In the present study, we investigated the effects of reproductive status, size of follicles and plasma progesterone concentrations of mares at PRID insertion on the efficacy of the treatment, estrous cycle patterns, plasma concentrations of progesterone and LH. The progesterone-releasing device (PRID) was administered intravaginally to 28 Haflinger mares for 11 days at different reproductive stages: anestrus (n=6), estrus (n=11) and diestrus (n=11). Plasma concentrations of progesterone at insertion (Day 1) of PRID differed among treatment groups (anestrus: 0.2-0.6 ng mL(-1), estrus: 0.2-0.5 and diestrus: 1.6-10.8 ng mL(-1); P<0.001). Total secretion of progesterone (area under curve (AUC)) during treatment period revealed highest values in diestrus (38.2+/-3.1 ng mL(-1)h(-1)) followed by estrus (25.1+/-2.7) and anestrus (21.0+/-0.4 ng mL(-1)h(-1); P<0.05). Progesterone area under curve (AUC) was positively correlated with initial progesterone concentrations (R=0.5; P<0.05), but it did not correlate with the interval from PRID removal to ovulation. Plasma concentrations of LH during treatment period, were significantly lower in anestrous mares (184.6+/-28.6 ng mL(-1)h(-1)) when compared to estrous and diestrous mares (349.7+/-53.3 and 370.5+/-40.3 ng mL(-1)h(-1); P<0.05). Follicular size at PRID insertion had no effects on the intervals from PRID removal to subsequent estrus and ovulation. Follicle diameters at removal of PRID were significantly correlated with the interval from coil removal to estrus (R=-0.55, P<0.05) and ovulation (R=-0.72, P<0.0004) in cyclic mares. In anestrus 0 of 6 (0%) mares, in estrus 5 of 11 (45.5%) and in diestrus 6 of 11 (54.5%) mares ovulated within a defined interval of 1 day before to 1 day after mean interval from PRID removal to ovulation. In cyclic mares, response to treatment was significantly higher when compared to anestrous mares: almost all mares responded with estrus and ovulation independent from the stage of the estrous cycle at the start of treatment. However, accuracy of synchronization was still unsatisfactory. In cyclic mares, the plasma progesterone concentrations at insertion of PRID seem to be more important for the efficacy of the treatment than the assignment to estrous cycle stages.