6 Betty Croshaw Coates (18) has reviewed the microbiological requirements of a preservative. While screening tests in solid or liquid media for the evaluation of the antimicrobial activity of potential preservatives are valuable in selecting agents worthy of further study they will not necessarily predict the activity in a formulation. This is readily understandable as other factors are involved, particularly the o/w partition coefficient since any effective preservative must be present in the aqueous phase. Thus, a compound whose m.i.c. is 1000 [tg/ml may protect an emulsion at a concentration of 0-1-0.2• while another agent which inhibits growth at 20 [tg/ml in screening tests can fail to protect an emulsion even at a level of 0-45/o (21). On the other hand, Charles and Carter (22) found that the parabens, esters of p-hydroxybenzoic acid, were more effective in finished formulations than had been expected from their performance in component testing. Our own work with some formulations supports this a potential preservative must be further evaluated in different formulations at various concentrations to determine its efficacy under dif- ferent conditions. Bactericidal, rather than bacteriostatic, activity is an added advantage although in most cases a preservative may be considered to be effective if it holds small numbers of non-pathogenic organisms in a quiescent state. Even a low level of preservative may be slowly bactericidal and a higher level, resulting in a more rapid bactericidal effect, may be undesirable because of skin irritancy and, indeed, may not be necessary. The number of organisms challenging a preservative system must also be considered and the effect of inoculum size on the activity of any preservative should always be determined in the initial evaluation tests. In a formulated product an overwhelming microbial contamination may swamp the preservative system or, at best, may lengthen the time necessary to reduce the number of organisms to a safe, low level. Resistance studies should also be carried out at an early stage on any potential preservative. Any agent which rapidly selects out insensitive strains of microorganisms is unlikely to be a satisfactory preservative. Toxicological considerations The ideal preservative must be safe in use that is, it must be non-irritant, non-sensitizing and preferably non-poisonous. No chemical can ever be listed as ever being totally non-allergenic or totally non- sensitizing (23). Even the parabens, which have been and still are widely used for the preservation of toiletries and cosmetics, cannot be regarded as completely safe. Paraben allergy was recognized in Europe in the early 1960s and later in the United States when a sensitization index of 0.8• was documented (24). Sorbic acid, which was recommended as a preservative for formulations containing nonionic surface active agents (25) prior to the introduction of some of the newer pre- servatives, has a sensitization index probably slightly lower than that of the parabens (24). The phenolic group of preservatives also has its own problems (24). Two phenolics which are used as preservatives, chlorocresol (p-chloro-m-cresol) and especially chlor- oxylenol (p-chloro-m-xylenol), are known sensitizers and cross sensitization has been reported. The sensitization index for chlorocresol is 0-5•o and in a survey of contact allergy by Calnan (26) chloroxylenol was second only to mercury in terms of sensitivity produced. Formaldehyde, which has been widely used in shaml•oos (5, 7), constitutes a greater

Preservatives for cosmetics and toiletries 7 sensitization hazard than the parabens and the organic mercurials present the greatest hazard (24). Quaternary ammonium compounds are relatively weak sensitizers. The sensitization potential of some of the newer preservatives is not yet fully estab- lished from the clinical point of view. It has been suggested that 2-bromo-2-nitropropane- 1, 3-diol (Bronot•ol) may be used safely at the concentrations normally used in formulations (27). This compound was found to be a mild irritant when applied in soft yellow paraffin on to i49 eczematous patients at a concentration of 0'25% but no evidence of sensitization was seen in this study nor was there any suggestion of cross- sensitization with any other substance particularly formalin. Imidazolidinyl urea (Germall 115) is reported to be non-toxic, non-irritating and non-sensitizing (28) [rgasan DP 300 is said to be free of allergenic or photoallergenic potential on the basis of animal and human tests (29). There is some evidence that 6-acetoxy-2, 4-dimethyl-m- dioxane (Dioxin) is a sensitizer (24). Water solubility The ideal preservative should be readily water soluble at the effective concentration since microbial growth occurs in the aqueous phase. The low water solubility of the organic mercurials, Irgasan DP 300 and the parabens are marked disadvantages. Where water solubility is low, micellar solubilization of preservative molecules by compatible surfactants can be used to increase the amount of preservative in the system (18). Effect of oeH, temt•erature and storage times The pH tolerance limits for microorganisms is between pH 2-11 and since cosmetic and toiletry formulations can cover a wide pH range the ideal preservative should be effective and stable in solution over this range. It should also be stable at temperatures likely to be encountered during the manufacture and storage life of the product. It is well known that only the undissociated molecules of benzoic, dehydroacetic, salicylic and sorbic acids are active against microorganisms and that activity is lost with o , ß increasing pH. The parabens do not have the same pH dependence as benzoic acid but like phenols they show greater activity on the acid side of neutrality. Preservative dissociation as a function of pH has been reviewed by Wedderburn (3). The ideal preservative should be non-volatile at temperatures used during manufac- : ture as well as at normal storage temperatures. Volatility is a disadvantage of chloro- cresol, chloroxylenol and formaldehyde, for example. ß ::::i:: • Effect of oil/water t•artition coefficient .• Since microorganisms multiply in the aqueous phase of formulations the preservative . must be available in an effective concentration in this phase. Aqueous solutions such as :': shampoos are therefore relatively easy to preser•Je since all the preservative is available ::• providing that there is no chemical or physical incompatibility between it and the for- mulation or its container. The situation is obviously more complicated when the for- mulation is a cream or an emulsion as Bean and his colleagues (6, 30, 31) have pointed out the failure of preservatives in this situation may frequently be attributed to the fact that