Retention of the foetal membranes (RFM) and post-partal endometritis are common problems in dairy cows. Among other things, the disease is characterized by a bacterial endometritis with aerobic as well as anaerobic bacteria. From an endocrine perspective, cows with RFM have high levels of 15-ketodihydro-PGF(2 alpha) (PG-metabolite) immediately after parturition but these levels fall rapidly within 2 weeks post-partum (early PG-metabolite elevation). After this decline, the PG-metabolite levels increase again and the levels (at this time of a lower magnitude) remain elevated during the period of uterine infection (late PG-metabolite elevation). The aim of this study was to investigate the PG-metabolite profiles in cows with retained placenta and post-partal endometritis treated with the prostaglandin synthesis inhibitor flunixin (F), either alone or in combination with oxytetracycline (T). The study was accomplished over 2 years with 12 primiparous cows in each experiment. As a model for RFM, preterm parturition was induced in late-pregnant heifers by injecting PGF(2 alpha) (25 mg i.m) twice with a 24 h interval. In each experiment, the cows were divided into four groups and treated with either T (10 mg/kg b.w. i.m. once daily), F (2.2 mg/kg b.w. p.o. twice per day), a combination of T and F (dosage, as above), or conservatively (0). The treatment periods lasted from day 11 to day 14 post-partum (pp) in experiment 1 (after placental shedding, groups T1, F1, TF1 and 0) and from day 3 to day 6 pp in experiment 2 (before placental shedding, groups T2, F2, TF2 and 0). Jugular vein blood samples were collected for analyses of PG-metabolite and flunixin. Uterine biopsies were collected twice weekly for investigation of endometrial microbiology. Rectal palpation and ultrasonographic examinations were performed three times per week for investigations of uterine and cervical involution and ovarian activity. No attempts were made to remove the placentas manually. The experiment lasted until day 56 pp. The induction of parturition was successful in all heifers and 22 of 24 animals had RFM. All RFM cows had bacterial endometritis, based on bacteriological examinations. Flunixin treatment (F1, TF1, F2 and TF2) suppressed PG-metabolite levels significantly (p=0.006) during the period of treatment in both experiments. However, the early flunixin treatment only suppressed PG synthesis partially. Late oxytetracycline treatment (T1) did not influence the PG-metabolite levels but oxytetracycline treatment (T2 and TF2) before placental shedding significantly altered the kinetics of the early PG-metabolite elevation compared with other treatments. Late PG-metabolite elevation was significantly correlated to duration of uterine infection and cervical involution. In conclusion, flunixin treatment of cows with retained placenta either before or after placental shedding suppresses prostaglandin synthesis. However, early treatment, when the release of prostaglandins is high, might need more intensive treatment in order to prevent the PG synthesis effectively. Oxytetracycline treatment during the period immediately after parturition before placental shedding might influence the PG-metabolite profile and suggests a bacteriological contribution to the high levels of PG-metabolite seen during the first 2 weeks pp in cows with retained placenta. In this study, a correlation between prostaglandin release, the final cervical involution and the end of infection was found. This suggests a link between uterine endocrinology, bacteriology and involution in cows with retained placenta.