Vitamin D deficiency is increasingly recognised in neonates, especially those born to high-risk mothers.1 Term infants at high risk of vitamin D deficiency have higher mean alkaline phosphatase (ALP) activity than controls, indicating increased bone turnover.1 Despite this, ALP activities remain within the normal range in these high-risk infants. There is little research on vitamin D deficiency in very-low-birthweight (VLBW) infants, who are at high risk of bone disease due to osteopenia of prematurity (OP) and potentially maternal or nutritional vitamin D deficiency. The increase in breast feeding with inadequate supplementation in mother and infant with vitamin D may have led to a coincident increase in infant vitamin D deficiency (28% in 1990; 50% in 2000), as concentrations of vitamin D in breast milk are low (12–50 U/l).2 Serum ALP activity increases during the first postnatal weeks and correlates negatively with quantitative tibial ultrasound evaluation.3 We hypothesised that raised ALP activity may be predictive of abnormal vitamin D concentrations in VLBW infants.

Convalescent, stable VLBW infants (birth weight <1500 g) admitted to a tertiary referral neonatal intensive care unit were eligible for inclusion if their ALP activity was >1000 U/l in routine, convalescent “growing” bloods. 25-Hydroxycholecalciferol (25-hydroxyvitamin D₃) concentrations were determined in these infants using radioimmunoassay. Forty-two VLBW infants (median (range) gestation 27.6 (24–34) weeks, birth weight 970 (840–1480) g and 18 (42.8%) male) had ALP determinations in the first month of life. Twenty-three infants with ALP >1000 U/l had their vitamin D concentrations measured. Only three had adequate vitamin D concentrations (⩾70 nmol/l) (table 1). In 16, concentrations were <50 nmol/l, and in four they were 50–69 nmol/l. ALP activity did not correlate with vitamin D, calcium or phosphate concentrations. There was no correlation between vitamin D concentration and ALP activity (p = 0.28) or vitamin D and phosphate concentrations (p = 0.08) (Spearman’s correlation). Although there was a correlation between vitamin D and calcium concentration (p = 0.002), the range of calcium concentrations was essentially normal and therefore this finding may not be a clinically useful indication of vitamin D deficiency (table 1).

Because of the increased risk of OP in VLBW infants, early recognition and treatment of vitamin D deficiency is vital. ALP, calcium and phosphate are not adequate surrogates for OP and/or vitamin D deficiency. In addition, vitamin D supplementation in infancy may protect against type I diabetes mellitus and is involved in immunomodulation.4 Recently, the association between vitamin D deficiency and infantile heart failure has also been reported.5 Sixty-four percent of infants in whom vitamin D was determined were markedly deficient (<50 nmol/l), and 87% had some degree of deficiency (<70 nmol/l). We suggest that vitamin D concentrations should be assessed in all VLBW infants, as vitamin D deficiency is a readily treatable cause of bone disease.