J. Soc. Cosmet. Chem., 33, 75-84 (March/April 1982) The parabens: Bacterial adaptation and preservative capacity JOHNJ. O'NEILL and CATHERINE A. MEAD, Avon Products, Inc., Division St., Suffern, NY 10901. Received May 6, 1981. Presented at the SCC Annual Scientific Seminar, IVashington, D.C., May 21-22, 1981. Synopsis The initial rates of kill of Escherichia coli by methyl, ethyl and propyl parabens are similar at equal saturation fraction in a buffered salts-glucose medium. This is as expected from the Ferguson principle of equal activity of homologs, but the survival curves soon diverge strongly, apparently because the rate of adaptation increases markedly with increasing molecular weight E. coli adapts readily to propyl paraben, hardly at all to methyl. As a result, the parabens clearly rank methyl ethyl propyl butyl benzyl in their antimicrobial capacities at equal saturation fraction. The same ranking holds when equal weight concentrations are challenged with other bacteria in oil/water emulsions and in a shampoo base. Performance in these systems is discussed in terms of apparent solubilities. Mixtures of parabens are not superior except when multiple saturation is possible. In overall performance, only methyl paraben appears to be a good preservative. INTRODUCTION In two previous papers (1,2) we showed on theoretical grounds that the choice among methyl, ethyl and propyl parabens of a preservative for any cosmetic or pharmaceutical product could be based on solubility measurements. The paraben least soluble in the product should be the one most effective in both microbial kill rate and inoculum capacity in comparison with an equal weight of the more soluble homologs or with any equal weight combination of any two of the three parabens. This was an extension of W. P. Evans' argument (3) that as preservatives for simple emulsions, paraben mixtures can never be more effective than an equal weight of the less soluble one if the Ferguson principle (4) is applicable. According to the Ferguson principle, the members of many homologous series of biologically active materials are equally potent at their respective saturation concentrations or at any given fraction of saturation. Allawala and Riegelman (5) have shown that this principle applies quite uniformly to their measurements of the efficacy of a variety of phenols representing several distinct homologous series. If it applies to the parabens, then in water, propyl paraben at its saturation concentration of 0.3 grams per liter should be as potent as ethyl paraben at its solubility limit of one gram per liter or methyl paraben at two grams per liter. This prediction is in accord with numerous reports of minimum inhibitory concentrations (MIC's) in aqueous broth (see, for example, Reference 6) for a wide variety of bacteria 75

76 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS the MIC's increase in the order: butyl propyl ethyl methyl and the ratios, as would be predicted, are roughly the ratios of the solubilities. In cosmetic ingredients such as vegetable oils, the solubility order is the reverse of that in water methyl paraben is least soluble and therefore should be the most efficient preservative for oil-rich emulsions. Evans (3) showed that for simple oil/water mixtures the best preservative may be propyl paraben at low oil/water ratios or methyl paraben at high oil/water ratios but that methyl/propyl mixtures are less efficient in both cases. There seems to be a contradiction here between theory and practice. Parabens are almost always used in combinations in preserving cosmetics. A search of the literature, however, yielded no data unequivocally showing synergism in either aqueous broths or complex products. In our own experimental work we first attempted to demonstrate the applicability of the Ferguson principle to the parabens in simple, well-defined aqueous solutions as a step toward resolving the question of the utility of mixtures and also to support our theoretical proposal that single parabens be selected according to solubility. The earliest of these experiments (7) showed that the parabens do not follow the Ferguson principle to a useful extent at saturation their antimicrobial potencies are not equal. In fact, they drop sharply in the order: methyl ethyl propyl butyl (and benzyl paraben, not a member of the homologous series, is less potent yet). The ranking of the• parabens is evident from the way the survival curves of E. coli change as the inoculation level and saturation fraction are varied. At levels of 103 per ml or less the curves are roughly log-linear with about the same slope for methyl, ethyl and propyl parabens at saturation the bacterial population is extinguished in a day or two and no survivors are detected thereafter for as long as three weeks. The Ferguson principle is clearly applicable under these conditions. With methyl paraben at saturation the survival curve remains log-linear to extinction with the same slope, as the inoculation level is increased to over 107 per ml as its saturation fraction is decreased the rate of kill decreases but kill is persistent and appears to be complete in all cases until the saturation fraction is reduced to less than one-half, where the initial slope of the survival curve approaches zero. With propyl paraben the initial kill rate at saturation remains the same as the inoculation level is increased but at levels of about 105 per ml the survival curve becomes concave up within hours of inoculation and in some cases it passes through a deep minimum followed shortly by regrowth at about the same rate as in the unpreserved control. At still higher inoculation levels the minimum is shallow and occurs so early that the initial killing phase (if it occurs at all) is not detected and only a lag relative to the unpreserved control is noticed. The performance of ethyl paraben is intermediate but qualitatively more similar to that of propyl paraben: the transition from persistent kill to the kill-minimum-regrowth pattern occurs but it takes place at higher inoculation levels and lower saturation fractions than with propyl paraben. We found the same paraben ranking in experiments with Pseudomonas aeruginosa ATCC//0721 but with this organism the superiority of methyl paraben is much more striking at saturation it extinguishes inoculations as high as 107 per ml in less than one day while the ethyl and propyl esters cause only transient reductions in survivor counts at inoculations as low as 104 per mi.