Detection and Quantification of Dental Unit Water Line Contamination by Oral Streptococci
Objective. (1) To investigate the prevalence of oral streptococci (OS) and biological indicators of water contamination by oral fluids in water from dental unit water lines (DUWs) by detection and quantification and of saprophytes indigenous to the oral cavity. (2) To test whether measurement of the total cultivable mesophilic flora (TCF), the parameter commonly used to monitor water quality in DUWs, is an effective predictor for OS contamination.
Design. Survey of 21 dental units equipped with antiretraction devices. Water samples were collected from air‐water syringes, cup fillers, tap water, and before and during the working day.
Setting. Units were from 7 public dental offices selected for convenience from among those in proximity of the microbiological laboratory.
Methods. For detection of OS, samples were plated on an enriched medium, to revitalize the organisms. Colonies were subcultured on a selective medium and biochemically identified (lower detection limit, 1 cfu/mL). For measurement of the TCF, samples were plated on a nutrient‐poor medium. Cultures with colony counts greater than 200 cfu/mL were considered to be TCF positive. The sensitivity and specificity of TCF positivity in predicting OS detection was calculated.
Results. Prevalence rates for OS contamination and for TCF positivity were, respectively, 34.4% (11 of 32 samples) and 25.0% (8 of 32 samples) for syringes, 27.8% (10 of 36 samples) and 8.3% (3 of 36 samples) for cup fillers, and 0.0% (0 of 7 samples) for tap water. OS contamination levels ranged from 1 to 6 cfu/mL. No statistically significant differences were found between samples obtained before and during the working day. TCF positivity did not predict OS contamination effectively, because of low sensitivity.
Conclusions. Given the absence of OS in tap water, the reported prevalence of OS contamination suggests that oral fluids are aspirated during dental therapy with relatively high frequency and that DUWs can potentially expose successive patients to bloodborne cross‐infections.
Received December 19, 2003; accepted June 21, 2004; electronically published April 20, 2006.
The contamination of dental unit water lines (DUWs) with microorganisms is documented by a preponderance of scientific evidence.1,2 Most of these microorganisms are aquatic bacteria suspended in the incoming water, which colonize the walls of the DUWs and form a biofilm. The initial adhesion of these bacteria is promoted by water stagnation during periods of inactivity of the dental unit, because of the characteristics of the tubes and the presence of minerals—such as calcium carbonate—deposited on water‐bearing surfaces. During periods when the unit is in use, small particles of the biofilm may break off because of the water flow and heavily contaminate the water. Human opportunistic pathogens—such as Pseudomonas aeruginosa, Legionella pneumophila, and nontubercular Mycobacterium species—have frequently been recovered from DUWs at levels that suggest a moderate risk from exposure to these microorganisms for immunocompromised patients and for dental healthcare personnel.3
Nevertheless, the criteria to assess the microbial contamination of DUWs are not based on the detection of such opportunistic pathogens, which is generally considered to be counterproductive and unnecessary, but on the investigation of the concentration of biofilm‐forming heterotrophic bacteria, a nonspecific indicator for contamination, which should be lower than a certain threshold.1,4‐7
The microorganisms that contaminate DUWs can also derive from the oral cavities of dental patients undergoing treatment. These microorganisms may gain access to the water lines by aspiration of oral fluids during the transient negative pressure that occurs when the drill stops rotating.8 A retraction of up to 1 mL of oral fluids has been reported in old equipment9,10 and, under experimental conditions, in some new dental units.11‐13 This aspiration can be limited by antiretraction valves fitted distally to hand pieces,3 but these are also subject to clogging during use.14 The primary biological fluids are saliva and other oral and pharyngeal secretions; however, in cases of invasive therapy, blood could also be aspirated. Retraction of oral secretions and blood into DUWs might be worrisome with respect to the risk of transmission of bloodborne infections to successive patients.3,6,7,10,15
DUW contamination by saliva and, possibly, blood could be investigated with the same strategy used to assess the quality of potable water; namely, by testing of DUWs to detect and quantify saprophyte microorganisms indigenous to the oral cavity and, therefore, associated with oral fluids, such as oral streptococci (OS).16 These microorganisms have two peculiar characteristics. First, their habitat is almost exclusively the oral cavity and the upper respiratory tract.17 Specifically, they can survive on other nonanimal substrates but not in some stages of their life cycle, as demonstrated by their instability and pleomorphism when they are detected in culture substrates.18 Second, all humans are colonized by these bacteria and at high levels.17,19 If we assume that the presence of OS is a biological indicator of contamination by oral secretions, their detection in DUWs suggests that either oral fluids were previously aspirated from the dental patients or that tap water was already contaminated by such fluids. If saliva and blood are truly retracted into DUWs, then the detection of OS in DUWs can be associated with the presence of bloodborne pathogens, in the event that a patient is a carrier for these microorganisms. As a consequence, the detection of OS in DUWs would support the hypothesis that the risk for bloodborne infections for consecutive patients is higher than nil.3,6,7,10,15
The few data on the frequency of recovery of OS from DUWs have shown low prevalence values, ranging between 0% and 7%.15,20,21 This low prevalence could be related to the type of dental therapy performed—namely, the invasiveness and duration of the treatment and the type of hand instrument used. However, it could also be because of the difficulty of culturing such microorganisms, which live under adverse conditions. In fact, in cultures of water samples collected from DUWs, colony counts of cultivable bacteria are less than 4% of the counts obtained when direct microscopic observation is used; therefore, it has been proposed that these microorganisms be classified as viable but not cultivable.21,22
The present study sought to investigate the prevalence of OS in DUWs using a 2‐step culture method (ie, revitalization and isolation). In addition, a secondary aim was to test whether the parameter most commonly used to detect microbial contamination of water from DUWs—that is, detection of mesophilic, heterotrophic, cultivable flora at a concentration greater than a given threshold—could also be an effective predictor for the detection of OS.
Methods
Water Samples
Water samples were collected from 21 dental units from 7 multichair public dental offices (3 chairs per office), in the city of Rome, where the almost exclusive activity was conservative dentistry; this is a branch of dentistry that focuses on restoring decayed teeth after the removal of carious lesions. This option was chosen because, in conservative dentistry, the use of the high‐speed drill, which is potentially responsible for the retraction of oral fluid into DUWs,3 is frequent and prolonged. The dental offices were selected for convenience among those in close proximity to the laboratory where the microbiological analyses were performed. The units were chosen from among those that had been in service from 1 to 7 years and equipped with antiretraction devices. They regularly underwent maintenance procedures, according to what was determined by the dentists responsible for the offices. The various dental units and instruments were of the most common brands in the national and international dental market. The units were conventionally connected to municipal water, in which the level of residual chlorine ranged between 0.08 and 0.25 mg/L. The day before water collection, at the end of the working day, the dental assistants or dentists were asked to purge the DUWs by means of a 10‐minute flush; this duration is generally considered to be sufficient for a temporary reduction in bacterial load.3
Water samples (150 mL) were aseptically obtained with sterile bottles after a 1‐minute flush. The residual chlorine in the water was neutralized by sodium thiosulphate at a final concentration of 0.01% (wt/vol). Water was not collected from the tube of the high‐speed drill, which is supposed to be the main route of DUW contamination by oral fluids and, therefore, was expected to be frequently contaminated by OS.3,11,21,23 Therefore, for every unit, water was sampled from 2 different operative sites, located at progressively greater distances from the tube of the high‐speed drill—the air‐water syringe and the cup filler. For every office, water was also collected from a control, nonoperative site (ie, the adjacent tap used for hand washing and connected to the same water supply as the DUWs). At every episode of sampling, 7 water samples per office were collected. Two sampling episodes per office took place during the working day, early in the morning before starting work (time 0) and during the working period in the middle of the morning (time 1). Cumulatively, this procedure would result in 98 water samples. However, for practical reasons, only 87 samples were collected: from tap water, 7 each at time 0 and time 1; from the syringes, 18 at time 0 and 16 at time 1; and, from the cup fillers, 21 at time 0 and 18 at time 1.
Microbiological Procedures
The collected water samples were processed within 1 hour and vortexed for 3 minutes. Then, 10 mL of the sample was diluted from 10−1 to 10−5 in 0.9% (wt/vol) NaCl solution, and 1 mL of undiluted water and its dilutions were plated onto various media, for the investigation of the following parameters. (1) OS were isolated using a 2‐step method. First, microorganisms were revitalized using Columbia agar (Oxoid) supplemented with 5% (vol/vol) sheep blood (Oxoid). Plates were incubated for 3 days at 37°C in anaerobic conditions. Second, colonies were subcultured onto Mitis Salivarius agar (Becton Dickinson), a selective medium for OS, and incubated for 2 more days at 37°C in anaerobic conditions. The colonies were counted, Gram stained, observed under an optical microscope, and biochemically identified by API STREP (API System). (2) Measurement of the total heterotrophic, mesophilic, cultivable flora was performed by isolation on Plate Count agar (Oxoid). The plates were incubated at 22°C for up to 7 days in aerobic conditions.21 The samples were considered to be positive for OS if at least 1 identified colony was observed on the selective medium (lower limit of detection, 1 cfu/mL). Samples were considered TCF‐positive if the microbial concentration was greater than the American Dental Association threshold of 200 cfu/mL.6
Statistical Analysis
The prevalence of OS‐positive samples among syringes, cup fillers, and tap water at time 0 and time 1 was calculated. In the event that OS were truly aspirated into DUWs and did not come from the external water supply, we expected to detect OS‐positive samples from the operative sites (ideally, more frequently from syringes than from cup fillers) and not to detect them from the control site. To test this hypothesis, the difference in the proportion of OS‐positive samples between operative and nonoperative sites was statistically analyzed by means of the McNemar test, by matching of each sample collected from an operative site with a sample collected at the same time from the control site in the same office.
Moreover, to investigate whether the OS isolated were able to form a biofilm in DUWs, as do heterotrophic mesophilic flora, the differences in the proportions of OS‐positive samples between time 0 and time 1 for each specific site were analyzed using the same statistical test by matching of each sample collected from one site at time 0 with a sample collected from the same site at time 1. In the event that the OS isolate formed a biofilm, it was expected that the prevalence of OS‐positive samples was higher at time 0 than at time 1, because of overnight growth of bacteria and their partial removal during use of the dental unit. For comparison, the differences in prevalence of TCF‐positive samples between operative and nonoperative sites and between time 0 and time 1 for each site were also analyzed, using the same procedures described above.
In the present study, the choice to compare only matched observations resulted in a decrease in the overall number of usable samples to 82: 14 from tap water, 32 from air‐water syringes, and 36 from cup fillers. Because of the low power of the statistical analysis, which resulted from the small sample size and the use of the McNemar test, we interpreted the results cautiously.
To test whether TCF positivity was able to predict the detection of OS in cup fillers and syringes, findings for all samples (not necessarily matched) collected from these sites were classified either as true positive (TCF and OS positive), true negative (TCF and OS negative), false positive (TCF positive/OS negative), or false negative (TCF negative/OS positive). Sensitivity and specificity values were calculated, and the predictive power was considered to be effective enough if the sum of the sensitivity and specificity values was greater than 160%.24
Results
OS were isolated from 25 of 73 water samples collected from cup fillers and air‐water syringes, with an overall point prevalence of 34% (Table 1), but they were not detected in tap water (data not shown). Streptococcus salivarius was the most frequently detected species (7 isolates), followed by S. intermedius and S. anginosus (4 isolates each). A nonoral species that usually inhabits the human intestine (Enterococcus durans) was isolated once. In all positive samples, only 1 OS species was isolated, with a streptococcal load ranging between 1 and 6 cfu/mL. The prevalence of OS positivity was significantly higher among samples from the operative parts of the dental unit than among samples of tap water (Table 2).
No statistically significant differences in the prevalence of OS positivity were observed between samples collected before starting (time 0) and during (time 1) dental activity. However, although the prevalence remained unchanged from time 0 to time 1 among samples from cup fillers (28%), an increase from 25% to 44% was observed among samples from syringes (Table 2).
The overall prevalence of TCF positivity was 25% among samples from syringes, 8% among samples from cup fillers, and 0% among samples of tap water (Table 3), with medians of 81.0 cfu/mL (range, 3‐864 cfu/mL), 2.0 cfu/mL (range, 0‐744 cfu/mL), 1.0 cfu/mL (range, 0‐64 cfu/mL), respectively (data not shown). The difference between samples from syringes and from tap water was statistically significant, whereas that between samples from cup fillers and from tap water was not. Similar to what was reported for OS, no statistically significant change occurred between time 0 and time 1. However, although the number of TCF‐positive samples remained almost unchanged, increasing from 1 (5.6%) to 2 (11.1%) for samples from cup fillers, it decreased 3‐fold, from 6 (37.5%) to 2 (12.5%), for samples from syringes, showing a trend opposite to that of OS positivity (Table 2).
The effectiveness of TCF positivity as predictor of detecting OS was low; the highest value for sensitivity plus specificity was 139.5% (Table 4). The high proportion of samples with false‐negative findings, particularly among samples from cup fillers (12 [30.8%] of 39 samples), led to low sensitivity values.
Discussion
In the present study, one‐third of the water samples from the operative parts of the dental units were contaminated by OS. This prevalence value is higher than values reported by other investigators, whose findings have ranged from 0% to 7%.15,20,21 This discrepancy could be explained by the fact that, in the present study, the principal activity in the selected dental offices was conservative dentistry, which is characterized by the frequent and prolonged use of a high‐speed drill, a hand instrument considered to be the main tool responsible for oral‐fluid retraction.3,11,21,23 Another explanation that helps to reconcile the discrepancy could lie in the use of different microbiological methods among various studies. The isolation method used in the present study could result in an increased proportion of samples with false‐positive findings, compared with other studies, because of the misclassification of other bacteria as OS. This hypothesis is consistent with the finding that, in 1 positive sample, a nonoral Enterococcus species was isolated on Mitis Salivarius agar (Table 1). On the other hand, it is also likely that the isolation method used in other studies caused an artificial decrease in the prevalence of OS. In fact, in these studies, the sampled water was inoculated only on enriched medium. This medium is useful to revitalize but not to differentiate among OS species, which are biologically unstable on culture and demonstrate pleomorphism after having lived under adverse conditions. Therefore, some OS‐positive samples could have been misclassified as OS‐negative because of the atypical aspect of some OS colonies grown on blood agar plates. This hypothesis is corroborated by the fact that, in the present study, in which almost all colonies grown on blood agar were subcultured onto the selective medium, it was not uncommon that morphologically atypical colonies on the blood agar, initially misclassified as non‐OS, were further classified as OS after subculturing.
Nevertheless, the already high prevalence of OS reported in the present study could have been even higher, given that water was not collected from the tubes of the high‐speed drill and that the sample was collected after a 1‐minute flush—a time interval double the 30 seconds recommended by the American Dental Association.6 This hypothesis is corroborated by the extremely high prevalence (80% of samples with a mean OS level of 
cfu/mL) reported by Shephard et al.,23 who collected water directly from the tubes of high‐speed drills and without a previous flush. These sampling procedures were helpful in overcoming the problem of culturing OS, which were not left in difficult environmental conditions in DUWs for long periods. Therefore, the microorganisms were presumably more stable, did not require previous revitalization on blood agar, and could be plated onto the selective medium directly.
The results of the present investigation, which shows that OS were detected in water from the operative parts of the dental units and never in water from control sites (Tables 1 and 2), suggest that OS are not present in incoming water but are aspirated into DUWs during dental therapy. Therefore, it is likely that saliva, and possibly blood, are retracted into DUWs during dental activity, which corroborates the hypothesis that a certain risk for bloodborne cross‐infections between successive patients really exists.3,6,7,10,15,25
A more complicated issue would be to assess the level of such risk. Given the reported low concentration of OS in DUWs, which suggests that only a small volume of oral fluids is aspirated, this risk ought to be low. In fact, to our knowledge, no cases of parenteral transmission of infection via DUW have been reported,7,25‐28 and only 1 case of patient‐to‐patient transmission of hepatitis B virus in the dental setting has been ascertained, so that the risk for cross‐infection between patients is considered to be low.29 However, given that the route of transmission was not documented in that case, and because, among healthcare personnel, who are occupationally exposed to the risk for bloodborne infections, only about one third of hepatitis B infections can be explained,7 it has been suggested that ascertaining the route of transmission of bloodborne infections in dentistry is not easy. Therefore, in the event that sporadic cases of such infections transmitted via DUWs truly occurred, the fact that they are not documented would not be surprising.
A secondary, less important aim of the present study was to test whether measurement of the TCF, the parameter commonly used to assess water quality from DUWs, is also a good predictor of OS contamination and, thus, of the presence of oral fluids. The present data must be interpreted with caution, given the low power of statistical analysis. However, they seem to suggest that the effectiveness of measuring TCF is not high enough, mainly because of low sensitivity. Another indication of this ineffectiveness is that, in water samples from syringes, an operative site very close to the potential source of oral fluid aspiration, the rate of detection of OS almost doubled from time 0 to time 1 (from 4 of 16 to 7 of 16 samples positive), whereas high levels of TCF decreased by two thirds (from 6 of 16 to 2 of 16 samples positive). On the other hand, the situation at the cup filler was more stable, with fewer observable changes. This lack of association between OS detection and TCF positivity could be partly ascribed to the culture method used to assess TCF, which is based on the use of nutrient‐poor, nonselective media and long, room‐temperature incubation periods in aerobic conditions. Such conditions favor the development of aquatic bacteria more than oral flora.30,31 Therefore, it is plausible that, in the event that OS also accounted for a considerably high fraction of the TCF measured in water samples from DUWs, their growth on such media and under such conditions would be difficult, if not impossible.
In conclusion, the high prevalence of the detection of OS in water from DUWs suggests that the aspiration of oral fluids through dental instruments occurs frequently during dental therapy and raises the doubt that there could be a certain undefined risk for bloodborne cross‐infections during dental therapy. In the event that the level of such risk was high enough, the data suggest that the parameter commonly used to monitor the microbial quality of water from DUWs (measurement of the TCF) is not an effective indicator of oral fluid aspiration. Therefore, to assess water quality thoroughly, it would be advisable to measure levels of OS or another group of indigenous oral microorganisms, in addition to heterotrophic mesophilic flora.
References
- 1. Mills SE. The dental unit waterline controversy: defusing the myths, defining the solutions. J Am Dent Assoc 2000; 131:1427‐1441.
- 2. De Paola LG, Mangan D, Mills SE, Costerton W, Barbeau J, Shearer B, Barlett J. A review of the science regarding dental unit waterlines. J Am Dent Assoc 2002; 133:1199‐1206.
- 3. Pankhurst CL, Johnoson NW, Woods RG. Microbial contamination of dental unit waterlines: the scientific argument. Int Dent J 1998; 48:359‐368.
- 4. van der Wende E, Characklis WG, Smith DB. Biofilms and bacterial drinking water quality. Wat Res 1989; 23:1313‐1322.
- 5. European Community. Council directive 98/83/EC of 3 November 1998 on the quality of water intended for human consumption. Off J Eur Community 1998;L330:32‐54.
- 6. American Dental Association–Council on Scientific Affairs. Dental unit waterlines: approaching the year 2000. J Am Dent Assoc 1999; 130:1653‐1664.
- 7. Kohn WG, Collins AS, Cleveland JL, Harte JA, Eklund KJ, Malvitz DM, Centers for Disease Control and Prevention (CDC). Guidelines for infection control in dental health‐care settings—2003. MMWR Recomm Rep 2003;52(RR‐17):1‐61.
- 8. Pankhurst CL, Philpott‐Howard JN. The microbiological quality of water in dental chair units. J Hosp Infect 1993; 23:167‐174.
- 9. Bagga BSR, Murphy RA, Anderson AW, Punwani I. Contamination of dental unit cooling water with oral microorganisms and its prevention. J Am Dent Assoc 1984; 109:712‐716.
- 10. Lewis DL, Arens M, Appleton SS, et al. Cross‐contamination potential with dental equipment. Lancet 1992; 340:1252‐1254.
- 11. Berlutti F, Testarelli L, Vaia F, De Luca M, Dolci G. Efficacy of anti‐retraction devices in preventing bacterial contamination of dental unit water lines. J Dent 2003; 31:105‐110.
- 12. Del Nero F. Letter to the editor [Efficacy of anti‐retraction devices in preventing bacterial contamination of dental unit water lines]. J Dent 2004; 32:169‐170.
- 13. Berlutti F. Reply to the letter to the editor [Efficacy of anti‐retraction devices in preventing bacterial contamination of dental unit water lines]. J Dent 2004; 32:171.
- 14. Williams JF, Molinari JA, Andrews N. Microbial contamination of dental unit waterlines: origins and characteristics. Compendium 1996; 17:538‐550.
- 15. Walker JT, Bradshaw DJ, Bennett AM, Fulford ML, Martin MV, Marsh PD. Microbial biofilm formation and contamination of dental‐unit water systems in general dental practice. Appl Environ Microbiol 2000; 66:3363‐3367.
- 16. Hackney RW, Crawford JJ, Tulis JJ. Using a biological indicator to detect potential sources of cross‐contamination in the dental operatory. J Am Dent Assoc 1998; 129:1567‐1577.
- 17. Hardie JM, Whiley RA. The genus Streptococcus—oral. In: Balows A, Truper HG, Dworkin M, Harder W, Schleifer KH, eds. The Prokaryotes, 2nd ed. New York: Springer Verlag; 1991:1421‐1449.
- 18. Gledhill WE, Casida LE. Predominant catalase‐negative soil bacteria. I. Streptococcal population indigenous to soil. Appl Microbiol 1969; 17:208‐213.
- 19. Petti S, Tarsitani G. Intra‐individual variations of salivary microbial levels in young adults. Eur J Oral Sci 1998; 106:616‐622.
- 20. Fitzgibbon EJ, Bartzokas CA, Martin MV, Gibson MF, Graham R. The source, frequency and extent of bacterial contamination of dental unit water systems. Br Dent J 1984; 157:98‐101.
- 21. Barbeau J, Tanguay R, Faucher E, et al. Multiparametric analysis of waterline contamination in dental units. Appl Environ Microbiol 1996; 62:3954‐3959.
- 22. Barbeau J, ten Bokum L, Gauthier C, Prevost A. Cross‐contamination potential of saliva ejectors used in dentistry. J Hosp Infect 1998; 40:303‐311.
- 23. Shepherd PA, Shojaei MA, Eleazer PD, van Stewart A, Staat RH. Clearance of biofilms from dental unit waterlines through the use of hydroperoxide ion–phase transfer catalysts. Quintessence Int 2001; 32:755‐761.
- 24. Kingman A. Statistical issues in risk models for caries. In: Bader JD, ed. Risk Assessment in Dentistry, 1st ed. Chapel Hill: University of North Caroline Dental Ecology; 1990:193‐200.
- 25. Cottone JA, Puttaiah R. Viral hepatitis and hepatitis vaccines. In: Cottone JA, Terezhalmy GT, Molinari JA, eds. Practical Infection Control in Dentistry, 2nd ed. Meida, PA: Williams & Wilkins 1996:15‐47.
- 26. Yeh C‐K, Puttaiah R, Cottone JA. Human immunodeficiency virus infection, acquired immune deficiency syndrome, and related infections. In: Cottone JA, Terezhalmy GT, Molinari JA, eds. Practical Infection Control in Dentistry, 2nd ed. Media: Williams & Wilkins 1996:48‐74.
- 27. De Paola LG. Managing the care of patients infected with bloodborne diseases. J Am Dent Assoc 2003; 134:350‐358.
- 28. Leclerc H, Schwartzbrod L, Dei‐Cas E. Microbial agents associated with waterborne diseases. Crit Rev Microbiol 2002; 28:371‐409.
- 29. Daschner F. Saliva ejectors in dentistry. J Hosp Infect 1999; 42:78‐79.
- 30. Karpay RI, Plamondon TJ, Mills SE, Dove SB. Validation of an in‐office dental unit water monitoring technique. J Am Dent Assoc 1998; 129:207‐211.
- 31. Karpay RI, Plamondon TJ, Mills SE. Comparison of methods to enumerate bacteria in dental unit water lines. Curr Microbiol 1999; 38:132‐134.