Table 2.

Age at treatment initiation and long-term GH treatment outcomes in children born SGA

Study	N	Design/Duration	Agea/Model	Outcome (∆HSDS or statistical model)	PValue
Dahlgren 2005 [36]	77	Ob/prepubertal to FH	Prepubertal during >2 years GH therapy vs prepubertal during <2 years GH therapy	Mean gain FH SDS prepubertal for >2 years GH therapy = 1.7; prepubertal for <2 years = 0.9	<0.001
de Ridder 2008 [47]	150	Data from 2 previous GH trials; R, DB/5 years[38]; R, C/3 years[48]	Median age 7.5/Statistical model predicting HSDS at puberty	Age at start (−0.27 estimated coefficient)Other significant predictors: HSDS at start (0.71 estimated coefficient), target height SDS (0.13 estimated coefficient), GH dose X IGF-I SDS at start (−0.29 estimated coefficient), gender (−0.34 estimated coefficient)	<0.0001 = 0.009 to <0.0001
Ranke 2010 [45]	161	KIGS, clinical trials/7.7 years	Median age 7.8 years/Statistical models predicting AH SDS and ∆HSDS	70% of variability in adult height SDS: HSDS at GH start, ∆HSDS 1st year on GH, years on treatment [younger start, longer phase], maternal HSDS, length SDS at birth, SRS diagnosis	NR
				60% of variability in ∆HSDS: ∆HSDS 1st year GH, H-MPH SDS at GH start, years of GH treatment [younger start, longer phase]
van Pareren 2003 [37]	54	MC, R, DB/mean 7.8 years	Mean age 8.1/∆HSDS correlations	∆HSDS from start of GH treatment to AH negatively correlated with age at treatment start: r = −0.36	<0.01

^aAge at treatment initiation. AH, adult height; C, controlled; DB, double-blind; FH, final height; GH, growth hormone; HSDS, height SDS; ∆HSDS, change in height SDS; IGF-I, insulin-like growth factor; KIGS, Pharmacia International Growth Database; MC, multicenter; MPH, mid-parental height; NR, not reported; Ob, observational; R, randomized; SDS, standard deviation score; SRS, Silver-Russell syndrome.