Metabolic acidosis is common in preterm and term newborn infants secondary to various conditions such as asphyxia, sepsis and persistent pulmonary hypertension of the newborn (PPHN). Pediatricians have attempted to address metabolic acidosis by adding base to intravenous solutions for almost a century [1]. Effective management of metabolic acidosis and optimization of intravenous solutions for use in pediatrics has an interesting history. In this article, we have reviewed the history of use of base in newborns, the potential benefits and harms of adding base to correct metabolic acidosis in newborns and the current knowledge gaps.

History of intravenous solutions and base in Pediatrics

Sydney Ringer (1834–1910) worked for a short period of time at the Hospital for Sick Children in Great Ormond Street and Brompton Hospital for Diseases of the Chest. When his senior colleague, Sir William Jenner reprimanded him for holding posts in two hospitals, Ringer (unfortunately) resigned from the Children’s Hospital. Ringer was interested in the action of various salts on cardiac function. While experimenting on a frog’s heart [2,3,4], his assistant mistakenly substituted tap water (supplied by the New River Water Co which had 278.6 ppm of solids including 38.9 ppm of calcium) for distilled water [5, 6]. Being an astute observer, Sydney Ringer realized that the presence of small quantities of inorganic salts of sodium, potassium, calcium and chloride had an impact on cardiac function [6]. Ringer’s solution was formulated as a result of this discovery. He further emphasized the importance of small quantities of salt by showing that fish could not survive in distilled water but adding minute quantities of river water was sufficient to keep them alive.

Alexis Frank Hartmann (1898–1964) was an American pediatrician and clinical biochemist. He identified the need for alkali therapy to correct acidosis in children and advocated the use of sodium lactate as an option to reduce chloride concentrations. Sodium lactate would normally convert to NaHCO3 over a 2-h period (Fig. 1) and a liter of 1/6 M sodium lactate had the same acid neutralizing effect of 290 ml of 5% NaHCO3. By addition of lactate to the original Ringer’s solution, the Hartmann’s solution or “lactated” Ringer’s solution (LR) had the following composition: 131 mEq/L of sodium, 5 mEq/L of potassium, 4 mEq/L of calcium, 29 mEq/L of lactate and 111 mEq/L of chloride [1]. It is important to remember that the LR solution or the Hartmann solution as we currently use it, was formulated by a pediatrician with an intention to correct metabolic acidosis in sick children [7]. Sodium bicarbonate is hard to sterilize, interacts with calcium and produces rapid fluctuations in pH and is not an ideal constituent of bolus or maintenance solutions. However, sodium lactate when added to the hypotonic Ringer’s original solution was stable and allowed slow conversion to bicarbonate in the liver (Fig. 1) [8].

Fig. 1: Mechanism of action of bases in treatment of metabolic acidosis.

Administration of Ringer’s lactate solution is followed by the conversion of lactate to pyruvate by lactate dehydrogenase. Pyruvate is then converted to acetyl CoA, which enters the tricarboxylic acid (TCA) cycle (also known as the Krebs cycle), ultimately resulting in the generation of HCO3− that enters the circulation and CO2 that is exhaled by the lungs. The administration of intravenous acetate and oral citrate provides substrates for acetyl CoA and citric acid in the TCA cycle, respectively. Copyright Satyan Lakshminrusimha.

Although, NaHCO3 was initially introduced as a therapeutic agent for the management of malignant cholera [9], the widespread use of intravenous NaHCO3 in the neonatal intensive care unit (NICU) began during the 1950s and 1960s. Sodium bicarbonate was first commercially produced in the 1950s, and its use in premature neonates became prevalent for preventing azotemia, elevations in serum potassium levels, as well as for correcting metabolic acidosis (Usher-regimen) [10]. However, Corbet et al observed a slight increase in deaths in a controlled trial of NaHCO3 in high-risk preterm infants [11]. Usher himself recognized the association between rapid NaHCO3 therapy and intraventricular hemorrhage (IVH) in preterm babies [12]. Wigglesworth also suspected a direct link between alkali use and IVH in infants with hyaline membrane disease [13].

Animal studies showed that NaHCO3 and glucose infusions prolonged survival after asphyxia and possibly lessened the degree of cerebral injury [14]. These studies, along with others, suggested the utility of NaHCO3 in neonatal resuscitation and was listed in the Medications chapter of the initial editions of the Textbook of Neonatal Resuscitation. Starting in the 1970s, other studies emerged showing that NaHCO3 may not be beneficial and in fact may be harmful. For example, a study by Ostrea et al. showed the paradoxical effect of increasing PCO2 and decreasing intracellular pH after administration of NaHCO3 in the absence of adequate ventilation (Fig. 2) [15]. In 2006, a Cochrane review of randomized controlled trials (RCT) found insufficient evidence to determine whether the infusion of NaHCO3 during neonatal resuscitation reduces mortality and morbidity [16]. Currently, NaHCO3 is not recommended during neonatal resuscitation but is used selectively and sparingly in the NICU [17,18,19].

Fig. 2: Metabolic effects or rapid administration of intravenous sodium bicarbonate.

Although intravascular correction of pH and alkalosis occurs, due to better diffusion of carbon dioxide into the cells (compared to bicarbonate ion), intracellular acidosis can occur, especially in cardiomyocytes. High osmolar load and fluctuations in cerebral blood flow can contribute to intraventricular hemorrhage. Copyright Satyan Lakshminrusimha.

The use of acetate as a buffer to correct metabolic acidosis crept into neonatal practice towards the end of the 20th century. Use of sodium acetate is associated with higher pH and reduced need for NaHCO3 in preterm infants [20, 21]. Similar to sodium lactate, acetate has to be metabolized to carbon dioxide through the Krebs cycle and contributes to HCO3 formation (Fig. 1).

Acid-base homeostasis mechanism

Acid-base homeostasis is tightly regulated by neonates through several mechanisms including buffer systems, respiratory and renal regulation [5]. The etiologies for metabolic acidosis (Fig. 3 and Table 1) include, a) loss of base (gastrointestinal- including ostomy losses or renal, or post-operative losses), b) excessive intake of H+ ions beyond what the kidneys can eliminate, c) excessive generation of H+ ions from abnormal cellular metabolism, especially lactic acidosis in hypoxia and sepsis. In critically ill infants, severe tissue-level hypoxia secondary to either poor perfusion or oxygen delivery can shift metabolism towards anerobic glycolysis with the generation of H+ ions and lower ATP production, and subsequent production of lactic acid [22].

Fig. 3: Henderson–Hasselbach equation.

The Henderson–Hasselbalch equation helps assess the acid-base status of a patient and determine whether respiratory or metabolic factors are contributing to an acid-base imbalance. The yellow (excess acid production), green (loss of bicarbonate), and blue (respiratory acidosis) boxes depict the common causes of acidosis in newborns.

Table 1 Neonatal metabolic acidosis mnemonic.

A serum pH in the range of 7.35–7.45 provides the ideal environment for cellular metabolism, making the maintenance of a stable pH critical for homeostasis. Buffer solutions resist changes in pH when acid or alkali are added due to their ability to absorb or release H+ ions. The three main buffer systems are the HCO3, phosphate and protein systems [23]. The protein and phosphate systems are intracellular, where extracellular H+ is exchanged for sodium and potassium ions. The bicarbonate system is extracellular and consists of a mixture of carbonic acid (H2CO3) and NaHCO3. This extracellular buffer system works by combining excess H+ with HCO3−, forming H2CO3, which dissociates to give CO2 and H2O (Fig. 2). The CO2 is then excreted from the lungs through exhalation (\(^+_^\,\longleftrightarrow _}_\longleftrightarrow _O+C_\)). As shown in the Henderson–Hasselbach equation (Fig. 3), exhalation of CO2 shifts the reaction to the right allowing HCO3 to play an important role in buffering the H+ ion and reducing acidosis.

Interpretation of base excess on an arterial blood gas

Base excess on an arterial blood gas can help clinicians understand the overall acid-base balance in a neonate. However, base excess should not be interpreted in isolation when determining acid-base status of an infant. Base deficit is a theoretical estimate that is calculated by a blood gas analyzer utilizing measured values of pH, PaCO2 and hemoglobin. Some blood gas machines report two different values for base deficit or base excess: blood and ECF. Base deficit in blood is the amount of base that needs to be added or extracted (i.e., base excess) to a liter of blood to shift the plasma pH to 7.4 at a temperature of 37 °C and PaCO2 of 40 mm Hg [24]. As blood gas machines calculate base excess using measured values of pH and PCO2, acute changes in PCO2 without changing serum bicarbonate can alter calculated base excess values. For example, a term infant with perinatal acidosis may have an ABG with a pH of 7.1, PaCO2 of 40 mmHg, bicarbonate of 12 mEq/L and a calculated base excess of −15.7 mEq/L. If this infant develops acute endotracheal tube obstruction with an increase in PaCO2 to 60 mmHg without any change in serum bicarbonate, the pH decreases to 6.92 and calculated base excess will be −18.2 mEq/L.

On the contrary, the standard base deficit in the extracellular fluid (ECF) consists of blood volume diluted with the interstitial fluid with a standardized hemoglobin concentration of 5 g/dL (a low concentration of hemoglobin is utilized in this calculation to represent dilution by ECF). This base deficit in ECF refers to the metabolic (or non-respiratory) component of base deficit, as it is minimally affected by changes in PaCO2 due to dilution by interstitial fluid, compared to the base deficit in blood that is increased by an increase in PaCO2 and the conversion of CO2 to HCO3 by erythrocytes [24].

Various factors can affect the reported value of base excess, including inconsistencies between different equations used to calculate base excess, such as the Siggaard-Andersen equation (=0.9287 × (HCO3 − 24.4 + 14.83 (pH − 7.4)) and the Van Slyke equation (=cHCO3 − 24.4 + (2.3 × hematocrit + 7.7) × (pH − 7.4) × (1 − 0.023 × hematocrit)), as well as clinicians’ not knowing which equation is used by their clinical labratory [25]. Additionally, the unique aspects of fetal acid-base physiology, which are not accounted for in these equations, can lead to inaccurate values [25]. For example, these equations do not account for acid-base status of newborns at birth and rely on adult acid-base physiology norms (pH 7.4, pCO2 40 mmHg and HCO3 levels ranging from 24–24.8 mmol/L) as well as adult hemoglobin characteristics [25]. This approach fails to reflect the higher hemoglobin levels found in newborns (~17 g/dL in term newborns compared to ~12 g/dL in adults) and does not consider properties of fetal hemoglobin [25]. Hence it is important to check bicarbonate level on a serum chemistry sample in addition to base deficit on a blood gas before deciding to correct acidosis.

In infants with perinatal asphyxia, metabolic acidosis with pH <7 or base deficit >−12 is included as an essential criteria for the diagnosis of moderate-to severe HIE, along with early onset of encephalopathy, multisystem dysfunction, and exclusion of other etiology such as trauma, coagulation disorders, metabolic disorders and genetic causes [26]. Furthermore, Puthuraya et al demonstrated that time to recovery of base deficit was observed to be longer across the severity of HIE in a cohort study [27]. It is critically important that the umbilical arterial cord blood sample is utilized rather than umbilical venous to assess the base excess, as the CO2 is effectively cleared by the placenta in the blood returning to the fetus in the umbilical vein and the metabolic component is overestimated due to the lower PCO2 resulting in falsely elevated base excess [24].

Normal Saline vs. LR for neonatal resuscitation

Previous editions of the textbook of neonatal resuscitation recommended the use of normal saline or LR as a crystalloid for volume bolus. The more recent editions have eliminated LR as a choice. Compared to normal saline, LR is more physiological with a sodium and chloride concentration like that of plasma. The pH of normal saline is significantly lower, contributing to hyperchloremic metabolic acidosis. However, LR is not as commonly available as normal saline in the delivery room, and its use may transiently increase lactate levels (by ~0.27 mM/L following 10 mL/kg of LR). Ringer’s lactate has small quantities of calcium that can precipitate with simultaneous use of blood products. While saline is the preferred solution during delivery room resuscitation, we prefer to use LR to sodium chloride during the post-resuscitation phase, especially in the presence of hyperchloremic metabolic acidosis [28,29,30].

Bicarbonate is currently not recommended for neonatal resuscitation

Currently, the use of NaHCO3 is not routinely recommended as part of neonatal resuscitation. This is in part due to limited evidence of benefits in improving short-term outcomes following neonatal resucitation [4, 31, 32]. The only placebo-controlled RCT that evaluated use of NaHCO3 during neonatal resuscitation found no difference in survival to discharge or neurological outcome at discharge [32]. Moreover, infants who received NaHCO3 had a trend towards worse outcomes of encephalopathy, cerebral edema and increased need for inotropes. In particular, there is no benefit in correcting metabolic acidosis secondary to HIE with NaHCO3, nor is there evidence in treating respiratory acidosis with NaHCO3 when CO2 cannot be effectively eliminated from the lungs due to poor ventilation [15, 33, 34]. Given this background, current American Academy of Pediatrics- Neonatal Resuscitation Program (NRP) guidelines do not include routine NaHCO3 administration during the acute resuscitation phase, focusing instead on optimal ventilation techniques, chest compressions, and the use of intravenous epinephrine/ volume expanders. In 2000, the guidelines further stated that “the hyperosmolarity and CO2 generating properties of NaHCO3 may be detrimental to myocardial or cerebral function (Fig. 2)” [35, 36]. However, according to the European resuscitation guidelines, the use of sodium bicarbonate can be considered in prolonged resuscitation with adequate ventilation to reverse intracardiac acidosis, with a dosing recommendation of 1–2 mmol/kg administered by slow infusion [37].

Bicarbonate use to treat metabolic acidosis in the NICU

Metabolic acidosis can increase mortality in critically ill newborns by causing hemodynamic instability, arterial vasodilation, lowering myocardial contractility, reducing cellular oxygen supply and mitochondrial oxygen consumption as well as impairing catecholamine responsiveness [17, 38, 39]. A golden rule in the NICU for treating metabolic acidosis is to address the underlying cause (e.g. sepsis, hypovolemia, hypoxia, renal and/or gastrointestinal losses) [17, 18, 33, 35]. Reducing acid production by optimizing the delivery of oxygen to tissues is the ideal treatment for acidosis secondary to tissue-level hypoxia. However, in critically ill neonates such as those with critical congenital heart disease, severe PPHN or septic shock, severe metabolic acidosis may be ominous and may warrant correction while the primary etiology is being addressed. Acidosis may contribute to hemodynamic instability by reducing myocardial contractility, predisposing to arrhythmia, inducing vasodilation, and impairing the response to catecholamines by reducing cellular calcium influx, increasing extracellular potassium, and depleting pituitary vasopressin reserves [40].

The presence of severe metabolic acidosis with compensated respiratory alkalosis and hypocapnia in patients with birth asphyxia and HIE deserves further discussion. Infants with perinatal metabolic acidosis have tachypnea due to HIE and Kussmaul respirations with low PaCO2 levels [41]. Low PaCO2 levels in patients with HIE are associated with increased risk of death or disability at follow-up [42, 43]. Some of these infants are on minimal or no respiratory support. Infants with HIE with absent gag reflex might need an endotracheal tube to protect their airway. In such patients, ventilator adjustments such as low rate, increasing trigger sensitivity (making it difficult to trigger ventilator inflation), decreasing pressure support, increasing termination cut-off for flow cycling to shorten pressure supported breaths and use of sedation (and in extreme situations, use of muscle relaxants) may assist in increasing PaCO2. If these measures fail, a slow infusion of a small dose of NaHCO3 (1 mEq/kg) may correct metabolic acidosis and increase PaCO2. The effect of such small doses of NaHCO3 in HIE and hypocapnia on cerebral blood flow, short- and long-term outcomes needs further investigation.

Currently, in the United Kingdom, a study is underway to examine the short-term and long-term health and developmental effects of NaHCO3 use in preterm neonates with metabolic acidosis [44]. This study aims to recruit 3764 pre-term infants born between 23+0−30+6 weeks gestational age in about 45 NICUs in the U.K. over a 3-year period who have metabolic acidosis [44]. The infants will be randomized into two groups: one group will receive NaHCO3 for episodes of metabolic acidosis, while the other group will not receive NaHCO3 for metabolic acidosis [44]. The primary outcome is survival to discharge without major morbidities including bronchopulmonary dysplasia, retinopathy of prematurity (ROP) requiring treatment, severe IVH or periventricular leukomalacia, late onset sepsis, severe NEC and need for major surgery. These infants will be followed to assess survival without moderate-to-severe neurodevelopmental impairment at 2-years postnatal age corrected for prematurity [44]. Although the study does not set specific dosages of NaHCO3 and allows clinicians to use dosing per their unit protocol, it does recommend using the formula: Mmol of NaHCO3 = (0.3–0.6) × weight (kg) × base deficit (mmol/L) over 30 min to 4 h for NICUs without set dosing guidlines [44]. This study will be valuable in understanding the effects of NaHCO3 in the treatment of metabolic acidosis.

Current use of sodium bicarbonate in the NICUs

A international survey completed in 2023 showed that, out of a pool of 134 physicians, around 91.2% used NaHCO3 to treat metabolic acidosis in their neonatal practice [17]. Although there are known adverse effects associated with NaHCO3 administration, such as fluctuations in cerebral blood flow, cardiovascular hemodynamics and increased rates of severe IVH in preterm babies, its use remains a common treatment for metabolic acidosis in NICUs worldwide [17,18,19]. Another survey completed in Italy showed similar results and also pointed out that dosage and time of infusion varied widely across the centers [19]. With no standard guidelines on indications, dosing or speed of infusion, there is a risk of unsafe treatments, highlighting the need to clarify these dosages [19, 35]. Commonly the following equation has been used to determine the base deficit correction: Full Correction Dose (mmol) = 0.3 X base deficit (mmol/L) X the weight (kg) [17]. and a prevailing practice is to only administer half the dose and reassess the need for the remainder [17]. However, this dose may be excessive due to variations in base deficit as explained previously. We prefer the alternate approach to use 1-2 mEq/kg by slow infusions and reassess acid-base status and hemodynamics then repeat the dose as needed [37, 45, 46]. The use of a lower dose could potentially help reduce the risk of hypokalemia, hypernatremia, hypocalcemia, rebound alkalemia and water-sodium overload.

Effects of bicarbonate administration on the brain

As mentioned previously, a concerning adverse effect of NaHCO3 is fluctuations in cerebral blood flow. Rapid infusion of NaHCO3 over less than 5 min has been associated with both decreased and increased cerebral blood flow [47,48,49]. An RCT was performed in Netherlands to study the effects of rapid vs. slow infusion of NaHCO3 on cerebral hemodynamics and oxygenation in 29 preterm infants [48]. In this study, it was shown that a slow infusion of NaHCO3 is preferred to minimize fluctuations in cerebral hemodynamics as a rapid infusion of NaHCO3 increased cerebral blood volume [48]. A retrospective study by Katheria et al showed that infusion of NaHCO3 over 30-minutes administered in 36 preterm infants (23–31 + 6 weeks gestational age) in the first 24 h after birth was associated with increased pH, decreased base deficit and PaCO2 and increased cerebral oxygenation without an increase in oxygen extraction or cardiac output [49]. However, these patients also had a decrease in systemic blood pressure, so it is unclear if the increase in the cerebral oxygenation, despite no change in cardiac output, was a compensatory reaction to the decrease blood pressure or a direct effect of NaHCO3 [49].

Effect of bicarbonate administration on the heart

In the past, NaHCO3 was used during cardiopulmonary resuscitation in cardiac arrest based on the idea that acidemia impairs myocardial performance and attenuates catecholamine-induced increase in blood pressure, heart rate, and cardiac contractility. Thus, correcting the acidosis with NaHCO3 might improve outcomes [35]. However, studies do not support the use of buffers during cardiac arrest and have shown deleterious effects on myocardial performance [35]. An abrupt rise in serum HCO3 levels following the administration of NaHCO3 may cause paradoxical intracellular acidosis [50]. This can be a particular