IgG antibodies to cardiolipin and beta 2-glycoprotein I were looked for using an enzyme-linked immunosorbent assay (ELISA) in 19 patients with giant cell arteritis (meeting 1990 American College of Rheumatology criteria), including 16 with concomitant polymyalgia rheumatica (meeting Bird's criteria) and in three patients with isolated polymyalgia rheumatica. IgG anti-cardiolipin antibodies were demonstrated in eight patients (36%) and IgG anti-beta 2-glycoprotein I antibodies in two patients (9%) including one without anti-cardiolipin antibodies. Titers of anti-cardiolipin antibodies ranged from 27 to 190 units of IgG antiphospholipid antibodies (UGPL) (mean 71 UGPL). Of the eight patients with anti-cardiolipin antibodies, two had giant cell arteritis without polymyalgia rheumatica and six had polymyalgia rheumatica with clinical (n = 2) or histologic (n = 4) evidence of giant cell arteritis. None of the three patients with polymyalgia rheumatica but no giant cell arteritis had anti-cardiolipin or anti-beta 2 glycoprotein I antibodies. The VDRL was negative in the 14 patients who had this test. Tests for lupus anticoagulant were performed routinely, always with negative results. Among giant cell arteritis patients, those who tested positive for anticardiolipin antibody had significantly higher values for the erythrocyte sedimentation rate (p < 0.006) and for serum C-reactive protein (p < 0.03) and fibrinogen values (p = 0.05), and a trend toward higher platelet counts, as compared to those who tested negative for anticardiolipin antibody. The mean daily prednisone dose at the time of sampling was significantly lower in giant cell arteritis patients with anti-cardiolipin antibodies (p < 0.05); this difference may account for the apparent correlation between anti-cardiolipin antibodies and laboratory markers for inflammation. These data, as well as findings from serial measurements, suggest that anti-cardiolipin antibodies are present early in the course of giant cell arteritis and disappear within a few weeks of initiation of corticosteroid therapy in a dose of more than 25 mg prednisone per day. In this study, only one patient without anticardiolipin antibodies developed a cerebrovascular accident. Positive tests for anti-cardiolipin antibody or anti-beta 2 glycoprotein I antibody in a patient with polymyalgia rheumatica suggest a diagnosis of concomitant giant cell arteritis, which is usually symptomatic.