The aim of this study was to investigate and quantify the agreement between simultaneous and ipsilateral non-invasive finger artery blood pressure (Finapres) and intra-arterial radial blood pressure among 13 volunteer hypertensive patients, aged 36-71 years and taking cardiovascular medication, during steady-state fluctuation of arterial blood pressure and during an increase in blood pressure induced by static exercise. Eight patients were being treated with beta-blocking agents, eight with calcium antagonists, four with angiotensin-converting enzyme inhibitors, four with diuretics and one with prazosin in combination therapy. Their auscultatory brachial artery blood pressures ranged in systole from 142 to 206 mmHg and in diastole from 88 to 120 mmHg during the treatment. The mean systolic finger artery blood pressure deviated by -14 +/- 5 mmHg (P = 0.02, mean value +/- SEM) and the diastolic finger artery blood pressure deviated by 0.6 +/- 3 mmHg (P = 0.70) from the corresponding radial artery pressure. The maximal beat-to-beat difference between systolic and diastolic finger and radial artery pressure, respectively, showed that a range of less than 10 mmHg in the steady state after individual adjustment for bias. In general, neither systolic nor diastolic differences between the methods exceeded the limits of +/- 10 mmHg, and the bias did not significantly increase (P > or = 0.12) during a 10-mmHg increase in arterial blood pressure caused by static exercise. Among three subjects, an increase in bias and poorer agreement was associated with atrial fibrillation and steplike changes in the Finapres output after autocalibration. The results support usage of the Finapres technique to measure beat-to-beat changes of peripheral arterial blood pressure in hypertensive patients taking cardiovascular medication, with a feasible agreement with beat-to-beat radial artery blood pressure.