Forty-one subjects received GH replacement. Nine of them did not have follow-up imaging during GH therapy and were not included in the statistical analyses. The remaining 32 (22 males/10 females) received GH for a mean period of 6.3 +/- 4.6 years (median 5.1, range 0.8-22); 21 started during childhood (13 of them continued after the achievement of final height with an adult dose) and 11 during adult life. The mean duration of their follow-up (from surgery until last assessment) was 10.8 +/- 9.2 years (range 1.9-40). Fifty-three subjects had not received GH therapy (30 men/23 women). The mean duration of their follow-up (from surgery until last assessment) was 8.3 +/- 8.8 years (range 0.5-36). During the observation period, 4 patients treated with GH and 22 non-GH treated ones developed tumour recurrence. After adjusting for sex, age at tumour diagnosis and type of tumour therapy (gross total removal, partial removal, surgery + irradiation), GH treatment was not a significant independent predictor of recurrence (P = 0.06; hazard ratio = 0.309). Similar results were obtained when the impact of GH replacement was assessed according to its duration (P = 0.18; hazard ratio = 0.991/month of treatment). None of the nine patients with insufficient imaging data for inclusion in the statistical analyses [5 men/4 women, 3 treated with GH during childhood/6 during adult life, mean duration of GH therapy 2.9 +/- 2.4 years (median 1.8, range 0.4-7)] showed clinical features suggestive of recurrence during the period of GH replacement. CONCLUSION Based on the data of the craniopharyngiomas database in Oxford, there is no evidence that GH replacement is associated with an increased risk of tumour recurrence.

GH replacement in patients with NFA treated by surgery alone is not associated with an increased risk of tumour recurrence.

Ovarian morphology overlaps in PCOS cases and controls, emphasising the importance of considering clinical/biochemical presentation together with imaging ovarian morphology in the diagnosis of PCOS.