The association between hypoalbuminemia and microcirculation, endothelium, and glycocalyx disorders in children with sepsis

The objective of this study was to evaluate the association between serum albumin levels and microcirculation changes, glycocalyx degradation, and the clinical outcomes of interest.


| INTRODUC TI ON
Sepsis is a clinical syndrome characterized by potentially fatal organ dysfunction secondary to a dysregulated host response to infection. 1 Despite advances in diagnosis and early recognition, it still has a high mortality rate, especially in middle and low-income countries. 2It is common for multiple organ failure to be the final common pathway to death for the majority of these patients.The cardiovascular system is one of the most commonly involved organs both in macro and microcirculation.[5][6] The glycocalyx is a negatively charged carbohydrate-rich layer covering all endothelial cells.It is made up of proteoglycans, glycosaminoglycans, and glycoproteins.It was first described by Luft et al.   in 1966 using special stains that allowed it to be seen on electron microscopy. 7,8It is responsible for maintaining vascular permeability and mechanotransduction and for modulating the inflammatory response, as well as sustaining an endothelial "antiadherent" phenotype.Its components can be measured with plasma biomarkers (syndecan-1, endocan) or sublingual dark field video microscopy (Handheld Vital Microscopes -HVM). 8With the devices available today, the glycocalyx can be assessed indirectly in adults and children with sepsis with a high inter and intra-observer correlation index and good reproducibility in emergency room and intensive care settings. 5,9,10[12] In addition, endothelial inflammation and dysfunction in these patients may decrease albumin synthesis and increase its loss.There are several consequences of this hypoalbuminemia including abnormal transport of hormones and other substances, abnormalities in the ability to regulate osmotic pressure, or diminished antioxidant activity.4][15][16][17] Due to its negative charge, albumin is electrostatically repelled by the glycocalyx, which helps keep it within the intravascular space.However, in disease states, albumin moves to the interstitial space, and it is unclear whether it can alter capillary Starling forces.The current theories mention that the sub-glycocalyx (protein-free) space is one of the most important determinants in the fluid balance between the intravascular and interstitial space. 5,180][21][22] The consequences of hypoalbuminemia for the endothelium and endothelial glycocalyx in children with sepsis are unknown.Our hypothesis is that low albumin levels alter the microcirculation and foster glycocalyx degradation and, therefore, the inflammatory response, which may lead to worse clinical outcomes.The objective of our study was to evaluate the association between serum albumin levels and microcirculation changes, glycocalyx degradation, and the clinical outcomes of interest.

| Study design and context
An observational, analytical, prospective cohort study was carried out in children hospitalized in the PICU of Fundación Cardio-Infantil in Bogotá, Colombia, between January 2021 and June 2022.This study was approved by the hospital's ethics and research committees (CEIC-0366-0022) and all the parents or guardians signed informed consent prior to being included in the protocol.

| Patients eligibility
All children from 1 month to 18 years old with sepsis or septic shock who were admitted to the PICU due to clinical deterioration and on whom sublingual video microscopy and albumin serum levels were conducted within 6 h of admission were included.Patients with hypoalbuminemia and with normal serum albumin levels were included to evaluate microcirculation in both cohorts.Patients who had received 5% albumin boluses for fluid resuscitation within the 24 h prior to admission to the study; those with hyperglycemia, diabetic ketoacidosis, head trauma, or continuous renal support therapy; those in the postoperative period following cardiovascular surgery and those with a history of chronic kidney disease (defined as a glomerular filtration rate less than 60 mL/min per 1.73m 2 for more than 3 months) were excluded.
All children with a clinical syndrome characterized by potentially fatal organ dysfunction caused by a dysregulated host response to infection were considered to have sepsis.Septic shock was defined as sepsis with particularly severe circulatory, cellular and metabolic abnormalities, according to the recently recommended definitions. 2Hypoalbuminemia was defined as serum albumin less than 3.0 gr/dL. 13,17Children with low albumin during their PICU stay received replacement with 20% albumin at a dose of 1-2 gm/kg/day at the discretion of the attending physician.The severity of all patients was evaluated using the PIM-2 scale.Acute kidney injury was defined as abnormal creatinine for the patient's height, according to the Schwartz formula.Hyperchloremia was defined as serum chloride greater than 110 mEq/L, and hypernatremia as serum sodium greater than 150 mEq/L.Semiquantitative elevated procalcitonin (PCT) was defined as greater than 2.0 gm/ dL, elevated C-reactive protein (CRP) as 4 mg/dL, elevated ferritin as greater than 500 mg/dL, and abnormal D-dimer as more than 1.5 mg/L.

| Microcirculation, endothelial activation, and glycocalyx degradation measurement
The definition of microcirculation used was the one recommended in the consensus on sublingual microcirculation. 10Microvessels were defined as vessels with a diameter <20 μm, including arterioles, capillaries, and venules.Capillaries were defined as vessels <10 μm in diameter in which a single file of red blood cells (RBCs) can be observed.The main characteristics of venules are that they are vessels that collect blood from other vessels, and they have more RBCs in the lumen than the single-file RBCs seen in capillaries.The microcirculation and endothelial glycocalyx degradation were assessed in vivo using dark field video microscopy (Glycocheck System ® -Microvascular Health Solutions Inc 2014.)within 6 h of admission to intensive care.The measurement was repeated 24 h after being included in the study.This device measures sublingual microcirculation by evaluating 4-25 μm diameter vessels, using a dark field camera (CapiScope, HVCS, KK Technology United Kingdom) which emits stroboscopic green light diodes which detect RBCs by reflection.The machine amplifies the image 325 times with 720 pixel resolution and establishes 23 frames per second.
The software (Glycocheck System®) analyzes the measurements from high-quality images (in terms of movement, intensity and focus).In order to do this, it defines 10 μm vascular segments and records 40 frames (300 green segments, which are the ones with complete measurement).The operator would move the camera to five or ten different positions and could take up to 3000 vascular segments.The measurements were analyzed by the machine's software independently of the examiner and the investigators.This system analyzes the data and reports what has been termed the perfused boundary region (PBR) in μm, which is inversely related to the endothelial glycocalyx dimensions.In healthy individuals, the normal PBR is considered to be less than 2.0 μm. 8,9The software calculates the dynamic lateral red blood cell movement in the permeable region of the glycocalyx layer, which is expressed as the PBR (in μm).This measurement may vary depending on the red blood cell velocity and flow, and therefore the software also measures what is known as PBR dynamic or flow corrected. 8,9The flow-corrected evaluation is a complement to PBR and has been found to be higher in patients with sepsis compared with healthy controls. 8In addition, the video microscope measures the percentage of capillary blood volume (PBV-proportion of perfused blood vessels over the total number of vessels) and the capillary density of 4-6 μm vessels (CD 4-6 s).
Angiopoietin-2 (Ang-2 / Human Angiopoietin 2 ELISA Kit ANG 2; ab99971, Abcam Lab) was used as the biomarker for endothelial activation and increased vascular permeability.Plasma syndecan-1 (Human Syndecan-1 ELISA Kit CD138; ab46506; Abcam Lab, Cambridge, United Kingdom) was processed as the endothelial glycocalyx degradation biomarker, and a level under 80 ng/mL was considered normal. 12Annexin A5 (ab119503 -Annexin V Human ELISA Kit: Abcam Lab) was measured to evaluate cell death.All biomarker measurements were done in duplicate according to the manufacturer's instructions, simultaneously and within 6 h of admission to intensive care.The samples were 100 μL of citrated plasma which were centrifuged for 30 min at 1000 rpm, and the samples were stored at (−) 20°C for later processing using the enzyme-linked immunosorbent assay (ELISA) method.

Demographic analysis variables along with clinical data on
macrocirculation perfusion (heart rate, arterial pressure, pulse pressure) and microcirculation perfusion (capillary refill) were gathered on admission to the PICU and 24 h later.In addition, the need for vasoactive support and the severity of the disease according to the Pediatric Index of Mortality -2 (PIM-2) and Pediatric Logistic Organ Dysfunction-2 (PELOD-2) were also gathered within 6 h of admission.Laboratory tests were conducted on admission and 24 h later, including serum albumin, electrolytes, serum lactate, creatinine, D-dimer, and inflammatory biomarkers (ferritin, CRP, and PCT).

| Outcomes
The primary outcome was the association between hypoalbuminemia and microcirculation changes, endothelial activation and glycocalyx degradation measured with sublingual video microscopy and plasma biomarkers (Ang-2, syndecan-1).The secondary outcomes were the association between microcirculation abnormalities in patients with hypoalbuminemia and the presence of an inflammatory response or need for mechanical ventilation (MV).

| Statistical methods
Descriptive statistics were derived reporting each cohort with or without hypoalbuminemia.The categorical variables were reported as proportions and compared using Pearson's Chi 2 or Fisher's exact test according to the count per cell.For continuous variables, data were reported as means or medians according to their distribution, with their associated measures of dispersion.A bivariate analysis was performed according to the variable's distribution using Student's t-test for equal variances in the case of two variables.The Wilcoxon test was used to evaluate the differences in non-normally distributed biomarkers from dependent groups.Confounding factors were controlled for in the design, using the exclusion criteria to restrict patients who might have other reasons to explain the endothelial glycocalyx damage previously described in the literature, like ketoacidosis or trauma. 5,8,11Confounding (especially disease severity assessed by the PIM-2 scale, malnutrition, the use of vasoactive medications and the use of 20% albumin infusions) was also controlled for in the statistical analysis plan by carrying out multivariate analysis using logistic regression.The variables which fulfilled the Hosmer-Lemeshow criteria on bivariate analysis and which had biological plausibility were included in the model.The model was constructed using the forward method and was adjusted with the Omnibus test.Two-sided analyses were performed with a p value less than or equal to 0.05 considered to be statistically significant.Analysis was performed using SPSS (IBM® version 26 statistical package).

| Patients characteristics
During the study period, 125 patients with sepsis or septic shock were included (Table 1).The median age was 2 years (IQR 0.5-12.5).The participants were similarly distributed by sex (46% females).Of these, 48 children (38.4%) had hypoalbuminemia.The main causes of PICU admission were respiratory and gastrointestinal problems.Altogether, 57.6% of the patients had septic shock, with no difference in terms of severity with regard to the serum albumin level.We had 14 patients (11.2%) admitted to intensive care for SARS-CoV-2 related diseases including multisystemic inflammatory syndrome associated with COVID-19 (MIS-C).

| Microcirculatory changes associated with serum albumin
Children with sepsis were found to have microcirculatory changes associated with the serum albumin level (Table 2).Of the patients with a serum albumin under 3.0 gm/dL in the first 24 h, 76% had a PBR flow corrected greater than 2.0 μm.Children with hypoalbuminemia had more 4-6-micron capillaries recruited than patients with normal albumin (60% vs. 40%; p = .04),with no differences in the percentage of blood volume (PBV) in the capillary (p = .42)(Figure 1A-C).A moderate correlation was found between PBR flow corrected and the number of CD 4-6 s recruited (rho 0.28; p < .01)(Figure 1B).In fact, those with the worst PBR had the highest number of CD 4-6 s recruited (rho 0.29; p < .01).
We found no relationship between the syndecan-1 level and serum albumin levels (p = .21),although the patients with hypoalbuminemia had more kidney injury that could affect the serum levels of this biomarker.On admission, patients with hypoalbuminemia had 48% higher Ang-2 (p = .04)and annexin A5 (p = .03).Twentyfour hours after admission the microcirculation disorders persisted in patients with low albumin.The hypoalbuminemic group's PBR flow corrected was higher (2.2 μm IQR 1.81-2.49)than that of the group with a normal albumin level at 24 h (1.79 μm IQR 1.75-1.84)(p = .04).

| Microcirculatory changes and inflammatory response
Of the total population, 57.6% (72/125) had elevated PCT.Spe- There was also a negative correlation between albumin levels and     2).

| DISCUSS ION
In this study, we found that patients with sepsis have microcirculation changes associated with hypoalbuminemia.These children have endothelial glycocalyx degradation more often, a greater density of small 4-6 μm capillaries recruited, and increased redistribution of blood flow toward the microcirculation.In addition, we found that children with hypoalbuminemia have more elevated biomarkers of increased endothelial permeability, greater inflammatory response and require MV more often.The group who received albumin replacement for hypoalbuminemia showed recovery of the microcirculation and glycocalyx variables and had a shorter PICU stay and lower mortality.
The microcirculation is the terminal effector site of the cardiovascular circulatory system, where the supply of oxygen to the tissues and elimination of metabolic waste are coupled and controlled.
In the capillaries, the passage of fluid from the intravascular to the interstitial space is controlled by hydrostatic pressures, capillary oncotic pressure and sub-glycocalyx pressure. 5,6,18In the microcirculation, albumin, besides regulating the flow of liquid between the different spaces, can have immunomodulating effects in patients with sepsis.Studies in animal models have found that normal albumin levels favor sphingosine-1-phosphate transport in the RBCs and platelets. 22,235][26] In our study, we found greater glycocalyx degradation in children with sepsis and hypoalbuminemia, which was associated with a greater inflammatory response and elevated cell death biomarkers.
In sepsis, hypoalbuminemia has been associated with worse outcomes, including an up to 23 times greater risk of death.4][15][16][17] Recently, with artificial intelligence techniques and machine learning-based models which perform nonlinear analyses, low albumin has been found to be one of the most useful serum biomarkers for predicting survival in critically and chronically ill patients. 17Albumin, besides being the most important protein for maintaining the colloid-osmotic pressure of plasma, is responsible for carrying many endogenous molecules (antiinflammatory and immunomodulatory molecules) and exogenous molecules (like antibiotics). 14,15It has important anti-inflammatory and antioxidant properties. 20,21The group with the lowest albumin in our study, in addition to being associated with microcirculation disorders, had a greater inflammatory response and worse outcomes, like a greater need for MV.This could be explained by increased recruitment of small pulmonary capillaries in children with hypoalbuminemia associated with increased vascular permeability due to sepsis (we found elevated angiopoietin-2), which would favor more capillary leakage toward the alveolar units, a higher risk of pulmonary edema, and therefore, a longer duration of MV support.We also found a higher frequency of acute kidney injury in the group with hypoalbuminemia, which could favor albuminuria and predispose to worsening and greater glycocalyx degradation in sepsis.
As has been stated, serum albumin has anti-inflammatory and antioxidant properties. 20,21In this study, we found that patients In an animal model of hemorrhagic shock, 5% albumin resuscitation, which is similar to the body's protein content, restores the glycocalyx when compared with crystalloids. 30These effects may be mediated by greater sphingosine-1-phosphate release from the RBCs and the platelets.In a recent study, Fernández-Sarmiento et al. 31 found that, in children with sepsis, saline solution was associated with greater endothelial glycocalyx degradation, and that patients with hypoalbuminemia and albumin infusion replacement had a lower risk of microcirculation disorders (aOR 0.56 95% CI 0.31-0.98).Serum albumin replacement may have beneficial effects on the respiratory, cardiovascular, and neurological systems as well as the circulatory status. 16 this regard, the Saline versus Albumin Fluid Evaluation (SAFE) study compared the impact on mortality in critically ill adults of receiving 0.9% normal saline solution versus 4% albumin as replacement fluid during fluid resuscitation. 32No differences were found in terms of mortality, but patients who received fluid resuscitation with albumin required a 30% lower volume of fluids during resuscitation.Later, the ALBIOS study evaluated the effect of simultaneous administration of 20% albumin and crystalloids to maintain blood albumin levels equal to or greater than 3.0 gm/dL. 33Patients who had better albumin levels had their vasopressors or inotropes discontinued sooner, less positive balances, and greater macrocirculation stability.Recently, Raghunathan K et al. 34 found that patients with sepsis and acute kidney injury who received albumin infusions had shorter hospital stays than those who did not receive replacement (hazard ratio, 1.83; 95% CI, 1.56-2.15;p < .001).Albumin replacement and maintaining levels greater than 3.0 gm/dL in critically ill patients has been considered a cost-effective approach, showing an almost 20% reduction in care costs for patients with septic shock. 35In our study, we found that patients who received albumin replacement experienced endothelial glycocalyx degradation recovery, and this group had a shorter hospital stay and lower mortality.7][38] This hypothesis must be confirmed in future clinical trials or studies with an appropriate design for determining causality.
There is increasing evidence of the association between hypoalbuminemia and clinical outcomes in critically ill patients.Leite H et al. found that a 1.0 gm/dL increase in serum albumin on admission was related to a 73% reduction in the risk of death (HR 0.27; 95% CI 0.14-0.51;p < .01). 39A 1 gm/dL increase in serum albumin was independently associated with a 33% greater probability of early discharge from the PICU (HR 1.33; 95% CI 1.07-1.64;p = .008) and an increase in ventilator-free-days (OR 1.86; 95% CI 0.56-3.16;p = .005).In our study, we found that patients in whom hypoalbuminemia was corrected had less positive fluid balances and a shorter PICU stay.
We consider that our study has several limitations.First, it was performed at a single reference center for highly complex patients.
This could lead to the included patients being much more ill.However, the severity and organ dysfunction scales were similar in the groups with and without hypoalbuminemia, despite changes found in the microcirculation.However, the group with normal serum albumin had significantly more acute kidney injury.This involvement could have affected the serum levels of syndecan-1, which we did not find to be different between the two groups. 40In addition, we did not differentially analyze patients with malnutrition.This group could have underlying conditions associated with chronic microcirculation changes and worse outcomes.We tried to control for this variable in the multivariate analysis.However, we are not aware of studies evaluating how nutritional status affects endtissue perfusion and microcirculation.Another limitation of our study is that PBR is an indirect measure of glycocalyx damage.Although the equipment has low inter and intra-observer variability, it is unknown if the measurements could be affected by fluid boluses or the use of vasopressors.In addition, we found an altered PBR with elevated inflammation biomarkers or a greater need for MV.While the groups in which we found this association are small, they had statistical significance, and we must continue to study the relationship between the inflammatory response and microcirculation abnormalities.Finally, due to the available budget, we did not measure cytokines nor conduct long-term follow up of the acid-base status or chloride levels in our patients to determine if there was a relationship between albumin levels, microcirculation changes, and glycocalyx degradation which could be explained by an abnormal inflammatory response or acid-base balance or hyperchloremia. 41,42

| CON CLUS I ON AND PROS PEC T
In children with sepsis, an association was found between hyopalbuminemia and microcirculation changes.These patients have endothelial glycocalyx degradation more often, greater small capillary recruitment, and increased blood flow redistribution toward the microcirculation.In addition, the group with hypoalbuminemia had elevated endothelial activation, inflammatory response and apoptosis biomarkers.These microcirculation disorders in patients with hypoalbuminemia were associated with a greater need for and duration of MV, a longer hospital stay and greater mortality.We should carry out clinical studies evaluating the role of albumin infusions in outcomes like endothelial dysfunction, glycocalyx degradation and possible patient subgroups who could benefit more than others, decreasing the time on MV or mortality.

| PER S PEC TIVE S
Children with hypoalbuminemia have important changes in microcirculation, increased endothelial activation, glycocalyx degradation, and elevation of inflammatory biomarkers.These microcirculation disorders were associated with a greater need for and duration of PCT on admission (rho: [−] 0.16; p = .04)and after 24 h (rho [−] 0.20; p = .04).These patients with high PCT did not have more 4-6 μm capillaries recruited, regardless of the albumin level (p = .61).We found a negative correlation between albumin levels and CRP on admission (rho: (−) 0.22; p = .01)and after 24 h (rho [−] 0.26; p < .01).Patients with normal albumin and high CRP had a greater blood volume percentage in the recruited blood capillaries (72% vs. 44%; aOR 3.18 95% CI 1.12-3.12;p = .02)with no alterations TA B L E 2 Microcirculation changes associated with the serum albumin level.

F I G U R E 2
with higher PCT levels had more glycocalyx degradation, lower serum albumin, and elevated angiopoieitin-2.Recently, Brabenec L et al. found that elevated PCT was associated with microvascular dysfunction in patients after heart surgery, favoring increased vascular permeability.27,28These findings were confirmed with an animal model in which endothelial junction damage was induced by injecting intravenous PCT in a preclinical model.They identified how PCT-induced damage requires activation of a dipeptidylpeptidase 4 (DPP4) which, when inhibited with medications, was correlated with reduced capillary leakage and improved vascular integrity.This is the first study in human beings that suggests that PCT may be the cause of increased inflammation and capillary leakage in inflammatory states.In addition, they found that modulating elevated PCT levels could have protective effects on the vascular barrier.The findings of our study in children with sepsis are along the same lines.Hyperprocalcitoninemia was associated with low albumin, glycocalyx degradation and increased vascular permeability biomarkers (Ang-2).These mutually associated phenomena may favor the inflammatory response, with consequences for the microcirculation and microvascular permeability.In fact, we believe that the presence of hypoalbuminemia in sepsis may condition a compensatory response in the microcirculation leading initially to greater capillary recruitment, contrary to what is described in adults.This may be a temporary compensatory microvascular change in the early stages of the disease.As sepsis advances, this compensatory capacity for capillary recruitment may be lost, resulting in low vascular density which, if it persists after 48 h of PICU admission, has been associated with higher mortality.29Studies are needed to examine these aspects further and conduct mediumterm follow up of the microvascular changes associated with sepsis.Alterations in microcirculation in patients who underwent hypoalbuminemia correction.(A) Perfused boundary region corrected to flow on admission and at 24 hours in patients with and without hypoalbuminemia correction.(B) Percentage of Capillary density of 4-6 microns.(C) Fluid balance percentage greater than 10% with and without hypoalbuminemia correction.(D) Angiopoietin-2 and hypoalbuminemia correction.
Multivariate analysis of the factors associated with microcirculation alterations in children with sepsis and hypoalbuminemia evaluated through video microscopy.Persistence of altered PBR despite correction of hypoalbuminemia.
TA B L E 3 a