Since 1991, the Ontario Laboratory Proficiency Testing Program has assessed the analytical performance of creatine kinase (CK; EC 220.127.116.11) isoenzyme-2, using fresh human serum supplemented with purified human CK isoenzymes. In Ontario, the 142 laboratories licensed to analyze CK-2 use a variety of methods: electrophoresis-based, immunoinhibition, and mass assays. During a 1992 survey, duplicate CK-2 samples with different total CK activities showed poorer precision when analyzed after electrophoretic separation than by any other method. A 1993 survey designed to validate these observations conclusively showed that electrophoresis-based assays are subject to a bias proportional to the total CK activity. These survey results were confirmed by studies with selected patients' specimens. We therefore conclude that electrophoresis-based assays may not warrant their reputation as the gold standard for CK isoenzyme measurement.