Abstract
Introduction
The Emergency Severity Index (ESI) is a widely used triage tool in emergency departments (EDs) and is associated with adverse outcomes. Lactate has also been associated with mortality in various ED populations. This study evaluated whether lactate levels measured after triage at the treating physician’s discretion are associated with in-hospital mortality across ESI levels in admitted ED patients.
Methods
This study is a secondary analysis of prospectively collected data from adult patients presenting to the ED of a Swiss tertiary care center in 2013, 2015, 2017, and 2019. The study included admitted patients who underwent post-triage lactate measurement at the treating physician’s discretion. Only the first lactate value measured in the ED was analyzed and categorized as < 2.0, 2.0–3.9, and ≥ 4.0 mmol/L. The primary outcome was in-hospital mortality; secondary outcomes were 30-day mortality and intensive care unit (ICU) admission. Associations were assessed using multivariable logistic regression models adjusted for age and sex. Model discrimination was compared using receiver operating characteristic curve analysis.
Results
Among 3,909 patients admitted from the ED who had a lactate drawn (of 17,327 ED presentations during the study period; 22.6%), the median age was 72 years and 2,062 (52.8%) were male. In-hospital mortality occurred in 168 (4.3%), 30-day mortality in 258 (6.6%), and ICU admission in 698 (17.9%). In ESI levels 2 (n = 1,826) and 3 (n = 1,722), crude in-hospital mortality was similar (59 [3.2%] and 44 [2.6%], respectively), but varied after stratification by lactate level. In ESI level 2 patients, in-hospital mortality occurred in 24 (2.0%) among those with lactate < 2.0 mmol/L (n = 1,210) and in 18 (17.8%) among those with lactate ≥ 4.0 mmol/L (n = 101), corresponding to an age- and sex-adjusted absolute risk difference of 16.0% (95% CI, 8.8–23.3). In ESI level 3, in-hospital mortality occurred in 20 (1.6%) at lactate < 2.0 mmol/L (n = 1,254) and in 8 (15.1%) at lactate ≥ 4.0 mmol/L (n = 53), corresponding to a risk difference of 12.5% (95% CI, 3.7–21.3). Discrimination for in-hospital mortality was higher when lactate was considered alongside ESI (area under the curve [AUC] 0.844 versus 0.799; ΔAUC 0.045, 95% CI, 0.021–0.069), with smaller increases for 30-day mortality and ICU admission.
Conclusion
In admitted ED patients who had lactate measured after triage at physician discretion, higher lactate levels were associated with increased in-hospital mortality across ESI levels, particularly among patients triaged as ESI levels 2 and 3.
INTRODUCTION
Triage is one of the main tasks of an emergency department (ED), in which patients are categorized based on acuity and prognosis.1 The Emergency Severity Index (ESI) is a well-established, reliable, and valid five-level triage tool. Patients are classified according to presenting symptoms, resource needs, and vital signs.2 Despite being primarily an acuity-based triage tool, the ESI is known to be associated with hospital admission and mortality.3–5 Its discriminatory performance for adverse outcomes can be enhanced by incorporating additional parameters. Adding physician disease severity ratings to any ESI level was associated with higher mortality in ED patients.6 Similarly, markers of vulnerability, such as impaired mobility or the Clinical Frailty Scale, were more strongly associated with 30-day mortality than the ESI among patients aged 65 years and older, highlighting their potential utility in triage decision-making.7,8 Another study found that combining ESI with the quick Sequential Organ Failure Assessment (qSOFA) score improved discriminatory accuracy for in-hospital mortality and intensive care unit (ICU) admission compared to ESI alone.9 Likewise, adding the Peripheral Perfusion Index or the Shock Index to the ESI improved discrimination of high-acuity unselected ED patients.10
Lactate has been shown to be associated with adverse outcomes in both unselected ED patients and those presenting with specific acute conditions, including infection, sepsis, trauma, pneumonia, pulmonary embolism, and stroke.11–18 To date, no study has evaluated whether lactate levels are associated with outcomes across ESI levels in admitted ED patients. Given that incorporating lactate into triage would represent a substantial operational shift, we aimed to assess whether lactate levels obtained after triage at the discretion of the treating physician are associated with in-hospital mortality across ESI levels and may provide additional information beyond ESI.
METHODS
Study Design and Setting
This study is a secondary analysis based on prospectively collected data from the EMERGE cohort, which serves quality control purposes and benchmarking between Swiss EDs. The study was conducted in the ED of a tertiary care center in Switzerland, with an annual ED census of approximately 55,000 patients aged 16 years and older. Obstetric, pediatric, and ophthalmologic patients are treated in separate facilities on campus. All patients presenting to the ED were assessed for inclusion 24 hours a day, 7 days a week. Data were collected from 21 October to 11 November 2013 (EMERGE cohort 1), 1 February to 23 February 2015 (EMERGE cohort 2), 30 January to 19 February 2017 (EMERGE cohort 3), and 18 March to 20 May 2019 (EMERGE cohort 4). The study was approved by the local ethics committee (EKNZ-236/13). The results are reported in accordance with the STROBE guidelines.19
Population Health Research Capsule.
What do we already know about this issue?
Both the Emergency Severity Index (ESI) and lactate are associated with mortality in emergency department (ED) patients, but the association of lactate across ESI levels is unclear.
What was the research question?
Is post-triage lactate associated with mortality across ESI levels in admitted ED patients?
What was the major finding of the study?
In ESI 3 patients, mortality rose from 1.6% with lactate < 2 to 15.1% with ≥ 4 mmol/L (RD 12.5%; 95% CI, 3.7–21.3).
How does this improve population health?
Lactate could help identify high-risk ESI 3 patients who may benefit from more timely evaluation or closer monitoring.
Selection Of Participants
All consecutive patients presenting to the ED during the study periods were screened for eligibility. Patients were excluded if they left without being seen, were directly referred to other departments, declined informed consent, or had repeated ED visits during the study periods (only the index visit was included). For the primary analysis, we included adult patients who were admitted to the hospital and had a lactate measurement obtained during the ED stay at the discretion of the treating physician. Patients discharged from the ED or with missing ESI or lactate values were excluded. The rationale for excluding patients discharged from the ED is their generally very low risk of mortality.3,20,21 In addition, lactate measurements are typically obtained in clinically sicker patients, who are more likely to require admission. Therefore, the analysis was restricted to admitted patients to focus on a higher risk population. For descriptive comparisons of ESI distributions and crude in-hospital mortality according to lactate measurement status, admitted patients without lactate measurements were included.
Data Collection
Upon arrival, all patients were registered by the study team, and an electronic health record (EHR) was created. The study team consisted of trained medical students. Data on demographics (age, sex), ESI triage level, patient disposition (discharge, hospital admission, or ICU admission), in-hospital and 30-day mortality were prospectively and consecutively collected. Clinicians ordered nonprotocolized lactate measurements at their discretion during routine clinical care. Consequently, this secondary analysis used data extracted from the EHR for a select subset of patients who underwent testing.
Triage Process
All patients were triaged by a triage clinician using the German version of the ESI, version 4.3 The ESI stratifies patients into five acuity categories based on their presenting symptoms, anticipated resource requirements, and vital signs. High-acuity patients are classified as ESI levels 1–2, intermediate-acuity patients as ESI level 3, and low-acuity patients as ESI levels 4–5. Triage personnel assign an ESI level 1 to patients requiring immediate life-saving interventions (eg, airway or respiratory support, emergency medications, or hemodynamic interventions, such as fluid resuscitation or blood products). Patients in high-risk situations (such as those at high risk for deterioration, exhibiting acutely altered mental status, or experiencing severe pain or distress) are assigned an ESI level 2. The remaining patients are classified into ESI levels 3–5 based on the expected number of required resources (more than one in ESI level 3, one in ESI level 4, and none in ESI level 5). Examples for resources are laboratory tests, electrocardiograms, or radiographs. For patients triaged as ESI level 3, vital signs must also be measured. In the presence of danger-zone vital signs (heart rate greater than 100 beats/min, respiratory rate greater than 20 breaths/min, or oxygen saturation less than 92% in adult patients), reassignment to ESI level 2 should be considered.22 Notably, the fifth version of the ESI manual (released in 2023) integrated a complete set of vital signs for all low-acuity patients; however, its benefit remains uncertain.2,23
Lactate Measurement
Venous lactate was measured in the treatment area after triage at the discretion of the treating physician and was not part of a standardized protocol. In routine clinical practice, lactate measurement was typically performed shortly after triage, with earlier measurement in higher acuity patients (ESI levels 1–2) and later measurement in lower acuity patients depending on ED workflow. Accordingly, lactate measurements were obtained in a select subset of patients. Clinicians performed bedside lactate measurements using a point-of-care analyzer in the ED (Radiometer ABL90 FLEX PLUS; Radiometer Medical ApS, Brønshøj, Denmark). This analysis evaluated only the first lactate level obtained in the ED. To improve clinical interpretability, stratified lactate levels were used instead of continuous lactate levels. Based on thresholds used in previous studies, lactate levels were categorized as low (< 2.0 mmol/L), intermediate (2.0–3.9 mmol/L), and high (≥ 4.0 mmol/L).24–27
Outcomes
The primary outcome was in-hospital mortality. Admission was defined as transfer from the ED to any hospital ward with a minimum stay of one night. Patients who died in the ED prior to an overnight stay were considered admitted and counted as in-hospital deaths. Secondary outcomes included 30-day mortality and ICU admission. Thirty-day mortality was defined as death occurring within 30 days after presentation to the ED. Follow-up was performed by reviewing the hospital information system; contacting patients, proxies, or primary care physicians; and consulting health insurance companies or civil registries. Patients lost to follow-up were considered alive. Admission to the ICU was defined as transfer to the ICU, intermediate care unit, or stroke unit during the hospital stay.
Statistical Analysis
For categorical variables, data are presented as frequencies and percentages; continuous variables are reported as medians with interquartile ranges (IQRs). Given that lactate testing was performed selectively at physician discretion, patients without a lactate value were considered not to have undergone measurement rather than to have conventional missing data. Consequently, the primary analyses were restricted to admitted patients with available ESI and lactate data, and missing lactate values were not imputed. Descriptive analyses comparing patients with and without lactate measurements were performed without imputation. To assess whether lactate provided additional information beyond ESI within this selected cohort, logistic regression models were fitted for ESI alone and for the combined ESI-by-lactate specification; adjusted models included age and sex. Nested model comparisons were performed using likelihood-ratio tests (two-sided P < .05). Discrimination was assessed using the area under the receiver operating characteristic curve (AUC) with 95% confidence intervals (CIs). Changes in discrimination (ΔAUC) were reported with 95% CIs; statistical inference for nested models relied on likelihood-ratio tests, with ΔAUC estimates presented descriptively. Patients lost to follow-up at 30 days were coded as survivors in the primary analysis, with sensitivity analyses excluding them or recoding them as deaths (Supplementary Table S1). Due to small sample sizes, ESI levels 4 and 5 were combined. Adjusted analyses were considered primary and are presented in the main manuscript, whereas unadjusted analyses are provided in the Supplementary Materials (Supplementary Tables S2, S3, and S4 and Figure S1). Because this study was a secondary analysis of a prospectively collected cohort, no formal sample size calculation was performed a priori. The sample size was determined by the number of eligible patients available during the study periods. We did not perform a post hoc power calculation, because effect estimates and 95% CIs provide a more informative indication of estimate precision. All analyses were conducted in R, version 4.5.1 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
Baseline Characteristics
During the study period, 17,327 patients presented to the ED. Of 14,440 enrolled patients, 10,531 were excluded, mainly due to discharge from the ED (9,439 patients). The final cohort included 3,909 admitted patients (Figure 1).
Figure 1.
Flowchart of admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion in a study that evaluated whether lactate levels were associated with in-hospital mortality across Emergency Severity Index levels.
ED, emergency department; ESI, Emergency Severity Index.
In these, median age was 72 years (IQR 55–83), and 52.8% were male. In-hospital mortality occurred in 168 patients (4.3%), 30-day mortality in 258 (6.6%), and 698 (17.9%) were admitted to the ICU; 76 (1.9%) were lost to follow-up and considered alive. Most patients were triaged as ESI level 2 (1,826/3,909; 46.7%) or level 3 (1,722/3,909; 44.1%). In-hospital mortality was highest in ESI level 1 (64/286; 22.4%), followed by level 2 (59/1,826; 3.2%) and level 3 (44/1,722; 2.6%), with similar patterns for 30-day mortality and ICU admission. In ESI level 4 and 5 patients (n = 75), adverse outcomes were rare, with one in-hospital death (1.3%) and four ICU admissions (5.3%). Most patients had a lactate < 2.0 mmol/L (2,641/3,909; 67.6%), followed by 2.0–3.9 mmol/L (1,041/3,909; 26.6%) and ≥ 4.0 mmol/L (227/3,909; 5.8%). In-hospital mortality increased with higher lactate levels: 2.3% (< 2.0 mmol/L), 4.8% (2.0–3.9 mmol/L), and 25.1% (≥ 4.0 mmol/L), again with similar patterns for 30-day mortality and ICU admission. Detailed distributions are presented in Table 1.
Table 1.
Baseline characteristics and outcomes of admitted adult emergency department patients stratified by Emergency Severity Index and lactate category.
| N (%) | Age median (IQR) | Male % | In-hospital mortality n (%) | 30-day mortality n (%) * | ICU admission n (%) | |
|---|---|---|---|---|---|---|
| All patients | 3,909 | 72 (55–83) | 52.8 | 168/3,909 (4.3%) | 258/3,909 (6.6%) | 698/3,909 (17.9%) |
| ESI level 1 | 286/3,909 (7.3%) | 71 (57–80) | 59.8 | 64/286 (22.4%) | 75/286 (26.2%) | 171/286 (59.8%) |
| ESI level 2 | 1,826/3,909 (46.7%) | 70 (54–81) | 55.3 | 59/1,826 (3.2%) | 100/1,826 (5.5%) | 370/1,826 (20.3%) |
| ESI level 3 | 1,722/3,909 (44.1%) | 75 (58–84) | 49.7 | 44/1,722 (2.6%) | 81/1,722 (4.7%) | 153/1,722 (8.9%) |
| ESI level 4–5 | 75/3,909 (1.9%) | 65 (41–82) | 34.7 | 1/75 (1.3%) | 2/75 (2.7%) | 4/75 (5.3%) |
| Lac < 2.0 mmol/L | 2,641/3,909 (67.6%) | 72 (54–83) | 50.4 | 61/2,641 (2.3%) | 120/2,641 (4.5%) | 401/2,641 (15.2%) |
| Lac 2.0–3.9 mmol/L | 1,041/3,909 (26.6%) | 73 (58–83) | 57.1 | 50/1,041 (4.8%) | 74/1,041 (7.1%) | 213/1,041 (20.5%) |
| Lac ≥ 4 mmol/L | 227/3,909 (5.8%) | 71 (58–80) | 59.9 | 57/227 (25.1%) | 64/227 (28.2%) | 84/227 (37.0%) |
Data are presented for admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion (N = 3,909).
Patients lost to follow-up (n = 76) were considered alive.
ESI, Emergency Severity Index; ICU, intensive care unit; IQR, interquartile range; Lac, lactate.
Primary and Secondary Outcomes
Although crude in-hospital mortality was similar between patients triaged as ESI level 2 and 3 (3.2% and 2.6%, respectively), differences became more pronounced after stratification by lactate level. In ESI level 2 patients, in-hospital mortality increased from 2.0% among those with lactate < 2.0 mmol/L to 17.8% among those with lactate ≥ 4.0 mmol/L, corresponding to an age- and sex-adjusted absolute risk difference of 16.0% (95% CI, 8.8–23.3). In ESI level 3 patients, in-hospital mortality increased from 1.6% at lactate < 2.0 mmol/L to 15.1% at lactate ≥ 4.0 mmol/L, corresponding to a risk difference of 12.5% (95% CI, 3.7–21.3) (Table 2, Figure 2).
Table 2.
In-hospital mortality of admitted adult emergency department patients stratified by Emergency Severity Index and lactate category.
| N (%) | In-hospital mortality | ||
|---|---|---|---|
| n (%) | aOR (95% CI) | ||
| ESI level 1 | 286/3,909 (7.3%) | 64/286 (22.4%) | |
| Lac < 2.0 mmol/L | 121/286 (42.3%) | 16/121 (13.2%) | Ref |
| Lac 2.0–3.9 mmol/L | 93/286 (32.5%) | 17/93 (18.3%) | 1.30 (0.59–2.87) |
| Lac ≥ 4 mmol/L | 72/286 (25.2%) | 31/72 (43.1%) | 5.27 (2.53–10.97) |
| ESI level 2 | 1,826/3,909 (46.7%) | 59/1,826 (3.2%) | |
| Lac < 2.0 mmol/L | 1,210/1,826 (66.3%) | 24/1,210 (2.0%) | Ref |
| Lac 2.0–3.9 mmol/L | 515/1,826 (28.2%) | 17/515 (3.3%) | 1.56 (0.82–2.95) |
| Lac ≥ 4 mmol/L | 101/1,826 (5.5%) | 18/101 (17.8%) | 11.75 (6.01–22.97) |
| ESI level 3 | 1,722/3,909 (44.1%) | 44/1,722 (2.6%) | |
| Lac < 2.0 mmol/L | 1,254/1,722 (72.8%) | 20/1,254 (1.6%) | Ref |
| Lac 2.0–3.9 mmol/L | 415/1,722 (24.1%) | 16/415 (3.9%) | 2.22 (1.14–4.32) |
| Lac ≥ 4 mmol/L | 53/1,722 (3.1%) | 8/53 (15.1%) | 10.57 (4.28–26.10) |
| ESI level 4–5 | 75/3,909 (1.9%) | 1/75 (1.3%) | |
| Lac < 2.0 mmol/L | 56/75 (74.7%) | 1/56 (1.8%) | Ref |
| Lac 2.0–3.9 mmol/L | 18/75 (24.0%) | 0/18 (0.0%) | NA |
| Lac ≥ 4 mmol/L | 1/75 (1.3%) | 0/1 (0.0%) | NA |
Data are presented for admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion (N = 3,909). Adjusted odds ratios were estimated from age- and sex-adjusted logistic regression models. Within each ESI level, patients with lactate < 2.0 mmol/L served as the reference group. Adjusted odds ratios were not estimable for strata with zero events.
aOR, adjusted odds ratio; ESI, Emergency Severity Index; Lac, lactate; Ref, reference.
Figure 2.
In-hospital mortality across Emergency Severity Index and lactate categories in admitted adult ED patients with lactate measurements obtained after triage at physician discretion, adjusted for age and sex. Values represent absolute risk estimates (95% confidence intervals) derived from logistic regression models. Confidence intervals extending below zero reflect limited precision due to small cell sizes in these strata.
ED, emergency department; ESI, Emergency Severity Index.
Among patients without lactate measurements, crude in-hospital mortality was also similar between ESI levels 2 and 3 (1.2% and 1.4%, respectively) (Table 3).
Table 3.
Emergency Severity Index levels and in-hospital mortality stratified by lactate measurement status in admitted emergency department patients.
| All admitted patients (N = 4,745) | With lactate measurements (n = 3,909) | Without lactate measurements (n = 836) | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| n (%) | IHM, n (%) | n (%) | IHM, n (%) | n (%) | IHM, n (%) | |
| ESI level 1 | 306/4,745 (6.4%) | 67/306 (21.9%) | 286/3,909 (7.3%) | 64/286 (22.4%) | 20/836 (2.4%) | 3/20 (15%) |
| ESI level 2 | 1,994/4,745 (42.0%) | 61/1,994 (3.1%) | 1,826/3,909 (46.7%) | 59/1,826 (3.2%) | 168/836 (20.1%) | 2/168 (1.2%) |
| ESI level 3 | 2,235/4,745 (47.1%) | 51/2,235 (2.3%) | 1,722/3,909 (44.1%) | 44/1,722 (2.6%) | 513/836 (61.4%) | 7/513 (1.4%) |
| ESI level 4–5 | 210/4,745 (4.4%) | 1/210 (0.5%) | 75/3,909 (1.9%) | 1/75 (1.3%) | 135/836 (16.1%) | 0/135 (0%) |
Lactate measurements were obtained after triage at physician discretion. All patients refer to admitted index-visit patients with non-missing ESI; nine admitted patients with missing ESI were excluded from this ESI-stratified table.
ESI, Emergency Severity Index; IHM, in-hospital mortality.
Across ESI levels 1–3 patients, lactate ≥ 4.0 mmol/L was associated with increased in-hospital and 30-day mortality, while associations with ICU admission varied by ESI level. In ESI level 1 and 2 patients, lactate values between 2.0–3.9 mmol/L were not associated with in-hospital mortality, 30-day mortality, or ICU admission. In contrast, in ESI level 3 patients, lactate values between 2.0–3.9 mmol/L were associated with in-hospital and 30-day mortality, but not with ICU admission (Tables 2 and 4).
Table 4.
Thirty-day mortality and intensive care unit admission of admitted adult emergency department patients stratified by Emergency Severity Index and lactate category.
| N (%) | 30-day mortality | ICU admission | |||
|---|---|---|---|---|---|
| n (%) | aOR (95% CI) | n (%) | aOR (95% CI) | ||
| ESI level 1 | 286/3,909 (7.3%) | 75/286 (26.2%) | 171/286 (59.8%) | ||
| Lac < 2.0 mmol/L | 121/286 (42.3%) | 21/121 (17.4%) | Ref | 65/121 (53.7%) | Ref |
| Lac 2.0–3.9 mmol/L | 93/286 (32.5%) | 21/93 (22.6%) | 1.26 (0.60–2.65) | 53/93 (57.0%) | 1.06 (0.61–1.84) |
| Lac ≥ 4 mmol/L | 72/286 (25.2%) | 33/72 (45.8%) | 4.47 (2.15–9.30) | 53/72 (73.6%) | 2.35 (1.24–4.46) |
| ESI level 2 | 1,826/3,909 (46.7%) | 100/1,826 (5.5%) | 370/1,826 (20.3%) | ||
| Lac < 2.0 mmol/L | 1,210/1,826 (66.3%) | 55/1,210 (4.5%) | Ref | 228/1,210 (18.8%) | Ref |
| Lac 2.0–3.9 mmol/L | 515/1,826 (28.2%) | 26/515 (5.0%) | 1.02 (0.63–1.66) | 119/515 (23.1%) | 1.26 (0.98–1.62) |
| Lac ≥ 4 mmol/L | 101/1,826 (5.5%) | 19/101 (18.8%) | 5.52 (3.02–10.12) | 23/101 (22.8%) | 1.23 (0.75–1.99) |
| ESI level 3 | 1,722/3,909 (44.1%) | 81/1,722 (4.7%) | 153/1,722 (8.9%) | ||
| Lac < 2.0 mmol/L | 1,254/1,722 (72.8%) | 42/1,254 (3.3%) | Ref | 104/1,254 (8.3%) | Ref |
| Lac 2.0–3.9 mmol/L | 415/1,722 (24.1%) | 27/415 (6.5%) | 1.80 (1.09–2.97) | 41/415 (9.9%) | 1.19 (0.81–1.75) |
| Lac ≥ 4 mmol/L | 53/1,722 (3.1%) | 12/53 (22.6%) | 8.78 (4.13–18.69) | 8/53 (15.1%) | 1.89 (0.87–4.11) |
| ESI level 4–5 | 75/3,909 (1.9%) | 2/75 (2.7%) | 4/75 (5.3%) | ||
| Lac < 2.0 mmol/L | 56/75 (74.7%) | 2/56 (3.6%) | Ref | 4/56 (7.1%) | Ref |
| Lac 2.0–3.9 mmol/L | 18/75 (24.0%) | 0/18 (0.0%) | NA | 0/18 (0.0%) | NA |
| Lac ≥ 4 mmol/L | 1/75 (1.3%) | 0/1 (0.0%) | NA | 0/1 (0.0%) | NA |
Data are presented for admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion (N = 3,909). Adjusted odds ratios were estimated from age- and sex-adjusted logistic regression models. Within each ESI level, patients with lactate < 2.0 mmol/L served as the reference group. Adjusted odds ratios were not estimable for strata with zero events.
aOR, adjusted odds ratio; ESI, Emergency Severity Index; ICU, intensive care unit; Lac, lactate; Ref, reference.
Among the 1,722 patients triaged as ESI level 3, a lactate level ≥ 2 mmol/L captured 24 of 44 in-hospital deaths (54.5%) and 49 of 153 ICU admissions (32.0%), whereas a lactate level ≥ 4 mmol/L captured 8 of 44 in-hospital deaths (18.2%) and 8 of 153 ICU admissions (5.2%). The number of patients who would need to be flagged to capture one event was 19.5 vs 6.6 for in-hospital mortality and 9.6 vs 6.6 for ICU admission for the ≥ 2 mmol/L and ≥ 4 mmol/L thresholds (Supplementary Table S5). After adjustment for age and sex, models including both ESI and lactate showed higher discrimination for in-hospital mortality compared with ESI alone (AUC, 0.844 vs 0.799; ΔAUC, 0.045), with smaller differences for 30-day mortality (ΔAUC, 0.025) and ICU admission (ΔAUC, 0.006) (Table 5, Figure 3).
Table 5.
Discriminatory performance of Emergency Severity Index (ESI) and ESI–lactate models in admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion.
| Outcome | AUC (95% CI): ESI | AUC (95% CI): Lac | AUC (95% CI): ESI×Lac | ΔAUC (95% CI) | LR χ2 | LR P |
|---|---|---|---|---|---|---|
| In-hospital | 0.799 (0.76–0.83) | 0.794 (0.76–0.83) | 0.844 (0.81–0.87) | 0.045 (0.021–0.069) | 88.43 | < .001 |
| 30-day | 0.791 (0.76–0.82) | 0.776 (0.75–0.80) | 0.816 (0.79–0.84) | 0.025 (0.009–0.041) | 74.85 | < .001 |
| ICU | 0.709 (0.69–0.73) | 0.603 (0.58–0.63) | 0.715 (0.69–0.74) | 0.006 (0.001–0.012) | 16.94 | .03 |
AUCs are derived from age- and sex-adjusted logistic regression models. ΔAUC indicates the change in AUC after adding lactate to the ESI. Likelihood-ratio tests compare nested ESI-only and combined ESI×lactate logistic regression models.
AUC, area under the curve; χ2, chi-square statistic; ESI, Emergency Severity Index; ICU, intensive care unit; LR, likelihood ratio; Lac, lactate.
Figure 3.
Receiver operating characteristic curves comparing age- and sex-adjusted models based on Emergency Severity Index (ESI) alone, lactate alone, and combined ESI–lactate model for (a) in-hospital mortality, (b) 30-day mortality, and (c) intensive care unit admission in admitted adult emergency department patients with lactate measurements obtained after triage at physician discretion.
ESI, Emergency Severity Index; ICU, intensive care unit.
For in-hospital mortality, stratified analyses showed consistent benefits across age groups, with AUC increasing from 0.834 to 0.900 in patients < 65 years and from 0.746 to 0.796 in those ≥ 65 years (Supplementary Table S6).
DISCUSSION
This study showed that among admitted ED patients who had lactate measured after triage at the treating physician’s discretion, higher lactate levels were associated with increased in-hospital mortality and 30-day mortality across ESI levels, whereas associations with ICU admission were weaker and varied by ESI level. Differences in outcomes across lactate categories were particularly evident among patients triaged as ESI levels 2 and 3, in whom overall mortality rates were similar.
The ability of the ESI to discriminate adverse outcomes is limited by its design as an acuity-based tool that relies in part on the subjective clinical judgment of triaging clinicians. In the present study, most patients were triaged as ESI level 2 or 3, which is consistent with previous findings.4,28 As expected, the highest in-hospital mortality was observed among patients triaged as ESI level 1. Among admitted patients with lactate measurements obtained after triage at the treating physician’s discretion, crude mortality rates for patients in ESI levels 2 and 3 were similar, at 3.2% and 2.6%, respectively. This may reflect differences in case mix and the selective use of lactate measurement after triage, as lactate was preferentially obtained in older and more severely ill patients, whereas those without lactate measurements were younger and had lower mortality (Supplementary Table S7). Notably, crude mortality rates were also similar between ESI levels 2 and 3 among admitted patients without lactate measurements (Table 3), supporting the interpretation that this pattern is not solely attributable to selection related to lactate measurement, but may also reflect limited discrimination of the ESI within these categories.
When stratified by lactate level, differences in mortality became more apparent. In ESI level 3 patients, a lactate threshold of ≥ 2 mmol/L identified a large proportion of patients who died, whereas a threshold of ≥ 4 mmol/L identified fewer patients with higher mortality. These findings suggest that lactate may help to further characterize differences in risk within this intermediate-acuity group.
These results are consistent with previous studies showing that higher lactate levels are associated with adverse outcomes and may provide additional information when considered alongside established clinical scores. For example, studies evaluating the National Early Warning Score (NEWS) have reported that higher lactate levels were associated with increased in-hospital mortality and ICU admission in general ED populations.29 Similarly, in patients with suspected sepsis, lactate ≥ 2 mmol/L has been associated with higher sensitivity for mortality or prolonged ICU stay when considered alongside qSOFA.30 Likewise, higher lactate levels have been associated with mortality across established severity scoring systems such as Acute Physiology and Chronic Health Evaluation II, SOFA, and Mortality in Emergency Department Sepsis in septic ED patients.31
In the present study, discrimination for 30-day mortality and ICU admission was lower than for in-hospital mortality, likely reflecting the impact of postadmission factors and institutional decision-making beyond the patient’s initial physiological condition. In older patients, previous studies have reported that lactate, when considered alongside different triage scales or clinical scores such as the Modified Early Warning Score (MEWS), NEWS, and qSOFA, was not associated with differences in discrimination of poor outcomes.32 In our study, however, increases in discrimination for in-hospital mortality were observed in older patients when lactate was considered alongside ESI, which may reflect differences in study populations and the selective use of lactate measurements.
The optimal lactate threshold associated with adverse outcomes in unselected adult ED patients remains uncertain, with reported cutoffs ranging from 2.0 to 4.0 mmol/L.12,26,33 In this study, even moderately elevated lactate levels (2.0–3.9 mmol/L) were associated with adverse outcomes, consistent with prior research.34 However, the association of intermediate lactate elevations with adverse outcomes differed across triage categories. While lactate levels of 2.0–3.9 mmol/L were not associated with adverse outcomes in ESI levels 1 and 2, they were associated with both in-hospital and 30-day mortality in ESI level 3 patients. This suggests that moderately elevated lactate levels may help to further characterize differences in risk within this heterogeneous triage level.
If lactate were measured systematically at the time of triage, it might enhance the discrimination of the ESI for patients ultimately admitted. In this context, point-of-care lactate testing may be of interest for identifying more subtle high-risk presentations, particularly among patients triaged as ESI level 3 who may otherwise appear clinically stable but could warrant more timely evaluation or closer monitoring. However, this potential role remains hypothetical and requires prospective evaluation.
LIMITATIONS
Several limitations should be acknowledged. First, our study was conducted at a single tertiary care center in Europe, which limits its external validity. Second, the exclusion of discharged patients may have introduced selection bias toward a higher risk population, thereby limiting generalizability. In our cohort, 30-day mortality was 0.3% among discharged patients compared to 5.9% among admitted patients, consistent with previous reports.35 Third, lactate levels were measured at the treating physician’s discretion, potentially introducing selection bias. Patients without lactate measurements (and the small number with missing ESI) had lower rates of in-hospital mortality (1.4% vs 4.3%), 30-day mortality (2.8% vs 6.6%), and ICU admission (9.6% vs 17.9%) compared to included patients (Supplementary Tables S7 and S8), suggesting that less severely ill patients were underrepresented in the final cohort. Fourth, only the first lactate level was analyzed, without considering subsequent changes or lactate clearance, which may be more strongly associated with adverse outcomes than a single value.36–38 However, the aim was to evaluate the association of lactate during initial evaluation.
Fifth, the influence of potential confounding factors on lactate levels, such as alcohol consumption, liver disease, seizure, and medication, was not investigated and may have affected lactate measurements.39 Such information may be unavailable or unreliable during triage. Sixth, although lactic acidosis was previously reported to be more strongly associated with in-hospital mortality than hyperlactatemia, we did not differentiate between them to maintain a simple and clinically applicable approach.40,41 Seventh, details regarding prehospital care were not provided. Nevertheless, patients referred by general practitioners usually receive no prehospital treatment, and those brought in by emergency medical services have short transport times of less than 10 minutes due to the hospital’s central urban location. Eighth, the presence of hypotension was not analyzed separately. However, blood pressure is not included in the ESI algorithm version 4 to determine triage levels. In addition, previous studies have shown that the association between lactate levels and adverse outcomes is independent of blood pressure and the presence of shock.24,42
Finally, outcomes such as ICU admission may be influenced by factors including advance directives, institutional protocols, and resource availability. In addition, elevated lactate levels may directly influence clinical decision-making and disposition, since high lactate values are often perceived as markers of critical illness and may prompt ICU admission independent of the patient’s underlying physiological status. This may introduce confounding by indication and should be considered when interpreting associations between lactate levels and ICU admission.
CONCLUSION
Among admitted ED patients with lactate measurements obtained after triage at the discretion of the treating physician, higher lactate levels were associated with in-hospital mortality across ESI levels, particularly in patients triaged as ESI levels 2 and 3. Given the single-center design, these findings should be validated in prospective studies across diverse ED settings.
Supplementary Information
Footnotes
Section Editor: Christopher R. Tainter, MD
Full text available through open access at http://escholarship.org/uc/uciem_westjem
Conflicts of Interest: By the WestJEM article submission agreement, all authors are required to disclose all affiliations, funding sources and financial or management relationships that could be perceived as potential sources of bias. No author has professional or financial relationships with any companies that are relevant to this study. There are no conflicts of interest or sources of funding to declare.
REFERENCES
- 1.Bingisser R, Nickel CH. The last decade of symptom-oriented research in emergency medicine: triage, work-up, and disposition. Swiss Med Wkly. 2019;149:w20141. doi: 10.4414/smw.2019.20141. [DOI] [PubMed] [Google Scholar]
- 2.Agency for Healthcare Research and Quality. Emergency Severity Index Handbook. 5th ed. 2023. [Accessed January 18, 2026]. Available at: https://californiaena.org/wp-content/uploads/2023/05/ESI-Handbook-5th-Edition-3-2023.pdf.
- 3.Grossmann FF, Nickel CH, Christ M, et al. Transporting clinical tools to new settings: cultural adaptation and validation of the Emergency Severity Index in German. Ann Emerg Med. 2011;57(3):257–264. doi: 10.1016/j.annemergmed.2010.07.021. [DOI] [PubMed] [Google Scholar]
- 4.van der Wulp I, Schrijvers AJP, van Stel HF. Predicting admission and mortality with the Emergency Severity Index and the Manchester Triage System: a retrospective observational study. Emerg Med J. 2009;26(7):506–509. doi: 10.1136/emj.2008.063768. [DOI] [PubMed] [Google Scholar]
- 5.Wuerz R. Emergency severity index triage category is associated with six-month survival. Acad Emerg Med. 2001;8(1):61–64. doi: 10.1111/j.1553-2712.2001.tb00554.x. [DOI] [PubMed] [Google Scholar]
- 6.Bingisser R, Baerlocher SM, Kuster T, et al. Physicians’ disease severity ratings are non-inferior to the Emergency Severity Index. J Clin Med. 2020;9(3):762. doi: 10.3390/jcm9030762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kaeppeli T, Rueegg M, Dreher-Hummel T, et al. Validation of the Clinical Frailty Scale for prediction of thirty-day mortality in the emergency department. Ann Emerg Med. 2020;76(3):291–300. doi: 10.1016/j.annemergmed.2020.03.028. [DOI] [PubMed] [Google Scholar]
- 8.Nieves-Ortega R, Brabrand M, Dutilh G, et al. Assessment of patient mobility improves the risk stratification of triage with the Emergency Severity Index: a prospective cohort study. Eur J Emerg Med. 2021;28(6):456–462. doi: 10.1097/MEJ.0000000000000845. [DOI] [PubMed] [Google Scholar]
- 9.Kwak H, Suh GJ, Kim T, et al. Prognostic performance of Emergency Severity Index (ESI) combined with qSOFA score. Am J Emerg Med. 2018;36(10):1784–1788. doi: 10.1016/j.ajem.2018.01.088. [DOI] [PubMed] [Google Scholar]
- 10.Daş M, Bardakci O, Siddikoglu D, et al. Prognostic performance of peripheral perfusion index and shock index combined with ESI to predict hospital outcome. Am J Emerg Med. 2020;38(10):2055–2059. doi: 10.1016/j.ajem.2020.06.084. [DOI] [PubMed] [Google Scholar]
- 11.Jo S, Jeong T, Lee JB, et al. Initial hyperlactatemia in the ED is associated with poor outcome in patients with ischemic stroke. Am J Emerg Med. 2012;30(3):449–455. doi: 10.1016/j.ajem.2011.12.019. [DOI] [PubMed] [Google Scholar]
- 12.Park YJ, Kim DH, Kim SC, et al. Serum lactate upon emergency department arrival as a predictor of 30-day in-hospital mortality in an unselected population. PLoS One. 2018;13(1):e0190519. doi: 10.1371/journal.pone.0190519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Seker YC, Bozan O, Sam E, et al. The role of the serum lactate level at the first admission to the emergency department in predicting mortality. Am J Emerg Med. 2021;45:495–500. doi: 10.1016/j.ajem.2020.09.088. [DOI] [PubMed] [Google Scholar]
- 14.Shapiro NI, Howell MD, Talmor D, et al. Serum lactate as a predictor of mortality in emergency department patients with infection. Ann Emerg Med. 2005;45(5):524–528. doi: 10.1016/j.annemergmed.2004.12.006. [DOI] [PubMed] [Google Scholar]
- 15.Cheng H-H, Chen F-C, Change M-W, et al. Difference between elderly and non-elderly patients in using serum lactate level to predict mortality caused by sepsis in the emergency department. Medicine (Baltimore) 2018;97(13):e0209. doi: 10.1097/MD.0000000000010209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hung KKC. BET 2. Serum lactate as a marker for mortality in patients presenting to the emergency department with trauma. Emerg Med J. 2009;26(2):118–119. doi: 10.1136/emj.2008.070797. [DOI] [PubMed] [Google Scholar]
- 17.Chen Y-X, Li C-S. Lactate on emergency department arrival as a predictor of mortality and site-of-care in pneumonia patients: a cohort study. Thorax. 2015;70(5):404–410. doi: 10.1136/thoraxjnl-2014-206461. [DOI] [PubMed] [Google Scholar]
- 18.Vanni S, Viviani G, Baioni M, et al. Prognostic value of plasma lactate levels among patients with acute pulmonary embolism: the thrombo-embolism lactate outcome study. Ann Emerg Med. 2013;61(3):330–338. doi: 10.1016/j.annemergmed.2012.10.022. [DOI] [PubMed] [Google Scholar]
- 19.Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. Epidemiology. 2007;18(6):805–835. doi: 10.1097/EDE.0b013e3181577511. [DOI] [PubMed] [Google Scholar]
- 20.Obermeyer Z, Cohn B, Wilson M, et al. Early death after discharge from emergency departments: analysis of national US insurance claims data. BMJ. 2017;356:j239. doi: 10.1136/bmj.j239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gabayan GZ, Derose SF, Asch SM, et al. Patterns and predictors of short-term death after emergency department discharge. Ann Emerg Med. 2011;58(6):551–8e2. doi: 10.1016/j.annemergmed.2011.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gilboy N, Tanabe P, Travers DA, et al. Emergency Severity Index (ESI): Implementation Handbook. 2012 ed. Agency for Healthcare Research and Quality; 2011. [Google Scholar]
- 23.Espejo T, Grossmann FF, Riedel HB, et al. The Emergency Severity Index (ESI) Version 5: simulation of predictive validity and triage level distribution. J Emerg Med. 2025;78:57–70. doi: 10.1016/j.jemermed.2025.07.035. [DOI] [PubMed] [Google Scholar]
- 24.Mikkelsen ME, Miltiades AN, Gaieski DF, et al. Serum lactate is associated with mortality in severe sepsis independent of organ failure and shock. Crit Care Med. 2009;37(5):1670–1677. doi: 10.1097/CCM.0b013e31819fcf68. [DOI] [PubMed] [Google Scholar]
- 25.Pedersen M, Brandt VS, Holler JG, et al. Lactate level, aetiology and mortality of adult patients in an emergency department: a cohort study. Emerg Med J. 2015;32(9):678–684. doi: 10.1136/emermed-2014-204305. [DOI] [PubMed] [Google Scholar]
- 26.Datta D, Walker C, Gray AJ, et al. Arterial lactate levels in an emergency department are associated with mortality: a prospective observational cohort study. Emerg Med J. 2015;32(9):673–677. doi: 10.1136/emermed-2013-203541. [DOI] [PubMed] [Google Scholar]
- 27.Contenti J, Occelli C, Lemoel F, et al. Blood lactate measurement within the emergency department: a two-year retrospective analysis. Am J Emerg Med. 2019;37(3):401–406. doi: 10.1016/j.ajem.2018.05.065. [DOI] [PubMed] [Google Scholar]
- 28.Patel MD, Lin P, Cheng Q, et al. Patient sex, racial and ethnic disparities in emergency department triage: a multi-site retrospective study. Am J Emerg Med. 2024;76:29–35. doi: 10.1016/j.ajem.2023.11.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Jo S, Jeong T, Park B. Early clinical outcome prediction based on the initial National Early Warning Score + Lactate (NEWS+L) score among adult emergency department patients. Emerg Med J. 2023;40(6):444–450. doi: 10.1136/emermed-2022-212654. [DOI] [PubMed] [Google Scholar]
- 30.Shetty A, MacDonald SP, Williams JM, et al. Lactate ≥2 mmol/L plus qSOFA improves utility over qSOFA alone in emergency department patients presenting with suspected sepsis. Emerg Med Australas. 2017;29(6):626–34. doi: 10.1111/1742-6723.12894. [DOI] [PubMed] [Google Scholar]
- 31.Chen Y-X, Li C-S. Arterial lactate improves the prognostic performance of severity score systems in septic patients in the ED. Am J Emerg Med. 2014;32(9):982–986. doi: 10.1016/j.ajem.2014.05.025. [DOI] [PubMed] [Google Scholar]
- 32.Gosselin M, Mabire C, Pasquier M, et al. Prevalence and clinical significance of point-of-care elevated lactate at emergency admission in older patients: a prospective study. Intern Emerg Med. 2022;17(6):1803–1812. doi: 10.1007/s11739-022-03005-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Barfod C, Lundstrøm LH, Lauritzen MMP, et al. Peripheral venous lactate at admission is associated with in-hospital mortality, a prospective cohort study. Acta Anaesthesiol Scand. 2015;59(4):514–523. doi: 10.1111/aas.12503. [DOI] [PubMed] [Google Scholar]
- 34.Puskarich MA, Illich BM, Jones AE. Prognosis of emergency department patients with suspected infection and intermediate lactate levels: a systematic review. J Crit Care. 2014;29(3):334–339. doi: 10.1016/j.jcrc.2013.12.017. [DOI] [PubMed] [Google Scholar]
- 35.Baker M, Clancy M. Can mortality rates for patients who die within the emergency department, within 30 days of discharge from the emergency department, or within 30 days of admission from the emergency department be easily measured? Emerg Med J. 2006;23(8):601–603. doi: 10.1136/emj.2005.028134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Wang J, Ji M. The 6-hour lactate clearance rate in predicting 30-day mortality in cardiogenic shock. J Intensive Med. 2024;4(3):393–399. doi: 10.1016/j.jointm.2024.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Régnier M-A, Raux M, Le Manach Y, et al. Prognostic significance of blood lactate and lactate clearance in trauma patients. Anesthesiology. 2012;117(6):1276–1288. doi: 10.1097/ALN.0b013e318273349d. [DOI] [PubMed] [Google Scholar]
- 38.Shadvar K, Nader-Djalal N, Vahed N, et al. Comparison of lactate/albumin ratio to lactate and lactate clearance for predicting outcomes in patients with septic shock admitted to intensive care unit: an observational study. Sci Rep. 2022;12(1):13047. doi: 10.1038/s41598-022-14764-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Wardi G, Brice J, Correia M, et al. Demystifying lactate in the emergency department. Ann Emerg Med. 2020;75(2):287–98. doi: 10.1016/j.annemergmed.2019.06.027. [DOI] [PubMed] [Google Scholar]
- 40.Lee S-W, Hong Y-S, Park D-W, et al. Lactic acidosis not hyperlactatemia as a predictor of in-hospital mortality in septic emergency patients. Emerg Med J. 2008;25(10):659–665. doi: 10.1136/emj.2007.055558. [DOI] [PubMed] [Google Scholar]
- 41.D’Abrantes R, Dunn L, McMillan T, et al. Evaluation of the prognostic value of lactate and acid-base status in patients presenting to the emergency department. Cureus. 2021;13(6):e15857. doi: 10.7759/cureus.15857. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Howell MD, Donnino M, Clardy P, et al. Occult hypoperfusion and mortality in patients with suspected infection. Intensive Care Med. 2007;33(11):1892–1899. doi: 10.1007/s00134-007-0680-5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.