Uterine infections (i.e., endometritis) can have a major economic impact on dairy production. Identifying cows that are susceptible to endometritis and improving the diagnosis of endometritis could lead to a reduction in the impact of such infections. Thus, we used Holstein cows to determine whether postpartum changes in 13, 14-dihydro-15-keto-PGF2alpha (PGFM), a metabolite of PGF2alpha, could be used to identify cows that are susceptible to endometritis and to improve the diagnosis of endometritis. Cows were assigned to three treatments. 1) Control (n = 10) had no clinical or bacteriological signs of endometritis during the study. 2) Treated (n = 11) developed endometritis spontaneously and were treated i.m. with 25 mg of PGF2alpha immediately after clinical diagnosis (d 17.6 +/- 0.8 postpartum; mean +/- SEM). 3) Untreated (n = 10) developed endometritis spontaneously and were not treated after diagnosis (d 20.0 +/- 0.5). Examinations of external and internal genitalia and bacteriological data were used to diagnose endometritis. From d 0 (calving) until approximately d 63 postpartum, jugular blood was collected three times weekly. Progesterone and PGFM were quantified in plasma. For PGFM, the treatment x day interaction was significant (P < 0.01). Overall PGFM profiles for Control and Treated differed (P < 0.05), but the Untreated profile did not differ from either Control or Treated. To better understand the interaction, PGFM data from d 0 to 35 postpartum were partitioned into consecutive 7-d periods, and d-36 and greater data were partitioned into one period. Effects of treatment, day, and the treatment x day interaction were then evaluated within period. Except for the d-15 to -21 period, PGFM was greater (P < 0.03) in Control than in Treated and Untreated. In Treated and Untreated, PGFM increased during the d-15 to -21 period. For progesterone, treatment did not affect the profiles, but day was significant (P < 0.001). Progesterone concentrations were basal from d 0 until approximately d 12, and they generally increased after d 12. Onset of endometritis was associated with increased progesterone concentrations. Treatment did not affect the interval from calving to first detected estrus (29.5 +/- 4.9 d) or from calving to AI (73.3 +/- 8.7 d). We conclude that PGFM measures have the potential to be used to identify cows that are more likely to develop endometritis and that PGFM may aid in the diagnosis of endometritis.