ESPE Abstracts

Background: The predictive value of a random cortisol measurement in the neonatal population with suspected adrenal insufficiency is unknown.

Objective and hypotheses: i) determine the indications for measuring cortisol; ii) review subsequent management; and iii) establish predictors for adrenal insufficiency.

Method: A laboratory database search identified cortisol results in babies <1 year from three neonatal units between September 2010 and May 2013. Gender, gestational age, birthweight, time of test, CRIB II score, and sodium level were investigated as potential predictive factors. A random cortisol >100 nmol/l was accepted as normal. A short Synacthen test (SST) peak cortisol >450 nmol/l indicated an adequate response.

Results: 60 infants (M40:F20) had cortisols analysed, 58 had random cortisol and two a SST directly. Indications included: prolonged postnatal steroid use (n=4), conjugated hyperbilirubinaemia (n=22), hypoglycaemia (n=24), midline defect (n=6), hyponatraemia (n=4), and ‘other’ (n=2); two patients had two risk factors. In total 86 random cortisol levels were analysed; 46 were normal (209.0 nmol/l (138.3-340.8)) and 40 were low (50.5 nmol/l (30.0–71.5)). Twenty-four SST were carried out in 20 infants; ten infants (41.7%) had suboptimal cortisol peaks. In 8/10 cases, hydrocortisone was commenced, 1/10 hydrocortisone when unwell and 1/10 no plan for unclear reasons (Table 1).

Conclusion: No statistically significant predictive factors for adrenal insufficiency were found from this study. There was a trend for lower gestational age, weight at birth and at time of test, and male sex (9/10 infants), suggesting potential predictive factors. There is a need for clear guidelines for the management of sub-optimal cortisol levels in neonates.