We included all the adolescent boys who were seeking treatment
at our growth center between January 2004 and September 2009 with
the following criteria: i) testis volume N 12 ml and testosterone
levels N 4 ng/ml (young adult levels) [19]; ii) an informative track of
height measurements during puberty allowing calculation of growth
velocity precisely; iii) no prior rhGH treatment; iv) a slow growth velocity
(b2 cm/6 months); and v) a PAH at or under −2.5 SDS. We chose
this height because a quality-of-life (QOL) study identified −2.5 SDS
as the threshold for observing a negative impact of short stature on
male adults [7]. Mid-parental height was not used to predict adult
height because it does not take into account the individual tempo of puberty.
Insteadwe developed a personalized algorithmallowing an accurate
prediction of height (see “Modeling of growth”) [20]. GHdeficiency
(GHD) was excluded by a stimulation test with a GH peak N15 ng/ml. A
peak GH concentration less than 10 ng/ml after any one or more of
series of secretagogues and assayed by RIA (radioimmunoassay) has
traditionally been used to support the diagnosis of GHD [38]; however,
the equivalent level in monoclonal-based assays is considered to be
7 ng/ml [4]. We chose 15 ng/ml as an arbitrary cut-off to safeguard
the study from including adolescents with GHD. In addition, none of
the participants had a basal serum IGF-1 concentration b−1.2 SDS.
Subtle forms of dyschondrosteosis or other chondrodysplasia were excluded
by examination of radiographs of the forearm, pelvis and spine.
Thyroid stimulating hormone (TSH) levelswere normal. All adolescents
were healthy and clinical examination was normal.