j. Cosmet. Sci., 54, 1-7 (January/February 2003) Evaluation of preservative systems in a sunscreen formula by the linear regression method , NADIA A. BOU-CHACRA, TEREZINHA DE JESUS A. PIN,TO, C and MITSUKO TABA OHARA, Department of Pharmacy, Faculty of Pharmaceutical Sdences, Sago Paulo University, Sago Paulo, SP, Brazil 66355. •'•i • • Accepted j•r publication, May 29, 2002. Poster presented at the XIII Congreso Latino-Americano e Ibgrico de Qu/micos Cosmgticos, Acapulco, Mexico, September 1997. Synopsis A sunscreen formula with eight different preservative systems was evaluated by liftear regression, pharma- copeial, and the CTFA (Cosmetic, Toiletry and Fragrance Association) methods. The preparations were tested against Staphylococcus aureus, Burkholderia cepacia, Shewandla putrefaciens, Escherichia coli, and Bacillus sp. The linear regression method proved to be useful in the selection of the most effective preservative system used in cosmetic formulation. INTRODUCTION Preservative efficacy testing is performed to determine the effectiveness of any added antimicrobial preservative in multiple-use cosmetic and pharmaceutical products. Meth- ods and specifications for the preservative efficacy test for pharmaceutical products are described in pharmacopeias (1-3), and the standardized test for cosmetics, which is similar to the pharmacopeial ones, was proposed by the Cosmetic, Toiletry and Fragrance Association (4). These methods consist of a challenge test that takes at least 28 days. The linear regression method was then proposed as a faster alternative (5-10). In spite of being the same as the technique proposed by the CTFA or the pharmacopeial meth- ods, it demands a shorter period of time to obtain results, through the calculation of D-values. Even though advantageous, this technique has been regarded as failing, as it uses the same principles of kinetical microbial killing, which leads to a first-order reaction, not necessarily followed by the preservative systems (11,12). Another vulner- able aspect of this method concerns the possibility of microbial growth during a longer length of time, impossible to be detected in the rapid method. Despite the criticism, the need to get quick results led to this work, which evaluated the linear regression method when applied to the selection of preservative systems for cosmetics.

2 JOURNAL OF COSMETIC SCIENCE MATERIALS AND METHODS MATERIALS The preservative systems of the nine evaluated formulas were as follows: 1. Absence of preservative. 2. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.3% (v/v) methylparaben 0.2% (w/v) propylparaben 0.1% (w/v) disodium ethylenediamine-tetraacetate (EDTA) 0.2% (w/v). 3. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.0% (v/v) methylparaben 0.2% (w/v) propylparaben 0.1% (w/v) triclosan 0.2% (w/v). 4. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.3% (v/v) isopropylparaben (and) isobutylparaben (and) butylparaben (Liqua par ©) 0.3% (v/v) disodium EDTA 0.2% (w/v). 5. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.3% (v/v) dichlorobenzyl alcohol 0.15 % (v/v) disodium EDTA 0.2% (w/v). 6. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.3% (v/v) methylparaben 0.2% (w/v) propylparaben 0.1% (w/v) dichlorobenzyl alcohol 0.15% (w/v) disodium EDTA 0.2% (w/v). 7. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.3% (v/v) isopropylparaben (and) isobutylparaben (and) butylparaben (Liqua par ©) 0.3% (v/v) dichlorobenzyl alcohol 0.15% (w/v) disodium EDTA 0.2% (w/v). 8. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.0% (v/v) methylparaben 0.2% (w/v) propylparaben 0.1% (w/v) benzoic acid 0.3% (w/v) disodium EDTA 0.2% (w/v). 9. Phenoxyethanol (and) methylparaben (and) ethylparaben (and) propylparaben (and) butylparaben (Phenonip ©) 1.0% (v/v) isopropylparaben (and) isobutylparaben (and) butylparaben (Liqua par ©) 0.3% (v/v) benzoic acid 0.3% (w/v) disodium EDTA 0.2 % (w/v). METHODS Preparation and standardization of the inoculum. The following microorganisms, some from the American Type Culture Collection (ATCC) and others isolated from the product, were used in the test: Burkholderia cepacia ATCC 17759, Escherichia coli ATCC 10536, Staphylococcus aureus ATCC 6538, Shewanella putrefaciens, and Bacillus sp. These strains were used to prepare suspensions, obtained from the harvest of microbial growth of a 24-hour-culture incubated at 30ø-35øC, on the surface of tryptic soy agar (TSA, Difco Laboratories, Detroit, MI). The number of colony-forming units per milliliter (CFU/ml) of suspension was deter- mined by the pour-plate count method using the above medium and incubation con- ditions. The concentration of microorganisms in the suspensions had to be suitable to provide from 10 5 to 10 6 microorganisms per milliliter in the test preparation imme- diately after the inoculation. The standardized suspensions were used to inoculate the samples.