ABSTRACT. Transsphenoidal surgery (TSS) is an effective treatment for canine Cushing's disease, as well as human Cushing's disease. In humans, only the pituitary adenoma tissue is resected by TSS. However, in dogs, the whole pituitary including normal tissue is resected. Hence, central diabetes insipidus (CDI) may complicate the postoperative course in almost all dogs with Cushing's disease treated by TSS. However, it is difficult to assess the duration of the postoperative CDI, and whether it may be transient or permanent. In this study, postoperative CDI in 21 dogs with Cushing's disease and its predicted prognosis by preoperative parameters was investigated. In this study, CDI after TSS was classified as either transient or permanent based on the requirement for desmopressin. Preoperative circulating serum cortisol concentrations and pre-and post-operative plasma AVP concentrations were not significantly different between the transient CDI dogs and permanent CDI dogs. The duration of postoperative CDI was not correlated to the signal intensity ratio (posterior lobe of the pituitary/cerebral cortex), which is obtained from preoperative magnetic resonance imaging (MRI). However, the pituitary height/brain area (P/B) ratio of the permanent CDI dogs was significantly greater than that of the transient CDI dogs. In addition, there was a significant difference between the enlarged-pituitary (P/B ratio > 0.31) transient CDI dogs and the permanent CDI dogs. These results suggest that the incidence of postoperative permanent CDI is strongly influenced by the pituitary size.KEY WORDS: arginine vasopressin, dog, magnetic resonance image, pituitary tumor, signal intensity.