BSPED2021 Oral Communications Oral Communications 9 (6 abstracts)
1Medical Sciences Division, University of Oxford, Oxford, United Kingdom; 2Department of Paediatric Endocrinology, University Hospitals Bristol NHS Foundation Trust, Bristol, United Kingdom; 3Department of Paediatric Endocrinology, Oxford University Hospitals NHS Foundation Trust, Oxford, United Kingdom
Background: Insulin tolerance test (ITT) is the gold-standard investigation in children with suspected growth hormone deficiency (GHD). Despite their benefits, ITTs are resource intensive, potentially dangerous and can result in unnecessary treatment due to false positive results in children that are at low risk for GHD.
Methods: Retrospective analysis of all paediatric patients undergoing an ITT in two tertiary hospitals in the UK (centre A between 2017-2021, centre B 2016-2019). We determined baseline patient characteristics and proportion of normal tests by centre and patient group (high risk [HR]: pituitary abnormality including oncology patients; low risk [LR]: isolated short stature).
Results and discussion: Proportions of female patients were 28%(A) and 35%(B) (P = .2). Patients at centre A were 1.4 years older (mean ages 13.4, 12 years; P < .01) and had slightly lower BMI SDS (mean BMI SDS -0.49, -0.09; P = .047). Only 3 patients (2 centre A) had BMI SDS >3. The proportion of HR patients was lower in centre A (16% vs 39%; P < .001). Only ITTs where blood glucose <2.2 mmol/l had been achieved and no growth hormone (GH) values were missing were included (n = 108 centre A, n = 222 centre B). GH cut off for a normal test was ≥6.7µg/l or ≥3µg/l for patients who had completed growth. In centre A, 58% of ITTs were normal, compared to 39% in centre B (P = .002). When only LR patients (n = 227) were considered, the difference between centres remained (63% vs 48% normal ITTs, P = .039). The difference still stood when ITT results only from young (<12 years) LR children (n = 103) were analysed (61% vs 35% normal ITTs, P = .03). Among HR patients (n = 103) 35% and 26% of ITTs were normal, respectively (P = .6).
Conclusions: When ITTs were performed on low risk patients, centre A had a significantly higher proportion of normal results. This was not explained by baseline patient demographics. Although consensus in diagnostic GH cut-off levels has been reached, variability in clinical practice between centres can explain the observed difference and highlight the importance of using robust pretest clinical criteria especially in children at low risk for GHD.