APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 1989, p. 3113-3118 Vol. 55, No. 120099-2240/89/123113-06$02.00/0Copyright C) 1989, American Society for Microbiology

In Vivo Protocol for Testing Efficacy of Hand-Washing Agentsagainst Viruses and Bacteria: Experiments with

Rotavirus and Escherichia coliSHAMIM A. ANSARI,1 SYED A. SATTAR,1* V. SUSAN SPRINGTHORPE,1 GEORGE A. WELLS,2 AND

WALTER TOSTOWARYK2Department of Microbiology and Immunology, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada

KIH 8M5,1 and Division of Biometrics, Laboratory Centre for Disease Control, Health and Welfare Canada,Ottawa, Ontario, Canada KIA OL22

Received 13 June 1989/Accepted 8 September 1989

Ten antiseptic formulations, an unmedicated liquid soap, and tap water alone were compared for theircapacities to eliminate human rotavirus from the finger pads of adult volunteers; three of the antiseptics, thesoap, and the tap water alone were also tested against Escherichia coli. A fecal suspension of virus or bacteriumwas placed on each finger pad and air dried. The contaminated site was exposed to the test product for 10 s,rinsed in tap water, and dried on a paper towel. The residual virus or bacterium was then eluted. Selectedagents were also tested by an analogous whole-hand method by which the entire palm surfaces of both handswere contaminated. Alcohols (70%) alone or with Savlon reduced the virus titer by >99%, whereas thereductions by Proviodine, Dettol, and Hibisol ranged from 95 to 97%. Aqueous solutions of chlorhexidinegluconate were significantly less effective for virus removal or inactivation than 70% alcohol solutions.Furthermore, Savlon in water (1:200) was found to be much less effective in eliminating the virus (80.6%) thanthe bacterium (98.9%). The tap water alone and the soap reduced the virus titers by 83.6 and 72.5% and thebacterial titers by 90 and 68.7%, respectively. The results of the whole-hand method agreed well with those ofthe finger pad protocol. We conclude that the finger pad method is a suitable model for testing the in vivoefficacy of hand-washing agents and emphasize the need for using appropriate test viruses and bacteria.

Hands are among the principal vehicles for transfer ofinfectious agents in the community in general and in institu-tional settings in particular (4, 23, 25, 32). Regular andproper hand washing is, therefore, considered an importantstep in preventing the spread of infectious diseases (15, 20),and guidelines for this purpose are available for health careworkers (12). The continued occurrence of institutionaloutbreaks of many enteric infections, particularly thosecaused by rotaviruses (11, 20), is, however, believed to bedue to the lack of compliance with hand-washing guidelinesrather than the inadequacy of the hand-washing agents (16).

Earlier studies using the suspension test have demon-strated that rotaviruses are not readily inactivated by manytypes of commonly used chemical disinfectants (30); theseviruses proved to be even more resistant to disinfectionwhen they were dried on nonporous inanimate surfaces (18).Since products found to be ineffective in these in vitrostudies included certain commonly used antiseptics, weconsidered it important to examine them and other similarproducts for their abilities to eliminate rotaviruses fromhuman hands.Even though the testing of hand-washing agents against

viruses on experimentally contaminated human hands hasbeen studied before (8, 13, 14, 28), at present there is nostandardized protocol available for such testing. So far, invivo testing of such products has been conducted almostexclusively with bacteria (3, 7, 19, 24). This study was,therefore, undertaken to design and test a simple in vivoprotocol using finger pads for testing the rotavirus-elim-inating efficiency of hand-washing agents. The protocol wasthen applied to compare the rotavirus-eliminating efficien-

* Corresponding author.

cies of the hand-washing agents with their abilities to removeor inactivate Escherichia coli. Finally, experiments wereconducted with selected hand-washing agents to demon-strate that the results with the finger pad protocol arerepresentative of those obtained when the whole hand isexperimentally contaminated with the virus or the bacteriumand washed in the normal manner.

MATERIALS AND METHODS

Cells. The MA-104 line of rhesus monkey kidney cells wasused for the growth and quantitation of the rotavirus. Themethods used for the cultivation, maintenance, and passageof these cells have been described in detail elsewhere (26).For the virus plaque assay, cell monolayers were prepared in12-well plastic plates (Costar, Cambridge, Mass.) with Eagleminimal essential medium (Flow Laboratories, Inc.,McLean, Va.) containing 5% fetal bovine serum (GIBCOLaboratories, Grand Island, N.Y.) and gentamicin (Cidomy-cin; Roussel, Montreal, Quebec, Canada) at a final concen-tration of 50 ,ug/ml. The seeded plates were sealed in plasticbags (Philips, Toronto, Ontario, Canada) and incubated at37°C for 48 h for monolayer formation.

Fecal sample. A sample of formed feces was obtained froma healthy 1-year-old child. A 10% (wt/vol) suspension of thesample was prepared in normal saline. The suspension wascentrifuged at 1,000 x g for 15 min to remove coarseparticulate matter, and the supernatant was collected forstorage at -20°C. It was found to be nontoxic to MA-104cells, noninhibitory to the rotavirus, and free from anyagents that could produce cytopathic effects or plaques inthe cell line (2).

Virus. A cell culture-adapted strain (Wa) of human rota-virus obtained from R. G. Wyatt (National Institutes of

3113

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

3114 ANSARI ET AL.

TABLE 1. Hand-washing agents

Active ingredient(s) % Concn Dilution Product (manufacturer)

Isopropanol 70.0 (vol/vol) None NAaEthanol 70.0 (vol/vol) None NA"Chlorhexidine gluconate and 1.5 (wt/vol) and 15.0 (wt/vol) 1:30 in 70% isopropanol Savlon (Ayerst, Montreal, Quebec,

cetrimide Canada)Chlorhexidine gluconate and 1.5 (wt/vol) and 15.0 (wt/vol) 1:30 in 70% ethanol Savlon

cetrimideChlorhexidine gluconate and 1.5 (wt/vol) and 15.0 (wt/vol) 1:30 and 1:200 in tap water Savlon

cetrimideChlorhexidine gluconate and 0.5 (wt/vol) and 70.0 (wt/vol) None Hibisol (Ayerst)

isopropanolChlorhexidine gluconate and 2.0 (wt/vol) and 4.0 (wt/vol) None Cida-Stat (Huntington, Bramalea,

isopropanol Ontario, Canada)Povidone-iodine 10.0 (wt/vol) (1% free 12) None Proviodine (Rougier, Montreal,

Quebec, Canada)PCMX 4.8 (wt/vol) and 9.4 (vol/vol) None Dettol (Reckitt & Colman,

Lachine, Quebec, Canada)Liquid soap 100 1:10 in tap water Ivory (Procter & Gamble,

Toronto, Ontario, Canada)Tap water

a NA, Not applicable.

Health, Bethesda, Md.) was used in this study. The proce-dures for the preparation of the virus pools and plaque assayhave been described earlier (2). For contamination of thefinger pads or hands, the virus suspension was prepared bydiluting it 10-fold in the fecal sample.

Bacterium. The strain of E. coli selected for this study wasisolated from the fecal sample used for suspending therotavirus. For the initial isolation and quantitation of E. coli,mFC agar (Difco Laboratories, Detroit, Mich.) was used.The test suspension of the bacterium was prepared bymixing an 18-h broth culture with the fecal sample at a 1:10ratio.

Volunteers. Permission of the Ethics Committee of theUniversity of Ottawa was obtained for the use of humanvolunteers; one adult male and two adult females partici-pated in this study.Hand-washing agents. Ten antiseptic formulations, an un-

medicated liquid soap, and tap water were tested for theirefficacy in removing the rotavirus from experimentally con-taminated finger pads. Details of the hand-washing agentstested are given in Table 1. Isopropanol (70%), Savlon in70% ethanol (1:30), Savlon in tap water (1:200), liquid soap,and tap water alone were also tested against E. coli.Paper towels. Rolls of ordinary unbleached paper towels

(Boudreault, Inc., Hull, Quebec, Canada) were purchasedlocally and used for drying treated finger pads or hands as

specified in the protocols.Finger pad protocol. A modification of the protocol devel-

oped earlier in our laboratory to study the survival ofrotaviruses on human hands was used (2). The details of thisprocedure are given in Fig. 1, but only the results for themethod which included paper towel drying are reportedhere. The same basic procedure was also used with thebacterium, but E. coli was quantitated as CFU on mFC agarplates. Elution fluid for both the rotavirus and E. coli was

Earle balanced salt solution containing 20% (vol/vol) tryp-tose phosphate broth, and quantitation of the organisms was

performed immediately after elution. At the end of eachexperiment, the contaminated finger pads of the volunteerswere thoroughly rinsed with 70% ethanol and then washedwith soap and water.Whole-hand protocol. The following whole-hand washing

protocol was used to ascertain that the results obtained with

our finger pad protocol were representative of the normalhand-washing situation. The volunteers did not wash or

decontaminate their hands before the start of the experi-ment. A 0.5-ml portion of the fecal suspension of the virus orthe bacterium was placed on the palm surface of one hand.The volunteer was then required to spread the inoculum over

the entire palm surfaces of both hands by gently rubbingthem together, and the inoculum was allowed to dry for 20min under ambient conditions. To obtain the base titer of the

5 fingerpads of one hand were each inoculated withtOoL nirs suspension and atlowed to dry ton 20 min

thumb other fingers

FIG. 1. Procedure for in vivo testing of hand-washing agents.

hands washed with tap water only (10 sec),rinsed with ethanol and air dried

I I

iingerpads marked with vial rim

--41

APPL. ENVIRON. MICROBIOL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

HAND-WASHING AGENTS AGAINST VIRUSES AND BACTERIA 3115

TABLE 2. In vivo efficacy of hand-washing agents against humanrotavirus in finger pad protocola

Hand-washing agent % Reduction Tukey(concn) (mean + SD) grouping

Isopropanol (70%) 99.8 ± 0.18 AEthanol (70%) 99.8 ± 0.19 ASavlon in 70% isopropanol (1:30) 99.4 ± 0.84 ASavlon in 70% ethanol (1:30) 99.1 ± 0.50 AProviodine 96.5 ± 2.65 A, BDettol 96.2 ± 3.62 A, B, CHibisol 95.3 ± 2.15 A, B, CLiquid soap 86.9 ± 2.42 D, B, CSavlon in water (1:30) 86.4 ± 4.07 D, CTap water 83.6 ± 3.49 DCida-Stat 80.3 ± 3.51 DSavlon in water (1:200) 78.3 ± 5.55 D

a Agents with the same Tukey groupings do not have significantly differentpercentages of reduction (a = 0.05).

test organism present at the end of this drying period, 20 mlof the eluent (20% tryptose phosphate broth in Earle bal-anced salt solution) was poured on the hands while theywere being rubbed together over a plastic funnel 27 cm indiameter in order to bring the eluent in contact with theentire contaminated surface. Generally, more than 18 ml ofthe eluate was collected by this procedure.To test the efficacy of a hand-washing agent, the hands

were contaminated with the test organism and dried asdescribed above. Then 0.5 ml of the test agent was placed onthe palm surface of one hand, and the volunteer was requiredto rub his or her hands for 10 s, thereby covering the entirecontaminated surface. The hands were then washed bypouring over them 500 ml of tap water at 40°C and dried witha paper towel. The residual test organism was then elutedfrom the washed and dried hands, as described above. Theeluates from the rotavirus-contaminated hands (controls andtests) were passed through separate columns of a SephadexLH-20 gel (Pharmacia, Uppsala, Sweden) to remove anyresidual cytotoxicity due to antiseptic chemicals (5). Thefiltrates from the columns were then centrifuged at 10,000 xg to remove bacteria and fungi before the virus plaque assay.Eluates from bacterium-contaminated hands were titratedimmediately on mFC agar plates.

Statistical analyses. For the finger pad protocol, the testingof each hand-washing agent with each microorganism wasconducted at least three times by using two finger pads fortreatment with the test agent in each trial. Therefore, theresults presented in Table 2 are the mean of six observationsfor each hand-washing agent with one volunteer, and thosegiven in Table 3 are based on the mean of six observations

TABLE 4. Comparison of whole-hand and finger pad protocolsa

Hand-washing agent % Reduction (mean + SD)and protocolb E. coli Rotavirus

70% IsopropanolWH >99.9 ± 0.1 99.9 ± 0.1FP 99.0 ± 1.5 99.8 0.2

Savlon in 70% ethanol (1:30)WH >99.9 ± 0.1 98.9 ± 0.6FP 98.7 ± 0.3 99.1 0.5

Savlon in tap water (1:200)WH 96.6 ± 2.2 84.8 + 6.8FP 99.6 ± 0.1 78.3 ± 5.6

Liquid soap (1:10)WH 94.5 ± 3.2 93.8 ± 3.9FP 60.8 ± 36.8 86.9 ± 2.4

Tap waterWH 95.1 ± 4.4 85.5 ± 12.0FP 90.0 ± 1.9 83.6 ± 1.4

a Differences between whole-hand and finger pad protocols were notsignificant for any agents tested.

b WH, Whole hand; FP, finger pad.

with three volunteers. For the whole-hand washing protocol(Table 4), each experiment was repeated three times and thevolunteer used was the same as in experiments summarizedin Table 2.

Results were statistically analyzed by using two-wayanalysis of variance based on the arcsine transformation ofthe raw data, i.e., the percentage of organisms removed. Thearcsine transformation was utilized in order to stabilize thevariance. Comparisons of the means of results obtained withthe various hand-washing agents were made by the Tukey wprocedure (31). This procedure allows for multiple compar-isons of means while controlling for the error rate at thedesired level.

RESULTS

The challenge titers of the test organisms in the finger padexperiments were (1.5 + 0.4) x 104 and (9.9 + 1.6) x 104 forthe rotavirus and E. coli, respectively. These titers weremeasured at the end of a 20-min drying period, the minimumtime for a visually dry appearance of the inoculum underambient conditions, and they represent 57 and 7% of theoriginal inoculum for rotavirus and E. coli, respectively. Inthe whole-hand experiments, after 20 min of drying, (1.8 +

0.9) X 104 and (5.7 + 3.2) x 104 were the challenge titers left

TABLE 3. Comparison of in vivo efficacy of hand-washing agents against human rotavirus and E. coli in finger pad protocol

Mean % reduction in infectivity titer with:

Volunteer 70% Isopropanol Savlon in 70% ethanol Savlon in water Liquid soap Tap water(1:30) (1:200)

Rotavirus E. coli Rotavirus E. coli Rotavirus E. coli Rotavirus E. coli Rotavirus E. coli

99.1 98.7 77.5 99.698.8 99.1 77.3 97.699.4 99.0 87.0 99.5

78.171.468.1

60.880.664.7

83.682.284.9

90.088.191.8

Overall mean 99.6 + 0.2 99.3 + 0.4 99.1 + 0.3 98.9 + 0.2 80.6 5.5 98.9 + 1.1 72.5 ± 5.1 68.7 ± 10.5 83.6 1.4 90.0 ± 1.9± SD

123

99.899.699.5

99.099.799.2

VOL. 55, 1989

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

3116 ANSARI ET AL.

on the hands. These amount to 22.9 ± 6.0 and 1.4 ± 1.3% ofthe original inoculum for the rotavirus and E. coli, respec-tively.The results of testing hand-washing agents and tap water

against the rotavirus are shown in Table 2. Seventy percentisopropanol, 70% ethanol, and Savlon containing 70% ofeither one of the alcohols were each able to reduce theinfectivity titer of the virus by >99.0%. Hibisol, a waterlessproduct containing chlorhexidine gluconate and 70% isopro-panol, eliminated 95.0% of the virus. Dettol and Proviodineachieved 96.2 and 96.5% virus removal, respectively. Cida-Stat and Savlon in water (1:30 and 1:200) produced thelowest levels of virus removal of all the antiseptics tested, asthey were able to reduce the virus titer by only 80.3, 86.4,and 78.3%, respectively. Surprisingly, in terms of the meanpercentages of virus removal, the liquid soap (86.9%) and tapwater alone (83.6%) were as good as, if not better than,Cida-Stat and Savlon in water (1:200).The percentages of virus removal by the various hand-

washing agents were statistically compared by using theTukey w procedure. Results of the multiple comparisonsmade by the w procedure of the means of the 12 hand-washingagents are displayed in Table 2. Those means designated bythe same letter under the Tukey grouping were considerednot significantly different from each other at the ot = 0.05level. No Jifferences were seen among the hand-washingagents represented by the letter A. Likewise, no differenceswere demonstrated among those with the letter D. Group A,primarily the alcohol-containing formulations and those so-lutions with high concentrations of iodine or p-chloro-m-xylenol (PCMX), was significantly different from group D,which included the aqueous chlorhexidine salts, soap andwater, and water alone. However, multiple comparisons alsoshowed the existence of intermediate groups B and C, whichshowed some overlap with the means for hand-washingagents in groups A and D. Despite these overlaps, thealcohols or alcohol-containing formulations were clearlysuperior to water and the aqueous solutions of chlorhexidinesalts in reducing the levels of contaminating virus.

In a separate set of experiments using three volunteers,four hand-washing agents and tap water alone were com-pared for their abilities to remove the rotavirus and E. colifrom experimentally contaminated finger pads (Table 3). Thedifferences in the mean percent removals for the virus andthe bacterium were not statistically significant for the alco-hols, the alcoholic solution of Savlon, the soap, or the tapwater alone. However, there was a statistically significantdifference (P = <0.0001) between the mean percent remov-als of E. coli (98.9%) and the rotavirus (80.6%) by Savlon inwater (1:200).

Statistical analyses of the data summarized in Table 3 alsoshowed that the person-to-person variation in mean percentreductions by a hand-washing agent against a given organismwas not significant (P = 0.14).

In order to determine that the finger pad protocol wasrepresentative of the normal hand-washing practice, selectedhand-washing agents and tap water alone were also tested bythe whole-hand protocol for their abilities to remove the twotest organisms. The results of these experiments are sum-marized in Table 4. Isopropanol (70%) and the alcoholsolution of Savlon reduced the titers of both the organismsby at least 98.0% in both the protocols tested. Even thoughthe overall removal of the virus and the bacterium by theother hand-washing agents and tap water alone was not ashigh as that seen with the alcohol solutions, there was nosignificant difference between the results obtained with the

whole-hand washing and the finger pad protocols. The widevariations observed between the finger pad and whole-handexperiments with liquid soap and E. coli were not found tobe significant at the at = 0.05 level. For Savlon in water(1:200), the results with the whole-hand experiments weresimilar to those with the finger pad protocol; this solutionwas found to be significantly less effective in the eliminationof the virus (84.8%) than of the bacterium (96.6%).

DISCUSSION

Hand washing is regarded as the single most importantprocedure for preventing the transmission of infections (23).However, the impact of hand washing depends not only onthe regularity and thoroughness of the procedures used butalso on the type of hand-washing agent selected. The con-tinued occurrence of disease outbreaks, such as rotaviralgastroenteritis (9, 20), in health care settings suggests poorcompliance with hand-washing guidelines (1, 10). However,a high degree of compliance alone may not be sufficient forproper infection control if the hand-washing agent is noteffective in the removal or inactivation of important noso-comial pathogens. The findings of this study indicate thatsome commonly used hand-washing agents may be limited intheir abilities to deal effectively with viruses which areknown to cause disease outbreaks in institutions.Most of the in vivo evaluation of hand-washing agents has

been conducted with bacteria; there are very few publishedreports in which the agents have been tested against viruseson experimentally contaminated hands (8, 13, 14, 28, 29).Furthermore, there is no standard in vivo protocol to test thevirucidal properties of hand-washing agents.Whereas in commonly used suspension and carrier testing

of germicides a disinfectant is required to reduce the infec-tious titer of the test organism by at least 99.9% to beconsidered effective, there is no such generally recognizedcriterion of efficacy for the in vivo testing of antiseptics forviruses. Schurmann and Eggers (28) and Carter et al. (6)reported that antiseptics were more effective in vitro than invivo. In view of this, it is considered necessary to test otherformulations and additional representative viruses and bac-teria before deciding on a reasonable criterion for efficacy forantiseptics tested by in vivo protocols.Although surveys on health care personnel have shown

that the mean duration for total hand washing is about 8 s(21), in this study, the 10-s exposure of the contaminatedarea to the agent being tested was based on the hand-washing guidelines of the U.S. Centers for Disease Control(12). This relatively short exposure time is also consideredmore suitable for testing antiseptics used for hygienic handdisinfection (21, 23).The rinsing of hands may be an important step in hand

antisepsis; tap water is normally used in the field, and wefollowed this practice in the testing protocol. Although it isacknowledged that the composition of tap water may varygeographically and temporally, neither distilled nor standardhard water is believed to be more predictive than tap waterof the results which would be obtained under natural condi-tions. Moreover, the organic or inorganic content of tapwater could affect the elution of the test organism from thefingers. It can be argued that the residual chlorine in tapwater plays a role in the inactivation of pathogens oncontaminated hands. We do not believe this to have been animportant factor in this study, as the residual chlorine in ourtap water was low (0.05 to 0.2 ppm [0.05 to 0.2 ,ug/ml]).Previous work in our laboratory has shown that rotavirus is

APPL. ENVIRON. MICROBIOL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

HAND-WASHING AGENTS AGAINST VIRUSES AND BACTERIA 3117

relatively resistant to chlorine in tap water (22). Further-more, the feces used for suspending the test organisms mayhave interfered with the action of such small amounts ofchlorine by neutralization of the chlorine or protection of thetest organisms or both. However, inactivation by tap watercould occur for some other viruses (8, 22) and bacteria.Therefore, we suggest that dechlorinated tap water may be a

suitable rinse water for routine in vivo testing of antiseptics,but this needs further investigation.The selection of a human rotavirus for this study was

based on several facts. Not only are these viruses among themajor causes of acute gastroenteritis, but they are frequentlyimplicated in disease outbreaks, even when chemical disin-fection and antisepsis are regularly used for infection control(20). They are also among the very few human pathogenicviruses whose minimal infective dose, as judged from humanvolunteer studies (33), has been shown to be 1 cell cultureinfective unit. The viruses are capable of surviving on humanhands for several hours (2) and on nonporous inanimatesurfaces for several days (27). Infectious rotaviruses can betransferred readily between hands and inanimate objects (2),and both viable viruses (15) and their antigens (25) have beendemonstrated on the hands of care givers. Considerableevidence on the relative resistance of rotaviruses to chemicaldisinfectants is already available (18, 30).

E. coli was chosen as the representative bacterium be-cause it is a widely accepted indicator of fecal contaminationand is also a common pathogen transmitted by hands.Previous studies on antisepsis have used strains of thisorganism as models (3, 24). The challenge levels of both ofthe test organisms and the organic load were realistic fornormal hand contamination in the field, although it is recog-nized that higher levels of challenge may exist from time totime.

In a field situation, antiseptics should be suitable for theinactivation of both viral and bacterial pathogens on contam-inated skin. This may be particularly true when the minimalinfective dose for some viral agents is extremely low.Furthermore, some viruses may survive on human skin morereadily than many bacteria. In this study, E. coli was clearlymore susceptible to inactivation during drying than was therotavirus since a smaller proportion of its initial inoculumwas recovered after 20 min of drying on either finger pads or

hands. Recovery of both the virus and the bacterium fromthe whole-hand surface after 20 min was less than when onlythe finger pads were contaminated. However, it is not knownwhether this was due to the decreased efficiency of recoveryor the increased rate of inactivation of the test organisms orboth.Use of the whole hand for regular testing of antiseptics

against viruses is not considered feasible. Apart from therelatively large volumes of a high-titered virus pool requiredfor each experiment, determination of the base titer of theorganism and treatment of hands with the test product mustbe conducted at different times. For viruses, a detoxification(disinfectant removal) step is also required. Moreover, re-

covery of inoculated organisms from the whole hand isgenerally lower and more variable than from finger pads.These factors make the protocol inherently difficult to con-

trol. On the other hand, the finger pad method describedhere represents a simpler and better way of testing antisep-tics against viruses as well as bacteria. It not only permitsthe use of microliter quantities of the infectious agent butalso allows the inclusion of proper controls and sufficientnumbers of replicates in the same test. The results obtainedwith this protocol were reproducible and did not show any

significant person-to-person variation; they were also inagreement with the findings based on the whole-hand tech-nique. The suitability of using fingers instead of whole handsin antisepsis experiments has also been noted by others (28).Studies in progress in our laboratory have also successfullyapplied the finger pad protocol for the testing of hand-washing agents against rhinoviruses.

Alcohols and formulations containing 70% alcohol werethe most effective in reducing rotavirus and E. coli contam-ination of hands. This is in agreement with findings of otherinvestigators (3, 24). In The Federal Republic of Germany,60% isopropanol is used as an index for comparing the invivo bactericidal efficacy of antiseptics (24). Much of the lossof infectivity of the virus and the bacterium on the skinsurface due to these antiseptics may be the result of in situinactivation. In the case of the other agents tested, it mayhave been predominantly due to wash-off, as discussedbelow. Furthermore, the addition of emollients to alcoholsor alcohol solutions of chlorhexidine salts makes them lessastringent and more widely acceptable (17). Although bothHibisol and Savlon contain 70% alcohol, Hibisol contains a

higher concentration of chlorhexidine gluconate than Savlondoes at the recommended use dilution. Despite this, Savlonperformed better under the test conditions used. This mayhave been due to the presence in Savlon of cetrimide as anadditional active ingredient.

Dettol was chosen as a product containing PCMX. Theconcentration of PCMX in the undiluted product tested washigher than is usually found in PCMX-based hand-washingagents (17). Like Dettol, Proviodine is an antiseptic notroutinely used for hygienic hand disinfection. However, itwas selected because previous studies of rotavirus disinfec-tion (30) showed it to be the most effective of the iodine-based formulations tested. Even though these products were

found to be more effective than soap and aqueous chlorhex-idine salts, the level of rotavirus removal by the more dilutesolutions of PCMX or iodine-based products formulated forhygienic hand washing may not be as high.Aqueous solutions of chlorhexidine gluconate (Savlon and

Cida-Stat) are commonly used for hand washing; the resultsof this study clearly demonstrate that they may not bereliable antiseptics for pathogens such as rotaviruses. Infact, virus removal by tap water alone was comparable tothat achieved by both dilutions of Savlon in water. Thissuggests that the reduction in titer levels by this product maybe due primarily to wash-off. Similar results for the action ofaqueous chlorhexidine salts have been obtained in in vitrostudies of rotavirus disinfection (18, 30).The routine use of plain soap and water is considered by

many to be quite adequate for the decontamination of handsin infection control. This is not borne out by the resultsreported here. Washing of contaminated hands with almostany agent is clearly desirable, but it should be noted that use

of a less effective product could actually lead to the spread oflocalized contamination over the entire surfaces of bothhands (29). This phenomenon is under investigation in our

laboratory.Regular and proper washing of hands by health care

personnel may be inadequate if ineffective products are

relied on. Therefore, selection of an antiseptic with demon-strated broad-spectrum efficacy is considered essential inconjunction with efforts to increase compliance with goodhand-washing practice. Proper testing of hand-washingagents against both viruses and bacteria by an in vivoprotocol is clearly desirable, and the finger pad protocolpresented here may be a suitable model for such standard

VOL. 55, 1989

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

tests. Further evaluation of the finger pad protocol withother viral and bacterial pathogens is under way in ourlaboratory.

ACKNOWLEDGMENTS

Financial support for S.A.A. has been provided by a scholarshipfrom the Ministry of Education of the Government of Pakistan.Complimentary supplies of Cidomycin were received from Rous-

sel. We are grateful to the volunteers who agreed to participate inthis study.

LITERATURE CITED1. Albert, R. K., and F. Condie. 1981. Handwashing patterns in

medical intensive care units. N. Engl. J. Med. 304:1465-1466.2. Ansari, S. A., S. A. Sattar, V. S. Springthorpe, G. A. Wells, and

W. Tostowaryk. 1988. Rotavirus survival on human hands andtransfer of infectious virus to animate and nonporous inanimatesurfaces. J. Clin. Microbiol. 26:1513-1518.

3. Ayliffe, G. A., J. R. Babb, J. G. Davies, and H. A. Lilly. 1988.Hand disinfection: a comparison of various agents in laboratoryand ward studies. J. Hosp. Infect. 11:226-243.

4. Black, R. E., A. C. Dykes, K. E. Anderson, J. G. Wells, S. P.Sinclair, G. W. Gary, Jr., M. H. Hatch, and E. J. Gangarosa.1981. Handwashing to prevent diarrhea in day-care centers.Am. J. Epidemiol. 113:445-451.

5. Blackwell, J. H., and J. H. S. Chen. 1970. Effects of variousgermicidal chemicals on H.Ep.2 cell culture and herpes simplexvirus. J. Assoc. Off. Anal. Chem. 53:1229-1236.

6. Carter, C. H., J. 0. Hendley, L. A. Mika, and J. M. Gwaltney,Jr. 1980. Rhinovirus inactivation by aqueous iodine in vitro andon skin. Proc. Soc. Exp. Biol. Med. 165:380-383.

7. Casewell, M. W., M. M. Law, and N. Desai. 1988. A laboratorymodel for testing agents for hygienic hand disinfection: hand-washing and chlorhexidine for the removal of Klebsiella. J.Hosp. Infect. 12:163-175.

8. Cliver, D. O., and K. D. Kostenbader, Jr. 1984. Disinfection ofvirus on hands for prevention of food-borne disease. Int. J.Food Microbiol. 1:75-87.

9. Di Matteo, A., A. Sarasini, M. S. Scotta, M. Parea, G. Licardi,and G. Gerna. 1989. Nosocomial outbreak of infant rotavirusdiarrhea due to the appearance of a new serotype 4 strain. J.Med. Virol. 27:100-104.

10. Donowitz, L. G. 1987. Handwashing technique in a pediatricintensive care unit. Am. J. Dis. Child. 141:683-685.

11. Ford-Jones, E. L. 1987. The special problems of nosocomialinfection in the pediatric patient, p. 494-540. In R. P. Wenzel(ed.), Prevention and control of nosocomial infections. TheWilliams & Wilkins Co., Baltimore.

12. Garner, J. S., and M. S. Favero. 1985. Guideline for handwash-ing and hospital environmental control, 1985. Centers for Dis-ease Control, Atlanta.

13. Hayden, G. F., D. Deforest, J. 0. Hendley, and J. M. Gwaltney,Jr. 1984. Inactivation of rhinovirus on human fingers by viru-cidal activity of glutaric acid. Antimicrob. Agents Chemother.26:928-929.

14. Hendley, J. O., L. A. Mika, and J. M. Gwaltney, Jr. 1978.Evaluation of virucidal compounds for inactivation of rhinovi-rus on hands. Antimicrob. Agents Chemother. 14:690-694.

15. Keswick, B. H., H. L. Pickering, H. L. Dupont, and W. E.Woodward. 1983. Survival and detection of rotaviruses on

environmental surfaces in day care centers. Appl. Environ.Microbiol. 46:813-816.

16. Larson, E. 1987. Skin cleansing, p. 250-256. In R. P. Wenzel(ed.), Prevention and control of nosocomial infections. TheWilliams & Wilkins Co., Baltimore.

17. Larson, E. 1988. APIC guidelines for infection control practice.Guideline for use of topical antimicrobial agents. Am. J. Infect.Control 16:255-266.

18. Lloyd-Evans, N., V. S. Springthorpe, and S. A. Sattar. 1986.Chemical disinfection of human rotavirus-contaminated inani-mate surfaces. J. Hyg. 97:163-173.

19. Ojajarvi, J. 1980. Effectiveness of handwashing and disinfectionmethods in removing transient bacteria after patient nursing. J.Hyg. 85:193-203.

20. Pacini, D. L., M. T. Brady, C. T. Budde, M. J. Connell, V. V.Hamparian, and H. J. Hughes. 1987. Nosocomial rotavirusdiarrhea: pattern of spread on wards in a children's hospital. J.Med. Virol. 23:359-366.

21. Quraishi, Z. A., M. McGuckin, and F. X. Blais. 1984. Durationof handwashing in intensive care units: a descriptive study. Am.J. Infect. Control 12:83-87.

22. Raphael, R. A., S. A. Sattar, and V. S. Springthorpe. 1987. Lackof human rotavirus inactivation by residual chlorine in munici-pal drinking water systems. Rev. Int. Sci. Eau 3:67-69.

23. Reybrouck, G. 1986. Handwashing and hand disinfection. J.Hosp. Infect. 8:5-23.

24. Rotter, M. 1988. Are models useful for testing hand antisepsis?J. Hosp. Infect. ll(Suppl. A):236-243.

25. Samadi, A. R., M. I. Huq, and Q. S. Ahmed. 1983. Detection ofrotavirus in the handwashings of attendants of children withdiarrhea. Br. Med. J. 286:188.

26. Sattar, S. A., M. K. Ijaz, C. M. Johnson-Lussenburg, and V. S.Springthorpe. 1984. Effect of relative humidity on the airbornesurvival of rotavirus SA-11. Appl. Environ. Microbiol. 47:879-881.

27. Sattar, S. A., N. Lloyd-Evans, V. S. Springthorpe, and R. C.Nair. 1986. Institutional outbreaks of rotavirus diarrhea: poten-tial role of fomites and environmental surfaces as vehicles forvirus transmission. J. Hyg. 96:277-289.

28. Schurmann, W., and H. J. Eggers. 1983. Antiviral activity of analcoholic hand disinfectant. Comparison of the in vitro suspen-sion test with in vivo experiments on hands, and on individualfingertips. Antiviral Res. 3:25-41.

29. Schurmann, W., and H. J. Eggers. 1985. An experimental studyon the epidemiology of enteroviruses: water and soap washingof poliovirus 1-contaminated hands, its effectiveness and kinet-ics. Med. Microbiol. Immunol. 174:221-236.

30. Springthorpe, V. S., J. L. Grenier, N. Lloyd-Evans, and S. A.Sattar. 1986. Chemical disinfection of human rotaviruses: effi-cacy of commercially available products in suspension tests. J.Hyg. 97:139-161.

31. Steel, R. G. D., and J. H. Torrie. 1960. Principles and proce-dures of statistics. McGraw-Hill Book Co., Toronto.

32. Steer, A. C., and G. F. Mallison. 1975. Handwashing practicesfor the prevention of nosocomial infections. Ann. Intern. Med.83:683-690.

33. Ward, R. L., D. I. Bernstein, E. C. Young, J. R. Sherwood, D. R.Knowlton, and G. M. Schiff. 1986. Human rotavirus studies involunteers: determination of infectious dose and serologicalresponse to infection. J. Infect. Dis. 154:871-880.

3118 ANSARI ET AL.

on March 9, 2020 by guest

http://aem.asm

.org/D

ownloaded from