Effects of volume resuscitation on the microcirculation in animal models of lipopolysaccharide sepsis: a systematic review

Background Recent research has identified an increased rate of mortality associated with fluid bolus therapy for severe sepsis and septic shock, but the mechanisms are still not well understood. Fluid resuscitation therapy administered for sepsis and septic shock targets restoration of the macro-circulation, but the pathogenesis of sepsis is complex and includes microcirculatory dysfunction. Objective The objective of the study is to systematically review data comparing the effects of different types of fluid resuscitation on the microcirculation in clinically relevant animal models of lipopolysaccharide-induced sepsis. Methods A structured search of PubMed/MEDLINE and EMBASE for relevant publications from 1 January 1990 to 31 December 2015 was performed, in accordance with PRISMA guidelines. Results The number of published papers on sepsis and the microcirculation has increased steadily over the last 25 years. We identified 11 experimental animal studies comparing the effects of different fluid resuscitation regimens on the microcirculation. Heterogeneity precluded any meta-analysis. Conclusions Few animal model studies have been published comparing the microcirculatory effects of different types of fluid resuscitation for sepsis and septic shock. Biologically relevant animal model studies remain necessary to enhance understanding regarding the mechanisms by which fluid resuscitation affects the microcirculation and to facilitate the transfer of basic science discoveries to clinical applications.


Introduction
Sepsis is a syndrome induced by infection that is characterized by physiologic, pathologic, and biochemical abnormalities [1], thus causing a substantial primary disease burden and co-morbidity [2]. Treatment guidelines globally recommend correction of haemodynamic abnormalities via the rapid administration of fluid boluses and blood transfusion to

Intensive Care Medicine Experimental
Doppler-shifted reflected light [45]. Sidestream dark-field (SDF) imaging technology is based on improved orthogonal polarized spectral (OPS) imaging, whereby concentrically placed light-emitting diodes (LEDs) provide pulsed synchronous illumination of the microcirculation at a central wavelength of 530 nm for optimal absorption by the haemoglobin in red blood cells, independent of the oxygenation state [38]. Volume resuscitation for treatment of septic shock has been the standard of care recommended by guidelines following evidence of reduction in persistent hypovolaemia and improved survival [7,9,46,47]. These guidelines have been mainly based on observational studies and expert opinion, in the absence of supportive evidence from randomized controlled trials. Consequently, different types of fluid are used in resuscitation including crystalloids, colloids and blood transfusion.
Generally, there is a paucity of animal model studies on effects of fluid resuscitation on the microcirculation. A comparison of crystalloid and colloid in cecal ligation and puncture-induced sepsis in rodents concluded that crystalloid infusion produced better microcirculatory function as well as mortality benefit [48], with growing interest in comparison of fluid types [49,50] and early versus late resuscitation in experimental animal models [51,52]. Clinically, there is still controversy on the choice of resuscitation fluids. The trial on Early Goal Directed Therapy (EGDT) for treatment of sepsis and septic shock [47] was found to have no mortality benefit in subsequent trials [53][54][55]. However, lower mortality in septic patients resuscitated with albumin-containing solutions had been reported in a meta-analysis [56] but there was no significant mortality benefit of albumin compared to saline in the Fluid Expansion as Supportive Therapy (FEAST) randomized controlled trial [57]. A Cochrane meta-analysis comparing colloid to crystalloid fluid resuscitation, in a heterogeneous group of critically-ill patients, failed to demonstrate any tangible benefits on mortality [58]. Emerging research data highlights an urgent need to review volume resuscitation strategies and adjunct vasopressor use in sepsis and septic shock.
Mechanistic research is therefore essential to define cause and effect. However, whilst animal research is advocated for investigating mechanisms of disease process and testing therapeutic interventions, discordant comparisons of treatment effects between animal experiments and clinical trials have arisen, due to the failure of animal models to adequately mimic clinical disease [59].
The aim of this review was to identify clinically relevant experimental animal models that have been used to assess the microcirculatory effects of fluid resuscitation treatment for lipopolysaccharide (LPS)-induced sepsis. We therefore examined the evidence available on the effects of fluid resuscitation on the microcirculation in LPSinduced sepsis and septic shock.

Search strategy
A systematic search was conducted in two indexed online databases-PubMed/MED-LINE and EMBASE-for articles describing the microcirculation in sepsis and septic shock, published from 1 January 1990, through 31 December 2015, in accordance with PRISMA guidelines [60]. In PubMed/MEDLINE, the search terms used in [MeSH Terms] or [All Fields] were 'sepsis' OR 'septicaemia' OR 'septic shock' AND 'microcirculation'. In EMBASE, the keywords used were '(microcirculation AND sepsis)' , '(microcirculation AND septicaemia)' , and '(microcirculation AND septic shock)'.

Selection criteria
All abstracts identified through the searches were compiled in Endnote® (Thomson Reuters) and screened for relevance, after removing duplicates. Publications were eligible for inclusion in the review if they were conducted on animal models of LPSinduced sepsis and septic shock and compared different types of fluid resuscitation, with or without a control group. Full published manuscripts of studies considered relevant were retrieved and reviewed. Cited publications within the retrieved articles were also screened for relevance. We restricted our review to studies assessing the effects of fluid resuscitation on the microcirculation in sepsis/septic shock and articles either written in English or translated into English.

Results
A total of 2346 articles were retrieved from our search of published literature. Initial screening to remove duplicates and studies with no apparent relevance yielded 912 unique articles ( Fig. 1). There has generally been a gradual increase in the number of published papers on sepsis and the microcirculation over the last 25 years (Fig. 2). After excluding non-experimental review articles, clinical studies and pre-clinical experiments investigating therapeutic interventions other than volume resuscitation, 11 studies were found to be relevant to the assessment of microcirculatory effects of volume resuscitation in animal models of LPS-induced sepsis. The methods used for quantification of microcirculation function, type and volume of resuscitation fluid administered and the resultant microcirculatory effects in the retrieved studies are presented in Table 1.

Microcirculatory assessment and outcome
In the small rodent models, microcirculatory function was assessed primarily by intravital fluorescence microscopy [61][62][63], with one study using laser speckle imaging [52]. Laser Doppler flowmetry and/or videomicroscopic sidestream dark-field (SDF) imaging were largely used for microcirculatory assessments in the larger mammalian models [15,51,65,67]. One study utilized indirect quantification of oxygen extraction to assess microcirculatory function [66]. Six studies reported improved blood flow in the microcirculation and reduced extravasation of plasma following fluid resuscitation [61-64, 66, 67]. One study reported improvement of microvascular blood flow and oxygenation with overall negative fluid balance in sepsis [65], whilst three studies documented persistent microcirculatory dysfunction with volume resuscitation [15,51,52]. One study exhibited improved sublingual and serosal intestinal microcirculation, but persistent dysfunction in the intestinal mucosal villi within the same sepsis model resuscitated with hydroxyethyl starch (HES) [68]. None of the studies presented in this review used blood transfusion for volume resuscitation. Differences in the animal species, size of the study arms, methods of microcirculatory assessment, types and volume of fluid resuscitation administered and the experimental time-points for the induction of sepsis and resuscitation precluded any quantitative meta-analysis.

Discussion
Preclinical animal research is a necessary adjunct to clinical trials as we seek to improve our mechanistic understanding of the pathophysiology of sepsis and the consequent effects of volume resuscitation. Different outcomes in the animal studies presented in this review reflect a state of equipoise regarding the effectiveness of volume resuscitation on the microcirculation in the management of sepsis and septic shock.
The aetiology of microcirculatory dysfunction seen during sepsis is multi-factorial, including increased blood viscosity, reduced red blood cell deformability [69], neutrophil activation [70] and impaired vascular auto-regulation [20], leading to inadequate oxygen delivery to tissues. The mechanical properties of red blood cells are altered by effects of endotoxin binding directly to the red blood cells [71]. Increased vascular permeability occurs as a result of endothelial cell damage. Disruption of the endothelial- glycocalyx barrier exposes endothelial cell surface adhesion molecules that trigger the activation of mast cells, and the adhesion of platelets and white blood cells, thereby causing further inflammation [72] and nitric oxide production [73]. Nitric oxide, in turn, induces the relaxation of vascular smooth muscle, leading to increased blood flow shear stress and the production of pro-inflammatory cytokines, further worsening endothelial inflammation. Rapidly administered volume resuscitation therapy in the setting of a damaged endothelial-glycocalyx barrier alters cardiovascular haemodynamics [51] and may further exacerbate interstitial oedema directly by increasing intravascular hydrostatic pressure and indirectly through osmotic diffusion across the concentration gradient of leaked solutes.
Three of the studies presented in this review used intravital microscopy to assess the microcirculation, two demonstrating an improvement with fluid administration [61,63] and one an improvement with hydroxyl-ethyl starch but not saline administration [62]. Only one study used laser speckle microcirculatory imaging [52] and showed no improvement following fluid administration. Three studies presented in this review used laser Doppler imaging and showed improved microcirculatory reactivity with fluid administration [64,65,67]. Microcirculatory assessment by sidestream dark-field imaging presented in this review exhibited heterogeneous effects in different tissues in one study [68], but no improvement in another [51]. Comparisons of different microcirculatory imaging techniques have been reported in literature [41,[74][75][76]. Investigators who conducted a comparative study on cochlear blood flow concluded that intravital microscopic measurements were more sensitive than laser Doppler measurements [77]. However, comparing the microcirculatory results attained with different methods is a       Volume of resuscitation fluid administered not described significant challenge. We considered some of these, including videomicroscopic techniques, whilst offering direct visualization of the microcirculation, the tissue contact element is largely uncontrolled and may cause pressure-induced artefacts.
On the other hand, laser techniques are susceptible to motion artefacts, so that combining techniques might enhance the accuracy of microcirculatory assessments [76]. Only one study presented in this review used a combination of techniques to assess the microcirculation (i.e., laser Doppler flowmetry and sidestream dark-field imaging) and reported persistent acidosis, microcirculatory and mitochondrial dysfunction despite resuscitation with hydroxyethyl starch (HES) fluid [15]. Routine microcirculation assessment is rarely performed owing to heterogeneity of microvascular beds in different organs and complexity of assessment techniques. There was significant heterogeneity in the type and volume of resuscitation fluids used in the studies presented in this review. Clinically, substantial controversy still surrounds the choice of fluid for resuscitation. The Saline versus Albumin Fluid Evaluation (SAFE) study in adults similarly failed to identify any significant difference in the sepsis sub-group of patients [78], despite a lower unadjusted relative risk of death for albumin versus saline compared with non-septic patients [79].
The Crystalloid versus Hydroxyethyl Starch Trial (CHEST) revealed an increased rate of renal replacement therapy and adverse events (pruritus, rash) in intensive care unit patients resuscitated using colloid (6% hydroxyethyl starch, HES) compared to crystalloid (0.9% saline), with no significant difference in the 90-day mortality between the two groups [80]. For paediatric sepsis, the evidence base for fluid resuscitation upon definitive choices of fluids for resuscitation in severe infection and shock remains very weak [81], with most studies providing low quality of evidence or focusing on malaria and dengue. The FEAST trial documented a 45% relative (95% confidence interval, CI 13-86%) increase in 48-h mortality compared to control, with no significant difference between the saline and albumin bolus arms [57]. A subsequent systematic review formally assessing the evidence for bolus fluid resuscitation included 13 studies which met the inclusion criteria (4 general shock, 4 malaria, 4 dengue and 1 severe malnutrition). None were conducted in high-income countries and the only study to include a control arm (i.e. no fluid bolus arm; FEAST trial), found at 48 h a lower mortality in controls compared to those receiving saline or colloid boluses (relative risk for sepsis 0.69, 95% CI 0.54-0.89 and for malaria 0.64, 95% CI 0.46-0.91, respectively) [82]. Similarly, other studies have reported higher rates of morbidity and mortality with positive fluid balance in septic shock [83][84][85][86][87][88][89] and one retrospective pilot study reported good prognosis in patients who achieved a negative fluid balance within the first 3 days of septic shock [90]. Previous research showing transient pulmonary arterial hypertension that was induced by transfusion with donor blood, which had been stored for 35-42 days, is suggestive of a two-hit hypothesis that can be extrapolated to sepsis (first-hit) and volume resuscitation (second-hit) [91]. The pathogenesis of sepsis is complex, and evidence of higher rates of mortality with volume resuscitation therapy has led to the development of a two-hit hypothesis, highlighting the need for further mechanistic studies. Similarly, the heterogeneity in monitoring targets to guide resuscitation in septic shock has been recently highlighted [53][54][55]. Novel approaches and consensus in monitoring targets are required to preserve and evaluate the microcirculation in order to improve the treatment of sepsis.

Conclusions
Microcirculatory resuscitation is an important therapeutic goal, as restoration of the macro-circulation alone fails to improve microvascular function. Transferring wetbench discoveries to the clinical setting requires biologically relevant animal model studies to understand both the mechanisms by which fluid resuscitation affects the microcirculation and the role of the endothelial-glycocalyx in sepsis. One important priority for mechanistic research in sepsis is to compare the microcirculatory indices and outcomes achieved with different resuscitation strategies. Further research with defined therapeutic end-points and standardized fluid and microcirculatory assessment protocol is required as it remains unclear whether the heterogeneity that was seen was a result of the model used, resuscitation fluids used or the microcirculatory assessment techniques.