Ketosis in early lactation dairy cattle can negatively impact health events, milk yield and reproductive performance causing a loss in profit for dairy producers. The objective of this retrospective observational study was to identify factors such as management, metabolic profile and milk production that contribute to ketosis in early lactation. One hundred and thirty-two multiparous dry and lactating Holstein cows from a California commercial dairy herd were bled weekly from 3 wk before calving to the dry period via coccygeal venipuncture into evacuated sterile tubes containing sodium fluoride and potassium oxalate. Milk yield, fat and protein percentage were measured monthly by Tulare County Dairy Herd Improvement Association using a Bentley Instrument ChemSpec 150 (Chaska, MN). Blood metabolite data including nonesterified fatty acids (NEFA) using a commercial kit from Wako Chemicals Inc. (Richmond, VA), glucose and BHB using a Precision Xtra handheld meter (Abbott Diabetes Care Inc., Alameda, CA), and metabolic data using an Abaxis Vet Scan analyzer Large animal profile rotor (VetScan®, Abaxis, Inc., Union City, CA) were measured weekly. Management variables and milk production from Dairy Comp 305 (Valley Agricultural Software, Tulare, CA) were collected monthly in the previous and current lactation. Using the General Linear Model Procedure of SAS (SAS Institute v.9.4, 2018), BHB were regressed on blood metabolites, management data and milk production. Variables were eliminated from the regression if P > 0.05. The significant management variables that remained in the regression were total days in milk in previous lactation and days dry in previous lactation (P < 0.05, R² = 0.70). When metabolic variables such as aspartate amino transferase, globulin, blood urea nitrogen, NEFA, and glucose from −7 to 14 d relative to calving were added, and milk production variables such as total solids and total fat in the previous lactation were added to the regression, P < 0.05 and R² = 0.92. Residuals were normally distributed and unbiased. These equations indicate that cows at risk for ketosis could be identified before calving to help dairy producers manage ketosis before it negatively affects health and production performance.