The aim of this study was to identify factors predictive of individual final height (FH) in subjects born small for gestational age (SGA). All full-term singleton subjects born SGA (birth weight and/or length <3rd percentile) during the period 1971-1978, matched with appropriate birth weight for gestational age (AGA) subjects (birth weight between 25th and 75th percentile) were followed from birth to FH and evaluated before puberty at a mean age +/- SD of 6.1 +/- 0.7 y and after puberty at a mean age of 20.8 +/- 2.0 y (subjects born SGA, n = 213; born AGA, n = 272). When adjusted for target height, a significant deficit in final height (p < 0.0001) was found in SGA as compared with AGA subjects for both male subjects (-3.99 cm with 95% confidence interval from -5.6 to -2.4) and female subjects (-3.64 cm with 95% confidence interval from -5.0 to -2.3), with 13.6% of subjects in the SGA population presenting short final stature. In a multiple regression analysis, target height and studied group (SGA or AGA) were found to be the strongest predictors of individual FH (p < 0.0001, r2 = 0.35 for male subjects, p < 0.0001, r = 0.40 for female subjects). For SGA subjects and according to a multiple stepwise linear regression model, 31% of the variability of individual FH [SD score (SDS)] and 58% of the variability of individual height gain SDS could be explained at birth from mother's height, father's height, and birth length SDS. No other variables were found to be predictive such as sex, gestational age (from 37 to 42 wk), birth weight SDS, ponderal index at birth, or risk factors during pregnancy associated with intrauterine growth retardation such as pregnancy-induced hypertension, smoking, or a history of SGA in offspring. Although a significant increase of body mass index SDS was documented before and after puberty in SGA subjects, puberty was not found to have any influence on growth outcome.