The aim of this study was to retrospectively describe the outcome of 127 dogs with naturally occurring diabetic ketoacidosis (DKA) and to examine the association between outcome of canine DKA and clinical and clinicopathologic findings. Eightytwo (65%) dogs were diagnosed with DKA at the time of initial diagnosis of diabetes mellitus (DM). Eighty-seven dogs (69%) had one or more concurrent disorders diagnosed at the time of hospitalization. Commonly identified concurrent conditions included acute pancreatitis (52, 41%), urinary tract infection (21, 20%), and hyperadrenocorticism (19, 15%). Dogs with coexisting hyperadrenocorticism were less likely to be discharged from the hospital (P = .029). Of 121 treated dogs, 89 dogs (70%) survived to be discharged from the hospital, with a median hospitalization of 6 days. Nonsurvivors had lower ionized calcium concentration (P < .001), lower hematocrit (P = .036), lower venous pH (P = .0058), and larger base deficit (P = .0066) than did survivors. Time from admission to initiation of subcutaneous insulin therapy was correlated with lower serum potassium concentration (P = .0056), lower serum phosphorus concentration (P = .0043), abnormally high white blood cell count (P = .0060), large base deficit (P = .0015), and low venous pH (P < .001). Multivariate analysis showed that base deficit was associated with outcome (P = .021). For each unit increase in the base deficit, there was a 9% greater likelihood of discharge from the hospital. In conclusion, the majority of dogs with DKA were not previously diagnosed with DM. Concurrent conditions and electrolyte abnormalities are common in DKA and are associated with length of hospitalization. Survival was correlated to degree of anemia, hypocalcemia, and acidosis.