Comparative efficacy of fluid resuscitation strategies in a rat model of traumatic cardiac arrest Authors: Xue Liu, Jingru Li, Jianjie Wang, Bihua Chen, Yongqin Li Published: 2026 ABSTRACT Highlights•A rat model combining hemorrhage and asphyxia mimics TCA pathophysiology.•Hemorrhage exceeding 30% with asphyxia requires fluid resuscitation after CPR.•Whole blood yields the best outcomes, while colloid offers a comparable survival.•Lactated Ringer’s and normal saline prolong survival, but do not improve outcomes.Keywords: Traumatic cardiac arrest, Fluid resuscitation, Resource-limited settings, Whole blood, Hydroxyethyl starch Introduction Traumatic cardiac arrest (TCA) remains a major clinical challenge in acute care, with survival rates of only 2–5%.1, 2, 3 Unlike non-traumatic cardiac arrest (NTCA), where primary cardiac events predominate, TCA typically arises from severe hypovolemia secondary to hemorrhage,4 which is often compounded by respiratory compromise resulting from airway obstruction, maxillofacial or tracheal injuries, or tension pneumothorax.5, 6, 7 The combination of hypovolemia and hypoxia creates a lethal synergy that distinguishes TCA from other forms of cardiac arrest. Due to the unique pathophysiology of TCA, resuscitation strategies differ significantly from those for NTCA. Blood transfusion is pivotal for hypovolemia.8 But in resource-limited settings such as disaster zones, battlefields, or remote areas, prehospital storage and transport of blood products are often infeasible.9 In some countries, less than 1% of civilian systems have implemented prehospital blood transfusion plans.8 Therefore, fluid resuscitation with readily available volume expanders is prevalent.10 Although current guidelines recommend cardiopulmonary resuscitation (CPR) combined with fluid resuscitation for major trauma treatment, how to choose fluid types remains unclear, as no systematic study has provided direct evidence on the impact of different fluids on resuscitation outcomes.7, 11, 12, 13 Animal models are essential tools for investigating resuscitation strategies in TCA. However, most existing models only focus on hemorrhage and fail to capture the exacerbated injury resulting from the combination of hemorrhage and hypoxia.14, 15, 16 Conversely, models that induce cardiac arrest with ventricular fibrillation (VF) primarily represent a cardiac event that is uncommon in trauma.16, 17, 18 Consequently, there is currently no established animal model that can effectively simulate the combined pathophysiological effects of massive hemorrhage and respiratory compromise. To address this critical issue, we developed a rat TCA model combining controlled hemorrhage and subsequent asphyxia. Using this model, we aimed to compare the effects of four commonly used fluids, normal saline (NS), lactated Ringer's solution (LR), hydroxyethyl starch (HES) and whole blood (WB), on post-resuscitation outcomes. We hypothesized that different fluids would differentially improve survival and neurological function while attenuating brain and cardiac injury, with WB potentially being the most effective. These findings aim to provide evidence-based guidance for fluid resuscitation in resource-limited settings. Methods Ethical statement This study was approved by the Laboratory Animal Welfare and Ethics Committee of the Army Medical University (AMUWEC20223378). The study complied with the ARRIVE guidelines and all animals received humane care in compliance with the Principles of Laboratory Animal Care and Guide for the Care and Use of Laboratory Animals. Experimental animals A total of 132 male Sprague-Dawley rats (250–350 g) were sourced from the Laboratory Animal Center at the Army Medical University. 28 rats were used to study the relationship between blood loss and cardiac arrest. Another 28 rats were used to determine the hemorrhage threshold requiring fluid resuscitation in the combined model. The remaining 76 rats were assigned to the final experiment to study the effect of fluid resuscitation. Of these, 40 were used for evaluating resuscitation outcomes including 96-h survival, neurological recovery and pathology; 24 for blood gas and biomarker analysis; 12 as SHAM-operated controls (6 for pathology, 6 for biomarkers). Detailed group allocations are described in Sections ‘Experimental procedures’ and ‘Measurements’. Animals were assigned to groups by block randomization (block size = 8, 2 animals/group/block) to control for batch effects, using sequential drawing of lots without replacement. Animal preparation Animals were group-housed with wood shavings bedding under standard conditions (22.0 ± 2.0 °C; 50 ± 10% humidity; 12-h light/dark cycle) with ad libitum chow and water. After an acclimatization of at least 3 days, rats were fasted overnight with free access to water.19 Anesthesia was induced by a single intraperitoneal injection of 2% pentobarbital (50 mg/kg; dissolved in sterile saline) and maintained with intravenous supplements (10 mg/kg) as required. The animals were placed supine. As illustrated in Fig. 1A, subcutaneous needles were inserted for electroencephalogram (EEG) recording, and limb electrodes for electrocardiogram (ECG) recording. A 14-gauge cannula was inserted into the trachea for mechanical ventilation. The right femoral artery and left femoral vein were cannulated with PE-50 catheters for arterial pressure monitoring and fluid/medication administration. The catheters were flushed intermittently with heparinized saline (2.5 IU/mL). Animals received mechanical ventilation (0.65 mL/100 g, 65 breaths/min; RWD Life Science, Shenzhen, China) when spontaneous breathing was insufficient (<50 breaths/min, irregular breathing, heart rate <300 bpm, or discontinuous EEG). Esophageal temperature was maintained at 37.0 ± 0.5 °C using a heating lamp throughout preparation, hemorrhage, asphyxia and post-resuscitation. Animals equilibrated for at least 10 min post-surgery, and baseline data were collected once hemodynamic parameters stabilized (∼1 h). A priori exclusion criteria included death from anesthesia overdose or surgical errors, failure to achieve target blood loss within the prescribed time, or death before fluid resuscitation. Predefined humane endpoints for intensive care included resuscitation failure, persistent hypotension (mean arterial pressure (MAP) < 20 mmHg for 20 min), or severe respiratory distress during spontaneous breathing (e.g., persistent cyanosis, gasping, or apnea > 2 min), leading to euthanasia via pentobarbital (150 mg/kg). Defining the relationship between hemorrhage volume and shock/cardiac arrest. (A) Schematic illustration of the preparation of hemorrhage and asphyxial cardiac arrest in rats. ECG: electrocardiogram; EEG: electroencephalogram; MAP: mean arterial pressure. (B) Examples of ECG, blood pressure (BP) and EEG values after hemorrhage in four groups. (C-F) Measurements of heart rate (HR) (C), mean arterial pressure (D), cardiac output (CO) (E) and left ventricular ejection fraction (LVEF) (F) before and after resuscitation. (G) Weighted-permutation entropy analysis of EEG. (H) Kaplan-Meier survival curve used for cumulative overall survival analysis during the 96-h observation period. H0: control without hemorrhage; H15: 15% hemorrhage; H30: 30% hemorrhage; H40: 40% hemorrhage. BL: baseline. n = 7 animals/group, with LVEF and CO measured in duplicate. *P < 0.05 and **P < 0.01 compared with H0; †P < 0.05 and ††P < 0.01 compared with H40. Experimental procedures Defining the relationship between hemorrhage volume and shock/cardiac arrest To characterize the relationship between graded blood loss and shock or cardiac arrest severity in hemorrhage-only models, animals were randomized into four groups (n = 7): control without hemorrhage (H0); 15% blood loss (H15); 30% blood loss (H30); and 40% blood loss (H40).20 Groups 2–4 underwent blood removal from the right femoral artery at 1 mL/min using an infusion pump (Longer Precision Pump, Baoding, China), with the total blood volume estimated as 64 mL/kg.21 Shock was defined as a MAP of less than 65 mmHg for 10 min.22, 23 The animals were monitored for 2 h. Intermittent normal saline boluses (0.5 mL/30 min, ∼3 mL/kg/h) were given for fluid maintenance and catheter patency. Animals in group H40 were monitored until death or for six hours to record ECG during cardiac arrest. Determining the lethal threshold in a combined hemorrhage-asphyxia TCA model Based on the graded blood loss parameters, this experiment aimed to establish a combined hemorrhage-asphyxia model and determine the lethal threshold requiring fluid resuscitation. The animals were randomized into four groups (n = 7): asphyxial cardiac arrest (ACA) without hemorrhage as control (H0 + AS); 15% hemorrhage and asphyxia (H15 + AS); 30% hemorrhage and asphyxia (H30 + AS); and 40% hemorrhage and asphyxia (H40 + AS). Following hemorrhage, asphyxia was induced by clamping the endotracheal tube 30 s after the administration of pipecuronium bromide (0.2 mg/kg).24 CA was confirmed by a MAP of less than 20 mmHg. After 5 min of untreated CA, ventilation (0.65 mL/100 g, 90 breaths/min) and manual CPR with 100% O2 were initiated. Two-finger chest compressions at 240/min and one-third anteroposterior depth maintained a MAP above 20 mmHg. Epinephrine (0.03 mg/kg) was administered intravenously 30 s after CPR initiation and repeated every 2 min (maximum of 3 doses) until the return of spontaneous circulation (ROSC). A single 2 J defibrillation (Zoll Medical, Chelmsford, MA, USA) was attempted if the rhythm was shockable. CPR was continued for up to 5 min. ROSC was defined as a supraventricular rhythm with a MAP greater than 50 mmHg for more than 5 min. Animals received intensive care for 5 h (0.65 mL/100 g, 65 breaths/min for 3–4 h; 100% O2 for 2 h, then 21% O2; fluid maintenance as in Section ‘Defining the relationship between hemorrhage volume and shock/cardiac arrest’). Data from H40 + AS beyond 180 min were excluded due to gradual death. Comparing fluid resuscitation strategies following CPR The 30% hemorrhage + asphyxia model was used to study the effects of fluid resuscitation. The animals were randomized into four groups (n = 10): normal saline as control (NaCl 9.0 g/L, HA-NS); lactated Ringer’s solution (C3H5NaO3 3.1 g/L, NaCl 6.0 g/L, KCl 0.3 g/L, CaCl2·2H2O 0.2 g/L, HA-LR); hydroxyethyl starch (hydroxyethyl starch 130/0.4 60.0 g/L, NaCl 9.0 g/L, HA-HES); and whole blood (heparin sodium 2 U/mL, HA-WB). Hemorrhage, CA induction and CPR were performed as in Sections ‘Defining the relationship between hemorrhage volume and shock/cardiac arrest’ and ‘Determining the lethal threshold in a combined hemorrhage-asphyxia TCA model’. After ROSC, a volume of fluid (20–25 °C) equal to blood loss was administered via the left femoral vein at 0.5 mL/min (Longer Precision Pump, Baoding, China). Intensive care was as in Section ‘Determining the lethal threshold in a combined hemorrhage-asphyxia TCA model’. Post-operative monitoring and care Following catheter removal, extubation and suturing, animals were returned to cages for 96-h observation. Prophylactic antibiotics were omitted due to aseptic technique. Analgesics were withheld to avoid confounding resuscitation outcomes.25 Clinical checks assessing respiratory status, wound integrity and signs of complications were performed thrice daily. Interventions included clearing secretions and re-suturing wounds. Mobility-impaired rats received 5% glucose (1–2 mL, 2–3 times/day) orally and chow placed on the cage floor until recovery. Measurements The primary outcome was 96-h survival. The secondary outcomes were neurological recovery and serum levels of cardiac troponin I (cTn-I) and S100 calcium-binding protein B (S100B). Arterial pressure, ECG and EEG were continuously recorded via a data acquisition system (DATAQ Instruments, Akron, OH, USA). Left ventricular ejection fraction (LVEF) and cardiac output (CO) were assessed noninvasively using an echocardiograph (Shantou Ultrasonic Electronics, Shantou, China). LVEF and CO were measured in duplicate at each time point and averaged. Weighted-permutation entropy (WPE) of the EEG was analyzed as previously described,26 with data normalized to [0,1] using (x − xmin)/(xmax − xmin). Neurological deficit scores (NDS) were assessed by blinded investigators at 24, 48, 72 and 96 h after resuscitation.27 The NDS evaluates consciousness (100), respiration (100), cornea reflex (40), gag reflex (30), auditory reflex (30), righting reflex (100) and behavior (100), where 0 represents normal and 500 represents moribund/death. Survival was monitored for 96 h, with death (cessation of breathing and heartbeat > 2 min) recorded as the experimental endpoint.28 After 96-h observation, surviving and sham rats were anesthetized with pentobarbital (150 mg/kg) and transcardial perfused with cold phosphate-buffered saline (pH 7.4). Hearts and brains were harvested, fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned (4 µm) as described.24 Myocardial sections were stained with hematoxylin and eosin (H&E) and Masson’s trichrome to evaluate morphology and fibrosis. Brain sections were stained with H&E and Fluoro-Jade B (FJB) combined with 4′,6-diamidino-2-phenylindole (DAPI) to assess morphology and neuronal degeneration. All staining reagents were obtained from Servicebio (Wuhan, China). Images of the left ventricle (10 fields per sample) and the hippocampal mid-dorsal CA1 region (2 fields per sample) were captured at ×400 magnification. The collagen volume fraction and the proportion of FJB-positive cells were quantified by blinded investigators using ImageJ (National Institutes of Health, Bethesda, MD, USA). Data from the same sample were averaged. Following the procedure in Section ‘Comparing fluid resuscitation strategies following CPR’, 24 rats (6 per group) and 6 sham-operated rats were anesthetized with pentobarbital (150 mg/kg) 5 h after ROSC or sham surgery, respectively. Blood gases (0.1 mL per sample) were measured at baseline and 5, 30, 120, 300 min after resuscitation (Abbott Point of Care, Abbott Park, IL, USA). Sampling was omitted during the hemorrhage period to prevent hemodynamic instability right before CA induction. Arterial blood samples were obtained 5 h after ROSC prior to catheter removal, and from the sham-operated rats. Serum cTn-I29 and S100B30 were measured using an enzyme-linked immunosorbent assay (ELISA kit, Elabscience Biotechnology, Wuhan, China) according to the manufacturer’s instructions. Each sample was assayed in duplicate and averaged. Statistical analysis The sample size was calculated using the Lakatos normal approximation method based on the Log-rank test for survival analysis (StatBox, AI-Driven Clinical Trials, Shijiazhuang, China). Instead of the calculated n = 5, n = 7 was adopted for study ‘Determining the lethal threshold in a combined hemorrhage-asphyxia TCA model’ to ensure robust hemodynamic data and consistency with ‘Defining the relationship between hemorrhage volume and shock/cardiac arrest’. The normality was tested using the Kolmogorov-Smirnov test. Normal data were presented as mean ± SD, whereas non-normal data as median including interquartile ranges (IQRs). A one-way analysis of variance (ANOVA) with Bonferroni correction was employed for the physiological variables and biochemical markers. Non-parametric NDS analysis used the Kruskal-Wallis with Bonferroni correction. Survival was analyzed using the Kaplan-Meier with Log-rank test. P < 0.05 was considered statistically significant. All analyses were performed using SPSS (IBM, Armonk, NY, USA). Results As all animals survived for at least half an hour after hemorrhage or CPR. No animals met the priori exclusion criteria and none were excluded. Severe hemorrhage induces shock and cardiac arrest We first examined the correlation between hemorrhage volume and cardiac arrest in the absence of asphyxia (Fig. 1). Baseline characteristics were similar among the groups (Table 1). Dynamic changes in physiological signals are shown in Fig. 1B. Hemodynamic parameters are presented in Fig. 1C–G. Shock (MAP < 65 mmHg) occurred in both H30 and H40 groups. In H30, 5/7 animals developed shock 10 min after blood loss (59.4 ± 10.6 mmHg), then MAP recovered to 69.1 ± 12.4 mmHg by 30 min. In H40, all animals developed shock, with MAP remaining below 65 mmHg throughout the 2-h observation (51.9 ± 6.7 mmHg at 30 min, 55.5 ± 14.3 mmHg at 120 min). Overall, hemorrhage exceeding 30% induced shock in 86% of animals. Table 1Baseline characteristics for assessing hemorrhage volume and cardiac arrest, determining the hemorrhage threshold requiring fluid resuscitation in the combined model and comparing fluid resuscitation strategies.VariableAssessing hemorrhage volume and cardiac arrestDetermining the hemorrhage threshold requiring fluid resuscitationComparing fluid resuscitation strategiesH0H15H30H40H0 + ASH15 + ASH30 + ASH40 + ASHA-NSHA-LRHA-HESHA-WBBody weight, g293.2 ± 16.4305.7 ± 39.5295.7 ± 41.7303.6 ± 23.3299.7 ± 40.3306.5 ± 15.1307.7 ± 28.9302.1 ± 14.8296.4 ± 24.2307.1 ± 29.9302.4 ± 19.3301.3 ± 21.5Temperature, ℃37.1 ± 0.437.1 ± 0.337.1 ± 0.337.0 ± 0.336.9 ± 0.637.0 ± 0.237.1 ± 0.437.1 ± 0.237.1 ± 0.337.0 ± 0.337.0 ± 0.336.9 ± 0.4Heart rate, bpm420.3 ± 31.7416.2 ± 56.4414.1 ± 15.1413.0 ± 30.6421.6 ± 20.4412.0 ± 28.8392.1 ± 22.4403.3 ± 23.2403.4 ± 30.9392.6 ± 39.6426.6 ± 39.0409.3 ± 25.2MAP, mmHg118.7 ± 15.5115.4 ± 17.4126.9 ± 7.1125.0 ± 7.6126.7 ± 10.4127.1 ± 3.5120.4 ± 6.7128.9 ± 7.2125.3 ± 8.6124.4 ± 6.2130.4 ± 4.2128.9 ± 8.4CO, L/min0.26 ± 0.020.25 ± 0.050.24 ± 0.030.24 ± 0.030.22 ± 0.020.28 ± 0.040.22 ± 0.020.25 ± 0.020.20 ± 0.030.22 ± 0.020.22 ± 0.030.22 ± 0.02LVEF, %79.7 ± 2.281.6 ± 4.280.7 ± 3.680.1 ± 2.884.0 ± 1.580.7 ± 1.982.5 ± 3.481.1 ± 1.583.8 ± 2.982.1 ± 3.881.8 ± 2.383.7 ± 3.3Blood loss, mL–3.0 ± 0.45.7 ± 0.87.9 ± 0.6–3.0 ± 0.25.9 ± 0.67.7 ± 0.45.7 ± 0.55.9 ± 0.65.8 ± 0.45.8 ± 0.4CA induction, s––––173.5 ± 22.8146.4 ± 16.3116.9 ± 26.3**91.7 ± 18.3**121.9 ± 24.7109.0 ± 27.3105.8 ± 40.2128.4 ± 12.5CPR duration, s––––45.3 ± 4.744.3 ± 5.354.7 ± 7.951.7 ± 10.955.0 ± 12.851.6 ± 6.955.3 ± 13.253.5 ± 11.3Epinephrine, μg––––9.3 ± 1.210.0 ± 0.09.5 ± 0.910.0 ± 0.010.0 ± 2.010.3 ± 0.810.8 ± 3.310.5 ± 1.6MAP: mean arterial pressure; CO: cardiac output; LVEF: left ventricular ejection fraction; CA: cardiac arrest; CPR: cardiopulmonary resuscitation. H0: control without hemorrhage; H15: 15% hemorrhage; H30: 30% hemorrhage; H40: 40% hemorrhage; H0 + AS: asphyxial cardiac arrest without hemorrhage; H15 + AS: 15% hemorrhage and asphyxia; H30 + AS: 30% hemorrhage and asphyxia; H40 + AS: 40% hemorrhage and asphyxia; HA-NS: Traumatic cardiac arrest (TCA) with normal saline resuscitation; HA-LR: TCA with lactate Ringer’s resuscitation; HA-HES: TCA with hydroxyethyl starch resuscitation; HA-WB: TCA with whole blood resuscitation. n = 7 animals for groups in determination of hemorrhage volume and establishment of the TCA model; n = 10 animals for groups in evaluation of fluid resuscitation strategies.**P < 0.01 compared with H0 + AS. Baseline characteristics for assessing hemorrhage volume and cardiac arrest, determining the hemorrhage threshold requiring fluid resuscitation in the combined model and comparing fluid resuscitation strategies. MAP: mean arterial pressure; CO: cardiac output; LVEF: left ventricular ejection fraction; CA: cardiac arrest; CPR: cardiopulmonary resuscitation. H0: control without hemorrhage; H15: 15% hemorrhage; H30: 30% hemorrhage; H40: 40% hemorrhage; H0 + AS: asphyxial cardiac arrest without hemorrhage; H15 + AS: 15% hemorrhage and asphyxia; H30 + AS: 30% hemorrhage and asphyxia; H40 + AS: 40% hemorrhage and asphyxia; HA-NS: Traumatic cardiac arrest (TCA) with normal saline resuscitation; HA-LR: TCA with lactate Ringer’s resuscitation; HA-HES: TCA with hydroxyethyl starch resuscitation; HA-WB: TCA with whole blood resuscitation. n = 7 animals for groups in determination of hemorrhage volume and establishment of the TCA model; n = 10 animals for groups in evaluation of fluid resuscitation strategies. P < 0.01 compared with H0 + AS. All H0, H15 and H30 animals survived for 96 h, whereas only two H40 animals did (Fig. 1H). The cumulative 96-h survival was significantly lower in H40 than in H0 (29% vs. 100%, P = 0.007). In H40, cardiac arrest occurred 7.5 ± 7.8 h after blood loss. Pulseless electrical activity (PEA) was observed in four animals that died within six hours (see Fig. 1B for example). 30% blood loss as the lethal threshold requiring fluid resuscitation in the combined hemorrhage-asphyxia model Asphyxia was induced following hemorrhage to mimic combined injuries (Fig. 2A). Baseline parameters were similar among the groups (Table 1). CA was successfully induced in all animals, and ROSC rates were 100% in all groups, with no significant differences in epinephrine dosage or CPR duration. The CA induction time was significantly shorter in H30 + AS and H40 + AS than in H0 + AS (Table 1). Determining the hemorrhage threshold requiring fluid resuscitation in the combined model. (A) The scheme of experimental procedures for determination of the hemorrhage threshold requiring fluid resuscitation. CPR: cardiopulmonary resuscitation; ROSC: return of spontaneous circulation. (B) An example of dynamic changes of ECG, BP and EEG values before and after TCA in the H30 + AS group. ECG: electrocardiogram; BP: blood pressure; EEG: electroencephalogram; TCA: traumatic cardiac arrest; red arrows: the time point to clamp the trachea. (C-F) Measurements of heart rate (HR) (C), mean arterial pressure (MAP) (D), cardiac output (CO) (E) and left ventricular ejection fraction (LVEF) (F) before and after resuscitation. (G) Weighted-permutation entropy analysis of EEG. (H) Kaplan-Meier survival curve used for cumulative overall survival analysis during the 96-h observation period. H0 + AS: asphyxial cardiac arrest without hemorrhage; H15 + AS: 15% hemorrhage and asphyxia; H30 + AS: 30% hemorrhage and asphyxia; H40 + AS: 40% hemorrhage and asphyxia. BL: baseline. n = 7 animals/group, with LVEF and CO measured in duplicate. *P < 0.05 and **P < 0.01 compared with H0 + AS; †P < 0.05 and ††P < 0.01 compared with H40 + AS. Cardiac arrest was confirmed by an isoelectric ECG, a MAP below 20 mmHg, and the disappearance of EEG activity (Fig. 2B). Fig. 2C–F shows the hemodynamic parameters. Compared with H0 + AS, H30 + AS and H40 + AS had significantly lower MAP and CO during recovery (Fig. 2D, E). LVEF was higher in H30 + AS at 300 min and in H40 + AS at 30 min (Fig. 2F). WPE of EEG was lower in H30 + AS and H40 + AS after resuscitation (Fig. 2G). All H0 + AS and H15 + AS animals survived for 96 h. Only one H30 + AS animal survived, and none in H40 + AS survived past 5 h (Fig. 2H). The median survival times were 96.0 [96.0, 96.0], 96.0 [96.0, 96.0], 7.0 [5.6, 18.5] and 1.0 [0.7, 2.1] h, respectively. Cumulative 96-h survival was significantly lower in H30 + AS (14% vs. 100%, P = 0.001) and H40 + AS (0% vs. 100%, P < 0.001) than in H0 + AS. Survival was significantly higher in H15 + AS (100% vs. 0%, P < 0.001) and H30 + AS (14% vs. 0%, P < 0.001) than in H40 + AS. Thus, in the combined hemorrhage-asphyxia model, 30% blood loss represented the lethal threshold, requiring fluid resuscitation in addition to CPR. Subsequent experiments were based on this model. Fluid resuscitation after CPR improves outcomes, with WB and HES being most effective To improve survival, we tested the resuscitation effects of different fluids in the 30% hemorrhage + asphyxia model (Fig. 3A). Baseline parameters were similar among the groups (Table 1). CA was successfully induced in all animals, and the ROSC rate was 100%. Comparing fluid resuscitation strategies following CPR. (A) The scheme of experimental procedures for the fluid resuscitation strategies. TCA: taumatic cardiac arrest; CPR: cardiopulmonary resuscitation; ROSC: return of spontaneous circulation. (B–E) Measurements of heart rate (HR) (B), mean arterial pressure (MAP) (C), cardiac output (CO) (D) and left ventricular ejection fraction (LVEF) (E) before and after resuscitation. (F) Normalized weighted-permutation entropy analysis of electroencephalogram (EEG). (G–H) Blood gas measurements of HCO3– (G) and lactate (H). (I–J) The serum concentration of cardiac troponin I (cTn-I) (I) and S100 calcium-binding protein B (S100B) (J). (K) Total neurological deficit score during the 96-h observation period. (L) Kaplan-Meier survival curve used for cumulative overall survival analysis during the 96-h observation period. HA-NS: Traumatic cardiac arrest (TCA) with normal saline resuscitation; HA-LR: TCA with lactate Ringer’s resuscitation; HA-HES: TCA with hydroxyethyl starch resuscitation; HA-WB: TCA with whole blood resuscitation; SHAM: sham-operated group. BL: baseline. n = 10 animals/group for B-F, K-L and 6 animals/group for G-J, with LVEF, CO, cTn-I and S100B measured in duplicate. *P < 0.05 and **P < 0.01 compared with HA-NS; †P < 0.05 and ††P < 0.01 compared with HA-WB. Only lactate and bicarbonate are shown, as pH and base excess were highly correlated with bicarbonate, and oxygenation parameters were primarily influenced by the oxygen intervention. The hemodynamic parameters are shown in Fig. 3B–E. Compared with HA-NS, HA-WB had a significantly higher MAP during recovery (Fig. 3C), while HA-HES and HA-WB had a significantly higher CO (Fig. 3D). LVEF was lower in HA-HES at 240 min and in HA-WB at 300 min post-resuscitation (Fig. 3E). WPE did not differ among the groups (Fig. 3F). Compared with HA-WB, HA-LR had a lower MAP during recovery (Fig. 3C), a decreased CO before 180 min (Fig. 3D) and a higher LVEF at 300 min (Fig. 3E). Fig. 3G and H shows the blood gas measurements. Compared with HA-NS, HCO3– was significantly higher at 30 and 300 min in HA-LR, HA-HES and HA-WB (Fig. 3G). Lactate was significantly lower in HA-LR, HA-HES and HA-WB during recovery, except for HA-LR at 300 min (Fig. 3H). The NDS and survival data are shown in Fig. 3K and L. HA-WB animals had significantly lower NDS scores than HA-NS animals at 48, 72 and 96 h (144 [123, 173] vs. 410 [236, 500], P = 0.041; 101 [77, 149] vs. 500 [238, 500], P = 0.044; 78 [65, 119] vs. 500 [221, 500], P = 0.031). HA-HES showed lower scores than HA-NS, but not significantly (Fig. 3K). The numbers of animals surviving for 96 h was 3, 6, 8 and 9 in HA-NS, HA-LR, HA-HES and HA-WB, respectively (Fig. 3L). The median survival times were 34.5 [6.5, 87.3], 96.0 [11.4, 96.0], 96.0 [96.0, 96.0] and 96.0 [96.0, 96.0] h, respectively. Cumulative 96-h survival was significantly higher in HA-HES (80% vs. 30%, P = 0.027) and HA-WB (90% vs. 30%, P = 0.004) than in HA-NS. Cardiac and cerebral injuries are shown in Figs. 3I, J and 4. Compared with HA-NS, HA-LR, HA-HES and HA-WB all had lower cTn-I and S100B levels (Fig. 3I, J). Consistent with H&E staining, Masson staining showed significantly higher collagen deposition in HA-NS compared to HA-LR, HA-HES and HA-WB (Fig. 4A, C). Hippocampal CA1 neurons were highly susceptible to TCA, with HA-WB showing significantly attenuated neuronal degeneration relative to HA-NS in FJB staining (Fig. 4B, D). Histologically stained myocardial tissue and hippocampal CA1 neurons at 96 h after resuscitation. (A) Representative micrographs of H&E- and Masson-stained myocardium. (B) Representative micrographs of H&E-, DAPI- and FJB-stained hippocampal CA1 region. (C) Collagen volume fraction (10 quantified fields/animal). (D) The proportion of FJB-positive cells (2 quantified fields/animal). H&E: hematoxylin and eosin; DAPI: 4′,6-diamidino-2-phenylindole; FJB: Fluoro-Jade B; HA-NS: Traumatic cardiac arrest (TCA) with normal saline resuscitation; HA-LR: TCA with lactate Ringer’s resuscitation; HA-HES: TCA with hydroxyethyl starch resuscitation; HA-WB: TCA with whole blood resuscitation; SHAM: sham-operated group. n = 3 animals for HA-NS, 6 animals for HA-LR, 8 animals for HA-HES, 9 animals for HA-WB and 6 animals for SHAM. *P < 0.05 and **P < 0.01 compared with HA-NS; ††P < 0.01 compared with HA-WB. When comparing the fluid resuscitation groups with the non-resuscitation group (H30 + AS), no significant differences were found between HA-NS and H30 + AS except for WPE. The differences between HA-NS and the other groups (HA-LR, HA-HES and HA-WB) were consistent with those between H30 + AS and the other groups. The WPE in H30 + AS was lower than in all four fluid resuscitation groups 120 min after resuscitation. Discussion Using the established rat TCA model combining hemorrhage and asphyxia, we demonstrate that fluid resuscitation is necessary when blood loss exceeds 30%. Of the four different fluids compared, WB yielded the best outcomes. HES was comparably effective despite lower blood pressure. LR and NS extended survival but did not improve outcomes. In TCA, hemorrhage accounts for approximately 48% of deaths, while respiratory compromise contributes to 10–20%.4, 5, 6 These insults often coexist, as polytrauma accounts for 27% of major trauma.5 However, existing TCA models are limited in that hemorrhage-only or hemorrhage-with-organ-damage models may produce ‘pseudo-PEA’ with residual cardiac output,14, 16, 31, 32 whereas VF following hemorrhage mimics arrhythmias not typical of trauma.17, 18, 33 While polytrauma involves tissue injury, our model isolates the lethal synergy of hypovolemia and hypoxia to capture true arrest.5, 7 Based on 30% blood loss (class III hemorrhage in humans20), 5-min asphyxia, and only one animal surviving, this model represents a moderate-to-severe TCA insult. Using this model, we found that 30% blood loss which was tolerated by hemorrhage-only animals (100% survival; Fig. 1H), became lethal when combined with asphyxia (14% survival; Fig. 2H). The high ROSC rate with a low 96-h survival of 14% mirrors clinical TCA, where ROSC is often achievable, but long-term survival is below 10%.1, 34 As resuscitation guidelines vary, with the TCCC advising against the use of crystalloids and colloids,13 while the European guidelines recommend blood products and crystalloids,11, 12 a systematic comparison is necessary. Survival increased in the order of NS, LR, HES and WB (Fig. 3L). NS lacks both oncotic pressure and oxygen-carrying capacity,35 resulting in poor outcomes. LR improves blood gas via buffering36, 37 but fails to restore hemodynamics, though survivors exhibited better late-stage neurological function (Fig. 3K). HES provides oncotic pressure, thereby improving cardiac output.38 However, its lower viscosity compared to WB impairs microvascular perfusion, leading to a lower blood pressure (Fig. 3C).39 Despite high variability, HES showed a narrower NDS IQR than crystalloids (Fig. 3K), indicating more consistent efficacy. WB corrects coagulopathy and oxygen deficit to ameliorate cardiovascular and metabolic dysfunction,40 yielding the best survival. Notably, WB achieved superior neurological recovery with a smaller IQR at 72–96 h than HES (Fig. 3K), indicating more predictable efficacy. This consistent neuroprotection, particularly in consciousness and behavior, is vital as it directly links resuscitation quality to patient-centered outcomes, ensuring better long-term cognitive function and quality of life. The clinical relevance of these findings is evident in resource-limited settings where WB is often unavailable.9 In such environments, HES can serve as a bridging therapy to maintain perfusion until blood products become available. Although LR and NS are less effective, they prolong survival, allowing time for evacuation and transfusion. This study provides direct evidence to guide fluid choices in austere TCA care. This study has several limitations. Firstly, analgesics were withheld because opioids cause respiratory depression,41 while many agents (NSAIDs, opioids, aspirin and acetaminophen) possess neuroprotective effects.42, 43, 44, 45 These factors would confound survival and neurological recovery assessments.25 Consequently, humane endpoints were not applied due to deaths between checks and the statistical bias of informative censoring on survival curves.46 To mitigate potential distress, we provided intensive supportive care as detailed in Section ‘Post-operative monitoring and care’. Secondly, 30% blood loss is severe20 but lower than exsanguination (40%),16, 32 which may underestimate catastrophic collapse. Thirdly, the resuscitation protocol was limited to a 1:1 vol ratio and immediate infusion. Future studies should be performed to evaluate fluid volume and timing. Fourthly, we did not directly explore the mechanisms underlying the differential effects of fluids. Finally, the use of only male rats limits generalizability to females, despite the male predominance in trauma epidemiology. Conclusions In this rat TCA model, fluid resuscitation was required when blood loss exceeded 30%, a lower threshold than in hemorrhage-only models. WB produced the best outcomes. HES was comparably effective, albeit with lower blood pressure. LR and NS extended survival, enabling prolonged care. These findings support the management of major trauma in resource-limited settings. CRediT authorship contribution statement Xue Liu: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Jingru Li: Writing – original draft, Visualization, Investigation, Formal analysis. Jianjie Wang: Investigation, Formal analysis. Bihua Chen: Writing – original draft, Investigation, Funding acquisition, Conceptualization. Yongqin Li: Writing – review & editing, Writing – original draft, Visualization, Supervision, Methodology, Funding acquisition, Conceptualization. Funding This work was supported by the National Natural Science Foundation of China (62271490) and the Natural Science Foundation Project of Chongqing (CSTB2025NSCQ-GPX0670, CSTC2021JCYJ-MSXMX0521). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Contributor Information Bihua Chen, Email: windice1201@tmmu.edu.cn. Yongqin Li, Email: lyq@tmmu.edu.cn. References