Optimal fluid resuscitation strategies in patients with hemorrhagic shock (HS) remain controversial, although both crystalloids (1) and fresh frozen plasma (FFP) remain key components of both pre and in-hospital resuscitation strategies. Maintaining end-organ oxygenation by ensuring adequate tissue perfusion pressures as well as reducing edema formation by minimizing fluid extravasation is key to the resuscitation success, regardless of fluid choice.

While the choice of resuscitation fluid has been the subject of intense investigation [1],[2], little is known of the effect resuscitation speed. The Advanced Trauma Life Support (ATLS) guidelines place emphasis on early restoration of adequate tissue perfusion by rapid infusion of crystalloids [3], thus theoretically minimizing the time to adequate end organ oxygenation. In contrast, several studies have indicated that rapid bolus infusion of crystalloids and artificial colloids may be associated with lower post resuscitation systemic blood pressures and higher mortality compared with slow or stepwise infusion in general trauma [4]–[6]. If higher infusion speeds are indeed associated with higher levels of fluid extravasation, this may be particularly detrimental if HS is combined with traumatic brain injury (TBI), owing to the confined nature of the brain in the cranial cavity, the susceptibility to hypoxia and the well established detrimental effects of the associated intracranial pressure (ICP) changes [7].

Using a large animal model of combined TBI and HS, our group has previously demonstrated that early bolus resuscitation with FFP attenuates both lesion size and fluid extravasation into the brain compared to bolus resuscitation with 0.9% normal saline (NS). Using the same model, we now hypothesize that stepwise infusion speed increment resuscitation with FFP or NS will be superior to bolus in reducing brain lesion size and edema formation.

All the research was conducted in compliance with the Animal Welfare Act and other Federal statutes and regulations relating to animals and experiments involving animals. The study adhered to the principles stated in the Guide for the Care and Use of Laboratory Animals, Institute for Laboratory Animal Research (1996) and was approved by the appropriate Institutional Animal Care and Use Committees. All the procedures were performed under the supervision of a veterinarian.

A total 23 Female Yorkshire swine (40–50 kg; Tufts Veterinary School, Grafton, MA) were used and allowed three days for acclimatization prior to surgery. Of these, 20 animals were primarily used for and included in other studies [8],[9] with an identical protocol while the remaining three animals were included for the purpose of the present study. The aims of these other studies were to compare the effects of different resuscitation fluids on brain injury. The present study thus presents secondary use of animal experiments previously published as well as novel data from animals not previously published.

Food was withheld the night before surgery, but access to water was allowed. Preanesthesia was administered with an intramuscular injection of Tiletamine/Zolazepam (Telazol, 50 mg/ml), 8 mg/kg (Fort Dodge Animal Health, Fort Dodge IA) and atropine sulfate 1.5 mg. Animals were weighed and anesthesia was subsequently induced with inhalation of 4% inspired fraction isoflurane in 100% oxygen. Animals were intubated with a 7.0 mm cuffed endotracheal tube and put on ventilator support (Narkomed-M, North American Dr\ager, Telford PA) with a tidal volume of 10 ml/kg, peak pressure of 20 cm H₂0 and a respiratory rate of 10 breaths per minute. No supplemental oxygen was administered following intubation. Tidal volumes and respiratory rate was adjusted to maintain a target end tidal PCO₂ of 40 mmHg. Isoflurane was adjusted to 1%-3% inspired fraction for maintenance of anesthesia.

Hemorrhage and resuscitation protocol

Model outline is provided in Figure 1. Total blood volume was estimated, and 40% of it was withdrawn through the femoral arterial catheter using a Masterflex pump, Model L/S Computerized Drive with a MF easy load II Pump-head, Model 77201–60 (Cole-Palmer, Vernon Hills, IL). Bleeding was started concurrent with TBI at a rate of 3.15% total blood volume/min and was captured in a Terumo blood collection bag (CPDA and AS-5). Isoflurane was decreased with the onset of hypotension. If MAP dropped < 30 mmHg, hemorrhage was briefly held and a small volume of saline was infused through the femoral venous line. Once the MAP reached 35 mmHg, saline infusion was stopped and hemorrhage was restarted. Using this protocol, MAP was maintained between 30 mmHg and 35 mm Hg until 40% of the estimated blood volume was withdrawn in a controlled fashion. Following hemorrhage, animals were left in shock for 120 minutes and MAP was maintained between 30–35 mmHg by titrating the dose of inhaled isoflurane. After the 2 hours of shock, animals were resuscitated as follows: Prior to the experiment, animals had been randomly assigned to 1 of 4 resuscitation groups (Figure 1): 1) bolus infusion with FFP 1*shed blood volume (n = 5) at 50 ml/min, 2) Stepwise infusion of FFP 1*shed blood with infusion speeds starting at 2 ml/min and gradually increasing to 50 ml/min (n = 6), 3) stepwise infusion of NS*3 shed blood volume with infusion speeds staring at 6 ml/min and gradually increasing to 165 ml/min (n = 6) and 4) bolus infusion of NS *3 shed blood volume at 165 ml/min (n = 5). Bolus resuscitation protocols were based on previous experiments [9]. The stepwise resuscitation protocols were the result of several pilot experiments suggesting an optimal hemodynamic response while minimizing resuscitation time and the extent of the brain injury. Fluids were infused into the femoral vein using a Masterflex pump. Animals were monitored for six hours post resuscitation and were kept warm (Bair Hugger Model 505; Arizant Healthcare Inc., Eden Praire, MN). Electrolytes were corrected as needed. Following the six-hour observation period, animals were euthanized by intravenous injection of sodium pentobarbital 100 mg/kg (Euthasol, Virbac Corp, Fort Worth, TX).

An arterial blood gas was analyzed at baseline, following 2 hours of shock and the 6-hour observation period.

In this study we found that bolus resuscitation with both FFP and NS resulted in adverse outcome compared with stepwise resuscitation. In the NS group, bolus resuscitation was associated with a significant increase in both brain swelling and lesion size, despite the fact that peripheral oxygenation (spO₂) was higher in the bolus resuscitation group. Furthermore, a significant decrease in cardiac output coupled with an increase in systemic vascular resistance suggests that increased fluid extravasation may have compromised the resuscitative outcome.

In the FFP groups, these changes were mirrored although less pronounced. Bolus resuscitation was associated with increased fluid extravasation in the brain as well higher systemic vascular resistance, higher hemoglobin levels (suggesting less hemodilution and thus increased fluid extravasation) and lower cardiac output. These effects did, however, not appear to impact greatly on brain lesion size.

Overall, these findings are in line with previous reports suggesting that bolus resuscitation following hemorrhagic shock may be associated with adverse outcome [4]–[6]. Indeed, bolus infusion of crystalloids have been shown to increase bleeding and shorten survival times in a rodent model of hemorrhage following splenic injury compared to slow infusion [4]. Furthermore, while brisk infusion of normal saline may results in fluid extravasation in excess of 100 ml/min in humans, doubling the infusion time while maintaining equal fluid volume may result in a four time reduction of fluid extravasation rates, thus improving the volume load over time [11]. These results furthermore support the growing body of evidence suggesting favorable outcome when low volume resuscitation is used [12]–[14], but does raise the question of whether the beneficial effects are due to low volume or low infusion/perfusion pressures, or a combination.

Although this study was not designed for investigating the underlying pathophysiological mechanisms, some potential explanations may be considered. Following the classic line of thinking proposed by Starling [15], the greater increase in hydrostatic pressure during bolus resuscitation may drive a net fluid shift out of the vascular bed as oncotic pressures remain equal due to the equal amounts of identical fluids used. This is supported by findings indicating less early fluid extravasation in anesthetized compared to non-anesthetized patients receiving equal crystalloid volume loads [16]–[18], presumably due to vascular relaxation and concomitant lower perfusion pressures rather than choice of anesthetic agent [19]. In contrast, the degree of fluid extravasation following FFP infusion is largely unknown, but results from similar experiments using artificial colloids suggest that fluids extravasation may actually be increased during anesthesia [20]. The increased colloid osmotic pressure of FFP compared with NS may, however, counteract extravasation in the FFP groups with a resulting smaller difference in the hemodynamic response between bolus and slow resuscitation groups. Indeed, a differential effect on colloid osmotic pressure between crystalloids and artificial colloids has been demonstrated [21].

Recent evidence have challenged the classic Starling approach indicating a pivotal role of the semi-permeable layer formed by the endothelial glycocalyx [22]. This barrier coats the vascular endothelium and is comprised of membrane bound proteoglycans, glycoproteins and plasma proteins comprising a non-circulating intravascular layer of an estimated 700–1.500 ml [23],[24]. The integrity of the subglycolayx layer as well as the starling forces operating herein, more than intra and extravascular colloid osmotic and hydrostatic pressures seem determine net fluid shifts [25],[26]. It therefore follows that an intact and functional endothelial glycocalyx is pivotal for optimal resuscitative outcome.

Interestingly, volume loading with a concomitant increased stretch of the atrial walls release atrial natriuretic peptide (ANP), which directly induce shedding of the endothelial glycocalyx [27],[28]. Rapid reconstitution of the vascular volume may thus induce increased atrial stretching and ANP release, directly inducing glycococalyx shedding and increased fluid extravasation. Of note, recent evidence suggests that FFP may protect or even reconstitute the endothelial glycocalyx [29],[30], which may account for the less pronounced fluid extravasation observed in the FFP groups in this study.

Several limitations exist in this study and deserve to be acknowledged. We have chosen a 6-hour post-resuscitation observation time, which makes us unable to conclude on the long-term consequences of different resuscitation strategies. Indeed, the minor changes observed in ICP suggests that the observed brain swelling is likely not of clinical significance at this time point. The differential ICP and brain oxygenation trajectories evident from Figure 4 does, however, suggest that differences would only be exaggerated if the observation phase was extended.

The 6-hour observation time point does not allow us to conclude on the effects on clinically relevant outcome parameters such as neurofunctional status. The focus on brain lesion size, an end-point parameter of clear clinical relevance, still allows us to conclude on the clinical relevance of the differences observed in this model. A long-term survival model has been developed to address these important questions.

In line with this, the presented data does not allow us to conclude on the mechanistics underlying the observed differences. These issues are, however, the object of future more focused studies.

Secondly, neither NS nor FFP are usually employed as stand-alone resuscitative agents in the treatment of HS, bringing the clinical translatability of the model into question. Indeed, the use of packed red blood cells (PRBC’s) and platelets is standard of care in resuscitation of HS. We opted not to include PRBC’s and platelets in the resuscitation protocols since we wanted to examine the isolated effects of different resuscitation fluids rather than introduce the potential confounding effects of multiple fluids with unknown interactions. Indeed, the isolated effect of PRBC or platelet infusion speed on hemodynamics is unknown.

Third, the choice of bolus infusion speeds may be debatable. Indeed, we used maximum NS infusion speeds of 165/ml and FFP infusion speeds of 50 ml/min, which may appear higher than what is commonly used. It is, however, important to note that gravity alone can produce infusion crystalloid rates of 123 ml/min [31], pressure bags 257 ml/min [31] while rapid infusers such as the Level 1 (Level 1 technologies, Rockland, MA) and Rapid infusion system (Heamonetics Corp, Braintree, MA) can deliver infusion rates in excess of 800 ml/min [32].

In conclusion, in this study we found that bolus resuscitation with both NS and FFP following TBI and shock was associated with increased brain swelling in both groups as well as increased lesion size in the NS groups when compared to stepwise resuscitation. If these findings translate to humans it is concerning that widely used protocols such as the ATLS [3] still call for rapid infusion of crystalloids as the initial resuscitative adjunct in hypotensive patients.