Introduction
Traumatic injuries are a leading cause of mortality worldwide, with hemorrhage being
the primary cause of preventable deaths during pre-hospital and early resuscitation
stages.[1] Hemorrhage accounts for up to 40% of preventable deaths within 24 hours of injury.[2] Coagulopathy, which affects at least one in four seriously injured trauma patients,
has a direct relationship with injury severity, massive resuscitation, transfusion,
and hemorrhagic shock. Bleeding control and hemostatic resuscitation strategies have
proven effective in reducing mortality associated with hemorrhagic injuries. Early
hemostatic interventions, including novel dressings, agents, and techniques, are crucial
for mitigating hemorrhage and improving survival.
Large animal models are used to better understand the complex processes of coagulation
in trauma and to develop therapeutic interventions. Large animal models, especially
pigs, have become essential tools for studying these processes because of their close
anatomical and physiological similarities to humans. These models offer a more realistic
representation of bleeding and clotting compared with small animal models, making
them highly relevant for advancing our understanding of trauma and hemorrhage management.
Over the past several decades, considerable research has been dedicated to studying
hemorrhagic shock models to better comprehend trauma pathophysiology and explore various
treatment options. These studies often employ isolated models, such as controlled
or uncontrolled hemorrhage, or combined injury models involving hemorrhage, bone fractures,
and/or abdominal (liver or spleen). However, large animal models face challenges,
such as ethical considerations, experimental design variability, species differences,
and lack of suitable specific laboratory tests for different species. These difficulties
hamper translating findings and therapeutic interventions from animal studies to clinical
applications. Thus, there is a need for systematic evaluation of the clinical relevance
of existing large animal models in studying hemostasis and bleeding control.
This review article explores isolated or combined trauma models in large animals,
focusing on identifying more suitable models for investigating therapeutic approaches
to prevent and treat late posttraumatic complications. We discuss the advantages,
disadvantages, and challenges of various hemorrhagic shock models across different
animal species and purposes while highlighting their potential to improve mechanistic
understanding of pathophysiology, facilitate the development of novel therapeutic
strategies, and evaluate the effectiveness of existing and emerging interventions.
Results
After applying specific inclusion and exclusion criteria, 75 relevant articles were
identified and included in this systematic review involving four different animal
species. Additionally, nine studies were identified from the reference lists of other
research sources initially missed in the search process ([Fig. 1]). The porcine model was the most commonly used, with 72 studies, followed by nonhuman
primates (NHPs) with 8 studies, and sheep and dogs with two studies each.
Fig. 1 Schematic diagram of the review process. This figure provides a visual representation
of the systematic process used in this review article, outlining the steps taken from
database search to final analysis. It showcases the rigorous methodology employed
to identify, screen, and assess the relevant studies on large animal models in trauma
and bleeding research.
The studies under review utilized a range of parameters to assess the effectiveness
of interventions on hemostasis and bleeding control. Physiological responses to interventions
were evaluated by consistently measuring blood cell count, blood gas analysis, and
hemodynamic and laboratory parameters, such as mean arterial pressure (MAP) and heart
rate. Plasma-based coagulation assays, namely prothrombin time (PT), were measured
in 54 studies and activated partial thromboplastin time (aPTT) in 48 studies. Viscoelastic
assays, including rotational thromboelastometry (ROTEM) and thromboelastography (TEG),
were extensively used in 37 and 20 studies to assess coagulation in whole blood. Coagulation
activation markers, including thrombin–antithrombin (TAT) complex and D-dimer, were
assessed in 18 studies, while prothrombin fragment 1 + 2 and fibrinopeptide A were
evaluated in 3 studies each. Fibrinogen levels and thrombin generation (TG) were also
evaluated in 27 and 14 studies, respectively, providing crucial information on TG
during massive bleeding and fibrinogen availability plus consumption for clot formation.
The calibrated automated thrombogram (CT), according to Hemker,[107] was the most frequently used method for TG measurement.[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16] Platelet function was measured in 11 studies by Multiplate,[17]
[18]
[19]
[20]
[21]
[22]
[23] flow cytometry,[3]
[24] and platelet aggregometry.[25]
[26]
The majority of studies evaluated the effectiveness of different interventions in
the management of bleeding. Thereby a reduction in blood loss served usually as the
primary endpoint, with secondary endpoints encompassing assessments of coagulation
parameters, hemodynamics, and pathological evaluations. Researchers used various types
of hemorrhage models, controlled and uncontrolled hemorrhage models,[27] such as liver injury, femur fractures with or without hemorrhage, arterial hemorrhage,
blunt chest injury with liver laceration, traumatic brain injury, and soft tissue
injury. To study the impact of contribution factors causing coagulopathy, such as
hemodilution (10 studies),[14]
[18]
[28]
[29]
[30]
[31]
[32] hypothermia (10 studies),[5]
[6]
[33]
[34]
[35]
[36]
[37]
[38]
[39] and acidosis, some of the trauma models incorporated those. Hemodilution models,
frequently used alongside other hemorrhage models like liver injury,[5]
[6]
[8]
[10]
[40]
[41]
[42] provide a means of identifying potential interventions to enhance coagulation and
minimize bleeding. Targeted interventions, such as the substitution of fibrinogen
concentrate,[8]
[42]
[43] negative pressure wound therapy,[34] and prothrombin complex concentrate (PCC),[4]
[5] have been demonstrated to impact hemodilution-induced coagulopathy.
[Tables 1] and [2] summarize hemorrhage models employed in pigs, sheep, dogs, and NHPs, with detailed
information on primary and secondary endpoints and the number of animals utilized
in each study.
Table 1
Trauma model used in pigs
|
Model classification
|
Model of trauma
|
Principal outcome[a]
|
Peripheral outcome[a]
|
Coagulation tests and markers measurements
|
(Animals in each study)
|
|
Fixed-volume hemorrhage
|
Hemorrhage (30–40%)
|
Impact on coagulation
|
–
|
PT,[48]
[103]
[104] aPTT,[48]
[103]
[104] fibrinogen,[48]
[103]
[104] TEG,[48] ROTEM,[103]
[104] TAT,[48]
[104] (ACT, TG, coagulation factors)[48]
|
(60),[103] (48),[104] (16),[47] (14)[48]
|
|
Hemorrhage (>40%)
|
Blood loss/survival
|
Impact on coagulation
|
PT,[30] PTI,[49] aPTT,[30]
[49] fibrinogen,[30]
[49] TEG,[30] ROTEM[49]
|
(20),[30] (24),[49] (15)[50]
|
|
Fixed-pressure hemorrhage
|
Arterial hemorrhage
|
Blood loss/survival rate
|
Impact on coagulation
|
PT,[31]
[53]
[58] aPTT,[31]
[53]
[58] ACT,[52] fibrinogen,[31]
[53]
[56]
[58] D-dimer,[31]
[53] ROTEM,[31] (ATIII, TAT),[53] TEG,[54] ROTEM[56]
|
(20),[52] (40),[31] (30),[53] (8),[54] (16),[55] (8),[56] (36)[58]
|
|
Fixed-volume and fixed-pressure hemorrhage
|
Hemorrhage (35%)+ MAP < 50 mm Hg
|
Metabolomic changes
|
–
|
NA
|
(12)[44]
|
|
Uncontrolled hemorrhage
|
Liver injury (grade III)
|
Blood loss/survival time
|
Impact on coagulation
|
PT,[3]
[60] aPTT,[3]
[60] ROTEM,[3]
[60] fibrinogen,[3]
[60] (TAT, D-dimer, FPA, TG, FC),[3] TEG[60]
|
(36),[3] (14)[60]
|
|
Liver injury (grade V)
|
Blood loss/survival time
|
Impact on coagulation
|
TEG[61]
|
(27),[61] (19)[62]
|
|
Blunt liver injury (not specified grade)
|
Blood loss/survival/hemostatic effect
|
Impact on coagulation
|
PT, aPTT, ACT, ROTEM, TG, fibrinogen[9]
|
(18)[9]
|
|
Hemodilution + liver injury
|
Blood loss/survival
|
Impact on coagulation
|
PT,[5]
[6]
[8]
[10]
[40]
[41]
[42] aPTT,[5]
[6]
[8]
[40]
[41]
[42] fibrinogen,[6]
[8]
[40]
[41]
[42] TAT,[5]
[6]
[8]
[10]
[42] ROTEM,[5]
[6]
[8]
[40]
[41]
[42] TG,[5]
[6]
[8]
[10] FC,[8] (D-dimer, FII, TG),[10] ATIII[5]
|
(20),[8] (18),[40] (28),[6] (32),[10] (14),[5] (9),[41] (18)[42]
|
|
Spleen injury
|
Blood loss/survival
|
Impact on coagulation
|
ROTEM, TEG[67]
|
(44)[67]
|
|
Combined uncontrolled hemorrhage and soft tissue injury
|
Blunt liver injury + femur fractures
|
Blood loss/impact on coagulation
|
Hemodynamics/survival rate/thrombin generation
|
PT,[7]
[11]
[12]
[13]
[14]
[15]
[16]
[25] aPTT,[7]
[11]
[12]
[13]
[14]
[16]
[25] ROTEM,[7]
[11]
[12]
[13]
[14]
[16]
[25] fibrinogen,[7]
[11]
[12]
[13]
[14]
[15]
[16] D-dimer,[7]
[11]
[13]
[14]
[15]
[16]
[25] TAT,[7]
[11]
[12]
[14]
[15]
[16]
[25] TG,[11]
[12]
[13]
[14]
[15]
[16]
[25] TT,[25] ACT,[7] FPA,[7]
[11] ATIII,[12]
[13]
[14]
[15] TEG[14]
|
(48),[16] (21),[25] (32),[7] (45),[11] (28),[13] (21),[12] (63),[15] (28)[14]
|
|
Combined soft-tissue injury and fixed-volume hemorrhage
|
Hemorrhage (30–50%) + femur fractures
|
Blood loss
|
Impact on coagulation
|
aPTT,[81]
[82] fibrinogen,[81]
[82] ROTEM,[81] (PT, TEG)[82]
|
(18),[81] (33)[82]
|
|
Impact on coagulation
|
–
|
PT,[83]
[84]
[85] aPTT,[83]
[84]
[85] fibrinogen,[83]
[84]
[85] TEG,[83]
[84] ROTEM[85]
|
(23),[83] (24),[84] (30)[85]
|
|
Hemorrhage (>50%) + femur fractures
|
Blood loss/survival time
|
Impact on coagulation
|
PT,[86]
[89] aPTT,[86]
[89] fibrinogen,[86]
[87]
[89] ROTEM,[86] TEG,[87]
[89] coagulation factors[87]
|
(17),[86] (21),[87] (57)[89]
|
|
Fibrinogen synthesis
|
–
|
PT, aPTT, fibrinogen, TEG[88]
|
(14)[88]
|
|
Impact on coagulation
|
Survival
|
PT, aPTT, fibrinogen, D-dimer, coagulation factors, ROTEM[32]
|
(20)[32]
|
|
Hemorrhage (30–50%) + liver injury
|
Impact on coagulation
|
Systemic inflammation
|
PT[91]
|
(24)[91]
|
|
Blood loss
|
Hemodynamic response + rebleeding
|
NA
|
(22)[65]
|
|
Femur fracture + liver injury + hemorrhage to MAP (40 mm Hg)
|
Physiologic parameters
|
Impact on coagulation
|
PT, aPTT, fibrinogen, D-dimer, ATIII, TEG[105]
|
(22)[105]
|
|
Pulmonary contusion + liver injury + hemorrhage (30–50%)
|
Blood loss
|
Impact on coagulation
|
(PT, fibrinogen, ROTEM, Multiplate),[17] TEG[92]
[93]
|
(40),[17] (36),[92] (14)[93]
|
|
Blood loss/survival
|
Liver, renal functions, and impact on coagulation
|
PT[45]
|
(30)[45]
|
|
Pulmonary contusion + liver injury + blast groin injury + hemorrhage (60%)
|
Impact on coagulation
|
Viscoelastic parameters
|
PT, aPTT, fibrinogen, D-dimer, FC, ROTEM, tPA, PAI-1, aPC, prothrombin[24]
|
(26)[24]
|
|
Femur fracture + hemorrhage (60%) + liver injury + hypothermia
|
Blood loss/survival
|
Impact on coagulation
|
TEG,[33]
[36] fibrinogen,[33]
[35] INR,[36] coagulation factors,[35] (PT, aPTT, ACT)[33]
|
(20),[36] (52),[35] (32)[33]
|
|
Inflammation + impact on coagulation
|
Organ tissue damage
|
ROTEM, ATIII, aPC, procoagulant/anticoagulant gene markers[37]
|
(24)[37]
|
|
Impact on coagulation
|
Plasma proteome
|
NA
|
(9)[38]
|
|
Combined TBI and hemorrhagic shock/soft tissue injury
|
Trauma brain injury + hemorrhage (40%)
|
Impact on coagulation
|
–
|
(sTM, ATIII, D-dimer, aPC, vWF, PAI-1, TF, PF 1 + 2),[74] (aPC, tPA, PF 1 + 2, E-selectin, ICAM),[75] (PT, TEG)[76]
|
(33),[74] (15),[75] (13)[76]
|
|
Survival rate
|
–
|
NA
|
(12)[77], (22)[78]
|
|
Platelet function
|
–
|
P-selectin,[18]
[79] TGF-β1,[18]
[79] CD40L,[18]
[79] (PECAM-1, CD61, CD62P)[79] (Multiplate, TEG)[18]
|
(10),[79] (33)[18]
|
|
Trauma brain injury + liver injury
|
Blood loss/survival rate
|
CNS pathophysiology
|
PT, PFA, TEG[80]
|
(26)[80]
|
|
Uncontrolled hemorrhage (250 mL) + trauma brain injury + hemorrhage (40%) + rib fracture
|
Platelet function
|
–
|
Multiplate, TEG, P-selectin, TGF-β1, CD40L, VCAM-1, fibrinogen[19]
|
(12)[19]
|
|
Hemodilution with or without hypothermia
|
Hemodilution
|
Blood loss/survival rate
|
Impact on coagulation
|
PT, aPTT, FVII/VIIa, TG[4]
|
(26)[4]
|
|
Plasma fibrinogen level
|
Impact on coagulation
|
PTI, aPTT, ROTEM, fibrinogen[43]
|
(12)[43]
|
|
Hemodilution + hypothermia
|
Blood loss
|
Impact on coagulation
|
PT,[34]
[39] aPTT,[34]
[39] TEG,[39] fibrinogen[34]
|
(26),[39] (38)[34]
|
Methodology abbreviations: ACT, activated clotting time; aPTT, activated partial thromboplastin
time; FC, flow cytometry; INR, international normalized ratio; Multiplate, Multiplate
analyzer; NA, not applicable for this study; PT, prothrombin time; PTI, prothrombin
time index; ROTEM, rotational thromboelastometry; TBI, traumatic brain injury; TEG,
thromboelastography; TG, thrombin generation; TT, thrombin time.
Markers abbreviations: aPC, activated protein C; ATIII, anti-thrombin III; CD40L,
CD40 ligand; FII, coagulation factor II; FVII/VIIa, coagulation factor VII/activated
factor VII; ICAM, intercellular adhesion molecule 1; PAI-1, plasminogen activator
inhibitor-1; PECAM-1, platelet endothelial cell adhesion molecule; PF 1 + 2, prothrombin
fragment 1 + 2; PFA, fibrinopeptide A; sTM, soluble thrombomodulin; TAT, thrombin-antithrombin;
TF, tissue factor; TGF-β1, transforming growth factor-β; tPA, tissue plasminogen activator;
VCAM-1, vascular cell adhesion molecule 1; vWF, von Willebrand factor.
a The principal and peripheral outcomes were not explicitly stated in all the included
studies. Therefore, the outcomes reported in this table were inferred from the results
and conclusions of the studies. Ref = References.
Table 2
Most common trauma models in sheep, dogs, and nonhuman primates
|
Model of trauma
|
Species
|
Principal outcome[a]
|
Peripheral outcome[a]
|
Coagulation tests and markers measurements
|
(Animals in each study)
|
|
Lung lobe contusions + bilateral tibial fractures + soft tissue injury in hamstring
region + 20–30% hemorrhage
|
Sheep
|
Developing an ovine model of trauma and hemorrhage
|
–
|
PT, aPTT, D-dimer, fibrinogen, Multiplate, ROTEM, FV, FVIII, aPC, PAI-1, sTM[23]
|
(12)[23]
|
|
Paranasal sinus injury + open surgical carotid injury
|
Hemostasis achievement
|
–
|
PT, aPTT, fibrinogen, D-dimer[57]
|
(14)[57]
|
|
Liver + heart injury
|
Dog
|
Hemostasis effect on liver and heart bleeding
|
Adhesive effect of study material
|
CT, coagulation factors activity[106]
|
NA[106]
|
|
Spleen injury
|
Blood loss
|
Impact on coagulation
|
Hemostatic time[68]
|
(56)[68]
|
|
Uncontrolled liver hemorrhage
|
Nonhuman primate
|
Blood loss
|
Impact on coagulation
|
PT, aPTT, ATIII, fibrinogen, D-dimer, vWF, ROTEM, Multiplate[26]
|
(16)[26]
|
|
Midline laparotomy incision + an open mid-femur fracture
|
Blood loss/survival
|
Impact on coagulation
|
PT,[20]
[21] aPTT,[20]
[21] ATIII,[20]
[21] fibrinogen,[20]
[21] D-dimer,[20] vWF,[20] ROTEM,[20]
[21]
[70]
[71] Multiplate[20]
[21]
[70]
|
(24),[20] (12),[21] (40),[70] (12)[71]
|
|
Fibrinolysis phenotypes
|
–
|
PT, aPTT, ATIII, fibrinogen, D-dimer, vWF, ROTEM[72]
|
(24)[72]
|
|
Platelet function
|
Metabolic changes
|
Multiplate[22]
|
(27)[22]
|
|
Hemorrhage (45%)
|
Inflammation C3 blockade by Compstatin Cp40
|
Impact on coagulation + pathological changes
|
PT, aPTT, ROTEM[51]
|
(8)[51]
|
Methodology abbreviations: aPTT, activated partial thromboplastin time; CT, clotting
time; Multiplate, Multiplate analyzer; NA, not applicable for this study; PT, prothrombin
time; ROTEM, rotational thromboelastometry.
Markers abbreviations: aPC, activated protein C; ATIII, antithrombin III; FV, coagulation
factor V; FVIII, coagulation factor VIII; PAI-1, plasminogen activator inhibitor-1;
sTM, soluble thrombomodulin; vWF, von Willebrand factor.
a The principal and peripheral outcomes were not explicitly stated in all the included
studies. Therefore, the outcomes reported in this table were inferred from the results
and conclusions of the studies. Ref = References.
Models of Controlled Hemorrhage
Well-defined and reproducible trauma models are crucial in preclinical research to
allow for a better understanding of the pathophysiology of hemorrhagic shock and the
identification of potential treatments. Controlled hemorrhage models like fixed-volume
and fixed-pressure ones offer high reproducibility, leading to consistent results
and facilitating inter-study comparisons. Researchers often employ a hybrid strategy,
combining pressure and volume control, for instance starting with a volume-controlled
setup (e.g., withdrawing 35% of the total blood volume) and ceasing it when the MAP
drops below a predetermined threshold.[44] Some studies, however, adopt the reverse approach.[45]
Fixed-Volume Models
The fixed-volume model is widely utilized in hemodynamic research to investigate blood
loss responses. This model involves the controlled removal of a predetermined blood
volume, with model variations reflecting differences in blood volume removal, severity,
and duration of hemorrhage. Despite its use and contribution to our understanding
of shock, metabolic changes, hemodynamic imbalances, and their effects on coagulation
and fluid resuscitation, this model has limitations. It does not completely replicate
trauma-related tissue factor release and coagulation activation and considerations
such as species-specific hypoxia tolerance. Also, inconsistent relationships between
body weight and blood volume add further complexity.
The Advanced Trauma Life Support (ATLS) defines class IV hemorrhagic shock as a situation
in which the bleeding surpasses 40% of the estimated circulating blood volume. This
severe blood loss is linked to a mortality rate exceeding 30% in humans.[46] In response to the need for effective study methods on coagulopathy and shock physiology,
researchers have extensively used pig hemorrhage models. These models, which emulate
blood loss of 30 to 40%[47]
[48] and more than 40%,[30]
[49]
[50] have been proven successful in consistently inducing hypotension and organ dysfunction,
simulating the conditions in human patients. Various studies have examined the influence
of differing fluid resuscitation strategies on coagulation in these hemorrhage models.
One significant finding was that applying lactated ringers (LR) for resuscitation
led to a substantial decrease in coagulation factors and subsequent TG, even though
hemodynamics stabilized within 2 hours. Additionally, fibrinogen concentrations and
platelet counts experienced a sharp decline after hemorrhage.[30]
[48] Interestingly, fibrinogen levels rebounded more quickly than platelet counts, highlighting
the vital role of fibrinogen in restoring hemostasis and its functionality as an acute-phase
protein.[48]
In contrast to the studies conducted using pig hemorrhage models, researchers have
also been investigating therapeutic interventions in NHP model of 45% hemorrhagic
shock.[51] Specifically, employing compstatin Cp40, a potential therapeutic agent, was able
to improve immune response, coagulation, and organ function, while also preserving
organ-barrier integrity after traumatic hemorrhagic shock.[51] This suggests that Compstatin Cp40 may have promising therapeutic potential.[51]
Fixed-Pressure Models
In the fixed-pressure model, the procedure involves constant blood extraction until
a target average arterial blood pressure is reached. This model is extensively used
in arterial hemorrhage research to evaluate the effectiveness of various hemostatic
dressings and gauze in controlling bleeding following severe trauma.[31]
[52]
[53]
[54]
[55]
[56]
[57]
[58] Numerous studies have evidenced that certain hemostatic dressings (e.g., amylopectin,
chitosan, and micronized purified polysaccharide),[52] or the direct local infusions of recombinant human activated factor VII (rFVII)[54] into a major artery, can strengthen clots. This could widen the safety margin, even
in high blood pressure cases. Innovative approaches such as modified setons infused
with procoagulant[53] and cellulose sponges[55] have proven successful in preserving the balance between fibrinogenesis and fibrinolysis[53] and managing severe noncompressible hemorrhage.[55] Hemostatic dressings that utilize self-propelling tissue technology have successfully
stopped bleeding in sheep models with minimal side effects and negligible risks of
thrombosis.[57] Another development includes the FeiChuang hemostatic gauze, which was more effective
at reducing blood loss in a gunshot wound model than Combat or standard medical gauze.[58] In contrast, certain novel hemostatic gauzes proved to be just as effective as the
current standard for controlling bleeding at the point of injury.[31]
Models of Uncontrolled Hemorrhage
Compared with controlled hemorrhagic shock models, models of uncontrolled hemorrhage
allow a more realistic clinical scenario with major trauma and uncontrolled bleeding.
These models permit unrestricted bleeding independent of hemodynamics and are often
used to investigate various fluid resuscitation strategies, assessing their effects
on animal survival, blood loss, and hemodynamic parameters.[59] However, a significant drawback of these models is the inability to control the
extent of blood loss that occurs, which could potentially limit their usefulness in
specific experimental conditions.
Liver/Spleen Trauma Models
The induction of uncontrolled hemorrhage by liver injuries represents a commonly employed
model. Typically, these injuries involve liver injuries with different degrees of
severity, ranging from grade III[60] to V[61]
[62] injuries, double liver injury,[3]
[9] or combined injuries with hemodilution[5]
[6]
[8]
[10]
[40]
[41]
[42] or femur fractures.[7]
[11]
[12]
[13]
[14]
[15]
[16]
[25]
[63] Grade III liver injuries are characterized by deep parenchymal laceration exceeding
50% of the surface area of ruptured subcapsular, resulting in moderate to severe disturbances
in coagulation parameters.[64] Similarly, grade V liver injuries are associated with a high risk of exsanguination
and increased mortality rates, necessitating timely and effective management strategies.[64]
Various methods have been developed to induce liver injuries in pigs. These methods
include controlled mechanical impact with laparotomy,[3]
[5]
[6]
[7]
[8]
[9]
[11]
[12]
[13]
[14]
[15]
[16]
[25]
[40]
[41]
[42]
[60]
[63]
[65] closed-cavity injury models,[61]
[62] and surgical knives.[10] The closed-cavity injury model has been utilized to mimic noncompressible intra-abdominal
hemorrhage,[61]
[62] while surgical knives have been used to simulate uncontrolled bleeding by removing
a portion of the liver lobe.[10]
To make animal models more applicable to studying traumatic liver injury, a laparoscopic
hepatectomy model in NHP was investigated.[66] This model allowed for observing both acute and long-term responses and compatibility
with human-derived treatments, with physiologic, metabolic, coagulation, and inflammatory
changes resembling those seen in trauma patients.[66] Despite alterations in coagulation parameters being detected, the model did not
induce clinically significant coagulopathy.[66] Additionally, the same research group in another study found the absence of thrombocytopenia
in this model, emphasizing the importance of appropriate animal model selection for
hemostatic treatment studies.[26]
Grottke et al devised a model of blunt liver injury, enabling consistent injuries
of different severities that lead to substantial alterations in all coagulation parameters,
imitating the injury and treatment stages of severe trauma, including elements that
promote coagulopathy, such as hypothermia, acidosis, and hemodilution.[41] The severity of liver injury significantly impacted coagulation parameters, with
more severe injuries leading to greater disturbances.[41] This model has been used to investigate the effectiveness of various interventions,
including direct oral anticoagulants (DOACs)-specific reversal.[3]
[9] Due to the adaption of the severity of liver injury, this model allows simulation
of the continuous bleeding in anticoagulated pigs.
Spleen injury models are utilized to investigate noncompressible hemorrhage and evaluate
the long-term effects of hemorrhage and resuscitation.[67]
[68] Noncompressible hemorrhage is a leading cause of preventable death in trauma situations;
thus, this model is particularly relevant given the spleen's susceptibility to injury
in abdominal trauma.[69] One study demonstrated that tranexamic acid (TXA) was ineffective in reducing blood
loss in a pig model despite inhibiting fibrinolysis, suggesting that the blood loss
induced by this model may be too severe to investigate only an antifibrinolytic.[67] Another study employed ultrasound visualization in dogs to develop an accurate model
of splenic artery hemorrhage and reported promising results with microwave coagulation
therapy as an alternative to thrombin injection for treating splenic hemorrhage.[68] However, the spleen injury model is nonlethal, which may limit its ability to simulate
the severity of noncompressible hemorrhage in humans.
Combined Models of Hemorrhage
Hemorrhage rarely occurs in isolation in real-world medical scenarios. Therefore,
realistic trauma models should consider factors that influence injury severity and
response, replicating human injuries with significant blood loss, hemodynamic and
metabolic changes, including shock, coagulopathy, and immune responses. Clinically,
hemorrhagic shock following major trauma often occurs with other traumatic injuries,
leading to shock-related organ dysfunction due to the release of cytokines and other
mediators.[29] Combined models incorporating multiple injury types, including substantial bleeding,
present a more accurate approach to studying trauma and evaluating therapeutic strategies.
However, the development of these large animal models presents a challenge as it requires
a careful balance between the severity of injuries and the prevention of immediate
blood loss to facilitate the study of hemostatic interventions. Hence, these models
may be particularly valuable in clinical scenarios involving trauma patients.
Liver Injury Combined with Femur Fractures
To investigate coagulation and hemostasis in response to combined traumatic insults,
a combined liver injury model with femur fractures has been successfully used in pig[7]
[11]
[12]
[13]
[14]
[15]
[16]
[25] and rhesus macaques.[20]
[21]
[22]
[70]
[71]
[72] Anticoagulated pig polytrauma models have become prevalent in simulating life-threatening
bleeding in trauma patients undergoing anticoagulation treatments. Various treatments,
such as TXA,[16] fibrinogen concentrate,[16] nonspecific hemostatic agents (PCCs[7]
[11]
[15]
[16] and activated PCC [aPCC][25]), have been investigated in these models. Moreover, several studies have utilized
this model to explore various strategies for managing massive bleeding associated
with DOACs, specifically employing specific antidotes such as idarucizumab[11]
[12] or andexanet alfa.[14] An experimental model incorporating femur fractures and double liver trauma has
been utilized to assess the efficacy of two distinct treatments in managing bleeding
within this context.[13] Moreover, this model provides a useful tool for determining optimal treatment dosages
and assessing the safety of combining medications.[13] Additionally, an NHP model of trauma has been developed to test potential therapeutic
interventions, specifically investigating trauma-induced platelet dysfunction and
acute suppression of fibrinolysis in the presence of tissue injury during hemorrhagic
shock.[70]
Trauma Brain Injury Combined with Hemorrhage and/or Tissue Injury
Traumatic brain injury (TBI), accounting for 30% of all injury-related mortalities,
frequently occurs together with hemorrhagic shock.[73] This cooccurrence immediately triggers the coagulation and complement systems, leading
to subsequent endothelial shedding, activation of protein C, and inflammatory response.[74] To investigate the effects of TBI and hemorrhage on coagulation and complement systems,
several studies have used combined models, which produced a hypocoagulable state with
reduced clot strength and thrombocytopenia.[18]
[19]
[74]
[75]
[76]
[77]
[78]
[79]
[80] The fluid percussion (employs a swift fluid injection to damage the dura),[77]
[80] controlled cortical impact (using an electromagnetic device for brain penetration),[18]
[19]
[74]
[75]
[79] rotational acceleration head injury (utilizing a pneumatic device),[78] and blast injury models[76] are fundamental models in TBI research. Each model offers unique insights into different
forms of brain injury, furthering our understanding of injury impacts and potential
treatment efficacy. Collectively, they replicate a wide range of brain injuries, from
concussive impacts to rotational forces and blast injuries, reflecting diverse real-world
scenarios. Various resuscitation strategies have demonstrated promising results in
these models. FFP resuscitation enhances platelet function and clot strength.[19] Adding valproic acid to FFP results in early upregulation of platelet activation.[79] Moreover, using hemoglobin-based oxygen carriers has improved brain oxygenation
without negatively impacting coagulation and hemodynamics.[80] Permissive hypotension has also been observed to enhance survival, hemodynamics,
and the restoration of cerebral oxygenation in severe head injuries.[77] While previous large animal models of TBI and hemorrhagic shock have primarily used
controlled blood loss models that did not fully simulate diffuse injuries following
rapid cranium acceleration or deceleration, recent studies with a closed-head, dynamic
acceleration model have revealed diverse survival rates associated with varying levels
of controlled blood loss.[78]
Hemorrhage Combined with Femur Fractures
Simultaneous femur fracture and hemorrhage can cause significant tissue damage, inflammation,
and impaired coagulation due to the release of bone marrow debris and fat into the
bloodstream, activating the coagulation system and leading to disseminated intravascular
coagulation (DIC) syndrome. Preclinical studies have primarily focused on evaluating
the combined effects of trauma and hemorrhage on coagulation impairments, clotting
status, and blood loss.[32]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88] Hemorrhage severity greatly impacts the metabolic markers,[86]
[87] physiological and coagulation responses observed in animal models, with varying
percentages used to mimic different degrees of blood loss seen in trauma patients.
Studies have found that hemorrhage and tissue injury significantly reduced arterial
blood pressure, cardiac index, hemoglobin, hematocrit, and platelet count. Coagulation
function was affected by apheresis, hemorrhage, and resuscitation with different solutions,
with a reduction in fibrinogen levels resulting in a significant decrease in clot
strength.[81]
[82]
[83]
[84] Therapy with high-dose fibrinogen concentrate and the transfusion of platelets were
found to restore coagulation function.[85] Also, resuscitation with shed whole blood or FFP has demonstrated greater efficacy
than PlasmaLyte in maintaining blood pressure and TEG maximum amplitude.[89]
Hemorrhage Combined with Liver Injury
These models simulate hypovolemia and hemodynamic instability due to hemorrhage, severe
bleeding, and coagulopathy resulting from liver injury.[90] Studies have evaluated interventions under varying degrees of hemorrhage severity
to gain insight into their potential impact. Several experimental porcine studies
have compared the effectiveness of different perihepatic packing techniques[65] and coagulation factor therapies.[91] These interventions have shown promising results in reducing blood loss,[65] improving survival,[65] and attenuating the development of acute trauma coagulopathy and systemic inflammation.[91]
Blunt Chest Injury Combined with Hemorrhage and Liver Injury
This trauma/hemorrhage model is utilized to induce endotheliopathy of trauma, which
leads to coagulopathy and multiple organ failure, including acute respiratory distress
syndrome.[17]
[24]
[45]
[92]
[93] Previous studies utilizing this model successfully induced TIC, as evidenced by
significant differences in TEG parameters[92] and principal component analysis of ROTEM.[17] PCC induced more rapid clot propagation than FFP in treating coagulopathy in pigs
but resulted in lower clot strength and a higher degree of clot lysis at later time
points, indicating a delayed consumptive coagulopathy.[92] Furthermore, studies have demonstrated that the ovine model of traumatic coagulopathy
is essential for exploring the complex interactions between hemodynamic, metabolic,
and coagulation functions in sheep.[23]
Hemorrhage Combined with Femur Fractures, Liver Injury, and Hypothermia
Animal models that simulate multiple traumatic injuries are useful for studying the
complex interactions between injuries that can exacerbate tissue ischemia, organ dysfunction,
and failure. Porcine models have been widely used to investigate the effects of various
treatments such as intravenous vitamin C administration,[37]
[38] lyophilized plasma administration,[35]
[36] and resuscitation fluid regimens[33]
[36] on inflammation,[35]
[36] coagulation function,[33]
[35]
[36] plasma proteome,[38] and end-organ histology.[37] Studies using these models have shown that high-dose vitamin C[37] and lyophilized plasma resuscitation fluids[35]
[36] have beneficial effects on coagulation function and inflammatory markers. A 1:1
ratio of plasma to red blood cells[33] and rFVIIa administration[28] have also shown promise in reducing blood loss and restoring abnormal coagulation
function in coagulopathic pigs.
Discussion
In this review, we evaluated various animal models used in trauma and bleeding research
to identify areas for improvement and provide insights into more clinically relevant
and reliable models. Essential criteria for clinically relevant trauma models include
significant tissue injury, severe bleeding, shock, and a level of trauma severity
that closely reflects real-life clinical situations.[94] Ideally, these models also consider a realistic time frame before resuscitation
is initiated.[94] However, achieving these parameters is often challenging, particularly in models
simulating severe hemorrhage. The extent of blood loss in these models can significantly
affect the survivability of the animal, and without immediate intervention, such as
the administration of blood products like red blood cells, survival duration may be
markedly limited.
Several models have been used to mimic clinical scenarios and to understand the pathophysiological
aspect of traumatic insults. Controlled hemorrhage models that use MAP as the variable
are considered more clinically relevant and reliable than fixed-volume models. Although
fixed-volume models offer the advantage of assessing compensatory hemodynamic mechanisms,
they produce variable outcomes due to the undefined degree of hypotension.[29] On the other hand, while fixed-pressure models offer better control over hypotension,
uncontrolled hemorrhage models provide a more realistic representation of real-life
situations.[29] Standardized traumatic insults have been developed to improve the reliability of
uncontrolled hemorrhage models. Effective treatment for both models should address
alterations in coagulation through standard-of-care resuscitation.
While investigating pathophysiological changes following trauma and shock, it is also
important to consider differences between species. NHP models have been shown to share
many similarities with human physiology in critical areas such as pharmacokinetics,
pharmacodynamics, immunology, genetics, and hemodynamics. Also, Tarandovskiy et al
found that humans, baboons, and rhesus macaques had the most similar simultaneous
thrombin and plasmin generation assay parameter values compared with other species.[95] Despite this, only a small percentage (10%) of studies reviewed here used NHPs as
research subjects due to ethical considerations, animal welfare concerns, costs, and
availability.
Sheep have been identified as a valuable animal model for coagulation and hemostasis,
given their physiological similarities to humans in routine and certain coagulation
tests.[96] These similarities include comparable cardiorespiratory and hemostatic functions,
as evidenced by similar results in tests such as PT, aPTT, fibrinogen assays, as well
as ROTEM parameters.[23]
[96]
[97] Sheep models have been used successfully to mimic the trauma-induced coagulopathy
(TIC) and involvement of the activated protein C (aPC) pathway, characterized by hyperfibrinolysis
and depletion of factor V.[23] However; TG data show differences in endogenous TG potential between sheep and humans.
Sheep exhibit markedly accelerated TG kinetics compared with humans, likely due to
their higher sensitivity to the human tissue factor used in the assay.[97] Sheep exhibit distinct differences in secondary hemostasis compared with humans.
These include the rapid initiation of the contact activation pathway, elevated levels
of factor VIII, low levels of protein C, increased clot firmness, and reduced capacity
for clot lysis.[96] Limitations of using sheep as a model included cost, size, and the lack of breed-specific
reference ranges in coagulation. Although dogs are one of the species most similar
to humans in coagulation, especially in the acceleration of coagulation, there are
still significant differences between dogs and humans, particularly in the extrinsic
activation of coagulation.[98] Moreover, using dogs in experimental studies is restricted due to ethical concerns
about their status as companion animals.
Pigs are the predominant animal model used in hemostasis and coagulation research,
accounting for 85% of studies in this review. Pigs remain a popular choice in preclinical
research due to their similarities to humans in organ size, blood volume, some functional
coagulation proteins, and hemodynamic response.[29]
[59]
[94] In the context of TIC, research involving pigs has highlighted the critical role
of aPC and its response following TBI and hemorrhage.[75] These findings suggest aPC operates as a compensatory mechanism in response to activated
coagulation in TIC and inflammatory pathways, thereby affirming the relevance of pigs
in coagulation research.[74] However, inducing coagulopathy in pigs can be challenging. These arise from differences
in compensatory mechanisms, variations in vasopressor receptors, and differences in
inflammatory and immunological responses between pigs and humans.[29] Some studies have introduced hemodilution prior to injury to establish a standardized
coagulopathy.[41] However, this approach might not accurately represent what occurs in real clinical
situations, especially when permissive hypotension is part of the treatment process.
Compared with humans, pigs have shown a lower TG potential[95]
[99] and maximum clot lysis,[99] pointing to the differences between coagulation pathways of humans and pigs. Results
from ROTEM demonstrate variability in the coagulation state of the pig model as a
hypercoagulable species.[97]
[100] These variations depend on the specific test (EXTEM[100] or NATEM[97]; considering clotting time, maximum clot firmness (MCF), and clot formation time),
and other experimental factors such as pig breed, age, and the anesthesia protocol
used.[99]
[100] Furthermore, FibTEM testing has shown that pigs have higher fibrinogen levels with
less platelet contribution to clot strength compared with humans, leading to a higher
MCF.[99] Regarding the similarities of fibrinolytic pathways between pigs and humans, thromboelastographic
findings suggest a pronounced delay in clot lysis in porcine whole blood compared
with human blood.[101] However, other studies have noted similar D-dimer reference intervals in both species,
indicating comparable fibrinolysis.[100] When utilizing pigs as experimental models for fibrinolysis studies, researchers
must be cautious due to the species propensity for thrombus formation in smaller vessels
and arteries, as well as specific markers measured in the study. Moreover, variations
in study outcomes underscore the necessity for meticulous methodological considerations
when comparing results across studies. For instance, differing tissue factor concentrations
used in TG assays may complicate comparisons.
Large animal models allow researchers to conduct controlled experimental studies,
enabling them to adapt the experimental variables and assess interventions with high
precision. Nevertheless, it is crucial to recognize the limitations associated with
translating findings from animal models to real-world clinical scenarios. Animal models
cannot fully address genetic variability, environmental effects, and other crucial
aspects of human biology.[102] Improving animal models is crucial for understanding human traumatic injuries and
investigating therapeutic interventions. Current models primarily focus on the initial
response to injury, which is insufficient to capture the full clinical picture. As
complications may emerge later, more extended observation periods are needed to evaluate
treatment impacts during intensive care. The essential use of anesthesia in experimental
models can potentially obscure the natural physiological stress responses to hemorrhage
and resuscitation, such as effects on sensorimotor functions, cardiovascular actions,
and metabolic demands.[28] Early use of mechanical ventilation with positive pressure could alleviate the progression
of lung failure in models featuring experimental chest trauma.[27] These factors should be considered when evaluating the outcomes of such studies.
Further investigation, including the importance of genetic differences and data integration
from various models, is needed to address the complex pathophysiology of trauma and
hemorrhage.
Additionally, ethical considerations of animal research should also be taken into
account. Alternative approaches, such as advanced imaging, multiomics analysis, in
vitro models, computer simulations, and humanized animal models, can offer valuable
insights while ensuring animal welfare. While we should continue to improve and develop
these alternatives, the use of large animal models remains essential for certain areas
of biomedical research. Integrating these approaches with animal models and clinical
data allows for more comprehensive and robust development of effective treatments.
With the advances in animal welfare and ethical standards, we should aim to refine,
reduce, and replace (the 3Rs principle) the use of animal models where possible and
justified.
