Objectives were to compare the incidence of disease and days spent in the hospital for cows diagnosed with metritis treated either with ampicillin trihydrate (AMP) or ceftiofur hydrochloride (CEFT) and matched control herdmates without metritis. Holstein cows with fetid, reddish/brownish watery vaginal discharge were diagnosed with metritis. Cows with metritis were blocked by parity and type of metritis (with or without fever) and assigned randomly to receive 11 mg/kg of AMP (n = 259) or 2.2 mg/kg of CEFT (n = 269) daily for 5 d. A cohort of healthy cows (NOMET, n = 268) with the same days postpartum and parity was selected as controls. Postpartum disorders (displacement of abomasum, fever, mastitis, and pneumonia) were monitored for the first 60 d postpartum. Data were analyzed using GLIMMIX procedure of SAS. Cows with metritis tended to have a higher incidence (P = 0.09) of postpartum disorders than cows without metritis (41.3% vs. 29.5%). Multiparous cows had a higher incidence (P < 0.001) of postpartum disorders than primiparous cows (44.9% vs. 28.9%). The incidence postpartum disorders was greater (P = 0.01) in AMP (45.6%) cows than in CEFT (37.2%) and NOMET (29.5%) cows. There were no interactions among treatment and parity. Cows with metritis spent more days in the hospital than (P < 0.01) cows without metritis (6.0 vs. 2.7). The number of days in hospital was greater (P < 0.01) in AMP (8.6) cows than in CEFT (3.4) and NOMET (2.7) cows. Multiparous cows spent more days in the hospital (P < 0.01) than primiparous cows (5.8 vs. 4.1). Among the individual disorders, mastitis was increased in AMP (34.0%) when compared with CEFT (24.5%) and NOMET (19.0%). There was no interaction between treatment and parity. Metritis increased the number of days in the hospital and tended to increase postpartum disorder. Cows treated with AMP had more postpartum disorders, cases of mastitis and a greater number of days in the hospital than CEFT cows.