The middle turbinates (MT) are sometimes partially resected during endoscopic transnasal transsphenoidal pituitary surgery to improve surgical access. Some outfracture the middle turbinates instead. This article presents results of the first prospective randomised study, investigating the effects of such surgery on olfaction, and also comparing results between resection and lateralisation of middle turbinates. Computer-generated simple randomisation was used. In the MT-resection group, the lower halves of the MTs were trimmed while in the MT-reserved group, the MTs were fractured laterally. The following steps were then performed: resection of lower third of superior turbinates, sphenoidotomy, removal of posterior nasal septum (approximately 1cm) and sphenoid septum, diamond burr drilling to expose dura that is incised, tumour excision, synthetic dural substitutes applied for haemostasis and closure of dura, and packing the sphenoid with dissolvable nasal packing. A nasoseptal flap (raised 2cm from skull base) was only used in patients with high-flow intraoperative cerebrospinal fluid leak. A short course of hydrocortisone was administered postoperatively for possible hypocortisolism. Patients are followed up with the ENT team at one and two weeks, and one, two, three, and six months postoperatively. Patients are encouraged to perform nasal saline irrigation after the first follow-up, and endoscopic debridement is performed at two weeks follow-up. The Taiwan Smell Identification Test (TWSIT) scores were documented before surgery and at one, three and six months after. Endoscopic appearance was scored using the Lund-Kennedy endoscopy score (LKES). Patients included were among those undergoing such surgery for sellar tumours at Mackay Memorial Hospital, Taiwan. Of 49 patients enrolled, 23 had MT-resection and 26 MT-preserved. Results showed worsened olfaction postop (mean preoperative TWSIT score was 41+6.9 and was significantly worse at one month 36.9+ 11.3). However, olfaction eventually improved with TWSIT returning to baseline at three months 37.8+10.3. There was no significant difference between patients who had pre-existing preoperative normosmia (seven patients), hyposmia (35), or anosmia (two). LKES scores (preop 0.2+0.7) were worse at two weeks 2.5+0.9, one month 2.5+0.9 and improved at three months 0.7+0.8, eventually returning to baseline at six months 0.3+0.4. There was no significant difference between the MT-resection and MT-preserved groups. Postoperative loss of olfaction could be due to oedema and crusting. There was no significant difference between results of patients with or without nasoseptal flaps raised.