Nationwide estimates of potential lives saved with prehospital blood transfusions

Abstract

Introduction

Prehospital blood transfusions result in a significant reduction in mortality risk for injured patients in hemorrhagic shock; however, prehospital blood transfusions have not been widely implemented across the United States. Thus, a paucity of data surrounding the impact of achieving near‐complete population‐level access to this resource exists. We aimed to determine the number of lives that could potentially have been saved among injured patients in hemorrhagic shock between 2020 and 2023 had prehospital blood products (blood components or whole blood, pBP) been fully implemented.

Methods

We performed a retrospective review of the National Emergency Medical Services Information System (NEMSIS) from 2020 to 2023 for all trauma patients ≥16 years. Patients with prehospital systolic blood pressure <90 mmHg and heart rate >108 beats per minute, or a systolic blood pressure <70 mmHg, and who did not receive pBP products were included in the analysis. Potential lives saved were calculated using mortality and risk ratio estimates (RR) from previously published studies, assuming 100% nationwide access to pBP. A series of models were developed incorporating varying RR, mortality rate assumptions, and nationwide access to pBP to encompass a wide range of scenarios.

Results

A total of 260,472 patients met our inclusion criteria. Using a 22.1% 24‐h mortality rate and an RR of 0.629, 21,356 deaths over the four‐year study period could have potentially been saved with the nationwide implementation of pBP.

Conclusion

Transfusion of pBP offers the potential to save thousands of injured patients lives. Efforts toward making policy‐level interventions aimed at increasing the adoption and availability of pBP should be sought.

Abbreviations

COMBAT

control of major bleeding after trauma trial

EMS

Emergency Medical Services

NEMSIS

National Emergency Medical Services Information System

PAMPer

prehospital air medical plasma trial

pBP

prehospital blood products

RR

relative risk

TOWAR

type O whole blood and assessment of age during prehospital resuscitation trialp

US

United States

1. INTRODUCTION

Hemorrhagic shock is the leading cause of preventable deaths in traumatically injured patients. ¹ , ² , ³ Along with hemorrhage control, early transfusion of balanced blood components and/or whole blood for hemorrhagic shock has been associated with decreased mortality secondary to exsanguination, improved early mortality outcomes, and has been shown to be an effective intervention to improve survival in both prehospital and in‐hospital settings. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ For example, one large randomized controlled trial showed a nearly 10‐percentage point lower mortality rate among patients who had plasma added to their resuscitation regimen. ¹⁰ Balanced, blood product‐based resuscitation strategies have been validated and are advocated by both civilian and military clinical practice guidelines. ¹¹ , ¹² , ¹³ , ¹⁴

While blood products are readily available in the hospital setting, many barriers to prehospital blood transfusion exist, impeding timely access to this life‐saving intervention. It is estimated that only 1% of emergency medical services (EMS) within the United States have established prehospital blood programs, with faster rates of adoption being seen in agencies servicing suburban or rural populations. ¹⁵ As a result of this disparity, the vast majority of bleeding trauma patients must wait until arrival at a hospital to receive a blood transfusion. This delay in time‐sensitive care results in a high proportion of otherwise preventable prehospital and early in‐hospital deaths among patients in traumatic hemorrhagic shock. To achieve the goal of “Zero Preventable Deaths After Injury” set forth by the National Academy of Science, Engineering, and Medicine in 2016, many among the trauma care community have called for a population health approach to ensure widespread access to prehospital blood products. However, there is a paucity of data surrounding the impact of achieving near‐complete population‐level access to this proven lifesaving resource.

Therefore, we aimed to determine the number of lives that could potentially have been saved among injured patients in hemorrhagic shock had prehospital blood products (either blood components or whole blood, pBP) been fully implemented nationwide. This information is critical to inform the design, funding, and implementation of health policy initiatives aimed at understanding and mitigating barriers to timely access to blood transfusion among bleeding trauma patients.

2. METHODS

Prior to the literature review, data abstraction, and statistical analysis, local institutional review board approval was obtained. We initially conducted a focused literature review to assess the published treatment effects from multicenter randomized controlled trials of prehospital blood products. We selected studies that included early mortality at 24 h as an outcome, as that represented our primary focus. Military studies focusing on combat casualty care were not included in our study, as these encompass a trauma system, resources, and patient population inherently different from civilian trauma systems. Further, only studies assessing patients in the United States were assessed in efforts to maintain congruency with our targeted patient population. Our results yielded one study that met our search criteria: the Prehospital Air Medical Plasma (PAMPer) trial. Mortality and treatment effect data from the PAMPer Trial were used for our 24‐h mortality reduction models described later within the methods. ¹⁰

Data from the National Emergency Medical Services Information System (NEMSIS) database were analyzed between 2020 and 2023 to identify our target patient population. The NEMSIS database represents the largest aggregation of EMS data within the United States, capturing an estimated 91% of all nationwide injury‐related 911 dispatch calls and is maintained through collaborative efforts between the National Highway Traffic Safety Administration's Office of EMS and the University of Utah.

All adult (≥16 years) trauma activations with a blunt and/or penetrating mechanism of injury reported to the NEMSIS database were identified. Interfacility transfers were excluded in efforts to assess only primary trauma activations. Patients who arrived pulseless to the trauma bay, as well as any patients who received prehospital blood transfusions, were further excluded. Patients with a prehospital systolic blood pressure <90 mmHg and a heart rate >108 beats per minute, or a systolic blood pressure <70 mmHg, were identified to represent our target population based on inclusion criteria from previously published and ongoing randomized controlled trials assessing prehospital blood product administration (Figure 1). ¹⁰

FIGURE 1.

Study flow diagram.

To estimate potential lives saved relative to prehospital blood product transfusions, we matched the effect size from previously published literature to our included sample population using the following equation ¹⁰ , ¹⁶ :

$$\text{Lives saved}=\begin{array}{l}\left(1-\text{relative risk}\left[\mathrm{RR}\right]\right)\times \text{mortality}\mathrm{at}24\mathrm{h}\\ \times \%\text{availability of}\mathrm{EMS}\text{services with}\mathrm{pBP}.\end{array}$$

Within the model, relative risk was defined by the previously published literature and converted from odds ratios as necessary. The 24‐h mortality rate was defined from the respective previously published literature that was utilized to define relative risk. ¹⁰ Our model accounted for varying levels of EMS access to pBP, ranging from 10% to 100%, to explore a broad range of potential nationwide coverage scenarios. A second model was evaluated to examine the number of lives saved when prehospital times exceeded 20 min. This time period is based on data from a previously published post hoc analysis of PAMPer that demonstrated a survival benefit associated with prehospital plasma use for patients who had prehospital transport times greater than 20 min. ¹⁷ Finally, further theoretical models were derived over a series of relative risks and mortality rate assumptions to encompass a wide variety of potential scenarios. All analyses were performed using Microsoft Excel (Redmond, WA) and R (R Foundation).

3. RESULTS

A total of 25,540,544 patients were included in the NEMSIS database between 2020 and 2023. Following the application of our inclusion and exclusion criteria outlined in the Methods section, a total of 260,472 patients were available for analysis. Of the included patients, 54,619 patients met the inclusion criteria for 2020, 62,377 patients for 2021, 68,753 patients for 2022, and 74,723 patients for 2023. The average EMS time from on‐scene arrival to patient destination was 38 min.

Assuming a 24‐h mortality rate of 22.1% and RR of 0.63 for patients based on previously published data using prehospital blood component transfusions and similar inclusion data, a total of 21,356 lives nationwide may have been saved at 24 h with 100% availability of prehospital blood products over the four‐year study period. ¹⁰ Using variable pBP availability estimates ranging from 10% to 90% of nationwide coverage, 2136–19,221 deaths at 24 h may have been averted (Table 1) over the 4‐year period of the study.

TABLE 1.

Nationwide estimates (2020–2023) of lives saved at 24‐h based on PAMPer trial.

Treatment effect of prehospital blood	Mortality rate	Percentage of EMS agencies with access to prehospital blood	Potential lives saved at 24‐h
0.629	0.221	100%	21,356
		90%	19,221
		80%	17,085
		70%	14,950
		60%	12,814
		50%	10,678
		40%	8543
		30%	6407
		20%	4271
		10%	2136

Model adjustments accounting for variations in RR signify the robust nature of potential lives saved at 24 h under both more conservative and liberal treatment effect estimates of prehospital blood product transfusions on mortality. Similar adjustments for variations in pBP availability, as well as varied mortality assumptions, elucidated a strong potential for large numbers of lives saved at 24 h, even under the most modest mortality, pBP availability, and treatment effect estimates. Results from the adjusted models are highlighted in Table 2.

TABLE 2.

United States potential lives saved with pBP (2021–2024).

	33% mortality estimate				30% mortality estimate				25% mortality estimate				20% mortality estimate				15% mortality estimate				10% mortality estimate
	National access to pBP (%)				National access to pBP (%)				National access to pBP (%)				National access to pBP (%)				National access to pBP (%)				National access to pBP (%)
RR a	25	50	75	100	25	50	75	100	25	50	75	100	25	50	75	100	25	50	75	100	25	50	75	100
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
0.95	1074	2149	3223	4298	977	1954	2930	3907	814	1628	2442	3256	651	1302	1954	2605	488	977	1465	1954	326	651	977	1302
0.9	2149	4298	6447	8596	1954	3907	5861	7814	1628	3256	4884	6512	1302	2605	3907	5209	977	1954	2930	3907	651	1302	1954	2605
0.85	3223	6447	9670	12,893	2930	5861	8791	11,721	2442	4884	7326	9768	1954	3907	5861	7814	1465	2930	4395	5861	977	1954	2930	3907
0.8	4298	8596	12,893	17,191	3907	7814	11,721	15,628	3256	6512	9768	13,024	2605	5209	7814	10,419	1954	3907	5861	7814	1302	2605	3907	5209
0.75	5372	10,744	16,117	21,489	4884	9768	14,652	19,535	4070	8140	12,210	16,280	3256	6512	9768	13,024	2442	4884	7326	9768	1628	3256	4884	6512
0.7	6447	12,893	19,340	25,787	5861	11,721	17,582	23,442	4884	9768	14,652	19535	3907	7814	11,721	15,628	2930	5861	8791	11,721	1954	3907	5861	7814
0.65	7521	15,042	22,563	30,085	6837	13,675	20,512	27,350	5698	11,396	17,093	22,791	4558	9117	13,675	18,233	3419	6837	10,256	13,675	2279	4558	6837	9117
0.6	8596	17,191	25,787	34,382	7814	15,628	23,442	31,257	6512	13,024	19,535	26,047	5209	10,419	15,628	20,838	3907	7814	11,721	15,628	2605	5209	7814	10,419
0.55	9670	19,340	29,010	38,680	8791	17,582	26,373	35,164	7326	14,652	21,977	29,303	5861	11,721	17,582	23,442	4395	8791	13,186	17,582	2930	5861	8791	11,721
0.5	10,744	21,489	32,233	42,978	9768	19,535	29,303	39,071	8140	16,280	24,419	32,559	6512	13,024	19,535	26,047	4884	9768	14,652	19,535	3256	6512	9768	13,024

Note: National Access to pBP represents the percentage of EMS agencies with pBP. Grey shades are for ease visualization between the differing mortality estimate groups.

Abbreviations: pBP, prehospital blood products; RR, risk reduction.

^a RR represents the range of potential treatment effect of pBP on 24‐h mortality.

When assessing patients with EMS transport times greater than 20 min, 109,070 patients were identified, accounting for 41.9% of the study population. However, when taking EMS time on scene into consideration, a total of 232,623 patients (89% of the study population) had prehospital times (EMS on‐scene plus EMS transport) greater than 20 min. Using the previously described model, 19,073 patient deaths may have been averted at 24 h for patients with greater than 20 min of prehospital EMS time had pBP been fully available over the four‐year study period.

4. DISCUSSION

Every minute of delay in treatment for trauma patients with hemorrhagic shock has been shown to increase mortality, measured both in the hospital and prehospital settings. ⁸ , ⁹ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ Naturally, transfusing blood products at the earliest time possible for these patients makes logical sense as a major factor toward improving survival. Data suggest that for every minute of delay in pBP transfusion following injury, mortality increases by 11%. ⁹ Despite this, only a small fraction of EMS agencies within the United States administer blood products in the prehospital setting. ¹⁵ , ²⁴ , ²⁵ , ²⁶ Lack of access to lifesaving blood transfusions represents a nationwide public health issue capable of causing widespread harm at a population level. Our analysis estimates that 21,356 deaths within the United States over the course of the first 24 h from injury may have been prevented during the 4‐year study period with universally available pBP. Although previous studies have demonstrated pBP transfusions are associated with decreased mortality in hemorrhaging patients, this study represents the first nationwide analysis quantifying the impact of pBP for trauma patients at a population level.

The estimates provided in this manuscript suggest that thousands of deaths may potentially be prevented within the first 24 h following injury; however, it is important to understand the nature of the data and these models. Our estimates were derived using early mortality data from the PAMPer trial, the only US multicenter randomized controlled trial assessing pBP to date. This study demonstrated that trauma patients in hemorrhagic shock who received prehospital plasma transfusions had significantly lower mortality at 24 h and 30 days compared to those who underwent a standard resuscitation. ¹⁰

Notably, however, another smaller, single‐center randomized controlled trial, the Control of Major Bleeding After Trauma (COMBAT) trial, in the United States during the same period demonstrated conflicting results. ²⁷ Differences in prehospital time and transport methods have been theorized as a main driving force behind the discordant findings. ¹⁰ , ¹⁷ , ²⁷ A subsequent harmonized analysis incorporating both trials concluded an overall mortality benefit associated with prehospital plasma transfusions that was most pronounced in patients with prehospital transport times greater than 20 min. ¹⁷ Taking only patients with prolonged prehospital times into account, our analysis suggests a potential nationwide reduction of up to 19,073 trauma‐related deaths at the 24‐h mark.

Despite our models incorporating a treatment effect from PAMPer, which was designed to study the use of prehospital plasma, a secondary analysis of this trial suggests that the greatest mortality benefit was demonstrated for patients who received both packed red blood cells and plasma in the prehospital period. ²⁸ Extrapolating those findings to our patient population, the substantial 24‐h mortality reduction derived within our study may in fact be a conservative estimate. This, however, is only speculative, as there have been no US randomized controlled trials designed specifically to assess both packed red blood cells and plasma in the prehospital setting.

Several military studies have demonstrated the early mortality benefits associated with pBP transfusions, both packed red blood cells and plasma, highlighting the notion that early transfusions for patients in hemorrhagic shock significantly increase the probability of survival. ²³ , ²⁹ , ³⁰ , ³¹ Conversely, multiple studies performed throughout Europe and Australia have shown inconsistent findings with prehospital transfusions and their effect on mortality. ³² , ³³ , ³⁴ However, numerous differences between prehospital treatment systems, resources, resuscitation strategies, and healthcare practices within these settings suggest the findings from military and non‐US studies may not be directly applicable to our US civilian trauma systems and study population.

Importantly, our findings center around the treatment effect of utilizing blood components in the prehospital arena. As resuscitation practices continue to evolve, many hospitals have shifted to using whole blood‐based resuscitation strategies. ²⁴ , ³⁵ , ³⁶ , ³⁷ Select locations within the United States have incorporated whole blood into the prehospital setting, with retrospective data supporting its use. ⁶ , ⁷ , ³⁸ , ³⁹ To date, no randomized controlled trials have been published assessing the effects of prehospital whole blood; however, the Type O Whole Blood and Assessment of Age during Prehospital Resuscitation (TOWAR) trial is currently enrolling patients and should help shed light on this topic.

While our study provides estimates for lives saved at 24 h, there are multiple issues to consider when reflecting on these statistics. Our models incorporated the treatment effects demonstrated in PAMPer, a large, well‐funded, multicenter trial. It remains plausible that the true treatment effect across the population may differ from that of the study due to a variety of differences in trauma centers and access to care throughout the country. To address this notion, we assessed varying treatment effects to provide a series of theoretical estimates to encompass more conservative estimates, which may better represent the benefits seen in less‐resourced settings. While pBP resuscitation has the potential to improve mortality, access to in‐hospital blood products and definitive hemorrhage control remains a critical component toward optimizing survivability. Assuring timely access to these resources in various regions of the country has proved difficult and remains a key focus for improving trauma systems. ⁴⁰ , ⁴¹

Moreover, our results focus solely on 24‐h mortality implications. Although the majority of deaths secondary to hemorrhagic shock occur within the first 6 h from injury, later causes of mortality, such as traumatic brain injury and multiorgan failure, traditionally occur later in the hospital course. ⁴² Importantly, 24‐h survival does not equate to overall survivability, which should be considered when viewing our estimates. As such, it remains unclear how increased survival past 24 h will affect these later potential causes of mortality, hospital costs, and quality of life.

The estimates from this study were derived to highlight and expand upon the lack of access to pBP and begin to understand what impact this may have at a population level. Currently, there remain many barriers within the United States for prehospital blood transfusions by EMS. While multiple well‐funded prehospital agencies have published their pBP protocols and practice patterns, billing and reimbursement, the lack of linked national prehospital datasets to in‐hospital outcomes, nationwide shortages of blood products, as well as concerns over wastage, maintenance of proper cold‐chain storage techniques, training, state‐by‐state variations in EMS providers' scope of practice, and medico‐legal issues all currently hinder the widespread adoption of prehospital transfusions. ¹⁵ , ²⁵ , ²⁶ , ³⁹ As the trauma community continues to invest in efforts toward improving in‐hospital resuscitation, a focus on improving prehospital resuscitation and addressing transfusion barriers should be made.

We acknowledge that this analysis is not without its limitations. As stated above, these estimates may not be direct reflections of true reductions in mortality. These data were derived from the largest US multicenter randomized controlled trial assessing prehospital transfusions to date; however, as newer studies are published incorporating the use of more novel resuscitation techniques, such as prehospital whole blood, variations in mortality reduction and its impact on other trauma‐related outcomes may become more prevalent. Although PAMPer studied prehospital plasma use during air medical transport, the vast majority of prehospital care in the United States is represented by ground‐based EMS platforms, resulting in the possibility that the results of PAMPer may not be a direct representation for all trauma patients. With regard to our data, NEMSIS is believed to capture over 90% of all injury‐related 911 dispatch calls across the nation; however, reporting is voluntary leading to the potential for selection bias. That said, without a national prehospital database similar to the American College of Surgeons Trauma Quality Improvement Program that integrates prehospital with hospital outcome data, NEMSIS represents the best nationwide prehospital dataset available. Our inclusion criteria did not take into account injury patterns and only assessed physiologic parameters. This was by design in efforts to match the inclusion criteria utilized in the PAMPer and ongoing TOWAR trials; however, we recognize that these estimates may not completely capture all trauma patients at risk for hemorrhagic shock. Moreover, our estimates include prehospital blood product availability as a binary variable. We recognize that it is infeasible for EMS agencies to carry infinite amounts of blood products; therefore, more work is needed to identify reasonable quantities of blood products needed in the prehospital phases of care. Finally, our mortality reduction estimates only highlight 24‐h mortality. While overall survival represents the ideal outcome metric, prehospital transfusions address ongoing hemorrhage. As a result, our intent was to capture a time frame that would encompass deaths secondary to hemorrhagic shock while mitigating deaths from other competing later causes of mortality. As stated previously, we recognize that improved 24‐h mortality does not necessarily equate to overall survival; however, the impact of this has yet to be elucidated.

Despite these limitations, our data suggest that implementing the routine use of prehospital blood products throughout the United States may provide a lifesaving adjunct to hemorrhagic shock that is not currently widely utilized. Implementing these practices should be considered in all settings, not just major urban settings or Level 1 trauma centers, in order to minimize variation in resuscitation strategies and promote national equity in trauma care. Ultimately, our hope is that these data may be utilized to help policymakers reevaluate national healthcare policies and prehospital reimbursement structures to assure that where patients are injured does not dictate the level of care they receive. The intent of this manuscript is to highlight a critical gap in prehospital trauma care and its potential impact on the US population. While research is ongoing to derive the optimal prehospital resuscitation practices, an emphasis on future work seeking solutions for the widespread implementation of these principles should be made.

5. CONCLUSION

Prehospital blood transfusions in traumatically injured patients offer the potential to improve 24‐h survival for thousands of patients annually. Implementing prehospital blood product transfusion programs across the United States would help to overcome critical resuscitation gaps in trauma care and directly support the goal of “Zero Preventable Deaths After Injury.” Efforts toward making policy‐level interventions aimed at increasing the adoption and availability of prehospital blood product transfusions should be sought.

AUTHOR CONTRIBUTIONS

DL, RB, RH, JD, JM, ME, JB, JJ, ZH contributed to the study design. DL, RB, RH, JD, JC, JW, JM performed the literature search. DL, RB, JK, ZH performed the data collection. DL, JM, ME, JB, ZH, JJ performed the statistical analysis. DL, RB, RH, JD, JC, JW, JM, ME, JJ, JK, JH, ZH critically interpreted the statistical results. DL, RB, JM, ME, JB drafted the original manuscript. DL, RB, RH, JD, JC, JW, JM, ME, JJ, JK, JH, ZH critically revised the manuscript.

CONFLICT OF INTEREST STATEMENT

DL, RB, ME, JB, ZH, and JJ all have grant funding support through the Department of Defense on unrelated research projects. ZH has grant funding from NIMHD (awaiting notice of funding), the UAB Faculty Development Grant Program, RevMedx, and the National highway Traffic Safety Administration on unrelated research projects. JH is a co‐founder and on the Board of Directors of Decisio Health, on the Board of Directors of QinFlow and Zibria, a co‐inventor of the Junctional Emergency Tourniquet Tool, an adviser to Aspen Medical, and the Wake Forest Institute of Regenerative Medicine. JJ also has unrelated grant funding from NHLBI, NIHR, MTEC, study support from CSL Behring, RevMedX, Infrascan, and is a consultant for CSL Behring, Infrascan, Cellphire, and Octapharm. JM is a co‐founder of the Behind The Knife Podcast. DL, RH, JD, JC, JW, and JM are all active‐duty military members; however, the views in this paper are those of our own and do not reflect the views of the government or the Department of Defense. JK is currently the ACS COT President. The remaining authors declare no conflicts of interest.

Lammers DT, Betzold R, Henry R, Dilday J, Conner JR, Williams JM, et al. Nationwide estimates of potential lives saved with prehospital blood transfusions. Transfusion. 2025;65(Suppl. 1):S14–S22. 10.1111/trf.18174