Twenty-six cows that developed positive milk tests for ketosis in early lactation were randomly assigned to one of three groups: control, propylene glycol (125 ml/d), or 125 ml of propylene glycol plus 12 g of niacin daily. Treatments were administered for 7 d. Changes in milk production, milk composition, and feed intake were similar for all three groups. None of the cows in any of the groups developed clinical ketosis. During the 14-d monitoring period from the start of treatment there were similar increases in blood glucose in all groups. Declines in plasma beta-hydroxybutyrate and free fatty acids over the period were also similar for all treatments. Results reinforce the concept that optimum feeding and management reduces the need for additives for ketosis control.