Since the advent of psoralen-ultraviolet A (PUVA) therapy, the value of plasma 8-methoxypsoralen (8-MOP) concentrations to predict PUVA-induced erythema has been widely investigated. Plasma 8-MOP concentrations have not been proportional to, and cannot alone predict, the degree of PUVA-induced erythema. We assumed that PUVA-induced erythema was related more closely to psoralen concentrations in the skin tissue rather than those within blood vessels. This study was designed to investigate the correlations between the 8-MOP concentrations in suction blister fluid (SBF) and in plasma, with the degree of PUVA-induced erythema. 8-MOP concentrations in plasma and SBF were measured in 15 vitiligo patients and 11 volunteers. Blood and SBF samples were collected 2 h after taking 8-MOP, and 8-MOP concentrations in plasma and SBF were quantified using reverse-phase high-performance liquid chromatography. Eleven volunteers were phototested using a series of doses of ultraviolet A at the time of sampling. The erythema responses were estimated visually to determine the minimal phototoxic dose (MPD). SBF 8-MOP concentrations showed a weak positive correlation with plasma 8-MOP concentrations, which means that we could not predict the exact SBF 8-MOP concentrations using the plasma 8-MOP concentrations. The MPD showed a better correlation with the log of the SBF 8-MOP concentration than with that of the plasma 8-MOP concentration. These results show that plasma 8-MOP concentration cannot represent the exact SBF 8-MOP concentration, and that SBF 8-MOP concentrations, which are representative of the skin tissue 8-MOP level, are more closely related to the erythemal sensitivity during PUVA therapy.