Neonatal fluid overload—ignorance is no longer bliss

Abstract

Excessive accumulation of fluid may result in interstitial edema and multiorgan dysfunction. Over the past few decades, the detrimental impact of fluid overload has been further defined in adult and pediatric populations. Growing evidence highlights the importance of monitoring, preventing, managing, and treating fluid overload appropriately. Translating this knowledge to neonates is difficult as they have different disease pathophysiologies, and because neonatal physiology changes rapidly postnatally in many of the organ systems (i.e., skin, kidneys, and cardiovascular, pulmonary, and gastrointestinal). Thus, evaluations of the optimal targets for fluid balance need to consider the disease state as well as the gestational and postmenstrual age of the infant. Integration of what is known about neonatal fluid overload with individual alterations in physiology is imperative in clinical management. This comprehensive review will address what is known about the epidemiology and pathophysiology of neonatal fluid overload and highlight the known knowledge gaps. Finally, we provide clinical recommendations for monitoring, prevention, and treatment of fluid overload.

Introduction

Fluid overload in critically ill children and adults is associated with increased mortality, need for aggressive and prolonged ventilatory support due to pulmonary and chest wall edema, congestive heart failure, acute kidney injury (AKI), and prolonged hospitalization [1–6]. Emerging data on fluid overload in neonates suggests an increased risk of detrimental outcomes including mortality [7–10], bronchopulmonary dysplasia (BPD) [11, 12], intraventricular hemorrhage (IVH) [13], necrotizing enterocolitis (NEC) [14, 15], hemodynamically significant patent ductus arteriosus (PDA) [14, 16, 17], and AKI [5, 18, 19]. While evolving data about the negative impact of fluid overload in neonates seems to mirror pediatric data, it is imperative to understand these associations in the context of distinctive neonatal physiology [10].

Neonates are dependent upon clinicians to provide parenteral and enteral nutrition, medications, and blood products. Provision of these fluids should achieve the right fluid balance to prevent dehydration, maintain adequate tissue perfusion, while also avoiding fluid overload. Management of fluid balance homeostasis in any critically ill patient is challenging and may be even more difficult in neonates given the changes in fluid dynamics that evolve with developmental maturity [20–23]. The postnatal, physiologic, extracellular fluid volume (ECF) contraction is highly regulated by the kidney [24]. Premature infant kidneys have limited ability to strongly concentrate or dilute urine necessary to maintain fluid homeostasis and manage serum osmolality. Conditions such as AKI, sepsis, NEC, and cardiac dysfunction will further complicate the ability to provide nutrition, medications, and/or blood products without causing fluid overload.

Over the last two decades, much progress has been made to define, track, prevent, and treat fluid overload in adult and pediatric critical care units [25, 26]. Much of this work has been led by the pediatric critical care nephrologists. In 2001, Goldstein et al. coined the term fluid overload after demonstrating detrimental outcomes in critically ill pediatric patients requiring continuous kidney replacement therapy (CKRT) [25]. Since then, there have been multiple studies that have been summarized in a recent meta-analysis by Alobaidi et al., who report higher rates of in-hospital mortality with an independent odds ratio (OR) = 4.34 (95% confidence intervals (CI), 3.01–6.26) in those with fluid overload [2]. Only a small number of neonatal patients were included in this analysis. Since that meta-analysis, similar research in neonatal populations have begun to show similar outcomes.

The time has come for the neonatology and nephrology communities to improve our understanding and enhance strategies to prevent/treat fluid overload in critically ill neonates. With improved knowledge and recognition, clinicians will be better able to define, track, and study fluid overload. More importantly, early prevention and appropriate treatment strategies including timely kidney support therapy are likely to improve outcomes, but well-designed neonatal studies are needed. This review will highlight areas of active research and evolving gaps in knowledge as well as practical clinical guidelines for the prevention and management of fluid overload in neonates.

Postnatal fluid shifts and adaptations

After delivery, neonates undergo fluid volume contraction and cardiovascular adaptation as they transition from intra- to extra-uterine life. Preterm infants experience a more pronounced total body water (TBW) and ECF volume contraction [21, 22, 27]. This is in part due to differences in TBW, intracellular water (ICF), and ECF ratios that differ with gestational age. ECF composition at 24-week gestational age infant is ~ 60%, and it reduces to about ~ 40% in a full-term infant and transitions to ~ 20% as a young child (Fig. 1) [21, 28]. The TBW during the first postnatal weeks is influenced by multiple factors including gestational age, antenatal steroid exposure, growth restriction, kidney function, and transepidermal fluid losses [21]. Typically, full-term infants should not lose more than 10% of their birth weight in the first postnatal week, whereas between 10 and 20% weight loss may be beneficial among extremely preterm neonates [21, 27]. As a result, different targets of postnatal weight loss have been recommended depending on gestational age. The degree of salt and water diuresis occurring around 48–72 h after birth is influenced by enteral and parenteral intake, sodium supplementation, thermoregulation, and humidification [29].

Fig. 1.

The total body water shifts that occur with intracellular and extracellular fluid throughout gestation and into infancy. Re-printed with permission. Bell EF, Segar JL, William O (2016) Fluid and Electrolyte Management. https://obgynkey.com/fluid-and-electrolyte-management/. Accessed 15 November 2021

The clinician is responsible for fluid provision in preterm infants during the first postnatal weeks until they are able to ingest oral feeds on their own. The clinician must provide adequate nutrition, allow ECF volume contraction while avoiding hyponatremic fluid overload and hypernatremic dehydration. Higher body-surface area, a thin epidermis, dependence on caregivers, and limited ability to conserve water put neonates at higher risk of dehydration [21]. High humidification in modern incubators reduces insensible fluid losses. While fluid restriction may be beneficial in the first postnatal weeks, there is concern it can predispose an infant to a catabolic state through limited caloric intake. However, data has shown that appropriate caloric intake can still be achieved with lower total fluid intake [30]. Prescribed total fluid intake goals and daily fluid advancement remain controversial in practice and should be guided by serum sodium concentrations, weight change, and urine output as surrogate measures of fluid status [31]. Tight regulation and moderate restriction of fluid in premature neonates have been shown to be beneficial for long-term outcomes [13, 14, 31, 32].

Too much fluid administration affects postnatal cardiac and pulmonary adaptation. Warburton et al. performed serial echocardiograms in 73 very low birth weight infants randomized to receiving higher or lower volume maintenance fluids during the first 20 postnatal days [33]. This study showed that those randomized to higher volume fluids had larger left ventricular end-diastolic volumes and larger atrial to aortic root diameters [33]. These results suggest that increased fluid provision can impact cardiovascular function and prevent appropriate adaptation from uterine environment. In addition, fluid restriction strategies have been evaluated in a randomized clinical trial of 64 late preterm and full-term infants with transient tachypnea of the newborn (TTN). Fluid restriction with approximately 20 mL/kg/day less fluid in the first 24 postnatal hours, with a control standard of 80 mL/kg/day for premature and 60 mL/kg/day for term infants, shortened the duration of respiratory support among 26 infants with severe TTN by 38 h, but the duration of respiratory support comparing all study participants was not statistically significant [34], This study found no difference in serum sodium, creatinine, or weight loss between groups but urine output was 0.6 mL/kg/h lower among infants in the restricted group [34].

The association between increased fluid intake resulting in a lack of sufficient weight loss in the first postnatal week and the risk of hemodynamically significant PDA is well documented [14, 16, 17, 30]. A Cochrane review by Bell et al. included 5 randomized clinical trials comparing early fluid restriction versus liberal fluid intake among preterm infants in the early postnatal period. Four of the five articles commented on reduced risk of PDA with restricted fluid intake (relative risk (RR) 0.52, 95% CI 0.37 to 0.73; number needed to treat (NNT), 7; 4 trials, n = 526) [14]. In three of the trials, there was a higher percentage of postnatal weight loss with restricted fluid intake compared with liberal fluid intake (mean difference 1.94% of birth weight, 95% Cl 0.82 to 3.07; 3 trials, n = 326) [14]. In addition, the restricted volume group showed potentially clinically important but non-significant trends towards a lower risk of death (RR 0.81, 95% CI 0.54 to 1.23, 1 trial) [12], BPD (RR 0.85, 95% CI 0.63 to 1.14, 4 trials), and IVH (RR 0.74, 95% CI 0.48 to 1.14, 3 trials) [14]. This meta-analysis included relatively few infants less than 750-g birth weight and most of the studies were conducted in the era prior to surfactant, the widespread adoption of antenatal corticosteroids, and contemporary less invasive ventilatory support techniques [14]. Together, these classic studies conducted in neonates indicate improved outcomes with volume restriction to prevent fluid overload in the first days following birth among preterm infants.

Observational studies also indicate that early intervention to prevent fluid overload in the first postnatal days may help with the complex extrauterine transition and prevent poor outcomes such as hemodynamically significant PDA or BPD [13, 31, 32, 35]. Although there is a known association between PDA and BPD [35–38], the exact interplay between the two has not been fully elucidated. That being said, the timing of fluid restriction may be important in the development of BPD or PDA. Prolonged fluid restriction after the establishment of full enteral feeds has not been shown to be beneficial for prevention or long-term improvement of BPD [39]. A randomized controlled trial involving 60 infants (less than 1500 g and 32-week gestation) with chronic lung disease, as defined by oxygen requirement at postnatal day 28, calorically dense volume-restricted feedings did not improve short-term respiratory outcomes including the duration of ventilator support or oxygen therapy [40]. However, they did not report relevant clinical outcomes such as PDA. In addition, a recent randomized clinical trial of 224 very preterm infants found no increased risk of BPD or PDA among infants receiving higher volume enteral feedings after the establishment of full enteral feedings [41]. However, none of the infants enrolled in this study was less than 1000 g, who are at highest risk of BPD. The effect of volume restriction on major morbidities after the establishment of full enteral feedings in premature infants is unclear.

Defining and quantifying fluid overload in neonates

Calculation of fluid status relies on either a fluid balance–based approach or a weight-based approach. The fluid balance approach (cumulative fluid balance (%) = (cumulative fluid input (mL) – fluid output (mL))/weight (kg)) may be less practical in a neonatal intensive care unit setting as measurement of fluid balance in neonates is challenging [42, 43]. A cumulative fluid balance approach will not account for the insensible losses which can be a significant proportion of cumulative output in neonates. The degree of insensible losses is influenced by external factors including total body surface area exposure, ambient air or incubator temperature and/or humidification, and respiratory support [21, 42]. Furthermore, output can be difficult to accurately assess due to urine leakage around diapers, inaccurate measurement of urine mixed with stool, and emesis [42]. Placement of bladder catheters can be very difficult and urine collection bags may not be ergodynamically appropriate for premature neonates. Fluid balance–based approach may be useful in specific clinical settings such as the first postnatal week among extremely preterm infants prior to passing stool, but further validation of this method is needed [7]. Therefore, a weight-based approach to assess fluid status may be more practical in neonates.

A weight-based fluid balance formula (cumulative weight change (%) = (daily weight (kg) – birth weight (kg))/birth weight (kg)) has been shown in multiple studies to calculate fluid status and determine fluid overload accurately in neonates [32, 44–48]. However, the weight-based approach has its own challenges. Small fluctuations in weight from scale differences can have a significant impact. Critically ill neonates may be too unstable to be weighed daily. Both fluid balance–based and weight-based approaches may have inaccuracies. Each method may be useful for different patients.

In addition, besides determination of the cumulative fluid balance, the clinician must decide where the fluid is located and whether the fluid is maldistributed. Physical examination incorporating vital signs, evaluation of laboratory data, and imaging should be used to determine fluid balance assessment and evaluation of the location of the fluid (intravascular space vs. extravascular). The use of ultrasound has shown to be effective for identifying fluid overload in pediatric patients [49, 50]. Studies in neonates to determine whether technologies such as bioelectric impedance and ultrasound improve the assessment of fluid balance are needed.

Epidemiology of fluid overload in neonates: what is known?

Critically ill neonates

The Assessment of Worldwide Acute Kidney injury Epidemiology in Neonates (AWAKEN) was a 24-center international retrospective cohort study developed to further understand neonatal kidney disease [51]. The database included 2,189 neonates who received at least 2 days of intravenous fluids excluding infants who met preset criteria including congenital heart disease requiring urgent surgery or lethal congenital anomaly [51]. This database has been used to evaluate the short-term outcomes of early fluid overload in the neonatal population. Selewski et al. showed that a positive fluid balance was associated with mechanical ventilation on post-natal day 7 in both 645 infants ≥ 36-week gestation (OR 1.12, 95% CI 1.07 to 1.17) and in 1007 infants < 36-week gestation (OR 1.10, 95% CI 1.06 to 1.13) [45, 46]. Selewski et al. demonstrated a graded increase in the magnitude of fluid overload in relation to the number of neonates requiring mechanical ventilation [45].

Extremely low birth weight (ELBW) infants

Previous work among ELBW infants after the implementation of routine steroids and surfactant administration recognized that higher cumulative fluid balance early in the postnatal period is associated with the development of chronic lung disease [36–38]. Since that time, multiple retrospective studies, including more recent datasets and infants below 750 g, have also indicated that preterm infants with a high cumulative fluid balance and/or net weight gain in the first 3–10 postnatal days are at higher risk of adverse outcomes including prolonged mechanical ventilation [7, 45], higher rates of BPD [8, 10, 11, 35], and mortality [7–11]. Schmidt et al. performed a retrospective analysis of the data collected from the Trial of Indomethacin Prophylaxis in Preterm Infants (TIPP) which evaluated the efficacy of prophylactic indomethacin in ELBW infants. Unexpected outcomes of the trial showed that despite decreased PDA after prophylactic therapy the incidence of BPD was increased (43% vs. 30%, P = 0.015) [52]. Logistic regression analysis showed that infants treated with indomethacin had lower weight loss in the first week of life in comparison to those in the control group (4.8% (SD, 9.6) vs. 10.1% (SD, 7.6), P < 0.001) [52]. The authors postulated that decreased urine output, a known side effect of indomethacin, without any change in fluid administration may have resulted in fluid overload and an increased incidence of BPD [52].

Necrotizing enterocolitis (NEC)

Early fluid restriction may be helpful for the prevention of NEC. The Cochrane meta-analysis by Bell et al. included 5 randomized clinical trials comparing early postnatal fluid restriction in premature infants. Restricted fluid intake reduced the risk of NEC (typical RR 0.43, 95% CI 0.21 to 0.87; NNT, 20; 4 trials, n = 526) [14]. A randomized, controlled trial including 170 low birth weight infants from 751 to 2000 g demonstrated that increased fluid administration in the first postnatal week was associated with increased risk of NEC (18% vs. 0.04%, chi² = 8.53, P < 0.005) as diagnosed by radiographic or post-surgical histologic findings [15]. Although the exact pathophysiology remains unclear, several mechanisms have been proposed, such as increased intestinal wall edema, decreased perfusion due to aberrant flow through a ductus, and/or hypoxemia from excessive pulmonary fluid [15]. Mechanistic data on the role fluid overload plays in the development of NEC remain limited.

Fluid overload is a common complication among infants with a systemic inflammatory response to disease states such as NEC or sepsis. Sonntag et al. found that the degree of third-spaced fluid and multi-organ dysfunction were better predictors of outcome than the Bell classifications system [53]. Xie et al. performed a retrospective study on 172 neonatal patients with surgical NEC evaluating outcomes of low (< 25.87 mL/kg/h) versus high (> 25.87 mL/kg/h) intraoperative fluid administration based on a median fluid administration of 25.87 mL/kg/h [54]. Receiving more intraoperative fluid was associated with fewer postoperative complications including surgical site infections and delayed healing (OR = 0.54, 95% CI 0.29 to 0.99, P = 0.046) but higher mortality (P = 0.005) [54]. Of note, this study included only late preterm and term infants, and excluded neonates at highest risk of adverse outcomes from NEC [54]. Clinicians should monitor fluid balance in infants with NEC who are at high risk of fluid overload.

Extracorporeal membrane oxygenation (ECMO)

Pediatric patients supported by ECMO are at high risk of fluid overload and kidney dysfunction [55–58]. A number of retrospective, single-center, and multicenter studies of pediatric patients requiring ECMO have shown an association between the degree of fluid overload and mortality [19, 55, 56]. Therefore, fluid overload may be targeted as an independently modifiable risk factor for patients receiving ECMO. In a multicenter retrospective cohort study of 446 neonates, Murphy et al. suggested that underlying pathophysiologic processes place infants at different risks of fluid overload [19]. Infants with congenital diaphragmatic hernia are at particularly high risk of fluid overload [59] with higher peak percent of fluid overload in comparison to other disease states (51% vs. 28% cardiac vs. 32% respiratory; P < 0.01) [19]. Furthermore, given the high risk of concomitant kidney dysfunction in these critically ill infants, earlier use of adjunctive kidney support therapy (KST) may be beneficial. Gorga et al. evaluated the impact of fluid removal on mortality of 756 pediatric patients including 187 neonates supported by ECMO [58]. The degree of fluid overload at the time of initiation of KST was associated with increased mortality (adjusted odds ratio (aOR) 1.11, 95% CI 1.00 to 1.18, P = 0.05). It is possible that early intervention strategies to limit and treat fluid overload may improve outcomes among infants on ECMO.

Pathophysiology of fluid overload

Fluid overload is defined as pathologic accumulation of fluid. There are many potential causes of fluid overload in neonates. Regardless of the underlying cause, interstitial edema from extravasated fluid accumulates in multiple organs, which can impede capillary blood flow [60, 61] and cause organ dysfunction. As a result of organ congestion, organ perfusion can become compromised which can worsen the underlying disease process creating a positive feedback cycle [62]. Understanding the pathophysiology of fluid overload for each specific patient is crucial in implementing appropriate management and treatment strategies.

The kidney is intricately involved in maintaining cardiac and pulmonary function through regulation of preload, afterload, pH dynamics, and osmolality [62]. The effective circulating volume (preload) is maintained by kidney homeostatic mechanisms and erythropoietin production. The renin and angiotensin system modifies vascular tone (afterload). Acid–base balance is tightly controlled by the kidney to maintain organ cellular enzymatic function and oxygen delivery [62]. Kidney function can be compromised by fluid overload as interstitial edema can blunt overall function.

Growing evidence shows that AKI may negatively affect lung function through multiple mechanisms including suboptimal fluid homeostasis, regulation of inflammation, and maintenance of acid–base balance [62]. AKI is common in neonates, occurring in up to 48% of preterm infants 22 to < 29 weeks of gestation, 18% of preterm infants 29 to < 36 weeks, and in 37% of infants 36 weeks or older [63]. Evidence has shown that infants with AKI were more likely to have higher fluid balance in the first week of life [45, 46, 48, 64]. Fluid overload and AKI are both independently and synergistically associated with poor outcomes in critically ill children [5, 18, 19, 65]. It is important to delineate the two, as these are independently modifiable risk factors [5, 19, 48]. However among infants at risk of fluid overload, these data stress the importance of minimizing nephrotoxic insults, monitoring for signs of AKI, and preventing propagation of AKI.

Vascular wall integrity plays an important role in maintenance of intravascular volume. The endothelial glycocalyx (EGL), a luminal surface-bound collection of glycoproteins and proteoglycans, helps to maintain vascular integrity, regulate inflammation, and store non-circulating plasma volume [66, 67]. The EGL layer supports the oncotic pressure and attenuates plasma leak across the membrane while preventing any reabsorption from the surrounding interstitium [60, 67]. Most of the interstitial fluid resorption is facilitated by lymphatics. Breakdown of the EGL layer can occur through a number of processes including sepsis, hyperglycemia, surgery, and ischemia [66, 68]. Once the EGL is damaged, fluid efflux is dependent upon capillary hydrostatic pressure, while interstitial fluid reabsorption remains relatively constant. Fluid overload may result from third spacing in the setting of capillary leak syndrome due to systemic inflammatory response including complement activation and EGL breakdown [53]. This process leaves patients at high risk of intravascular fluid depletion with extravasation of proteins and accumulation of interstitial fluid [60]. A high mortality rate has been reported among neonates and children with capillary leak syndrome [53].

As further transvascular fluid loss occurs and intravascular volume falls, it is initially possible to restore kidney and end-organ perfusion with volume resuscitation. However, there is a delicate balance between intravascular volume replacement and propagation of extravasation of fluid interstitially in the setting of compromised vascular integrity and capillary leak [60]. Therefore, careful selection of volume replacement media is important. Continued use of isotonic fluids such as normal saline and albumin 5% beyond the initial resuscitation may subsequently accentuate fluid overload as these fluids leak interstitially. A number of studies have shown that 5% albumin is not superior to normal saline as a resuscitation fluid [69, 70]. A randomized controlled trial comparing 5% albumin to isotonic saline in 63 hypotensive premature infants showed that there was no difference in the volume required, inotropic support, or mortality. However, infants receiving the 5% albumin had significantly more weight gain (5.9% vs. 0.9%, P = 0.05) in the first 48 h after birth [70].

Neonates have an increased risk of hypoalbuminemia due to slowed synthesis of albumin from immature liver function and increased degradation. There are not clearly defined normal ranges of albumin in neonates, but levels are considered to be inversely proportional to gestational age. The most widely accepted definition of moderate hypoalbuminemia is < 2.5 mg/dL and severe is < 2.0 mg/dL. Moderate hypoalbuminemia has been associated with AKI [71] and mortality in premature infants with sepsis [72]. The clinician must weigh the risks and benefits of albumin infusions as there is a theoretical risk of exacerbating systemic hypertension, protein extravasation, and pulmonary edema [73]. A randomized controlled trial comparing 5 mL/kg of 20% albumin to 5 mL/kg of maintenance fluids in 40 critically ill ventilated preterm infants (median gestational age 29 weeks) with hypoalbuminemia (< 3.0 mg/dL) reported no difference for reduction in weight or ventilator requirement between the two [74].

It may be preferable to use fluids with higher oncotic pressure such as albumin 20%, fresh frozen plasma (FFP), or packed red blood cells (pRBC) to improve intravascular volume and decrease the requirement for isotonic fluid boluses [75, 76]. A double-blind, randomized controlled trial of 76 pediatric and infant patients with congenital heart disease requiring surgical management and cardio-pulmonary bypass compared high oncotic priming with 20% albumin versus control [77]. Patients receiving pre-treatment with higher oncotic fluid had less hypotension (8% vs. 54%, P = 0.02) and required fewer fluid boluses (6% vs. 54%, P = 0.01) [77]. These data, although referring to priming fluid instead of media for resuscitation, suggest that use of 20% albumin may be beneficial over normal saline for maintaining intravascular volume.

There is limited data on the use of FFP as volume replacement in symptomatic infants. A Cochrane review evaluated the effect of early prophylactic volume expansion with FFP in premature infants [78]. This meta-analysis reviewed 4 randomized trials comparing infusion of FFP versus control regardless of cardiovascular status. No improvement in mortality (RR 1.05, 95% CI 0.81–1.36, 3 trials, N = 654) or severe disability was noted suggesting that routine use of FFP for volume expansion is not beneficial. Blood transfusion may be preferred to maintain intravascular volume over normal saline or 5% albumin in symptomatic patients with anemia. A recent multicenter, double-blind, placebo-controlled randomized trial of 941 extremely preterm neonates showed that high dose erythropoietin (1000 U/kg) every 48 h for 6 doses followed by three times weekly maintenance dosing (400 U/kg) until 32 weeks post-menstrual age was safe and reduced the number of pRBC transfusions although it did not improve the incidence of death or severe neurodevelopmental impairment (RR 1.03; 95% CI 0.81 to 1.32; P = 0.80) [79]. In an ancillary project to that trial, infants randomized to receive erythropoietin did not have a reduction in AKI [80].

Management of fluid overload

The approach to the management of the infant with fluid overload is informed by the aforementioned pathophysiology of fluid overload in neonates. We used evidence-based potentially-better-practices to develop the management strategy for fluid overload summarized into a mnemonic to allow for ease of use in Table 1.

Table 1.

A mnemonic “CAN-U-P-LOTS” summarizing evidence-based potentially-better-practices that may be used in clinical practice to treat patients with fluid overload

CAN-U-P-LOTS Mnemonic
Cause	Determine underlying etiology   - Abdominal compartment syndrome   - Cardiac (congenital heart disease, patent ductus arteriosus, heart failure)   - Congenital anomalies of kidney and urinary tract (CAKUT)   - Dehydration   - Hypoalbuminemia   - Hyperuricemia   - Shock (sepsis, necrotizing enterocolitis, or hypoxic ischemic injury)
Albumin	Treat with 20–25% albumin:   - 2 g/kg/dose over 4 h if albumin < 2.0 mg/dL   - 1 g/kg/dose over 4 h if 2.0 to < 2.5 mg/dL
Nephrotoxicity	Assess medications   - Avoid or switch to less nephrotoxic medications   - Follow levels closely if nephrotoxic medications needed
Ultrafiltration	Kidney support therapy via peritoneal or extracorporeal approach
Perfusion	Determine and treat cause of shock   - Adrenal   - Cardiac   - Hypovolemia   - Neurologic   - Sepsis   - Titrate mean arterial pressure (MAP) goals to achieve urine output
Lasix stress test	Furosemide stress test to assess kidney response
Output	Assess urine output trends Place foley catheter Consider bladder obstruction
Total fluid intake	Determine dry weight for dosing and calculating fluid balance Set fluid goals Concentrate all medications and nutrition Maintain adequate nutrition and intravascular volume
Steroids	Add stress dose hydrocortisone for refractory hypotension

It is important to consider the differential diagnosis and determine the most likely causes of fluid overload, including potentially reversible causes of oliguria/anuria as demonstrated in Table 2. The pathophysiology and treatment recommendations may differ for a patient at risk of developing fluid overload depending upon the underlying etiology. Infants with abdominal diseases, such as NEC and spontaneous intestinal perforation, are at risk of abdominal compartment syndrome which can present as fluid overload and oliguria. In an infant with a distended abdomen, intraabdominal pressure can be assessed by transducing a catheter placed in the bladder. Cardiac causes include a PDA causing vascular steal, congenital heart disease, and heart failure can be assessed with echocardiography. Most cases of congenital anomalies of kidney and urinary tract (CAKUT) or bladder outlet obstruction are now diagnosed prenatally but can be suspected in infants with low urine output and signs of pulmonary hypoplasia or pulmonary hypertension after birth. Hypoalbuminemia can be primary or secondary and can cause or worsen fluid overload and AKI. Hyperuricemia is a rare cause of fluid overload and AKI that can be treated with rasburicase.

Table 2.

A list of potentially reversible causes of oliguria/anuria in neonates

Potential reversible causes of oliguria/anuria
Abdominal compartment syndrome
Bladder outlet obstruction
Hypoalbuminemia
Inadequate kidney perfusion due to cardiac dysfunction, dehydration, mineralocorticoid deficiency, shock
Nephrotoxic medications
Hyperuricemia

Table 3 outlines several different scenarios for at-risk patients. The underlying pathophysiology in each may differ from another but ultimately they all result in extravascular fluid accumulation and progression of cumulative fluid overload. Importantly, there are clinical scenarios that may not be classified into one certain scenario. For example, an infant with NEC may develop sepsis and a systemic inflammatory response resulting in hypotension, oliguric kidney failure, capillary leak syndrome, low oncotic pressure or hypoalbuminemia, and ultimately heart failure, depending on the severity of disease.

Table 3.

Clinical scenarios in which a patient is at high risk of fluid overload. The table outlines each different pathophysiologic reason for fluid overload that must be considered when developing a treatment plan

Scenarios	Intravascular fluid	Extravascular fluid	Capillary leak	Oncotic pressure	Systemic pressure	Treatment
Heart failure	High	Yes	No	Normal	High, normal, or low	Fluid removal
Oliguric kidney failure	High	Yes	No	Normal	High	Fluid removal
Sepsis	Low	Yes	Yes	Normal	Low	Prevention
Hypoalbuminemia	Low	Yes	Yes	Down	Low	Albumin

Setting and achieving fluid balance goals

A patient is at risk of developing fluid overload when their fluid balance is no longer in homeostasis. This can be due to increased intake or decreased output or a combination of both. The main ways to achieve a net negative fluid balance are through limited fluid intake and/or by increasing fluid removal. Prevention may be the best treatment for fluid overload. Prevention of fluid overload progression can be achieved with vigilance, mindful fluid provision, assuring that all fluids (including parenteral nutrition) is concentrated and daily tracking of cumulative fluid balance in relation to clinical parameters. In patients at risk of fluid overload, the clinician should consider the following:

1. Calculate and monitor the trends in the cumulative fluid balance (i.e., current weight – dry weight / dry weight × 100) over time.

2. Intravascular volume status should be assessed by physical examination including evaluation of perfusion and vital sign trends, as well as by investigations of electrolytes, BUN, creatinine, serum albumin, and imaging.

3. Develop fluid balance goals based on intravascular volume status. Re-evaluate frequently to ensure that goal is being met.

4. Evaluate all forms of fluid provision and eliminate or concentrate all fluids (including nutrition) to achieve lowest possible fluid intake as able.

5. Evaluate and manage reversible causes of oliguria/anuria as demonstrated in Table 2. Different treatments should be applied to each separate disease process.

6. Communicate with pediatric nephrologist early.

Enhanced fluid removal by improving urine output can be achieved by optimizing kidney perfusion, diuretics, or both. Evaluation of intravascular volume and kidney perfusion will help determine therapeutic goals. It is imperative to maintain adequate kidney perfusion through improved cardiac output and blood pressure. While volume resuscitation may be necessary to achieve adequate intravascular volume, once euvolemia is achieved further volume expansion can be harmful. Achieving effective intravascular volume requires evaluation of volume status, determination of the appropriate fluids for volume repletion, and knowledge of the underlying pathophysiology. A trial of volume expansion or a “bolus” to increase kidney perfusion may be helpful in improving urine output. If further volume is required beyond the initial resuscitation, consideration to give 20–25% albumin 1–2 g/kg if hypoalbuminemic or FFP 10–20 mL/kg to increase oncotic pressure. Our typical practice includes treatment of moderate hypoalbuminemia (< 2.5 mg/dL) with 1 g/kg of 20% albumin, or 2 g/kg of 20% albumin over 4 h to treat persistent severe hypoalbuminemia (< 2.0 mg/dL). Other blood products can be given judiciously based on individual clinical circumstances such as anemia, coagulopathy, thrombocytopenia, or hypogammaglobulinemia. Once optimal intravascular volume has been achieved and kidney perfusion has been optimized, the goal for fluid balance management should be aimed to limit fluid overload until removal of fluid is possible.

Diuretics are one of the most prescribed medications in the neonatal intensive care unit [81]. Despite frequent usage, there is little evidence to support routine use of diuretic therapy in neonates [28, 82, 83]. In the setting of progressive oliguria or anuria, a trial of diuretics can be attempted to improve or stimulate diuresis [28, 83]. Emerging data shows a “furosemide stress test” (FST) can be very valuable in predicting AKI progression [84]. To perform the FST, a standard dose of diuretics (i.e., 1 mg/kg of intravenous furosemide) is given to stimulate a diuretic response. If the subject has a good response (i.e., > 1 mg/kg/h) over the following 2 h, the prognosis for recovery is good. However, if after an FST there is not a good response, AKI progression is likely to occur and the clinical team should begin to consider other options such as improving perfusion and/or oncotic pressure as well as adopting strategies to prevent further fluid accumulation before further trials of diuretics are considered [84, 85]. Bumetanide has also been used to achieve negative fluid balance in critically ill children [86] and neonates if furosemide therapy has not improved urine output [87]. More evidence is needed for choice of therapy, optimal dosing, administration method (bolus vs. drip), duration of diuretic treatment, and optimization of a neonatal FST.

Aside from further volume expansion or diuretic therapy, earlier use of medication to improve cardiac output that improves kidney perfusion may be needed. Neonatal blood pressure standards are not well defined. Several studies have attempted to further define optimal blood pressure goals in neonates [88–91]. Current accepted ranges are mean arterial pressure (MAP) greater than gestational age [92]. In a patient with oliguria or anuria, targeting higher MAPs may be necessary to achieve adequate kidney perfusion, especially if there is an increase in abdominal pressure. Ongoing data is emerging about appropriate first line vasopressor in neonates [93–97]. Recent meta-analysis showed that dopamine had the highest success of improving MAP (RR = 0.88, 95% CI: 0.76 to 0.94; 12 studies; N = 163) in neonates [93]. Dobutamine is the second most commonly used and is thought to have less peripheral vasoconstrictive effects.

In a Cochrane review of dopamine in comparison to dobutamine, dopamine was more effective, as shown by reduced treatment failure (RR 0.41, 95% CI: 0.25 to 0.65), but there was no difference in mortality outcomes [94]. In a meta-analysis including the data from a Cochrane review and two randomized controlled trials comparing dopamine to epinephrine as a first-line vasopressor, there was no difference in treatment efficacy [93]. One study included in this meta-analysis noted more significant side effects with epinephrine use, including higher heart rates, elevated lactates, and hyperglycemia [93, 98]. This review does not include an exhaustive list of vasoactive medications useful to treat a hypotensive neonate: a more comprehensive review can be found via the aforementioned Cochrane review [94]. More data are needed about the best assessment of perfusion status and the optimal regimen and dosing of these medications to prevent and treat fluid overload in neonates.

In addition to inotropic support, hydrocortisone is beneficial in refractory hypotension [99, 100] and may improve cardiovascular function in the setting of fluid overload. Premature infants have a blunted adrenal response until approximately 32 weeks gestational age (wga) and may benefit from cortisol replacement therapy [101]. Cortisol may be of benefit in the setting of fluid overload through improved myocardial function, enhanced vascular responsiveness to angiotensin II, and decreased capillary leak through effects on luminal integrity [101]. A retrospective analysis by Tolia et al. evaluated the effect of stress dose hydrocortisone in the first 14 postnatal days in 1427 infants less than 30 wga [102]. Higher dose (> 2 mg/kg) in comparison to lower dose (≤ 2 mg/kg) hydrocortisone was associated with increased mortality (aOR 3.27, 95% CI 2.47 to 4.34, P < 0.001) [102]. While stress dose hydrocortisone may be helpful in hypotension and fluid overload, further data are needed to support dosing recommendations and long term outcomes.

Kidney support therapy in neonates

If fluid restriction, diuretic trials, and optimization of hemodynamics are not able to achieve the desired fluid balance goals, kidney support therapy (KST) should be considered [103]. Neonates can receive KST via a peritoneal dialysis (PD) or an extracorporeal approach to treat or prevent fluid overload refractory to diuresis. Like all other medical decisions, the evaluation of the potential benefits and risk of the procedures are needed to make clinical decisions. In a patient with balanced electrolytes and relatively normal kidney function, KST may be considered as the primary therapy for fluid removal over diuretic therapy if the fluid accumulation trend is severe and the patient is not able to achieve the fluid goals with diuretics. The right time to start KST is when the impending harm from the inability of the kidney to maintain fluid, electrolytes, and toxin homeostasis outweighs the risks of KST. In this context, it is vital to recognize the goals, risks, and benefits of KST.

The risks associated with extracorporeal KST have been greatly reduced with the advent of neonatal-specific kidney support devices. Until recently, use of extracorporeal KST in neonates had to be adapted from adult machines which presented a number of challenges due to the large filter/tubing volumes which lead to high priming volume in relation to patient size, difficulty in maintaining adequate blood flow for optimal circuit function, hemodynamic instability during initiation, and challenges related to vascular access [103]. Many of the recent advances in the reduction of risk in neonatal KST are due to smaller tubing/filter volumes, which allow for lower priming volumes [103, 104].

Reports of the use of KST in extremely premature infants have mostly been limited to PD, primarily due to patient size preventing catheter placement in small caliber blood vessels. However, the use of PD in the setting of fluid overload is often limited by the low, unpredictable ultrafiltration and increased intraabdominal permeability that predisposes them to severe hyperglycemia [105]. Although complications from PD are not very common, the associated risks include catheter malfunction, infection, electrolyte, and blood pressure abnormalities. Catheter malfunction and infection can be greatly reduced with appropriate surgical technique and by allowing time for the catheter insertion site to heal before use. PD is contraindicated in a number of patients at high risk of fluid overload, including perforated NEC and patients with congenital diaphragmatic hernia [106]. PD is commonly used in neonates after cardio-pulmonary bypass surgery. A recent review article evaluating data on early initiation of PD in neonates who undergo cardio-pulmonary bypass surgery noted a trend towards reduced fluid overload and mortality but inconclusive evidence about the optimal timing of initiation following surgery [107].

The CARdiorenal PEDIatric Emergency Machine (CARPEDIEM) (Medtronic, Minneapolis, MN) achieves fluid, electrolyte, and waste product clearance with minimal complications. Multiple studies have validated the use of this device in the neonatal population down to infants weighing 2 kg [105, 108]. This system has very precise scales, works in either dialysis or hemofiltration mode, must be changed every 24 h, and has been recently approved for use in the USA by the FDA for children between 2.0 and 9.9 kg.

The Aquadex system (Nuwellis Inc., Eden Prairie, MN) for ultrafiltration was adapted for use in the neonatal population after its original use in adults with congestive heart failure. This system has an extracorporeal volume of 33 ml (mL) and provides continuous hematocrit and mixed venous oxygen saturation monitoring during the treatment. In combination with an external fluid delivery system, Askenazi et al. first described the ability to remove waste products, balance electrolytes, and remove fluid in continuous venovenous hemofiltration mode in small children [109]. A more recent multicenter study evaluated the ability of Aquadex to provide effective prolonged-intermittent and slow-continuous KST. Importantly, they found hemodynamic instability in only 3% of circuit initiation, and the few episodes were mild in nature [110].

The Newcastle Infant Dialysis and Ultrafiltration System (NIDUS) (Newcastle, England), another KST machine developed for small children, has an extracorporeal volume of 6–10 mL (depending on the stroke volume used). The system uses a single lumen access, which may allow safe and effective therapy even in ELBW infants. The machine uses diffusive clearance, in contrast to the two earlier mentioned modalities [111]. As of 2021, the machine was not available for commercial use, awaiting the conclusion of a multi-center clustered wedge-shaped clinical study.

In larger children and adults, the degree of fluid overload at the time of initiation is independently associated with mortality [47, 55, 58, 65, 112]. Thus, it is very plausible that this is also true in neonates. As technology has improved, the risk of hemodynamic instability during KST in neonates has sharply decreased. Each modality may be useful for different patient scenarios depending upon clinical need. The benefits, limitations, and potential complications of KST options are displayed in Table 4. With a reduction in complications, the risk to benefit ratio changes the equation such that early use of early KST in neonates should be considered as a therapeutic option as part of a comprehensive strategy to prevent and treat fluid overload. Further investigation on the optimal timing of initiation and delivery of these therapies in neonates is warranted.

Table 4.

A table representing the capabilities and limitations of kidney support therapies available to neonates. NEC, necrotizing enterocolitis

Kidney support modality	Benefits	Limitations	Complications	Mode of clearance
Peritoneal dialysis	- Ease of use - No vascular access - Can be used if hemodynamically unstable	- Limited precision of the exact amount of fluid and waste product removal - Unable to use if intraabdominal complications (e.g., NEC, intrabdominal surgery, congenital diaphragmatic hernia) - Must wait until catheter heals to use for prolonged use	- Catheter migration - Catheter malfunction - Catheter infection - Peritonitis - Hyperglycemia - Hyponatremia - Hypoalbuminemia	Diffusion
Extracorporeal support (CARPEDIEM, Nidus, Aquadex)	- Highly efficient waste product clearance - Precise fluid removal - Low extracorporeal volume filter/tubing enables smaller catheters, and decreases hemodynamic instability compared to traditional larger KST circuits - May start therapy immediately after catheter placement	- Requires vascular access - May be restricted based on size of vessel - Requires specialized machines - Requires higher expertise/training - Requires anti-coagulation of the circuit	- Catheter migration - Catheter thrombosis - Catheter infections - Hemodynamic instability (although this is much lower than traditional KST machines)	Diffusion or convection (CARPEDIEM) Diffusion (Nidus) Convection (Aquadex)

Conclusions and future directions

Increased clinician awareness of the impact of fluid overload will help to improve recognition and implementation of therapies aimed at reducing fluid accumulation. Early and targeted management of fluid overload may help providers improve outcomes. In addition, KST with machines that have smaller extracorporeal volume can now be used to prevent or treat fluid excess in infants who do not respond to medical therapy. Basic, translational, and clinical studies are needed to close remaining knowledge gaps and improve our current limited understanding of neonatal fluid overload.