Background
Traumatic brain injury (TBI) and uncontrolled hemorrhage are the leading causes of death in multiple injured trauma patients [
1‐
4] and the coincidence of TBI and hemorrhage after trauma has been associated with an even worse overall morbidity and mortality [
5‐
7]. Several studies have shown that superimposed TBI to hemorrhage, e.g. the addition of a traumatic impact to the brain on top of hemorrhage, may result in impaired cardiovascular compensation [
8‐
12] and a reduced ability to modulate vascular tone [
13]. More recent studies have demonstrated disturbances in the autonomic response to hemorrhage if TBI is present [
8] and that TBI may attenuate or delay the physiological bradycardic and hypotensive response to acute extracranial blood loss [
9]. Additionally, the presence of TBI has been associated with an early increase in baroreflex sensitivity, another key component of cardiovascular hemostasis [
10].
In order to maintain tissue perfusion and to preserve physiological hemodynamics, an adequate function of the peripheral microcirculation within the different organ systems is fundamental [
14]. Extracranial hemorrhage is known to be associated with microcirculatory hypoperfusion and subsequent tissue damage by inducing vasoconstriction and decreasing arterial blood flow [
15‐
17]. However, this microcirculatory response to extracranial hemorrhage when coincidenced with a superimposed TBI, as frequently encountered in the clinical setting, is yet to be defined. Apart from directly inflicting and damaging the brain, experimental TBI itself has also been shown affect systemic circulation. Several studies have demonstrated that various models of TBI may create an acute and transient systemic hypertension and bradycardia which is followed by a hypodynamic state with heterogeneous hypoperfusion among organs at increasing magnitudes of injury [
18,
19].
In the present study, we have combined two well-established and characterized experimental models, i.e. i.) a pressure-controlled hemorrhagic shock model with ii.) a standardized and clinically relevant model of experimental TBI, to assess vasomotor responses and microcirculatory changes in third-order arterioles and venules of a spinotrapezius muscle preparation by using intravital microscopy.
Discussion
The mortality rates observed in the present study after infliction of both impacts paralleled the mortality rates reported by McMahon and colleagues when using a combined fixed-volume model of hemorrhagic shock combined with TBI in which 50% of all animals died within the first 90 minutes after the completion of hemorrhage [
9]. In the clinical scenario, hemorrhagic shock and TBI have frequently been reported to account for approximately 50% of all trauma-related deaths within the first 24 hours after hospital admission [
3,
4].
Previous reports have indicated that hemorrhagic shock may result in a progressive vasoconstriction of arterioles [
15‐
17]. Accordingly, in the present study acute blood loss during pressure-controlled hemorrhagic shock resulted in a sharp and progressive decrease in the diameter of the third-order arterioles and venules both, at initiation and up to 60 minutes of shock. Interestingly, when TBI was added this constriction was also evident but substantially less pronounced. Obviously, the superimposed TBI counteracted the vasoconstriction from pressure-controlled hemorrhage alone. Law and colleagues have reported similar results when assessing macrocirculatory changes in rats suffering from combined TBI and hemorrhage [
13]. These authors could demonstrate that TBI may lead to a reduced ability to modify vascular tone. In detail, brain-injured rats were able to reduce aortal conductance (arterial blood flow/MAP) once hemorrhage was initiated, but they were not able to maintain this vasoconstriction. Furthermore, the response to fluid resuscitation was limited as no relevant changes in aortal conductance were observed. This observation was supported by Yuan and colleagues who demonstrated that TBI may suppress spontaneous hemodynamic recovery from hemorrhage and also impede the efficacy of fluid resuscitation [
12]. In contrast to the present study, the studies by Law and colleagues [
13] were focused on macrocirculatory changes as reflected by aortic blood flow (ABF) and vascular conductance (ABF/MAP) and on shorter periods of shock, e.g. 30 minutes. The results presented by Yuan and colleagues [
12] were focused on cardiovascular responses after TBI and hemorrhage including heart rate, pre- and afterload as well as stroke volumes and indices.
In the present study, resuscitation efforts in shocked but also TBI-injured animals resulted equivocally only in a moderate change in vessel diameter, but simultaneously in a sharp increase in RBC velocity and flow rate. In this context, it is of note that MAP levels were comparable between both groups (HS only and TBI + HS) and therefore could not be made responsible for these microcirculatory changes. Likewise, there were no differences in Hb, pH and PaCO2 between the experimental groups observed on repeated blood gas analysis at baseline and during the entire experiment. Therefore, potential acidosis and hypercapnia could also not be made responsible for changes in velocities and vessel diameters.
One possible explanation for the observed changes in vasomotor response and microcirculation may be related to altered sympathetic nervous activity. Traumatic brain injury has been associated with an immediate sympathetic activation as a result of increasing intracranial pressure thus triggering a corresponding increase of circulating catecholamines within minutes after trauma [
40‐
42]. This sympathoadrenal activation with its massive release of vasoactive substances leads to an initial hypertension and an increase in vascular tone. The initial increase in blood pressure shortly after TBI was also observed during our experiments and is confirmed by previous studies [
13,
18,
19].
The idea that there may be sympathetic vasodilatator nerves to skeletal muscles is an old concept fitting with the “fight or flight” model of the sympathetic nervous system [
43]. The first evidence for vasodilatator nerves to skeletal muscles emerged when stimulation of skeletal brain stem areas was shown to evoke hypertension, tachycardia and skeletal muscle vasodilatation in terms of the so-called “defense reaction” and that these dilator nerves were cholinergic. Matsukawa and colleagues assessed internal diameter changes of arterial vessels of skeletal muscles evoked by activation of sympathetic cholinergic nerve fibers during stimulation of the hypothalamic defense area in cats. The authors observed a dilatation in small arteries ranging from 100 to 500 μm, but not in larger extramuscular arteries [
44]. More recently, the skeletal muscle dilator response to sympathoexitatory maneuvers in both, humans and animals, appears to be nitric oxid (NO)-dependent. Joyner and Diek concluded from their review, that most “sympathetic dilator” responses in human muscle are due to adrenaline or local cholinergic mechanisms acting to stimulate NO release from the vascular endothelium [
43].
On the other hand side, Nagai and Pleschka have previously demonstrated circumscribed brain stem sites to mediate adrenergic and non-adrenergic vasoresponses in dogs by electrically probing selected brain stem points to determine from which vasodilatation or vasoconstriction of the lingual and intraorbital arteries could be elicited [
45]. Vasodilatation in these vessels by stimulation could be induced from an area extending parasagittally through the ventral part of the brain stem from the hypothalamus to the upper pontine region. The most potent area was in the supraoptic area of the anterior hypothalamus, but no representation was found in the dorsolateral part of the central gray matter, one of the most excitable mesencephalic vasodilatator area in the defense reaction. Centrally elicited vasodilatation of the tongue has been shown to be due to non-adrenergic, non-cholinergic efferents with pre- and postganglionic synapses running apart from the cervical spine trunk. The anatomical origin of these efferentes suggests that they may belong to the parasympathethic section of the autonomic nervous system [
45]. It is endogenous to the LFP model to induce bilateral diffuse white matter damage and severely-graded injury to the brain stem including the hypothalamus and pontine regions. In the approach here, the authors have inflicted a TBI of moderate severity and it may be assumed that these triggering areas may have been affected (Figure
1).
The effect of the sympathetic system on vascular tone and microcirculation may also be vessel-diameter dependent. For example, significant hypotension with a mean MAP of 40 mmHg has been shown to induce a reflex response mediated by sympathetic nerves to cause constriction of larger arterioles (70–150 μm) [
46]. In small arterioles, however, both, constriction and dilatation, have been observed in response to the same level of hypotension [
15,
16]. As previously described, Matsukawa and colleagues described dilatation in smaller arteries (100–150 μm), but not in larger extramuscular arteries after hypothalamic defense area stimulation in cats [
44]. The underlying mechanism by which hemorrhage induces dilation of smaller arterioles is not fully understood. Humoral factors may play a pivotal role as epinephrine has been shown to induce vasodilatation of small arterioles (<40 μm) by acting through β2-adrenoreceptors [
47].
Traumatic brain injury-associated cytokines have also been suggested to promote microcirculatory effects. A whole variety of inflammatory mediators and cytokines are typically released into the systemic circulation after both extra- and intracranial trauma [
48‐
51]. In the cremaster muscle of rats, TNF-α has been shown to modulate arteriolar diameters via an immediate vasodilatory effect [
52]. Interestingly, in this experimental setting only third- and fourth-order arterioles were immediately dilated after administration of TNF-α. In contrast, larger vessels did not show any alteration in diameter under this condition [
52,
53]. It has been speculated that different characteristics of TNF-dependent receptors may mediate this vasodilatation in third-order arterioles, while larger vessels remain unaffected.
Similarly, Interleukin (IL-)1 and IL-6 have frequently been implicated in the decreased systemic vascular resistance in different forms of shock [
54]. By also using a cremaster in-vivo preparation in rats, Minghini and colleagues have reported a dose-dependent increase in third-order arteriolar diameters from 11% to 51% with 1 hour in-vivo exposure to IL-1 in increments of 0.01, 0.1, 1.0 to 20 ng/ml. Cytokine washout resulted in arteriolar return to baseline diameter. In vivo exposure to IL-6 (10, 50 and 250 U/ml) for 1 hour yielded similar results, but after washout arteriolar dilatation persisted. These data indicate that both cytokines are potent dilating agents for skeletal muscle resistance vessels under in-vivo conditions.
In the present analysis, resuscitation efforts resulted in an immediate and sharp increase of RBC velocity and arterial flow rate, particularly in TBI + HS animals where an overshooting response to resuscitation was observed. A previous study using the same model of pressure-controlled hemorrhagic shock demonstrated that higher RBC velocities and flow rates after resuscitation were associated with a better survival rate [
17]. At first glance, the present results seem contradictory as TBI + HS animals displayed a higher mortality rate despite high arterial blood flow. However, the induced hypoperfusion during the 60 minutes of hemorrhagic shock may result in acute tissue ischemia. It has been shown that the sudden rise in oxygen at the onset of reperfusion is associated with oxidative stress which itself may lead to cellular and organ damage [
55,
56]. Therefore, a gradual introduction of oxygen during resuscitation seems to be beneficial [
57]. It can be speculated, that the overshooting flow rate, as observed in TBI + HS animals, will lead to a faster reperfusion and therefore may be associated with increased organ damage. Nevertheless, it has to be acknowledged that combined animals had sustained a high magnitude of injury in general.
Limitations
Several limitations of this study have to be acknowledged. In the present analysis, we have focused on selected third-order vessels in a spinotrapezius muscle preparation only. These vessels were selected at random, so we cannot exclude some degree of selection bias. Since we have not assessed other vessel sizes or vascular beds, our results are restricted and interpretation needs caution. The tissue of interest from the critical care perspective is certainly the gastrointestinal tract with secondary complications resulting from mucosal hypoxia and hypoperfusion. Splanchnic ischemia with an intramucosal pH < 7.3 has commonly been reported after isolated TBI [
58] and different forms of shock including haemorrhage [
59]. Further studies are necessary to analyze the effects of TBI and haemorrhage on the hepatosplanchnic, renal, pulmonary and cardiac microcirculation, all which persist to fail following hemorrhagic shock and may account for the higher overall mortality rate in the combined group.
Furthermore, the present study has focused on the first 60 minutes after resuscitation only. Longer observation periods are warranted to elucidate changes in vessel diameter, RBC velocity and flow rate beyond these time points and to assess chronic effects. Increased levels of vasoactive substances (for example catecholamines, nitric oxide (NO)) as well as trauma-associated circulating mediators and cytokines have been speculated as one possible explanation for the observed differential vasomotor and microcirculatory responses after pressure-controlled hemorrhagic shock with and without superimposed TBI. Further investigations are needed to assess the precise role of these substances in the context of hemorrhagic shock and TBI.
The choice of anesthesia, e.g. sodium pentobarbital, used in the present experiment may have also introduced confounding factors. This agent has been shown to cause metabolic suppression via hypoperfusion and a reduction in cerebral blood flow (CBF) due to flow-metabolism coupling. Pentobarbital has relatively minor effects on cardiac output, arterial pressure and total peripheral resistance, but more important effects on left ventricular function and myocardial contractility [
60,
61]. For instance, 15 minutes after intravenous pentobarbital, cardiac output, arterial pressure, and total peripheral resistance were all essentially at the pre-anesthetic control levels, but stroke volume was reduced, as was myocardial contractility and the velocity of myocardial fiber shortening [
60]. Although this would have affected all investigated animals and the 15 minutes stabilization time may have abolished the effect, it could have been influencing the reactions of those animals subjected to combined TBI and hemorrhage in particular due to the disturbance in circulation caused by the experimental model. Kawaue and Iriuchijima have reported a drop in arterial pressure acutely from an average value of about 105 mmHg to a minimum of about 75 mmHg in about 5 min which then gradually recovered to an average level of about 90 mmHg in 30 min after intravenous injection of pentobarbital sodium [
62]. It has to be recognized that pentobarbital in these experiments was administered intervenously while in our study the agent was administered intraperitoneally.
Another limitation of the present analysis is that the experiments have been restricted to the measurement of hemodynamics only. With regard to TBI, behavioural measurements as well as histopathological analyses would have been an interesting addition and will surely be an integral component of our future studies in this field of interest.
Lastly, the findings described here represent our first observations and remain therefore entirely descriptive. The authors are aware that none of the potential mechanisms behind have been explored in detail yet. Additional experiments to close this gap via pharmacological inhibition experiments (for example NO-blocking), cytokine and catecholamine measurements and/or studies in vessels of different diameter are necessary and planned.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
BC, MMae and QH designed the study. BC and MMae carried out the experiments and analyzed the data obtained. BC and MMu performed the statistical analysis. MMu, QH and MMae drafted the manuscript. EN conceived of the study, participated in its design and coordination, and helped to draft the manuscript. All authors have read and approved the manuscript in its final version prior to submission.