Challenging convention: a secondary data analysis of circulation-first versus airway-first triage in trauma patients receiving blood transfusions Authors: Jacob Metheny, Gena V Topper, Bedda L Rosario-Rivera, Eva Zadorozny, Jared Plumb, Aara Sheth, Patrick Morris, Stephen R Wisniewski, Jason L Sperry, Francis X Guyette, Tanya Egodage Published: 2026 ABSTRACT AbstractBackgroundTrauma guidelines have historically recommended intubation prior to transfusion, yet intubation and mechanical ventilation can exacerbate hemodynamic instability. We hypothesized that injured patients resuscitated with a circulation-first approach would require fewer blood transfusions and would have lower mortality than those managed with an airway-first approach.MethodsThis secondary data analysis of the Linking Investigations in Trauma and Emergency Services (LITES) database compared trauma patients in hemorrhagic shock who received blood products and intubation within 30 minutes of each other. ‘ABC’ patients were intubated first, whereas ‘CAB’ patients received blood first. A subgroup analysis compared patients with penetrating injuries only. Outcomes included transfusion volume, whole blood (WB) utilization, mortality, and hospital length of stay (HLOS).Results35 CAB and 49 ABC patients were included. The ABC group suffered more blunt injuries (73.5% vs 42.9%, p=0.02) and were more likely to receive WB in the prehospital setting (75.5% vs 48.6%, p=0.02). Transfusion requirements were significantly higher in the CAB group (7195 mL vs 6090 mL, p=0.01), although this was not the case after removal of outliers. In the penetrating cohort (17 CAB, 11 ABC), CAB patients were less likely to be transferred directly from the scene (70.6% vs 72.7%), yet were more likely to receive WB in the prehospital setting (64.7% vs 27.3%). There were no differences with respect to mortality or HLOS.ConclusionWhen comparing a circulation-first to an airway-first approach to triage in a unique, severely injured cohort of trauma patients, there was no difference detected in transfusion requirements or mortality. Centers participating in LITES more frequently used CAB over ABC for patients with penetrating trauma. The sample size and characteristics of this cohort limit its generalizability and highlight the need for prospective multicenter trials to guide the triage and management of hemorrhagic shock.Level of evidenceIII, prognostic/epidemiological. WHAT IS ALREADY KNOWN ON THIS TOPIC WHAT THIS STUDY ADDS This is the first study of our knowledge to consider CAB versus ABC in a cohort with information on prehospital whole blood administration. This is the first study of our knowledge to consider CAB versus ABC in a cohort with information on prehospital whole blood administration. The findings demonstrate a shift within trauma systems and prehospital providers toward prioritizing circulation over airway. The findings demonstrate a shift within trauma systems and prehospital providers toward prioritizing circulation over airway. Our results show a numerical mortality benefit that failed to meet statistical significance, potentially due to our small sample in those triaged with CAB over ABC in patients with life-threatening hemorrhage, emphasizing the need for more research to characterize the appropriate use of each triage algorithm in different trauma patient populations. Our results show a numerical mortality benefit that failed to meet statistical significance, potentially due to our small sample in those triaged with CAB over ABC in patients with life-threatening hemorrhage, emphasizing the need for more research to characterize the appropriate use of each triage algorithm in different trauma patient populations. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY This study prompts further evaluation of prehospital provider decision-making in initiating CAB versus ABC triage algorithms, and in a larger study population with a wider range of injury severity. This study prompts further evaluation of prehospital provider decision-making in initiating CAB versus ABC triage algorithms, and in a larger study population with a wider range of injury severity. Our results support the current trend in clinical practice toward a circulation-first approach in trauma patients with life-threatening hemorrhage; however, these practices have preceded policy adjustments. Our results support the current trend in clinical practice toward a circulation-first approach in trauma patients with life-threatening hemorrhage; however, these practices have preceded policy adjustments. Background Injury is the leading cause of death among adults under 45 years old, and the fourth leading cause of mortality worldwide.1 2 In the USA, hemorrhage remains the most common cause of preventable death after trauma.3–7 Data from civilian and military studies suggest that up to 80% of preventable trauma death results from uncontrolled hemorrhage.7–9 The military has adopted the ‘MARCH’ algorithm (Massive hemorrhage, Airway, Respiration, Circulation, and Hypothermia) which prioritizes control of massive hemorrhage initially.10 Conversely, civilian physicians have traditionally used the Advanced Trauma Life Support (ATLS)-supported guideline, which, until recently, promoted the ‘ABCDE’ approach, prioritizing airway first (A), followed by breathing (B), circulation (C), disability (D), and exposure (E).11 12 Whereas the airway, breathing, circulation (ABC) sequence provides a standardized, pragmatic, and widely adopted framework, intubation in the setting of ongoing hemorrhage may be detrimental. Intubation and mechanical ventilation can decrease preload secondary to increased intrathoracic pressure,13 and anesthetic induction agents can cause vasodilation and further reduce preload, potentially worsening pre-existing hypotension and risking fatal outcomes.12 Even analgesics may result in profound hemodynamic changes, blunting a patient’s compensatory vasoconstriction.14 Consequently, there is growing interest in the ‘CAB’ (circulation, airway, breathing) approach, in which resuscitation precedes airway management in injured patients to mitigate these hemodynamic changes and optimize central perfusion, consistent with the military’s MARCH algorithm, which prioritizes addressing life-threatening hemorrhage first.10 A 2025 meta-analysis by Ferrada et al found that in trauma patients with exsanguinating injuries, the CAB approach led to a decrease in mortality when compared with the ABC approach.15 It is important to recognize that triage sequences are frequently initiated by pre-hospital providers, with few emergency medical service (EMS) systems possessing the infrastructure to support prehospital whole blood (WB) administration. There has been a recent effort to minimize prehospital transport time and medical interventions in the field in those with penetrating injuries more commonly, whereas patients with blunt injuries typically receive additional care at the scene and slower transport, even when injury severity is comparable.16 17 Additionally, penetrating injuries appear to cause less global inflammatory dysregulation and subsequent coagulopathy than blunt injuries, but may demonstrate more severe upfront blood loss, and therefore the physiologic response to either insult and subsequent blood transfusion is different.18–20 As such, injury mechanism remains an important consideration in evaluation of triage sequences as CAB and MARCH may already be a preferred approach for those with penetrating injuries. This study aims to compare early transfusion requirements and clinical outcomes of patients treated with either CAB or ABC sequencing and presenting in extremis where rapid treatment decisions and initiation of care have been shown to affect long-term outcomes with an exploratory analysis of those injured by a penetrating mechanism.21 We hypothesize that patients resuscitated using the CAB approach would require fewer blood transfusions within the first 4 hours of admission and have improved mortality over those triaged with the ABC algorithm. Methods This is a secondary data analysis of a prospective, multicenter, observational cohort study conducted at seven American College of Surgeons (ACS)-verified level 1 trauma centers participating in the Linking Investigations in Trauma and Emergency Services network (LITES—www.litesnetwork.org) Task Order 2, Shock, Whole Blood and Assessment of TBI (SWAT) study.22 The LITES network is a collaborative research network comprising medical professionals, prehospital providers, and emergency medical services dedicated to conducting research on injury care and outcomes. The goal of the network is to inform clinical practice guidelines and update existing standards for managing traumatic injuries. The primary objective of the SWAT study was to compare patient-level outcomes across early low-titer group O whole blood and component resuscitation groups. The participating level 1 trauma centers for the SWAT study included Denver Health, University of Texas Health Science Center at Houston, University of Miami, Oregon Health and Science University, University of Pittsburgh, University of Pennsylvania, and University of Texas Southwestern. The original cohort study enrolled patients from April 1, 2018 to September 1, 2021, and included patients at risk of massive transfusion defined by the presence of at least two high-risk features: systolic blood pressure (SBP) ≤90 mmHg, penetrating or blunt injury, positive Focused Assessment with Sonography in Trauma (FAST), heart rate (HR) ≥120 beats/min, receipt of a blood transfusion, or operative or interventional radiologic hemorrhage control within 60 minutes of arrival in the operating room or interventional radiology suite. Institutional Review Board (IRB) approval was obtained from the University of Pittsburgh and the Office of Human Research Oversight for initial data collection. In the secondary data analysis, we included patients who received both intubation and blood products within 30 minutes of arrival at the LITES participating trauma center. We excluded those with isolated traumatic brain injury (TBI), those who did not receive advanced airway management, or patients with transfusion and intubation times of zero minutes apart. Patients with missing data pertinent to the study were also excluded. All patients in the database meeting inclusion and exclusion criteria were included in the study. For the subgroup analysis of patients injured by a penetrating mechanism, patients with either blunt mechanism or concomitant blunt and penetrating mechanism were excluded. This secondary analysis was considered exempt from IRB approval due to the deidentified nature of the database. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines were adhered to for this study (online supplemental file 1). Patients were categorized into either the CAB or ABC group based on the sequence of intubation as compared with transfusion. The primary outcome for this secondary analysis was 4-hour total transfusion volume. Secondary outcomes included mortality at 4 hours, 24 hours, and 28 days, hospital length of stay (HLOS), and intensive care unit length of stay (ICULOS), pulmonary complications including acute lung injury and acute respiratory distress syndrome (ARDS), and thromboembolic events including deep vein thrombosis (DVT) and pulmonary embolism (PE). Demographic information including age, sex, and race was collected. Injury characteristics including mechanism (blunt vs penetrating), Injury Severity Score (ISS), and Abbreviated Injury Scale (AIS) for all body regions were also collected. Additionally, initial Glasgow Coma Scale (GCS), SBP, HR (in beats per minute (BPM)), and Shock Index (SI) were noted. Transfusion data were also collected, with particular attention to whether patients received WB and/or component therapy. Descriptive statistics for baseline and clinical characteristics were summarized by group (CAB vs ABC) as frequencies and percentages for categorical data or as the mean±SD or median and IQR for normally or non-normally distributed continuous data, as appropriate. Examination of normal distribution assumption for continuous data was determined using histograms. χ2 or Fisher’s exact test, as appropriate, were used to compare the frequency distribution of the categorical variables between the two groups. Two-sample t-test or Wilcoxon rank-sum tests were performed to determine the differences between the two groups for continuous data. An unadjusted logistic regression model was applied to assess the strength of the association between groups and the binary outcome of interest (eg, 4-hour mortality). For the primary outcome, an unadjusted linear regression model was used to determine the association between groups and the continuous outcome (eg, 4-hour blood volume). A sensitivity analysis was performed excluding five observations defined as potential outliers (observations that exceeded 1.5 times the IQR). For adjusted analyses, propensity score analysis was used to reduce the potential impact of confounding and selection bias. Propensity scores were estimated using a logistic regression model and considered characteristics defined a priori based on previous research. The baseline characteristics included in the primary propensity score model were age, sex, ISS, and injury type. A sensitivity propensity score analysis was performed considering characteristics chosen based on statistical analysis by examining baseline characteristics, which revealed imbalance between the groups and their characteristics associated with the outcome (figure 1). The baseline characteristics included in the sensitivity propensity score model were abdominal AIS and systolic blood pressure in the emergency department (ED). Stabilized inverse probability of treatment weighting (ie, stabilized IPTW weights) were computed from the propensity scores . Common support was assessed visually by box plots. The range of propensity scores overlapped substantially between the two groups for both the primary and sensitivity propensity scores. For baseline characteristic balance, we assessed standardized mean difference (SMD) before and after applying the stabilized IPTW weights. A SMD <0.20 was considered a small difference and that the two groups were balanced in baseline characteristics. For the primary outcome, 4-hour transfusion volume, sex, ISS, and injury type were imbalanced (SMD >0.20) after applying stabilized IPTW weights and were included in the adjusted model. A stabilized IPTW-adjusted linear model was used to determine whether a patient’s group was an independent predictor of the continuous primary outcome after controlling for potential confounders. These fitted regression models used survey regression methods to account for the weighting (ie, stabilized IPTW weights) when estimating the SE. Stabilized IPTW-adjusted models were also fitted with weight trimming at the 1st and 99th percentiles as a sensitivity analysis. This sensitivity analysis excluded two observations (one observation below the 1st percentile and one observation above the 99th percentile). Results for unadjusted and adjusted models were summarized as mean difference and the corresponding 95% CI. Propensity scores overlap for (A) clinical confounders and (B) statistical confounders. Because of limited sample size and number of events, analyses for secondary outcomes focused on unadjusted analysis only. For secondary outcomes, an unadjusted Cox regression model was used to determine the association between groups and time to event outcome (eg, HLOS). Kaplan-Meier survival curves and log-rank tests were used to assess HLOS and ICULOS probabilities between the two groups. Results for unadjusted models were summarized as OR or HR, and the corresponding 95% CI. All statistical tests were two-sided, and the significance level was set to p=0.05. Analyses were repeated comparing only CAB and ABC patients injured by penetrating mechanism specifically, with patients injured by both blunt and penetrating mechanisms excluded from this analysis. The penetrating injury subgroup analysis was exploratory, given the small sample size. All analyses were conducted using SAS statistical software (V.9.4, SAS Institute, Cary, NC). Results Among 1051 patients who received blood transfusions, 115 were intubated, and 91 of those intubations occurred within 30 minutes of arrival. Of these, 84 had a transfusion/intubation gap between 1 and 30 minutes and were included in this analysis; 35 were transfused first, making up the CAB group, and 49 were intubated first, comprising the ABC group (figure 2). Baseline demographics were similar between groups. A total of 28 patients were injured by a penetrating mechanism, with CAB patients being more likely to have penetrating injuries (48.6% vs 22.5%, p=0.02). No differences were observed in initial GCS, SBP, HR, or ISS between groups (table 1). CONSORT (Consolidated Standards of Reporting Trials) diagram of included patients in this study. Table 1Demographic and injury characteristics by resuscitation sequence, excluding subjects with isolated TBICharacteristicsCAB(n=35)ABC(n=49)P valueAge, y, median (IQR)29.0 (39.0–49.0)26.5 (35.5–54.5)0.92Sex, m, n (%)22 (62.9)36 (73.4)0.30Race, n (%)0.87 White22 (65.9)30 (61.2) Black5 (14.3)9 (18.4) Other8 (22.9)10 (20.4)Injury0.02 Blunt15 (42.9)36 (73.5) Penetrating17 (48.6)11 (22.5) Both3 (8.57)2 (4.08)Transfer origin, n (%)0.06 Scene (AEMS or GEMS)25 (71.4)43 (87.8) Outside emergency department (referring hospital)10 (28.6)6 (12.2)ISS, median (IQR)27.0 (22.0–38.0)34.0 (24.0–45.0)0.20 >15, n (%)32 (91.4)45 (91.8)0.95Traumatic brain injury (CT diagnosed), n (%)12 (34.3)23 (65.7)0.25Head AIS, median (IQR)0.0 (0.0–3.0)3.0 (0.0–4.0)0.15 >2, n (%)13 (37.1)25 (51.0)0.21Chest AIS, median (IQR)3.0 (2.0–4.0)3.0 (2.0–4.0)0.99 >2, n (%)25 (71.4)32 (65.3)0.55Abdomen AIS, median (IQR)3.0 (0.0–4.0)3.0 (0.0–4.0)0.99 >2, n (%)19 (54.3)26 (53.1)0.91Extremity AIS, median (IQR)2.0 (0.0–3.0)2.0 (1.0–3.0)0.13 >2, n (%)13 (37.1)24 (49.0)0.28Glasgow Coma Scale, median (IQR)3.0 (3.0–3.0)3.0 (3.0–3.0)0.34 <9, n (%)30 (100)46 (97.8)0.61Systolic blood pressure in ED, mm Hg, median (IQR)92.0 (80.0–125)82.0 (60.0–109)0.31 <90 mm Hg, n (%)10 (47.6)16 (55.2)0.60Heart rate, bpm, median (IQR)119 (100–137)125 (99–147)0.86 >100 bpm, n (%)21 (72.4)32 (74.4)0.85Shock index, median (IQR)1.1 (0.9–1.6)1.3 (1.2–2.0)0.17Whole blood in the receiving hospital (remainder had component blood), n (%)26 (74.3)36 (73.5)0.93Whole blood before the receiving hospital (remainder had component blood), n (%)17 (48.6)37 (75.5)0.02ABC, airway, breathing, circulation; AEMS, air emergency medical services; AIS, Abbreviated Injury Scale; CAB, circulation, airway, breathing; CT, computer tomography; ED, emergency department; GEMS, ground emergency medical services; ISS, Injury Severity Score; m, male; n, number; TBI, traumatic brain injury; y, years. Demographic and injury characteristics by resuscitation sequence, excluding subjects with isolated TBI ABC, airway, breathing, circulation; AEMS, air emergency medical services; AIS, Abbreviated Injury Scale; CAB, circulation, airway, breathing; CT, computer tomography; ED, emergency department; GEMS, ground emergency medical services; ISS, Injury Severity Score; m, male; n, number; TBI, traumatic brain injury; y, years. Transfusion requirements were significantly higher in the CAB group (7195 mL vs 6090 mL, adjusted mean difference=5129 mL, p=0.01). However, when outliers were excluded, there was no difference in volume of blood transfused between groups (5780 mL vs 5880 mL, adjusted mean difference 1998 mL, p=0.13) (table 2). The results were consistent on sensitivity analyses. Table 2Primary outcome of interest by resuscitation sequence, excluding patients with isolated TBIOutcomesMedian (IQR)UnadjustedAdjusted*CAB (n=35)ABC (n=49)Effect95% CIP valueEffect95% CIP value4-hour blood volume, mL7195 (3605– 13 000)6090 (3775–9265)−3995.8−7742.7 to −248.90.04−5,128.7−9084.4 to −1173.00.014-hour blood volume, mL†5780 (3340 – 9725)5880 (3600–9132.5)−1501−3903.8 to 901.90.22−1997.8−4580.2 to 584.50.13*Stabilized Inverse probability of treatment weighting (IPTW) weighted and imbalance covariate (sex, ISS, and injury type) adjusted linear regression model. Propensity score analysis based on confounders defined a priori and included age, sex, ISS, and injury type.†Models excluded five potential outliers (ie, observations that exceeded 1.5*IQR). ABC group was used as reference group.ABC, airway, breathing, circulation; CAB, circulation, airway, breathing; effect, regression coefficient; ISS, Injury Severity Score; mL, milliliter; TBI, traumatic brain injury. Primary outcome of interest by resuscitation sequence, excluding patients with isolated TBI Stabilized Inverse probability of treatment weighting (IPTW) weighted and imbalance covariate (sex, ISS, and injury type) adjusted linear regression model. Propensity score analysis based on confounders defined a priori and included age, sex, ISS, and injury type. Models excluded five potential outliers (ie, observations that exceeded 1.5*IQR). ABC group was used as reference group. ABC, airway, breathing, circulation; CAB, circulation, airway, breathing; effect, regression coefficient; ISS, Injury Severity Score; mL, milliliter; TBI, traumatic brain injury. Mortality at 4 hours, 24 hours, and 28 days did not differ between the CAB group as compared with ABC in unadjusted models. Similarly, ICULOS and complication rates were similar between groups (table 3). The median HLOS for CAB and ABC groups were 18 (IQR 7–42) days and 14 (IQR 7–31) days, respectively. Kaplan-Meier survival curves showed no differences in HLOS between the two groups (log-rank χ2 test=0.051, p=0.82; online supplemental figure 1). Table 3Secondary outcomes by resuscitation sequence, excluding patients with isolated TBIOutcomesMedian (IQR)or frequency (%)UnadjustedCAB (n=35)ABC (n=49)OR/HR95% CIP value4-hour mortality, n (%)2 (6.06)5 (10.6)0.540.09 to 3.130.4924-hour mortality, n (%)4 (12.1)6 (12.8)0.940.24 to 3.780.9328-day mortality, n (%)6 (17.1)14 (28.6)0.520.17 to 1.560.24ICU length of stay, d7.5 (3.0–28.0)8.0 (3.0–15.0)−0.96−7.97 to 6.040.78Hospital length of stay, d18 (7.0–42.0)14 (7.0–31.0)1.060.63 to 1.800.82Lung-related events, n (%)6 (17.1)6 (12.2)0.670.19 to 2.380.54Acute lung injury, n (%)4 (11.4)4 (8.16)0.690.15 to 3.080.62ARDS, n (%)2 (5.71)4 (8.16)0.680.11 to 4.140.67Thromboembolic events, n (%)6 (17.1)8 (16.3)0.940.27 to 3.100.92DVT, n (%)3 (8.57)4 (8.16)0.950.19 to 4.730.95PE, n (%)3 (8.57)5 (10.2)0.820.18 to 3.860.80ABC, airway, breathing, circulation; ARDS, acute respiratory distress syndrome; CAB, circulation, airway, breathing; d, days; DVT, deep vein thrombosis; ICU, intensive care unit; n, number; PE, pulmonary embolism; TBI, traumatic brain injury. Secondary outcomes by resuscitation sequence, excluding patients with isolated TBI ABC, airway, breathing, circulation; ARDS, acute respiratory distress syndrome; CAB, circulation, airway, breathing; d, days; DVT, deep vein thrombosis; ICU, intensive care unit; n, number; PE, pulmonary embolism; TBI, traumatic brain injury. Penetrating mechanism exploratory subgroup analysis There were 17 CAB patients and 11 ABC patients who were injured by a penetrating mechanism. Penetrating trauma patients were of similar age between the CAB and ABC groups, and both were predominantly male. The CAB group received similar 4-hour blood volumes and demonstrated similar 4-hour, 24-hour, and 28-day mortality. Patients treated with CAB were more likely to receive WB both in the prehospital setting and on arrival to the receiving hospital (table 4). 11 of 12 (91.7%) CAB patients transferred from the scene were transfused WB in the prehospital setting. Median head AIS was 0 in both groups with a greater percentage of patients in the CAB group having a head AIS >2, signifying significant head injury (17.7% vs 0%). Median admission GCS was 3 in both groups with all subjects in the penetrating group having a GCS <9 (table 5). Table 4Demographic and injury characteristics by resuscitation sequence in the subgroup of penetrating trauma patients, excluding subjects with isolated TBICharacteristicsCAB(n=17)ABC(n=11)P valueAge, y, median (IQR)33.0 (28.0–47.0)37.0 (25.0–46.0)0.85Sex, m, n (%)10 (58.8)9 (81.8)0.25Race, n (%)0.38 White7 (41.2)4 (36.4) Black5 (29.4)6 (54.6) Other5 (29.4)1 (9.09)Transfer origin, n (%)1.00 Scene (AEMS or GEMS)12 (70.6)8 (72.7) Outside emergency room (referring hospital)5 (29.4)3 (27.3)ISS, median (IQR)25.0 (16.0–33.0)26.0 (16.0–33.0)1.00 >15, n (%)14 (82.4)9 (81.8)1.00Traumatic brain injury (CT diagnosed), n (%)1 (5.88)0 (0)1.00Head AIS, median (IQR)0.0 (0.0–0.0)0.0 (0.0–0.0)0.17 >2, n (%)3 (17.7)0 (0)0.26Chest AIS, median (IQR)3.0 (0.0–4.0)3.0 (0.0–4.0)0.77 >2, n (%)10 (58.8)7 (63.6)1.00Abdomen AIS, median (IQR)3.0 (0.0–4.0)0.0 (0.0–4.0)0.32 >2, n (%)9 (52.9)5 (45.5)1.00Extremity AIS, median (IQR)0.0 (0.0–3.0)2.0 (0.0–3.0)0.03 >2, n (%)2 (11.8)4 (36.4)0.17Glasgow Coma Scale, median (IQR)3.0 (3.0–3.0)3.0 (3.0–3.0)0.45 <9, n (%)12 (100)9 (100)1.00Systolic blood pressure in ED, mm Hg, median (IQR)122.5 (83.0–134)109 (90.0–121)0.68 <90 mm Hg, n (%)4 (33.3)3 (42.9)1.00Heart rate, bpm, median (IQR)119.5 (109–137.5)125 (77–135)0.75 >100 bpm, n (%)10 (83.3)6 (66.7)0.61Shock Index, median (IQR)1.0 (0.9–1.5)1.2 (1.0–1.3)0.61Whole blood in the receiving hospital (remainder had component blood), n (%)12 (70.6)7 (63.6)1.00Whole blood before the receiving hospital (remainder had component blood), n (%)11 (64.7)3 (27.3)0.12ABC, airway, breathing, circulation; AEMS, air emergency medical services; AIS, Abbreviated Injury Scale; CAB, circulation, airway, breathing; CT, computer tomography; ED, emergency department; GEMS, ground emergency medical services; ISS, Injury Severity Score; m, male; n, number; TBI, traumatic brain injury; y, years. Demographic and injury characteristics by resuscitation sequence in the subgroup of penetrating trauma patients, excluding subjects with isolated TBI ABC, airway, breathing, circulation; AEMS, air emergency medical services; AIS, Abbreviated Injury Scale; CAB, circulation, airway, breathing; CT, computer tomography; ED, emergency department; GEMS, ground emergency medical services; ISS, Injury Severity Score; m, male; n, number; TBI, traumatic brain injury; y, years. Table 5Primary and secondary outcomes by resuscitation sequence in the subgroup of penetrating trauma patients, excluding subjects with isolated TBIOutcomeMedian (IQR) orfrequency (%)UnadjustedCAB (n=17)ABC (n=11)Effect/OR/HR95% CI4-hour blood volume, mL5780 (3340–10 155)4930 (3425–7985)5604−938.0 to 12 1454-hour mortality, n (%)1 (6.25)1 (9.09)1.500.07 to 34.824-hour mortality, n (%)1 (6.25)1 (9.09)1.500.07 to 34.828-day mortality, n (%)2 (11.8)1 (9.09)0.750.05 to 11.7ICU length of stay, d5.0 (2.5–27.0)4.0 (2.0–6.0)8.441.10 to 15.8Hospital length of stay, d13.0 (7.0–36.0)10.5 (7.0–14.0)2.670.97 to 7.33Lung-related events, n (%)3 (17.7)0 (0)–– Acute lung injury, n (%)1 (5.88)0 (0)–– ARDS, n (%)2 (11.8)0 (0)––Thromboembolic events, n (%)2 (11.8)1 (9.09)0.750.05 to 11.7 DVT, n (%)1 (5.88)0 (0)–– PE, n (%)1 (5.88)1 (9.09)1.600.07 to 36.6ABC group was the reference group.ABC, airway, breathing, circulation; ARDS, acute respiratory distress syndrome; CAB, circulation, airway, breathing; d, days; DVT, deep vein thrombosis; effect, regression coefficient; ICU, intensive care unit; mL, milliliter; n, number; PE, pulmonary embolism; TBI, traumatic brain injury. Primary and secondary outcomes by resuscitation sequence in the subgroup of penetrating trauma patients, excluding subjects with isolated TBI ABC group was the reference group. ABC, airway, breathing, circulation; ARDS, acute respiratory distress syndrome; CAB, circulation, airway, breathing; d, days; DVT, deep vein thrombosis; effect, regression coefficient; ICU, intensive care unit; mL, milliliter; n, number; PE, pulmonary embolism; TBI, traumatic brain injury. Discussion The traditional ATLS triage sequence was originally based on management of airway prior to hemorrhage.11 As our understanding of the physiologic changes associated with different traumatic injuries improves, processes to optimize outcomes for patients in hemorrhagic shock are being re-evaluated. In this secondary analysis of severely injured trauma patients requiring both early transfusion and intubation, resuscitation sequence was not associated with differences in adjusted transfusion requirements, mortality, or length of stay. The patients in this study were transfused 5.7 L, on average, of blood in the first 4 hours, median GCS on arrival was 3, median ISS was around 30, and yet, 24-hour mortality was only around 12%. This is a unique, severely injured cohort and these characteristics are important to be considered when interpreting the results and generalizability of these findings. CAB sequencing was found to be on the rise, was more frequently utilized in patients with penetrating trauma, and was strongly associated with prehospital whole blood administration. The newest version of ATLS released last summer now emphasizes the importance of addressing exsanguinating hemorrhage prior to other interventions.23 Our findings suggest that a CAB approach to resuscitation has already been adopted in some circumstances, particularly for patients with penetrating injuries, and reflecting a growing shift in some trauma systems toward a circulation-first resuscitation. Unfortunately, few studies have evaluated this theory. Ferrada et al published an algorithm for the management of civilian trauma patients in which hemorrhage control is addressed first.12 Per this algorithm, when hemorrhage was compressible, traditional mechanisms for control including direct pressure, packing, and tourniquet use with concomitant ongoing resuscitation demonstrated benefit.12 However, in the event of noncompressible hemorrhage, deferring advanced airway management in favor of expedited transfusion demonstrated improved outcomes. In a consensus statement, these authors thus recommended that airway management only occur once volume resuscitation is underway and surgical intervention is accessible at the time of intubation.15 Their meta-analysis further found that in trauma patients with exsanguinating injuries, the CAB approach led to a decrease in mortality when compared with the ABC approach.15 These findings are reflected in the 11th edition of ATLS, released recently.23 Whereas our study did not demonstrate significant differences in outcomes between the CAB and ABC groups, we identified numerical trends toward improved outcomes among CAB patients, including a trend toward lower 4-hour, 24-hour, and 28-day mortality, and shorter ICULOS; these were not found to be statistically significant. These results are consistent with an international study conducted via the American Association for the Surgery of Trauma (AAST) including 440 patients which also compared CAB versus ABC, and did not demonstrate a difference in their outcomes of interest (transfusion, lengths of stay, and mortality).24 The lack of significant difference between groups may likely be attributed to the small sample size of patients meeting inclusion criteria. Additionally, patients enrolled into the LITES-network SWAT database used for our analysis are unique; these patients were severely injured with a median ISS of 33. Furthermore, 81% were transported from the injury scene with a median shock index of 1.3, and a median SBP of 86 mm Hg. Included patients received a median of 6 liters of blood transfusions within the first 4 hours of arrival to a trauma center, again suggesting the profound degree of injury in this patient cohort. These metrics suggest that this patient population is among the most severely injured patients seen at participating trauma centers and is a sicker population than studied in previous trials.24–28 A recent meta-analysis evaluated six trials comparing ABC versus CAB, despite significant variability between studies, found that patients with a penetrating mechanism and low admission GCS would benefit from hemorrhage control first (ie, a CAB algorithm).15 However, the majority of the cohort suffered a blunt mechanism of injury. To contrast, in our study, patients with a penetrating mechanism of injury were more likely to be treated with the CAB algorithm, whereas ABC resuscitation was followed to a greater degree in patients experiencing blunt trauma. Notably, all patients with head AIS over 2 in the penetrating subgroup were treated with CAB, paralleling prior evidence from The Prehospital Air Medical Plasma (PAMPer) trial, which demonstrated an association between prehospital plasma administration and improved survival among patients with TBI.29 In our study, patients who received a transfusion before airway management arrived with higher SBP, lower HR, and lower shock index potentially from this transfusion. Although no mortality difference between groups was seen in our small sample size, higher SBP and lower HR on presentation are associated with lower mortality after TBI30; therefore, these physiologic differences after prehospital transfusion suggest that prehospital circulation-first resuscitation may confer a survival benefit detectable in a larger sample size, similar to the benefit observed in the PAMPer secondary analysis.29 These findings support that CAB sequencing may be advantageous in patients with concurrent hemorrhage and neurologic injury. The findings from the penetrating injury subgroup underscore the growing role of prehospital provider decision-making in determining resuscitation sequence. Patients with penetrating trauma were more likely to be transported directly from the scene to a LITES-participating trauma center, suggesting expedited transport with fewer prehospital interventions. This pattern may reflect evolving prehospital practice as blood product availability and hemorrhage-focused training, including programs such as Stop the Bleed, are becoming more widespread; over half of the patients in this overall study received transfusion prior to the receiving hospital and in the penetrating subgroup, over 90% of CAB patients transferred directly from the scene received prehospital WB, further demonstrating that prehospital providers’ decision-making is often the deciding factor between CAB and ABC. Alternatively, these decisions may be influenced by provider heuristics, as penetrating injuries are perceived as more immediately life-threatening than blunt mechanism injuries, a phenomenon well-described.31 32 Taken together, our findings suggest that prehospital providers in the LITES network may preferentially apply CAB or MARCH-based algorithms in penetrating trauma while maintaining an airway-first approach for blunt injury. Whether driven by protocol, perception, or experience, the impact of these prehospital decisions on downstream outcomes warrants focused investigation in larger, prospective studies. Additional limitations of our trial include the retrospective nature of this secondary analysis and the subsequent inability to determine causation. Patients were sorted into CAB or ABC by resuscitation sequence only determined by time stamps and procedures noted within a database; although still valuable data, this is a limited picture of initial patient evaluation and clinical decision-making. Despite the large primary database, few patients met criteria for inclusion in this study, introducing potential confounding factors and limiting the power of the trial. Further, the absence of data on crystalloid use and the influence of prehospital decision-making also presents challenges. The inclusion of transferred patients may have introduced additional unmeasured confounding, as assessment and resuscitation decisions made at referring institutions could not be fully accounted for and may have influenced subsequent intervention sequencing and clinical outcomes. Additionally, our population primarily suffered blunt mechanisms and were severely injured,33 perhaps limiting generalizability to all trauma patient cohorts. Utilization of this database involving LITES-participating centers confers a selection bias toward systems, which may have greater infrastructure that can support prehospital blood administration or early adoption of transfusion prior to airway intervention. However, our study contributes to the growing body of evidence suggesting feasibility and increasing utilization of CAB-based triage utilization. Conclusion Centers participating in LITES Task Order 2 (SWAT) more frequently used CAB over ABC for severely injured patients with penetrating trauma. Although no transfusion volume or mortality differences were demonstrated, CAB remains an area of interest and warrants further investigation. Prospective and multicenter trials are required to evaluate algorithms and refine guidelines for the triage and management of hemorrhagic shock. Supplementary material