Investigation of a patient with an unexpected prolonged activated partial thromboplastin time showed on repeat testing that the patient had normal coagulation times, but had been preceded on the automated coagulation analyzer run (ACL 300R--Instrumentation Laboratory, Lexington, MA) by a heparin-contaminated apheresis line sample taken from a different patient. Heparin carryover was suspected, and subsequent experiments showed that ex vivo heparinization of normal plasmas will produce similar effects in following normal plasma samples in a test series. This effect increases with the baseline activated partial thromboplastin time and the heparin concentration. It is recommended that coagulation analyzers that are susceptible to carryover be cleaned after testing of plasmas likely to have heparin concentration of more than 5 x 10(3) U/L. Furthermore, when using analyzers of this type, investigation of an unexplained activated partial thromboplastin time must include a review of the preceding samples tested, particularly when duplicate testing has been discontinued.