OBJECTIVE: To test a semi-automatic oscillometric blood pressure measurement device with a wrist cuff. METHODS: Two different test methods, partly based on European and American standards for clinical investigation of non-invasive sphygmomanometers, were worked out and compared. Mostly German guidelines were considered. The reference measurement was always performed using the standard auscultatory method on the upper arm. Altogether, 404 comparative measurements were performed in 108 patients. RESULTS: The mean differences (oscillometric wrist values minus auscultatory reference values) were -2.5+/-8.5 mmHg (mean +/- SD) for systolic blood pressure and -3.3+/-6.7 mmHg for diastolic blood pressure in method 1, and -3.6+/-6.9 mmHg for systolic blood pressure and -3.7+/-4.8 mmHg for diastolic blood pressure in method 2. The calculated correlation coefficients were betewen 0.77 and 0.89. A grading analysis was performed using the cumulative percentages of readings which showed a difference between test and reference measurements larger than 5, 10 and 15 mmHg. CONCLUSION: