Fluorescence Assay for Evaluating Microbicidal Activity of Hand Antiseptics

Abstract

We developed a fluorescent β-d-glucuronidase activity (BGA)-based assay for detecting and quantifying Escherichia coli in samples to assess the biocide efficacy of hand antiseptics. The fluorescence level is proportional to the number of viable E. coli organisms present. We compared our assay results to those of the E. coli plate count method specified by the European standard for testing hygienic hand rub disinfectant products (EN1500). The plate count method requires excessive handling and materials and is not valid if the number of organisms per plate is too low or high for counting in many of the samples. We optimized the fluorescent assay based on the cleavage of 4-methylumbelliferyl-β-d-glucuronide by adding 4-nitrophenyl-β-d-glucuronide, a nonfluorogenic BGA substrate, to induce glucuronidase activity and reduce assay time. Furthermore, our method can be automated and eliminates the need for multiple dilutions. Fluorescence was temporally monitored, and the time required to reach a specific value of fluorescence was correlated with the initial number of viable E. coli organisms on the samples. There was a positive correlation (P < 0.05) with a high correlation coefficient (R² = 0.82) between the E. coli counts by plate count and fluorescence methods. Reported effects in fluorescent BGA were compared to the EN1500 plate count method with five hand disinfectants. We found our method more advantageous, because it was as sensitive as the EN1500 method, requires less time to complete, and is less expensive and less laborious than conventional plating techniques.

INTRODUCTION

Health professionals' hands are the most common vehicles for transmitting microorganisms between patients and within the health care environment (1). Washing hands before and after patient contact is considered the most effective solution to prevent the spread of microorganisms responsible for health care-associated infections (HCAIs) (2). Aspects such as professional status, work area, staff number, glove use, and individual or social factors influence handwashing practices (1). Since handwashing is effective in reducing HCAIs, a high level of adherence to previously validated disinfection protocols is required (1, 3).

Although liquid antiseptic soap is effective in reducing microorganism numbers, solutions containing alcohol and other biocides for hand disinfection are widely used (4). Alcohol-based hand rubs are fast-acting bactericidal solutions, less skin irritating, and more efficacious than handwashing under some conditions (5, 6). International organizations provide guidelines for assessing the antimicrobial efficacy of hand disinfectants (7–10). European norms (EN), simulating practical conditions to test disinfectant products, include EN1499 (hygienic hand wash), EN1500 (hygienic hand rub), and EN12791 (surgical hand disinfection) (11). EN1499 and EN1500 are two phase 2/step 2 tests for postcontamination treatment performed on hands artificially contaminated with a strain of Escherichia coli. These norms use a crossover design to compare a product to a reference (nonmedicated soap or 60%, vol/vol, isopropanol, respectively) on artificially contaminated hands of healthy subjects.

EN1499 and EN1500 are culture based, and the analysis time is ≥24 h. The method involves immersing fingertips in an E. coli-contaminated liquid medium and determining the microorganism count on a petri dish before (prevalues) and after (postvalues) washing hands with a product or reference standard. The difference between both values is the reduction factor, expressed as log₁₀. For approval, the product should not be significantly less effective than the reference.

Product analysis according to EN1499 or EN1500 requires two bacterial counts (pre- and postvalues) of the right and left hands of 18 to 22 volunteers for both the test product and the reference standard. The work is time-consuming and requires a number of materials to test product effectiveness. Thus, techniques that allow some automation can significantly save materials and time and reduce errors.

Recently, several assays based on the enzymatic activity of microorganisms were developed to rapidly count cells (12). E. coli produces β-d-glucuronidase that hydrolyzes fluorogenic substrates such as 4-methylumbelliferyl-β-d-glucuronide, producing a fluorescence which depends on the number of cells present in the sample (13). The main objective of our study was to evaluate a β-d-glucuronidase activity (BGA)-based rapid assay to estimate the antimicrobial potential of hygienic hand rub products under controlled conditions by following EN1500 requirements. To validate the assay, we compared the in vitro efficacy of five widely used products for surgical hand antisepsis following the BGA-based assay described in this article and the plate count method (EN1500).

MATERIALS AND METHODS

Products tested.

Hand sanitizers tested were 3 ml alcohol gel (AG; 70%, wt/wt, ethyl and propyl alcohol; Deb Spain, Ltd.), 3 ml liquid soap presurgical (LSP; 0.8% chlorhexidine digluconate, 0.1% 2-bromo-2-nitropropane-1,3 diol; Técnicas de Higiene Ltd., Spain), 3 ml NDP dermal (ND; N-duopropenide 0.46%, ethanol 60%, 2-propanol 15%; Vesismin Ltd., Spain), and 3 ml Biostex (BT; 2-butanone peroxide 2%; Neochemical S.A., Spain). Liquid medium for neutralizing antimicrobials (NA), obtained from Scharlau Chemie, Spain (0.1% peptone, 0.1% l-histidine, 0.3% lecithin, 0.36% monopotassium phosphate, 0.72% disodium phosphate, 0.43% sodium chloride, and 3% polysorbate 80), was diluted in tryptic soy broth (TSB), validated according to EN1500, and used in all experiments (TSB-NA).

Hand disinfection procedure.

EN1500 is the European standard for demonstrating the efficacy of hand disinfection products, under practical conditions, compared with the reference, alcohol (60%, vol/vol, isopropanol) (14, 15). This norm requires 18 to 22 test volunteers and was rigorously followed in the study. Microorganisms were counted using two methods: the EN1500-specified plate count method and our fluorescence method.

The four hand sanitizers and 60% isopropanol each were tested using the same 22 student volunteers of the University of Malaga. Exclusion criteria included cuts, abrasions, or other hand skin disorders. Each product or reference was tested in two phases (half of the volunteers) on different weeks, with 1 week between each test. Participants were trained in EN1500 techniques prior to experiments. Written informed consents were obtained in advance.

Based on the EN1500 procedure, an overnight suspension of E. coli K-12 (NTCC10538) was adjusted with TSB to an optical density at 630 nm (OD₆₃₀) between 0.150 and 0.460 to obtain a contamination fluid (CF) made up of a cell suspension ranging between 1.5 × 10⁸ and 5 × 10⁸ CFU/ml. Hands of volunteers then were washed for 60 s with soft soap, dried, and immersed in the CF up to the midmetacarpals for 5 s with fingers spread. Fingers were air dried for 3 min, and fingertips of each hand were separately rubbed for 60 s in a petri dish containing 10 ml of TSB-NA (prevalue plate). After drying hands (30 s), a product (following manufacturer recommendations) or control (two alcohol reference spreads of 3 ml for 60 s each, according to EN1500) was applied to the hands of volunteers. Once dry (3 min), fingertips of each hand were rubbed for 60s in a petri dish containing 10 ml of TSB-NA (postvalue plate) as before. One-milliliter aliquots then were removed from pre- and postvalue plates of each experiment to estimate microorganism numbers using the EN1500-specified plate count method (PCM) and the E. coli glucuronidase-based fluorescence method (FM). According to EN1500, the time interval between sampling and seeding for PCM or FM was not more than 30 min.

Bacterial recovery and enumeration by plate count.

In PCM, serial dilutions were plated to calculate the number of microorganisms from pre- and postvalue plates. Three replicates (0.1 ml) from each dilution were plated on selective tryptic soy agar (TSA containing 0.5 g · liter⁻¹ of sodium deoxycholate). Colonies with morphology similar to that of E. coli were counted with a plate counting system (Kodak EDAS 290 image capture system with Kodak 1D image analysis software). EN1500 specifications were used for statistical analysis of the biocide activity of test products. The nonparametric Kruskal-Wallis test was used to compare the CFU of CF in the different tests with study products or reference alcohol. Differences between pre- and postvalues (expressed as log₁₀ CFU) were used to obtain the corresponding reduction factor (RF) of each sanitizer. The nonparametric Wilcoxon test (P ≤ 0.05) was used for paired comparisons between the RF of reference alcohol and tested products.

Fluorescence method.

The number of microorganisms was calculated using a slightly modified E. coli BGA-based fluorescence assay (16) by adding 4-nitrophenyl-β-d-glucuronide as an inducer of BGA. Fluorescence reagent was 5 mg/liter 4-methylumbelliferyl-β-d-glucuronide and 200 mg/liter 4-nitrophenyl-β-d-glucuronide (17), both from Sigma-Aldrich, in sterile distilled water.

In the procedure, 10 ml of CF was centrifuged (4,500 rpm, 10 min), and the precipitate was diluted in TSB-NA medium to obtain a cell suspension OD₆₃₀ (CS OD₆₃₀) of 0.1 ± 0.03. A standard calibration curve (SCC) next was prepared using six additional 10-fold serial dilutions in tubes containing the same medium, and 0.1 ml from each tube was transferred into a sterile polystyrene flat-bottomed 48-well microtiter plate (Iwaki Co., Japan). Concurrently, 0.1-ml aliquots obtained from the pre- and postvalue plates were transferred into different wells of the microtiter plate, and five replicates were carried out in each experiment. An aliquot of 0.9 ml of the fluorescent reagent was transferred into all wells of the plate and incubated in the dark at 37°C. The remaining wells, containing 0.9 ml of fluorescent reagent and 0.1 ml of TSB-NA, were used as blank controls to correct the background signal. The fluorescence (excitation wavelength [λ_ex], 340 nm; emission wavelength [λ_em], 445 nm) was measured as relative fluorescence units (RFU) every 15 min for 10 h using a microtiter plate reader FLx800 (BioTek). For each hand sanitizer tested in each assay with the 22 volunteers, a total of six 48-well microtiter plates were used. The number of CFU/ml in CS OD₆₃₀ was quantified on TSA using conventional plating techniques.

The time required to reach an RFU of 500 (RFU₅₀₀) and the number of CFU in CS OD₆₃₀ dilutions were used to set up the SCC in each test, as previously described (18). Initially, a correlation between E. coli concentration (as log₁₀ CFU/ml) and the geometric means of time required to reach RFU₅₀₀ was plotted for each experiment. The number of microorganisms estimated by fluorescence in pre- and postvalue plates then was deduced from the linear regression equation obtained in SCC for each experiment. Like the PCM, differences between pre- and postvalues (expressed as log₁₀ CFU) estimated by fluorescence were used to obtain the corresponding RF of each sanitizer. Data for each product and reference alcohol are presented as means and standard deviations. Finally, microorganism numbers determined using the fluorescence enzyme assay and PCM on pre- and postvalue plates were compared. Independent sample t tests were used for multiple comparisons. Analysis of variance was performed for pairwise multiple comparisons to determine significant differences using the statistics software program from SPSS Inc., Chicago, IL.

RESULTS

This work compares the in vivo efficacy of different hand antiseptics in five trials with 22 volunteers, using two methods: EN1500-specified PCM and the β-d-glucuronidase activity-based rapid fluorescence method (FM) developed in this study. The average number of cells in the hand contamination fluids of the five assays was 8.43 ± 0.55 log₁₀ CFU/ml. The PCM showed that the initial colony counts (prevalues) before disinfection ranged from 5.10 to 6.83 log₁₀ CFU/ml (means, 6.36 ± 0.50 log₁₀ CFU/ml). Table 1 shows the mean log₁₀ RF obtained for the hand antiseptics assayed. Overall, the formulations produced effects of typical magnitudes with RF between 2.48 ± 0.85 for presurgical liquid soap (LSP) and 4.10 ± 1.54 for alcohol gel (AG) in the PCM procedure.

TABLE 1.

Mean log₁₀ reduction factor obtained with five hand antiseptics according to EN1500 (PCM) and FM

Preparation	RF (mean log10 ± SD)
	PCM	FM
AG	4.10 ± 1.54	3.98 ± 1.30
BT	2.59 ± 0.63	2.63 ± 0.92
LSP	2.48 ± 0.85	2.39 ± 0.58
ND	2,92 ± 0.54	3.16 ± 0.72
ISP	3.91 ± 1.02	4.21 ± 1.49

^aLog ₁₀ RF is the log₁₀ prevalue minus the log₁₀ postvalue.

Simultaneously, a CS OD₆₃₀ was obtained for each experiment, and the standard calibration curve (SCC) for the FM was constructed by using 10-fold serial dilutions as mentioned above. No significant differences were found in the colony count CS OD₆₃₀ (mean values ranging from 7.76 to 8.79 log₁₀ CFU/ml) used in all experiments. As shown in Fig. 1, the fluorescence curves of BGA versus time moved to the right when the concentration of bacteria decreased. For each assay, a correlation between E. coli concentration and RFU was plotted, and the time to reach RFU₅₀₀ was calculated by regression analysis on the linear portion of the corresponding curve (between 200 and 900 RFU). The same procedure was used to calculate the time required to reach RFU₅₀₀ in the pre- and postvalue plates. The average time required to reach RFU₅₀₀ was similar for each test, ranging from 72.89 ± 5.58 min for the first dilution (about 10⁶ CFU/ml) to 507.01 ± 29.89 min for the last dilution, with no significant difference between the times required for each dilution. All SCCs revealed a linear relationship (Fig. 2) between the RFU₅₀₀ values and E. coli concentrations from 1 to 6 log₁₀ CFU/ml (R² > 0.90). The corresponding linear regression equation of the SCC then was used to calculate the E. coli concentration in pre- and postvalue plates from the RFU₅₀₀ values obtained for each disinfectant. According to FM, initial colony counts (prevalues) ranged from 5.88 to 6.73 log₁₀ CFU/ml (means, 6.40 ± 0.33 log₁₀ CFU/ml). No significant differences in the initial colony counts were found in the different experiments with study products or reference alcohol or between PCM and FM.

FIG 1.

Sample data set for E. coli BGA expressed in RFU. D1 to D6 are successive 10-fold dilutions from the OD₆₃₀ cell suspension (D1, 6.54 ± 0.53 log ₁₀ CFU/ml at the assay) corresponding to the standard curve. VI and VF are E. coli BGA of pre- and postvalue plates, respectively, obtained in the assay with ISP.

FIG 2.

Relationships between the time to reach RFU₅₀₀ and E. coli concentration of the calibration curve (◆) obtained in the assay with ISP.

The number of bacteria recovered from postvalue plates showed that all test products significantly reduced (P ≤ 0.001) the number of organisms in both the PCM and FM (Table 1). Comparing the RF of different products, ISP and AG showed maximum reductions at 99.99%. AG produced reductions similar to those of ISP (P > 0.09), which were significantly higher than those of BT and LSP (P < 0.008), in both methods. Furthermore, although the reduction with ND was less than that of ISP and AG, this difference was not statistically significant for either of the methods tested based on the EN1500 valuation criteria.

The regression lines in Fig. 3 compare the results obtained for each disinfectant using PCM and FM. Both methods showed good correlation (R² > 0.80) across the five trials. Furthermore, the effectiveness of the neutralizing agent on the antiseptics tested was validated using the EN1500 procedure.

FIG 3.

Relationship between the RF values (as log₁₀) obtained by PCM (log RF CFU) and FM (log RF RFU) with 5 hand antiseptics: AG, BT, FGC, ND, and ISP.

DISCUSSION

In this study, we tested the feasibility of replacing the PCM described in EN1500 with our fluorescence method (FM) based on E. coli BGA to estimate bacterial numbers and to measure the efficacy of hand antiseptics. The PCM technique requires up to 48 h to detect and enumerate E. coli while consuming considerable material and time. In contrast, FM is faster (the process can be completed in 12 h), consumes less material, and can be easily automated. A significant (P < 0.05) positive correlation with a high correlation coefficient (R² = 0.89) between BGA and E. coli bacterial counts was previously reported (19). Our study showed similar correlation coefficients (R² > 0.82) in each SCC of BGA in the 5 hand antiseptic trials. Furthermore, the small time differences required for each of the six serial dilutions used to reach RFU₅₀₀ among the SCCs were not statistically significant and were attributed to the small differences in the starting bacterial concentration of CS OD₆₃₀. According to EN1500, only counts between 330 and 14 CFU/plate are valid; thus, serial dilution is required to determine the number of E. coli organisms in pre- and postvalue plates. In contrast, FM does not require sample dilution, and aliquots from the pre- and postvalue plates can be directly analyzed. However, low E. coli counts (10¹ or 10² CFU/ml) often require long incubation periods to detect BGA. In previous studies using E. coli BGA (16), 18 h of incubation was required to detect E. coli concentrations as low as 4 × 10¹ CFU/ml. In this study, the addition to the medium of NAP, a known BGA inducer (17), reduced this time by approximately 5 h. The obvious advantage of the FM is the minor assay time and less need for materials, such as petri dishes and culture media, provided that a fluorescence reader is available.

In some studies, BGA estimated a higher number of cultivable E. coli organisms from environmental samples than from pure cultures (13, 20). The presence of active, noncultivable, or false-positive bacteria was postulated as the most likely cause (12, 13). This finding demonstrates that BGA may not be suitable to predict E. coli in environmental samples but is suitable for pure cultures. In our study, we found a strong positive correlation between FM and the conventional plate count method described in EN1500. A hand washing with soap, as described in EN1500, before E. coli contamination reduces the microbial load of hands (21), and several dilutions are made prior to counting E. coli. Although some common skin microorganisms are included in the plate count, their numbers usually do not skew the results. Furthermore, under these conditions, false positives due to BGA from other microorganisms are infrequent, and E. coli from CF is often in pure culture in the dilutions from the pre- and postvalue plates.

Although this study was not designed to determine the efficacy of surgical hand preparations, our findings showed significant differences between the products tested. The best results were obtained for the alcohol-based preparations, AG and ISP, with the RF being significantly better than those of LSP and BX; however, no significant difference was found for ND. This is not surprising, as the higher efficacy of alcohol-based surgical hand preparations is well known (22). These results clearly demonstrate that both PCM and FM are useful for counting E. coli using the protocols described. Nevertheless, the fluorescent method is quicker, avoids numerous dilutions, and prevents contamination during handling. Results are obtained automatically, and counts that are not between 330 and 14 CFU/plate are not discarded in the final calculation. A shortcoming is that the fluorescence assay was performed with a neutralizing agent suitable to the disinfectants tested that does not affect glucuronidase activity. Thus, the effect of another disinfectant neutralizer on the fluorescence assay must be considered.

A known limitation of the EN1500 standard is that the efficacy of disinfectants is determined only against E. coli (11, 23). Although the purpose of the rule is to compare various disinfectants under this norm, the effectiveness of disinfectants could be tested against other microorganisms of interest according to the EN1500 method and the fluorescent method by using fluorogenic assays suitable to other microorganisms (13, 24).

In the field of hand hygiene, European regulations of importance exist for assessing the antimicrobial efficacy of hand disinfectants. Testing products according to the norms simulating practical conditions are obligatory for each hand disinfectant. With EN1500, an antiseptic hand rub must not prove to be significantly inferior to treatment with a reference alcohol according to a very strict protocol for the same volunteers. The procedure described here based on BGA has been incorporated into a test method basically identical to EN1500, with the exception of the method for counting organisms. Although the data presented in this study demonstrate that FM is effective for evaluation of antiseptic hand rub and is less time-consuming, not being regulated by the appropriate agency, it should be considered only for research purposes, such as when different products are compared and especially when required to process simultaneously a large number of samples.