This retrospective analysis of traumatized patients’ data available in the HEMS and the University Hospital registers revealed that the shock index (AUC 0.88), reversed shock index (AUC 0.88) and pulse pressure (AUC 0.86) are suitable scoring systems for the identification of high-risk patients requiring PHBT.
The concept of pre-hospital transfusion administration is based on experience from the Vietnam war and has been optimized during further conflicts. The decrease in long-term mortality has been confirmed compared to patients who did not receive PHBT or received a late transfusion [17]. PHBT has also been introduced into some European Emergency Medical Services (EMS) procedures, respectively, the HEMS. The Czech Republic (resp. the HEMS of the Hradec Kralove Region) is one of eleven EMS/HEMS in European countries where this practice has been applied [18]. The pre-hospital teams face the challenge of identifying patients for PHBT. The identification criteria between countries differ. Based on the survey conducted among European countries, the main identifier for PHBT is major trauma, shock, and prolonged entrapment in unstable patients [18]. According to the review, which involved 22 PHBT studies by Shand et al., the physiological criterion most frequently assessed is systolic blood pressure (SBP) (varied between < 70 and < 90 mmHg), tachycardia (varied between > 108 and > 130/min) or no radial pulse. The mechanism of injury (penetrating injury or amputation above the knee/elbow) was included in 5 studies as an indication criterion. In 4 studies, the criteria for PHBT were not identified and in six studies, the criteria for PHBT were not quantified [19]. Rijnhout et al. in the review of 32 studies also recorded the lactate test (> 5), haemoglobin value (< 7 g/dl), the estimate of blood loss (> 500 ml), capillary return (> 2 s) and clinical gestalt as indication criteria [6].
The link to massive transfusion activation in the in-hospital phase has been proved in the scoring systems. Many of them, except physiological function values, use the results of lab examinations, imaging, or time-consuming numerical processes that are difficult to use in the pre-hospital setting, especially in EMS/HEMS that do not use these complements [13, 14]. Colleagues from Spain conducted a retrospective analysis in 2019 on the topic of prehospital prediction of massive bleeding using scoring systems. The best results were achieved by a score that had at least 6 variables, including BE, serum Hb, or FAST performed during transport to the ED: Emergency Transfusion Score (ETS; AUC 0.85), Trauma Associated Severe Haemorrhage (TASH) and the Prince of Wales Hospital score (AUC 0.82) [20]. Since many pre-hospital systems do not use imaging and laboratory complement, we focused on scoring systems using readily available vital signs values.
The shock index, defined as heart rate/SBP, is a better predictor of trauma outcome than vital signs alone [10, 11]. SI can be used, according to the European Guideline on the management of major bleeding and coagulopathy after trauma, to assess the seriousness of hypovolemic shock [21]. Vandrome and colleagues have proved the increasing risk of MT requirement if the SI value rises over 0.9 even in relatively normotensioned patients [22]. El-Menyar et al. defined the optimal cut-off point as 0.81 for predicting MT in trauma patients (sensitivity 85%, specificity 64%) [23]. Some clinicians prefer the characteristic of unstable hemodynamic status as a lower SBP than HR and not a higher HR than SBP. Kimura et al. considered reversed shock index (rSI – ratio SBP to HR) and GCS together (rSI–G: rSI multiplied value of GCS) and proved that rSI–G was a better predictor of in-hospital mortality and 24-h blood transfusion than SI [24]. Another research group from South Korea found that rSI–G is a strong predictor of massive transfusion initiation in the ED with median rSI–G 6.47 (IQR 25–75: 3.80–12.24) [25]. The next derivative of SI, the age-related shock index (AGE–SI), was introduced to improve the accuracy of SI. Rau et al. found the AGE–SI cut-off point 36.95 to predict the requirement for MT (AUC 0.627) [26]. Pulse pressure (PP) is defined as the difference between diastolic and systolic blood pressure. The value of PP narrows in bleeding patients as a response to decreased intravascular volume. The main purpose of the research led by Priestley et al. was to determine whether a narrowed PP in a normotensive patient (SBP ≥ 90 mmHg) is an independent predictor of bleeding. They found that the mean PP was significantly lower in the group with acute haemorrhage (AH) compared to the group without AH (39 ± 18 mmHg vs 53 ± 19 mmHg, p ˂ 0.0001). The analysis identified a significantly higher risk of AH at the PP cut-off of 55 mmHg (p = 0.005 AUC 0.955) in patients 61 years or older vs 40 mmHg (p < 0.0001, AUC 0.940) in patients from 16 to 60 years [27]. The median age value in our analysis was 44 (IQR 25–75: 24–60) and thus we consider the PP cut-off of 40 mmHg as referenced (sensitivity 75%, specificity 76%). The MGAP system (mechanism of injury, GCS, age, and systolic blood pressure) was introduced in France in 2010 for physician-staffed EMS crews and was originally defined for the prediction of in-hospital mortality. Based on the acquired values, three risk groups were defined: low (23–29 points), intermediate (18–22 points) and high risk (< 18 points). In the derivation cohort, the mortality was 2.8%, 15% and 48%, respectively [28]. The HEMS crew in the Czech Republic is staffed with a physician, therefore, this system was included in our analysis in order to verify its applicability in the pre-hospital phase. In the prediction of the patient's PHBT requirement, there was the lowest AUC value (AUC 0.66; 95% CI 0.57–0,76; cut-off 24.5) and the suitability of MGAP for PHBT was not demonstrated.
To the best of our knowledge, the comparison of the above-mentioned scoring systems has not been carried out in association with the initiation of PHBT.
The parameters in the pre-hospital phase show dynamics and variability, e.g., in connection to age and make the final decision difficult. Thus, whether a unified algorithm for PHBT is possible to develop arises. Nevertheless, the use of easily calculated SI and PP scoring systems allow the range of available criteria for the PHBT to be extended, the decision-making process to be optimized and minimize the risk of unnecessary administration (of an expensive and rare commodity) or, on the contrary, the miscalculation of patients who can profit from PHBT administration.
Limitations
Our study has several limitations. Firstly, it was a single-center study, thus these results are difficult to generalize in clinical practice. Secondly, it was a retrospective study with some patients missing data. The evaluated period was 30 months and involved a relatively small cohort of patients to whom PHBT was administered. PHBT was established at our workplace in the middle of 2018 and is considered a fairly new method, and the results presented reveal our first experience. Thirdly, we did not carry out stratification of patients according to blunt or penetrating injuries. We also did not differentiate the subgroups of patients with head trauma. These patients with serious head trauma and clinical manifestation of an increase in SBP and a decrease in heart rate could have an impact on score calculation. We did not perform further analysis in relation to emergency surgery or angioembolization, ICU length of stay or mortality. Given that the main objective was to identify patients who required PHBT prior to a hospital-administrated massive transfusion, we did not perform a detailed analysis of patients from the PHBT + /MT − subgroup (8 patients). Some of these patients may have benefited from one or two units of blood products administered in the pre-hospital phase and then no longer required transfusions after reaching the hospital. This may have limited our ability to identify predictors for patients who may benefit from just a prehospital blood transfusion. At the same time, we also did not consider the age of the patients, when a higher threshold for the physiological value of SBP can be assumed in higher age categories. We also did not consider comorbidities with associated medications (mainly antiarrhythmics) that could influence the predictive value of the indexes.