Preliminary results from disinfection of irreversible hydrocolloid

impressions

The virucidal efficacy of germicides acting on an irreversible hydrocolloid surface is not known.
Tests that are currently performed on germicides do not simulate the conditions under which the
germicides are often used. One major concern for the dental profession is the disinfection of
dental impressions, particularly irreversible hydrocolloid impressions. This study was designed to
test the biocidal action of germicides against an enveloped virus on an irreversible hydrocolloid
surface. The disinfection model, which was developed to simulate clinical conditions, specified the
use of vesicular stomatitus virus, an animal virus amenable to safe handling. A 0.6% sodium
hypochlorite spray inactivated the virus when the spray was allowed to remain on the impression
3 to 10 minutes. The iodophor disiniectant required a 3- to lo-minute immersion for total
inactivation. Although 2% glutaraldehyde achieved total viral inactivation in less than 1 minute,
the authors conclude that short disinfectant sprays, in general, are not an appropriate disinfection
method.

D ental practitioners, auxiliaries, and laboratory personnel are subject to significant risk with
respect to infectious disease, which can be spread by saliva or blood as droplets and aerosols, or
by direct contact.1-3 Underscored in this risk profile are hepatitis B virus (HBV), the AIDS retrovirus
(human immunodeficiency virus [HIV]), the herpes viruses (HSV), the tubercle bacillus,
staphylococci, streptococci, and viruses infesting the upper respiratory tract. The practices
recommended for the prevention of disease transmission include sterilization of some items,
disinfection of others, and partial isolation (barrier protection) of the practiti0ner.w 4 A major
concern is the problem of disinfecting dental impressions, particularly irreversible hydrocolloid
(hydrophilic) impressions. Irreversible hydrocolloid is a complex carbohydrate that imbibes water.
It is reasonable to assume that if pathogens are also imbibed into the irreversible hydrocolloid,
they would be less exposed to the action of germicides. Several studies have tested the effects of
various disinfectants on dimensional stability of irreversible hydrocolloid impressions. 5-10
However, few studies have examined the residual infectious potential of an impression after a
disinfection procedure. Rowe and Forrest? demonstrated the existence of bacterial contamination
on hydrophilic impressions, which can be transferred to and cultured from gypsum casts made
from contaminated impressions.12 These studies were not designed to identify viral
contamination, which may constitute a much greater challenge to disinfection than do vegetative
bacteria. To quantitatively measure the efficiency of virucides in a dental office application, a
model was developed that closely simulates clinical conditions. In preference to using a more
virulent human viral pathogen in this model, we chose a low-risk lipophilic enveloped virus,
vesicular stomatitus virus (VSV). VSV is one of the most studied viruses and is amenable to safe
handling.13J4 Preliminary tests with this virus showed that it is potentially a useful prototype for
testing disinfectants. Because the efficacy of a virucide may drop significantly when organic “soil”
(blood, saliva, exudate, and plaque) is present, this model was designed to represent the organic
soil present in the mouth by using human saliva for the suspension of VSV.15 The interference to
viral inactivation in this model is thus twofold: physical interference produced by possible
irreversible hydrocolloid imbibition and chemical interference from competing proteins in the
saliva. The most accurate way to measure viral inactivation when unknown variables are present,
is to take an inoculum from the impression and do direct infectivity testing on a susceptible host.
This model specifies this type of testing on a mammalian cell culture (plaque titration), which is
able to reveal the action of a single surviving virus.

MATERIAL AND METHODS

Tissue culture cells, medium, and virus recovery medium
The mammalian cell monolayers were baby hamster kidney transformed cells (BHK) which were
grown in Sigma Hybri Max tissue culture medium (Sigma Chemical Co., St. Louis, MO.)
supplemented with antibiotics and 10% fetal calf serum (FCS). Virus recovery medium was Sigma
Hybri Max medium buffered with 25 mM HEPES (pH 6.8 to 8.2) and with no FCS added.

Lipophilic enveloped virus

For this study, two stocks of VSV (Indiana Serotype- HR strain) was grown on BHK cells to yield a
titre of lOlo to lOi infectious virions, (or plaque-forming units [PFU]), per ml. The virus was stored
at -80’ C in 2 ml aliquots. Starter virus stock and BHK cells were obtained as a gift from Drs. P.
Plagemann and R. Peluso, Department of Microbiology, University of Minnesota.

Human saliva for suspension of the virus

Whole expectorated saliva was collected from the first two authors and mixed. To minimize
degradation of the salivary proteins, the saliva was collected in tubes placed in an ice beaker,
immediately sterile filtered (0.45 micron filter), and stored at -20’ C. Sterile filtration produced a
60 % reduction of total protein in the saliva sample by the removal of cells and mucin. However,
the protein concentration after filtration was 2.2 mg/ml (Lowry protein assay17), which is still
within reported values for unfiltered human saliva (1.4 to 6.5 mg/ml).l

Disinfectants tested
The following list describes the four disinfectants tested in the study. 1. Alkaline glutaraldehyde,
2% (Cidex-7, Johnson & Johnson, East Windsor, N.J.), freshly-mixed and 28- day-old solutions were
tested, but no difference was observed between them as to virucidal activity. 2. Sodium
hypochlorite (NaOCl), 0.5 % , was a 1:lO dilution of a 5% solution of NaOCl. 3. Iodophor solution
(ProMedyne-D, Cottrell Ltd., Parker, Colo.) was freshly diluted (1:213) with distilled water before
each trial to yield 75 ppm titratable iodine. 4. Two phenolics, one immersion type and one spray
type, were subjected to preliminary testing only: (a) 0-syl (Lehn and Fink, Montvale, N.J.)
disinfectantdetergent was diluted tc 0.78 % according to the manufacturer’s recommendation.
The active ingredients are 0-benzyl-p-chlorophenol, 0-phenylphenol, and isopropyl alcohol. (b)
Lysol Scent II spray (Lehn and Fink). The active ingredients are 0-phenylphenol (0.1%) and ethyl
alcohol (79%).

Experimental procedure
Irreversible hydrocolloid impressions (Jeltrate Impression Material, type II, normal set) were made
on sterile epoxy models of a mandibular posterior quadrant (first premolar to second molar). Each
impression was used just once for a single trial. Because the pH of irreversible hydrocolloid (pH
4.6) may create a problem for disinfectant testing (refer to Discussion), the impressions were
neutralized to pH 7 in a solution of distilled water and NaHCOs. VSV stock was thawed and 0.5 ml
was mixed with 4.5 ml room temperature sterile human saliva. The mixtures from the two virus
stocks assayed at 1 x lo9 and 5 x lOlo PFU/ ml, titres that were judged to be appropriately
representative of human serum virus titres, such as those observed in patients infected with
HBV.lg A sterile epoxy model was then dipped in the VSVlsaliva mixture and inserted into a
neutralized, rinsed, and randomly selected irreversible hydrocolloid impression for 1 minute. The
VSV-contaminated impression was rinsed briefly in distilled water to remove all visible red
medium (VSV/saliva mixture). A l-second rinse with water flowing at 35 ml/second was sufIicient
to eliminate the coloration. One investigator did all the rinses to maintain consistency. The
impression was then exposed to one of the following disinfection procedures:

 Spraying the impression four times, immediately followed by a rinse of 100 ml distilled
water
 Spraying the impression four times, a l-minute humid storage under wet paper towel, and
then the 100 ml distilled water rinse
 Spraying four times, a 3-minute storage and the 100 ml rinse
 Spraying four times, a lo-minute storage and the 100 ml rinse
 Immersing the impression for 1 minute in the disinfectant followed by a 100 ml rinse with
distilled water l Immersing for 3 minutes followed by the 100 ml rinse
 Immersing for 10 minutes followed by the 100 ml rinse

After disinfection and rinse, 5 ml of virus recovery medium was placed in the impression for
30 seconds to harvest the remaining viruses. The virus inoculum was recovered from the
impression by pipet, serially diluted to 10-l and 10s2, and tested for infectivity by plaque assay
on a monolayer of BHK cells.* This method gave the best estimate of the number of viable
virions that had survived the disinfection procedure and were still able to replicate in a
mammalian cell.

Cytotoxicity control
The BHK cells were subject to toxic effects from both the germicide and the irreversible
hydrocolloid carryover. The problem of germicide carryover has been encountered by a
number of investigators, who have used various chemical means, and dilution of the
inoculum, to reduce toxic effects on the mammalian cells.16r 22-26 We found that BHK
freshly grown to confluence were less susceptible to toxic effects of carryover products.
Inocula dilutions of 1:lO or 1:lOO were used when necessary to reduce toxic cell death to less
than 10 % after 24 hours. Cytotoxic controls were run with each series of trials by using
uncontaminated irreversible hydrocolloid impressions, which underwent the same
disinfection procedures.

Sample sizes
The research design involved testing six impressions in each cell assigned to the three primary
disinfectants of this study: 2 % glutaraldehyde, 0.5 % NaOCI, and the iodophor (75 ppm 12).
The purpose of this design was to establish trends and determine endpoints (no virus
recovered, [NVR]). When no virus was recovered after a simple spray with glutaraldehyde, it
was decided that the six remaining cells in the hierarchy of increasing exposure to this
disinfectant would not be tested. Glutaraldehyde served as a positive control for the spray
trials, bringing to 14 the number of impressions treated in this manner. NaOCl served as a
positive control for the immersion trials, but with reduced sample sizes of three.
Unexpectedly high ranges of outcomes in two cells required 10 observations. Two
intermediate cells in the iodophor column were also tested with samples of three.

Statistical analysis
Fisher’s exact test, which is appropriate for small sample sizes, was used to evaluate
differences between cells. The chi-square test statistic was used for statistical evaluation of
summary data between disinfectants. ANOVA, generated by the statistical analysis system
procedure PROC REG, was appropriate for one data set (iodophor sprays) with no zero
outcomes. The Wilcoxon rank sum test served to analyze the statistical association between
surviving virus counts and the two different virus stocks that were used in this study. This test
is not sensitive to extreme values such as those that were observed with the simple spray
technique.
RESULTS
Table I indicates that 2 % glutaraldehyde totally inactivated VSV in less than 1 minute. The
actual disinfection time was the time required to spray the impression four times and rinse it.
No PFUs were observed in 42 plaque assays done on the inocula recovered from 14
impressions. Spraying 0.5% NaOCl on the impression left seven of 10 impressions positive for
VSV. The erratic results (range of 0 to 5600; mean 646 + 1745) produced by a simple NaOCl
spray demonstrate the unreliability of using NaOCl in this manner. One measure of this
unpredictability was the large coefficients of variation associated with the NaOCl sprays (245
to 270). Spraying NaOCl on the impression and storing it for 10 minutes produced total
inactivation of VSV, as did immersion for 1 to 10 minutes. Iodophor spray did not inactivate
the virus. All impressions treated by a spray method (25 total) were positive for VSV, as was
the impression immersed for 1 minute in iodophor (six total). The iodophor achieved total
inactivation only with the lo-minute immersion. Table II gives the raw data from the simple
spray trials done with both NaOCl and the iodophor. A wide range in the results was noted
despite all attempts to perform uniform tests. The small number of trials done with the
phenolics suggested that a spraying technique was similarly ineffective (up to 3 minutes
storage), but the immersion type may be adequate with long immersion times (10 minutes). A
graphic presentation of the data in Table I is seen in Fig. 1. The statistical analysis done on
these data showed inactivation of VSV by glutaraldehyde to be statistically superior to that of
NaOCl and iodophor. NaOCl was statistically superior to iodophor for certain disinfection
protoshowed an overall difference between the three disinfectants: chi-square 25.600, p <
0.001. Fisher’s exact test was also used for the top row of Table I to compare ratios of VSV
positive impressions for glutaraldehyde versus NaOCl and for glutaraldehyde versus iodophor.
Both comparisons yielded p values of less than 0.001. Summary data for glutaraldehyde (0 for
14) were compared with summary data for NaOCl (9 for 31): chi-square 5.081, p 0.024. The
same comparison with the iodophor summary data (33 for 40) yielded chi-square 29.700, p <
0.001. These summary data comparisons are conservative because the effect of
glutaraldehyde at a low exposure level only is being compared to the data from NaOCl and the
iodophor, which include both low and higher level exposures. An obvious difference between
NaOCl and the iodophor is that NaOCl spray is able to inactivate the virus, and iodophor spray
is not. Statistical comparison of the spray disinfection techniques associated wth l-minute
storage, and 3-minute storage both showed NaOCl to be significantly superior (Fisher’s exact
test: p 0.0076), even when the conservative Bonferroni a posteriori test is applied, which
divides the nominal (r=O.O5 by the number of comparisons (6) to give (Y*= 0.0683. Even
though an iodophor (75 ppm Ix) spray does not totally inactivate VSV, a significant reduction
of VSV counts was found to be associated with increasing levels of exposure to the iodophor.
With log10 transformations done on the surviving virus counts and on the spray exposure
times (calculated as 0.1, 1.1, 3.1, and 10.1 minutes), PROC REG (SAS system) indicated a
regression slope of -0.514 (p < 0.003). In addition, a Fisher’s exact test (one-sided), comparing
a lo-minute immersion with a lo-minute exposure to spray, showed the immersion technique
to be significantly superior to the spray (p < 0.006). Statistical analysis by the Wilcoxon rank
sum test (twosided) was done to determine whether an association existed between surviving
virus counts and the two VSV stocks, one containing 1 x log PFU/ml, and the other with 5 x
lOlo PFU/ml (Table II). With alpha level 0.05, no statistical association was demonstrated. The
NaOCl data showed a p valve of 0.83 and the iodophor data resulted in p = 0.15.

DISCUSSION AND CLINICAL IMPLICATIONS

This study arose from the need to investigate the virutidal efficacy of germicides in conditions
that simulate a clinical situation as closely as possible. In 1972, Spaulding and Groschel,27 and
others since that time, have stated that in vitro tests do not always apply to the clinical
conditions in which disinfectants are used. McDuff and Gaustad2s noted that the disinfectant
tests that use suspensions of virus do not represent the surfaces on which disinfectants are
used. A steel cylinder was used as a carrier for the virus. Others have used porcelain carriers
or vinyl surfaces.2gr30 Whereas the use of a surface appears to have some definite
advantages for testing a surface disinfectant, the surfaces mentioned here do not adequately
represent the absorptive surfaces of hydrophilic impressions, which could retain and possibly
protect viruses from the action of a disinfectant. In this model VSV was chosen as the test
virus because it is the most studied enveloped virus and is often used for the study of viral
disease. Klein and Deforest3i classify VSV in the same group as HSV, HIV, and other viruses
that are characterized by a lipid-protein envelope and are predicted to have a high degree of
susceptibility to the chemical action of germicides. Although enveloped viruses may differ
considerably as to the ribonucleic acid (RNA) or deoxyribonucleic acid (DNA) in their cores, the
remaining chemical constituents (protein, lipid, and carbohydrate) and the overall proportions
of these constituents in the envelopes are fairly similar for most major groups. Some viruses,
such as HSV and the RNA tumor viruses, demonstrate greater membrane (envelope)
complexity than VSV does, but there is no evidence to suggest that the basic membrane
structure in a complex membrane is different in principle from that of simpler viral
membranes.32 It would, therefore, be expected that the reaction of VSV with germicides
would parallel that of other lipophilic enveloped viruses. In their 1963 study, Klein and
Deforest33 reported that 2 % glutaraldehyde inactivated in 1 minute or less their test viruses,
which included three lipophilic enveloped viruses. They also found NaOCl(200 ppm or 0.02 % )
and iodophor (75 to 150 ppm Is) broadly virucidal against all of the test viruses in 3 to 10
minutes. Klein and Deforest did their disinfectant tests on suspensions of virus instead of on a
surface, and with organic material present in their experimental model. It is apparent from
Table I that these germicides show the same patterns of reaction with VSV on an irreversible
hydrocolloid surface. It follows that broadspectrum germicides may be essentially as efficient
against viruses on an irreversible hydrocolloid surface as they are in suspensions of virus,
when organic soil is present in both cases. This hypothesis should be tested with other viruses,
including a hydrophilic nonenveloped virus. If this hypothesis is demonstrated to be true, then
the major challenge to the biocidal action of virucides on an irreversible hydrocolloid surface
is seen not as a problem of imbibition, but of the interfering organic material adhering to the
irreversible hydrocolloid surface. Of the disinfectants tested in this study, only glutaraldehyde
is not significantly affected by the presence of organic soil.15* 22 Irreversible hydrocolloid
itself, may inactivate some of the viral contamination. In this study, VSV, which is considered
to be relatively pH stable, incurred a 60% loss in PFU/ml as the pH of the virus recovery
inocuhrm dropped from 7.25 to 6.25 because of the low pH (4.6) of the irreversible
hydrocolloid.34 If the pH of irreversible hydrocolloid could totally inactivate adherent
infectious material, it would be of great clinical importance. This not being the case, the pH of
the irreversible hydrocolloid was, in the context of disinfectant testing, a confounder. It
obscured the quantitative relationship between germicides and surviving virus counts, and
therefore was eliminated from the test model. Focusing on NaOCl and iodophor, which are
both widely used in dental offices, an important question relates to the observed range and
variance in virus counts following a brief spray and rinse (Table II). It was necessary to
determine whether this variance was due to the unreliability of brief spray or to variation in
the test methodology. Of particular concern was a possible variation in the loss of virus and
organic soil during the first rinse step. This step was designed to simulate the technique that
dental clinicians have practiced for 50 years; that is, to remove blood and saliva from an
impression by gently rinsing it in cold water. Variation in this step would have had its greatest
effect on the outcomes recorded in Table II. Viral inactivation initially follows an exponential
curve with the rate of inactivation sometimes decreasing significantly as the endpoint is
approached. 35 The number of survivors present at any point during the exponential phase is
directly proportional to the initial virus titre. However, the number of survivors present during
endpoint kinetics is a function of other factors such as the organic soil present. In this study,
more than 99.5% of the virus was inactivated in seconds by each of the disinfectants, but the
question remained as to whether the results associated with a brief spray technique were
proportional to the initial virus titre. This question was addressed by using a second virus
stock that had 50 times more virus than the first. An analysis of the results of each virus stock
showed no statistical association. Therefore, it is unlikely that small variations attributable to
the rinse step could explain the ranges recorded in Table II. In addition, there is evidence for
the methodologic consistency in this study which includes (1) the symmetric distribution of
surviving virus counts within most cells of Table I and (2) a progressive reduction of mean
virus counts in each column up to the point of total inactivation (NVR). We conclude that the
procedure outlined for this lipophilic enveloped virus model is sufficiently reliable and
sensitive to demonstrate true differences between disinfectants and disinfection protocols.
Although caution should be exercised against the assumption that VSV is representative of all
viruses, there are essential clinical implications immediately evident from this model. Most
important is the demonstration of the effect of variables, such as organic soil, that can
constitute a challenge to NaOCl and iodophor for at least 3 minutes, and produce highly
unpredictable results. This model also demonstrates that iodophor can be considered reliable
only when it is used as an immersion disinfectant.
Much of the dental community is inclined to use short disinfection times. Watkinson reported
recently on a survey of 40 U.K. dental teaching hospitals. Twenty-five of the respondents
indicated that they routinely did no disinfection of impressions other than to rinse them in
water. The 15 respondents (37.5%) who did practice some form of disinfection predominantly
indicated disinfection times of 1 minute or less. That this would seem to be a dangerous
practice is supported by the present study. Some dental professionals may neglect
disinfection practices because of frustration over proposed disinfection times of 1 hour and
more, which appear to contradict traditionally accepted ways for handling impressions,
particularly irreversible hydrocolloid impressions. It does seem, however, that humid storage
of an irreversible hydrocolloid with the disinfectant on the surface is an acceptable alternative
to immersion in some cases,4* 6 and we find this to be a good application for NaOCl. In the
case of glutaraldehyde, dipping or immersion is strongly preferred to spraying, to avoid
inhalation of an aldehyde. We believe that the dental community needs to have a better idea
as to the extent of disinfection that actually takes place with any given disinfection protocol.

SUMMARY

This article describes an enveloped virus model for disinfection testing that is reliable and
sensitive. To the extent that it narallels the true clinical situation. the findings of this model
have immediate significance. It was found that 2 % glutaraldehyde inactivated the virus on an
irreversible hydrocolloid surface in less than 1 minute, a 0.5 % sodium hypochlorite spray was
effective in a 3- to lo-minute range, and iodophor (75 ppm Is) required a 3- to lo-minute
immersion for inactivation of the virus. It was also found that during an initial period, the
length of which depends on the disinfectant and the technique (spray versus immersion),
inactivation of viruses may be highly unpredictable. A simple spray disinfection and an
immediate rinse should not, in general, be considered an appropriate method.