Brief Report
Application of the shock index to the prediction of need for hemostasis intervention

https://doi.org/10.1016/j.ajem.2013.05.027Get rights and content

Abstract

Introduction

The traditional method to identify hemorrhage after trauma has been vital signs–based. More recent attempts have used mathematical prediction models, but these are limited by the need for additional data including a Focused Assessment with Sonography for Trauma exam, or an arterial blood gas. Shock Index (SI) is the mathematical relationship of the heart rate divided by the systolic blood pressure; the cutoff of > 0.9 has been associated with bleeding.

Methods

A total of 4292 trauma patients were identified in database over an 11 year period. Inclusion criteria included age > 16 years and initial presentation to our trauma center. Patients were excluded for incomplete data, traumatic brain injury, or transfer leaving 4277 patients for analysis. Patients were further subdivided by age, and by mechanism of injury (blunt versus penetrating). Finally, patients were divided into bleeding versus nonbleeding, and the SI formula was applied to their initial hospital vital signs.

Results

Across our dataset, using the standard SI cutoff of > 0.9 as the threshold for bleeding, the sensitivity is 54.5%, with a specificity of 93.6%. In the geriatric subanalysis, there was no difference for sensitivity between the age groups, but SI is more specific in the older patients. There was no difference in sensitivity using SI in blunt versus penetrating. Lowering the SI to ≥ 0.8 increases the sensitivity to 76.1%, with a specificity of 87.4%.

Conclusion

SI, at a lowered threshold of ≥ 0.8, can be used to identify trauma patients that will require intervention for hemostasis.

Introduction

Significant bleeding after trauma occurs only in a minority of patients [1], but represents a major cause of morbidity and mortality [2]. Early identification of significantly bleeding trauma patients traditionally has been a challenge [3] but would facilitate the application of aggressive therapies including angiography with embolization, emergent surgery, and early activation of a massive transfusion protocol. The expeditious and timely use of these therapies in the patient with significant bleeding has the potential to significantly impact outcomes.

The monitoring of traditional vital signs, including looking for tachycardia followed by hypotension as the indicator of blood loss, has been a primary method of identifying and treating patients in hemorrhagic shock [4]. In the last several years, there have also been attempts at mathematical prediction models of post traumatic bleeding such as the Assessment of Blood Consumption (ABC) Score [5], the McLaughlin Score [6], and the Trauma Assessment Severity of Hemorrhage score [7]. However, while they are based on vital signs, they also incorporate advanced modalities. These modalities include a Focused Assessment with Sonography for Trauma (FAST) scan in the case of the ABC Score, or a laboratory hemoglobin, and an arterial blood gas with the McLaughlin and Trauma Assessment Severity of Hemorrhage Score. The tests take precious time to perform and result, which limits their attractiveness for universal adoption in the trauma setting.

Shock Index (SI) is a relationship that can potentially be used for the early identification of significantly bleeding trauma patients [8], [9], [10]. It is based only on vital signs, namely the mathematical relationship of the heart rate divided by the systolic blood pressure. These vital signs are available in the field, and are repeated upon arrival to the emergency department (ED). The SI represents an initial modality that can be used to identify more critically injured patients at higher risk for significant bleeding [11]. The commonly used threshold is that a SI ≥ 0.9 is predictive of mortality [9]. The SI has also been applied to patients with septic shock [12].

As patients age, there are differences in vital signs, and these can be quite pronounced in the setting of hemorrhagic shock [13]. Maximum heart rate is known to decrease as patients get older [14]. While a systolic blood pressure less than 90 mmHg is considered hypotensive in younger patients after trauma, a measurement of less than 110 mmHg is considered hypotensive in the elderly [15]. This is secondary to the high prevalence of hypertension, and elderly patients can easily develop a “relative hypotension” that can go unrecognized unless age-adjusted normal values are used.

Our initial hypothesis is that the cutoff SI of 0.9 used with trauma patients may too high for the geriatric trauma subpopulation, which as explained above do not mount the same response in their vital signs to hemorrhage. These potential age differences have not been previously studied. A secondary question was if the SI cutoff should be different in blunt versus penetrating trauma, as the penetrating patients generally have a higher risk of bleeding.

Section snippets

Methods

Winthrop University Hospital is a tertiary care facility and a regional trauma center. Trauma data is collected by full time dedicated staff, and reported locally, statewide and nationally to the National Trauma Data Bank as well as the Trauma Quality Improvement Program databases. Data is collected in a retrospective fashion via chart review of all trauma patients at our hospital, and organized in the Trauma One database. This project was approved by our institution’s investigational review

Results

Table 1 presents the overall results as the cutoff SI is varied from 0.1 to 2.0 in increments of 0.1. With the commonly accepted SI ≥ 0.9 cutoff for the entire group of 4277 trauma patients, the sensitivity of the SI is 54.5%, with a specificity of 93.6%. In the 65 years old and older subpopulation of 2093 patients, using the SI ≥ 0.9 cutoff, the sensitivity drops to 41.2%, compared to a sensitivity of 57.5% in the younger than 65 years subgroup. The specificities at a cutoff SI of 0.9 are 95.7%

Discussion

The sensitivity and specificity of the ABC Score in the original validation study were 75% and 86%, respectively [3]. In comparison, the SI across our entire data set using the ≥ 0.9 threshold has a lower sensitivity (54.5%), but a higher specificity (93.6%). However, the McLaughlin Score had a sensitivity of 59.4%, and a specificity of 77.4% [4], and the SI, while marginally less sensitive, has a higher specificity.

While the traditional SI cutoff has been ≥ 0.9, lowering it to ≥ 0.8 increases the

Conclusion

This data set represents that largest analysis of the use of SI to predict bleeding in trauma patients. Our analysis suggests that the same ≥ 0.8 threshold should be used across all the adult age groups in both blunt and penetrating trauma patients. Lowering the SI threshold to ≥ 0.8, from the ≥ 0.9 suggested in previous research increases the sensitivity to bleeding, and more closely approximates the sensitivity and specificity thresholds of the ABC Score. From our analysis, the SI, used at a ≥ 

References (16)

There are more references available in the full text version of this article.

Cited by (56)

  • Prognostic Role of Shock Index in Traumatic Pelvic Fracture: A Retrospective Analysis

    2019, Journal of Surgical Research
    Citation Excerpt :

    It helps predicting early shock and correlates well with the central venous oxygen saturation and arterial lactate concentration in trauma patients.6,10-12 There are several studies that assessed the outcomes of hypovolemic shock, which necessitates blood transfusion based on different cutoff values of SI in general trauma.9,13-15 However, to date, we are not aware of studies that primarily evaluated the potential utility of SI in patients with pelvic fracture.

  • Clinical Practice Guideline: Massive Transfusion Scoring Systems

    2019, Journal of Emergency Nursing
    Citation Excerpt :

    The researchers concluded that the MSI and the ASI did not provide improved predictability, with the ASI impacted by comorbid factors including hypertension, diabetes, and coronary artery disease. DeMuor et al. (2013) found that while there were minimal differences in the sensitivity between age groups (< 65 years 57.5% and > 65 years 41.2%) with an SI ≥ 0.9, the specificity was 95.7% in the greater than 65-year-old population and 91.6% for those < 65, indicating fewer false positives. When the SI cutoff was decreased to 0.8, the sensitivity fell to 76.1% with a specificity of 87.4% for both populations, thus leading the researchers to conclude that decreasing the SI cut off to ≥ 0.8 was a simple system to implement as a CB indicator (DeMuro, Simmons, Jax, & Gianelli, 2013).

View all citing articles on Scopus
View full text