Quest for a Realistic In Vivo Test Method for Antimicrobial Hand-Rub Agents: Introduction of a Low-Volume Hand Contamination Procedure

Abstract

A novel method has been developed for the evaluation of alcohol-based hand rubs (ABHR) that employs a hand contamination procedure that more closely simulates the in-use conditions of ABHR. Hands of human subjects were contaminated with 0.2 ml of a concentrated suspension of Serratia marcescens (ATCC 14756) to achieve baseline contamination between 8 and 9 log₁₀ CFU/hand while allowing product to be applied to dry hands with minimal soil load. Evaluation of 1.5 ml of an ABHR gel containing 62% ethanol produced log₁₀ reductions of 2.66 ± 0.96, 2.40 ± 0.50, 2.41 ± 0.61, and 2.33 ± 0.49 (means ± standard deviations) after 1, 3, 7, and 10 successive contamination/product application cycles. In a study comparing this low-volume contamination (LVC) method to ASTM E1174, product dry times were more realistic and log₁₀ reductions achieved by the ABHR were significantly greater when LVC was employed (P < 0.05). These results indicate that a novel low-volume hand contamination procedure, which more closely represents ABHR use conditions, provides more realistic estimates of in-use ABHR efficacies. Based on the LVC method, log₁₀ reductions produced by ABHR were strongly dependent on the test product application volume (P < 0.0001) but were not influenced by the alcohol concentration when it was within the range of 62 to 85% (P = 0.378).

INTRODUCTION

Alcohol-based hand rubs (ABHR) are recommended as the primary means for hand hygiene in health care settings when hands are not visibly soiled (5, 21). They are also used as an adjunct to hand washing in food preparation settings (1). Because ABHR use is prescribed for situations when hands are dry and not soiled (5, 21), and in order to produce data that reflect reality, it is vital that excessive hand wetness and soiling are avoided during the in vivo testing of ABHR.

However, high soil and/or hand wetness are predominant features of in vivo methods for evaluating the efficacy of ABHR products (2, 6, 10, 19, 20). Several authors have noted the shortcomings of the current test methods and have called for development of in vivo protocols that more closely simulate real world conditions of use (18, 20, 21). Previous studies from our labs investigating the appropriateness of method ASTM E1174 for evaluating ABHR demonstrated that when hands are contaminated with 4.5 ml of a liquid bacterial culture, organic nutrients from the growth medium and hand wetness at the time of test product application, resulting from insufficient drying time, compromised the antibacterial activities of ABHR (15). Furthermore, hand wetness is exacerbated with repeated hand contamination, which dilutes the active ingredient and leads to a decline in efficacy over multiple hand contamination/product application cycles. Our studies have indicated that modification of the ASTM E1174 hand contamination procedure to decrease either the contaminant volume or reduce the soil load produces only modest changes to these efficacy trends. However, when both the soil load and hand wetness were significantly reduced simultaneously by employing a “contact contamination” procedure, a striking increase in efficacy, which did not decline over repeated contamination/product application cycles, was observed (15). However, the “contact contamination” method is complex, requires significant additional materials and labor, and does not consistently achieve acceptable baseline contamination levels.

The goals of this study were to develop a simple yet robust procedure to contaminate the hands without introducing significant soil load or wetness, to evaluate the efficacy of ABHR by this method in comparison to ASTM E1174, and to evaluate the influence of alcohol concentration and product volume on ABHR efficacy under conditions more reflective of actual use conditions.

MATERIALS AND METHODS

Test product formulations.

The test product used in this study was a commercially available alcohol-based hand rub gel based on 62% (vol/vol) ethanol. Three test formulations identical to the commercial formulation, but with ethanol levels of 0% (vehicle), 70%, and 85%, and a commercially available 4% chlorhexidine gluconate (CHG) handwash formulation were used as controls.

Clinical test laboratories.

Studies involving human test subjects were conducted at two separate laboratories. BioScience Laboratories (Bozeman, MT) conducted the studies summarized below in Table 1 (Lab A), Table 2, and Fig. 4 and 5. Hill Top Research (Miamiville, OH) conducted studies summarized in Table 1 (Lab B) and Fig. 3.

Table 1.

Efficacy of a 62% ethanol ABHR using the LVC procedure in four independent experiments at two laboratories^a

LVC expt no. (lab)	nb	Log10 baseline recoveryc	Log10 reductionc
1 (lab A)	16	8.32 ± 0.53	3.09 ± 0.66
2 (lab A)	12	8.61 ± 0.41	2.41 ± 0.58
3 (lab B)	12	9.21 ± 0.26	2.66 ± 0.80
4 (lab A)	13	8.21 ± 0.37	2.71 ± 0.60
All data	53	8.56 ± 0.56	2.74 ± 0.69

^aProduct application volume was 1.5 ml.

^bNumber of test subjects.

^cValues are means ± standard errors.

Table 2.

Influence of alcohol level on ABHR efficacy when evaluated using the low-volume contamination procedure^a

Test product active ingredient	n	Log10 reductionb
Vehicle, 0% ethanol	12	1.56 ± 0.46
Ethanol, 62%	12	2.41 ± 0.76
Ethanol, 70%	12	2.41 ± 0.96
Ethanol, 85%	12	2.61 ± 0.86

^aProduct application volume was 1.5 ml.

^bValues are means ± standard deviations.

Fig. 4.

Comparison of the performance of an ABHR when evaluated using the LVC procedure and method ASTM E1174. (A) Mean dry times for 1.5 ml of the 62% ethanol ABHR gel product. Errors bars represent standard errors. (B) Mean log₁₀ reductions of S. marcescens achieved with the ABHR product. Values are the means from 12 subjects, and error bars represent the standard errors.

Fig. 5.

Mean log₁₀ reductions of S. marcescens in response to different application volumes of a 62% ABHR gel using the LVC procedure. Values are averages from 12 volunteers at each application volume, and error bars represent the standard errors.

Fig. 3.

Mean log₁₀ reductions of S. marcescens achieved by 1.5 ml of a 62% ethanol hand sanitizer gel (A) or a 30-s wash with 5 ml of a 4% chlorhexidine gluconate handwash over multiple contamination/product application cycles using the LVC procedure (B). Values are the means from 12 subjects (A) and 6 subjects (B), and error bars represent the standard errors.

Human subjects.

All studies in which bacteria were used to contaminate the hands of human subjects were conducted according to an Institutional Review Board-approved protocol. The test populations were overtly healthy subjects of mixed gender and race and at least 18 years of age. All subjects were provided with and signed informed consent forms and were examined to ensure hands were free of clinically evident dermatoses and/or any other disorders that could have compromised the subject and the study. Subjects were instructed to avoid contact with antimicrobial products and harsh chemicals, such as acids, bases, and solvents, for at least 1 week prior to the test and for the duration of the test.

Microorganism and growth conditions.

The challenge bacterial species used in this study was Serratia marcescens ATCC 14756. Liquid cultures were grown in Trypticase soy broth (TSB) at 30°C without shaking, unless otherwise noted. For cultivation on solid medium, the bacteria were incubated at 30°C on Trypticase soy agar (TSA) with lecithin and polysorbate 80. Under these conditions, the strain forms a red pigment, so that it is easily distinguished from resident skin microflora.

Bacterial challenge preparation. (i) Standard challenge suspension.

Standard challenge suspensions were prepared essentially as described in ASTM E1174 (2). Briefly, stock cultures of S. marcescens were established by aseptically transferring contents of a lyophilized vial to approximately 5 ml of sterile TSB and incubating at 25°C for 24 h. Challenge suspensions were prepared by transferring 1 ml of the 24-h stock culture into 1,000 ml of TSB and incubating at 25°C for 24 h, which yielded bacterial titers of approximately 1.0 × 10⁹ CFU/ml.

(ii) High-titer suspension.

For preparation of high-titer suspensions, stock cultures of S. marcescens were prepared by aseptically transferring contents of a lyophilized vial to approximately 5 ml of sterile TSB and incubating at 35°C for 24 h. Five-hundred-milliliter Erlenmeyer flasks containing 125 ml of TSB were inoculated with 1 ml of the 24-h broth culture, and cultures were incubated at 35°C with vigorous aeration (250 rpm) for 25 h. This modification produced culture titers (10 log₁₀ CFU/ml) that were approximately 10-fold greater than achievable using the E1174 procedure. The cultures were then centrifuged at 7,000 × g for 10 min and resuspended to 1/10 the original volume with fresh TSB to yield a homogeneous suspension containing between 5.0 × 10¹⁰ and 1.0 × 10¹¹ CFU/ml. The suspension was assayed for the number of bacteria at the beginning and the end of a use period and was not used beyond 8 h.

Hand contamination. (i) E1174 broth contamination method.

A total volume of 4.5 ml of the standard S. marcescens challenge suspension was transferred onto volunteers' hands by using the procedure described in E1174 (Fig. 1A) (2). A 1.5-ml aliquot of the test bacteria suspension was dispensed onto each subject's cupped hands. Subjects rubbed the aliquot over all surfaces of their hands, up to the wrists, for 20 s and then held them motionless for 30 s to air dry. The dispensing, rubbing, and drying steps were then repeated twice more for a total volume of transfer of 4.5 ml. Following the third aliquot, the air-drying step was extended to 90 s.

Fig. 1.

Hand contamination procedure for method E1174 (A) and the LVC procedure (B).

(ii) Low-volume contamination.

For the low-volume contamination (LVC) method, a 0.2-ml aliquot of the high-titer suspension was dispensed into each subject's cupped hands. Subjects were instructed to distribute the inoculum evenly over all surfaces of both hands, not reaching above the wrist, for 30 s (Fig. 1B). Hands were completely dry by the end of this step.

Test product application. (i) ABHR test products.

For all experiments except those described for Fig. 4, 1.5-ml aliquots of ABHR test products were dispensed as a single pump from a commercially available pump bottle, confirmed prior to the study to consistently deliver the specified volume accurately. Within 10 s of the dispensing, subjects were instructed to rub the test product evenly over all surfaces of the hands up to and including the wrists until dry, exercising caution to retain test product in the hands. Subjects were then instructed to hold hands upright and motionless prior to bacterial recovery. For the results presented below in Fig. 4, specific volumes of product were dispensed from sterile disposable syringes and rubbed over the hands as described above. In some experiments the time required for a product to be rubbed dry on hands of each subject was recorded.

(ii) Handwash formulation.

Application of the 4% chlorhexidine gluconate handwash followed the procedure described in ASTM E1174-06 (2). Subjects' hands were wetted with a small amount of water, and 5.0 ml of the handwash formulation was dispensed into the cupped palm of one hand of each subject. Subjects washed all surfaces of both hands up to but not including the wrist vigorously for 30 s. Particular attention was paid to the area between the fingers, beneath the nails, and around the thumb. The hands and forearms were rinsed under running tap water for 30 s. Bacteria were recovered immediately from wet hands as described below.

Contamination, product application, and recovery schedule.

A subject's hands were contaminated with the challenge bacteria prior to the baseline recovery and prior to each test material application. Hands were sampled to recover surviving bacteria after the baseline contamination and after test product applications. For results from experiments presented below in Fig. 3 and Fig. 4, multiple hand contamination/product application cycles were performed, as described in the FDA's proposed rule, “Tentative Final Monograph for Health-Care Antiseptic Drug Products” (8). A series of 11 hand contaminations were performed. Hands were sampled for baseline recovery immediately after the first hand contamination, and test products were applied after each of the 10 subsequent contaminations. Hands were sampled for surviving bacteria after test product applications 1, 3, 7, and 10.

Bacterial recovery and enumeration.

Baseline and postapplication samples were recovered using the glove juice method and diluted, plated, incubated, and enumerated as described in ASTM E1174 (2). Neutralizer was included in the sampling fluid used for the final sampling, and all diluting solutions contained the neutralizer.

Qualitative visualization of hand contamination.

Hands of test subjects were contaminated according to either the E1174 broth contamination procedure or the low-volume contamination procedure. Immediately following contamination, the palm of the left hand was placed firmly onto a 24.5- by 24.5-cm bioassay dish containing TSA. The hand was lifted from the agar and the dorsum was immediately placed onto a second assay dish containing TSA. The hands of the test subject were decontaminated and contaminated again according to the second hand contamination procedure. Agar imprints were again taken, and all assay dishes were incubated for 24 h at 30°C prior to visualization. Figure 2, below, illustrates the results from a representative subject.

Fig. 2.

Qualitative visualizations of S. marcescens hand contamination by the E1174 and LVC procedures. Imprints of contaminated hands (palm and dorsum) were made onto TSA plates and visualized after overnight incubation.

Neutralizer validation.

A neutralizer validation study performed according to ASTM E1054 ensured that the neutralizing solution employed effectively neutralized the antibacterial properties of the test and control products and was not toxic to S. marcescens (3) (data not shown).

Calculations and statistical analysis.

Plate counts (CFU/hand) from each sample were converted to log₁₀ values, and reductions at each sample time were calculated using the following formula: log₁₀ reduction at posttreatment sample time = log₁₀ baseline population − log₁₀ population at sampling time. A one-way analysis of variance (ANOVA) with a Bonferroni post hoc analysis for multiple comparisons was performed using SPSS 16.0 (International Business Machines Corp., Armonk, NY) to compare the log₁₀ reductions of different treatment arms (α = 0.05). Linear regression analysis was applied to determine the relationships between alcohol concentration and log₁₀ reduction values and between product application volume and log₁₀ reduction values, using Prism 5.04 (GraphPad Software, Inc., San Diego, CA). If a significantly nonzero slope resulted (P < 0.05), then the relationship was considered significant.

Hand decontamination.

Upon completion of testing in all experiments, subjects rinsed their hands and forearms for 1 min with 70% ethanol, air dried them, and then performed a supervised 4-min wash with a 4% chlorhexidine gluconate formulation. A topical antibiotic ointment was applied to each subject's hands following the decontamination procedure.

RESULTS

Low-volume contamination method.

We previously demonstrated that wet, soiled hands resulting from the Healthcare Personnel Handwash (ASTM E1174) hand contamination procedure (Fig. 1A) so compromises the activity of ABHR as to render it inappropriate for evaluation of these products (15). A modified, simplified low-volume hand contamination procedure was developed that introduces minimal soil and that leaves the hands completely dry within 30 s (Fig. 1B) (see Materials and Methods). Figure 2 illustrates that the LVC procedure achieved a uniform contamination of the palmar and dorsal surfaces of the hands that was qualitatively indistinguishable from that produced by the E1174 procedure.

Evaluation of ABHR efficacy using the LVC method.

The LVC procedure was used for hand contamination in four studies conducted at two separate labs to evaluate the efficacy of 1.5 ml of an ABHR gel (Table 1). All other procedural aspects of the studies adhered to E1174, as described in Materials and Methods. Mean log₁₀ baseline recoveries ranged from 8.21 ± 0.37 to 9.21 ± 0.26 (± standard errors). Mean log₁₀ reductions achieved by the test ABHR ranged from 2.41 ± 0.58 to 3.09 ± 0.66, and the mean across all data sets was 2.74 ± 0.69. A statistical comparison of the 4 data sets revealed no significant differences between the mean log₁₀ reductions (P = 0.064).

Evaluation of ABHR efficacy over multiple hand contamination and product application cycles.

A study using the LVC procedure was then conducted to evaluate performance of the ABHR test product over the course of 10 consecutive hand contamination/product application cycles (Fig. 3). Log₁₀ reductions for the test product were 2.66 ± 0.96, 2.40 ± 0.50, 2.41 ± 0.61, and 2.33 ± 0.49 at applications 1, 3, 7, and 10, respectively. The difference between the mean log₁₀ reductions following first product application and final product application (0.33) was not statistically significant (P = 1.0). A handwash formulation containing 4% CHG achieved log₁₀ reductions that increased progressively from 2.48 ± 0.51 at application 1 to 4.72 ± 0.55 at application 10. These results were highly consistent with historical data observed for this product when tested according to ASTM E1174 (G. Mulberry and C. M. Beausoleil, unpublished data).

Comparison of ABHR performance evaluated by the LVC procedure and the ASTM E1174 procedure.

Figure 4 presents the results of a study that directly compared the performance of the ABHR test product evaluated by either LVC or E1174. When E1174 was used, the mean dry times of the test product were unrealistically long, increasing progressively from 129 s at application 1 to 241 s at application 10 (Fig. 4A). In contrast, the mean product dry times were more realistic (44 s at application 1) and did not significantly increase over the course of the study (47 s at application 10) when testing was performed according to the LVC procedure. Mean log₁₀ reductions were significantly greater at both application 1 (P = 0.001) and application 10 (P < 0.001) when the LVC method was used, although reductions resulting from both methods did decline over the course of the study (Fig. 4B).

Influence of product application volume and alcohol concentration on ABHR efficacy.

To determine the ability to detect differences in product performance when LVC is employed in testing, the influences of product application volume (Fig. 5) and alcohol concentration (Table 2) were examined. Mean log₁₀ reductions produced by the 62% ethanol test product were 1.96 ± 0.71, 2.41 ± 0.77, and 3.79 ± 0.83 at product application volumes of 0.75 ml, 1.5 ml, and 3.0 ml, respectively. The slope resulting from linear regression analysis (0.84 log₁₀ per ml) was significantly different from zero (P < 0.0001), indicating a strong positive relationship between volume of product applied and log₁₀ reduction.

The data in Table 2 indicate the lack of influence of alcohol concentration on ABHR efficacy. One-way ANOVA showed the log₁₀ reductions produced by 1.5 ml of each of the ABHR gel products to be significantly greater than with the 0% ethanol vehicle (P < 0.05). However, log₁₀ reductions produced by 62%, 70%, and 85% ethanol test products were statistically equivalent to each other (P = 0.65). Furthermore, linear regression analysis of the log₁₀ reductions produced by the alcohol-containing test products plotted against alcohol concentration identified a slope that did not differ significantly from zero (P = 0.378) (data not shown).

DISCUSSION

We have developed a simple, alternative procedure for contaminating hands when evaluating ABHR in vivo efficacy. This procedure offers several advantages over that used in ASTM E1174 (Table 3). By growing the challenge organism under conditions to maximize titer and then concentrating through centrifugation, high baseline populations (8 to 9 log₁₀ CFU/hand) were achievable with a single delivered aliquot of only 0.2 ml. This LVC procedure takes roughly 30 s to perform, whereas the contamination procedure employed in ASTM E1174 is executed in multiple repeated steps (Fig. 1), takes approximately 4.5 min to perform, and often becomes a bottleneck in the clinical procedure. By reducing the contamination volume from 4.5 ml to 0.2 ml, the soil load introduced onto the hands was dramatically reduced, relative to E1174, and hands were completely dry at the time of test product application over the course of multiple contamination/product application cycles. Therefore, the conditions under which ABHR are evaluated when the LVC procedure is used simulate much more closely the conditions under which ABHR are actually used (i.e., when hands are dry and not visibly soiled) (5). This conclusion is supported by the test product dry times presented in Fig. 4A. When a single pump (1.5 ml) was dispensed from a bottle of a commercially available ABHR, the average time for the test product to be rubbed dry following LVC was approximately 45 s and did not increase over the course of 10 contamination/product applications. In contrast, when the E1174 contamination procedure was used, average product dry times were unrealistically long and increased over multiple contamination/product application cycles (from 2 min to 4 min).

Table 3.

Comparison of key features of ASTM E1174 and the low-volume contamination procedure

Feature	ASTM E1174	Low-volume contamination
Contamination method	Liquid (4.5 ml)	Liquid (0.2 ml)
Contamination time	4.5 min	30 s
Baseline recovery	9 logs	8–9 logs
Soil load	High	Low
Hands adequately dry at time of product application	No	Yes
Consistent with recommended ABHR use patterns (5, 21)	No	Yes
Compatible with recommended use volumes of ABHR products (5, 21)	No	Yes
Compatible with recommended use ABHR product dry times (5, 21)	No	Yes

As a result of using the LVC method, log₁₀ reductions achieved by the ABHR test product were significantly higher than those achieved using E1174 procedures (Fig. 4B). These results are consistent with our previous studies of a contact hand contamination procedure that also eliminated hand wetness and reduced soil (15). Moreover, when the LVC procedure was used to evaluate an antimicrobial handwash, the log₁₀ reductions that resulted were highly consistent with historical data for the formulation when tested according to the E1174 method (Fig. 3). This result was not surprising, because hand wetness and soil load are not expected to influence the efficacy of antimicrobial handwash products, which are intended to be applied to wet, soiled hands. Taken together, these results clearly demonstrate that E1174 routinely underestimates the antibacterial activity of ABHR relative to that of handwashes and that the LVC provides a much more realistic approximation of the in-use efficacy of ABHR.

That the LVC procedure results in evaluations of ABHR that are less procedurally biased suggests it would more accurately detect efficacy differences arising from important test variables, such as alcohol concentration and product application volume. Consistent with previous reports, experiments using the LVC procedure revealed a strong relationship between the test product application volume and antimicrobial efficacy (Fig. 5) (14, 17, 20). Interestingly, no relationship was found between alcohol concentration and efficacy among otherwise-identical formulations ranging from 62% to 85% ethanol (Table 2). Although these results do not support the conclusions of previous reports based on testing per the E1174 or EN1500 methods, the reported dependence of ABHR efficacy on alcohol concentration may have been an artifact of the high soil load and/or hand wetness associated with those methods (13, 20). Furthermore, the data presented here are consistent with previous data from our labs when we used the contact contamination procedure, which also demonstrated that ABHR efficacy was not influenced by alcohol concentration within the range of 50% to 90% (16).

Further research is needed to understand the factors influencing interexperimental and interlaboratory variabilities observed in this study. As illustrated in Table 1, baseline reductions from 4 independent experiments in 2 separate labs ranged from 8.21 log₁₀ to 9.21 log₁₀, and application 1 reductions produced by the test product ranged from 2.41 log₁₀ to 3.09 log₁₀. Although the same method was used in each study, because they were conducted at different times of the year environmental conditions and the condition of subjects' hands varied dramatically. In addition, subtle differences between the two labs in the logistical execution of the entire clinical procedure may have contributed to the variability. Similar variability has been observed for the ASTM E1174 and EN1500 methods (11, 12; Mulberry and Beausoleil, unpublished). Experiments are under way to better understand test variables that may contribute to uncontrolled variability and to establish the repeatability and reproducibility of the LVC method.

The current success criteria set by the U.S. FDA for ABHR and antimicrobial handwash products are a 2-log₁₀ reduction after the first product application and a 3-log₁₀ reduction after the 10th product application when tested using ASTM E1174 (8). There have been numerous discussions regarding the relevance of these criteria, and alternate success criteria and statistical considerations have been proposed (9). Given the interexperimental and interlaboratory variabilities of ASTM E1174 and the LVC method described herein, it may be appropriate to include an internal reference standard, as is used in method EN1500, to control for these variabilities (6). Obviously, an appropriate internal reference would need to be identified, and our labs currently are working to identify appropriate controls and internal reference products that may aid in establishing standardized efficacy criteria. Whatever the success criteria may become, it is critical that they be more predictive of clinical effectiveness.

The LVC procedure described herein has been incorporated into a test method essentially identical to E1174, with the exception of challenge organism preparation and hand contamination procedures. This method was recently approved as a standard test method under the designation ASTM E2755-10 (4). E2755 also provides the option of using a methicillin-resistant Staphylococcus aureus strain as the challenge organism. Results of testing with this species have been recently published elsewhere (7). For regulatory purposes, we propose that E2755 be adopted for the routine evaluation of ABHR. Although the data presented in this paper demonstrate that LVC is effective for evaluation of antimicrobial handwash products and is less complex procedurally than is the E1174 contamination procedure, further studies must be conducted before concluding that a transition to E2755 for routine evaluation of handwash products is scientifically defensible.

In conclusion, a “low-volume” hand contamination (LVC) procedure has been developed that eliminates hand wetness, greatly reduces soil load, and eliminates the excessive ABHR dry times that result when testing ABHR using E1174 procedures. The LVC procedure is more appropriate for evaluating the efficacy of an ABHR because it more closely simulates actual contamination conditions appropriate to ABHR use, it enables testing of products using typical in-use volumes, and it more closely estimates the in-use efficacy of ABHR.