The overshoot rise in arterial pressure after release of Valsalva strain is a natural challenge for baroreflex regulation of heart rate. To assess the feasibility of the Valsalva manoeuvre in the determination of baroreflex sensitivity (BRS), we measured the slope of the linear relationship between the length of the RR interval and preceding systolic blood pressure value during the overshoot phase after the strain and compared this index of BRS to a standard phenylephrine test in 64 subjects, of whom 58 had coronary artery disease. The BRS slopes obtained with the Valsalva manoeuvre showed a good linear correlation with the phenylephrine test (r = 0.77 in the 27 patients with two Valsalva and phenylephrine tests and r = 0.56 in the whole cohort). The correlation coefficients of the BRS slopes were better than in the phenylephrine test (r = 0.89 vs r = 0.85, P < 0.05). The rise in systolic blood pressure in the slope calculation area was higher than with phenylephrine (41 +/- 18 vs 30 +/- 10 mmHg, P < 0.01). The reproducibility of BRS slopes in successive tests was comparable with both methods. These results suggest that non-invasive assessment of BRS using Valsalva strain to induce blood pressure rise is possible in patients with coronary artery disease.