In order to investigate the clinical value of anti-mitochondrial antibodies type 5 (anti-M5), we carried out a retrospective study on 48 patients with these antibodies. Seventeen of these 48 patients (35%) satisfied at least 4 criteria of the revised American Rheumatism Association classification of SLE. Twenty-nine (61%) had at least one clinical manifestation of anti-phospholipid syndrome; thirteen had symptoms consistent with primary anti-phospholipid syndrome; five had isolated recurrent thrombosis; five had Evans' syndrome; four had auto-immune haemolytic anaemia; two had immunologic thrombocytopenia. Two of the 48 patients had no clinical manifestations, but only anti-M5 and a false laboratory test for syphilis (FBTS). Our data confirm that patients with anti-M5 have a high prevalence of: 1) thrombosis (42% had three or more deep thromboses) and fetal loss (21%); 2) auto-immune cytopenia with idiopathic thrombocytopenic purpura (29%) and auto-immune haemolytic anaemia (54%); 3) laboratory markers of anti-phospholipid syndrome (lupus anticoagulant (71%), FBTS (95%) and anticardiolipin antibodies (aCL) (71%). For 32 patients with anti-M5, anti-beta 2 glycoprotein I antibodies were also tested; 12 (38%) were positive, all of whom had IgG aCL, ie none had anti-beta 2GPI antibodies without aCL. There was no association between the presence of anti-beta 2GPI antibodies and recurrent thrombosis among patients with anti-M5. All these findings suggest that anti-M5 is another marker of the antiphospholipid syndrome. Even though the prevalence of anti-M5 is low, especially in SLE, it was the only marker of the anti-phospholipid syndrome in two patients; this appears to justify routine screening for these antibodies.