Short-term growth response to GH treatment and considerations upon the limits of short-term growth predictions

OBJECTIVES: To investigate the impact of short-term growth measurements on predicting the individual growth response to GH treatment, and to elucidate the possible reasons for the limited accuracy of current growth prediction models for GH-treated children. METHODS: Short-term growth measurements by knemometry and stadiometer in 99 short, GH-treated children (27 girls, 72 boys), aged 10.3 +/- 2.3 years, from the Children's University Hospital, Leipzig, Germany. RESULTS: GH treatment significantly accelerated the mean height velocity (HV) from 4.3 +/- 1.0 to 8.1 +/- 1.8 cm/year during the first year of treatment, the average height standard deviation score (SDS) shifted by +0.52 SD. The variation in HV also increased, from S(2) = 1.0 before to S(2) = 3.4 cm(2)/year(2) during treatment. Lower leg length (LLL) velocity accelerated from 1.6 +/- 0.7 before treatment to 3.4 +/- 1.0 cm/year during the first 8 weeks of treatment. Four coefficients of correlation appeared clinically meaningful: (1) LLL velocity vs. body HV during the first year of GH treatment (r = 0.87), indicating that GH acts simultaneously on leg and rump growth; (2) early (first 8 weeks) LLL velocity vs. 1-year body HV during treatment, with r = 0.61 (R(2) = 0.38), indicating that 38% of the variation in HV during the first year of treatment is already predictable by an initial 8-week period of knemometry; (3) early (first 8 weeks) LLL velocity vs. 1-year LLL velocity during treatment, with r = 0.63 (R(2) = 0.39), and (4) early (first 8 weeks) LLL velocity vs. later LLL velocity, up to the end of the first year, with r = 0.53 (R(2) = 0.28) indicating that the early response on lower leg growth persists for at least 1 year of GH treatment. CONCLUSIONS: (1) Thirty-eight percent of the variation in HV during the first year of GH treatment is predictable by an initial 8-week period of knemometry. This parallels early changes in biochemical markers of bone turnover after GH treatment. (2) There is evidence for a baseline variability in HV both in healthy children and in children with growth disorders that make growth prediction difficult.