Iontophoretic assessment of skin microvascular function is complicated by the occurrence of electrically induced hyperaemia, especially at the cathode. Studies were performed to identify means of reducing such effects. Skin vasodilator responses were measured using a laser Doppler imager that controlled iontophoretic current delivery. A novel feature involved monitoring voltage across the iontophoresis chambers. Comparison between responses to vehicle (distilled H(2)O), acetylcholine (ACh) and sodium nitroprusside (SNP) showed electrically induced hyperaemia at the cathode associated with the vehicle, whose time course overlapped with that of the SNP response. Voltage across the chambers containing drugs dissolved in H(2)O was significantly (p = 0.018, n = 7) lower than the voltage profile of H(2)O alone. H(2)O iontophoresis was associated with cathodal hyperaemic responses in most subjects, whereas a 0.5% NaCl vehicle produced lower voltages and eliminated this artefact. Voltage.time integral rather than charge was the prime determinant of electrically induced hyperaemic responses. No significant correlation was found between skin fold thickness and either calculated skin resistance (r(2) = 0.0002) or vascular response to ACh (r(2) = 0.13). Smaller chamber size led to higher voltages and greater electrically induced hyperaemic responses. These appear to be prostaglandin dependent as they were ablated by cyclooxygenase inhibition. Use of a low-resistance vehicle combined with larger chamber sizes and lower currents can prevent such artefacts, thereby increasing the robustness of this methodology for clinical assessment of endothelial function.