Evaluating the Whole Blood Survival Advantage: How Citrate and Calcium Levels Drive Mortality in a Translational Swine Model of Polytrauma Authors: Jason M. Rall, Craig D. Nowadly, Jae-Hyek Choi, Heang Sundermann, Kaysie Sutton, Allyson A. Arana, Dena Norouzi, R. Madelaine Paredes, Joseph K. Maddry Published: 2026 ABSTRACT ObjectivesWhole blood (WB) is increasingly used for trauma resuscitation, but clinical comparisons with component therapy (COMP) have reported mixed survival results, and physiologic contributors to any differences remain incompletely defined. We evaluated the association between excess citrate exposure, calcium (Ca2+) supplementation, hemodynamic stability, and survival during massive transfusion.MethodsIn a randomized swine (Sus scrofa) polytrauma model, 32 castrated male and nonpregnant females underwent tibial fracture, blunt liver injury, 30% controlled hemorrhage, and a 30-minute shock period. Animals were assigned to WB without Ca2+ supplementation (WB/−Ca), WB with Ca2+ supplementation (WB/+Ca), simulated component therapy without Ca2+ supplementation (COMP/−Ca), or simulated component therapy with Ca2+ supplementation (COMP/+Ca). Resuscitation occurred during continued hemorrhage, followed by 4 hours of monitoring. The primary outcome was mean arterial pressure (MAP) after completion of transfusion. Secondary outcomes included survival, ionized and serum Ca2+, potassium, and lactate.ResultsBaseline values were similar across groups. Survival was 3 of 8 animals in COMP/−Ca and 8 of 8 animals in each other group (log-rank P < .001). MAP differed significantly between groups over time (P < .001), with lower MAP in COMP/−Ca compared with WB/−Ca (P = .0036), WB/+Ca (P = .0046), and COMP/+Ca (P = .0076). Ionized Ca2+ was lowest in COMP/−Ca.ConclusionsIn this polytrauma model, simulated COMP with excess citrate and without Ca2+ supplementation was associated with lower ionized Ca2+, hemodynamic instability, and decreased survival. These findings support further investigation into how citrate burden and Ca2+ supplementation strategies influence massive transfusion physiology during WB and COMP-based resuscitation. The Bottom LineWhole blood (WB) is increasingly used to treat traumatic bleeding, but clinical studies comparing WB with component therapy have shown mixed survival results. One possible difference between WB and component therapy is varying levels of citrate, an anticoagulant that can lower calcium during large-volume transfusion. In this swine trauma study, animals received WB or simulated component therapy, with or without calcium supplementation. Survival was lowest with simulated component therapy and no calcium supplementation: 3 of 8 animals survived, compared with 8 of 8 animals in the WB groups and the simulated component therapy with calcium supplementation group. These findings support further investigation into how citrate burden and Ca2+ supplementation strategies influence massive transfusion physiology during WB and component-based resuscitation. The Bottom Line Whole blood (WB) is increasingly used to treat traumatic bleeding, but clinical studies comparing WB with component therapy have shown mixed survival results. One possible difference between WB and component therapy is varying levels of citrate, an anticoagulant that can lower calcium during large-volume transfusion. In this swine trauma study, animals received WB or simulated component therapy, with or without calcium supplementation. Survival was lowest with simulated component therapy and no calcium supplementation: 3 of 8 animals survived, compared with 8 of 8 animals in the WB groups and the simulated component therapy with calcium supplementation group. These findings support further investigation into how citrate burden and Ca2+ supplementation strategies influence massive transfusion physiology during WB and component-based resuscitation. Introduction Background Massive hemorrhage remains a leading cause of preventable death in both civilian and military trauma.1,2 Modern resuscitation strategies have increasingly shifted toward the use of whole blood (WB), particularly cold-stored low titer O type whole blood (CS-LTOWB).3, 4, 5 This transition has been driven by evidence suggesting that both fresh whole blood and cold-stored whole blood (CS-WB) may improve short- and long-term survival, decrease total blood product utilization, and improve hemodynamic stability compared with traditional component therapy (COMP). Civilian and military trauma systems, including the Joint Trauma System (JTS) and the Armed Services Blood Program, have therefore integrated CS-WB as a primary resuscitation product during damage control resuscitation, with LTOWB serving as the predominant formulation and expanding use of type-specific WB where logistically feasible.6,7 Despite these observed advantages, the dominant physiologic mechanisms driving the apparent benefit of CS-WB over COMP for resuscitation remain unclear. Proposed explanations include differences in rate of product infusion, dilutional effects, cellular integrity, and storage age.8,9 However, one of the most compelling hypotheses centers on the biochemical interaction between citrate anticoagulant and ionized Ca2+, as COMP exposes patients to greater citrate loads than WB.10, 11, 12 The extent to which this citrate-Ca2+ relationship influences hemodynamic stability and survival remains poorly understood. Importance Citrate is a key component of modern anticoagulant solutions, such as citrate-phosphate-dextrose (CPD) and citrate-phosphate-dextrose-adenine (CPDA-1). It prevents stored blood from clotting by binding ionized Ca2+, which is required for multiple enzymatic steps in the coagulation cascade. This same mechanism that preserves donated blood products becomes problematic when large volumes are transfused, especially during traumatic resuscitation, as the infused citrate continues to chelate Ca2+ within the recipient. This iatrogenic hypocalcemia disrupts clot formation and may amplify the lethal diamond of acidosis, coagulopathy, hypothermia, and trauma-induced hypocalcemia.13,14 In addition to its effects on Ca2+, excess citrate can contribute to acid–base disturbances due to its acidic pH and metabolic consumption of bicarbonate. Clinical data indicate that more than 85% of patients receiving massive transfusion develop ionized hypocalcemia, and the severity of Ca2+ depletion correlates with both transfusion volume and mortality.15,16 Although Ca2+ supplementation is recommended during massive transfusion, timing and dosing can be inconsistent, especially in low-resource settings and early in traumatic resuscitation.17,18 Goals of This Investigation In this study, we sought to investigate the relationship between citrate exposure, Ca2+ supplementation, and hemodynamic stability during massive transfusion. Using a pragmatic, randomized, large-animal model of polytrauma and ongoing hemorrhage, we compared resuscitation with WB and simulated COMP that differed only in citrate content and Ca2+supplementation strategy. We hypothesized that excess citrate administration in the absence of Ca2+ supplementation would result in impaired hemodynamic recovery, metabolic stability, and survival. Furthermore, we hypothesized that Ca2+ supplementation would mitigate these effects. Methods Study Design This study was a prospective, randomized, blindly allocated, large-animal translational study. The local Institutional Animal Care and Use Committee approved this research. All research personnel involved in the study obtained appropriate training through the American Association for Laboratory Animal Science, and animal care was performed in strict compliance with the Guide for the Care and Use of Laboratory Animals in a facility accredited by Association for Assessment and Accreditation of Laboratory Animal Care, International. Healthy, adult, castrated male and nonpregnant female Yorkshire-cross (Sus scrofa) were allowed to acclimate for at least 5-days in temperature- and light-controlled pens with access to environmental enrichment. Animals weighed between 50 and 70 kg and were fasted the night before the experiment with free access to water. The protocol is illustrated in Fig 1. Before initiation of the randomized protocol, 3 animals were used for model development to refine the injury, hemorrhage, resuscitation, and transfusion procedures. These animals were not randomized and were not included in the final analysis.Figure 1Experimental design and protocol illustration. Animals underwent controlled hemorrhage to 30% total blood volume followed by a fixed 30-minute hemorrhagic shock period where no further hemorrhage occurred, after which they were randomized to resuscitation with WB or COMP with or without Ca supplementation during ongoing hemorrhage. Timepoints are referenced to the start of randomized resuscitation (T = 0 min). Blood removed is depicted below the timeline, and blood transfused is depicted above the timeline. COMP, simulated component therapy; EJ, external jugular; ETT, endotracheal tube; WB, whole blood. Experimental design and protocol illustration. Animals underwent controlled hemorrhage to 30% total blood volume followed by a fixed 30-minute hemorrhagic shock period where no further hemorrhage occurred, after which they were randomized to resuscitation with WB or COMP with or without Ca supplementation during ongoing hemorrhage. Timepoints are referenced to the start of randomized resuscitation (T = 0 min). Blood removed is depicted below the timeline, and blood transfused is depicted above the timeline. COMP, simulated component therapy; EJ, external jugular; ETT, endotracheal tube; WB, whole blood. Animal Preparation The animals were premedicated with intramuscular injections of atropine (2.0 mg), tiletamine-zolazepam (4 to 8 mg/kg), and buprenorphine HCl (0.01 mg/kg). Anesthesia was induced with a facemask and 2% to 4% isoflurane. Once intubated, isoflurane was adjusted to 1% to 3% during procedures. Throughout the experiment, animals were mechanically ventilated with tidal volumes of 8 to 10 mL/kg, 30% fraction of inspired oxygen (FiO2), a positive end-expiratory pressure of 5 cm H2O, and a respiratory rate of 10 to 15 breaths per minute, titrated to maintain end-tidal CO2 (EtCO2) at 40 ± 5 mm Hg and an oxygen saturation (SpO2) of >92%. A bolus of 7.5 mL/kg warm 0.9% sodium chloride was administered to ensure euvolemia. Animals were placed on an electric warming pad with a core temperature of 37 °C to 38.5 °C to be maintained throughout the experiment. All vascular access was obtained via Seldinger technique under ultrasound guidance. Both external jugular veins were cannulated with 8.5-Fr resuscitation catheter (Arrow-Flex; Teleflex, Wayne, PA). The right jugular vein was utilized for pulmonary catheter introduction (Swan-Ganz CCOmbo V, Edwards Lifesciences, Irvine, CA), and the left was accessed for infusion of resuscitation fluids. The right carotid artery was cannulated with a 5-Fr arterial catheter (Micropuncture Kit, Cook, Bloomington, IN) for continuous blood pressure measurement and blood sampling. The left femoral artery was accessed with a 7-Fr resuscitation catheter to facilitate controlled hemorrhage. Transurethral catheterization was performed on female animals, and a tube cystostomy was performed on male animals for continuous bladder draining. Injury, Controlled Hemorrhage, and Randomization A custom 4-inch plate captive bolt gun was discharged against the abdomen centered directly above the liver, and a 0.22 caliber powder-actuated tool (Ramset Mastershot, Powder Level 3, Glenview, IL) was fired twice against the right tibia. Simultaneously, controlled hemorrhage was initiated via a femoral arterial sheath to achieve 30% total blood volume loss using a modified Frankel protocol: 2.0 mL/kg/min for 7 minutes, then 1.2 mL/kg/min for 5 minutes. The target mean arterial pressure (MAP) during hemorrhage was 25 mm Hg to 30 mm Hg. Hemorrhage was paused for 60 seconds if MAP fell below 25 mm Hg and resumed at the slower rate (1.2 mL/kg/min) if MAP recovered to greater than 30 mm Hg; further hemorrhage was withheld if MAP remained below 30 mm Hg. To ensure complete collection of 30 % total blood volume, controlled hemorrhage was continued until the target volume was achieved. The time required to reach this threshold varied between animals due to physiologic differences. However, the subsequent hemorrhagic shock period was fixed at 30 minutes. For clarity, the experimental timeline is presented schematically, and all time points are referenced to the initiation of randomized resuscitation (T = 0 minutes). Shed blood was collected via the femoral arterial catheter into sterile collection bags containing citrate-phosphate-dextrose-adenine (CPDA-1) anticoagulant (Terumo BCT, Lakewood, CO). The volume collected from each animal was recorded, and bags were gently agitated during collection to ensure anticoagulant mixing. All collected blood was placed in a 35 °C (95 ℉) warmer and used for subsequent resuscitation of the same animal according to group allocation. Following initial hemorrhage, a 30 min hemorrhagic shock period was observed without additional intervention. During the shock period, animals were randomly assigned using the closed envelope method in a 1:1:1:1 ratio to 1 of 4 treatment groups (n = 8 in each):•WB without Ca2+ supplementation (“WB / −Ca”)•WB with Ca2+ supplementation (“WB / +Ca”)•simulated component therapy without Ca2+ supplementation (“COMP / −Ca”)•simulated component therapy with Ca2+ supplementation (“COMP / +Ca”) •WB without Ca2+ supplementation (“WB / −Ca”) WB without Ca2+ supplementation (“WB / −Ca”) •WB with Ca2+ supplementation (“WB / +Ca”) WB with Ca2+ supplementation (“WB / +Ca”) •simulated component therapy without Ca2+ supplementation (“COMP / −Ca”) simulated component therapy without Ca2+ supplementation (“COMP / −Ca”) •simulated component therapy with Ca2+ supplementation (“COMP / +Ca”) simulated component therapy with Ca2+ supplementation (“COMP / +Ca”) Intervention Resuscitation was initiated at 100 mL/min through the left external jugular vein using the autologous WB collected during the initial hemorrhage phase (T = 0). Concurrently, controlled hemorrhage was reinitiated at a rate of 50 mL/min to simulate ongoing blood loss during massive transfusion. Two additional units were sequentially collected and transfused during this phase, resulting in a total of 4 units of autologous blood administered to each animal over a 20-minute period. For animals randomly assigned to WB groups, transfused units consisted solely of autologous blood collected in CPDA-1 anticoagulant, with no further additives. For animals assigned to simulated COMP groups, 95 mL of sterile citrate-phosphate-dextrose solution (Cat# C7165, Sigma-Aldrich,St Louis, MO) was added to each unit immediately prior to transfusion to approximate the cumulative citrate content typically encountered with COMP. This approach was designed to isolate the physiologic effects of excess citrate exposure independent of other blood component differences. Animals assigned to Ca2+ supplementation groups received 1 g of calcium chloride intravenously per transfused unit. Animals not assigned to Ca2+ supplementation received no additional Ca2+ beyond baseline physiologic replacement. Following completion of transfusion of the 4th unit of blood, animals were monitored for a 4-hour observation period. Death was defined as a sustained mean arterial pressure <20 mm Hg and end-tidal CO2 <15 mm Hg for > 2 minutes. Laboratory Analysis Blood samples were obtained at baseline, at the end of hemorrhage, before resuscitation (T = 0 min), after 2 units of blood (T = 10), after 4 units of blood (T = 25), then at T = 40, 55, 80, 140, and 260 minutes after the start of resuscitation. Arterial blood gas analysis was used to measure pH, PaO2, PaCO2, potassium, ionized Ca2+, sodium, chloride, glucose, lactate, calculated bicarbonate, and calculated base excess (ABL 800 Flex blood gas analyzer, Radiometer America, Westlake, OH). Coagulation parameters were tested by rotational thromboelastometry (ROTEM, Werfen, Bedford, MA), prothrombin time, and activated partial thromboplastin time (Stago, Parsippany, NJ). Expanded blood chemistry tests included creatinine, total Ca2+, total magnesium, aspartate amino transferase (AST), alanine amino transferase (ALT), creatinine, creatinine kinase, and blood urea nitrogen. Outcomes The primary outcome was MAP after completion of transfusion of the fourth unit of blood. MAP was analyzed as a continuous outcome and compared between groups rather than dichotomized according to predefined success thresholds. Prespecified secondary outcomes included survival through the 4-hour observation period, ionized and serum Ca2+, potassium, and lactate. Ionized Ca2+, serum Ca2+, potassium, and lactate were analyzed as continuous laboratory outcomes across groups and over time. The prespecified mortality criterion was sustained MAP <20 mm Hg and end-tidal carbon dioxide <15 mm Hg for >2 minutes. Statistical Analysis Data are presented as mean ± standard deviation for normally distributed continuous variables and median with interquartile range for non-normally distributed continuous variables. Analysis of variance (ANOVA) with Tukey post hoc testing was used to compare continuous variables across groups when appropriate. Log-rank analysis was used to compare survival times among groups and visualized using a Kaplan–Meier curve. Repeated-measures outcomes were analyzed using mixed-effects models fit using restricted maximum likelihood, with treatment group and time included as model factors and individual animals included to account for repeated measurements within the same animal. Animals stopped contributing data after meeting predefined death criteria, and postdeath values were not imputed. Therefore, hemodynamic and laboratory comparisons after early mortality reflect observed values among surviving animals at each time point. This approach avoids treating postdeath values as physiologic measurements, although it introduces survivor bias for later time points. Significance was set at P < .05. Power analysis was based on the primary outcome, MAP, from a study that compared blood preservatives in a swine model of controlled hemorrhage.19 A sample of 8 animals in each group (32 total) is adequate to find an effect size 0.42 with 82% power and an alpha of 0.05 in an ANOVA. Results Baseline Characteristics and Shock Period A total of 37 animals were entered into the study: 3 were used for model development, 2 were excluded due to early demise prior to random assignment, and the remaining 32 animals (8 per group) were included in the analysis. At baseline, there were no significant differences among groups with respect to weight, sex, hemodynamics, or laboratory results (Table).TableBaseline characteristics and physiologic parameters before and after initial hemorrhage.WB/−CaWB/+CaCOMP/−CaCOMP/+CaP valuen8888.999Weight (kg)62.6 ± 3.159.9 ± 2.362.6 ± 4.363.8 ± 3.2.140Sex % male100% (8/8)87.5% (7/8)100% (8/8)87.5% (7/8).999BaselineMean Arterial Pressure (mm Hg)73.1 ± 9.567.0 ± 6.166.8 ± 8.468.0 ± 8.6.384End-tidal CO2 (mm Hg)43.5 ± 2.044.6 ± 1.843.9 ± 1.843.9 ± 2.2.698White blood cell (×103/μL)14.8 ± 3.915.0 ± 2.917.7 ± 4.116.5 ± 2.6.313Hemoglobin (g/dL)9.12 ± 0.708.76 ± 0.799.15 ± 0.979.17 ± 1.00.763pH7.425 ± 0.0207.437 ± 0.0427.424 ± 0.0267.405 ± 0.020.181PaCO2 (mm Hg)43.3 ± 2.844.3 ± 1.443.9 ± 2.343.7 ± 1.3.769Lactate (mmol/L)1.59 ± 0.241.45 ± 0.441.54 ± 0.521.63 ± 0.42.852Serum calcium (mg/dL)9.78 ± 0.759.43 ± 0.5610.0 ± 1.309.94 ± 1.00.624Ionized calcium (mmol/L)1.37 ± 0.051.37 ± 0.051.37 ± 0.651.38 ± 0.36.981Postinitial hemorrhageTotal hemorrhage volume (mL)1176 ± 1321130 ± 1121094 ± 2261235 ± 104.295Mean arterial pressure (mm Hg)36.4 ± 9.534.3 ± 9.930.9 ± 8.136.8 ± 7.2.527End-tidal CO2 (mm Hg)40.6 ± 2.140.5 ± 1.838.3 ± 2.440.9 ± 2.9.112Hemoglobin (g/dL)10.24 ± 1.289.36 ± 0.899.81 ± 0.899.73 ± 1.30.481Lactate (mmol/L)1.99 ± 0.471.75 ± 0.431.76 ± 0.492.11 ± 0.53.374pH7.427 ± 0.0297.429 ± 0.0227.448 ± 0.0317.413 ± 0.034.141PaCO2 (mm Hg)40.2 ± 2.939.7 ± 1.638.9 ± 1.841.1 ± 2.8.343Serum calcium (mg/dL)9.48 ± 0.679.14 ± 0.379.74 ± 1.219.78 ± 0.89.419Ionized calcium (mmol/L)1.36 ± 0.051.34 ± 0.61.31 ± 0.131.38 ± 0.04.283Ca, calcium; COMP, simulated component therapy; mm Hg, millimeters of mercury; PaCO2, arterial partial pressure of carbon dioxide; WB, whole blood.Baseline demographics, hemodynamic variables, and laboratory values are shown for all animals prior to injury and after completion of controlled hemorrhage, immediately before the hemorrhagic shock period. Data are presented by treatment group. TableBaseline characteristics and physiologic parameters before and after initial hemorrhage.WB/−CaWB/+CaCOMP/−CaCOMP/+CaP valuen8888.999Weight (kg)62.6 ± 3.159.9 ± 2.362.6 ± 4.363.8 ± 3.2.140Sex % male100% (8/8)87.5% (7/8)100% (8/8)87.5% (7/8).999BaselineMean Arterial Pressure (mm Hg)73.1 ± 9.567.0 ± 6.166.8 ± 8.468.0 ± 8.6.384End-tidal CO2 (mm Hg)43.5 ± 2.044.6 ± 1.843.9 ± 1.843.9 ± 2.2.698White blood cell (×103/μL)14.8 ± 3.915.0 ± 2.917.7 ± 4.116.5 ± 2.6.313Hemoglobin (g/dL)9.12 ± 0.708.76 ± 0.799.15 ± 0.979.17 ± 1.00.763pH7.425 ± 0.0207.437 ± 0.0427.424 ± 0.0267.405 ± 0.020.181PaCO2 (mm Hg)43.3 ± 2.844.3 ± 1.443.9 ± 2.343.7 ± 1.3.769Lactate (mmol/L)1.59 ± 0.241.45 ± 0.441.54 ± 0.521.63 ± 0.42.852Serum calcium (mg/dL)9.78 ± 0.759.43 ± 0.5610.0 ± 1.309.94 ± 1.00.624Ionized calcium (mmol/L)1.37 ± 0.051.37 ± 0.051.37 ± 0.651.38 ± 0.36.981Postinitial hemorrhageTotal hemorrhage volume (mL)1176 ± 1321130 ± 1121094 ± 2261235 ± 104.295Mean arterial pressure (mm Hg)36.4 ± 9.534.3 ± 9.930.9 ± 8.136.8 ± 7.2.527End-tidal CO2 (mm Hg)40.6 ± 2.140.5 ± 1.838.3 ± 2.440.9 ± 2.9.112Hemoglobin (g/dL)10.24 ± 1.289.36 ± 0.899.81 ± 0.899.73 ± 1.30.481Lactate (mmol/L)1.99 ± 0.471.75 ± 0.431.76 ± 0.492.11 ± 0.53.374pH7.427 ± 0.0297.429 ± 0.0227.448 ± 0.0317.413 ± 0.034.141PaCO2 (mm Hg)40.2 ± 2.939.7 ± 1.638.9 ± 1.841.1 ± 2.8.343Serum calcium (mg/dL)9.48 ± 0.679.14 ± 0.379.74 ± 1.219.78 ± 0.89.419Ionized calcium (mmol/L)1.36 ± 0.051.34 ± 0.61.31 ± 0.131.38 ± 0.04.283Ca, calcium; COMP, simulated component therapy; mm Hg, millimeters of mercury; PaCO2, arterial partial pressure of carbon dioxide; WB, whole blood.Baseline demographics, hemodynamic variables, and laboratory values are shown for all animals prior to injury and after completion of controlled hemorrhage, immediately before the hemorrhagic shock period. Data are presented by treatment group. Baseline characteristics and physiologic parameters before and after initial hemorrhage. Ca, calcium; COMP, simulated component therapy; mm Hg, millimeters of mercury; PaCO2, arterial partial pressure of carbon dioxide; WB, whole blood. Baseline demographics, hemodynamic variables, and laboratory values are shown for all animals prior to injury and after completion of controlled hemorrhage, immediately before the hemorrhagic shock period. Data are presented by treatment group. Following completion of the polytrauma and initial controlled hemorrhage, all animals had physiologic changes consistent with hemorrhagic shock, including a decrease in both MAP and EtCO2 at the end of hemorrhage (Table). Across groups, there were no differences in the total blood volume removed, hemodynamics, or laboratory results obtained immediately before transfusion at T = 0 min (Table). Survival Only 3 of the 8 animals in the COMP/−Ca group survived to the end of the experiment, whereas all animals in the WB/−Ca, WB/+Ca, and COMP/+Ca survived through the full observation period. Median time of survival in the COMP/−Ca group was 35 (23.75 to 260) min after the onset of resuscitation compared with 260 (260 to 260) min in the other 3 groups. The Kaplan–Meier analysis demonstrated a significant difference in survival among groups (log-rank; P < 0.001), with the deaths all occurring shortly after the completion of resuscitation (Fig 2).Figure 2Kaplan–Meier survival analysis by treatment group. Survival over the postresuscitation observation period is shown, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using log-rank analysis. Kaplan–Meier survival analysis by treatment group. Survival over the postresuscitation observation period is shown, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using log-rank analysis. Primary Outcome: MAP MAP differed significantly between groups over time during and after randomized resuscitation (Fig 3, P < .001). The COMP/−Ca group demonstrated an abrupt decline in MAP compared with the other 3 groups. Pairwise post hoc comparisons demonstrated significantly lower MAP in the COMP/−Ca group compared with WB/−Ca (P = .0036), WB/+Ca (P = .0046), and COMP/+Ca (P = .0076). The WB/+Ca group showed a slight increase in MAP compared with the WB/−Ca and COMP/+Ca groups, although these differences did not reach statistical significance. These MAP findings should be interpreted alongside the survival results above because early deaths in the COMP/−Ca group reduced the number of animals contributing data at later time points.Figure 3Hemodynamic responses by group over time. A, Mean arterial pressure (MAP) and B, end-tidal carbon dioxide (EtCO2) are shown over the experimental course, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. Hemodynamic responses by group over time. A, Mean arterial pressure (MAP) and B, end-tidal carbon dioxide (EtCO2) are shown over the experimental course, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. Additional Hemodynamics Consistent with the MAP findings, systolic blood pressure, diastolic blood pressure, and pulse pressure also differed significantly between groups over time (overall P < .001, P = .007, and P = .001, respectively; Fig S1). EtCO2 also demonstrated a marked decline in the COMP/−Ca group after transfusion and differed significantly between groups over time (overall P < .001). Among animals still contributing repeated-measures data, other hemodynamic parameters, including heart rate, pulmonary artery pressure, cardiac output, and mixed venous oxygen saturation, did not differ significantly between groups (Fig S1). Laboratory Findings Ionized Ca2+ declined abruptly after resuscitation in the COMP/−Ca group during the early postresuscitation period (Fig 4). In contrast, decreases in ionized Ca2+ were less pronounced in the COMP/+Ca and WB/−Ca groups, which showed overlapping values, while the WB/+Ca group maintained ionized Ca2+ near or slightly above baseline. Ionized Ca2+ differed significantly between groups overall, with pairwise differences observed between all groups except COMP/+Ca and WB/−Ca. Finally, another physiologically critical metal cation that also binds citrate, magnesium, did not significantly differ among groups despite an observable drop in the COMP/+Ca group upon resuscitation.Figure 4Cation concentrations by group over time. (A) Ionized calcium; (B) serum calcium; and (C) magnesium are shown over the experimental course, with T = 0 defined as the start of randomized resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. Cation concentrations by group over time. (A) Ionized calcium; (B) serum calcium; and (C) magnesium are shown over the experimental course, with T = 0 defined as the start of randomized resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. Acid–base parameters demonstrated distinct early temporal differences between groups (Fig 5). The COMP/−Ca group had a higher pH accompanied by a significant reduction in PaCO2 following initiation of resuscitation. Chloride concentrations were also lower in the COMP/−Ca group, whereas bicarbonate (approximated by Total CO2) and lactate levels did not differ between groups. Although base excess did not differ significantly, both citrate groups demonstrated a consistent trend toward more negative base excess over time compared with noncitrate groups.Figure 5Acid–base by group over time. (A) pH; (B) Arterial CO2; (C) Total CO2; (D) Total Sodium; (E) Total Chloride; and (F) Potassium, over the experimental course, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. Acid–base by group over time. (A) pH; (B) Arterial CO2; (C) Total CO2; (D) Total Sodium; (E) Total Chloride; and (F) Potassium, over the experimental course, with T = 0 defined as the start of randomly assigned resuscitation. Groups were compared using mixed-effects analysis, and overall P values are shown. Animals were excluded from analysis after meeting predefined death criteria. Significant differences between the COMP/−Ca group and the WB/−Ca, WB/+Ca, and COMP/+Ca groups are indicated by ∗. COMP, simulated component therapy; WB, whole blood. During the shock and early resuscitation period, laboratory values were largely similar across groups (Fig S2). All groups exhibited a biphasic change in potassium, with an initial increase following hemorrhage, a decrease after resuscitation, and a gradual rise over the remainder of the observation period. Lactate demonstrated a single increase following hemorrhage and the hemorrhagic shock period, which decreased after fluid resuscitation. Significant between-group differences were observed only in glucose, with the COMP/−Ca group demonstrating higher glucose during the resuscitation period than the other groups. No significant differences were identified among groups for any other laboratory values measured. Limitations Several limitations should be considered when interpreting these findings. First, early mortality in the COMP/−Ca group introduced survivor bias into later hemodynamic and laboratory comparisons because animals stopped contributing repeated-measures data after meeting predefined death criteria, and postdeath values were not imputed. As a result, later time points reflect only surviving animals and may underestimate between-group differences. Next, animals were resuscitated with autologous fresh WB rather than CS-LTOWB, which may differ in hemostatic function, have storage-related changes, and other metabolic changes. All animals sustained a blunt hepatic injury as part of the polytrauma model, which may have influenced citrate clearance and limits extrapolation to trauma patients without liver injury. The experimental design required reinfusion of collected blood, and transfused blood could theoretically have been subsequently hemorrhaged and retransfused, a constraint inherent to large-animal models with finite blood volumes. Finally, although swine are a well-established translational model for trauma and transfusion research, species-specific differences in Ca2+ handling, citrate metabolism, and cardiovascular physiology may limit direct generalizability to humans. Baseline ionized calcium concentrations in our swine cohort (1.37 ± 0.05 mmol/L) exceeded the accepted human reference range (1.05 – 1.30 mmol/L), consistent with published values,20 suggesting species differences in calcium homeostasis and susceptibility to hypocalcemia. This may limit direct extrapolation of calcium thresholds to humans. Discussion In this translational swine model of polytrauma with ongoing hemorrhage and massive transfusion, we found that excess citrate exposure without Ca2+ replacement produced early hemodynamic collapse and high mortality. Five of 8 animals in the COMP/−Ca group did not survive beyond the completion of transfusion, whereas all animals resuscitated with WB with or without Ca2+, and all animals receiving simulated COMP with Ca2+ supplementation, survived the full observation period. Death clustered in the immediate postresuscitation period, consistent with an acute physiologic failure that did not progress to the classic lethal diamond of trauma-associated acidosis, coagulopathy, hypothermia, and hypocalcemia. Instead, excess citrate exposure produced an early, dominant cardiovascular failure that preceded the development of overt coagulopathy or additional metabolic derangements. Citrate-associated hypocalcemia is a well-described risk during massive transfusion, but the magnitude and consistency of the associated mortality phenotype in this model were striking. The physiologic basis for improved outcomes with WB resuscitation remains incompletely defined, in part because clinical comparisons with COMP are confounded by injury severity, transfusion ratios, timing, and ongoing resuscitation interventions. In this study, we isolated citrate burden as a primary experimental variable by simulating COMP with excess citrate, while holding transfusion volume, rate, and injury severity constant. Under these conditions, citrate exposure alone was sufficient to precipitate a lethal phenotype in the absence of Ca2+ replacement, whereas WB resuscitation, even without exogenous Ca2+, was not. These findings suggest that differences in citrate load may contribute meaningfully to the survival advantage observed with WB in both military and civilian trauma systems. Animals in the COMP/−Ca group exhibited abrupt hemodynamic collapse characterized by a marked decline in MAP with only a modest, nonsignificant trend toward lower heart rate, occurring shortly after completion of transfusion. This pattern is inconsistent with hemorrhagic shock alone and suggests a superimposed failure of Ca2+-dependent vascular smooth muscle tone and myocardial contractility.21, 22, 23 Excess citrate exposure without Ca2+ supplementation resulted in profound ionized hypocalcemia that temporally coincided with MAP collapse and subsequent death. Similar hemodynamic signatures are observed in states of acute Ca2+ unavailability, such as severe Ca2+ channel blocker toxicity, in which hypotension is often disproportionate to changes in heart rate.24,25 Beyond acute hemodynamic collapse, animals in the COMP/−Ca group demonstrated a broader physiologic pattern consistent with severe metabolic stress. Although all animals sustained a blunt hepatic injury as part of the polytrauma model, AST elevations were more pronounced in the COMP/−Ca group, suggesting greater hepatic injury or ongoing cellular stress. Given the central role of the liver in citrate metabolism, impaired hepatic function may exacerbate and prolong ionized hypocalcemia during resuscitation.10,26 Additional laboratory abnormalities in the COMP/−Ca group, including lower chloride and paradoxical alkalosis, are notable, given that severe shock is more commonly associated with acidosis. We hypothesize that these findings may reflect the combined effects of citrate metabolism, absence of chloride-containing calcium supplementation, and impaired perfusion, which together may transiently alter strong ion balance and pH. Interpretation of later laboratory trends is limited by survivor bias due to early mortality in this group. However, the overall pattern supports a coherent physiologic response to excess citrate exposure in the setting of severe injury and impaired metabolic reserve. These findings offer a framework through which to interpret Ca2+ derangements observed during massive transfusion, particularly when comparing WB-based and component-based resuscitation strategies. The citrate burden used in the simulated COMP groups was intentionally selected to stress the underlying biologic question rather than to replicate any single clinical protocol. These animals received 2.5 times the citrate content of citrated WB. This dose was selected to model the upper range of citrate burden expected during balanced component transfusion while remaining consistent with published estimates of citrate exposure associated with common component ratios.10,17 Additionally, comparable citrate loads are plausible during rapid transfusion, high component ratios, and in patients with impaired hepatic clearance due to shock or liver injury.15,16 Ca2+ dosing in this study similarly exceeded current JTS Damage Control Resuscitation Clinical Practice Guideline recommendations and was based on prior molar-equivalence work to offset citrate exposure,27 not to propose a clinical dosing strategy.4 These design choices allowed isolation of citrate–Ca2+ interactions but require careful interpretation when placed alongside civilian and military trauma registry data demonstrating that both hypocalcemia and hypercalcemia are associated with increased mortality.13,28,29 Rather than contradicting these observations, the present findings reinforce Ca2+ as a narrow-therapeutic-window variable whose impact depends on timing, citrate burden, and metabolic reserve. Taken together, these observations raise the possibility that the optimal approach to Ca2+ supplementation during massive transfusion may depend on transfusion strategy and citrate exposure, rather than a uniform dosing paradigm. In this translational swine model of polytrauma, excess citrate exposure without Ca2+ replacement was associated with early hemodynamic collapse and high mortality, occurring shortly after transfusion. These effects were not observed in animals resuscitated with WB or in those receiving simulated component therapy with Ca2+ supplementation. These findings support citrate-Ca2+ interactions as an important physiologic consideration during massive transfusion and highlight ionized Ca2+ availability as a key variable influencing hemodynamic stability. Further work is needed to define how transfusion strategies and citrate burden should inform Ca2+ management in trauma care. Authors Contribution Jason Rall: Conceptualization, Methodology, Investigation, Resources, Writing – original draft, Writing – review & editing. Craig Nowadly: Data Analysis, Writing – original draft, Writing – review & editing. Dena Norouzi: Data Analysis, Investigation, Writing – review & editing. Kaysie Sutton: Methodology, Investigation, Resources. Heang Sundermann: Methodology, Investigation, Resources. Jae-Hyek Choi: Conceptualization, Methodology, Investigation. Allyson A. Arana: Statistics. R. Madelaine Paredes: Conceptualization, Methodology, Investigation, Writing – review & editing. Joseph Maddry: Conceptualization, Funding Acquisition, Writing – review & editing. Funding and Support By JACEP Open policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The study is supported by funding from the 10.13039/100009898Defense Health Agency as a Restoral grant under project number ST1181. Data Sharing Statement Reasonable requests for access to data may be directed to the corresponding author. Declaration of Generative AI and AI-Assisted Technologies in the Writing Process An artificial intelligence–based language model (ChatGPT 5.2, OpenAI) was used as a writing support tool during manuscript preparation. The tool assisted with editing for clarity, tone, organization, and grammar, and with refining phrasing in response to detailed author input. All study concepts, analyses, interpretations, and conclusions were developed by the authors. The AI did not generate original scientific ideas, perform analyses, make independent interpretive decisions, or generate citations. All content was reviewed, edited, and approved by the authors, who take full responsibility for the manuscript. Conflicts of Interest All authors have affirmed they have no conflicts of interest to declare.