-elderly patients stayed in ICU for more days, but the 28-day mortality was signi cantly greater in the elderly group. The baseline clinical and demographic characteristics of the survivors were compared to the non-survivors in each age group (Table 1). Analysis indicated the mortality rate increased with age within each age group, and that time in the ICU had a negative association with survivorship only in the elderly group. The non-survivors had a higher score of SAPSII, SOFA, LODS, OASIS, and lactate (p<0.001 for all). Analysis of major complications indicated the incidence of liver cirrhosis, chronic renal insu ciency and malignancy were signi cantly greater among non-survivors in each age group. Further analysis ( Table 2) showed that the lactate was similar for elderly and non-elderly survivors (2.20 vs. 2.10 mmol/L, P = 0.062), but was greater in non-elderly non-survivors than elderly non-survivors (3.20 vs. 2.40 mmol/L, P < 0.001).
Predictive values of lactate and some severity scoring systems for 28-day mortality Table 3 lists the predictive values of lactate and some severity scoring systems in the two age groups for 28-day mortality. Their ROC curves are shown in Fig.2 and Fig.3. The predictive value of lactate for 28-day mortality was 0.661 for the non-elderly group, and 0.553 for the elderly group. The predictive performance of QSOFA and SIRS for 28-day mortality increased after lactate was added, but there was no signi cant difference in the AUROC values of SAPSII SOFA and LODS before and after lactate addition in the two age groups.
Association of lactate with 28-day mortality for each age group In our study, there were 296 dead patients in the non-elderly group and 570 dead patients in the elderly group. We developed different models to examine the independent predicting value of lactate for 28-day mortality and used multivariate analysis to determine the impact of lactate on mortality in each group. The adjusted ORs (95%CIs) for lactate were 1.16 (1.09-1.23) and 1.03 (0.98-1.08) for 28-day mortality in the non-elderly and elderly group, respectively.
To test the nonlinear trend between lactate and 28-day mortality in septic patients of each age group, we converted the continuous variable of lactate into a categorical variable according to tri-segment quantile in the models and divided lactate into three levels. Lactate levels in the nonelderly group were probably consistent with those in the elderly group. We initially used univariate logistic regression analysis to identify variables related to 28-day mortality. The subsequent multivariate logistic regression analysis, in which patients with low level of lactate were used as the reference group, indicated increased lactate level was associated with 28-day mortality. A signi cant trend in higher mortality was observed in patients with elevated lactate level compared to patients with less level of lactate in the non-elderly group (p < 0.05 for all). Lactate level was positively associated with risk of death at 28 days in the non-elderly group (p for trend < 0.001), but there was no such correlation in the elderly group (p for trend = 0.830).

Subgroup Analyses
A strati ed analysis was conducted by baseline characteristics. In the subgroup analyses, we used sex, ethnicity, rst care unit, severity scoring systems, comorbidities and major source of infection as the strati ed variables to examined the associations between lactate and 28-day mortality (Table 5). There was no signi cant difference in relationship between lactate and risk of 28-day mortality in the elderly group in the subgroup analyses. We observed signi cant changes in SAPSII, SOFA, liver cirrhosis and intra-peritoneal infection (p < 0.05 for all) (Fig.4). It was worth mentioning that we found signi cant inverse association of liver cirrhosis with 28-day mortality among the non-elderly sepsis patients. The lactate of the non-elderly patients without liver cirrhosis had a higher risk of 28-day mortality (OR 1.28, p < 0.0001). The association between lactate and 28day mortality for sepsis patients without liver cirrhosis was stronger than for sepsis patients with liver cirrhosis in non-elderly group (OR 1.28 vs. OR 1.10, P =0.027 for the interaction lactate* liver cirrhosis for 28-day mortality).

Discussion
In this study, we retrospectively analyzed the clinical characteristics, 28-day mortality rate, and the relationship of plasma lactate level with the prognosis of sepsis patients in different age groups in the MIMIC-database. Our results suggested that higher lactate level was associated with higher 28-day mortality in the non-elderly sepsis patients, but there was no signi cant association in the elderly group.
The clinical mortality rate of sepsis is now higher than that of myocardial infarction and, except for heart disease, sepsis is the main cause of death in the ICU [15]. The ultimate cause of death from sepsis is organ dysfunction caused by the patient's reaction to the infection [16]. Because of their reduced immune responses and resistance, sepsis more common among the elderly. The prevalence and mortality of severe sepsis have increased signi cantly over time [17].
Most of these patients were from an emergency department, so those with high lactate levels were treated soon after admission. This may have contributed to the lower overall lactate levels in our study than in a previous study [18].
As a product of hypoxia and hypoperfusion, plasma lactate can be used to guide the judgment of uid resuscitation and curative effect in patients with sepsis, and it is also an effective predictor of sepsis mortality [19]. Lactate is a well-known predictor of infection and trauma in patients and has been recognized as a biomarker for risk-strati cation, especially in sepsis patients. Mikkelsen ME et al. [3] proposed initial serum lactate was associated with mortality independent of clinically apparent organ dysfunction and shock in patients admitted to the ED with severe sepsis. A previous study [20] reported that elevated lactate level was associated with poor prognosis for ICU patients after ruptured abdominal aortic aneurysm repair.
The elderly patients are more likely than the nonelderly to develop sepsis due to Gram-negative bacteria, especially among patients with pneumonia and fungal infections. Respiratory tract infections are also more common causes of sepsis in elderly patients [21]. Relative to the non-elderly, we found that elderly sepsis patients had a higher 28-day mortality rate and that sepsis was more likely to be caused by a respiratory tract infection, in agreement with previous studies [18,22]. In contrast, we found that sepsis in non-elderly patients was more likely to be caused by skin and soft tissue infections.
The major indicators of poor prognosis in elderly sepsis patients are shock, elevated plasma lactate, and organ failure (especially of the respiratory system or heart) [23]. In addition, previous research indicated that advanced age is an independent risk factor for severe sepsis and death from sepsis [24]. Our multivariate adjusted logistic regression analysis showed that lactate level was an independent risk factor for 28-day mortality for non-elderly sepsis patients, but this relationship was not signi cant for the elderly. This might be a result of blunted in ammatory responses in the elderly.
We were surprised to nd that the association between lactate and 28-day mortality for sepsis patients without liver cirrhosis was stronger than in sepsis with liver cirrhosis in non-elderly group. Some patients have high plasma lactate levels due to liver dysfunction leading to lactate removal dysfunction, but patients with normal or moderate lactate levels may have a higher risk of death[25], the interpretation of serum lactate levels is often complex. Therefore, a randomized controlled trial is necessary to examine whether liver insu ciency affects the relationship between lactate and the risk of death.
There are some limitations in this study. First, the patient information we used for analysis was only from the MIMIC-III database. The results may have bias, but the advantage of population size in our study may reduce the bias. Second, we could not analyze data from patients excluded due to missing lactate values. Although this is one of our potential limitations, we think that the characteristic of study with a relatively large number of patients may reduce the drawback. In addition, we compared the 28-day mortality in patients with lactate and without lactate values. There was no signi cant difference between the two groups (p=0.281). Finally, further forward-looking, large sample validation and risk grading studies are necessary to conduct to nd more simple and effective prediction methods and achieve targeted interventions to reduce the incidence of death.

Conclusions
A signi cant trend in higher mortality was observed in patients with elevated lactate levels compared to patients with less level of lactate in the nonelderly group. Lactate level was positively associated with risk of 28-day mortality of the non-elderly sepsis patients, but there was no signi cant association between lactate level and 28-day mortality in the elderly group.
Abbreviations ICU: Intensive Care Unit; MIMIC-III: the Medical Information Mart for Intensive Care III; ROC: receiver operating characteristic; AUROC: the area under the receiver operating characteristic curve; OR: odd ratio; SpO2, peripheral capillary oxygen saturation; MICU, medical intensive care unit; CCU, coronary care unit; TSICU, trauma surgical intensive care unit; CSRU, cardiac surgery recovery unit; SICU, surgical intensive care unit; SAPSII, simpli ed acute physiology score II;SOFA, sequential organ failure assessment; LODS: logistic organ dysfunction system; OASIS: oxford acute severity of illness score; QSOFA: quick sepsis related organ failure assessment; SIRS: systemic in ammatory response syndrome.

Declarations Acknowledgments
The authors would like to thank MIMIC program for access to the database.

Funding
No funding was obtained for this study.

Availability of data and materials
The datasets analyzed during the current study are available in https://github.com/MIT-LCP/mimic-code/tree/master/concepts.

Authors' contributions
YHD designed the methods and experiments, and contributed to the writing of manuscript. XYM and YFH cleaned the data. YYH, JC and YRL provided guidance and reviewed the manuscript critically. JYP supervised the study and revised the paper. All authors read and approved the nal manuscript.

Ethics declarations
Ethics approval and consent to participate Ethical consent was not required in this study, since the MIMIC data were analyzed namelessly.

Consent for publication
The manuscript does not include individual person's data.  Adjust II model adjust for: age; gender; rst care unit; SAPSII. Table 5. Subgroup analysis of the association between lactate and 28-day mortality in the two age groups.  Receiver operating characteristic curves for predicting 28-day mortality in non-elderly patients.

Figure 3
Receiver operating characteristic curves for predicting 28-day mortality in elderly patients. Figure 4