Abstract
Background
For decades, trauma has been recognized globally as a major cause of death. Reducing the mortality of patients with trauma is an extremely pressing issue, particularly for those with severe trauma. An early and accurate assessment of the risk of mortality among patients with severe trauma is important for improving patient outcomes.
Methods
We performed a retrospective medical record review of 582 patients with severe trauma admitted to the emergency department between July 2011 and June 2016. We analyzed the associations of in-hospital mortality with the baseline characteristics and initial biochemical markers of patients with severe trauma on admission.
Results
The overall in-hospital mortality rate was 14.9%. Multivariate logistic regression analysis showed that the patient’s Rapid Emergency Medicine Score (REMS; odds ratio [OR], 1.186; 95% confidence interval [CI], 1.018–1.383; p=0.029), Emergency Trauma Score (EMTRAS; OR, 2.168; 95% CI, 1.570–2.994; p<0.001), serum lactate levels (SLL; OR, 1.298; 95% CI, 1.118–1.507; p<0.001), and Injury Severity Score (ISS; OR, 1.038; 95% CI, 1.010–1.130; p=0.021) were significantly associated with in-hospital mortality.
Conclusion
The REMS, EMTRAS, and SLL can easily and rapidly be used as alternatives to the injury severity score to predict in-hospital mortality for patients who present to the emergency department with severe trauma.
Keywords: Wounds and injuries, Mortality, Risk factors, Injury Severity Score, Lactic acid
Introduction
Recent improvements in the management of trauma and in critical care have reduced mortality globally among patients with multiple trauma [1,2]. However, trauma is still recognized as one of the most important public health problems worldwide, since it is a major cause of death and disability [2,3]. Approximately 5.8 million people die each year due to trauma, accounting for 10% of worldwide mortality [3,4]. Reducing mortality among patients with trauma is an extremely pressing issue, particularly in patients with severe trauma [5]. Mortality can be reduced by promptly identifying patients with injuries associated with high mortality rates at an early stage and providing adequate treatment, including intensive care, promptly [3,5]. Strategies implemented to identify patients with injuries associated with high mortality include the development of severity scoring systems. Several trauma scores have been developed to predict mortality among patients with trauma, of which the Injury Severity Score (ISS) is the most commonly used [6,7]. The ISS is associated with mortality, and severe trauma is defined as an ISS >15 [6]. While it is commonly used to predict mortality, there are limits to the use of the ISS as a decision tool in the emergency setting [7]. As the calculations needed for the ISS are complex and time-consuming, they are generally used for auditing and research purposes rather than for clinical decision-making [7,8]. To overcome this limitation, several studies have sought to identify biochemical markers or to develop new scoring systems (e.g., the Rapid Emergency Medicine Score [REMS] and Emergency Trauma Score [EMTRAS]) that can be used to predict mortality [3,7–11]. Unlike the ISS, using new scoring systems and biochemical markers to predict mortality has the advantage of generating scores rapidly [7–9]. However, not many studies have assessed the ability of these new scoring systems and biochemical makers to predict the mortality rate in severe trauma patients. Thus, we analyzed the usefulness of these new scoring systems and various biochemical values as predictors of in-hospital mortality among patients with severe trauma (ISS >15) and compared the discriminating power of those values to that of the ISS.
Methods
1) Study population and ethics approval
We retrospectively collected data from all patients admitted to the emergency department (ED) of South Korea’s Gyeongsang National University Hospital with an ISS >15 who were treated for severe trauma between July 2011 and June 2016. The following exclusion criteria were applied: ISS ≤15, age <15 years, penetrating injuries, patients discharged from the ED, death on arrival to hospital, or incomplete medical data. Overall, 582 patients were included in the study. Their medical records and electronic laboratory results were reviewed and the following information was extracted: patient demographics, mechanism of injury, ISS, biochemical markers on arrival at the ED, and in-hospital mortality. Due to the retrospective nature of this study, the requirement for informed consent was waived. The study protocol was approved by the Institutional Review Board of Gyeongsang National University Hospital (IRB approval no., GNUH 2018-03-008).
2) Definitions
The ISS is an injury severity scoring system that generates an overall score for patients with multiple trauma, with severe trauma defined as an ISS >15 [6]. An abbreviated injury scale (AIS) score is assigned to each injury in 1 of 6 body regions: head, face, chest, abdomen, distal (including pelvis), and external [6]. The scores for the 3 most severely injured body regions are squared and then summed to obtain the ISS score (range, 0 to 75) [6]. If an injury is assigned an AIS of 6 (unsurvivable injury), the ISS is automatically scored as 75 [6]. When considering mortality, we did not assess the specific cause of death.
The REMS is a new prognostic tool for in-hospital mortality assessment among patients admitted to the ED; it was developed by Olsson et al. [10] in 2004. It is a simplified version of the Acute Physiology and Chronic Health Evaluation (APACHE) II that allows for a rapid score calculation in 20 minutes [10]. The REMS is calculated based on patients’ Glasgow Coma Scale (GCS) score, respiration rate (RR), oxygen saturation, mean arterial pressure (MAP), and heart rate (HR); higher scores are associated with worse prognoses [10]. Table 1 shows the method of calculating the REMS.
Table 1.
The REMS scoring system
| Variable | REMS score | ||||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| 0 | +1 | +2 | +3 | +4 | +5 | +6 | |
| Age (yr) | <45 | 45–54 | 55–64 | 65–74 | >74 | ||
|
| |||||||
| Mean arterial pressure (mm Hg) | 70–109 | 110–129 | 130–159 | >159 | |||
| 50–69 | ≤49 | ||||||
|
| |||||||
| Heart rate (bpm) | 70–109 | 110–139 | 140–179 | >179 | |||
| 55–69 | 40–54 | ≤39 | |||||
|
| |||||||
| Respiratory rate (breaths/min) | 12–24 | 25–34 | 6–9 | 35–49 | >49 | ||
| 10–11 | ≤5 | ||||||
|
| |||||||
| O2 saturation (%) | >89 | 86–89 | 75–85 | <75 | |||
|
| |||||||
| Glasgow Coma Scale | 14–15 | 11–13 | 8–10 | 5–7 | 3–4 | ||
REMS, Rapid Emergency Medicine Score.
The EMTRAS was developed by Raum et al. [11] in 2009; it combines 4 early parameters from the ED, specifically age, GCS score, base excess (BE), and prothrombin time (PT). Raum et al. [11] suggested that the EMTRAS can be calculated within 30 minutes and accurately predicts mortality. Table 2 shows the method used to calculate the EMTRAS; higher scores are associated with worse prognoses [11].
Table 2.
The EMTRAS scoring system
| Variable | Category | EMTRAS scoring weight |
|---|---|---|
| Age (yr) | <40 | 0 |
|
| ||
| 41–60 | 1 | |
|
| ||
| 61–75 | 2 | |
|
| ||
| >75 | 3 | |
| Glasgow Coma Scale | 13–15 | 0 |
|
| ||
| 10–12 | 1 | |
|
| ||
| 6–9 | 2 | |
|
| ||
| 3–5 | 3 | |
|
| ||
| Base excess (mmol/L) | >−1 | 0 |
|
| ||
| −1 to −5 | 1 | |
|
| ||
| −5.1 to −10 | 2 | |
|
| ||
| <−10 | 3 | |
|
| ||
| Prothrombin time (%) | >80 | 0 |
|
| ||
| 50–80 | 1 | |
|
| ||
| 20–49 | 2 | |
|
| ||
| <20 | 3 | |
EMTRAS, Emergency Trauma Score.
3) Data analysis
We calculated p-values using the Mann-Whitney U-test for continuous variables and used the Fisher exact test or the Pearson chi-square test for categorical variables. Significance was set at p<0.05. To evaluate the risk factors for in-hospital mortality, we used logistic regression analysis. The efficacy of the various scores and biochemical parameters analyzed in this study for predicting mortality we included relevant variables with p<0.20 in the univariate analysis. We calculated the associations between the variables included in the multivariate analysis, with significance set at p<0.05. Sensitivity and specificity were defined using receiver operating characteristic (ROC) curves. The discriminant power of the REMS, EMTRAS, ISS, and serum lactate level (SLL) were compared using area under the ROC curve (AUC) analyses. Statistical analyses were performed using IBM SPSS software ver. 24.0 (IBM Corp., Armonk, NY, USA) and R software ver. 3.3.4 for Windows (R Foundation for Statistical Computing, Vienna, Austria).
Results
Of the 582 patients included in this study, 87 (14.9%) died. Of the deaths, 58 (66.7%) were in patients with head injuries, 15 (17.2%) were due to thoracic injuries, 8 (9.2%) were due to abdominal injuries, 5 were due to cervical injuries (5.7%), and 1 person (1.1%) died of a pulmonary embolism. The median age was 59 years (interquartile range [IQR], 46–78 years) and the majority of patients were male (72.2%). Falls were the main cause of trauma (29.0%), followed by car accidents (23.2%), motorcycle accidents (18.9%), pedestrian accidents (15.1%), and slips (13.7%). The median ISS, REMS, and EMTRAS were 22 (IQR, 17–25), 6 (IQR, 3–8), and 4 (IQR, 2–5), respectively. Patients who survived had a median age of 57 years (IQR, 46–70 years) with a median ISS of 20 (IQR, 17–25), and 13.1% were involved in pedestrian accidents. Patients who died had a median age of 69 years (IQR, 55–78 years) with a median ISS of 25 (IQR, 20–29), and 26.4% were involved in pedestrian accidents. There were significant differences between survivors and non-survivors in age (p<0.001), involvement in a pedestrian accident (p=0.003), REMS (p<0.001), EMTRAS (p<0.001), and ISS (p<0.001). However, there were no significant differences in other variables (Table 3). Differences in biochemical markers between survivors and non-survivors were studied. We assessed the partial pressure of carbon dioxide (PCO2), partial pressure of oxygen (PO2), BE, hemoglobin, hematocrit, platelets, PT/international normalized ratio (PT/INR), blood urea nitrogen (BUN), creatinine, albumin, aspartate aminotransferase, alanine aminotransferase, lactate, C-reactive protein (CRP), creatine kinase (CK), and troponin I. Table 4 sets out the crude associations between initial biochemical markers on arrival at the ED and in-hospital mortality for patients with severe trauma. Statistically significant differences between survivors and non-survivors were found for almost all biochemical makers (except PCO2, PO2, BUN, CRP, and CK). Table 5 shows the results of the logistic regression analysis. Univariate logistic regression analysis revealed that in-hospital mortality of patients with severe trauma was significantly associated with the following: age, involvement in a pedestrian accident, ISS, REMS, EMTRAS, BE, hemoglobin, hematocrit, platelets, PT/INR, creatinine, albumin, and lactate. In the multivariate logistic regression analysis, in-hospital mortality was independently significantly associated with ISS (odds ratio [OR], 1.068; 95% confidence interval [CI], 1.010–1.130; p=0.021), REMS (OR, 1.186; 95% CI, 1.018–1.383; p=0.029), EMTRAS (OR, 2.168; 95% CI, 1.570–2.994; p<0.001), and SLL (OR, 1.355; 95% CI, 1.155–1.591; p<0.001).
Table 3.
Comparison of baseline characteristics between survivors and non-survivors
| Variable | Total (n=582) | Survivors (n=495) | Non-survivors (n=87) | p-value |
|---|---|---|---|---|
| Age (yr) | 59 (46–78) | 57 (46–70) | 69 (55–78) | <0.001 |
| Sex (male) | 420 (72.2) | 359 (72.5) | 61 (70.1) | 0.697 |
| Mechanism | ||||
| Pedestrian accident | 88 (15.1) | 65 (13.1) | 23 (26.4) | 0.003 |
| Slip | 80 (13.7) | 73 (14.7) | 7 (8.0) | 0.127 |
| Motorcycle accident | 110 (18.9) | 98 (19.8) | 12 (13.8) | 0.235 |
| Car accident | 135 (23.2) | 120 (24.2) | 15 (17.2) | 0.170 |
| Falls | 169 (29.0) | 139 (28.1) | 30 (34.5) | 0.249 |
| Past medical history | ||||
| Hypertension | 85 (14.6) | 75 (15.2) | 10 (11.5) | 0.415 |
| Diabetes mellitus | 43 (7.4) | 39 (7.9) | 4 (4.6) | 0.375 |
| Cerebrovascular attack | 9 (1.5) | 7 (1.4) | 2 (2.3) | 0.561 |
| Acute coronary syndrome* | 5 (0.9) | 4 (0.8) | 1 (1.1) | 0.761 |
| Cancer | 4 (0.7) | 3 (0.6) | 1 (1.1) | 0.600 |
| Injury Severity Score | 22 (17–25) | 20 (17–25) | 25 (20–29) | <0.001 |
| Rapid Emergency Medicine Score | 6 (3–8) | 5 (3–7) | 10 (8–12) | <0.001 |
| Emergency Trauma Score | 4 (2–5) | 3 (2–5) | 7 (5–8) | <0.001 |
Values are presented as median (interquartile range) for continuous values or number (%) for categorical values.
Included unstable angina, non-ST elevation myocardial infarction, and ST elevation myocardial infarction.
Table 4.
Comparison of biochemical markers between survivors and non-survivors
| Variable | Total (n=582) | Survivors (n=495) | Non-survivors (n=87) | p-value |
|---|---|---|---|---|
| PCO2 (mm Hg) | 35 (30 to 40) | 35 (31 to 40) | 32 (27 to 42) | 0.104 |
| PO2 (mm Hg) | 83.0 (53 to 102) | 84 (67 to 103) | 77 (52 to 102) | 0.058 |
| Base excess (mmol/L) | −2.9 (−11.1 to −0.6) | −2.7 (−5.4 to −0.5) | −6.1 (−10.9 to −1.7) | <0.001 |
| Hemoglobin (g/dL) | 12.7 (11.2 to 14.0) | 12.9 (11.4 to 14.1) | 12.0 (9.9 to 13.7) | <0.001 |
| Hematocrit (%) | 38 (33 to 44) | 38 (34 to 42) | 36 (30 to 41) | <0.001 |
| Platelet (×103/mm3) | 228 (182 to 270) | 234 (189 to 278) | 189 (165 to 238) | <0.001 |
| Prothrombin time/international normalized ratio | 1.10 (1.02 to 1.56) | 1.08 (1.02 to 1.19) | 1.35 (1.13 to 1.84) | <0.001 |
| Blood urea nitrogen (mg/dL) | 15.6 (12.1 to 18.7) | 15.6 (12.5 to 18.8) | 15.0 (11.5 to 18.6) | 0.185 |
| Creatinine (mg/dL) | 0.85 (0.68 to 1.03) | 0.84 (0.67 to 1.00) | 0.96 (0.80 to 1.16) | 0.001 |
| Albumin (g/dL) | 4.0 (3.5 to 4.3) | 4.0 (3.6 to 4.3) | 3.7 (3.0 to 4.1) | <0.001 |
| Aspartate aminotransferase (U/L) | 50 (31 to 114) | 48 (30 to 105) | 72 (40 to 227) | 0.001 |
| Alanine aminotransferase (U/L) | 35 (20 to 64) | 33 (20 to 63) | 43 (23 to 117) | 0.012 |
| Lactate (mmol/L) | 2.70 (1.48 to 7.91) | 2.4 (1.4 to 4.2) | 4.7 (3.0 to 8.7) | <0.001 |
| C-reactive protein (mg/L) | 0.9 (0.4 to 2.7) | 1.0 (0.5 to 2.9) | 0.8 (0.4 to 2.3) | 0.179 |
| Creatine kinase (U/L) | 314 (179 to 547) | 304 (175 to 555) | 373 (212 to 541) | 0.242 |
| Troponin I (ng/mL) | 0.10 (0.10 to 0.13) | 0.10 (0.10 to 0.11) | 0.10 (0.10 to 0.26) | 0.003 |
Values are presented as median (interquartile range) for continuous values.
PCO2, partial pressure of carbon dioxide; PO2, partial pressure of oxygen.
Table 5.
Logistic regression analysis of risk factors for mortality in patients with severe trauma (n=582)a)
| Risk factors | Univariate | Multivariate | ||
|---|---|---|---|---|
|
|
|
|||
| OR (95% CI) | p-value | OR (95% CI) | p-valueb) | |
| Age | 1.036 (1.020–1.052) | <0.001 | 0.990 (0.959–1.022) | 0.526 |
|
| ||||
| Sex (male) | 0.889 (0.539–1.465) | 0.644 | ||
|
| ||||
| Pedestrianaccident | 2.377 (1.381–4.093) | 0.002 | 1.058 (0.391–2.862) | 0.912 |
|
| ||||
| Slips | 0.056 (0.225–1.139) | 0.100 | 0.281 (0.069–1.138) | 0.075 |
|
| ||||
| Motorcycleaccident | 0.648 (0.339–1.239) | 0.190 | 0.631 (0.219–1.818) | 0.394 |
|
| ||||
| Car accident | 0.651 (0.360–1.178) | 0.156 | 0.845 (0.298–2.396) | 0.375 |
|
| ||||
| Falls | 1.348 (0.831–2.186) | 0.226 | ||
|
| ||||
| Hypertension | 1.378 (0.682–2.784) | 0.371 | ||
|
| ||||
| Diabetes mellitus | 1.779 (0.619–5.109) | 0.285 | ||
|
| ||||
| Cerebrovascular attack | 0.611 (0.125–2.990) | 0.543 | ||
|
| ||||
| Acute coronary syndrome | 0.702 (0.078–6.357) | 0.753 | ||
|
| ||||
| Cancer | 0.525 (0.054–5.110) | 0.579 | ||
|
| ||||
| Injury Severity Score | 1.071 (1.037–1.107) | <0.001 | 1.068 (1.010–1.130) | 0.021 |
|
| ||||
| Rapid Emergency Medicine Score | 1.535 (1.405–1.678) | <0.001 | 1.186 (1.018–1.383) | 0.029 |
|
| ||||
| Emergency Trauma Score | 2.063 (1.788–2.380) | <0.001 | 2.168 (1.570–2.994) | <0.001 |
|
| ||||
| PCO2 | 1.015 (0.991–1.039) | 0.220 | ||
|
| ||||
| PO2 | 0.998 (0.992–1.004) | 0.531 | ||
|
| ||||
| Base excess | 0.900 (0.865–0.937) | <0.001 | 1.007 (0.912–1.112) | 0.894 |
|
| ||||
| Hemoglobin | 0.809 (0.735–0.890) | <0.001 | 0.861 (0.360–2.058) | 0.737 |
|
| ||||
| Hematocrit | 0.932 (0.901–0.964) | <0.001 | 1.293 (0.858–1.949) | 0.220 |
|
| ||||
| Platelet | 0.992 (0.988–0.996) | <0.001 | 0.994 (0.998–1.000) | 0.063 |
|
| ||||
| Prothrombin time/international normalized ratio | 1.416 (1.121–1.789) | 0.004 | 0.910 (0.726–1.142) | 0.417 |
|
| ||||
| Blood urea nitrogen | 0.997 (0.963–1.032) | 0.866 | ||
|
| ||||
| Creatinine | 1.625 (1.131–2.332) | 0.009 | 1.293 (0.858–1.949) | 0.220 |
|
| ||||
| Albumin | 0.444 (0.325–0.606) | <0.001 | 0.955 (0.419–2.177) | 0.912 |
|
| ||||
| Aspartate aminotransferase | 1.001 (1.000–1.002) | 0.071 | 0.999 (0.997–1.002) | 0.570 |
|
| ||||
| Alanine aminotransferase | 1.001 (1.000–1.002) | 0.110 | 1.001 (0.998–1.005) | 0.451 |
|
| ||||
| Lactate | 1.314 (1.224–1.412) | <0.001 | 1.298 (1.118–1.507) | <0.001 |
|
| ||||
| C-reactive protein | 0.991 (0.979–1.003) | 0.134 | 0.997 (0.987–1.006) | 0.505 |
|
| ||||
| Creatine kinase | 1.000 (1.000–1.000) | 0.963 | ||
|
| ||||
| Troponin I | 1.168 (0.945–1.444) | 0.150 | 0.901 (0.599–1.355) | 0.618 |
OR, odds ratio; CI, confidence interval; PCO2, partial pressure of carbon dioxide; PO2, partial pressure of oxygen.
The multivariate analysis included variables with significant associations in the univariate analysis (p<0.2).
Boldface indicates significant differences in the statistical comparisons of baseline characteristics.
ROC curves were used to estimate the sensitivity, specificity, and cutoff values of the ISS, REMS, EMTRAS, and SLL to predict in-hospital mortality (Fig. 1). A REMS value of 8 was set as the cutoff for in-hospital mortality. The in-hospital mortality rate for patients with severe trauma with a REMS >8 was 47.8%, compared with an in-hospital mortality rate of 5.1% for those with a REMS ≤8 (p<0.001; OR, 16.894). The REMS had a sensitivity of 77.0% and a specificity of 79.0% for in-hospital mortality. The positive predictive value (PPV) was 39.2%, but its negative predictive value (NPV) was higher than expected, at 95.1%. An EMTRAS value of 5 was set as the cutoff for predicting in-hospital mortality. The in-hospital mortality rate for patients with severe trauma with an EMTRAS >5 was 47.1%, compared with the in-hospital mortality rate of 5.4% for those with an EMTRAS ≤5 (p<0.001; OR, 15.676). The EMTRAS had a sensitivity of 83.9% and a specificity of 72.5% for in-hospital mortality. The PPV was 34.9% and the NPV was again higher than expected, at 96.2%. A SLL of 3.5 mmol/L was set as the cutoff value for predicting in-hospital mortality, as 28.4% of patients with a SLL >3.5 mmol/L died versus 8.4% with a SLL ≤3.5 mmol/L (p<0.001; OR, 4.276). At this cutoff value, SLL as a predictor of mortality had a sensitivity of 64.4%, specificity of 65.3%, PPV of 24.6%, and NPV of 91.2%. An ISS value of 24 was set as the cutoff for predicting in-hospital mortality, as 20.0% of patients with an ISS >24 died versus 11.6% with an ISS ≤24 (p=0.006; OR, 1.909). At this cutoff value, the sensitivity and specificity values for ISS as a predictor of in-hospital mortality were 56.8% and 60.9%, respectively. For the ISS, the PPV was 20.3% and the NPV was 88.9% (Table 6). In a pairwise AUC comparison, REMS (p=0.002), EMTRAS (p<0.001), and SLL (p=0.043) were each significantly better predictors of in-hospital mortality than the ISS (Table 7, Fig. 1).
Fig. 1.
Receiver operating characteristic curves for mortality in patients with severe trauma. AUC, area under the receiver operating characteristic curve; EMTRAS, Emergency Trauma Score; ISS, Injury Severity Score; REMS, Rapid Emergency Medicine Score.
Table 6.
Sensitivity, specificity, PPV, and NPV of REMS, EMTRAS, and lactate for in-hospital mortality
| Variable | REMS >8 | EMTRAS >5 | Lactate >3.5 mmol/L | ISS >24 |
|---|---|---|---|---|
| Mortality (%) | 47.8 | 47.1 | 28.4 | 20.0 |
| Sensitivity (%) | 77.0 | 83.9 | 64.4 | 56.8 |
| Specificity (%) | 79.0 | 72.5 | 65.3 | 60.9 |
| PPV (%) | 39.2 | 34.9 | 24.6 | 20.3 |
| NPV (%) | 95.1 | 96.2 | 91.2 | 88.9 |
PPV, positive predictive value; NPV, negative predictive value; REMS, Rapid Emergency Medicine Score; EMTRAS, Emergency Trauma Score; ISS, Injury Severity Score.
Table 7.
Pairwise comparison of the area under the curve for various predictors
| Variable | Difference between the areas | Standard error | Z | p-value |
|---|---|---|---|---|
| REMS–ISS | 0.211 | 0.039 | 5.490 | <0.001 |
| EMTRAS–ISS | 0.234 | 0.0038 | 6.304 | <0.001 |
| Lactate–ISS | 0.099 | 0.043 | 2.277 | 0.023 |
REMS, Rapid Emergency Medicine Score; ISS, Injury Severity Score; EMTRAS, Emergency Trauma Score.
Discussion
Despite advances in injury prevention and medical care, trauma-related mortality remains a major public health problem worldwide [1,2]. Trauma is the leading cause of death in the United States among people under 44 years of age and among the top 5 causes of death in people over 45 years of age [12]. Prompt provision of appropriate treatment for patients with trauma is crucial for overall survival and treatment outcomes [3]. In order to improve overall survival and treatment outcomes, it is important to promptly and accurately determine the severity of patients with trauma in the ED [1,3]. Therefore, we assessed the initial risk factors for in-hospital mortality among patients with severe trauma in the ED. We conducted a retrospective medical record review of 582 patients with severe trauma (ISS >15) between July 2011 and June 2016, and analyzed the associations of in-hospital mortality with the baseline characteristics and initial biochemical markers of patients with severe trauma on admission. We found that the REMS, EMTRAS, and SLL could easily and rapidly be used to predict in-hospital mortality of patients with severe trauma in the ED.
Several scoring systems have been developed to objectively measure the initial condition of a patient with trauma, and these may also serve as a prognostic indicator for specific patients [11]. The ISS, developed by Copes et al. [6] in 1988, is the most commonly used injury scoring system. The ISS is an anatomical scoring system based upon the AIS, and an ISS >15 has traditionally been considered as a marker of severe injuries [3,6]. Several studies have shown that the ISS is a valid predictor of mortality, length of stay in the hospital or the intensive care unit (ICU), and cost of trauma care [13–15]. Recent studies evaluating the medical records of over 50,000 patients with trauma found that the ISS was a better discriminator of patients with a high risk of mortality [15,16]. We found that the ISS (OR, 1.038; 95% CI, 1.010–1.130; p=0.021) was a statistically significant risk factor for in-hospital mortality in patients with severe trauma, corroborating the results of the above studies.
The APACHE II is a trauma severity scoring system for use in patients in the ICU. The APACHE II is assessed within 1 day of the patient’s admission to the ICU; higher APACHE II scores are associated with higher mortality [17]. However, the APACHE II score is calculated using the patient’s age and 12 routine physiological measurements; therefore, it has limited use as a rapid scoring system in the ED [10]. The REMS is an abbreviated version of the APACHE II, and includes the GCS score, RR, oxygen saturation, MAP, and HR [10]. The REMS can be rapidly determined in 20 minutes and has been shown to correlate with mortality in patients with trauma in several previous studies [16]. Imhoff et al. [16] showed that a higher REMS was associated with increased mortality (OR, 1.51; 95% CI, 1.45–1.58; p=0.001) in American patients with trauma. Nakhjavan-Shahraki et al. [18] suggested that the REMS could be used for predicting mortality (AUC=0.93) and poor outcomes (p=0.001) in patients with trauma in emergency settings. Our findings for patients with severe trauma with an ISS >15 are consistent with these previous findings. We found that the in-hospital mortality of patients with severe trauma was significantly associated with the REMS (OR, 1.186; 95% CI, 1.018–1.383; p=0.029), suggesting that a higher REMS might be an accurate, rapid predictor of in-hospital mortality for patients with severe trauma; in fact, in a pairwise AUC comparison, the REMS was significantly better than the ISS as a predictor of in-hospital mortality (p<0.001).
The EMTRAS was developed by Raum et al. [11] in 2009 as an easy-to-use tool to predict mortality among patients with trauma in emergency settings. The EMTRAS is based on 4 parameters (age, GCS score, BE, and PT) and has been found to correlate with the mortality of trauma patients in previous studies [3,11,19]. Park et al. [3] showed that the EMTRAS predicted in-hospital mortality for patients with trauma more easily and accurately than other trauma scores (p<0.001, AUC=0.957). A comparison of the EMTRAS with other scores was performed among 150 patients in Italy [19]. Those researchers found that the EMTRAS correlated well with mortality (AUC=0.809) and suggested that it should be performed before other examinations, such as computed tomography, in patients with trauma because it provides simple and rapid findings [3]. Our study of patients with severe trauma with an ISS >15 showed similar results. We found that in-hospital mortality of patients with severe trauma was significantly associated with the EMTRAS (OR, 2.168; 95% CI, 1.570–2.994; p<0.001); in a pairwise AUC comparison, the EMTRAS was a significantly better predictor of in-hospital mortality than the ISS (p<0.001).
Lactate is constantly produced from pyruvate during normal metabolism and exercise, and the lactate level does not increase until the production rate exceeds the removal rate [20]. Metabolism from pyruvic acid to lactate is increased in tissue with inadequate perfusion, and higher lactate concentrations than in the normal state are associated with severe trauma, post-traumatic respiratory failure, and death [20,21]. Several studies have investigated SLL as predictor of mortality among patients with trauma [21,22]. Abramson et al. [22] studied SLL in 76 patients with multiple trauma admitted to the ICU and concluded that SLL was a useful value for predicting the prognosis of patients with severe trauma. Parsikia et al. [23] reviewed 1,941 patients with trauma between 2007 and 2012 in a prospectively maintained database and found that the median SLL among non-survivors was 32 mg/dL (IQR, 17–62 mg/dL), versus 21 mg/dL (IQR, 14–32 mg/dL) for survivors (p<0.001). In their multivariate analysis, SLL was a significant covariant risk factor for mortality (OR, 1.010; 95% CI, 1.002–1.019; p<0.001) [23]. Our findings are consistent with the above studies, which found that SLL was significantly associated with treatment outcomes in patients with trauma. In our study, SLL (OR, 1.038; 95% CI, 1.010–1.130; p=0.021) was a statistically significant risk factor associated with mortality among patients with severe trauma; in a pairwise AUC comparison, SLL was a significantly better predictor of in-hospital mortality than the ISS (p=0.023).
Our study has several limitations. First, we evaluated a relatively small number of patients with severe trauma and larger studies are required to validate our findings. Second, our study evaluated patients from a single hospital, which may have introduced selection bias. Third, our study was a retrospective evaluation and, as with all trauma registries, the accuracy of the recorded data may vary [24]. Accurate prospective and larger-population studies are needed to support our findings.
In conclusion, this study found that the REMS, EMTRAS, SLL, and ISS were statistically significant risk factors for mortality among patients with severe trauma. However, the ISS calculations are complex, time-consuming, and are not used to guide early resuscitation in practice [11,16]. Unlike the ISS, the REMS, EMTRAS, and SLL are based on measurements that are easier to obtain and were found to be significantly better predictors of in-hospital mortality than the ISS upon pairwise comparisons of the AUC. Therefore, we suggest that the REMS. EMTRAS, and SLL be used as easy and rapid predictors of mortality among patients with severe trauma on admission to the ED.
Acknowledgments
This study was supported by a Grant of the Samsung Vein Clinic Network (Daejeon, Anyang, Cheongju, Cheonan; Fund no. KTCS04-131). We acknowledge the outstanding contributions of the technicians and nursing staff at Gyeongsang National University Hospital, Korea.
Footnotes
Conflict of interest
No potential conflict of interest relevant to this article was reported.
References
- 1.Schoeneberg C, Schilling M, Burggraf M, Fochtmann U, Lendemans S. Reduction in mortality in severely injured patients following the introduction of the “treatment of patients with severe and multiple injuries” guideline of the German society of trauma surgery: a retrospective analysis of a level 1 trauma center (2010–2012) Injury. 2014;45:635–8. doi: 10.1016/j.injury.2013.11.024. [DOI] [PubMed] [Google Scholar]
- 2.Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ. Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet. 2006;367:1747–57. doi: 10.1016/S0140-6736(06)68770-9. [DOI] [PubMed] [Google Scholar]
- 3.Park HO, Kim JW, Kim SH, et al. Usability verification of the Emergency Trauma Score (EMTRAS) and Rapid Emergency Medicine Score (REMS) in patients with trauma: a retrospective cohort study. Medicine (Baltimore) 2017;96:e8449. doi: 10.1097/MD.0000000000008449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lendrum RA, Lockey DJ. Trauma system development. Anaesthesia. 2013;68(Suppl 1):30–9. doi: 10.1111/anae.12049. [DOI] [PubMed] [Google Scholar]
- 5.Jelodar S, Jafari P, Yadollahi M, et al. Potential risk factors of death in multiple trauma patients. Emerg (Tehran) 2014;2:170–3. [PMC free article] [PubMed] [Google Scholar]
- 6.Copes WS, Champion HR, Sacco WJ, Lawnick MM, Keast SL, Bain LW. The Injury Severity Score revisited. J Trauma. 1988;28:69–77. doi: 10.1097/00005373-198801000-00010. [DOI] [PubMed] [Google Scholar]
- 7.Jang HN, Park HO, Yang TW, et al. Biochemical markers as predictors of in-hospital mortality in patients with severe trauma: a retrospective cohort study. Korean J Crit Care Med. 2017;32:240–6. doi: 10.4266/kjccm.2017.00360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sammour T, Kahokehr A, Caldwell S, Hill AG. Venous glucose and arterial lactate as biochemical predictors of mortality in clinically severely injured trauma patients: a comparison with ISS and TRISS. Injury. 2009;40:104–8. doi: 10.1016/j.injury.2008.07.032. [DOI] [PubMed] [Google Scholar]
- 9.Mikhail J. The trauma triad of death: hypothermia, acidosis, and coagulopathy. AACN Clin Issues. 1999;10:85–94. doi: 10.1097/00044067-199902000-00008. [DOI] [PubMed] [Google Scholar]
- 10.Olsson T, Terent A, Lind L. Rapid Emergency Medicine score: a new prognostic tool for in-hospital mortality in nonsurgical emergency department patients. J Intern Med. 2004;255:579–87. doi: 10.1111/j.1365-2796.2004.01321.x. [DOI] [PubMed] [Google Scholar]
- 11.Raum MR, Nijsten MW, Vogelzang M, et al. Emergency trauma score: an instrument for early estimation of trauma severity. Crit Care Med. 2009;37:1972–7. doi: 10.1097/CCM.0b013e31819fe96a. [DOI] [PubMed] [Google Scholar]
- 12.Centers for Disease Control and Prevention. Cost of injuries and violence in the United States [Internet] Atlanta (GA): Centers for Disease Control and Prevention; [cited 2018 Mar 13]. Available from: http://www.cdc.gov/injury/wisqars/overview/cost_of_injury.html. [Google Scholar]
- 13.Palmer C. Major trauma and the injury severity score: where should we set the bar? Annu Proc Assoc Adv Automot Med. 2007;51:13–29. [PMC free article] [PubMed] [Google Scholar]
- 14.Yeen W, Carrillo EH, Bragg S, Marsh W, Gilson S, Meltzer T. Cost analysis of vertical deceleration injury in the acute setting. Eur J Trauma Emerg Surg. 2007;33:251–5. doi: 10.1007/s00068-007-6057-1. [DOI] [PubMed] [Google Scholar]
- 15.Brown JB, Gestring ML, Leeper CM, et al. The value of the injury severity score in pediatric trauma: time for a new definition of severe injury? J Trauma Acute Care Surg. 2017;82:995–1001. doi: 10.1097/TA.0000000000001440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Imhoff BF, Thompson NJ, Hastings MA, Nazir N, Moncure M, Cannon CM. Rapid Emergency Medicine Score (REMS) in the trauma population: a retrospective study. BMJ Open. 2014;4:e004738. doi: 10.1136/bmjopen-2013-004738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wong DT, Crofts SL, Gomez M, McGuire GP, Byrick RJ. Evaluation of predictive ability of APACHE II system and hospital outcome in Canadian intensive care unit patients. Crit Care Med. 1995;23:1177–83. doi: 10.1097/00003246-199507000-00005. [DOI] [PubMed] [Google Scholar]
- 18.Nakhjavan-Shahraki B, Baikpour M, Yousefifard M, et al. Rapid acute physiology score versus rapid emergency medicine score in trauma outcome prediction: a comparative study. Emerg (Tehran) 2017;5:e30. [PMC free article] [PubMed] [Google Scholar]
- 19.Mangini M, Di Valvasone S, Greco C, et al. Validation of the new proposed Emergency Trauma Score (EMTRAS) Crit Care. 2010;14(Suppl 1):252. doi: 10.1186/cc8484. [DOI] [Google Scholar]
- 20.Goodwin ML, Harris JE, Hernandez A, Gladden LB. Blood lactate measurements and analysis during exercise: a guide for clinicians. J Diabetes Sci Technol. 2007;1:558–69. doi: 10.1177/193229680700100414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kellum JA, Kramer DJ, Lee K, Mankad S, Bellomo R, Pinsky MR. Release of lactate by the lung in acute lung injury. Chest. 1997;111:1301–5. doi: 10.1378/chest.111.5.1301. [DOI] [PubMed] [Google Scholar]
- 22.Abramson D, Scalea TM, Hitchcock R, Trooskin SZ, Henry SM, Greenspan J. Lactate clearance and survival following injury. J Trauma. 1993;35:584–8. doi: 10.1097/00005373-199310000-00014. [DOI] [PubMed] [Google Scholar]
- 23.Parsikia A, Bones K, Kaplan M, et al. The predictive value of initial serum lactate in trauma patients. Shock. 2014;42:199–204. doi: 10.1097/SHK.0000000000000208. [DOI] [PubMed] [Google Scholar]
- 24.Owen JL, Bolenbaucher RM, Moore ML. Trauma registry databases: a comparison of data abstraction, interpretation, and entry at two level I trauma centers. J Trauma. 1999;46:1100–4. doi: 10.1097/00005373-199906000-00025. [DOI] [PubMed] [Google Scholar]