Viruses are present in all water types including finished water. Bacterial indicators in general do not reflect virological water quality ( Bosch A 1998). As early as 1981 concerns were raised about the presence of viruses in drinking water. When studying treated water from a drinking water treatment plant drawing sewage-contaminated river water, Payment et al. found typically 1-10 cytopathogenic viral particles per liter drinking water, some samples even contained more than one ( Payment P 1981). No coliforms were present in these samples and residual chlorine levels were maintained. Obviously these viruses had not been inactivated or removed by water treatment procedures. This was the basis for later studies on the efficiencies of different treatment steps during the preparation of drinking water. Great differences were observed. Payment et al. monitored the elimination of viruses and indicator bacteria at each step of treatment at seven water treatment plants in Canada ( Payment et al. 1985). Whereas the overall total coliform population was reduced by 6 log10 units, the overall virus reduction was about 4 log10 units. Viruses were still detected by cell culture in 7% of the finished water samples at an average density of 0.0006 MPNCU/liter. The highest virus density measured were 0.02 MPNCU/liter. The average cumulative reduction in overall viruses was 95.15% after sedimentation, and 99.97% after filtration. Interestingly, ozonation or final chlorination did not result in a significant further reduction.
Part of the problems with viruses in treated water lies in the distribution systems. Skraber et al. point out in a review article that pathogenic viruses might become trapped in biofilms in drinking water distribution systems ( Skraber et al. 2005). There is a chance that viruses might accumulate in biofilms over time. In addition to acting as an attractive surface, biofilms might also provide a sheltered environment protecting viruses from disinfection. Problems would arise in the case that infectious viruses detached with portions of the biofilms by shearing forces. In contrast to bacterial pathogens, viral pathogens have however no potential for regrowth in water aquatic environments due to the obligate host requirements ( Fong and Lipp 2005).
Human enteric viruses are the most important viral group in terms of water contamination. This group comprises enteroviruses, caliciviruses, adenoviruses, rotaviruses, reoviruses, and others. Circoviruses, picornaviruses, and polyomaviruses are considered to be emerging pathogens ( Fong and Lipp 2005). They enter the water by fecal contamination and have been reported to be shed in the feces of infected people in extremely high numbers between 105 and 1011 viruses per gram of stool ( Farthing MJG 1989). Once in the water, they are difficult to remove as they can be highly resistant to standard water disinfection procedures. At the same time, the infectious doses are typically very low. In the case of rotaviruses, one infectious particle causes disease in 1% of healthy adults without antibodies to the virus ( Schiff et al. 1984). In a model to describe experimental dose-response data, Haas et al. hypothesized that the risk of infection by consumption of contaminated drinking water is 10- to 10,000-fold greater with viruses compared to bacterial pathogens ( Haas et al. 1993). The low infectious doses represent a great challenge to detection methodology in terms of sensitivity. For more information about enteric viruses we refer to a detailed review by Fong and Lipp ( Fong and Lipp 2005).
Traditional detection methodology is based on cell culture. The culturable virus assay or plaque assay is required in the United States for regulatory purposes ( Hwang et al. 2007). The method has the advantage that only viruses which are infectious are detected. When using cell culture, the choice of the cell line is critical for detection. Viral contamination might be underestimated if a non-optimal cell line is chosen. An example was provided by Chapron et al. when studying enteric viruses in surface waters: When using Caco-2 cells for detection of enteroviruses, only 17% of 29 samples tested positive, whereas 59% of samples tested positive when using BGMK cells ( Chapron et al. 2000). When studying a larger group of viruses, such as enteric viruses, the combination of different cell lines might be advantageous as some viruses might not reproduce in a single specific cell line. The combination of BGMK and A549 cells increased the frequency of simultaneous enterovirus and adenovirus detection in 40 river water samples from 50 to 65% compared to detection using BGMK cells alone ( Lee et al. 2004a).The higher detection rate was mainly due to the fact that adenoviruses show a higher cytopathogenic effect with A549 cells than with BGMK cells. A wider range of cell lines thus increases the range of virus types which can be detected and results in higher assay sensitivity. Another approach to achieving higher sensitivity is the use of genetically engineered cell lines. Huang et al. reported BGML cells expressing human decay-accelerating factor (hDAF) which had been reported to be a cellular receptor for several enteroviruses ( Huang et al. 2002). The engineered cell line was reported to show increased susceptibility and sensitivity to several enterovirus types compared to the wild-type cell line. To improve viral detection even further, the BGMK-hDAF cells were used in a mixed monolayer with CaCo-2 cells.
Despite the widespread use of cell culture techniques and their ability to detect infectious particles, novel detection methods would be highly desirable. The main reason, apart from the analysis costs and labor intensity, is the great time demand for analysis, which does not allow preventive measures in case of positive results and defeats the purpose of the most important aspect of the monitoring effort. The use of molecular tools holds great promise to significantly shorten the analysis time although the necessity to specifically detect infectious particles still requires the combination with cell culture. Detection which is solely based on the amplification of nucleic acids does not provide any information about infectivity and tends to overestimate the health risk. The results from detection of genome copies have therefore to be interpreted with care as the numbers of PCR detectable units (PDUs) does not necessarily reflect the numbers of infectious particles. Numbers of infectious particles can be notably lower than the number of PDUs. This problem of assessing infectious risk based on the presence of genomes was addressed by Gassilloud et al. using Poliovirus 1 (PV1) and Feline calicivirus f9 (FCV-f9) as model pathogens ( Gassilloud et al. 2003). Studying the survival of these viruses, it was shown at different temperatures that the infectious particles in mineral water degraded more rapidly than viral genomes. At 35°C, Poliovirus 1 infectivity was reduced by 4 log10 units after 19 days, whereas it took 75 days to detect an equivalent reduction based on detectable genomes. Contradictory survival data were reported dependent whether survival is measured by infectivity or by viral genomes. Whereas PV1 was more resistant to elevated temperature than FCW-f9 based on infectivity measurements over time, the opposite was suggested based on the enumeration of genomes as a function of time. The lack of correlation between presence of viral genomes and viral infectivity demands cautious interpretation of molecular data. Despite the fact that PCR detection of viral sequences cannot be equated with epidemiological risk, it can inform about the history of viral presence of a certain sample.
Results of studies on viral survival in water are greatly influenced by experimental conditions and the water source being used. The presented examples thus only provide a rough idea about viral persistence in water. Survival is influenced by a wide variety of factors including the properties of the virus itself as well as abiotic and biotic parameters. Abiotic factors affecting viral persistence include temperature, dissolved oxygen, pH, nutrient concentrations, and the presence of solid particles. Some of those factors might have an indirect effect as they also influence the persistence of indigenous microorganisms which also have been shown to influence viral survival ( Yates et al. 1988; Gordon and Toze 2003). Temperature seems to affect viral persistence and survival to a greater extent than it does for coliform bacteria. Survival and susceptibility to disinfection can be greatly impacted by adhesion of viruses to solid organic and inorganic particles in the water. Attachment to suspended material and also to cells results in most cases in higher persistence and resistance, although the effects appear to be virus-dependent. Whereas adsorption to solids leads to a protective effect for hepatitis A and poliovirus, the opposite seems to hold true for bacteriophage PRD-1 ( John and Rose 2005). Examples are given in the individual pathogen sections. In addition to affecting survival patterns, adsorption to suspended solids can also affect hydrotransportation. Viral particles can be carried over long distances in natural waters. Schaub and Sagik mention that it is not uncommon that detectable viral particles are carried over several miles in fresh and estuarine waters ( Schaub and Sagik 1975). Another important factor influencing viral persistence is the presence of the indigenous water flora. Survival in non-sterile conditions tends to be significantly shorter than under sterile conditions.