Objective To compare in two epochs of differing phosphate provision serum calcium, phosphate, potassium, and sodium concentrations and the frequency of abnormality of these electrolytes and of sepsis in preterm infants who received an optimised higher amino acid-content formulation.

Design and setting Retrospective cohort study at a single tertiary-level neonatal unit.

Patients Preterm infants given parenteral nutrition (PN) in the first postnatal week during two discrete 6-month epochs in 2013–2014.

Interventions In epoch 1 the Ca²⁺:PO₄ molar ratio of the PN formulation was ~1.3–1.5:1 (1.7 mmol Ca²⁺ and 1.1 mmol PO₄ per 100 mL aqueous phase) and in epoch 2 was 1.0:1 via extra phosphate supplementation (1.7 mmol Ca²⁺ and 1.7 mmol PO₄ per 100 mL).

Main outcome measures Peak calcium and nadir phosphate and potassium concentrations, and proportions with severe hypercalcaemia (Ca²⁺ >3.0 mmol/L), hypophosphataemia (PO₄<1.5 mmol/L), and hypokalaemia (K⁺ <3.5 mmol/L) within the first postnatal week.

Results In epoch 2, peak calcium concentrations were lower than in epoch 1 (geometric means: 2.83 mmol/L vs 3.09 mmol/L, p value<0.0001), fewer babies were severely hypercalcaemic (10/49, 20%, vs 31/51, 61%, p value<0.0001); nadir plasma phosphate concentrations were higher (means: 1.54 mmol/L vs 1.32 mmol/L, p value=0.006), and there were fewer cases of hypophosphataemia (17/49, 35% vs 31/51, 61%, p value=0.009) and hypokalaemia (12/49, 25% vs 23/51, 45%, p value=0.03).

Conclusions Reverting from a PN Ca²⁺:PO₄ molar ratio of 1.3–1.5:1 to a ratio of 1.0:1 was associated with a lower incidence and severity of hypophosphataemia and hypercalcaemia. For preterm infants given higher concentrations of amino acids (≥2.5 g/kg/day) from postnatal day 1, an equimolar Ca²⁺:PO₄ ratio may be preferable during the first postnatal week.