Testing strategies for diagnosing lupus anticoagulant: Decision analysis

Clinicians commonly evaluate patients with thrombosis or a prolonged activated partial thromboplastin time (aPTT) for the presence of a lupus anticoagulant (LA). We evaluated strategies for detecting LA, in three clinical settings, with decision-modeling techniques. A decision tree was constructed with 12 strategies, using a combination of aPTT and dilute Russell viper venom times (dRVVT) with confirmatory tests, tissue thromboplastin time (TTI), platelet neutralization procedures, and mixing studies. Probabilities and costs of adverse events and test costs were obtained from the literature. Patient preference for each strategy was evaluated by assigning utilities to each outcome. On the basis of assay results in 90 healthy people and 77 patients, we calculated sensitivities and specificities for each strategy, with true positives defined as suggested by the International Society on Thrombosis and Haemostasis. The least costly strategy for evaluation of patients with a prolonged aPTT, or with thrombosis, is not to test and to assume that LA is absent. For patients with systemic lupus erythematosus (SLE), it is least expensive not to test, although testing with TTI alone can also be considered an efficient strategy. The strategy of highest utility to patients with SLE is testing with TTI, followed by dRVVT. On the basis of these cost and utility results, clinicians' strategies for detecting LA may need modification. These strategies would then optimally be tested in clinical trials.