Facial skin flush is the most frequent adverse effect induced by acipimox (ACX), a nicotinic acid analogue used in the management of hyperlipidaemia. The aims of the study were to evaluate the frequency, magnitude and reproducibility of the ACX flush in previously unexposed healthy subjects and to assess any possible relationship with the dose and plasma level of ACX. Seventy four healthy subjects received, on two different mornings, ACX 250 mg and placebo (P), according to a single blind, randomized, cross-over design; 33 had a clear flush after ACX and not after P.25 of those subjects were retested on five different mornings, with P, and with ACX 31.2, 62.5, 125.0, 250.0 mg, according to a double blind, randomized, cross-over design. Any increase in the local skin temperature was recorded by a thermocouple fixed to the left check. Subjective and objective assessment of the flush were strongly correlated with thermographic recordings. They indicated that a 120 min flush occurred after doses of ACX greater than 62.5 mg. In 12 of the 25 subjects, 6 with the highest and 6 with the lowest thermographic recordings, plasma ACX levels were determined. Subjects with different thermographic records had superimposable plasma ACX levels after all doses of ACX. Only the 6 subjects with the highest skin temperatures showed a significant relationship between the thermographic record and the plasma ACX. The data indicate that flush is a frequent, reproducible and dose-related adverse effect of ACX.