Pituitary Volume Is a Better Predictor of Growth During Growth Hormone Therapy Than the Growth Hormone Stimulation Test

Abstract Background: Patients with diminished GH secretion are candidates for GH therapy (GHT). The GH stimulation test (GHST) is considered the gold standard for the diagnosis of GH deficiency (GHD), yet the cutoff of 10 ng/mL has not been well validated statistically. Another proposed method to define GHD has been to measure patients’ pituitary volumes (PV), as the size of the gland may correlate with the amount of GH produced. Objective: This study seeks to ascertain whether the GHST or PV is a better predictor of response to GHT, and determine which method can better define true GHD. Patients and Methods: A database at a Pediatric Endocrinology center was queried for patients aged 6-18 yrs who underwent a GHST, MRI, and GHT between 1/2018 - 6/2019. Patients with relevant comorbidities, those with GHST peak ≥ 10.0 ng/mL, and patients that were non-adherent to their GHT were excluded. Clonidine and L-dopa were stimulants for the GHST. MRIs were acquired on a Philips 1.5 or 3.0 T scanner (1mm slices) and PV was calculated using the ellipsoid formula (LxWxH/2). 87 patients met these criteria for analysis. PV was converted to standard deviation scores (SDS) based on age and sex using the largest data set of pituitary volumes available in the literature. To account for sex-related growth rate differences by age, heights at the initial and subsequent time points were also converted to SDS based on age and sex using parameters provided by the National Center of Health Statistics. Response to treatment was defined as change in height SDS over the assessed time interval. The initial height was included as a covariate. R statistical software was utilized to analyze the correlation between response to GHT and GHST peak value, as well as response to GHT and PV. The relationship between GHST peak value and PV SDS was analyzed with a Spearman correlation. Results: The GHST peak was not a significant predictor of growth response to treatment in both the first or second intervals (r= -0.01, p= 0.207 and r= 0.00, p= 0.815 respectively). GHST peak and PV SDS were not correlated (r=0.08, 95% CI: -0.14, 0.28). Lower SDS of PV significantly predicted growth response to therapy in the first 1 to 8.7 months of treatment (n= 87, model r2=0.231, b=-0.05, SE=0.02, P=0.012). This association in the second interval between 7.8 and 17.4 months of treatment was neither as strong as the first interval nor was it statistically significant (n=62, model r2=0.145, b=-0.05, SE=0.03, P=0.127). Within-person growth velocity was greater in the first interval (mean = 0.37, SD = 0.17) than in the second interval (mean = 0.20, SD = 0.16). Conclusion: Our data indicates that PV can be a valuable tool in defining GHD and should be considered a criterion for determining eligibility for GHT. To our knowledge this is the first study to determine that PV is a better predictor of growth response to GHT than the GHST.