Cyclo-oxygenase 2 (COX-2) is implicated in the regulation of aromatase transcription in malignant breast tissue and has been considered as a potential target for tissue specific aromatase suppression. We initiated a randomised controlled pre-surgical study of celecoxib versus no treatment in women with primary breast cancer to determine the effects of COX-2 inhibition on markers of biological response. Postmenopausal women (50-80 years of age) with stage I or II, primary breast cancer, were randomised 2:1 to receive 400 mg/day celecoxib or no treatment for 14 days prior to surgery. A core biopsy was obtained pre- and post-treatment. Paired baseline and endpoint biopsies were analysed for Ki67, apoptosis, COX-2, CD31, estrogen receptor (ER) and progesterone receptor (PgR). Comparisons between the treatment groups were conducted using the Mann-Whitney test with a two-sided 5% significance. Of the 25 patients treated, 23 had evaluable data and 19 (83%) were ER positive. Overall the geometric mean change in Ki67, the primary end point, relative to baseline in the celecoxib arm was -16.6% (P = 0.056). The change in the no-treatment group was -8.1% (P = 0.24). There was no statistically significant difference in the change between the two groups. Celecoxib did not significantly affect apoptosis, COX-2, ER or PgR expression. There is only modest evidence for a biological effect of celecoxib in primary breast cancer. However, the trend towards a reduction in Ki67 in ER-positive breast cancer warrants further investigations in a larger cohort of patients.