Determinants of neurosurgical outcome in pituitary tumors

Objective: Neurosurgery is one of the main therapies for pituitary tumors; optimising outcome is highly desirable for the patient and the health system. We have analysed predictors of outcome in surgically treated pituitary adenomas operated in this centre. Design and patients: A total of 289 patients underwent neurosurgery for a pituitary tumor, by the same two neurosurgeons, between 1982 and 2001. Their records were examined to find predictors of post-surgical outcome. Thirty-eight percent were males, with a median age of 40.8 (8–82.7) yr; 51.9% had been operated since 1992, 92.2% by the transsphenoidal route. Most tumors (70.2%) were macroadenomas; 28.4% were non-functioning, 27.3% secreted PRL, 26.3% GH of which 14 (4.8%) also secreted PRL, 17.3% ACTH, 0.3% FSH and 0.3% TSH. Results: A stepwise, forward logistic regression analysis revealed tumor size as the only significant predictor of radiological cure [odds ratio (OR) for macroadenoma 0.16 vs microadenoma, p=0.0005]. Hormonally, PRL-secretion by the tumor was a predictor of poor prognosis (OR 3.29 for cure of non-PRL-secreting tumors, p=0.005), as was tumor size (OR 0.45 for cure of macroadenomas, p=0.005). Considering simultaneous radiological and hormonal remission, tumor size (OR 0.35 for macroadenoma, p=0.0002), and operation date (OR 0.40 for up to 1991, p=0.0002) were the only significant predictors. Conclusions: PRL secretion, tumor size and operation date are the main predictors of neurosurgical outcome in pituitary tumors, the latter suggesting that neurosurgical experience plays an important role.