Pituitary pars intermedia dysfunction (PPID) is an age-related neurodegenerative disorder, affecting >20 % of older horses. There is a need for improved endocrine tests for early disease detection, and the thyrotropin-releasing hormone (TRH) stimulation test has been recommended for diagnosis of early or mild cases. However, it is currently not recommended for year-round use due to marked seasonal variability. The aims of this cohort study were to evaluate effects of month and season on adrenocorticotropic hormone (ACTH) responses to TRH stimulation and to derive monthly cut-offs for PPID diagnosis. Sixty-three horses were assigned to control (n = 17), subclinical PPID (n = 21) and clinical PPID (n = 25) groups, based on a composite reference standard that combined clinical history and examination findings with endocrine test results. TRH stimulation tests were performed monthly for a 12-month period. Circannual changes were evaluated with one- and two-way repeated-measures analysis of variance and receiver operating characteristic curve analysis was used to derive cut-off values for basal and TRH-stimulated ACTH. TRH-stimulated ACTH concentrations were lowest in February-May and highest in August-October. Specificity of both basal and 30 min post-TRH ACTH was generally higher than sensitivity, and TRH stimulation had improved diagnostic accuracy compared to basal ACTH, although its sensitivity was not significantly greater year-round. TRH stimulation tests yielded considerably more positive results than basal ACTH in the subclinical group, but few additional positive results in clinical PPID cases. There were large differences between cut-offs that maximised sensitivity or specificity for TRH-stimulated ACTH, highlighting the importance of considering clinical presentation alongside test results in diagnostic decision-making.