To the Editor: We read the article by Floyd R. Sallee, M.D., Ph.D., et al. (1) with great interest. The authors used the growth hormone response to the α2-adrenergic antagonist yohimbine as a measure of presynaptic norepinephrine activity in a study comparing anxious children with comparison subjects. There are, however, methodological problems inherent in the use of growth hormone as a measure of norepinephrine activity, largely because the release of growth hormone from the pituitary is inhibited by cortisol (2). The authors appeared to measure a single baseline blood sample for each subject for the estimation of cortisol content and reported that there was no significant difference in baseline cortisol levels between comparison children (mean=6.5 g/ml, SD=2.8) and children with anxiety disorders (mean=9.1 g/ml, SD=8.4). Measuring single blood samples of cortisol is often unreliable, and integrated measures of cortisol output are preferred (3). In view of the effect of cortisol on growth hormone release, we believe it would be prudent for these authors to examine a robust measure of cortisol as a covariate of growth hormone response. We further suggest that all studies in which neurotransmitter function is inferred from pituitary hormone measures must measure corticosteroids as a potentially confounding covariate.