In clinical settings, when the reproductive history of a near-term bitch is limited to mating dates, the possibility to accurately assess whether pregnancy is at term could be very useful in order to be able to plan a correct management of parturition or to safely perform an elective Caesarean section. The aim of this study was to assess the diagnostic efficacy of a single progesterone determination, measured by chemiluminescent immunoassay (CLIA), in predicting the occurrence of parturition on the following day. At least one blood sample was collected from 51 pre-partum bitches during the 3 days before parturition and on day of parturition. The efficacy of progesterone as a marker of the end of pregnancy was tested using a receiver operating characteristic (ROC) analysis. Youden's index was calculated to select the optimal cut-off value (with 95% confidence interval), aiming at maximizing the correct identification of negative events, so not to risk to diagnose as full term a bitch which is not. Progesterone concentration lower than 3.4 ng/ml correctly identified the bitches whelping the following day; however, because of the obliged prudential approach, sensitivity was low (46.88%), and 17 of 32 full-term bitches were missed. Due to a very large individual variation, a single progesterone determination has low diagnostic efficacy, although it can represent a useful first screening.