Hypophosphataemia is an electrolyte disturbance that has been overlooked by neonatology reference books and protocols for the management of ill newborns until very recently.1,2 Before year 2000, reports of hypophosphataemia were limited to cases associated with the bioavailability of the inorganic phosphates formerly used in parenteral nutrition,3 and it was not until 2006 that cases of hypophosphataemia in newborns potentially associated with refeeding syndrome were first reported.4

In our unit, we started administering amino acids (Primene 10%®) parenterally to preterm newborns with birth weights of less than 1500g from the first day of life in 2001. Around the middle of that decade, hypophosphataemia started to be detected in newborns as a chance finding in blood tests performed for unrelated reasons. In 2010, due to the frequency and severity of some of the episodes of hypophosphataemia, our hospital developed a protocol for screening at-risk newborns and for the prevention and management of hypophosphataemia. The lack of relevant research in neonatology posed a challenge in the development of this initial protocol, for which we resorted to the extrapolation of data on treatment of adults with refeeding syndrome.5–8 The first reports of hypophosphataemia in very low birth weight preterm newborns were published in 2012, and different authors started to propose refeeding syndrome as a potential aetiology.9–11 Subsequent publications have described its pathophysiology, an association with a high amino acid intake and a potential increase in the risk of sepsis that would compound the direct and possibly fatal effects of severe hypophosphataemia.12–14

The objectives of our study were to define the risk groups for hypophosphataemia in our hospital and to determine the frequency, severity and temporal distribution of early episodes of hypophosphataemia in the first 2 weeks of life.

Phase 1: of a total of 1,394 newborns admitted to the NICU in this period, 34 (2.4%) had received a diagnosis of hypophosphataemia during their stay. Of these, 26 (76%) had been born at 32 weeks’ gestation or earlier, and 26 also weighed less than 1500g at birth. All the rest had birth weights below the 10th percentile and had been born between 33 and 40 weeks’ gestation, and the highest birth weight among the patients with a diagnosis of hypophosphataemia was 2300g.

Phase 2: of a total of 73 preterm newborns with birth weights of less than 1500g or born before 32 weeks’ gestation, 12 (16.4%) developed hypophosphataemia, which was severe in 7 (9.6%) and critical in 5 (6.8%). All cases of critical hypophosphataemia (serum phosphate <2mg/dL) occurred in the first 3 days post birth. Nine cases (75%) occurred in patients with a history of intrauterine growth restriction with abnormalities in blood flow distribution. The other three corresponded to patients born with weights of less than 1000g. Hypophosphataemia was associated with hypokalaemia in 5 cases (42%) and hypercalcaemia in 3 (25%).

Phase 3: a total of 37 patients met the inclusion criteria, and we received parental consent for the participation for 20, of who 16 had birth weights of less than 1500g. Two patients died during their stay due to morbidity associated with prematurity. Table 1 summarises the characteristics of the patients in the cohort.

Of all participants, 45% had at least one episode of hypophosphataemia (9 out of 20). Table 2 presents a description of these 9 patients. Severe or critical hypophosphataemia occurred in 5 of the patients (25%), adding up to 8 episodes. Table 3 shows the temporal distribution of the episodes of hypophosphataemia.

Of all patients with hypophosphataemia, 55% also had mild hypokalaemia, and 44% changes in serum calcium levels (half of them significant hypocalcaemia and the other half mild hypercalcaemia). Table 4 presents the results of the remaining serum measurements. We found statistically significant increasing trends in the levels of calcium, phosphate and alkaline phosphatase and decreasing trends in the levels of sodium, chloride and creatinine with the passage of time. Twenty percent of the patients developed mild hypernatraemia, with a peak sodium level of 155mEq/L.

The birth weight and the z-score for birth weight were significantly lower in patients with hypophosphataemia (Mann–Whitney U test, P=0.001 and 0.014, respectively), but there was no difference in gestational age (P=0.18).

We did not find differences between patients with hypophosphataemia and patients without based on sex, type of life support, history of maternal preeclampsia or prenatal administration of magnesium sulphate.

Patients with hypophosphataemia received smaller enteral nutrition volumes in the first 2 weeks and therefore higher doses of parenteral nutrition (Table 5), with higher intravenous doses of amino acids, glucose and lipids. However, we did not find differences in the intravenous doses of calcium and phosphorus in the first 5 days (Table 6). Starting at 6 days, the size of the differences rendered the analysis inconsequential. None of the patients received insulin.

Recent studies have consistently identified foetal malnutrition, birth weight of less than 1000g and high intake of amino acids early after birth as risk factors for the development of hypophosphataemia in the first week of life.10,12,14 Our study corroborated these associations and found a high incidence of hypophosphataemia in the prospective phase, with low phosphate levels detected in approximately one third of the blood tests performed at days 1, 3 and 7 post birth, of which 22% were in the range of critical hypophosphataemia. The incidence was highest in newborns with birth weights of less than 1000g or with birth weights of less than 1200 and a history of foetal malnutrition. Data from adult studies describe the development of symptoms when levels drop below 1.5mg/dL, or before if the decline is rapid.8 In newborns, serum phosphate levels are higher compared to adults, and they can drop very quickly after cord clamping unless phosphorus is administered with enteral or parenteral nutrition. We observed some cases in extremely small newborns with serum phosphate levels of 0.5mg/dL that required doses of up to 3.5mmol/kg/day to correct the hypophosphataemia, as reported by other authors.9 These patients are at immediate risk of losing their lives, but other patients with hypophosphataemia are also at risk of suffering the clinical repercussions described in adults and critically ill children15 as well as newborns.12,16

This surge in hypophosphataemia seems to have 2 well-described causes. On the one hand, there is refeeding syndrome, as observed in adult critical patients in the first 3 days of refeeding, which in newborns would result from intrauterine malnutrition14 in the context of rapid foetal growth combined with aggressive parenteral nutrition after birth, and especially with the administration of high doses of amino acids.17,18 On the other hand, there is the practice of withholding administration of intravenous sodium or potassium in the first 48h of life in very low birth weight preterm newborns, which limits the intake of phosphate salts in the first 2 days and exacerbates the hypophosphataemia. There is also the calcium-to-phosphorus molar ratio of 1:1.6 that has been used this far, which is probably inadequate. Retaining 1g of amino acids requires 10mg (0.33mmol) of phosphorus, so in the situation that we have described, a molar ratio of 1.2:1 to 1:1 seems more appropriate from a physiological perspective. Needless to say, once hypophosphataemia develops, the necessary replacement of phosphorus will invert this ratio.

There were several limitations to our study, including it having been conducted in a single centre, the small size of the prospective sample, the high proportion of parents that refused to participate (probably due to the greater number of blood samples drawn in study participants compared to standard care) and the retrospective design of phases 1 and 2. It is very likely that hypophosphataemia was underdiagnosed in these two retrospective phases, and it was not possible to establish when hypophosphataemia had started in the patients included in phase 1. Although it is likely that the pattern was similar to the one observed in phase 3, we found cases of late-onset hypophosphataemia associated with sepsis or refeeding after enterocolitis. Furthermore, there were conspicuously high proportions of maternal disease during gestation and of twin pregnancies in the phase 3 cohort. Lastly, the morbidity in this small group cannot be deemed representative of overall morbidity in the population of infants born with weights of less than 1500g or before 32 weeks’ gestation in our unit. Due to all of the above, it is difficult to infer the clinical repercussions of hypophosphataemia in this small sample of newborns.

Despite these limitations, our results warranted changes to our practices regarding parenteral nutrition, electrolyte supplementation and performance of blood chemistry tests in our patients. First of all, we established routine monitoring of serum phosphate levels in the first week of life in this at-risk group (birth weight <1000g and newborns with a history of IUGR with a birth weight z-score <−2)19 and in critical patients undergoing refeeding after the first week of life. It was also decided that initiation of sodium and phosphorus supplementation would not be delayed past 24h post birth. In cases where an intake of 2–3g of amino acids is desirable since birth, supplementation with phosphorus should probably also start in the first day of life with doses of 0.5–1mmol/kg/day. Although further research is required to evaluate the safety and efficacy of these changes, there is also evidence that a moderate intake of sodium is well tolerated since birth18 and is not associated with an increase in the incidence of hypernatraemia as long as electrolyte and fluid balances are properly controlled.

To conclude, this small study demonstrates that hypophosphataemia frequently develops in newborns weighing less than 1000 or 1200g and with a history of foetal malnutrition receiving amino acid supplementation in parenteral nutrition from the first day of life. Hypophosphataemia may reach critical levels, so monitoring of phosphate levels should probably become routine practice in this group of patients, with consideration of electrolyte supplementation in the first 2 days of life.

The authors have no conflicts of interest to declare.