To test the hypothesis that GH deficiency might explain the low growth velocity of some LHRH agonist (LHRHa)-treated children with central precocious puberty, we measured stimulated (n = 81) and spontaneous (n = 32) GH levels during or after LHRHa treatment. GH stimulation tests in the children who were receiving LHRHa treatment were performed after 2 days of ethinyl estradiol administration. Thirty-one of 81 children (38%) who underwent GH stimulation tests had subnormal responses (less than or equal to 7 micrograms/L) to all tests administered (at least 2 stimuli), including 22 of 67 (33%) who had precocious puberty that was idiopathic or associated with hypothalamic hamartoma. Eleven of 32 children (34%) who underwent measurement of the mean nighttime spontaneous GH level had levels below the normal range for prepubertal children (less than 1.2 microgram/L). Despite the high incidence of subnormal GH levels, there appeared to be no relationship between the GH levels of these children and their growth characteristics. The height, growth velocity, bone maturation rate, predicted height, and insulin-like growth factor-I levels were not different between the children with low GH levels and the children with normal GH levels. Conversely, the GH levels were not different between the children with subnormal growth rates and the children with normal growth rates. Thus, variation in the growth rates of these LHRHa-treated children with central precocious puberty could not be explained by variation in the stimulated or spontaneous secretion of GH. In attempting to understand the high incidence of low GH levels in children with precocious puberty, we examined the relationship between GH level and body mass index (BMI). Both the stimulated (r = -0.33; P less than 0.002) and the spontaneous (r = -0.61; P less than 0.0002) GH levels were inversely related to BMI. Moreover, the children with precocious puberty as a group had significantly elevated BMI [1.2 +/- 0.1 (+/- SE) SD units] compared to normal children of the same age (P less than 0.0001). Thus, increased body mass may explain the high incidence of subnormal GH levels in these patients, and normative GH levels adjusted for body mass are needed before it can be concluded that the apparently subnormal GH levels in LHRHa-treated children with precocious puberty are in fact low.