1. Pulse-wave analysis (PWA) is an established method to assess arterial wave reflections and arterial vasoreactivity in humans. A high short-term reproducibility of baseline augmentation index (AIx) has been reported. However, the short-term reproducibility of AIx changes following pharmacological challenge with either inhaled salbutamol (endothelium-dependent vasodilatation) or sublingual glyceryl trinitrate (GTN; endothelium-independent vasodilatation), using appropriate statistical methods, is largely unknown. 2. Baseline AIx and GTN- and salbutamol-mediated changes in AIx (all corrected for a heart rate of 75 b.p.m.) were measured on two separate occasions, 1 h apart, in 22 healthy controls (mean (+/-SD) age 52.0 +/- 13.4 years) and 11 elderly patients with chronic heart failure (CHF; 73.1 +/- 8.7 years). Reproducibility was assessed by measuring intraclass correlation coefficients (ICC), coefficients of variation (CV) and Bland-Altman plots. 3. Baseline AIx showed good short-term reproducibility with high ICC in both the control and CHF groups (0.90 and 0.87, respectively). In contrast, in the control and CHF groups, the ICC of GTN- (0.58 and 0.17, respectively) and salbutamol-mediated (0.18 and 0.04, respectively) changes in AIx were substantially low. The CV was relatively low for baseline AIx in control and CHF groups (25.0 and 22.5%, respectively), but not for GTN- (22.3 and 59.8%, respectively) or salbutamol-mediated (45.1 and 184.0%, respectively) changes in AIx. Bland-Altman analysis revealed poor reproducibility, with limits of agreement beyond either +15% or -15% for changes in AIx after GTN and salbutamol for both control and CHF groups. The changes in blood pressure and heart rate following pharmacological challenge were similar between the two measurements. 4. The poor reproducibility of changes in AIx following pharmacological challenge questions the use of this method in acute studies.