MA is an acid-base imbalance that can occur acutely or chronically, characterized by a primary reduction in the serum concentration of bicarbonate (HCO3−), a secondary decrease in the arterial partial pressure of carbon dioxide (PaCO2), and a reduction in blood pH . MA is a common finding in the critically ill infant and can be determined by a number of different etiologies, other than renal and gastrointestinal bicarbonate losses (in which the deficit correction with SB is indicated). The results of our survey show that more than half of Italian NICUs (54.7%) routinely use SB to treat MA. In the above-mentioned European survey, 42.2% of interviewed neonatologists would have administered SB in the case of an asphyxiated term neonate with severe combined metabolic and respiratory acidosis, with remarkable differences in practices among European countries (the rate of SB use in Italy corresponded to about 40%) .
The practice of treating neonatal MA with intravenous SB, was first reported in 1963 by Usher , who showed that an early infusion of a 10% glucose solution plus 5–15 mEq/dl of SB resulted in a considerable reduction in mortality among treated neonates. From that moment, this type of infusion, defined as the “Usher regimen”, started to be widely adopted in neonatal care. Nonetheless, in the following years, further researches demonstrated that the administration of SB in asphyxiated neonates was not affecting the acid–base balance in the first 24 hours , and that there was insufficient evidence from randomized controlled trials to support or refute the use of SB during resuscitation of infants at birth . Furthermore, studies conducted on adult patients evidenced that administration of SB during cardio-pulmonary resuscitation could be detrimental to the myocardial function, due to the worsening of intracellular acidosis related to carbon dioxide accumulation, increased hyperosmolality, extracellular alkalosis and reduced coronary perfusion pressure [3, 11]. Thus, since the 2000 update, the International guidelines for neonatal resuscitation did not suggest to use SB during neonatal resuscitation any longer . However, one shall also mention that other authors are still permissive towards SB therapy .
Further researches also showed how the administration of SB could increase the risk of death and intraventricular hemorrhage in preterm infants [4, 5, 14]. In a recent retrospective study, Katheria et al. showed how SB administration in extremely preterm infants does not act on the cardiac output in the short term, but leads to transient fluctuations in cerebral and cardiovascular hemodynamics that could cause dangerous effects on the weak brain vessels of such patients . An increased concentration of CO2 (a potent cerebral vasodilator) and greater blood osmolality (resulting in a flow of intracellular water into the extracellular space) have been suggested as possible co-factors for the cerebral blood flow modifications following SB infusion .
Our survey evidences the lack of well-defined pH, BE and lactate thresholds adopted by Italian NICUs for the correction of MA with SB.
Such variability could also be explained by means of different criteria used to define neonatal MA in studies thereof [17–19]. Moreover, normal arterial lactate values are influenced by the hours of life of the infant, with described thresholds ranging from more than 3.8 mmol/l (at 48 hrs.) to over 1.5 mmol/l (after 96 hrs.) .
Around 70% of units that commonly treat MA, when needed, administer “half dose” of the SB amount determined by the classic formula proposed in 1960 by Astrup to correct the BE deficit . In addition, the heterogeneity of SB dilution and administration practices observed reflects the largely arbitrary recommendations that can be found in literature [3, 4, 17, 22].
Intravenous SB solutions are highly hypertonic. The 1:1 dilution is the most used by the surveyed centers and frequently recommended [17–22], yet still being hypertonic.
Evidence is conflicting even with reference to the suggested duration of SB infusions. A systematic review in 2002 reported no differences when comparing rapid vs. slow or no correction , whereas, a few years later, other researchers suggested a slow infusion of SB, over a 30-min period, to minimize fluctuations in cerebral blood flow of preterm infants .
Our survey showed that there is a relevant number of centers (45%) opting for other strategies for the management of MA, rather than the routine administration of SB. Such approach may be considered as being appropriate, also taking into account that the treatment of neonatal MA should preferably rely on the correction of its primary cause, as also suggested by many experts [1, 3, 6, 11]. Another strategy to deal with chronic MA of preterm infants could be to add buffer salt, such as potassium lactate or sodium acetate, to the PN  - a practice reported as being used by 37.6% of the surveyed NICUs.
Preterm infants can be frequently exposed to an excessive chloride intake  with subsequent hyperchloremia that can constitute a cause of MA, particularly in premature infants, because of the negative effect on the neonatal kidney capability of eliminating acid load . Thus, limiting chloride infusion and providing sodium and potassium as organic phosphate, sodium acetate-citrate or potassium acetate-citrate within the PN preparation might prevent hyperchloremic metabolic acidosis [28–30]. Moreover, some authors reported as earlier/higher parenteral aminoacid and lipid intakes raised the risk of metabolic acidosis, particularly in babies born less than 24–26 weeks of gestation. Nevertheless, stronger evidence is still needed to further support the mentioned practices.
However, our study includes some limitations. Questions did not distinguish between acute or chronic MA nor between occurrence of such condition in term or preterm infants. Thus, some answers may have been arbitrary and may not have adequately described different strategies adopted. The question regarding the "use of SB only for some pathologies" was not fully clear, and we have received answers both from the units that were using SB as well as those not routinely using it. Finally, answers given by the directors of the NICUs did not always reflect attitudes of every clinician within the unit.