Background

Fluid bolus therapy (FBT) is a first line therapy for resuscitation of different types of shock and has been a recommendation of international guidelines for about two decades. Nevertheless, fluid resuscitation remains controversial in pediatric patients. Finding the best type of fluid, volume, and infusion rate is still a major therapeutic challenge.

Many types of fluids have been used as FBT: isotonic crystalloids, hypertonic saline, natural (albumin) and synthetic colloids (gelatin, dextrans, and hydroxyethyl-starches), as well as different combinations of colloids and hypertonic crystalloids. However, none of them has proven to be the ideal resuscitation fluid, capable of producing effective, long-lasting blood volume expansion with few side effects.

Isotonic crystalloids, such as normal saline (NS) or Lactated Ringer’s solution (LR), have been considered the fluid of choice for initial resuscitation in pediatric patients, due to their wide availability and few side effects1,2. However, larger quantities of liquid infusion are required in comparison to other fluids, and the duration of its effect is short, since 75% of the infused volume easily diffuses into the interstitial and intracellular spaces3. Hyperchloremic metabolic acidosis, hypothermia and coagulopathy have been described when infusing large volumes of NS. Therefore, studies of fluid restrictive strategies in adults and children have emerged, and international sepsis guidelines recommend not to exceed 3 L of fluid in the first 6 h of resuscitation in adult patients4,5 and 60 ml/kg during the first hour of resuscitation in children with septic shock1. However, fluid overload (FO) has also been described, even when following these restrictive fluid recommendations. A FO greater than 5% at the end of the first day of admission to the PICU has been associated with greater morbidity and mortality6. Furthermore, in certain resource-limited settings, fluid resuscitation in children with fever and poor perfusion without severe hypotension has been related to increased mortality7. This is why even more restrictive strategies are being explored.

Hypertonic resuscitation fluids such as hypertonic saline (HS) produce a greater increase in preload and urine output with smaller volumes than isotonic fluids, as they have greater tonicity than plasma, and can attract intracellular water to the intravascular space. Further benefits of HS include potentially beneficial immunomodulatory effects8,9. The main side effects of hypertonic solutions include hypernatremia and hyperchloremia. Nonetheless, a recent meta-analysis compared the performance of crystalloid and colloid hypertonic fluids in adult patients with hypovolemic shock. Hypotensive patients who were resuscitated with HS or hypertonic dextran had lower mortality than those who received isotonic crystalloids10.

Colloids have greater molecular weights than crystalloids and are therefore capable of increasing plasma oncotic pressure and thus attract water from the interstitial space to the intravascular space. Therefore, the infusion of smaller volumes of colloids has a greater and longer lasting effect when compared to crystalloid boluses. However, some colloids as starches and dextrans can cause anaphylactic reactions, coagulation disorders and renal dysfunction11,12.

Albumin is the major determinant of plasma colloid osmotic pressure. Therefore, fluid resuscitation with albumin, a natural colloid, has fewer side effects than synthetic colloids. In addition, albumin has protective endothelial effects, acts as a free radical scavenger, and has immunomodulatory and anti-inflammatory properties, decreasing vascular permeability13. The combination of a natural colloid and a hypertonic solution (such as 3% hypertonic saline albumin) could maximize the beneficial effects of longer lasting intravascular fluid replacement with minimal infusion volumes. This could be especially useful in patients with intravascular hypovolemia and extravascular fluid overload6,14,15.

Recently published articles reported a relationship between the administration of albumin and decreased mortality in adults with septic shock16,17. Another study, however, reported higher mortality in a subgroup of patients with traumatic brain injury18, but these results were probably due to low osmolality of the albumin solution that was used19. Some disadvantages of albumin include its high price20 and the need to be stored in glass containers.

An experimental model of controlled hemorrhage in piglets reported greater improvement in hemodynamicand tissue oxygenation parameters in the group receiving HSA than those with twice the volume of NS21,22. Severe hyperchloremic acidosis was not observed.

The main objective of this study was to evaluate the short-term hemodynamic effects and the safety profile of HSA fluid bolus resuscitation in critically ill children.

Methods

Study design and treatment protocol

An observational, prospective, pilot study was conducted in the 17-bed medical-surgical pediatric intensive care unit (PICU) of a tertiary hospital in Madrid, Spain. It is a national referral center for congenital and acquired heart disease and it counts with an advanced cardiac surgery program that includes heart transplant, extracorporeal membrane oxygenation (ECMO) and ventricular assist device (VAD) programs.

All patients between 0 and 16 years of age who received HSA boluses during PICU admission between October 2018-May 2021 were included, regardless of their diagnosis on admission. Informed consent was obtained from all subjects and/or their legal guardian(s).

A local protocol regarding HSA bolus therapy was designed before the study started (Supplementary material-1). HSA was indicated in patients with suspected hypovolemia and persistent hypotension after conventional FBT with more than 20 ml/kg of colloid or crystalloid solutions, especially in patients that had right ventricular dysfunction or a positive net fluid balance and intravascular hypovolemia despite fluid overload in the cellular and extravascular spaces. HSA fluid bolus therapy was decided at the discretion of the prescribing physician. HSA was contraindicated in patients with serum sodium ≥ 150 mEq/L, serum chloride ≥ 110 mEq/L, pH < 7.15, or serum albumin ≥ 4 0.5 mg/dl.

In addition to the study fluid, patients received maintenance fluids, enteral or parenteral nutrition, and blood products at the discretion of the treating clinicians. The monitoring of central venous pressure, invasive arterial blood pressure, and other aspects of patient care were performed at the discretion of the treating clinicians.

Preparation and administration of HSA

HSA was prepared by adding 22 ml of 20% NaCl and 50 ml of 20% albumin (Albutein Serum Human 20% Albumin, Grifols, Spain) to 128 ml of 0.9% saline, to achieve 200 ml of HSA, as described in previous studies21,22. The resulting fluid is equivalent to a mixture of 5% albumin and 3% hypertonic saline. An osmolarity analysis of the HSA solution was performed using the OsmoSpecial 1 osmometer (Astori Tecnica SR, Italy). Three osmolarity tests were conducted, obtaining a median fluid osmolarity of 980 mOsm/L (Standard Deviation, SD ± 2 mOsm/L).

HSA boluses were infused by a central intravenous access at a dose of 5 ml/kg over a 15-minute period. If necessary, repeated 5 ml/kg boluses were administered, especially if hypovolemia was suspected (active bleeding, excessive fluid output such as high diuresis, increased output trough surgical drains….) and if a positive response was observed previously.

Baseline assessment and follow-up data collection

Demographic, clinical, and analytical variables were recorded before and 1, 2, 6, 12 and 24 h after each bolus infusion. Patients were followed until discharge from the PICU or death. Seriated blood tests that included liver and renal function, electrolyte tests (sodium, potassium, chloride, magnesium, phosphorus and calcium), blood gas profiles and coagulation were performed daily during the first 48 h after infusion.

The hemodynamic parameters evaluated included arterial blood pressure, heart rate, central venous blood pressure, vasoactive index score, lactate, central venous oxygen saturation, urinary output and other parameters related with end-organ perfusion. Hypovolemia was indirectly assessed by clinical findings (hypotension, tachycardia, peripheral perfusion, low urine output, elevated fluid loss via chest drains or elevated lactate), as recommended by the international guidelines on hemodynamic monitoring in critically ill children23.

HSA safety profile evaluation included potential adverse effects such as metabolic acidosis, increased risk of bleeding, anaphylactic reactions, acute kidney injury, electrolyte disbalance or pulmonary edema.

Fluid overload was assessed by clinical findings (interstitial pulmonary edema, subcutaneous edema, pleural effusion or ascites). When the study was performed, we didn’t use a routine ultrasound protocol to check FO, such as Venous Excess Ultrasoung Grading System (VExUS) score24.

Pediatric Risk of Mortality Score III (PRISM-III) was calculated for each patient on admission or, at least, 72 h prior to the first HSA infusion. Blood pressure (BP) was monitored invasively in all patients. Since BP reference values vary widely in the pediatric population, the variable “variation in BP” was calculated using the following formula: (initial BP – final BP) / 100. This was used for all BP measurements (systolic, mean, and diastolic BP at each time point of the study).

Net fluid balance (total fluid input - total fluid output) was expressed in ml/kg at the end of the three following work shifts after each HSA bolus infusion. The morning shift includes the period from 8 a.m. to 3 p.m. (7 h), the afternoon shift from 3 p.m. to 10 p.m. (7 h) and night shift from 10 p.m. to 8 a.m. (10 h).

The use of mechanical ventilation, transfusion requirements and use of renal-replacement therapy were recorded daily until discharge from the ICU or death.

Statistical analysis

IBM SPSS Statistics 25.0 system (IBM SPSS Statistics, Armonk, NY, USA) was used for statistical analysis. Quantitative variables were expressed as mean and standard deviation or median and interquartile range, depending on whether the sample was normally distributed. Qualitative variables were expressed as percentages. The Chi-square and Fisher tests were used to compare categorical variables. Student’s T and repeated measures ANOVA tests were used for continuous variables. P < 0.05 was considered significant.

Results

Sixty-four HSA fluid boluses were analyzed in 23 patients. The most common reason for bolus administration was acute hypotension (62.5%), followed by suspected hypovolemia without hypotension (29.7%). A mean volume of 5.7 ml/kg (SD 2.3 ml/kg) was delivered per bolus, over 15 min.

Of the 23 patients, 69.6% had been admitted after surgical correction of a congenital heart disease and 13% after a heart transplant. 47.8% of them associated right ventricular dysfunction according to cardiological evaluation. None of them received an HSA bolus in the context of traumatic brain injury. 17.3% patients required circulatory support (ECMO or VAD), and 8.7% of them were receiving continuous renal replacement techniques (CRRT) at the time of the first bolus infusion. Baseline characteristics of the patients are shown in Table 1.

Table 1 Patient baseline characteristics before first HSA bolus infusion (n = 23).
Full size table

Patients had received at least one fluid bolus in the 6 h prior to each HSA bolus in 92% of cases, with a mean prior volume infusion of 15.1 ml/kg (SD 13.6 ml/kg). The most frequent resuscitation fluid was balanced gelafundin solution, according to our local protocol. Table 2 shows the difference between baseline parameters and those 1h, 2h, 6h, 12h and 24 h after each HSA infusion.

Table 2 Variation in clinical and analytical parameters at baseline, 1 h, 2 h, 6 h, 12 h and 24 h after each HSA bolus (n = 64). Analytical parameters were only asessed every 24 hours.
Full size table

Repeated boluses

Fourteen patients, representing more than half of the HSA boluses recorded (64%), received more than one HSA bolus over time (36.6% within the next 24 h and 22% between the next 24–48 h).

The mean cumulative volume that was infused in patients who received repeated boluses was 22.2 ml/kg (SD 15.7). Patients with CRRT (4 patients, 23 boluses) received more fluids than those without CRRT (28.3 ml/kg (SD 15.0) vs. 16 ml/kg (SD 8.8), p < 0.05).

A patient with septic shock who required mechanical ventilation, inotropic support, and CRRT received up to 10 HSA boluses (64.6 ml/kg) over an 18-day period (with a maximum of 2 boluses every 24 h).

Hemodynamic effects

A significant increase in systolic, mean, and diastolic BP was observed after the administration of HSA (p < 0.05) (Fig. 1). This finding was evident within the first hour and lasted up to 24 h after the infusion. Likewise, a significant decrease in the vasoactive-inotropic score (VIS) was observed at 24 h (39 (SD 29) vs. 45 (SD 29) at baseline, p < 0.05). No differences were found in heart rate or central venous pressure in the first 24 h.

Fig. 1
figure 1

General linear model of the variation (percentage increase) of systolic blood pressure (SBP), mean arterial pressure (MBP) and diastolic blood pressure (DBP) after the infusion of hypertonic saline albumin. Av: Average. SD: Standard Deviation. Statistically significant values are marked with an asterisk (*).

Full size image

Most patients (70.3%) required further fluid boluses within the next 6 h after the administration of HSA, with a mean accumulated volume of 8.7 ml/kg (SD 9.6). Repeated boluses occurred a mean of 8.2 h later (SD 7.4 h). Nevertheless, the amount of volume infused after the HSA bolus was significantly lower than before (15.1 ml/kg in 6 h, SD 13.6, p < 0.05).

A change in fluid balance was observed after HSA infusion. Mean fluid balance during the shift in which HSA was administered was + 15.8 ml/kg (SD 24 ml/kg), and it decreased in the following shifts: -1 ml/kg, (SD 25 ml/kg) in the shift immediately after HSA infusion, and + 4.3 ml/kg (SD 24 ml/kg) in the shift after that (p < 0.05).

Thirty-eight patients (58%) had signs of fluid overload prior to HSA infusion while 40 patients (61%) showed FO 24 h after the bolus (chi-square 0.60, p = 0.40). The proportion of patients with pulmonary edema before and after the HSA bolus was 19% and 10% respectively (chi-square 1.91, p = 0.16).

Three patients (13%) required initiation of CRRT in the 24 h following HSA infusion. All of them were in the first 24 h after cardiac surgery for transposition of the great arteries (two after a Nikaidoh procedure and one after a Rastelli procedure) and had right ventricular dysfunction and FO. One of these patients had chronic renal failure prior to admission. Two of these patients required two HSA boluses before CRRT each, and the other received only one bolus.

Metabolic effects

An increase in sodium (Na) and chloride (Cl) serum levels was observed after HSA infusion. Their peak concentration was reached one hour later (Na 143 mEq/L (SD 3.5, p = 0.08) and Cl 109.7 mEq/L (SD 6, p = 0.02) respectively). The patient with the highest levels reached a peak of 153 mEq/L (Na) and 126 mEq/L (Cl) two hours after infusion. Both sodium and chloride levels returned to baseline in less than 24 h in patients with and without CRRT (Fig. 2). The increase in sodium and chloride levels did not have a clinical impact in any of the patients.

Fig. 2
figure 2

General linear model of the variation in sodium and chloride levels (mEq/L) after the administration of hypertonic saline albumin. Av: Average. SD: Standard Deviation. Statistically significant values are marked with an asterisk (*).

Full size image

Blood pH, bicarbonate, and albumin were significantly higher (p < 0.05) and lactic acid levels were significantly lower (p < 0.05) twenty-four hours after HSA infusion. No differences were observed regarding blood creatinine or urea concentrations. Urine output remained stable after each bolus (Table 2). Renal disfunction was observed in five patients surrounding the HSA infusion: three patients required initiation of CRRT, as mentioned above, and the rest of them (2 patients) developed acute kidney injury (AKI) stage 1 according to the KDIGO stratification.

Coagulation effects and transfusions

Fibrinogen slightly increased after HSA (p < 0.05), but no changes were observed in INR and aPTT, as shown in Table 2. 30% of patients received red blood cell (RBC) transfusions, 28% received platelets and 25% received plasma in the twenty-four hours following HSA bolus.

Effects on global prognosis

Side effects such as anaphylactic reactions or acute lung edema related to the infusion of HSA were not observed. Mortality rate in our cohort was 13% (3 patients), due to multiorgan failure as a worsening in their clinical status. None of the deaths could be directly attributed to the use of HSA.

Discussion

This is the first study to evaluate hemodynamic effects and safety of fluid bolus resuscitation with hypertonic saline albumin in pediatric patients. According to the results of this pilot study, HSA seems to be an effective resuscitation fluid, safe enough to continue performing a larger study.

There is a lack of solid evidence regarding which resuscitation fluid is better. In terms of safety, the FEAST trial is the only study that compares mortality between a crystalloid (NS) and a colloid solution (albumin) for FBT in pediatric patients. No differences were found between both groups25. Concerning effectiveness, Schroth et al. found an improvement in cardiac output and a decrease in vasoactive drugs in children who received a fluid bolus with hypertonic starch in the immediate postoperative period of cardiac surgery26. Experimental studies in infant piglets also support this beneficial hemodynamic effect, as the infusion of HSA resulted in a greater improvement in hemodynamic parameters than isotonic and hypertonic crystalloids in two experimental randomized studies of pediatric hemorrhagic shock, infusing half of the volume than NS21,22. Prospective clinical trials with larger sample sizes are needed to assess the effectiveness and safety of HSA and compare it to that of other fluids in pediatric patients.

The patients in our study required multiple fluid boluses before the administration of HSA. The need for new fluid boluses decreased significantly after the infusion of 5 ml/kg boluses of HSA. A significant increase in systolic, mean, and diastolic BP as well as a decrease in VSI was also observed. Nevertheless, the lack of a control group makes it difficult to ascribe this improvement solely to the HSA, as many confounding factors are present in critically ill children during the resuscitation phase of their admission. These findings, however, are consistent with the study by Jelenko et al. in adult burnt patients, as they required smaller volumes of fluid and presented a more rapid normalization of physical, physiological, and biochemical parameters than those who received hypotonic or hypertonic fluids without albumin27.

Almost half of our cohort had right ventricular dysfunction, a common feature in the post-operative period of a congenital heart surgery28,29. Fluid management of these patients is often difficult, as both hypovolemia and fluid overload with interstitial edema may coexist and can have detrimental effects on blood pressure, organ perfusion, and cardiac function30. In patients with low blood pressure and without signs of fluid overload, an increase in preload by administering small isotonic boluses is recommended, as an adequate ventricular filling is an essential step to achieve an adequate hemodynamic status23,31,32. However, fluid overload must be cautiously assessed, as an excessive preload may worsen right ventricular impairment30.

In patients with severe ventricular dysfunction, capillary leak or excessive fluid losses, a fluid bolus might be insufficient. Besides, it has not been demonstrated that the exclusive use of vasoactive drugs is capable of completely spare the need for fluid boluses. The ideal fluid would be one that, with the minimum amount of volume, achieves the longest intravascular permanence and the shift of interstitial fluid to the intravascular space with the minimum side effects on the endothelium and coagulation. The use of HSA administered at a dose of 5 ml/kg can meet several of the requirements of this ideal fluid.

According to the ESPNIC Recommendations for hemodynamic monitoring for critically ill children, it is recommended to observe the patient’s clinical situation, physical examination, and various perfusion indicators suggesting hypovolemia before considering fluid loading and to administer fluid therapy (5–10 ml/kg) as part of early resuscitation in unstable patients guided by the effect on blood pressure and/or cardiac output23, as we did in our study.

Hypernatremia, hyperchloremia and acidosis are well described side effects of hypertonic saline solutions33. Sodium and chloride blood levels mildly and transiently increased in our patients but returned to normal levels after a few hours. Metabolic acidosis after HSA was not observed. Moreover, pH improved over the 24 h following the infusion.

AKI has also been reported after the infusion of hyperoncotic colloids34. In this study, 22% (n = 5) of our patients developed AKI according to the KDIGO stratification: two had transient stage 1 AKI, and three of them were stage 3, as they required the initiation of CRRT within the 24 h following HSA bolus infusion. Nevertheless, AKI cannot be attributed to HSA alone, as they were all in the immediate postoperative period of complex cardiac surgery, hemodynamically unstable and had significant FO. Furthermore, the reported incidence of AKI in critically ill children reported by Kaddourah et al. in 2017 is 26.9%, and 11.6% for severe AKI35. The incidence of AKI in our unit is currently 25%, according to our local registry (unpublished data). Therefore, the incidence of AKI in this study is similar to that expected in this cohort. However, HSA bolus safety is still inconclusive and needs further research to conclude.

Hyperchloremia and the infusion of high amounts of chloride has been associated with AKI in critically ill children36,37. However, this association was not found in a retrospective study in 250 critically ill children38. On the other hand, albumin seems to have a protective effect on the glycocalyx and the kidney39. Therefore, the net effect of the potential benefits of using albumin and lower infusion volumes as opposed to the potential deleterious effects of hyperchloremia is unknown.

Given the fact that this is an observational pilot study, it is difficult to determine whether HSA bolus therapy could influence the management of net fluid balance. In this study, fluid balance switched from being largely positive in the shift prior to hypertonic albumin bolus to neutral in subsequent shifts. This finding cannot be attributed solely to HSA infusion; however, it seems to indicate that HSA bolus therapy does not pose a risk to the patients. Still, further studies are required regarding this topic.

Some colloids like starches or dextrans can cause coagulation and bleeding disorders by inhibiting platelet aggregation and the formation of fibrin polymers11,12,40. These alterations have not been described with albumin. Coagulation disorders were not observed in our study. 83% of our patients were admitted after cardiac surgery. Transfusion requirements in our study are similar to what Hanson et al. published from the REDS-IV-P study concerning the use of blood products in pediatric patients after cardiac surgery: 26% of patients required RBC, 22% required platelets and 23% of them received plasma infusion41.

Our cohort mortality rate was 13%, which is high considering that our global annual mortality rate is between 2 and 3%. However, this is a selected cohort with high severity scores and that required high vasoactive, respiratory and often circulatory support, and therefore, a higher need of fluid bolus administration.

Our study has several limitations. First, it is a single-center pilot study with a small sample size. As it is an observational study, it only evaluates the effect of HSA but does not allow comparisons between HSA and other fluids. Moreover, our patients were mainly children with a congenital heart disease, many of them in the immediate postoperative period of cardiac surgery. Therefore, the results might not be extrapolated to other populations of critically ill patients.

On the other side, the evaluation of intravascular volume has been performed indirectly through the evaluation of clinical signs. As we have mentioned before, the assessment of intravascular volume and the indication of fluid boluses in these patients is a complex decision, although we proceeded as recommended in the international guidelines on hemodynamic monitoring for critically ill children23.

Another major limitation is the presence of many confounding factors that affect hemodynamic status, such as vasoactive drug management. This is, therefore, a preliminary study that requires further consideration for future research.

Conclusions

Hypertonic saline albumin seems to be an effective and safe resuscitation fluid in critically ill children. It could have a special indication in hemodynamically unstable patients who do not tolerate high volumes of fluids and in those with fluid overload. However, our results must be interpreted with caution, as larger studies with different treatment groups, longer observation times and confounding factor control are needed.