The hydrochloride salts of 5-aminolevulinic acid (ALA) and its methyl ester (m-ALA), respectively, were dissolved in a lipid sponge phase comprising monoolein, propylene glycol and aqueous buffer at concentrations of approximately 0.25% and 16% w/w m-ALA. The iontophoretic and passive delivery of ALA and m-ALA from this formulation through porcine skin in vitro were measured and compared to formulations used in clinical practice, 20% w/w ALA in Unguentum M and Metvix (a cream containing 16% w/w m-ALA). A sponge phase with 16% w/w m-ALA showed a higher passive flux (approximately 140 nmol cm(-2) h(-1) at 5 h) but a lower iontophoretic flux (approximately 800 nmol cm(-2) h(-1) at 5 h) compared to the clinically used products but the differences are hardly significant due to large standard deviations. ALA and m-ALA in sponge phase formulation showed iontophoretic fluxes in the range 80-100 nmol cm(-2) h(-1) at 3 h, i.e. values comparable to the passive fluxes from the more concentrated vehicles. The results demonstrate that the lipid sponge phase, a thermodynamically stable liquid with amphiphilic character, may have potential as a transdermal drug delivery vehicle.