a b s t r a c t

Background: The shock index (heart rate divided by systolic blood pressure) of trauma patients upon emergency department arrival predicts blood loss and death. However, some patients with normal shock indices (0.4 < shock index <0.9) upon emergency department arrival also have poor prognoses. This study aimed to de- termine whether abnormal prehospital shock indices in trauma patients with normal shock indices upon emer- gency department arrival were predictors of a high risk of mortality.

Methods: We conducted a retrospective cohort study of emergency department-admitted trauma patients from 2004 to 2017. The study included 89,495 consecutive trauma patients aged >=16 years, with Abbreviated Injury Scale score of >=3, who were transported to the emergency department directly from the field and had a normal shock index upon emergency department arrival. According to the prehospital shock index scores, the patients were categorized into low shock index (<= 0.4), normal shock index, and high shock index (>=0.9) groups. Odds ra- tios and 95% confidence intervals were calculated using logistic regression analysis.

Results: The 89,495 patients had a median age of 64 (interquartile range: 43-79) years, and 55,484 (62.0%) of the patients were male. There were 1350 (1.5%) 24-h deaths in total; 176/4263 (4.1%), 1017/78,901 (1.3%), and 157/ 6331 (2.5%) patients were in the low, normal, and high prehospital shock index groups, respectively. The ad- justed odds ratios for 24-h mortality compared with the normal shock index group were 1.63 (95% confidence interval: 1.34-1.99) in the low shock index group and 1.62 (95% confidence interval: 1.31-1.99) in the high shock index group.

Conclusion: Trauma patients with abnormal prehospital shock indices but normal shock indices upon emergency department arrival are at a higher risk of 24-h mortality. Identifying these indices could improve triage and targeted care for patients.

(C) 2023 Published by Elsevier Inc.

1. Background

* Corresponding author at: Department of Healthcare Epidemiology, School of Public Health, Graduate School of Medicine, Kyoto University, Yoshida Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan.

E-mail address: [email protected] (Y. Yamamoto).

Approximately 4.4 million people die from traumatic injuries each year [1] accounting for approximately 8% of all annual deaths world- wide. The main causes of Trauma deaths include road accidents, suicides, homicides, and falls. There are nearly 70,000 yearly trauma- related deaths in Japan, accounting for about 5% of all deaths [2], and their main causes are unintentional accidents and suicide. According

https://doi.org/10.1016/j.ajem.2023.05.008 0735-6757/(C) 2023 Published by Elsevier Inc.

to the Tokyo Fire and Disaster Management Agency, trauma accounts for 27% of emergency transportation cases [3]. Therefore, identifying trauma patients with a higher risk of mortality or patients who need im- mediate diagnosis and treatment in the emergency department (ED) is crucial.

Vital signs are used to predict the severity of injury and prognosis of trauma patients [4,5]. The shock index (SI; heart rate divided by systolic blood pressure) [6] is a predictor of visible and hidden blood loss, a need for blood transfusion, injury severity, and mortality [7]. The SI is useful in patients without obvious Vital sign abnormalities [8,9], has a greater Predictive ability than any vital sign [5,10,11], and is easier to calculate than the other indices, such as the age shock index [10,12], the reverse shock index [13], or the Trauma and Injury Severity Score [14]. It is known that a high SI (e.g., >= 0.9) at prehospital or upon ED arrival is as- sociated with increased mortality risk [15-17], whereas a low SI (e.g., <= 0.4; high blood pressure and low heart rate) has been suggested as a predictor of a serious head injury, leading to an increased mortality risk [18,19]. However, a low SI cut-off value has rarely been considered in previous studies.

The middle SI range (0.4 < SI < 0.9) could be considered “normal”

and may indicate a good prognosis. However, some patients may have a normal SI range upon ED arrival but have a poor prognosis. Vital sign changes over time due to physiological compensatory mechanisms could help detect patients at high risk of mortality [20]. The SI may also vary depending on when the vital signs are measured [21]. Sometimes, an abnormal SI measured immediately after an injury could become normal upon ED arrival. Therefore, prehospital SI might be useful to fur- ther stratify mortality risk among patients with normal SI upon ED ar- rival. However, the association between prehospital SI and prognosis in such patients has not been examined. Therefore, this study aimed to determine whether an abnormal prehospital SI (SI >= 0.9 or <=0.4) was associated with a higher risk of 24-h mortality than normal prehos- pital SI (0.4 < SI < 0.9) among trauma patients with normal SI upon ED arrival.

1. Methods
   1. Study design and participants

We performed a retrospective cohort study of ED-admitted trauma patients using anonymized data from the Japan Trauma Data Bank (JTDB). The JTDB was approved by the ethics committee of the National Defence Medical College. The ethics committee of Kyoto University ap- proved our research (approval number: R2601). The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) state- ment was used to ensure the proper reporting of methods, results, and discussion.

The eligibility criterion was as follows: trauma patients aged

>=16 years, transported to the ED directly from the field and had normal SI upon arrival. The exclusion criteria were as follows: Burn injury [4,22], prehospital cardiopulmonary arrest, prehospital fluid infusion, hypotension (systolic blood pressure < 90 mmHg) [11,23] upon ED ar- rival, bradycardia (heart rate <= 40 bpm) [24] upon ED arrival, as well as missing data for the systolic blood pressure, heart rate, or outcome. We excluded patients with burns according to the methods of previous studies because their treatment differs from that for other causes of trauma [4,22]. Patients who presented with hypotension and bradycar- dia upon ED arrival were excluded because they usually needed prompt examination or intervention.

• 1. Setting and data sources

We used the JTDB registry data from 2004 to 2017 [25]. The JTDB is a nationwide prospective registry of trauma cases in Japan, established by the Japanese Association for the Surgery of Trauma and the Japanese As- sociation for Acute Medicine. A total of 264 emergency hospitals across

Japan participate in the registry, comprising approximately 70% of the government-certified tertiary emergency and critical care centers [26]. This registry enrolled approximately 300,000 trauma patients who pre- sented to an ED with an Abbreviated Injury Scale score of >=3 for any part of their body. Paramedics and medical staff measured the pre- hospital vital signs and those upon ED arrival, treatments, diagnoses, in- jury severity, and in-hospital mortality, and compiled to form the registry data.

• 1. Measurement
     1. Exposure

The exposure was prehospital SI measured by the emergency medi- cal services in the field. The SI was categorized into low (<=0.4), normal (0.4 < SI < 0.9), and high (>=0.9) SI groups. These categories were se- lected because the association between SI upon ED arrival and mortality has been reported to follow a U-shaped curve, with SIs of 0.4 and 0.9 being associated with approximately equal mortality rates with the lowest mortality rate found in between these values [19].

• • 1. Outcome

The primary outcome was mortality within 24 h of ED arrival. The secondary outcomes were invasive Hemostatic interventions (tho- racoabdominal surgery, endoscopic surgery, surgical hemostasis, angiostomy, and transcatheter arterial embolisation), blood transfusion within 24 h, head surgery, and in-hospital mortality.

• • 1. Other factors to be adjusted

In the multivariable analyses, adjustment was performed using the factors below: the patients’ sex, age, Glasgow Coma Scale (GCS) upon ED arrival, respiratory rate upon ED arrival, year of ED arrival, transpor- tation time (time of departure from the field to ED arrival), type of injury (blunt, penetrating, unknown, and other injuries), cause of in- jury (unintentional accident, occupational accident, suicide attempt, assault by others, unknown, and other causes), and comorbidities (respiratory, cardiovascular, digestive, metabolic, central nervous system, mental, or immunodeficiency diseases and cancer). All the variables were recorded by paramedics and medical staff at each participating hospital.

• 1. Statistical analysis
     1. Descriptive analysis

The eligible patient characteristics were summarized for the entire cohort and each SI group. The continuous variables were presented as the medians and interquartile ranges (IQRs), while the categorical vari- ables were presented as numbers and percentages.

• • 1. Primary analysis

We calculated the odds ratios (ORs) and 95% confidence intervals (CIs) for the 24-h mortality of the prehospital low and high SI groups and compared them with those of the normal SI group using a logistic regression analysis after adjusting for the abovementioned factors. Model 1 included the covariates usually available at the time of pa- tients’ ED arrival: sex, age, GCS, respiratory rate, year of ED arrival, transportation time, and type of injury. The other factors, such as the cause of injury and comorbidities, were not necessarily obtain- able during ED arrival. These respective factors were added in Model 2 and Model 3.

• • 1. Secondary analysis

We used logistic regression analyses to calculate the ORs for the sec- ondary outcomes: invasive hemostatic interventions, blood transfusion within 24 h, head surgery, and in-hospital mortality. The same indepen- dent variables used in the primary analysis were used to determine whether the prehospital SI could predict the secondary outcomes.

• • 1. Subgroup analysis

We also performed the same analysis performed in the primary analysis for the following subgroups: patients with isolated serious head injuries (defined as a head AIS score >= 3 and AIS score < 3 in other body parts) and patients without serious head injuries (AIS score >= 3 in other body parts and a head AIS score < 3).

• • 1. Sensitivity analysis

We performed three sensitivity analyses. Firstly, we changed the normal SI definition to 0.4 < SI < 1.0 and 0.5 < SI < 0.7. No definite cri- terion exists for a normal SI range; however, an SI “below 1.0” or “0.5 to 0.7” has also been considered to be a normal range in clinical settings and previous studies [8,23,27]. Secondly, we added the adjustment for SI upon ED arrival as a covariate into Model 1, 2 and 3 of the primary multivariable logistic regression analysis. This analysis was undertaken in consideration of the possibility that the level of SI upon ED arrival may be a stronger predictor than prehospital SI, despite the study pop- ulation falling within the normal range of SI upon ED arrival. Thirdly, we performed a multiple imputation (MI) for the missing adjusted vari- ables in the primary multivariable logistic regression analysis, based

on the assumption that the data were missing at random. The MI proce- dure imputed the missing values using chained equations with factors of all the variables used in Model 1. We created 20 imputed datasets and performed a logistic regression analysis for each. The results were integrated using Rubin’s rule.

All the analyses were conducted using STATA/MP, Version 15.1 (StataCorp, TX, USA). Except for the sensitivity analysis with MI, only the cases without any missing covariate values were included in the analyses (complete case analysis).

1. Results
   1. Patient characteristics

A total of 113,494 adult (age >= 16) trauma patients were transported directly to the ED from the field and had an SI within the normal range upon ED arrival (Fig. 1). The eligibility criteria were met by 89,495 (78.9%) patients, of whom 55,484 (62.0%) were male. The median age was 64 (IQR: 43-79) years (Table 1). Blunt and penetrating inju- ries accounted for 96.7% and 1.9% of all trauma cases, respectively. In

Fig. 1. Patient selection flowchart. Abbreviations: ED, emergency department.

[Model 1] was adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, and type of injury.

[Model 2] was adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, type of injury, and cause of injury.

[Model 3] was adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, type of injury, cause of injury, and comorbidities.

Table 1

Patient characteristics.

				Prehospital SI
Parameter	Total			SI <= 0.4			0.4 < SI < 0.9			SI >= 0.9
	N = 89,495			N = 4263			N = 78,901			N = 6331
Male	55,484	(62.0)		2456	(57.6)		48,821	(61.9)		4207	(66.5)
Missing	27	(0.0)		1	(0.0)		23	(0.0)		3	(0.0)
Age [year]a	64	[43-79]		74	[62-83]		65	[44-79]		50	[31-69]
At prehospital Shock index [bpm/mmHg]a	0.60	[0.51-0.72]		0.37	[0.34-0.39]		0.60	[0.52-0.70]		1.00	[0.94-1.10]
Heart rate [bpm]a	84	[73-95]		66	[60-72]		84	[74-93]		102	[90-115]
Systolic blood pressure [mm Hg]a	139	[120-160]		180	[161-200]		140	[121-159]		100	[87-111]
At ED arrival
Shock index [bpm/mmHg]a	0.58	[0.50-0.69]	0.48		[0.43-0.55]	0.58		[0.50-0.68]	0.74		[0.64-0.82]
Heart rate [bpm]a	82	[73-93]	75		[67-86]	82		[73-92]	92		[81-104]
Systolic blood pressure [mm Hg]a	140	[124-159]	155		[136-173]	140		[124-159]	129		[115-144]
Glasgow Coma Scalea	15	[14-15]	15		[13-15]	15		[14-15]	14		[13-15]
Missing	5075	(5.7)	245		(5.7)	4565		(5.8)	265		(4.2)
Respiratory rate [/min]a	20	[17-24]	20		[16-24]	20		[17-24]	21		[18-26]
Missing	8224	(9.2)	425		(10.0)	7353		(9.3)	446		(7.0)
Year
2004-2009	14,842	(16.6)	687		(16.1)	12,870		(16.3)	1285		(20.3)
2010-2014	48,751	(54.5)	2324		(54.5)	42,979		(54.5)	3448		(54.5)
2015-2017	25,902	(28.9)	1252		(29.4)	23,052		(29.2)	1598		(25.2)
Transportation time [min]a	13.1	[6.6-19.7]	13.1		[8.7-19.7]	10.9		[6.6-19.7]	13.1		[8.7-19.7]
Missing	7063	(7.9)	347		(8.1)	6244		(7.9)	472		(7.5)
Type of injury
Blunt injury	86,582	(96.7)	4152		(97.4)	76,523		(97.0)	5907		(93.3)
penetrating injury	1672	(1.9)	44		(1.0)	1282		(1.6)	346		(5.5)
Unknown	533	(0.6)	35		(0.8)	468		(0.6)	30		(0.5)
Others	136	(0.2)	3		(0.1)	119		(0.2)	14		(0.2)
Missing	572	(0.6)	29		(0.7)	509		(0.6)	34		(0.5)
Cause of injury
Unintentional accident	78,252	(87.4)	3828		(89.8)	69,477		(88.1)	4947		(78.1)
Occupational accident	5015	(5.6)	227		(5.3)	4436		(5.6)	352		(5.6)
Suicide attempt	2716	(3.0)	46		(1.1)	1965		(2.5)	705		(11.1)
Assault by others	993	(1.1)	23		(0.5)	824		(1.0)	146		(2.3)
Unknown	1161	(1.3)	65		(1.5)	997		(1.3)	99		(1.6)
Others	644	(0.7)	39		(0.9)	575		(0.7)	30		(0.5)
Missing 714		(0.8)	35		(0.8)	627		(0.8)	52		(0.8)
Comorbidities
Respiratory	4147	(4.6)	157		(3.7)	3655		(4.6)	335		(5.3)
Cardiovascular	27,116	(30.3)	1837		(43.1)	24,112		(30.6)	1167		(18.4)
Digestive	7055	(7.9)	378		(8.9)	6210		(7.9)	467		(7.4)
Metabolic	12,094	(13.5)	704		(16.5)	10,801		(13.7)	589		(9.3)
Central nervous system / mental	15,340	(17.1)	782		(18.3)	13,349		(16.9)	1209		(19.1)
Immunodeficiency / cancer	9151	(10.2)	508		(11.9)	8159		(10.3)	484		(7.6)
Body part with serious injury
Isolated serious head injury	24,925	(27.9)	1560		(36.6)	22,119		(28.0)	1246		(19.7)
Without serious head injury	55,032	(61.5)	2235		(52.4)	48,714		(61.7)	4083		(64.5)
Head and other body part injury	9538	(10.7)	468		(11.0)	8068		(10.2)	1002		(15.8)

Abbreviations: SI, shock index; ED, emergency department.

n (%), unless otherwise specified.

“Isolated serious head injury” was defined as a head Abbreviated Injury Scale score >= 3 and AIS score < 3 for other body parts.

“Without serious head injury” was defined as a head AIS score < 3 and AIS score >= 3 for other body parts.

“Head and other body part injury” were defined as AIS score >= 3 for both the head and other body parts.

^a Median [interquartile range].

total, 87.4% of the patients had unintentional accidents, 5.6% had oc- cupational accidents, and 3.0% attempted suicide. The proportions of the low, normal, and high SI groups were 4.8%, 88.2%, and 7.1%, respectively.

• 1. Primary analysis: association between prehospital SI and 24-h mortality

Overall, 1350 (1.5%) 24-h deaths occurred, including 176/4263 (4.1%), 1017/78,901 (1.3%), and 157/6331 (2.5%) in the low, normal, and high SI groups, respectively (Fig. 2A). Compared with the normal SI group, the unadjusted ORs for 24-h mortality in the low and high SI groups were 3.30 (95% CI: 2.80-3.88) and 1.95 (95% CI: 1.64-2.31), re-

spectively. The corresponding adjusted ORs for 24-h mortality in the low and high SI groups were 1.63 (95% CI: 1.34-1.99) and 1.62 (95%

CI: 1.31-1.99) in Model 1 (Fig. 2B), 1.65 (95% CI: 1.35-2.01) and 1.50

(95% CI: 1.21-1.85) in Model 2, and 1.63 (95% CI: 1.34-2.00) and 1.49

(95% CI: 1.21-1.85) in Model 3. The ORs and 95% CIs for the adjusted fac- tors are shown in the Supplementary (Table S1).

• 1. Secondary analysis: association between prehospital SI and invasive hemostatic interventions, blood transfusion within 24 h, head surgery, and in-hospital mortality

Similar to the primary analysis, the low and high SI groups showed higher ORs for in-hospital mortality than the normal SI group (Fig. S1). Compared with the normal SI group, the ORs for invasive he- mostatic interventions were higher in the high SI group but lower in the low SI group (Fig. 3A). The OR for blood transfusion within 24 h was higher in the high SI group than in the normal SI group (Fig. 3B).

Fig. 2. Primary analysis: odds ratios of the prehospital shock index for 24-h mortality based on the logistic regression analysis. Abbreviations: OR, odds ratio; CI, confidence interval.

Graph A is from the unadjusted model, and Graph B is from the model adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, and type of injury [Model 1].

The low SI group showed a higher OR for head surgery, and the high SI group showed a lower OR than the normal SI group (Fig. 3C).

• 1. Subgroup analysis: association between prehospital SI and 24-h mortality in patients with and without head injury

The low SI group in the subpopulation of patients with isolated seri- ous head injuries was associated with a higher risk of 24-h mortality than the normal SI group; however, there were no substantial differ- ences between the high and normal SI groups in this subgroup analysis (Fig. 4A). Among the patients without serious head injuries, the high but not the low SI group showed a higher risk of 24-h mortality than the normal SI group (Fig. 4B).

• 1. Sensitivity analysis: prehospital SI and 24-h mortality with altered cut- off values, additional adjustment for SI upon ED arrival as a covariate, and MI

Sensitivity analyses with the normal range of the SI set to 0.4-1.0 or 0.5-0.7, additional adjustment for the SI upon ED arrival, and MI for the missing values showed results similar to those of the primary analysis (Fig. S2-S4). In the analysis of the entire cohort with the MI, 20.4% of the patients had missing values, with the most common factors being respiratory rate (9.2%), transportation time (7.9%), and GCS (5.7%; Table 1).

1. Discussion

In this retrospective cohort study of 89,495 patients with a normal SI upon ED arrival recorded in the JTDB, we found an association between

prehospital SI abnormalities (SI <= 0.4 or >=0.9) and 24-h mortality (Fig. 2). We confirmed the robustness of the results using three sensitiv- ity analyses; alternative definitions of the normal SI, additional adjust- ment for the SI upon ED arrival, and analysis with MI showed results similar to those of the primary analysis (Fig. S2-S4). Invasive hemostatic interventions and blood transfusion within 24 h were performed more frequently in the high SI group than in the normal SI group, while head surgery was more frequent in the low SI group (Fig. 3). The sub- group analysis of patients with isolated serious head injuries showed that the low SI group was associated with a higher risk of 24-h mortality than the normal SI group, while the high SI group without serious head injuries had a higher risk of 24-h mortality (Fig. 4).

The two leading causes of trauma-related deaths were bleeding and neurological damage. When injuries in the body trunk or limbs cause Massive bleeding, blood pressure drops due to hypovolaemia. Compen- satory mechanisms work to maintain cardiac output and blood flow to vital organs [20]. The sympathetic nervous system is activated to in- crease the heart rate and, consequently, the cardiac output. This reaction increases the SI (high heart rate and low blood pressure), indicating an increased risk of death from haemorrhagic shock. Additionally, the sym- pathetic nervous system constricts the peripheral blood vessels to raise the blood pressure, resulting in normal SI values in some cases. In con- trast, a serious head injury with intracranial haemorrhage could in- crease intracranial pressure, resulting in bradycardia and high blood pressure (low SI), known as the Cushing reflex [18], potentially explaining the secondary analysis results, in which the low SI group was associated with a higher rate of head surgeries than the normal SI group (Fig. 3). These mechanisms may temporarily alter the vital signs and SI [21], sometimes normalizing the SI value [20], leading clinicians to misestimate the risk of death.

Fig. 3. Secondary Analysis: adjusted odds ratios of the prehospital shock index for each treatment based on the logistic regression analysis. Abbreviations: OR, odds ratio; CI, confidence interval.

Adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, and type of injury [Model 1].

“Invasive hemostatic interventions” comprised thoracoabdominal surgery, endoscopic surgery, surgical hemostasis, angiostomy, and transcatheter arterial embolisation. Graph A, B and C are from independent logistic regression models, see “Primary analysis” in Method section.

Previous studies have attempted to improve the Prognostic ability of the SI by considering the difference between the prehospital SI and ED arrival SI [15,28,29]. Patients with different prognoses may have been classified into the same group in these studies. Briefly, the prognosis in patients with a similar increase in SI but different prehospital SI values could be different. For example, the prognosis of patients with low prehospital SI and normal SI upon ED arrival might differ from that of patients with normal prehospital SI and high SI upon ED arrival. In clinical settings, physicians rush to treat patients with high SI upon ED arrival, regardless of their prehospital SI. Our study focused on pa- tients with a normal SI upon ED arrival as they are generally considered to have a good prognosis. The results allowed us to identify patients at high risk of death, which may require therapeutic interventions based on their prehospital SI.

The SI was used to predict blood loss, indicating the need for blood transfusions following trauma [7,17], and the risk of death [7,11,23]. In

our study, patients with a high prehospital SI had a higher mortality risk, as reported previously [17]; furthermore, they had a higher risk of undergoing an invasive hemostatic intervention or blood transfusion within 24 h than patients who had normal prehospital SI, possibly due to severe blood loss following Organ injuries (Fig. 3). Without a serious head injury, the 24-h mortality in the low SI group was rare, and there were no substantial differences between the low and normal SI groups (Fig. 4). Therefore, after excluding trauma patients with a serious head injury, a traditional single high SI cut-off value (e.g., 0.9) might be suffi- cient to predict mortality and the need for therapeutic interventions. However, patients in the low SI group required head surgery more fre- quently than those in the normal SI group. When trauma was confined to a serious head injury, mortality in the low but not high SI group was more frequent than in the normal SI group. This finding suggested that a low SI cut-off value should be set for trauma patients with isolated seri- ous head injuries.

Fig. 4. Subgroup analysis: adjusted odds ratios of the prehospital shock index for 24-h death with/without serious head injury based on the logistic regression analysis. Abbreviations: OR, odds ratio; CI, confidence interval.

Adjusted for age, sex, Glasgow Coma Scale, respiratory rate, year of emergency department arrival, transportation time, and type of injury [Model 1]. Graph A and B are from different populations; Graph A. is from the population with isolated serious head injury defined as a head Abbreviated Injury Scale (AIS) score >= 3 and AIS score < 3 for the other body parts, and Graph B is from the population without serious head injury defined as a head AIS score < 3 and AIS score >= 3 for the other body parts.

This study had several strengths. First, we used one of the largest multicenter trauma registries in the world, resulting in an adequately large sample. Second, we confirmed the robustness of the results asso- ciated the prehospital SI with 24-h mortality using sensitivity analyses. Finally, unlike previous studies, we excluded patients with normal SI who were in shock states upon ED arrival. This eligibility criterion could identify patients whose prognoses were uncertain and who needed further risk estimations in clinical settings.

This study also had several limitations. First, the extrapolation of our results to populations with less Serious injuries could be challenging be- cause this registry enrolled only trauma patients with serious injuries (AIS score >= 3). Second, it was unclear whether our results may be ap- plied to patients in countries where Penetrating injuries are more com- mon [30]. It is often difficult for physicians to determine whether patients require further examinations or interventions after a blunt injury with visible and hidden blood loss. Our findings may provide useful information for clinical decision-making in patients with Blunt injuries. Third, there were missing covariates and outcomes in this study. However, we found no apparent differences in the pa- tient characteristics among the eligible patients, those with missing covariate data, and those who were excluded due to missing out- comes (Table S2). Fourth, we did not have data on the use of drugs intimately associated with SI (e.g., vasoactive drugs such as beta blockers or calcium blockers). As a substitute, adjustment was per- formed using comorbidities, for which patients were likely to use such kinds of drugs. Last, this was an observational study; therefore, we cannot exclude the effect of unknown factors which may have affected the observed relationship.

1. Conclusion

Among the trauma patients with normal SI upon ED arrival (0.4 < SI < 0.9), abnormal prehospital SI (SI >= 0.9 or SI <= 0.4) was asso- ciated with higher 24-h mortality than normal prehospital SI. This study contributes to a more effective triage of trauma patients with normal SI upon ED arrival.

Funding

This research did not receive any grant from funding agencies in the public, commercial, or non-profit organizations.

Ethics approval

The ethics committee of Kyoto University approved this study (ap- proval number: R2601). This study was conducted in accordance with Japan’s ethical guidelines for medical and biological research involving human subjects.

Patient consent for publication

Not applicable.

Prior presentations

Annual Meeting of the Japanese Association for Acute Medicine, Gifu Japan, November 18, 2020.

CRediT authorship contribution statement

Yoshie Yamada: Writing – original draft, Visualization, Project ad- ministration, Methodology, Formal analysis, Conceptualization. Sayaka Shimizu: Writing – review & editing, Validation, Methodology, Formal analysis. Shungo Yamamoto: Writing – review & editing, Methodology, Conceptualization. Yoshinori Matsuoka: Writing – review & editing, Validation, Conceptualization. Yusuke Tsutusmi: Writing – review & editing, Resources, Conceptualization. Asuka Tsuchiya: Writing – review & editing, Resources, Conceptualization. Tsukasa Kamitani: Writing – review & editing, Resources, Conceptualization. Hajime Yamazaki: Writing – review & editing, Methodology, Conceptualiza- tion. Yusuke Ogawa: Writing – review & editing, Methodology, Concep- tualization. Shunichi Fukuhara: Writing – review & editing, Supervision, Conceptualization. Yosuke Yamamoto: Writing – review & editing, Supervision, Conceptualization.

Data availability

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

Declaration of Competing Interest

Y. Yamamoto has received consultancy fees from Nippon Shinyaku Co., Ltd, and personal fees from Sun Pharma, Asahi Kasei Pharma, TORAY, and Ono, outside the submitted work. The other authors state that they have no conflict of interest with the present work.

Acknowledgments

We would like to thank the members of the JTDB for the data of this research and Editage (www.editage.com) for English language editing.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2023.05.008.

References

1. World Health Organization. Injuries and Violence. https://www.who.int/news- room/fact-sheets/detail/injuries-and-violence; 2021. accessed October 28 2022.
2. Ministry of Health LaW. Fact Sheet of Monthly Demographic Statistics. https://www.

mhlw.go.jp/toukei/saikin/hw/jinkou/geppo/nengai19/index.html; 2019. [in

Japanese] (accessed October 28 2022).

1. Tokyo Fire Department. Current Activities of the Emergency Medical Service. https:// www.tfd.metro.tokyo.lg.jp/hp-kyuukanka/katudojitai/R01.html; 2020. [in Japanese] (accessed October 28 2022).
2. Shiraishi A, Otomo Y, Yoshikawa S, Morishita K, Roberts I, Matsui H. Derivation and validation of an easy-to-compute trauma score that improves prognostication of mortality or the Trauma Rating Index in age, Glasgow coma scale, respiratory rate and systolic blood pressure (TRIAGES) score. Crit Care. 2019;23(1):365.
3. Totten AM, Cheney TP, O’Neil ME, et al. Physiologic predictors of severe injury: systematic review. AHRQ Comparative Effectiveness Reviews; 2018 18-EHC008-EF.
4. Allgower M, Burri C. Schockindex. DMW. 1967;92(43):1947-50. [in German].
5. King RW, Plewa MC, Buderer NM, Knotts FB. Shock index as a marker for significant injury in trauma patients. Acad Emerg Med. 1996;3(11):1041-5.
6. Vandromme MJ, Griffin RL, Kerby JD, McGwin Jr G, Rue 3rd LW, Weinberg JA. Iden- tifying risk for massive transfusion in the relatively normotensive patient: utility of the prehospital shock index. J Trauma. 2011;70(2):384-8.
7. Birkhahn RH, Gaeta TJ, Terry D, Bove JJ, Tloczkowski J. Shock index in diagnosing early acute hypovolemia. Am J Emerg Med. 2005;23(3):323-6.
8. Zarzaur BL, Croce MA, Magnotti LJ, Fabian TC. Identifying life-threatening shock in the older injured patient: an analysis of the National Trauma Data Bank. J Trauma. 2010;68(5):1134-8.
9. Haider AA, Azim A, Rhee P, et al. Substituting systolic blood pressure with shock index in the National trauma triage Protocol. J Trauma Acute Care Surg. 2016;81 (6):1136-41.
10. Bruijns SR, Guly HR, Bouamra O, Lecky F, Lee WA. The value of traditional vital signs, shock index, and age-based markers in predicting trauma mortality. J Trauma Acute Care Surg. 2013;74(6):1432-7.
11. Wu SC, Rau CS, Kuo SCH, Chien PC, Hsieh HY, Hsieh CH. The reverse shock index multiplied by Glasgow coma scale score (rSIG) and prediction of mortality outcome in adult trauma patients: a cross-sectional analysis based on registered trauma data. Int J Environ Res Public Health. 2018;15(11):2346.
12. Tohira H, Jacobs I, Mountain D, Gibson N, Yeo A. Systematic review of predictive per- formance of injury severity scoring tools. Scand J Trauma Resusc Emerg Med. 2012; 20:63.
13. Cannon CM, Braxton CC, Kling-Smith M, Mahnken JD, Carlton E, Moncure M. Utility of the shock index in predicting mortality in traumatically injured patients. J Trauma. 2009;67(6):1426-30.
14. McNab A, Burns B, Bhullar I, Chesire D, Kerwin A. A prehospital shock index for trauma correlates with measures of hospital resource use and mortality. Surgery. 2012;152(3):473-6.
15. Olaussen A, Blackburn T, Mitra B, Fitzgerald M. Review article: shock index for pre- diction of critical bleeding post-trauma: a systematic review. Emerg Med Australas. 2014;26(3):223-8.
16. Yumoto T, Mitsuhashi T, Yamakawa Y, et al. Impact of Cushing’s sign in the prehos- pital setting on predicting the need for immediate neurosurgical intervention in trauma patients: a nationwide retrospective observational study. Scand J Trauma Resusc Emerg Med. 2016;24(1):147.
17. Odom SR, Howell MD, Gupta A, Silva G, Cook CH, Talmor D. Extremes of shock index predicts death in trauma patients. J Emerg Trauma Shock. 2016;9(3):103-6.
18. Abou-Khalil B, Scalea TM, Trooskin SZ, Henry SM, Hitchcock R. Hemodynamic re- sponses to shock in young trauma patients: need for invasive monitoring. Crit Care Med. 1994;22(4):633-9.
19. Chen L, Reisner AT, Gribok A, Reifman J. Exploration of prehospital vital sign trends for the prediction of trauma outcomes. Prehosp Emerg Care. 2009;13(3):286-94.
20. Oyetunji TA, Chang DC, Crompton JG, et al. Redefining hypotension in the elderly: normotension is not reassuring. Arch Surg. 2011;146(7):865-9.
21. Singh A, Ali S, Agarwal A, Srivastava RN. Correlation of shock index and modified shock index with the outcome of adult trauma patients: a prospective study of 9860 patients. N Am J Med Sci. 2014;6(9):450-2.
22. Kusumoto FM, Schoenfeld MH, Barrett C, et al. 2018 ACC/AHA/HRS guideline on the evaluation and management of patients with bradycardia and cardiac conduction delay: executive summary: a report of the American College of Cardiology/ American Heart Association task force on clinical practice guidelines, and the Heart Rhythm Society. J Am Coll Cardiol. 2019;74(7):932-87.
23. The Japanese Association for the Surgery of Trauma. Japan Trauma Data Bank. https://www.jtcr-jatec.org/traumabank/index.htm; 2017. [in Japanese] (accessed October 28 2022).
24. Ministry of Health LaW. Evaluation of the Quality of Critical Care Medical Center. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000188907.html; 2017. [in

Japanese] (accessed October 28 2022).

1. Jehan F, Con J, McIntyre M, et al. Pre-hospital shock index correlates with transfu- sion, resource utilization and mortality; the role of patient first vitals. Am J Surg. 2019;218(6):1169-74.
2. Schellenberg M, Strumwasser A, Grabo D, et al. Delta shock index in the emergency department predicts mortality and need for blood transfusion in trauma patients. Am Surg. 2017;83(10):1059-62.
3. Joseph B, Haider A, Ibraheem K, et al. Revitalizing vital signs: the role of delta shock index. Shock. 2016;46:50-4.
4. Soreide K. Epidemiology of major trauma. Br J Surg. 2009;96(7):697-8.