Impact of IGF-1 Normative Datasets on Indication and Outcome of Growth Hormone Stimulation Testing

Abstract

In the Netherlands, the diagnosis of growth hormone deficiency in children follows the Dutch national guidelines for Triage and Diagnosis of Growth Disorders in Children. Initial biochemical evaluation includes an IGF-1 measurement as screening parameter for growth hormone deficiency. Based on the clinical probability of growth hormone deficiency and the IGF-1 Z-score, a growth hormone stimulation test is performed if serum IGF-1 Z-score is < 0 SD in case of a high probability and if serum IGF-1 Z-score is < -1 SD in case of low probability. An IGF-1 Z-score > 0 SD virtually excludes a growth hormone deficiency disorder. The interpretation of growth hormone stimulation testing is dependent on both the peak growth hormone concentration, but also on the baseline IGF-1 Z-score, particularly in cases of partial deficiency. Although, nation wide, Dutch laboratories have harmonized their measurement for IGF-1 (as was previously done for growth hormone), a Dutch harmonized normative data set has not been widely adopted. Moreover a clinical evaluation of the implementation of this dataset based on dynamic testing has not been published. To assess the impact of choice of a particular normative dataset on the diagnosis of growth hormone deficiency we recalculated Z-scores of IGF-1 measurements between 2016 and 2019, using our home reference values based on de normative dataset by Elmlinger (E)1, and using the normative datasets defined by Bidlingmaier (B)2 and by the Dutch IGF-1 harmonization program (NL). Based on these three Z-scores, the outcomes of growth hormone stimulation tests performed in this period (n=86) were reassessed according to the interpretation described in the Dutch guideline. Using all three normative datasets the same 4 patients were identified as likely to have a growth hormone deficiency, whereas 10(E), 10(B), or 8(NL) patients were identified as possible partial growth hormone deficiency. In 70(E), 66(B) or 72(NL) patients the growth hormone stimulation test was unaffected. Using normative dataset B, 6 patients displayed a pattern associated with a possible growth hormone resistance, or of bio-inactive growth hormone syndromes, which based on its incidence would be unlikely for a secondary care setting. A striking observation was however, that of all patients with a normal stimulation test 9 (E)/16 (B) or 30 (NL) had a IGF-1 Z-score of > 0 SD. This implies that, for the diagnosis of growth hormone deficiency, it is safe to implement the Dutch harmonized dataset, which in addition could result in a reduction in the number of growth hormone stimulation tests that have to be performed.