The use of Insuject-X for NPH insulin was tested by 21 children, age less than or equal to 10 years with insulin dependent diabetes mellitus less than or equal to 1 year, on conventional treatment with NPH insulin once or twice daily. After a preliminary period of one month the syringe injection regimen was changed to injections by Insuject-X for a four-month "pen" period. We observed no significant changes in metabolic control. The mean potency of NPH insulin in partly used cartridges was 103.6 +/- 5.6 IU/ml (range 91.0-118.1, n = 39) with no correlation to the residual volume or days of use. Measurements over 110 IU/ml were seen at the end of the study indicating the need of repeated instruction in use of the pen by children. No microbial contamination of the cartridges, nor any local reactions at injection sites were found. We observed more technical problems than in adult studies indicating that children handle the pen more roughly. The accept of Insuject-X was good as a more convenient means of injecting. All of the children preferred to continue using it.