Brain MRIs may be of low value in most children diagnosed with isolated growth hormone deficiency

Objectives Brain MRIs are considered essential in the evaluation of children diagnosed with growth hormone deficiency (GHD), but there is uncertainty about the appropriate cut-off for diagnosis of GHD and little data about the yield of significant abnormal findings in patients with peak growth hormone (GH) of 7–10 ng/mL. We aimed to assess the frequency of pathogenic MRIs and associated risk factors in relation to peak GH concentrations. Methods In this retrospective multicenter study, charts of patients diagnosed with GHD who subsequently had a brain MRI were reviewed. MRIs findings were categorized as normal, incidental, of uncertain significance, or pathogenic (pituitary hypoplasia, small stalk and/or ectopic posterior pituitary and tumors). Charges for brain MRIs and sedation were collected. Results In 499 patients, 68.1% had normal MRIs, 18.2% had incidental findings, 6.6% had uncertain findings, and 7.0% had pathogenic MRIs. Those with peak GH<3 ng/mL had the highest frequency of pathogenic MRIs (23%). Only three of 194 patients (1.5%) with peak GH 7–10 ng/mL had pathogenic MRIs, none of which altered management. Two patients (0.4%) with central hypothyroidism and peak GH<4 ng/mL had craniopharyngioma. Conclusions Pathogenic MRIs were uncommon in patients diagnosed with GHD except in the group with peak GH<3 ng/mL. There was a high frequency of incidental findings which often resulted in referrals to neurosurgery and repeat MRIs. Given the high cost of brain MRIs, their routine use in patients diagnosed with isolated GHD, especially patients with peak GH of 7–10 ng/mL, should be reconsidered.