In a controlled field trial, 178 dairy cows with chronic endometritis and at least 21 days in lactation were randomly assigned to four different treatment groups: prostaglandin F2alpha intramuscularly (PG, 5 mg dinoprost (5 ml Dinolytic), n = 51), intrauterine antibiotics (AB; 400 mg ampicillin + 800 oxacillin (20 ml Totocillin), n = 49), intrauterine antiseptics (AS; 100 ml 4% Lotagen, n = 50); control (C, no initial treatment, n = 28). Before treatment, uterine swabs for bacteriologic examination and blood samples for determination of serum progesterone concentrations were collected. Two weeks following the first treatment, cows were reexamined. In case no clinical cure was diagnosed, treatment was repeated and control cows were treated for the first time with one of the three treatments mentioned above. The four treatment groups did not differ with respect to the clinical cure or reproductive performance. Therefore, factors that might have an influence on clinical cure and fertility were evaluated. With increasing duration of lactation, the clinical cure after a single treatment increased significantly over all treatment groups from 59.5% (treatment before day 42 postpartum) to 79.6% (treatment following day 42 postpartum) (P < 0.05). Within the PG group, a statistically significantly higher cure rate after a single treatment and first service conception rate and a lower pregnancy index were obtained when the treatment was performed following day 42 postpartum (P < 0.05). This was not the case in the other treatment groups. A retarded involution of the uterus based on the size had a negative effect on clinical cure over all groups (first treatment clinical cure: 68.2% (small uteri) vs 44.4% (large uteri); P < 0.05). Within groups, this effect was also detected, but only as a trend (P > 0.05). Isolation of Arcanobacterium (A.) pyogenes negatively influenced first treatment clinical cure over all treatment groups (79.0% vs 31.5%) and within treatment groups (P < 0.05). In the AB group, the pregnancy index, days open and the interval from first insemination to conception increased compared with the other treatment groups, when A. pyogenes was detected. Isolation of unspecific bacteria and the presence or absence of a corpus luteum only had minor effects over all and within the PG, AS and C group. Within the AB group, presence of luteal tissue was connected with a higher pregnancy index and increased days open and interval from first insemination to conception (P < 0.05). A multivariate analysis revealed that the presence of A. pyogenes and no initial treatment (control group) had a statistically significant independent and negative influence on the first treatment clinical cure. Ill-smelling vaginal discharge and no treatment had a negative independent effect on first service conception rate. Furthermore, days open were negatively influenced by the interval from calving to first examination and the presence of ill-smelling vaginal discharge. In conclusion, under the conditions of this field study, it seems preferrable to treat chronic endometritis. However, under the condition of a herd health program with continuous observation of sick and healthy animals, a wait-and-see attitude might also be justified.