We observed the new onset of severe obstructive sleep apnoea syndrome (OSAS) in an adult male patient during human growth hormone (hGH) replacement therapy. This prompted us to evaluate the potential influence of hGH substitution therapy on sleep in middle-aged men. A longitudinal study. Five male patients (aged 44-56 years, median age 54 years) with postoperative pituitary insufficiency given hGH replacement therapy for 1-2 years (median dose 2.0 U/day; median IGF-I serum concentration 351 microg/l) and 6 months after cessation of hGH treatment (median IGF-I level 77 microg/l - 1 microg/l = 0.131 nmol/l). Polysomnographic studies were performed, and the following parameters were determined: time in bed (TIB), sleep period time (SPT), total sleep time (TST), sleep efficiency (SE = TST/TIB), sleep stage 1 onset latency (SL), different sleep stages [W (wake), S1, S2, SWS (slow wave sleep = S3 + S4) and REM; % of SPT], stage shifts per hour of SPT (SS/h), stage shifts to W/h of SPT [A/h (awakening)], index of apnoea and hypopnoea events per hour of TST (AH/h), arousals from apnoea and hypopnoea per hour of TST (Ar/h), index of obstructive (OAH/h), central (CAH/h) and mixed (MAH/h) events of apnoea and hypopnoea per hour of TST and minimal desaturation (MD). Median baseline results were: TIB, 479 min; SPT, 465 min; TST, 405 min; SE, 77%; SL, 8.5 min; W, 18.9%; S1, 8.2%; S2, 52.7%; REM, 13.5%; SS/h, 17.7; A/h, 2.8; AH/h, 11.9; Ar/h, 4.4; MD, 80%. These parameters did not change significantly after cessation of hGH treatment. In contrast, median SWS decreased significantly from 33 min (7.1%) to 7.5 min (1.8%; P = 0.03). Median OAH/h decreased significantly from 4.4 to 0.1 (P = 0.03) whereas CAH/h increased from 6.3 to 14.6 (P = 0.03) after cessation of hGH. Correspondingly, one patient with OSAS improved markedly whereas another patient developed new and asymptomatic central SAS after cessation of hGH. This study showed that hGH replacement therapy influenced sleep reaction in a complex way in middle-aged men; cessation of treatment was associated with a significant decrease in slow wave sleep and a shift from obstructive to central apnoea and hypopnoea.