OBJECTIVE Children with GH deficiency (GHD) show a general decrease in electroencephalographic (EEG) arousability represented by a significant global decrease in Cyclic Alternating Pattern (CAP). The aim of the present study was to evaluate if sleep structure is influenced by GH substitutive therapy by analyzing the classical sleep architecture parameters and sleep microstructure by means of CAP. METHODS Laboratory polysomnographic sleep recordings were obtained from five children affected by GHD (two girls and three boys; mean age: 4.6 ± 3.1 years), at baseline and after GH therapy, and from 10 normal healthy children (four girls and six boys, mean age: 5.6 ± 2.2 years). RESULTS Compared to controls, GHD subjects showed a reduced total sleep time with increased wakefulness and a consequent decrease in sleep efficiency; GH therapy was associated with an increase of the awakenings/hour and a large effect size was evident for sleep latency, sleep efficiency, and stage N3, which were decreased, and for stage W, which was increased. CAP appeared to be globally reduced and all phase A subtypes and CAP cycle showed a longer duration in GHD children vs. controls. GH substitutive treatment was followed by an increase in CAP rate (total, in N2, and in N3); additionally, A1 index was also significantly increased, especially during stage N3, with a very large effect size. On the other hand, A2 and A3 index and CAP cycle mean duration were reduced. CONCLUSIONS Sleep stage architecture seems to be influenced by the GH status, but the analysis of sleep microstructure by means of CAP reveals an enhancement of EEG slow oscillations in GHD patients (demonstrated by an increase in CAP rate and A1 index during N3) after the start of GH replacement therapy. These findings deserve to be verified in a larger trial.