Prediction Model May Identify Children Eligible for Growth Hormone Treatment

In children eligible for growth hormone treatment, a Swedish model can exclude 98% of poor responders, according to the results of a recent study.  

­To validate a prediction model designed to accurately identify children who may benefit from treatment with growth hormone, the researchers studied 121 prepubertal children with growth hormone deficiency who began growth hormone treatment from 2004 to 2016 at Queen Silvia Children’s Hospital in Gothenburg, Sweden, and had at least 1 year of data after treatment was initiated. Observed heights of the children, who were primarily boys (64%), were compared with heights predicted by the Gothenburg model. The model predicted a first-year increase in height of at least a 0.7 standard deviation score (SDS) in children treated with a growth hormone dose of 33 μg/kg/d.

During the first year of treatment of this pediatric cohort, median growth hormone dose was 30 μg/kg/d (range, 10-43 μg/kg/d), and median age at treatment initiation was 5.3 years (range, 3.0-11.8 yr). The researchers found that the model was able to exclude 119 of 121 participants (98%) as poor responders, defined as those with height gain of less than 0.5 SDS in a year.

Nevertheless, the researchers found that the model underestimated relatively low predictions and overestimated relatively high predictions (residual standard deviation, 0.3 SDS).

The researchers also found that the 9 study participants who were born small for their gestational age did not differ from the main cohort in initial height or weight SDS or response to growth hormone treatment.

“We found that by using a validated prediction model in a clinical context, unnecessarily treating short children with an expected poor height outcome can be avoided,” the researchers concluded.

—Ellen Kurek


Ly HJ, Fors H, Nilsson S, Dahlgren J. A prediction model could foresee adequate height response in children eligible for growth hormone treatment. Acta Paediatr. 2022;111(2):346-353. doi:10.1111/apa.16070