Critical Care

Association of admission lactate with mortality in adult patients with severe community-acquired pneumonia

a b s t r a c t

Purpose: The present study was conducted to investigate the association of admission lactate with mortality in severe community-acquired pneumonia (SCAP).

Methods: We performed a retrospective, observational, cohort study on adult SCAP patients admitted to intensive care unit (ICU) in West China Hospital of Sichuan University between December 2011 and December 2018. The primary outcome was hospital mortality. Univariate and then multivariate analysis were performed to identify independent risk factors for hospital mortality. The association of admission lactate categories with hospital mor- tality was examined in three logistic regression models and Kaplan-Meier plots. We also applied restricted cubic splines to estimate the potential non-linear associations.

Results: In total, 2275 SCAP patients were included. Admission lactate remained a significant factor for mortality after multivariate regression (OR: 1.085; 95% CI: 1.033,1.141; by continuous variable). After lactate was catego- rized into quartiles and the confounders were fully adjusted, compared with the quartile 1, ORs (95% CIs) of hospital mortality for quartile 2, quartile 3 and quartile 4 were 1.001 (0.759-1.321), 1.153 (0.877-1.516) and

1.593 (1.202-2.109), respectively (P for trend =0.001). survival curves indicated that elevated lactate was associated with poor prognosis (P < 0.001). Moreover, this association was non-linear, indicating that increased lactate has the most notable impact on mortality within the range of 1.5 to 4 mmol/L (P non-linear: 0.029 for hos- pital mortality; 0.004 for ICU mortality). Conclusion: Elevated admission lactate has a significant, independent, and potentially non-linear association with increased mortality in SCAP patients.

(C) 2022

  1. Background

Abbreviations: SCAP, severe community-acquired pneumonia; ICU, intensive care unit; OR, odds ratio; 95% CI, 95% confidence interval; AUC, area under the curve; PSI, pneu- monia severity index; AST, aspartate aminotransferase; CRP, C-reactive protein.

* Corresponding author at: Department of Respiratory and Critical Care Medicine, West China Hospital, Sichuan University, No. 37 Guoxue Alley, Chengdu, 610041, Sichuan, China.

?? Corresponding author at: Institute of Clinical Pathology, Key Laboratory of Transplant Engineering and Immunology, NHC, West China Hospital, Sichuan University, No. 37 Guoxue Alley, Chengdu, 610041, Sichuan, China.

E-mail addresses: [email protected] (Y. Shi), [email protected] (Z. Liang).

1 Dong Huang, Dingxiu He, Rong Yao and Wen Wang contributed equally to this work.

community-acquired pneumonia , defined as Acute infection of the lung parenchyma, is among the leading causes of substantial morbidity around the world during recent decades. Meanwhile, there is consistent evidence that the rate of hospitalization, the proportion of patients admitted to intensive care unit (ICU) due to severe community-acquired pneumonia (SCAP) and the overall mortality all remain enormously high [1]. According to previous reports from the United States and Europe, the annual incidence of pneumonia was

24.8 cases per 10,000 adults and 21% of CAP patients required intensive care [2]. Moreover, the ICU mortality and hospital mortality of SCAP vary from 30% to 50% [3].

https://doi.org/10.1016/j.ajem.2022.12.036

0735-6757/(C) 2022

The pneumonia severity index and the CURB-65 score are two most commonly used Severity assessment tools in CAP. However, a meta-analysis concluded that the negative and positive predictive values of them to predict mortality are probably limited [4]. Considering the prognostic relevance of early adequate evaluation, more researches about association between various admission indicators and mortality in SCAP patients are warranted.

As a product of anaerobic metabolism and tissue hypoxia, lactate level >2 mmol/L is a well-known biomarker in patients with sepsis currently. [5] Recently, a systematic review revealed that a better out- come was associated with decreasing blood lactate concentrations in critically ill patients, which was not limited to septic patients [6]. It has been also shown that implementation of severe Sepsis Bundles could clearly reduce mortality from severe CAP [7]. Furthermore, fol- lowing the Sepsis-3 flowchart involving lactate could resulted in better identification of CAP patients at high risk of mortality [8].

The association of lactate with mortality in SCAP patients has been investigated in previous studies. However, the conclusions are still conflicting and lack external validity. For instance, a study with 86 CAP patients admitted to ICU suggested that increased lactate (>2 mmol/L) was a significant risk factors for in-ICU mortality [9]. In another single-center, prospective cohort study of 107 ICU patients with SCAP, a model based on lactate on day 3, rather than day 1, had good discrimi- nation for primary outcome (death or absence of improvement within 5 days of treatment) [10]. However, in a study of 614 ICU patients with pneumococcal CAP, from Bedos et al., only lactate >4 mmol/L was an independent predictors of hospital mortality (adjusted OR [95% CI] for lactate >4: 2.41 [1.27-4.56]; adjusted OR [95% CI] for lactate 2-4: 1.25 [0.68-2.30]; vs. lactate <2 mmol/L) [11]. Furthermore, one recent retrospective study with 160 ICU patients reported that lactate was not independently associated with in-hospital mortality of severe pneumonia after multivariate regression analysis in a Korean popula- tion cohort [12].

Early SCAP management is primarily based on severity assessment. As a convenient, widely available and low-cost Diagnostic biomarker, more comprehensive and detailed investigation and understanding the association of lactate with mortality of SCAP might be helpful for the stratification and identification of high-risk SCAP patients. Thus, the aim of the present study was to investigate the association of admis- sion lactate with mortality of SCAP in a larger cohort and after adjusting for more confounders.

  1. Methods
    1. Study design and cohort

We performed a retrospective, observational, cohort study in a 215-bed ICU of a large tertiary-care teaching hospital in Sichuan, China in accordance with the amended Declaration of Helsinki. The study was approved by the West China Hospital of Sichuan University Biomedical Research Ethics Committee (No. 2021-828). A waiver for written informed consent was obtained due to the retrospective noninterventional design.

With approximately 30-40 variables potentially associated with hospital mortality in SCAP patients, the minimum sample size required 800 deaths to follow the principle of at least twenty outcome events per variable (EPV) in the regression analysis [13]. Considering that the mortality of SCAP was approximately 40% in previous reports [3], the sample size of patients with SCAP was estimated to be approximately 2000. Therefore, adult SCAP patients with documented complete base- line clinical data admitted to ICU between December 2011 and Decem- ber 2018 were recruited.

According to the Infectious Diseases Society of America (IDSA)/ American Thoracic Society (ATS) guidelines, SCAP was defined as fulfilment of at least 1 major criterion (septic shock with need for vaso- pressors; respiratory failure requiring mechanical ventilation) or 3

minor criteria (respiratory rate >= 30 breaths/min; PaO2/FiO2 ratio <= 250; multilobar infiltrates; confusion /disorientation; blood urea nitrogen level >= 20 mg/dL; white blood cell count <4000 cells/uL; platelet count <100,000/uL; core temperature < 36 ?C; hypotension requiring aggressive fluid resuscitation) [14].

The exclusion criteria were as follows: (1) under 18 years old;

(2) pregnancy; (3) residents of long-term care facilities/nursing homes, or prior hospitalization within 30 days of study enrollment;

(4) unclear clinical outcomes; (5) severe immunosuppression: human immunodeficiency virus infection, autoimmune diseases, or immuno- suppressive therapy; (6) repeated admission.

All patients received Standard care and antibiotic therapy at the discretion of the ICU attending physician and based on the CAP guidelines [14].

    1. Study outcomes and measurements

The following clinical data within 24 h of admission to the ICU were collected anonymously from electronic medical records: demographic characteristics, co-morbidities, vital signs, and laboratory examinations, including hematological data, Biochemical parameters, inflammatory markers, coagulation indicators, etc. Lactate levels were obtained within the routinely performed admission arterial blood gas analysis. The first value was recorded for analysis if any laboratory examination was repeated more than once within 24 h of admission. Two trained and experienced physicians reviewed the medical records and completed the data collection by using a standardized data collection form independently. The physicians were blinded to the study hypothesis and their performances were monitored by the research team during data collection. Data were checked by a third reviewer if there was any disagreement.

Patient follow-up was until hospital discharge. The primary outcome established for this analysis was hospital mortality, and the secondary outcome was ICU mortality.

    1. Statistical analysis

Data were analyzed using IBM SPSS Statistical version 23.0 (SPSS, Chicago, IL, USA) and R software 4.1.2 (R Foundation for Statistical Com- puting). A two-sided p < 0.05 was considered statistically significant. Data are presented as medians (interquartile range, IQR) for continuous variables and counts (%) for categorical variables as appropriate. The nonparametric Kruskal-Wallis test, chi-square analysis and Fisher’s exact test were used to test for differences between groups as appropri- ate. Multiple imputation (MI) was applied to account for missing data by using Bayesian methods in SPSS.

To explore which variables could potentially modify the association of lactate with hospital mortality, the variables with P < 0.10 in univar- iate logistic regression analysis were included in the multivariate logis- tic regression analysis with forward LR (forward stepwise selection based on maximum likelihood estimation) to identify independent risk factors for hospital mortality. The results were reported as odds ra- tios (ORs) and 95% confidence intervals (95% CIs). A Spearman correla- tion analysis was carried out to test the correlations of the continuous variables among these factors.

All baseline clinical data were presented and compared stratified by lactate categories after lactate was categorized into quartiles. The P values for trend through the lactate categories were calculated by treating lactate as an ordinal variable with the lowest quartile as the ref- erence group. The linear trend between continuous variables and the different quartile of lactate was evaluated by Linear regression analysis. The trend between dichotomous variables’ positive rate and lactate was based on linear-by-linear association test in chi-square analysis.

The association of admission lactate categories with hospital mortal- ity was tested in three logistic regression models to systematically account for confounding factors. First, Model 1 was adjusted for no

covariates. Then, the demographic characteristics and comorbidities were adjusted in Model 2. Third, Model 3 was additionally adjusted for the vital signs and laboratory examinations based on Model 2. P values for trend were also tested by entering the lactate quartiles as a continuous variable into the regression models. Only the independent risk factors for hospital mortality in demographic characteristics, comorbidities, vital signs and laboratory examinations were selected for adjustment. Taking into account that the need for vasopressors is a main diagnostic criterion for both SCAP and septic shock [5,14], we did subgroup analyses to explore whether the association of lactate with mortality varied across patients with or without need for vasopres- sors during hospitalization.

Then, Kaplan-Meier plots with log-rank statistics were used to assess differences in survival across quartiles of lactate. Finally, we also applied restricted cubic splines to flexibly model and visualize the association of lactate level with mortality on a continuous scale and to estimate the potential non-linear associations [15]. It was performed using the Regression modeling Strategies (rms) package in R. To balance best fit and overfitting in the splines for hospital and ICU mortality, the locations of the knots were set at the 5th, 35th, 65th, and 95th percen- tiles with 4 knots. The number of knots used in the overall splines were also applied in splines for subgroup analyses. Analyses were multivariate-adjusted for all covariates listed in Model 3. The lactate was analyzed as a continuous variable, and ORs and 95% CIs were calcu- lated using the median lactate level as the reference value.

  1. Results
    1. Independent risk factors for hospital mortality

In total, 2513 SCAP patients were identified in the current study. Then, 238 patients were excluded according to the exclusion criteria. Among the remaining 2275 SCAP patients, 925 (40.7%) patients died during hospitalization (Fig. 1). The median age was 69 years old, and 67.3% of study cohort was male. Detailed comparison of baseline clinical characteristics in survival group and in-hospital death group was shown in supplementary materials Table S1.

A total of 32 variables with P < 0.10 in univariate logistic regression analysis were included in the multivariate analysis, which revealed that the independent risk factors for hospital mortality included demographic characteristics (age), comorbidities (chronic hepatic dis- eases, chronic renal diseases, chronic cardiovascular diseases and chronic pulmonary diseases), vital signs on admission (diastolic blood

SCAP patients (n= 2513)

Patients excluded (n = 238)

-under 18 years old (n = 27)

  • pregnancy (n =8)
  • residents of long-term care facilities/nursing-home, or prior hospitalisation within 30 days of study enrolment (n =70)
  • unclear clinical outcomes (n =85)
  • severe immunosuppression (n =16)
  • repeated admission (n=32)

Patients included in analysis (n =2275)

Fig. 1. Study population. SCAP: severe community-acquired pneumonia.

pressure and heart rate), and laboratory examinations (creatinine, Troponin T, platelet and lactate). Of note, lactate remained as a signifi- cant factor (adjusted OR: 1.085; 95% CI: 1.033-1.141; P = 0.001 by continuous variable). The detailed ORs and 95% CIs in univariate and multivariate analysis were summarized in Table 1. In Fig. 2, the lactate was especially, positively correlated with heart rate and negatively correlated with diastolic blood pressure and platelet significantly (P for spearman correlation analysis <0.05).

    1. Clinical characteristics of study cohort

Of the entire cohort, the median lactate was 1.5 mmol/L. The range of lactate levels were 0-1.1 mmol/L, 1.2-1.5 mmol/L, 1.6-2.2 mmol/L,

2.3-20 mmol/L in quartile 1, quartile 2, quartile 3, quartile 4, respectively.

As demonstrated in Table 2, significant linear trends across quintiles of lactate were observed for some clinical characteristics (P for trend

<0.05). For instance, SCAP patients in the highest quintile (quartile 4) had higher levels of heart rate, neutrophil, D-dimer, bilirubin, myoglobin, procalcitonin but lower levels of blood pressure, albumin, platelet compared with those in the lowest quartile (quartile 1). In keeping with expected trends, crude risks of hospital and ICU mortality were progressively higher over increasing quartiles of lactate (P value

<0.001; P for trend <0.001).

    1. Lactate is associated with mortality

Compared with individuals from quartile 1 of lactate levels, the unadjusted and multivariable adjusted ORs for hospital mortality for in- dividuals from quartile 2-4 in Model 1-3 were displayed in Fig. 3. The association was no longer statistically significant after adjustment for increased potential confounders in quartiles 2 and 3, but remained significant for quartile 4. In Model 3, Compared with the quartile 1, ORs (95% CIs) of hospital mortality for quartile 2, quartile 3 and quartile

4 were 1.001 (0.759-1.321), 1.153 (0.877-1.516) and 1.593

(1.202-2.109), respectively. On the other hand, we detected significant linear trends of mortality across quartiles of lactate in all Model 1-3 (P for trend <0.05).

In supplementary materials Table S2, the results of subgroup analysis indicated that the need for vasopressors did not significantly modify the association between lactate quartiles and hospital mortality for most of results. However, in those patients without need for vaso- pressors during hospitalization, increased lactate was not significantly associated with higher mortality in Model 3 (P for trend = 0.108).

As illustrated in Fig. 4A, elevated quartiles of lactate level were associated with a remarkably poor prognosis (P < 0.001). Meanwhile, similar estimate was observed when patients were divided into two groups in terms of median value of lactate (1.5 mmol/L) (Fig. 4B).

    1. The nonlinear association of lactate with mortality

As shown in Fig. 5, the estimated associations between lactate and hospital mortality and ICU mortality were both “J” shaped on a continuous scale, which indicated significant nonlinear relationships (P non-linear =0.029 for hospital mortality; P non-linear =0.004 for ICU mortality). We found no evidence of association between lactate and mortality for individuals with lactate under 1.5 mmol/L because the 95% CI included the OR of 1.0. We observed a marked rapidly in- crease in risk of mortality with lactate ranging from 1.5 to about 4 mmol/L, and then mild elevation in risk at higher lactate levels after- wards. Especially for individuals with lactate above 4 mmol/L, relatively flat curves demonstrated that the increase in risk of ICU mortality was less pronounced.

In the subgroup analysis in which lactate is considered as a continu- ous variable, the correlation of lactate with mortality was stronger and more non-linear among patients with need for vasopressors during

Univariate and multivariate logistics regression analysis of independent risk factors for hospital mortality in SCAP patients.

Risk factors

Univariate analysis

Multivariate analysis

OR (95% CI)

P

OR (95% CI)

P

Demographic characteristics

Age (years old)

1.014 (1.008,1.019)

<0.001

1.017 (1.010,1.023)

<0.001

Comorbidities

Chronic hepatic diseases (%)

1.520 (0.959,2.410)

0.075

1.810 (1.056,3.102)

0.031

Chronic renal diseases (%)

3.779 (2.729,5.232)

<0.001

3.615 (2.466,5.299)

<0.001

Chronic cardiovascular diseases (%)

2.268 (1.816,2.832)

<0.001

2.130 (1.631,2.782)

<0.001

Chronic pulmonary diseases (%)

1.803 (1.491,2.179)

<0.001

1.776 (1.421,2.219)

<0.001

Cancer (%)

0.653 (0.521,0.819)

<0.001

Chronic hematological diseases (%)

2.483 (1.514,4.070)

<0.001

Diabetes (%)

1.323 (1.058,1.654)

0.014

Hypertension (%)

0.578 (0.475,0.702)

<0.001

chronic cerebrovascular diseases (%)

2.740 (1.388,5.409)

0.004

Vital signs

Diastolic blood pressure

0.991 (0.986,0.996)

<0.001

0.993 (0.987,0.999)

0.027

Heart rate

1.008 (1.004,1.011)

<0.001

1.008 (1.003,1.012)

<0.001

Systolic blood pressure (mmHg)

0.997 (0.994,1.000)

0.054

Unconsciousness or insanity (%)

0.809 (0.648,1.009)

0.061

Laboratory examinations

Creatinine (umol/L)

1.002 (1.001,1.002)

<0.001

1.001 (1.001,1.002)

0.001

Troponin T (ng/L)

1.001 (1.000,1.001)

0.001

1.000 (1.000,1.001)

0.038

Platelet (x10 9 /L)

0.997 (0.997,0.998)

<0.001

0.999 (0.998,1.000)

0.006

Lactate (mmol/L)

1.109 (1.065,1.154)

<0.001

1.085 (1.033,1.141)

0.001

Neutrophil (x10 9 /L)

1.033 (1.018,1.048)

<0.001

Lymphocyte (x10 9 /L)

0.842 (0.757,0.937)

0.002

BUN (mmol/L)

1.029 (1.019,1.039)

<0.001

Hemoglobin (g/L)

0.997 (0.993,1.000)

0.060

APTT (s)

1.014 (1.007,1.020)

<0.001

PT (s)

1.012 (1.001,1.023)

0.029

Myoglobin (ng/mL)

1.000 (1.000,1.000)

<0.001

BNP (pg/mL)

1.000 (1.000,1.000)

<0.001

Uric acid (umol/L)

1.001 (1.001,1.002)

<0.001

CRP (mg/L)

1.003 (1.001,1.004)

<0.001

Procalcitonin (ng/mL)

1.009 (1.002,1.016)

0.012

Albumin (g/L)

0.981 (0.965,0.997)

0.019

IL-6 (pg/mL)

1.000 (1.000,1.000)

<0.001

Calcium (mmol/L)

0.419 (0.269,0.650)

<0.001

SCAP: Severe Community-acquired Pneumonia; OR: odds ratio; 95% CI: 95% confidence interval.

The age, diastolic blood pressure, heart rate, creatinine, Troponin T, platelet,and lactate were continuous variables. The chronic hepatic diseases, chronic renal diseases, chronic cardio- vascular diseases, chronic pulmonary diseases were categorical variables.

Image of Fig. 2

Fig. 2. A Spearman correlation analysis for the continuous variables among independent risk factors of mortality.

hospitalization than among patients without vasopressors (supplemen- tary materials Fig. S1 and Fig. S2).

  1. Discussion

To the best of our knowledge, this is the first study with over 2000 SCAP patients to report the association between admission lactate, both as a continuous and categorical variable, and mortality. In the pres- ent study, we identified 11 independent risk factors at admission for hospital mortality of SCAP. An increase in lactate level was significantly associated with increased risk of mortality even after adjusting for age, comorbidities, diastolic blood pressure, heart rate, creatinine, Troponin T and platelet. Moreover, this association was non-linear, indicating that increased lactate has the most notable impact on mortality within the range of 1.5 to 4 mmol/L. When further stratifying SCAP patients, we found that the correlation was more significant in patients with need for vasopressors during hospitalization.

The utilization of lactate in risk stratification and prognosis prediction among CAP patients is increasing rapidly. For example, a ret- rospective study showed admission lactate predicts poor prognosis in- dependently of the CURB-65 scores with an optimal cutoff of >1.8 mmol/L (area under the curve [AUC]: 0.67) [16]. In another retrospec- tive study consisting of 1031 patients with CAP, multivariable analysis identified lactate >1.7 mmol/L (OR: 2.59, 95% CI: 1.53-4.38) as a

Table 2

Baseline clinical characteristics across quintiles of lactate in SCAP patients.

Variables

Lactate Q1

(0-1.1 mmol/L),

n = 668 (29.4%)

Lactate Q2

(1.2-1.5 mmol/L),

n = 540 (23.7%)

Lactate Q3

(1.6-2.2 mmol/L),

n = 534 (23.5%)

Lactate Q4

(2.3-20 mmol/L),

n = 533 (23.4%)

P

P for trend

Demographic Characteristics

Age (years old)

69.00 (55.75, 79.00)

71.00 (58.75, 81.00)

69.00 (56.00, 79.00)

67.00 (53.00, 78.00)

0.004

0.054

Sex: male (%)

460 (68.9)

364 (67.4)

361 (67.6)

345 (64.7)

0.5

0.160

Comorbidities

Cancer (%)

116 (17.4)

107 (19.8)

85 (15.9)

101 (18.9)

0.35

0.869

Chronic hematological diseases (%)

22 (3.3)

11 (2.0)

13 (2.4)

23 (4.3)

0.132

0.349

Diabetes (%)

107 (16.0)

86 (15.9)

91 (17.0)

89 (16.7)

0.95

0.648

Chronic hepatic diseases (%)

20 (3.0)

16 (3.0)

20 (3.7)

19 (3.6)

0.84

0.453

Chronic renal diseases (%)

39 (5.8)

46 (8.5)

42 (7.9)

59 (11.1)

0.012

0.003

Chronic cardiovascular diseases (%)

105 (15.7)

96 (17.8)

89 (16.7)

97 (18.2)

0.661

0.341

Hypertension (%)

192 (28.7)

151 (28.0)

140 (26.2)

136 (25.5)

0.576

0.165

Chronic pulmonary diseases (%)

160 (24.0)

143 (26.5)

146 (27.3)

137 (25.7)

0.575

0.402

Chronic cerebrovascular diseases (%)

10 (1.5)

9 (1.7)

6 (1.1)

12 (2.3)

0.528

0.481

Vital Signs

Respiratory rate (breath/min)

20 (16, 24)

20 (16, 25)

21 (17, 25)

22 (18, 27)

<0.001

<0.001

Systolic blood pressure (mmHg)

131.00 (114.00, 148.00)

128.00 (111.00, 143.00)

124.00 (108.00, 143.25)

119.00 (102.00, 138.00)

<0.001

<0.001

Diastolic blood pressure (mmHg)

72 (63, 85)

70 (60, 82)

69 (59, 80)

68 (57, 79)

<0.001

<0.001

Temperature (?C)

36.6 (36.3, 37.0)

36.7 (36.4, 37.1)

36.6 (36.3, 37.0)

36.8 (36.4, 37.4)

<0.001

<0.001

Heart rate (beat/min)

96.0 (81.0, 110.0)

98.5 (83.0, 114.0)

100.0 (88.0, 118.0)

107.0 (90.0, 123.5)

<0.001

<0.001

Unconsciousness or insanity

121 (18.1)

93 (17.2)

109 (20.4)

85 (15.9)

0.277

0.661

Laboratory Examinations White blood cell (x10 9 /L)

10.07 (7.27, 13.91)

9.89 (6.58, 13.79)

9.95 (6.69, 14.03)

9.89 (6.75, 13.66)

0.592

0.473

Neutrophil (x10 9 /L)

7.25 (4.70, 10.83)

7.39 (4.73, 10.65)

7.77 (4.92, 12.19)

8.28 (5.09, 12.89)

0.013

<0.001

Lymphocyte (x10 9 /L)

0.86 (0.55, 1.36)

0.86 (0.51, 1.30)

0.76 (0.46, 1.14)

0.71 (0.36, 1.19)

<0.001

0.083

BUN (mmol/L)

8.60 (5.70, 14.64)

7.96 (5.50, 13.70)

8.90 (5.79, 13.87)

9.70 (6.20, 15.75)

0.023

0.853

D-dimer (mg/L)

3.90 (2.02, 8.50)

4.40 (2.30, 9.86)

4.33 (2.14, 10.79)

5.33 (2.51, 11.63)

<0.001

<0.001

Total bilirubin (u mol/L)

9.90 (6.60, 14.90)

10.00 (6.32, 14.88)

10.70 (7.30, 16.50)

11.40 (7.30, 19.00)

<0.001

<0.001

direct bilirubin (u mol/L)

5.30 (3.40, 8.20)

5.05 (3.10, 8.47)

6.00 (3.80, 10.10)

6.20 (3.80, 11.93)

<0.001

<0.001

Globulin (g/L)

24.50 (20.60, 28.90)

25.10 (21.40, 28.80)

24.95 (21.37, 28.52)

24.70 (20.80, 29.17)

0.581

0.885

Monocyte (x10 9 /L)

0.42 (0.26, 0.64)

0.39 (0.23, 0.60)

0.37 (0.19, 0.60)

0.32 (0.15, 0.54)

<0.001

0.969

Hemoglobin (g/L)

102 (84, 120)

102 (82, 117)

101 (82, 117)

98 (82, 116)

0.104

0.014

APTT (s)

33.30 (29.08, 39.42)

34.10 (28.50, 40.73)

34.20 (29.45, 41.90)

35.30 (30.00, 45.50)

<0.001

<0.001

PT (s)

13.10 (12.07, 14.53)

13.30 (12.20, 14.70)

13.40 (12.20, 15.20)

13.90 (12.40, 16.30)

<0.001

<0.001

Fibrinogen (g/L)

3.85 (2.83, 4.98)

3.69 (2.65, 4.88)

3.85 (2.71, 5.12)

3.45 (2.17, 4.93)

<0.001

0.009

Creatinine (umol/L)

71.00 (51.00, 140.90)

68.00 (50.00, 104.55)

72.00 (51.00, 121.00)

76.00 (54.00, 137.00)

0.026

0.326

Myoglobin (ng/mL)

99.11 (47.36, 259.70)

103.80 (46.09, 261.00)

108.65 (50.77, 370.35)

170.20 (68.60, 446.65)

<0.001

<0.001

ALT (IU/L)

21 (13, 44)

22 (13, 43)

24 (13, 45)

26 (15, 51)

0.003

0.110

AST (IU/L)

30.0 (20.0, 51.0)

30.0 (20.0, 55.0)

34.5 (22.0, 62.0)

42.0 (24.0, 79.0)

<0.001

0.001

Troponin T (ng/L)

40.20 (20.52, 80.52)

37.70 (19.60, 81.30)

42.20 (22.40, 99.38)

50.85 (23.20, 122.07)

0.001

0.463

BNP (pg/mL)

1451.5 (531.5, 4970.0)

1555.0 (545.5, 5229.0)

1811.0 (677.0, 5070.0)

2301.0 (662.0, 7396.0)

0.011

0.061

Uric acid (u mol/L)

205.90 (122.80, 336.10)

192.50 (128.00, 304.75)

211.00 (132.80, 325.25)

230.70 (144.00, 349.00)

0.014

0.194

Glucose (mmol/L)

7.51 (5.92, 10.20)

7.65 (5.80, 10.39)

7.73 (6.16, 10.63)

8.39 (6.30, 11.50)

0.001

<0.001

CRP (mg/L)

54.05 (15.33, 104.00)

58.00 (17.78, 108.00)

56.50 (14.70, 116.00)

78.30 (26.75, 144.00)

<0.001

<0.001

Procalcitonin (ng/mL)

0.29 (0.10, 1.15)

0.30 (0.11, 0.98)

0.45 (0.12, 1.71)

0.88 (0.18, 4.06)

<0.001

<0.001

Albumin (g/L)

30.80 (28.00, 34.40)

30.40 (27.50, 33.70)

30.10 (26.60, 33.92)

29.50 (26.30, 33.10)

<0.001

<0.001

IL-6 (pg/mL)

35.54 (14.93, 107.45)

42.37 (15.14, 111.72)

48.93 (16.12, 161.55)

76.64 (22.84, 320.60)

<0.001

<0.001

Platelet (x10 9 /L)

179.0 (112.0, 265.5)

176.0 (104.0, 258.0)

168.0 (97.0, 263.0)

129.5 (61.0, 212.5)

<0.001

<0.001

Calcium (mmol/L)

2.08 (1.98, 2.19)

2.08 (1.96, 2.19)

2.08 (1.94, 2.20)

2.04 (1.92, 2.17)

0.003

<0.001

Clinical outcomes ICU mortality (%)

193 (28.9)

183 (33.9)

177 (33.1)

239 (44.8)

<0.001

<0.001

Hospital mortality (%)

230 (34.4)

213 (39.4)

214 (40.1)

268 (50.3)

<0.001

<0.001

Data are shown as median with interquartile range (IQR) for continuous variables or number with percentage for categorical variables. SCAP: Severe Community-acquired Pneumonia; n: numbers; BUN: blood urea nitrogen; APTT: activated partial thromboplastin time; PT: prothrombin time; ALT: alanine aminotransferase; AST: aspartate aminotransferase; BNP: Brain Natriuretic Peptide; CRP: C-reactive protein;

prognostic factor for the prediction of mortality [17]. Moreover, a prospective observational clinical study of 1641 patients in emergency department indicated lactate is superior to CURB-65 in predicting mor- tality, hospitalization and ICU admission in pneumonia patients in the emergency department with the cut-off values from 3 to 6 mmol/L [18]. As a continuous variable, it has also been reported as a significant factor (OR 1.24; 95% CI 1.01-1.53) for inpatient mortality in multivari- able analysis after adjusting for laboratory variables and PSI [19]. The combination of lactate with other biomarkers has also been investi- gated. In a cohort of 8284 CAP hospitalizations, a prediction model including lactate >2 mmol/L showed high predictive value of need for ICU with an AUC of 0.91 [20]. Nevertheless, the conflicts still exist. Zhou et al. reported that admission lactate did not predict well the out- comes or improve the common severity scores in 340 enrolled septic

patients with CAP [21]. The diverse inclusion criteria of population, different baseline features or outcome definitions, heterogeneity in detection time of lactate, residual confounding might partly explain the discrepant results among previous studies, such as the independent association and Predictive power.

Taking into count that the heterogenous severity of CAP varies from mild symptoms to life-threatening critical illness, it remains necessary to evaluate and validate the association of admission lactate with hospital and ICU mortality of SCAP among different populations with adjustment for more confounders. To date, a limited number of studies have examined mortality in relation to lactate in SCAP patients. Never- theless, controversies are ongoing around previous results [9-12]. Moreover, the optimal cut-off values, which were often calculated via receiver operating characteristic curves, were not uniform in

Image of Fig. 3

Fig. 3. Association of quartiles of lactate with hospital mortality. Model 1: unadjusted. Model 2: adjusted for age, chronic hepatic diseases, chronic renal diseases, chronic cardiovascular diseases and chronic pulmonary diseases. Model 3: adjusted for age, chronic hepatic diseases, chronic renal diseases, chronic cardiovascular diseases, chronic pulmonary diseases, diastolic blood pressure, heart rate, creatinine, Troponin T and platelet. OR: odds ratio; 95% CI: 95% confidence interval.

previous studies. Therefore, it is generally recommended, based on practice guidelines of sepsis, that elevated lactate levels (>2 mmol/L) measured during routine assessment should prompt careful clinical evaluation of septic organ dysfunction or shock, and the need for fluid resuscitation in SCAP [22,23].

Our study cohort has been shown to be broadly representative of the general SCAP population in terms of baseline clinical characteristics [9-12]. In agreement with one previous research reporting that serum lactate level was positively correlated with respiratory rate, blood

pressure, neutrophil, procalcitonin, etc. in CAP patients [24], our results also demonstrated the notable correlation between increased lactate and higher heart rate, lower diastolic blood pressure and lower platelet. Furthermore, the linear associations between lactate quartiles and more baseline characteristics, such as D-dimer, AST, CRP, etc. were found. However, these findings still need to be proved by further investiga- tions, which might be useful for comprehensive evaluation of acute change of pathophysiological status in SCAP patients. Another interest- ing result of the current study is that none of the variables included in

Image of Fig. 4

Fig. 4. Survival curves of SCAP patients at 90-day follow-up. (A) grouped by quartiles of lactate; (B) grouped by median value of lactate.

Image of Fig. 5

Fig. 5. Association of admission lactate with hospital mortality (A), and with ICU mortality (B). The restricted cubic spline regression models with 4 knots. ORs and 95% CIs were estimated using the median lactate level as the reference value (1.5 mmol/L). The horizontal dashed line represents the reference OR of 1.0. The model was adjusted for age (continuous), chronic hepatic diseases (categorical), chronic renal diseases (categorical), chronic cardiovascular diseases (categorical), chronic pulmonary diseases (categorical), diastolic blood pressure (con- tinuous), heart rate (continuous), creatinine (continuous), Troponin T (continuous) and platelet (continuous). OR, odds ratio; 95% CI, 95% confidence interval.

qSOFA (respiratory rate, mental status and systolic blood pressure) was an independent risk factor for mortality from SCAP. More studies about the sensitivity and discriminative ability of qSOFA in SCAP patients are needed.

In the current study, the admission lactate was obtained prior to the initiation of vasopressors and stabilization of blood pressure in most patients. The median lactate level of the SCAP patients in the current study was only 1.5 mmol/L, which means that only less than half the SCAP patients have hyperlactatemia at the time of admission to ICU. However, 62.3% of patients received vasopressors to maintain the mean arterial pressure during hospitalization, which potentially indi- cated possible inadequate, insufficient, or undetected elevation of lac- tate at admission. Nevertheless, we still observed the significant association of admission lactate with mortality. Our findings suggest that there is evidence to support for early personalized evaluation for disease status of SCAP patients based on admission lactate level, and that attention should be also paid to patients with admission lactate levels under 2 mmol/L. Further studies which could account for the as- sociation of variation trends of lactate after admission with clinical out- comes are warranted. Furthermore, the necessity and benefits of repeated lactate measurement in which subgroup of SCAP population needs to be elucidated.

We analyzed lactate as a continuous variable using restricted cubic spline to avoid any loss of information about its association with risk death, which has not been reported in the literature. We observed important non-linearity in the association even after accounting for potential confounding factors. In the restricted cubic spline, the SCAP patients with lower lactate levels than 1.5 mmol/L seem to have increased mortality. However, the relationship isn’t significant. We speculated that one possible reason is that the splines are normalizing to the mean when there isn’t much data. But most importantly, the sig- nificantly Rapid increase in risk of mortality in the range of 1.5-4 mmol/L could be attributed to the role of lactate as an indicator of acid-base homeostasis, perfusion status, and organ dysfunction. The non- significant increase in mortality risk in the higher lactate range (>=4) is likely due partly to small proportion of patients or underlying other health conditions which might strongly influence mortality for those especially critical ill patients. Further large-scale, multicenter, prospec- tive studies are required to determine, adjust and recalibrate evidence-based optimal cut-off values for admission lactate in SCAP.

In this work, we stratified SCAP patients according to the need for vasopressors. One unexpected finding is that the independent and

non-linear association was more intensified in patients with hypoten- sion. However, we should be cautious about explaining this result. As an early marker of organ failure and hypoperfusion, admission hyperlactatemia has been reported to be associated with increased mortality in ICU and emergency department, irrespective of presence of hypotension, organ failure or shock [25,26]. Potential biological mechanisms that might explain the different association between ad- mission lactate level and increased risk of death in different subgroups remain to be explored.

Several limitations need to be acknowledged while interpreting the results. First, this was a single-center, retrospective study with routinely collected data in a limited number of patients. The interobserver reliability was not tested or measured. The observational design did not permit to establish causalities among high lactate, severity of illness and mortality. Some potential study population who had incomplete or unavailable clinical data were not included. In addition, we did not per- form subgroup analysis in patients with and without respiratory failure necessitating mechanical ventilation considering that the proportion of patients without mechanical ventilation is especially low. We did not have more external validation study data for our results. These condi- tions could have led to bias in our results. Second, we cannot completely rule out the possibility of unmeasured or undetected confounders which may account for the associations observed in the present study, such as interventions and antibiotic therapies. In addition, in multivari- ate logistic regression analysis, we used stepwise selection which might produce biased parameter estimates sometimes. It probably gives mis- leading results if irrelevant nuisance variables could be coincidentally significant. Thus, this regression model might fit the data well in- sample but do poorly out-of-sample [27]. Third, we lacked dynamic data of lactate levels after admission and failed to perform follow-up after discharge due to the scarcity of relevant information.

  1. Conclusions

In summary, the hospital mortality of SCAP was 40.7% and the median value of admission lactate was 1.5 mmol/L. We identified 11 in- dependent risk factors on admission for hospital mortality of SCAP. Our results indicated that an increase in lactate levels was significantly asso- ciated with elevated risk of death among SCAP patients. Compared with patients in the lowest quartile of lactate, patients in the highest quartile were at the highest risk of hospital mortality. Moreover, the non-linear association was most pronounced in patients with lactate ranging from

1.5 to 4 mmol/L. However, our conclusions still need to be confirmed further in future studies.

Ethics approval and consent to participate

Ethics approval was obtained from the West China Hospital of Sichuan University Biomedical Research Ethics Committee (No.2021-828), with waiver of written informed consent due to retro- spective non-interventional design. All patient data was maintained with confidentiality.

Consent for publication

Consent for publication was provided by all authors.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Funding

This work was supported by the National Natural Science Foundation of China (82072156), the Science and Technology Depart- ment of Sichuan Province (2019YFS0443, 2018JY0389, 2022NSFSC1313), key program of Sichuan Provincial Health Commis- sion (18ZD002), and the Fundamental Research Funds for the Central Universities (2022SCU12055).

CRediT authorship contribution statement

Dong Huang: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Conceptual- ization. Dingxiu He: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Rong Yao: Writing – review & editing, Supervision, Funding acquisition, Con- ceptualization. Wen Wang: Writing – review & editing, Data curation. Qiao He: Writing – review & editing, Data curation. Zhenru Wu: Writ- ing – review & editing, Conceptualization. Yujun Shi: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Zongan Liang: Writing – review & editing, Supervision, Resources, Methodol- ogy, Data curation, Conceptualization.

Declaration of Competing Interest

The authors declare that they have no competing interests.

Acknowledgements

None.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2022.12.036.

References

  1. Torres A, Chalmers JD, Dela Cruz CS, et al. Challenges in severe community-acquired pneumonia: a point-of-view review. Intensive Care Med. 2019;45(2):159-71. https://doi.org/10.1007/s00134-019-05519-y.
  2. Jain S, Self WH, Wunderink RG, et al. Community-acquired pneumonia requiring hospitalization among U.S. Adults. N Engl J Med. 2015;373(5):415-27. https://doi. org/10.1056/NEJMoa1500245.
  3. Woodhead M, Welch CA, Harrison DA, Bellingan G, Ayres JG. Community-acquired pneumonia on the intensive care unit: secondary analysis of 17,869 cases in the

ICNARC case mix Programme database. Crit Care. 2006;10 Suppl 2(Suppl. 2):S1. https://doi.org/10.1186/cc4927.

  1. Loke YK, Kwok CS, Niruban A, Myint PK. Value of severity scales in predicting mor- tality from community-acquired pneumonia: systematic review and meta-analysis. Thorax. 2010;65(10):884-90. https://doi.org/10.1136/thx.2009.134072.
  2. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus def- initions for Sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):801-10. https:// doi.org/10.1001/jama.2016.0287.
  3. Vincent JL, Quintairos E, Silva A, Couto Jr L, Taccone FS. The value of blood lactate ki- netics in critically ill patients: a systematic review. Crit Care. 2016;20(1):257. Pub- lished 2016 Aug 13. https://doi.org/10.1186/s13054-016-1403-5.
  4. Guo Q, Li HY, Li YM, et al. Compliance with severe sepsis bundles and its effect on patient outcomes of severe community-acquired pneumonia in a limited resources country. Arch Med Sci. 2014;10(5):970-8. https://doi.org/10.5114/aoms.2014. 46216.
  5. Ranzani OT, Prina E, Menendez R, et al. New sepsis definition (Sepsis-3) and community-acquired pneumonia mortality. A validation and clinical decision- making study. Am J Respir Crit Care Med. 2017;196(10):1287-97. https://doi.org/ 10.1164/rccm.201611-2262OC.
  6. Celikhisar H, Dasdemir Ilkhan G, Arabaci C. Prognostic factors in elderly patients ad- mitted to the intensive care unit with community-acquired pneumonia. Aging Male. 2020;23(5):1425-31. https://doi.org/10.1080/13685538.2020.1775192.
  7. Pereira JM, Laszczynska O, Azevedo A, et al. Early prediction of treatment failure in severe community-acquired pneumonia: the PRoFeSs score. J Crit Care. 2019;53: 38-45. https://doi.org/10.1016/j.jcrc.2019.05.020.
  8. Bedos JP, Varon E, Porcher R, et al. Host-pathogen interactions and prognosis of crit- ically ill immunocompetent patients with pneumococcal pneumonia: the nation- wide prospective observational STREPTOGENE study. Intensive Care Med. 2018;44 (12):2162-73. https://doi.org/10.1007/s00134-018-5444-x.
  9. Baek MS, Park S, Choi JH, Kim CH, Hyun IG. Mortality and Prognostic prediction in very elderly patients with severe pneumonia. J Intensive Care Med. 2020;35(12): 1405-10. https://doi.org/10.1177/0885066619826045.
  10. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373-9. https://doi.org/10.1016/s0895-4356(96)00236-3.
  11. Metlay JP, Waterer GW, Long AC, et al. Diagnosis and treatment of adults with community-acquired pneumonia. An official clinical practice guideline of the American Thoracic Society and Infectious Diseases Society of America. Am J Respir Crit Care Med. 2019;200(7):e45-67. https://doi.org/10.1164/rccm.201908-1581ST.
  12. Harrell Jr Frank E. Regression modeling strategies, with applications to linear models, logistic and ordinal regression, and survival analysis. Cham: Springer; 2015.https://doi.org/10.1007/978-3-319-19425-7.
  13. Frenzen FS, Kutschan U, Meiswinkel N, Schulte-Hubbert B, Ewig S, Kolditz M. Admis- sion lactate predicts poor prognosis independently of the CRB/CURB-65 scores in community-acquired pneumonia. Clin Microbiol Infect. 2018;24(3):306.e1-6. https://doi.org/10.1016/j.cmi.2017.07.007.
  14. Ahn JH, Choi EY. Expanded A-DROP Score: A New Scoring System for the Prediction of Mortality in Hospitalized Patients with Community-acquired Pneumonia. Sci Rep. 2018;8(1):14588. Published 2018 Oct 1. https://doi.org/10.1038/s41598-018-32

750-2.

  1. Chen YX, Li CS. Lactate on emergency department arrival as a predictor of mortality and site-of-care in pneumonia patients: a cohort study. Thorax. 2015;70(5):404-10. https://doi.org/10.1136/thoraxjnl-2014-206461.
  2. Gwak MH, Jo S, Jeong T, et al. Initial serum lactate level is associated with inpatient mortality in patients with community-acquired pneumonia. Am J Emerg Med. 2015; 33(5):685-90. https://doi.org/10.1016/j.ajem.2015.03.002.
  3. Gearhart AM, Furmanek S, English C, Ramirez J, Cavallazzi R. Predicting the need for ICU admission in community-acquired pneumonia. Respir Med. 2019;155:61-5. https://doi.org/10.1016/j.rmed.2019.07.007.
  4. Zhou HJ, Lan TF, Guo SB. outcome prediction value of National Early Warning Score in septic patients with community-acquired pneumonia in emergency department: a single-center retrospective cohort study. World J Emerg Med. 2020;11(4):206-15. https://doi.org/10.5847/wjem.j.1920-8642.2020.04.002.
  5. Phua J, Dean NC, Guo Q, Kuan WS, Lim HF, Lim TK. Severe community-acquired pneumonia: timely management measures in the first 24 hours. Crit Care. 2016;20

(1):237. Published 2016 Aug 28. https://doi.org/10.1186/s13054-016-1414-2.

  1. Kolditz M, Braeken D, Ewig S, Rohde G. Severity assessment and the immediate and long-term prognosis in community-acquired pneumonia. Semin Respir Crit Care Med. 2016;37(6):886-96. https://doi.org/10.1055/s-0036-1592127.
  2. Ning P, Zheng Y, Luo Q, et al. Metabolic profiles in community-acquired pneumonia: developing assessment tools for disease severity. Crit Care. 2018;22(1):130. Pub- lished 2018 May 14. https://doi.org/10.1186/s13054-018-2049-2.
  3. Juneja D, Singh O, Dang R. Admission hyperlactatemia: causes, incidence, and impact on outcome of patients admitted in a general medical intensive care unit. J Crit Care. 2011;26(3):316-20. https://doi.org/10.1016/j.jcrc.2010.11.009.
  4. Mikkelsen ME, Miltiades AN, Gaieski DF, et al. Serum lactate is associated with mor- tality in severe sepsis independent of organ failure and shock. Crit Care Med. 2009; 37(5):1670-7. https://doi.org/10.1097/CCM.0b013e31819fcf68.
  5. Smith G. Step away from stepwise. J Big Data. 2018;5:32. https://doi.org/10.1186/ s40537-018-0143-6.