This prospective study was designed to investigate D-dimer concentrations in clinically healthy dogs, clinically ill dogs without thromboembolic disease (TE), and dogs with TE. The goals of this study were to determine whether the coagulation cascade is activated in nonembolic metabolic and inflammatory conditions and whether differentiation from TE is possible. Group 1 consisted of 30 clinically healthy dogs presented for routine care. Group 2 consisted of 67 clinically ill dogs without TE. This group was subdivided into the following categories: postoperative surgical procedures, congestive heart failure, renal failure, hepatic disease, and neoplastic disease. Group 3 consisted of 20 dogs diagnosed with TE. A CBC and a measurement of prothrombin time (PT), activated partial thromboplastin time (PTT), fibrinogen degradation product (FDP) concentration, and plasma D-dimer concentration was performed on dogs in all groups. D-dimer concentrations were highest in dogs with TE; next highest was the hepatic disease group. Only these 2 groups had median D-dimer concentrations markedly different from clinically healthy dogs. The frequency of platelet abnormalities was markedly greater for the TE and neoplastic disease groups. The sensitivity of D-dimer concentrations >500 ng/mL for predicting TE was 100%; however, the specificity of D-dimer for TE at that concentration was 70%. The specificity of D-dimer concentrations >1,000 ng/mL to predict TE was 94% (sensitivity, 80%), and the specificity of D-dimer concentrations >2,000 ng/mL was 98.5% (sensitivity, 36%). FDPs were not high in any TE patient; thus, they may be an insensitive indicator of thromboembolism, with or without overt disseminated intravascular coagulation (DIC).