In iontophoresis experiments, a 'non-specific' current-induced vasodilatation interferes with the effects of the diffused drugs. This current-induced vasodilatation is assumed to rely on an axon reflex due to excitation of cutaneous nociceptors and is weaker and delayed at the anode as compared to the cathode. We analysed whether these anodal specificities could result from a break excitation of nociceptors. Break excitation is the generation of action potentials at the end of a square anodal DC current application, which are generally weaker than those observed at the onset of a same application at the cathode. In eight healthy volunteers, we studied forearm cutaneous laser Doppler flow (LDF) responses to: (1) anodal and cathodal 100 microA current applications of 1, 2, 3, 4 or 5 min; (2) 100 microA anodal applications of 3 min with a progressive ending over 100 s (total charge 23 mC); these were compared to square-ended 100 microA anodal applications of the same total charge (23 mC) or duration (3 min); (3) a 4 min 100 microA anodal application with a 333 msec break at half time. Results (mean +/- S.D.) are expressed as percentage of heat-induced maximal vasodilatation (%MVD). Onset (T(vd)) and amplitude (LDF(peak)) of vasodilatation were determined. We observed that: T(vd) was linearly related to the duration of current application at the anode (slope = 1.01, r(2) = 0.99, P < 0.0001) but not at the cathode (slope = 0.03, r(2) = 0.02, n.s.). Progressive ending of anodal current did not decrease LDF(peak) (63.3 +/- 24.6 %MVD) as compared to square-ending of current application of the same duration (36.9 +/- 22.2 %MVD) or the same total charge (57.1 +/- 23.5 %MVD). A transient break of anodal current did not allow for the vasodilatation to develop until current was permanently stopped. We conclude that, during iontophoresis, anodal break excitation alone cannot account for the delay and amplitude of the vascular response.