- Email: [email protected]
Comparison of two in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics
Journal of Hospital Infection (2004) 58, 115–121
www.elsevierhealth.com/journals/jhin
Comparison of two in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics S. Messagera, P.A. Goddardb, P.W. Dettmarb, J.-Y. Maillardc,* a
Welsh School of Pharmacy, Cardiff University, King Edward VII Avenue, Cardiff CF10 3XF, Wales, UK Reckitt Benckiser Healthcare UK Ltd, Hull HU8 7DS, UK c School of Pharmacy and Biomolecular Sciences, University of Brighton, Cockcroft Building, Moulsecoomb, Brighton BN2 4GJ, UK b
Received 24 April 2003; accepted 12 December 2003 Available online 3 September 2004
KEYWORDS Antiseptics; Skin; Ex vivo; Antibacterial activity
Summary An ex vivo test was adapted to mimic the in vivo conditions of testing antiseptic activity on human forearms and in the European Standard Hygienic Handwash Test (BSEN 1499). The study was to validate the ex vivo protocols using 4.8% (w/v) para-chloro-meta-xylenol (PCMX, neat Dettol), 0.5% (w/v) triclosan in 70% (v/v) isopropanol, and 2% (v/v) povidone-iodine against a high bacterial inoculum (. 108 cfu/mL) of Escherichia coli NCTC 10538. Two ex vivo tests using human skin samples, including one introducing a mechanical rubbing effect, were compared with two corresponding in vivo tests (the forearm test and the BSEN handwashing test). All antiseptics assessed in vivo (forearm and handwash tests) produced reductions in bacterial counts that were significantly greater than those for the nonmedicated soft soap control. When assessed ex vivo without rubbing, only PCMX and povidone-iodine achieved reductions significantly greater than soft soap. When assessed ex vivo with mechanical rubbing, only PCMX and triclosan achieved reductions significantly greater than soft soap. Overall, the antiseptics at the concentrations tested were more active when tested in vivo than ex vivo. The addition of a mechanical effect, either in vivo by the volunteers washing their hands or ex vivo by a drill rubbing two skin samples against each other, produced a significantly greater reduction in bacterial concentrations. The ex vivo tests were easily adapted to mimic in vivo protocols. The value of such tests, particularly the one that includes a rubbing effect, may be significant as they avoid the need for human volunteers. Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
*Corresponding author. Tel.: þ44-1273-642105; fax: þ 441273-642674. E-mail address: [email protected] 0195-6701/$ - see front matter Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2004.06.001
116
Introduction Contamination of the hands of healthcare personnel with antibiotic-resistant organisms is known to play a role in the epidemiology of nosocomial infections.1,2 Handwashing is generally accepted as the most effective action to prevent the spread of pathogens.3 Many skin cleansing agents are commercially available. However, there are limited ways to study the efficacy of these agents on skin.1 A variety of in vivo tests have been reported.4 – 6 Protocols not only vary in what is sampled, but also in sampling methods, which are mainly variations on the glove juice/scrub rinse technique and impression plate/hand stamping procedures. 7 Moreover, handwashing tests are limited for various reasons. Nosocomial and other pathogens should not be tested on human volunteers for ethical reasons. Handwashing tests reflect the total reduction of organisms from the skin, as opposed to transferable organisms, and they may reflect handwashing technique rather than product efficacy.1 Different populations of volunteers may produce different mean bacterial reductions with the same antiseptic preparation, even using a wellstandardized test method.8 As reported in previous work, in vivo studies with volunteers can show discrepancies in results.9,10 To address this problem, an ex vivo method, previously developed to allow the testing of antiseptics directly on human skin11 was adapted to mimic the in vivo conditions of the European Standard Handwash Method (BSEN 14994) and a test on forearms of volunteers. The aim of this study was to validate the ex vivo protocols with high antiseptic and bacterial concentrations.
Materials and methods The following antiseptics were tested: aqueous 2% (v/v) povidone-iodine (Sigma, Poole, UK), aqueous 4.8% w/v para-chloro-meta-xylenol (PCMX; Dettol, Reckitt Benckiser, Hull, UK), and 0.5% (w/v) triclosan (Ciba Specialty Chemicals, GrenzachWyhlen, Germany) in 70% (v/v) isopropanol. Solutions were freshly made using sterile ultrapure water (Nanopure, Barnstead, UK) when appropriate. Soft soap made of linseed oil (100 g/L; Sigma, Poole, UK), potassium hydroxide (19 g/L; Sigma, Poole, UK), ethanol (14 g/L; Fisher, Loughborough, UK) and as much distilled water as required (BSEN 14994) was used as a reference agent. Neutralizing solutions based on the BSEN 14994 were freshly made in sterile ultrapure water (Nanopure, Barnstead, UK)
S. Messager et al.
and are shown in Table I. Previous experiments validated the efficacy of these neutralizers to quench the activity of their respective antiseptics, and showed that they had no toxic effects on the test strain of Escherichia coli.11 The in vivo tests were carried out with 12 female and three male volunteers (aged 18 – 25 years). Participants had short fingernails and intact skin and were instructed not to apply antibacterial toiletries onto their forearms for 24 h before testing. The volunteers were thoroughly informed about the experimental details, aims and objectives, and signed a consent form. Ethical approval was received for the studies.
In vivo forearm test An E. coli K12 (NCTC 10538) inoculum was prepared by re-suspending a single colony from a tryptone soya agar (TSA, Oxoid, Basingstoke, UK) plate into 10 mL tryptone soya broth (TSB, Oxoid, Basingstoke, UK). After overnight incubation at 37.1 ^ 0.4 8C in a shaking water-bath (Gallenkamp, Loughborough, UK; 90 rpm), the inoculating suspension contained approximately 1.17 £ 109 cfu/mL. The in vivo forearm protocol was based on a method described by Payne and Eddowes6 for testing antiseptics. Briefly, five circles of 36 mm diameter were marked on the back of the forearms using plastic circles fixed with sticky tape (three on one arm and two on the other). Each circle was identified by its position (left or right arm) and test. Preparations were assigned to sites using a Latin square arrangement. Twenty microlitres of bacterial inoculum was spread inside each circle using a sterile glass rod and allowed to dry for 3 min at room temperature. Thirty microlitres of antiseptic or soft soap was spread over the appropriate circle with a sterile glass rod. One circle was left untouched as a control. After 1 min, 950 mL of the appropriate neutralizers were added and mixed on the test circle surfaces, and 980 mL of TSB was added to the control circle. The resultant fluid was recovered with an Eppendorf pipette from each circle. The number of surviving micro-organisms was counted using a drop counting method (DCM).12 Briefly, a 0.1 mL aliquot was serially diluted in phosphate-buffered saline (PBS; Sigma, Poole, UK) and 10 mL drops were plated onto the surface of a dried TSA plate. After overnight incubation at 37 8C, colonies were counted and the viable count was expressed as colony-forming units per millilitre.
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 117
Table I Neutralizers used (BSEN 1499) Test agent
Polysorbate 80 (Sigma)
Egg lecithin (Sigma)
Sodium thiosulphate (Fisher)
Bovine albumen (Sigma)
L -histidine (Sigma)
30 30 30 150
3 3 3 15
5 5 5 25
1 – – –
1 1 1 5
Povidone-iodine Triclosan in isopropanol Soft soap PCMX
Figures are in grams per litre dissolved in tryptone soya broth.
In vivo handwashing test An overnight E. coli K12 (NCTC 10538) culture was inoculated into 1 L TSB and incubated for 18 – 24 h at 37.1 ^ 0.4 8C in a shaking water-bath. This contaminating fluid contained 1.33 – 1.67 £ 109 cfu/mL. The in vivo handwashing test followed the European Standard BSEN 14994 Hygienic Handwash recommendations. In this test, the number of test organisms released from artificially contaminated hands is assessed before and after hygienic handwashing. Therefore, the number of colony-forming units sampled from the skin before treatment (i.e. pre-value) and the number of colony-forming units sampled from the skin after treatment (i.e. postvalue) were determined for all the products tested. The order of testing was determined by a Latin square design. The bacterial reduction factors (RFs) produced by the procedures were calculated as follows: Log10 RF ¼ log10 pre-value 2 log10 post-value where RF is a measure of the antimicrobial activity of the disinfectant tested. After performing the handwashing test, volunteers were asked to wash their hands for 2 min with a Hibiscrubw solution (Zeneca, Oslo, Norway).
microlitres of antiseptic or soft soap was then spread over the skin with a sterile glass rod. After 1 min, 950 mL of appropriate neutralizer was added and the surviving bacteria were re-suspended and counted. As a control, a 20 mL bacterial sample was inoculated on skin, dried for 3 min and re-suspended in 980 mL TSB. Surviving bacteria were enumerated using the DCM.
Ex vivo protocol to mimic the in vivo handwashing test Fresh skin samples were cut into 2 £ 2 cm pieces with sterile scissors. The pieces were glued (superglue; Niceday, London, UK) on to individual plastic cap holders (4 –5 cm diameter Petri dishes, Fisher, Loughborough, UK) with the skin surface exposed (Figure 1). Two pieces of mounted skin were needed for each sample being tested. Ten millilitres of an overnight E. coli K12 (NCTC 10538) inoculum containing approximately 7.5 £ 108 cfu/mL was poured into a sterile Petri dish, in which both skin samples were immersed for 5 s. They were then removed and air-dried for 3 min. Immediately after drying, both samples were rubbed for 1 min on the base of a Petri dish containing 5 mL TSB without neutralizer, to assess the release of test organism before antiseptic treatment (i.e. pre-value). Serial
Ex vivo protocol to mimic the in vivo forearm test This ex vivo protocol was based on the method of Messager et al.11 Fresh samples of skin were obtained with patients’ informed consent after plastic surgery (breast and abdominal reduction) and were cut to fit in diffusion cells (JB and DW Jones, Loughborough, UK) especially designed for that experiment with a donor compartment of 36 mm diameter.11 The receptor compartment was filled with Earle’s balanced salt solution (EBSS, Sigma, Poole, UK). A 20 mL aliquot of a E. coli K12 (NCTC 10538) inoculum containing approx. 8 £ 108 cfu/mL was spread over the skin with a sterile glass rod and the suspensions were allowed to dry for 3 min at room temperature. Thirty
Figure 1 Mounted human skin: top (on the left) and bottom (on the right) pieces of skin. The plastic cap holder, onto which the top piece of skin is mounted, has been previously glued onto a steel rod that will be attached to the drill.
118
dilutions of the sampling fluid were prepared in sterile TSB and the number of bacteria released was counted. Immediately after the pre-value sampling and without recontamination of the samples, they were pre-moistened by a brief immersion in sterile water in a Petri dish. Surplus water was allowed to drain back into the dish. The plastic cap on to which the bottom skin sample had been glued was itself mounted on a polystyrene holder on top of a balance to measure the pressure exerted on skin (Figure 2). A 1 mL aliquot of either soft soap or test antiseptic was then applied to the surface of the bottom skin sample. The plastic cap (top piece of skin) was fixed to a steel rod to allow mechanical rotation with a drill (IKA, Labortechnik, Staufen, Germany). Both pieces of skin were put in contact with, the top sample exerting a weight of approximately 100 g on the surface of the bottom sample. The top skin sample was then rotated against the bottom sample for 1 min at 200 rpm (Figure 2). After rinsing the skin samples thoroughly in a Petri dish containing 5 mL of sterile water for 5 s (one Petri dish per sample), they were rubbed on the
Figure 2 Picture showing the ex vivo protocol used. Bottom piece of skin, previously mounted on a plastic holder, was fixed on a weighing machine. The top piece, previously mounted on a plastic holder, itself glued to a steel rod, was pulled down on to the bottom piece to exert a weight of approximately 100 g and the two pieces were rubbed against each other for 1 min at the minimum speed of the drill used: 200 rpm.
S. Messager et al.
base of a Petri dish containing 5 mL of the appropriate neutralizer for 1 min (one Petri dish per sample). The sampling fluids were diluted in neutralizer and the surviving bacteria (i.e. postvalue) were counted. Statistical analyses were performed using Minitabw software and analysis of variance was conducted at the 95% confidence intervals.
Results In vivo and ex vivo testing of antiseptics without rubbing When the antibacterial activity of antiseptics was assessed in vivo on volunteers, forearms, the number of bacteria recovered from the control (untreated) circles was marginally lower (,tenfold reduction in number; P , 0:05) than that of the original bacterial inoculum (Figure 3). The number of bacteria recovered from the treated circles (soft soap and antiseptics) was significantly lower ðP , 0:05Þ than that recovered from the control circles. The concentration of bacteria recovered from the antiseptic treated samples was also significantly lower ðP , 0:05Þ than that recovered from the softsoap-treated samples (Figure 3). Point five percent (w/v) triclosan in 70% (v/v) isopropanol and 2% (v/v) povidone-iodine achieved the highest RFs, followed by 4.8% (w/v) PCMX and finally the soft soap. The antibacterial activity of povidone-iodine was not significantly different ðP . 0:05Þ to that of triclosan in isopropanol (Table II). PCMX was significantly less active ðP , 0:05Þ than triclosan in isopropanol and povidone-iodine (Table II).
Figure 3 Log10 counts of E. coli (cfu/mL) recovered from the forearms of volunteers, depending on the test performed: antiseptic or soft soap compared with the untreated sites (control) and with the original bacterial inoculum (mean values, bars represent 1 standard deviation).
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 119
Table II Results obtained in vivo and ex vivo: log10 reduction in Escherichia coli K12 concentration (cfu/mL) after soft soap or antiseptic treatment for 1 min depending on the test carried out Mean log10 reduction in bacterial concentration (standard deviation)
Test without rubbing effect In vivo forearms Ex vivo Test with a rubbing effect In vivo handwashing Ex vivo
Soft soap
Povidone-iodine
Triclosan in isopropanol
PCMX
0.41 (0.42) 0.06 (0.22)
2.73 (0.93) 0.76 (0.51)
3.45 (1.38) 0.47 (0.58)
1.80 (0.92) 0.49 (0.40)
2.24 (0.521) 1.98 (0.36)
3.41 (0.83) 1.49 (1.29)
3.79 (0.89) 2.46 (0.66)
4.41 (1.26) 2.89 (0.37)
In the corresponding ex vivo test, the bacteria count from the control samples was marginally lower (, tenfold reduction in number; P , 0:05) than that of the original inoculum (Figure 4). There was no difference ðP . 0:05Þ in bacterial counts between treatment with soft soap and the control. The number of bacteria recovered after treatment with antiseptics was marginally lower (although statistically significant; P , 0:05) than that recovered from the control samples (Figure 4). When the results of both in vivo and ex vivo tests without a rubbing effect were compared, the reduction in E. coli achieved after treatment with soft soap was not significantly different (P . 0.05) whether it was performed with the ex vivo test or with the in vivo forearm test (Table II). Overall, the reduction in bacterial concentration after exposure to antiseptics was lower ðP , 0:05Þ when tested with the ex vivo test, regardless of the treatment. Finally, the variability in results obtained with the ex vivo protocol was much lower than that of the in vivo forearm test (Table II).
Figure 4 Log10 counts of E. coli recovered from skin samples treated with soft soap or antiseptics compared with that of the control (untreated) samples and with the original inoculum (mean values, bars represent 1 standard deviation).
In vivo and ex vivo testing of antiseptics with a rubbing effect Counts from the BSEN in vivo handwashing test required validation. Requirements for acceptance were achieved: valid results from all 15 participants were available and the overall mean of the log10 pre-values for reference and test procedures was at least 5 (data not shown). In addition, the number of bacteria recovered from the left or the right hands of each volunteer was not significantly different (P . 0.05) in any test (data not shown). The mean RFs for each antiseptic were all significantly greater than the RF of soft soap (Table II). The antibacterial activity of 0.5% (w/v) triclosan in 70% (v/v) isopropanol was not significantly different (P . 0.05) to that of 2% (v/v) povidone-iodine or 4.8% (w/v) PCMX (Table II). However, PCMX was significantly more active ðP , 0:05Þ than povidone-iodine. Overall, the antimicrobial activity of the antiseptics tested against E. coli was greater than that of soft soap (Table II). When assessed with the corresponding ex vivo test with a rubbing effect, there was no significant difference ðP . 0:05Þ between the concentration of E. coli recovered from the top skin pieces and the concentration of bacteria recovered from the bottom pieces of skin, including both the pre- and post-treatment values (data not shown). Point five percent (w/v) triclosan in 70% (v/v) isopropanol and 4.8% (w/v) PCMX achieved comparable ðP . 0:05Þ Log10 RFs (2.46 ^ 0.66 and 2.89 ^ 0.37, respectively), which were significantly higher ðP , 0:05Þ than that from soft soap (Table II). The log10 RF achieved by 2% (v/v) povidone-iodine was not significantly ðP . 0:05Þ different from soft soap (1.49 ^ 1.29 log10 and 1.98 ^ 0.36, respectively; Table II). When the results of both protocols with rubbing were compared, the antibacterial activity achieved by the soft soap and povidone-iodine was not significantly dependent ðP . 0:05Þ upon the test
120
performed. However, the number of bacteria recovered after exposure to PCMX or triclosan in isopropanol was significantly greater ðP , 0:05Þ when assessed ex vivo (Table II). Furthermore, with the exception of povidone – iodine, variability in results was lower with the ex vivo protocol (Table II).
Discussion Antiseptics can be assessed directly on human skin. However, in vivo studies are limited by several factors including ethical limitations, variability in results, and cost. For many years, there has been a search for internationally acceptable tests for hand disinfecting agents that could be performed by manufacturers and research laboratories alike.1,13,14 The objective of the present study was to compare two novel ex vivo tests with their reciprocal in vivo protocols. The bacterial inoculum dried on the skin surface was inactivated to a lesser extent when assessed ex vivo than when tested in vivo. In essence, this might be an advantage as it makes the ex vivo test more stringent. Differences in inactivation observed between the ex vivo and in vivo protocols might be explained partially by the ‘self-sterilizing’ effect of the skin of volunteers,15,16 due to sebaceous and sweat gland secretions.17 Rubbing was important in reducing the bacterial concentration on the skin. With soft soap, it was clear that this rubbing effect had a significant ðP , 0:05Þ outcome in the reduction of viable bacteria recovered from the skin surface (Table II). Overall, in terms of the rubbing effect, results from the ex vivo protocols correlated well with their respective in vivo protocols, both giving an increased RF (Table II). The physical removal of bacteria from the skin as a result of volunteers rubbing their hands (or fingertips) has been attributed to both the inherent surfactancy of the product used and the mechanical action of rubbing, i.e. breaking up the bacterial colonies on the skin.18 The reduction of bacterial concentration, often referred to as the ‘killing effect’, by exposure to an antiseptic on skin in vivo is therefore due to a combination of physical removal and antimicrobial activity. In addition, in both protocols with a rubbing effect, skin samples or hands were immersed into the bacterial inoculum as opposed to the bacterial inoculum being spread on to the surface in the forearm in vivo and the corresponding ex vivo tests. Spreading the micro-organisms on to the skin surface might also account for the lower
S. Messager et al.
antiseptic activity observed. Some studies showed that rubbed-in organisms are more difficult to kill than dried-on micro-organisms.19,20 The results from our in vivo studies confirmed results from the literature, whereby exposure to an antiseptic reduced the number of bacteria recovered from the skin surface significantly.21 – 25 Further comparison of activity was difficult to ascertain because of differences in methodology; especially products (active ingredient and concentration) used, or lack of information, especially with the forearm protocol. However, despite the use of different protocols, results from the literature on handwashing tests and fingertip experiments provide valuable information. For example, when the bactericidal activity of different antiseptics was tested in vivo, povidone – iodine (1% available iodine) was more effective than 3% chloroxylenol in reducing the overall skin bacterial flora.21 Another study showed that 7.5% povidoneiodine achieved the same reduction of bacterial skin flora concentration after a 30 s wash.22 Munton and Prince23 showed that 5% Dettol achieved a log10 reduction of . 4.95 in Staphylococcus aureus, Enterococcus faecalis, E. coli and Pseudomonas aeruginosa within 5 min in an in vivo investigation using artificially contaminated skin. The low activity of PCMX has been reported elsewhere, although in this case, a low concentration of the phenolic was used.24 When assessed in vivo on hands, 0.5% triclosan in 70% isopropanol (Manuseptw) was significantly more active against E. coli than 60% isopropanol alone after 1 min of treatment.25 These findings correlated with our results that showed that triclosan in isopropanol achieved a log10 RF of 3.45 ^ 1.38 after 1 min without the rubbing effect. The three antiseptics tested here were significantly more effective ðP , 0:05Þ at removing E. coli from the hands than the soft soap. The mean log10 reduction obtained in this study was comparable with that found in an in vivo study using participants’ fingertips.26 Overall, in the handwashing study, the three antiseptics achieved a log RF of . 3 (PCMX achieving . 4 log10). However, the corresponding ex vivo test failed to achieve this level of bacterial reduction. Ayliffe et al.27 reported that after heavy contamination, a bacterial agent should achieve log10 RFs of 3 – 4. Finally, the ex vivo protocols performed in our study showed that, overall, a better reproducibility in results was obtained when compared with the in vivo tests. We believe that an even greater reproducibility would have been obtained if a higher number of replicates was performed. Unfortunately, skin supply limitation (especially large samples) limited us three to four replicates.
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 121
In conclusion, the ex vivo protocols described in this study, and particularly the ex vivo test with a rubbing effect, were easily adapted to mimic in vivo protocols and offered the potential to assess antiseptics’ efficacy directly on human skin. However, the ex vivo protocols appear to be more stringent than in vivo protocols since bacterial inactivation achieved ex vivo was lower than that achieved in vivo. This might be considered as an added value for the testing of bactericidal activity of new formulations. The added value of such ex vivo testing models for handwashing and other in vivo protocols is their ability: (1) to simulate a wide variety of skin contamination (e.g. level of microbial contaminants and pathogenicity of micro-organisms); and (2) to control parameters inherent to treatment conditions effectively. In particular, the addition of a mechanical rubbing effect is significant as it imitates the rubbing of both hands against each other during handwashing procedures. In our ex vivo protocol, the rubbing effect can be fully controlled and parameters (i.e. speed of rotation and pressure) can be adapted to correspond to different handwashing habits. Considerations of health, safety and expenses preclude conducting the type of testing presented here on humans. The methods used are easy and inexpensive to perform, and pathogenic micro-organisms, as well as new formulations, can now be assessed effectively on skin. The only caveat is to ensure that informed consent is obtained from the tissue donors for the experiments. This study demonstrated the usefulness and applicability of new testing procedures that could be further standardized for use as reference tests for the preliminary testing of antiseptic formulations on skin.
References 1. McDonnell G, Haines K, Klein D, Rippon M, Walmsley R, Pretzer D. Clinical correlation of a skin antisepsis model. J Microbiol Methods 1999;35:31—35. 2. Pittet D, Boyce JM. Hand hygiene and patient care: pursuing the Semmelweis legacy. Lancet Infect Dis 2001;April:9—20. 3. Larson E. Hygiene of the skin: when is clean too clean? Emerg Infect Dis 2001;7:225—230. 4. BSEN 1499 (European Standard). Chemical disinfectants and antiseptics—hygienic handwash. Test method and requirement (phase 2/step 2). European Standard EN, 1499; 1997. 5. Ayliffe GAJ, Babb JR, Quoraishi AH. A test for hygienic hand disinfection. J Clin Pathol 1978;31:923—928. 6. Payne DN, Eddowes S. An in vivo method for assessing the antimicrobial properties of topical antiseptics using artificially contaminated skin. Lett Appl Microbiol 1990;10: 101—104. 7. Larson E, Rotter ML. Handwashing: are experimental models a substitute for clinical trials? Two viewpoints. Infect Control Hosp Epidemiol 1990;11:63—66.
8. Rotter ML. Hygienic hand disinfection. Infect Control Hosp Epidemiol 1984;5:18—22. 9. Nagai I, Ogase H, Takechi M, Kadota M, Kumamoto R. Evaluation of the disinfectant effect of a quick drying rubbing type povidone-iodine alcoholic solution by the glove juice method. Postgrad Med J 1993;69:S33—S38. 10. Rotter ML, Simpson RA, Koller W. Surgical hand disinfection with alcohols at various concentrations: parallel experiments using the new proposed European standard method. Infect Control Hosp Epidemiol 1998;19:778—781. 11. Messager S, Goddard PA, Dettmar PW, Maillard J-Y. Antibacterial efficacy of several antiseptics tested on skin using the ex vivo test. J Med Microbiol 2001;50:284—292. 12. Maillard J-Y, Messager S, Veillon R. Antimicrobial efficacy of biocides tested on skin using an ex vivo test. J Hosp Infect 1998;40:313—323. 13. Ayliffe GAJ. Standardization of disinfectant testing. J Hosp Infect 1989;13:211—216. 14. Sattar SA, Springthorpe VS. New methods for efficacy testing of disinfectants and antiseptics. In: Rutala A, editor. Disinfection, sterilization and antisepsis: principles and practices in healthcare facilities. Washington, DC: Association for Professionals in Infection Control and Epidemiology, Inc. Publishers; 2000. p. 73—186. 15. Chikakane K, Takahashi H. Measurement of skin pH and its significance in cutaneous diseases. Clin Dermatol 1995;13: 299—306. 16. Cove JH, Eady EA. Cutaneous antimicrobial defense. Clin Dermatol 1998;16:141—147. 17. Selwyn S, Ellis H. Skin bacteria and skin disinfection reconsidered. BMJ 1972;1:136—140. 18. Baquero F, Patron C, Canton R, Martinez Ferrer M. Laboratory and in vitro testing of skin antiseptics: a prediction for in vivo activity? J Hosp Infect 1991;18:5—11. 19. Lilly HA, Lowbury EJL. Transient skin flora—their removal by cleansing or disinfection in relation to their mode of deposition. J Clin Pathol 1978;31:919—922. 20. Bush LW, Benson LM, White JH. Pig skin as test substrate for evaluating topical antimicrobial activity. J Clin Microbiol 1986;24:343—348. 21. Aly R, Maibach HI. Comparative antibacterial efficacy of a 2min surgical scrub with chlorhexidine gluconate, povidoneiodine, and chloroxylenol sponge brushes. Am J Infect Control 1988;16:173—177. 22. Leyden JJ, McGinley KJ, Kaminer MS, et al. Computerized image analysis of full-hand touch plates: a method for quantification of surface bacteria on hands and the effect of antimicrobial agents. J Hosp Infect 1991;18:13—22. 23. Munton TJ, Prince J. The bactericidal activity of Dettol on skin artificially contaminated with micro-organisms using the replica plating technique. Reckitt & Colman SORL BL 75/12 (technical report); 1975. 24. Soulsby ME, Barnett JB, Maddox S. Brief report: the antiseptic efficacy of chloroxylenol-containing vs. chlorhexidine gluconate-containing surgical scrub preparations. Infect Control 1986;7:223—226. 25. Bartzokas CA, Gibson MF, Graham R, Pinder DC. A comparison of triclosan and chlorhexidine with 60 per cent isopropyl alcohol for hygienic hand disinfection. J Hosp Infect 1983;4: 245—255. 26. Rotter ML, Koller W. Test models for hygienic handrub and hygienic handwash: the effects of two different contamination and sampling techniques. J Hosp Infect 1992;20: 163—171. 27. Ayliffe GAJ, Babb JR, Davies JG, Lilly HA. Hand disinfection: a comparison of various agents in laboratory and ward studies. J Hosp Infect 1988;11:226—243.
www.elsevierhealth.com/journals/jhin
Comparison of two in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics S. Messagera, P.A. Goddardb, P.W. Dettmarb, J.-Y. Maillardc,* a
Welsh School of Pharmacy, Cardiff University, King Edward VII Avenue, Cardiff CF10 3XF, Wales, UK Reckitt Benckiser Healthcare UK Ltd, Hull HU8 7DS, UK c School of Pharmacy and Biomolecular Sciences, University of Brighton, Cockcroft Building, Moulsecoomb, Brighton BN2 4GJ, UK b
Received 24 April 2003; accepted 12 December 2003 Available online 3 September 2004
KEYWORDS Antiseptics; Skin; Ex vivo; Antibacterial activity
Summary An ex vivo test was adapted to mimic the in vivo conditions of testing antiseptic activity on human forearms and in the European Standard Hygienic Handwash Test (BSEN 1499). The study was to validate the ex vivo protocols using 4.8% (w/v) para-chloro-meta-xylenol (PCMX, neat Dettol), 0.5% (w/v) triclosan in 70% (v/v) isopropanol, and 2% (v/v) povidone-iodine against a high bacterial inoculum (. 108 cfu/mL) of Escherichia coli NCTC 10538. Two ex vivo tests using human skin samples, including one introducing a mechanical rubbing effect, were compared with two corresponding in vivo tests (the forearm test and the BSEN handwashing test). All antiseptics assessed in vivo (forearm and handwash tests) produced reductions in bacterial counts that were significantly greater than those for the nonmedicated soft soap control. When assessed ex vivo without rubbing, only PCMX and povidone-iodine achieved reductions significantly greater than soft soap. When assessed ex vivo with mechanical rubbing, only PCMX and triclosan achieved reductions significantly greater than soft soap. Overall, the antiseptics at the concentrations tested were more active when tested in vivo than ex vivo. The addition of a mechanical effect, either in vivo by the volunteers washing their hands or ex vivo by a drill rubbing two skin samples against each other, produced a significantly greater reduction in bacterial concentrations. The ex vivo tests were easily adapted to mimic in vivo protocols. The value of such tests, particularly the one that includes a rubbing effect, may be significant as they avoid the need for human volunteers. Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
*Corresponding author. Tel.: þ44-1273-642105; fax: þ 441273-642674. E-mail address: [email protected] 0195-6701/$ - see front matter Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2004.06.001
116
Introduction Contamination of the hands of healthcare personnel with antibiotic-resistant organisms is known to play a role in the epidemiology of nosocomial infections.1,2 Handwashing is generally accepted as the most effective action to prevent the spread of pathogens.3 Many skin cleansing agents are commercially available. However, there are limited ways to study the efficacy of these agents on skin.1 A variety of in vivo tests have been reported.4 – 6 Protocols not only vary in what is sampled, but also in sampling methods, which are mainly variations on the glove juice/scrub rinse technique and impression plate/hand stamping procedures. 7 Moreover, handwashing tests are limited for various reasons. Nosocomial and other pathogens should not be tested on human volunteers for ethical reasons. Handwashing tests reflect the total reduction of organisms from the skin, as opposed to transferable organisms, and they may reflect handwashing technique rather than product efficacy.1 Different populations of volunteers may produce different mean bacterial reductions with the same antiseptic preparation, even using a wellstandardized test method.8 As reported in previous work, in vivo studies with volunteers can show discrepancies in results.9,10 To address this problem, an ex vivo method, previously developed to allow the testing of antiseptics directly on human skin11 was adapted to mimic the in vivo conditions of the European Standard Handwash Method (BSEN 14994) and a test on forearms of volunteers. The aim of this study was to validate the ex vivo protocols with high antiseptic and bacterial concentrations.
Materials and methods The following antiseptics were tested: aqueous 2% (v/v) povidone-iodine (Sigma, Poole, UK), aqueous 4.8% w/v para-chloro-meta-xylenol (PCMX; Dettol, Reckitt Benckiser, Hull, UK), and 0.5% (w/v) triclosan (Ciba Specialty Chemicals, GrenzachWyhlen, Germany) in 70% (v/v) isopropanol. Solutions were freshly made using sterile ultrapure water (Nanopure, Barnstead, UK) when appropriate. Soft soap made of linseed oil (100 g/L; Sigma, Poole, UK), potassium hydroxide (19 g/L; Sigma, Poole, UK), ethanol (14 g/L; Fisher, Loughborough, UK) and as much distilled water as required (BSEN 14994) was used as a reference agent. Neutralizing solutions based on the BSEN 14994 were freshly made in sterile ultrapure water (Nanopure, Barnstead, UK)
S. Messager et al.
and are shown in Table I. Previous experiments validated the efficacy of these neutralizers to quench the activity of their respective antiseptics, and showed that they had no toxic effects on the test strain of Escherichia coli.11 The in vivo tests were carried out with 12 female and three male volunteers (aged 18 – 25 years). Participants had short fingernails and intact skin and were instructed not to apply antibacterial toiletries onto their forearms for 24 h before testing. The volunteers were thoroughly informed about the experimental details, aims and objectives, and signed a consent form. Ethical approval was received for the studies.
In vivo forearm test An E. coli K12 (NCTC 10538) inoculum was prepared by re-suspending a single colony from a tryptone soya agar (TSA, Oxoid, Basingstoke, UK) plate into 10 mL tryptone soya broth (TSB, Oxoid, Basingstoke, UK). After overnight incubation at 37.1 ^ 0.4 8C in a shaking water-bath (Gallenkamp, Loughborough, UK; 90 rpm), the inoculating suspension contained approximately 1.17 £ 109 cfu/mL. The in vivo forearm protocol was based on a method described by Payne and Eddowes6 for testing antiseptics. Briefly, five circles of 36 mm diameter were marked on the back of the forearms using plastic circles fixed with sticky tape (three on one arm and two on the other). Each circle was identified by its position (left or right arm) and test. Preparations were assigned to sites using a Latin square arrangement. Twenty microlitres of bacterial inoculum was spread inside each circle using a sterile glass rod and allowed to dry for 3 min at room temperature. Thirty microlitres of antiseptic or soft soap was spread over the appropriate circle with a sterile glass rod. One circle was left untouched as a control. After 1 min, 950 mL of the appropriate neutralizers were added and mixed on the test circle surfaces, and 980 mL of TSB was added to the control circle. The resultant fluid was recovered with an Eppendorf pipette from each circle. The number of surviving micro-organisms was counted using a drop counting method (DCM).12 Briefly, a 0.1 mL aliquot was serially diluted in phosphate-buffered saline (PBS; Sigma, Poole, UK) and 10 mL drops were plated onto the surface of a dried TSA plate. After overnight incubation at 37 8C, colonies were counted and the viable count was expressed as colony-forming units per millilitre.
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 117
Table I Neutralizers used (BSEN 1499) Test agent
Polysorbate 80 (Sigma)
Egg lecithin (Sigma)
Sodium thiosulphate (Fisher)
Bovine albumen (Sigma)
L -histidine (Sigma)
30 30 30 150
3 3 3 15
5 5 5 25
1 – – –
1 1 1 5
Povidone-iodine Triclosan in isopropanol Soft soap PCMX
Figures are in grams per litre dissolved in tryptone soya broth.
In vivo handwashing test An overnight E. coli K12 (NCTC 10538) culture was inoculated into 1 L TSB and incubated for 18 – 24 h at 37.1 ^ 0.4 8C in a shaking water-bath. This contaminating fluid contained 1.33 – 1.67 £ 109 cfu/mL. The in vivo handwashing test followed the European Standard BSEN 14994 Hygienic Handwash recommendations. In this test, the number of test organisms released from artificially contaminated hands is assessed before and after hygienic handwashing. Therefore, the number of colony-forming units sampled from the skin before treatment (i.e. pre-value) and the number of colony-forming units sampled from the skin after treatment (i.e. postvalue) were determined for all the products tested. The order of testing was determined by a Latin square design. The bacterial reduction factors (RFs) produced by the procedures were calculated as follows: Log10 RF ¼ log10 pre-value 2 log10 post-value where RF is a measure of the antimicrobial activity of the disinfectant tested. After performing the handwashing test, volunteers were asked to wash their hands for 2 min with a Hibiscrubw solution (Zeneca, Oslo, Norway).
microlitres of antiseptic or soft soap was then spread over the skin with a sterile glass rod. After 1 min, 950 mL of appropriate neutralizer was added and the surviving bacteria were re-suspended and counted. As a control, a 20 mL bacterial sample was inoculated on skin, dried for 3 min and re-suspended in 980 mL TSB. Surviving bacteria were enumerated using the DCM.
Ex vivo protocol to mimic the in vivo handwashing test Fresh skin samples were cut into 2 £ 2 cm pieces with sterile scissors. The pieces were glued (superglue; Niceday, London, UK) on to individual plastic cap holders (4 –5 cm diameter Petri dishes, Fisher, Loughborough, UK) with the skin surface exposed (Figure 1). Two pieces of mounted skin were needed for each sample being tested. Ten millilitres of an overnight E. coli K12 (NCTC 10538) inoculum containing approximately 7.5 £ 108 cfu/mL was poured into a sterile Petri dish, in which both skin samples were immersed for 5 s. They were then removed and air-dried for 3 min. Immediately after drying, both samples were rubbed for 1 min on the base of a Petri dish containing 5 mL TSB without neutralizer, to assess the release of test organism before antiseptic treatment (i.e. pre-value). Serial
Ex vivo protocol to mimic the in vivo forearm test This ex vivo protocol was based on the method of Messager et al.11 Fresh samples of skin were obtained with patients’ informed consent after plastic surgery (breast and abdominal reduction) and were cut to fit in diffusion cells (JB and DW Jones, Loughborough, UK) especially designed for that experiment with a donor compartment of 36 mm diameter.11 The receptor compartment was filled with Earle’s balanced salt solution (EBSS, Sigma, Poole, UK). A 20 mL aliquot of a E. coli K12 (NCTC 10538) inoculum containing approx. 8 £ 108 cfu/mL was spread over the skin with a sterile glass rod and the suspensions were allowed to dry for 3 min at room temperature. Thirty
Figure 1 Mounted human skin: top (on the left) and bottom (on the right) pieces of skin. The plastic cap holder, onto which the top piece of skin is mounted, has been previously glued onto a steel rod that will be attached to the drill.
118
dilutions of the sampling fluid were prepared in sterile TSB and the number of bacteria released was counted. Immediately after the pre-value sampling and without recontamination of the samples, they were pre-moistened by a brief immersion in sterile water in a Petri dish. Surplus water was allowed to drain back into the dish. The plastic cap on to which the bottom skin sample had been glued was itself mounted on a polystyrene holder on top of a balance to measure the pressure exerted on skin (Figure 2). A 1 mL aliquot of either soft soap or test antiseptic was then applied to the surface of the bottom skin sample. The plastic cap (top piece of skin) was fixed to a steel rod to allow mechanical rotation with a drill (IKA, Labortechnik, Staufen, Germany). Both pieces of skin were put in contact with, the top sample exerting a weight of approximately 100 g on the surface of the bottom sample. The top skin sample was then rotated against the bottom sample for 1 min at 200 rpm (Figure 2). After rinsing the skin samples thoroughly in a Petri dish containing 5 mL of sterile water for 5 s (one Petri dish per sample), they were rubbed on the
Figure 2 Picture showing the ex vivo protocol used. Bottom piece of skin, previously mounted on a plastic holder, was fixed on a weighing machine. The top piece, previously mounted on a plastic holder, itself glued to a steel rod, was pulled down on to the bottom piece to exert a weight of approximately 100 g and the two pieces were rubbed against each other for 1 min at the minimum speed of the drill used: 200 rpm.
S. Messager et al.
base of a Petri dish containing 5 mL of the appropriate neutralizer for 1 min (one Petri dish per sample). The sampling fluids were diluted in neutralizer and the surviving bacteria (i.e. postvalue) were counted. Statistical analyses were performed using Minitabw software and analysis of variance was conducted at the 95% confidence intervals.
Results In vivo and ex vivo testing of antiseptics without rubbing When the antibacterial activity of antiseptics was assessed in vivo on volunteers, forearms, the number of bacteria recovered from the control (untreated) circles was marginally lower (,tenfold reduction in number; P , 0:05) than that of the original bacterial inoculum (Figure 3). The number of bacteria recovered from the treated circles (soft soap and antiseptics) was significantly lower ðP , 0:05Þ than that recovered from the control circles. The concentration of bacteria recovered from the antiseptic treated samples was also significantly lower ðP , 0:05Þ than that recovered from the softsoap-treated samples (Figure 3). Point five percent (w/v) triclosan in 70% (v/v) isopropanol and 2% (v/v) povidone-iodine achieved the highest RFs, followed by 4.8% (w/v) PCMX and finally the soft soap. The antibacterial activity of povidone-iodine was not significantly different ðP . 0:05Þ to that of triclosan in isopropanol (Table II). PCMX was significantly less active ðP , 0:05Þ than triclosan in isopropanol and povidone-iodine (Table II).
Figure 3 Log10 counts of E. coli (cfu/mL) recovered from the forearms of volunteers, depending on the test performed: antiseptic or soft soap compared with the untreated sites (control) and with the original bacterial inoculum (mean values, bars represent 1 standard deviation).
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 119
Table II Results obtained in vivo and ex vivo: log10 reduction in Escherichia coli K12 concentration (cfu/mL) after soft soap or antiseptic treatment for 1 min depending on the test carried out Mean log10 reduction in bacterial concentration (standard deviation)
Test without rubbing effect In vivo forearms Ex vivo Test with a rubbing effect In vivo handwashing Ex vivo
Soft soap
Povidone-iodine
Triclosan in isopropanol
PCMX
0.41 (0.42) 0.06 (0.22)
2.73 (0.93) 0.76 (0.51)
3.45 (1.38) 0.47 (0.58)
1.80 (0.92) 0.49 (0.40)
2.24 (0.521) 1.98 (0.36)
3.41 (0.83) 1.49 (1.29)
3.79 (0.89) 2.46 (0.66)
4.41 (1.26) 2.89 (0.37)
In the corresponding ex vivo test, the bacteria count from the control samples was marginally lower (, tenfold reduction in number; P , 0:05) than that of the original inoculum (Figure 4). There was no difference ðP . 0:05Þ in bacterial counts between treatment with soft soap and the control. The number of bacteria recovered after treatment with antiseptics was marginally lower (although statistically significant; P , 0:05) than that recovered from the control samples (Figure 4). When the results of both in vivo and ex vivo tests without a rubbing effect were compared, the reduction in E. coli achieved after treatment with soft soap was not significantly different (P . 0.05) whether it was performed with the ex vivo test or with the in vivo forearm test (Table II). Overall, the reduction in bacterial concentration after exposure to antiseptics was lower ðP , 0:05Þ when tested with the ex vivo test, regardless of the treatment. Finally, the variability in results obtained with the ex vivo protocol was much lower than that of the in vivo forearm test (Table II).
Figure 4 Log10 counts of E. coli recovered from skin samples treated with soft soap or antiseptics compared with that of the control (untreated) samples and with the original inoculum (mean values, bars represent 1 standard deviation).
In vivo and ex vivo testing of antiseptics with a rubbing effect Counts from the BSEN in vivo handwashing test required validation. Requirements for acceptance were achieved: valid results from all 15 participants were available and the overall mean of the log10 pre-values for reference and test procedures was at least 5 (data not shown). In addition, the number of bacteria recovered from the left or the right hands of each volunteer was not significantly different (P . 0.05) in any test (data not shown). The mean RFs for each antiseptic were all significantly greater than the RF of soft soap (Table II). The antibacterial activity of 0.5% (w/v) triclosan in 70% (v/v) isopropanol was not significantly different (P . 0.05) to that of 2% (v/v) povidone-iodine or 4.8% (w/v) PCMX (Table II). However, PCMX was significantly more active ðP , 0:05Þ than povidone-iodine. Overall, the antimicrobial activity of the antiseptics tested against E. coli was greater than that of soft soap (Table II). When assessed with the corresponding ex vivo test with a rubbing effect, there was no significant difference ðP . 0:05Þ between the concentration of E. coli recovered from the top skin pieces and the concentration of bacteria recovered from the bottom pieces of skin, including both the pre- and post-treatment values (data not shown). Point five percent (w/v) triclosan in 70% (v/v) isopropanol and 4.8% (w/v) PCMX achieved comparable ðP . 0:05Þ Log10 RFs (2.46 ^ 0.66 and 2.89 ^ 0.37, respectively), which were significantly higher ðP , 0:05Þ than that from soft soap (Table II). The log10 RF achieved by 2% (v/v) povidone-iodine was not significantly ðP . 0:05Þ different from soft soap (1.49 ^ 1.29 log10 and 1.98 ^ 0.36, respectively; Table II). When the results of both protocols with rubbing were compared, the antibacterial activity achieved by the soft soap and povidone-iodine was not significantly dependent ðP . 0:05Þ upon the test
120
performed. However, the number of bacteria recovered after exposure to PCMX or triclosan in isopropanol was significantly greater ðP , 0:05Þ when assessed ex vivo (Table II). Furthermore, with the exception of povidone – iodine, variability in results was lower with the ex vivo protocol (Table II).
Discussion Antiseptics can be assessed directly on human skin. However, in vivo studies are limited by several factors including ethical limitations, variability in results, and cost. For many years, there has been a search for internationally acceptable tests for hand disinfecting agents that could be performed by manufacturers and research laboratories alike.1,13,14 The objective of the present study was to compare two novel ex vivo tests with their reciprocal in vivo protocols. The bacterial inoculum dried on the skin surface was inactivated to a lesser extent when assessed ex vivo than when tested in vivo. In essence, this might be an advantage as it makes the ex vivo test more stringent. Differences in inactivation observed between the ex vivo and in vivo protocols might be explained partially by the ‘self-sterilizing’ effect of the skin of volunteers,15,16 due to sebaceous and sweat gland secretions.17 Rubbing was important in reducing the bacterial concentration on the skin. With soft soap, it was clear that this rubbing effect had a significant ðP , 0:05Þ outcome in the reduction of viable bacteria recovered from the skin surface (Table II). Overall, in terms of the rubbing effect, results from the ex vivo protocols correlated well with their respective in vivo protocols, both giving an increased RF (Table II). The physical removal of bacteria from the skin as a result of volunteers rubbing their hands (or fingertips) has been attributed to both the inherent surfactancy of the product used and the mechanical action of rubbing, i.e. breaking up the bacterial colonies on the skin.18 The reduction of bacterial concentration, often referred to as the ‘killing effect’, by exposure to an antiseptic on skin in vivo is therefore due to a combination of physical removal and antimicrobial activity. In addition, in both protocols with a rubbing effect, skin samples or hands were immersed into the bacterial inoculum as opposed to the bacterial inoculum being spread on to the surface in the forearm in vivo and the corresponding ex vivo tests. Spreading the micro-organisms on to the skin surface might also account for the lower
S. Messager et al.
antiseptic activity observed. Some studies showed that rubbed-in organisms are more difficult to kill than dried-on micro-organisms.19,20 The results from our in vivo studies confirmed results from the literature, whereby exposure to an antiseptic reduced the number of bacteria recovered from the skin surface significantly.21 – 25 Further comparison of activity was difficult to ascertain because of differences in methodology; especially products (active ingredient and concentration) used, or lack of information, especially with the forearm protocol. However, despite the use of different protocols, results from the literature on handwashing tests and fingertip experiments provide valuable information. For example, when the bactericidal activity of different antiseptics was tested in vivo, povidone – iodine (1% available iodine) was more effective than 3% chloroxylenol in reducing the overall skin bacterial flora.21 Another study showed that 7.5% povidoneiodine achieved the same reduction of bacterial skin flora concentration after a 30 s wash.22 Munton and Prince23 showed that 5% Dettol achieved a log10 reduction of . 4.95 in Staphylococcus aureus, Enterococcus faecalis, E. coli and Pseudomonas aeruginosa within 5 min in an in vivo investigation using artificially contaminated skin. The low activity of PCMX has been reported elsewhere, although in this case, a low concentration of the phenolic was used.24 When assessed in vivo on hands, 0.5% triclosan in 70% isopropanol (Manuseptw) was significantly more active against E. coli than 60% isopropanol alone after 1 min of treatment.25 These findings correlated with our results that showed that triclosan in isopropanol achieved a log10 RF of 3.45 ^ 1.38 after 1 min without the rubbing effect. The three antiseptics tested here were significantly more effective ðP , 0:05Þ at removing E. coli from the hands than the soft soap. The mean log10 reduction obtained in this study was comparable with that found in an in vivo study using participants’ fingertips.26 Overall, in the handwashing study, the three antiseptics achieved a log RF of . 3 (PCMX achieving . 4 log10). However, the corresponding ex vivo test failed to achieve this level of bacterial reduction. Ayliffe et al.27 reported that after heavy contamination, a bacterial agent should achieve log10 RFs of 3 – 4. Finally, the ex vivo protocols performed in our study showed that, overall, a better reproducibility in results was obtained when compared with the in vivo tests. We believe that an even greater reproducibility would have been obtained if a higher number of replicates was performed. Unfortunately, skin supply limitation (especially large samples) limited us three to four replicates.
Comparison of in vivo and two ex vivo tests to assess the antibacterial activity of several antiseptics 121
In conclusion, the ex vivo protocols described in this study, and particularly the ex vivo test with a rubbing effect, were easily adapted to mimic in vivo protocols and offered the potential to assess antiseptics’ efficacy directly on human skin. However, the ex vivo protocols appear to be more stringent than in vivo protocols since bacterial inactivation achieved ex vivo was lower than that achieved in vivo. This might be considered as an added value for the testing of bactericidal activity of new formulations. The added value of such ex vivo testing models for handwashing and other in vivo protocols is their ability: (1) to simulate a wide variety of skin contamination (e.g. level of microbial contaminants and pathogenicity of micro-organisms); and (2) to control parameters inherent to treatment conditions effectively. In particular, the addition of a mechanical rubbing effect is significant as it imitates the rubbing of both hands against each other during handwashing procedures. In our ex vivo protocol, the rubbing effect can be fully controlled and parameters (i.e. speed of rotation and pressure) can be adapted to correspond to different handwashing habits. Considerations of health, safety and expenses preclude conducting the type of testing presented here on humans. The methods used are easy and inexpensive to perform, and pathogenic micro-organisms, as well as new formulations, can now be assessed effectively on skin. The only caveat is to ensure that informed consent is obtained from the tissue donors for the experiments. This study demonstrated the usefulness and applicability of new testing procedures that could be further standardized for use as reference tests for the preliminary testing of antiseptic formulations on skin.
References 1. McDonnell G, Haines K, Klein D, Rippon M, Walmsley R, Pretzer D. Clinical correlation of a skin antisepsis model. J Microbiol Methods 1999;35:31—35. 2. Pittet D, Boyce JM. Hand hygiene and patient care: pursuing the Semmelweis legacy. Lancet Infect Dis 2001;April:9—20. 3. Larson E. Hygiene of the skin: when is clean too clean? Emerg Infect Dis 2001;7:225—230. 4. BSEN 1499 (European Standard). Chemical disinfectants and antiseptics—hygienic handwash. Test method and requirement (phase 2/step 2). European Standard EN, 1499; 1997. 5. Ayliffe GAJ, Babb JR, Quoraishi AH. A test for hygienic hand disinfection. J Clin Pathol 1978;31:923—928. 6. Payne DN, Eddowes S. An in vivo method for assessing the antimicrobial properties of topical antiseptics using artificially contaminated skin. Lett Appl Microbiol 1990;10: 101—104. 7. Larson E, Rotter ML. Handwashing: are experimental models a substitute for clinical trials? Two viewpoints. Infect Control Hosp Epidemiol 1990;11:63—66.
8. Rotter ML. Hygienic hand disinfection. Infect Control Hosp Epidemiol 1984;5:18—22. 9. Nagai I, Ogase H, Takechi M, Kadota M, Kumamoto R. Evaluation of the disinfectant effect of a quick drying rubbing type povidone-iodine alcoholic solution by the glove juice method. Postgrad Med J 1993;69:S33—S38. 10. Rotter ML, Simpson RA, Koller W. Surgical hand disinfection with alcohols at various concentrations: parallel experiments using the new proposed European standard method. Infect Control Hosp Epidemiol 1998;19:778—781. 11. Messager S, Goddard PA, Dettmar PW, Maillard J-Y. Antibacterial efficacy of several antiseptics tested on skin using the ex vivo test. J Med Microbiol 2001;50:284—292. 12. Maillard J-Y, Messager S, Veillon R. Antimicrobial efficacy of biocides tested on skin using an ex vivo test. J Hosp Infect 1998;40:313—323. 13. Ayliffe GAJ. Standardization of disinfectant testing. J Hosp Infect 1989;13:211—216. 14. Sattar SA, Springthorpe VS. New methods for efficacy testing of disinfectants and antiseptics. In: Rutala A, editor. Disinfection, sterilization and antisepsis: principles and practices in healthcare facilities. Washington, DC: Association for Professionals in Infection Control and Epidemiology, Inc. Publishers; 2000. p. 73—186. 15. Chikakane K, Takahashi H. Measurement of skin pH and its significance in cutaneous diseases. Clin Dermatol 1995;13: 299—306. 16. Cove JH, Eady EA. Cutaneous antimicrobial defense. Clin Dermatol 1998;16:141—147. 17. Selwyn S, Ellis H. Skin bacteria and skin disinfection reconsidered. BMJ 1972;1:136—140. 18. Baquero F, Patron C, Canton R, Martinez Ferrer M. Laboratory and in vitro testing of skin antiseptics: a prediction for in vivo activity? J Hosp Infect 1991;18:5—11. 19. Lilly HA, Lowbury EJL. Transient skin flora—their removal by cleansing or disinfection in relation to their mode of deposition. J Clin Pathol 1978;31:919—922. 20. Bush LW, Benson LM, White JH. Pig skin as test substrate for evaluating topical antimicrobial activity. J Clin Microbiol 1986;24:343—348. 21. Aly R, Maibach HI. Comparative antibacterial efficacy of a 2min surgical scrub with chlorhexidine gluconate, povidoneiodine, and chloroxylenol sponge brushes. Am J Infect Control 1988;16:173—177. 22. Leyden JJ, McGinley KJ, Kaminer MS, et al. Computerized image analysis of full-hand touch plates: a method for quantification of surface bacteria on hands and the effect of antimicrobial agents. J Hosp Infect 1991;18:13—22. 23. Munton TJ, Prince J. The bactericidal activity of Dettol on skin artificially contaminated with micro-organisms using the replica plating technique. Reckitt & Colman SORL BL 75/12 (technical report); 1975. 24. Soulsby ME, Barnett JB, Maddox S. Brief report: the antiseptic efficacy of chloroxylenol-containing vs. chlorhexidine gluconate-containing surgical scrub preparations. Infect Control 1986;7:223—226. 25. Bartzokas CA, Gibson MF, Graham R, Pinder DC. A comparison of triclosan and chlorhexidine with 60 per cent isopropyl alcohol for hygienic hand disinfection. J Hosp Infect 1983;4: 245—255. 26. Rotter ML, Koller W. Test models for hygienic handrub and hygienic handwash: the effects of two different contamination and sampling techniques. J Hosp Infect 1992;20: 163—171. 27. Ayliffe GAJ, Babb JR, Davies JG, Lilly HA. Hand disinfection: a comparison of various agents in laboratory and ward studies. J Hosp Infect 1988;11:226—243.