In a prospective study assessing haemostatic functions, the activated partial thromboplastin time was prolonged in 134 out of 10,229 patients studied, without an increase in the prothrombin or thrombin times; this abnormality persisted in only 37 of them on a new blood sample. A retrospective analysis was made of 265 patients who had such an isolated prolongation of the activated partial thromboplastin time on two successive blood samples: the causal abnormality remained unexplained in 135 patients; a well defined coagulation disorder without abnormal bleeding tendency was present in 110 patients (1 severe factor XII deficiency, 58 partial factor XI or XII deficiencies and 51 lupus anticoagulants); a bleeding disorder was diagnosed in 20 patients (8 haemophilias, 8 Von Willebrand's diseases, 4 factor VIII inhibitors). The well-iron efficacy of the activated partial thromboplastin time for detecting coagulation abnormalities is counter-balanced by some disadvantages such as the delay for biologic conclusions. In the preoperative assessment of haemostatic functions, rather than taking a routine approach, it would seem better to determine for each patient the need and the extent of biological testing according to the type of planned surgery, the clinical status of the patient and possible bleeding symptoms.