Shedding UV light on the Cryptosporidium Threat.

Abstract Protecting the public from waterborne outbreaks of cryptosporidiosis requires barriers beyond those provided by filtration. While sequential chemical disinfection has shown promise, concern over disinfection by-products may limit its use. This paper re-examines past evaluations of the effect ultraviolet light has on Cryptosporidium oocysts and reviews recently generated data on the topic. The studies demonstrate that ultraviolet light could have promising effects on Cryptosporidium in drinking-water applications.

Editor's note: Through NEHA's long-standing and excellent relationship with NSF International, NEHA was granted permission by NSF International to share with the Journal's readership various papers that were presented January 12-15, 2000, at the "NSF International Small Drinking Water and Wastewater Systems International Symposium and Technology Expo" in Phoenix, Arizona. This paper "Shedding UV Light on the Cryptosporidium Threat," is one of them.

It is important to note that these papers were screened by an NSF International advisory committee prior to their presentation at the conference, but they have not been peer reviewed by NEHA's Journal program for technical accuracy.

Because these papers contain useful and interesting ideas and information that may be either delayed or lost if the papers were sent through the Journal's normal peer review process, NEHA has decided to publish them as presented, with only minor editorial modifications.

We hope you look forward to more of these papers in future issues of the Journal!

Cryptosporidium Control Through Water Treatment

The public health effects of waterborne Cryptosporidium oocysts are well recognized. As a response to this health threat, significant public policy changes have been made to regulate and control the presence of this pathogen in drinking-water supplies. Like Giardia, Cryptosporidium oocysts can be controlled by physical removal through filtration processes, although oocyst removals can be expected to be lower than for Giardia, because of their smaller size. Cryptosporidium removal has been assessed with a variety of treatment techniques, and removals range from 2 to 4 log in conventional systems (Ongerth and Pecoraro, 1995; Nieminski and Ongerth, 1995; Plummer et al., 1995; Hall et al., 1995). Diatomaceous earth filtration was shown to provide 3.8- to 6-log oocyst removal in bench-scale studies (Ongerth and Hutton, 1997). Membrane processes (ultrafiltration and microfiltration) have been shown to provide high levels (greater than 6 log) of oocyst removal (Jacangelo et al., 1995).

Unlike Giardia, however, Cryptosporidium oocysts that escape the filtration process are resistant to chlorine-based disinfectants at the concentrations and contact times practical for water treatment (Korich et al., 1990). This makes the physical removal process (coagulation, sedimentation, filtration) the most critical step in conventional water treatment in plants using chlorine for disinfection. Alternative disinfectants do exist.

Ozone is highly effective for Cryptosporidium control. Korich et al. (1990) demonstrated that exposure to ozone at 1 milligram per liter (mg/L) reduced oocyst viability from 84 percent to 0 percent after five minutes at 35[degrees] Exposure times of five and 10 minutes resulted in 90 percent to 99.9 percent reduction in neonatal mouse infectivity. These data are further supported by Finch et al (1993), who also used ozone to treat oocysts and found it to be highly effective for oocyst inactivation. Further work by Finch et al. (1997) has shown that a synergistic effect occurs with combinations of certain disinfectants, with a higher-log inactivation of oocysts occurring when the chemicals are applied sequentially than when each one is used individually. For example, an initial residual of 2.0 mg/L chlorine for 240 minutes resulted in a 0.4-log inactivation of Cryptosporidium; an ozone dose with an initial residual of 0.75 mg/L for 3.7 minutes resulted in a 1.6-log inactivation; but treating the oocysts with ozone followed by chlorine resulted in a 2.9-log inactivation. Other combinations--ozone/monochloramine (1.8-log inactivation), chlorine dioxide/free chlorine (2.9-log inactivation), and even free chlorine/monochloramine (0.6-log inactivation)--have proved more useful than each oxidant applied individually. The discovery that sequential disinfection is effective provided the first real option for water suppliers to provide a final barrier to Cryptosporidium after filtration. Concerns exist, however, about the health effects of chemical disinfection by-products.

Studies of UV Light Effects on Cryptosporidium

Ultraviolet (UV) light may be the newest addition in the war against Cryptosporidium, but it was almost overlooked in the search for ways to inactivate oocysts. In retrospect, it appears that the early discovery of oocyst resistance to chlorine-based disinfectants left us with the prejudice that oocysts were extremely resistant to anything and everything. A review of the literature, however, indicates that the work done with UV light was positive but was not immediately capitalized upon as data became available. Only recently has the scientific community begun to accept that UV light may be a highly effective tool for Cryptosporidium control. The history of Cryptosporidium inactivation with UV light is brief, as is the review of the published literature that follows.

Lorenzo-Lorenzo et al. (1993) used mouse infectivity experiments in a bench top system to assess inactivation of Cryptosporidium parvum oocysts. There is a lack of clarity in the experimental details, and the paper is difficult to interpret. Dr. Andrew Campbell communicated directly with the authors to ascertain the details and interpret the findings (Andrew Campbell, personal communication), Lorenzo-Lorenzo used a single inoculum (2.5 X 104 oocysts) and assessed overall infection intensity after treatment by using low-pressure UV light. The results estimated a greater than 3-log inactivation at 100-300 microjoules per square centimeter (mJ/[cm.sup.2]), although these data cannot be gleaned from the published paper.

Following this work, Campbell et al. (1995) examined the Safe Water Solutions, Ltd., Cryptosporidium Inactivation Device (CID) designed for oocyst inactivation in clean water (less than 1 nephelometric turbidity unit [NPU]). These researchers demonstrated 2- to 3-log inactivation of oocysts using the 4'-6'-diamidino-2-phenylindole and propidium iodide (DAPI/PI) vital dyes assay and excystation with a low-pressure UV dose of 8,700 mJ/[cm.sup.2]. The researchers postulated that their ability to assess the limits of inactivation was restricted by the limits of the standard enumeration procedures that are used with in vitro assays. Their study also tested the CID in a static mode rather than the flowing mode for which it is designed. Therefore, they recommended a definitive study using mouse infectivity and operating the unit in the flow-through mode matching its design.

In 1996, the American Water Works Association Research Foundation (AWWAPF) and the Electric Power Research Institute/Community Environmental Center (EPRI/CEC) jointly funded a study to assess innovative electrotechnologies for inactivation of Cryptosporidium. This research was undertaken to determine if any commercially available, innovative electrotechnologies were capable of inactivating Cryptosporidium oocysts in drinking water. Five electrotechnologies were challenge-tested on live oocysts under carefully controlled laboratory or field conditions. The five systems tested were

1. advanced UV light (represented by the CID),

2. pulsed UV light,

3. conventional UV light,

4. acoustic shock, and

5. resonant electric current.

Only advanced and pulsed UV light were shown to inactivate oocysts under the experimental conditions. In retrospect it can be seen that the study design might entirely have missed the potential of UV light to inactivate oocysts had lower UV doses been applied (Clancy et al., 1998).

In the AWWARF/EPRI study the objective was to demonstrate whether a given electrotechnology had any potential for oocyst inactivation. The study did not assess levels of inactivation, experiment with various doses of energy, or attempt to optimize any technology--it was a quick look-see to screen the various electrotechnologies for further study. Therefore, the research team decided to use the in vitro surrogate assays (DAPI/PI, SYTO(r)-9, SYTO(r)-59, or maximized in vitro excystation) to make initial measurements since these assays are simple to perform, are inexpensive, and were thought to correlate to animal infectivity. For those electrotechnologies that showed promise, additional studies would be conducted with animal infectivity thought by the U.S. Environmental Protection Agency (U.S. EPA) and North American researchers to be the "gold standard" for demonstrating loss of infectivity in disinfection trials. The vendors supplying equipment for assessment were asked to "give it their best shot aka provid e a high dose," as the assessment was an initial demonstration and the electrotechnology had to pass this hurdle for future consideration.

Two electrotechnologies appeared to be successful at oocyst inactivation--pulsed UV and advanced UV (low pressure over an extended exposure period). Pulsed UV (Innovatech, Inc.) at 1,900 mJ/[cm.sup.2] in a 10-gallon-per-minute (gpm) system provided greater than 2-log oocyst inactivation. The CID with 8,700 mJ/[cm.sup.2] at 400 gpm full scale provided greater than 4-log inactivation. Initial studies showed promise with the in vitro surrogates, and animal studies supported the results. Conventional low-pressure UV light (180 mJ/[cm.sup.2]) appeared to have no effect on oocyst viability as measured with the surrogates alone, and on the basis of data accrued with in vitro viability assays, Clancy et al. (1998) reported that it was ineffective for oocyst inactivation.

At the same time, Clancy et al. (1999) were involved in another AWWARF study--in conjunction with the U.K. Drinking Water Inspectorate (DWI)--to determine which of the four in vitro surrogate assays (DAPI/PI, SYTO-9, SYTO-59, or maximized in vitro excystation) most closely predicted results of animal infectivity. The objectives were to identify one or more in vitro surrogates that correlated well with animal infectivity, allowing disinfection research to continue with an inexpensive, faster, but equally reliable assay. The study was large, had a robust statistical design, and involved two U.S. and two U.K. labs so that identical trials could be conducted in both countries. The results of the UV trials using the Innovatech pulsed-UV system showed that all four of the surrogate assays significantly underestimated oocyst inactivation when compared with oocyst inactivation as measured by mouse infectivity. It appeared that doses as low as 40 mJ/[cm.sup.2] for fresh oocysts and 14 mJ/[cm.sup.2] for aged oocysts p rovided over 2-log inactivation by mouse infectivity, while the surrogates showed less than 0.5-log inactivation. Further work on the project has shown that this phenomenon--lack of correlation between the in vitro assays and animal infectivity--extends to oocysts exposed to ozone as well.

Meanwhile, Finch et al. (1997), examining sequential chemical disinfection, also conducted experiments with UV and used mouse infectivity to assess inactivation. The unit used was a bench-scale low-pressure system, and the researchers calculated that UV doses of 1,280 and 41,400 mJ/[cm.sup.2] were applied. The oocysts were stirred in a batch reactor--a Wheaton glass bottle with UV applied 11 cm from the glass. Finch et al. (1997) reported "no detectable loss of infectivity" but also mentioned that the "data were not comparable with oocysts exposed in thin layers in a Petri dish or membrane filter." They went on to say that their results were "consistent with those reported by others where UV is not a very effective control of cysts or oocysts." The seeming inconsistency of these results with those of Clancy et al. (1998) can easily be explained. In the Finch experiments, the UV never reached the oocysts, as glass is an effective barrier to UV. Although high UV doses, as calculated by lamp output, were though t to be delivered, in reality no UV reached the oocysts.

Clancy Environmental Consultants, in conjunction with Calgon Carbon Corporation, conducted the next work on medium-pressure UV, using bench-scale collimated-beam units followed by a demonstration project at greater than 200 gpm that used the Calgon Carbon [Sentinel.sup.TM] demonstration-scale unit. The objectives of this work were to

1. determine the UV dose required from medium-pressure lamps for 3- to 5-log inactivation of Cryptosporidium oocysts in finished water;

2. establish a dose-response curve for oocyst inactivation by using a collimated-beam apparatus at bench scale;

3. conduct demonstration-scale studies and compare oocyst inactivation data from the bench-scale studies with data obtained from the demonstration-scale studies; and

4. compare the in vitro surrogate assays with animal infectivity assays.

Oocyst viability was assessed with in vitro assays (DAPI/PI and maximized in vitro excystation) and in vivo assays (neonatal mouse infectivity). For the neonatal-mouse-infectivity assay, the bench-scale studies showed greater than 4-log inactivation at UV doses as low as 41 mJ/[cm.sup.2]; the in vitro surrogate assays showed little or no inactivation at this dose or at higher UV doses. The in vitro assays, which indicate oocyst viability, grossly overestimated the UV doses required to prevent oocyst infection in susceptible hosts. The demonstration studies, carried out under the NSF/U.S. EPA Environmental Technology Verification (ETV) program, provided results that agreed with the bench-scale results and showed that a UV dose as low as 19 mJ/[cm.sup.2] provided 3.9-log inactivation of Cryptosporidium oocysts (Bukhari et al., 1999). Recent work by Finch and Belosevic (unpublished data) using collimated-beam studies appears to support data from Bukhari et al. Further work on low- and medium-pressure UV conduct ed at Clancy Environmental Consultants with collimated-beam bench-scale studies has shown that doses in the range of 6 to 9 mJ/[cm.sup.2] provide oocyst inactivation at greater than 3.5 log both in deionized water and in filter backwash supernatant with a turbidity of 11 NTU (unpublished data). The UV exposure time was adjusted to account for the lower UV transmittance of backwash water. Additional work at Clancy Environmental Consultants with a low-pressure UV system developed by WaterHealth International, Inc., operating at 4 gpm and delivering over 100 mJ/[cm.sup.2], found that the unit achieved 6-log inactivation of Cryptosporidium oocysts (unpublished data).

The promise of significant UV disinfection of Cryptosporidium apparently comes with the additional benefit that UV does not create harmful disinfection by-products (Malley et al., 1996).

Suggestions for Further Research

Several issues relating to the biology of Cryptosporidium still must be resolved:

1. Are there differences in strains or differences based on oocyst production, processing, or storage that affect UV susceptibility?

2. Are there differences in the mouse model (different mouse strains, infectivity assays, etc.) that need to be assessed?

3. Is oocyst reactivation, as seen in bacteria, possible? If so, what are the minimal UV doses at which reactivation can be prevented?

4. Can UV disinfection of surrogate organisms (MS2 coliphage, Bacillus subtilis, etc.) be used to predict effectiveness against Cryptosporidium?

5. Can alternatives to animal infectivity studies be used to demonstrate disinfection effectiveness, allowing significantly more research to be conducted?

Issues related to engineering and operations also must be resolved:

1. Are oocysts that penetrate conventional filters shielded by persistent coagulant or other particles, thereby reducing the effects of UV irradiation?

2. How can UV systems best be monitored to ensure consistent and effective operation in varying water quality or over time?

It appears that UV may be able to control Cryptosporidium in drinking water. If this method proves to be as effective as the early studies indicate, it will constitute an additional tool that water suppliers can use.

(Adapted with permission from NSF Proceedings of the Small Drinking Water and Wastewater Systems International Symposium and Technology Expo, January 12-15, 2000, Phoenix, Arizona.)

REFERENCES [*]

Bukhari, Z., et al. 1999. Medium-pressure UV light for oocyst inactivation. J. AWWA, 91(3):86-94.

Campbell, A.T. et al. 1995. Inactivation of oocysts of Cryptosporidium parvum by ultraviolet irradiation. Water Research 29: 2583-2586.

Clancy, J.L. et al. 1998. Inactivation of Cryptosporidium parvum oocysts in water using ultraviolet light. J. AWWA 90: 92-102.

Clancy,J.L., et al. 1998 Comparison of viability and infectivity methods for Cryptosporidium in the US and UK. Proc AWWA WQTC 1998. San Diego, CA.

Finch, G.R. et al. 1993. Ozone inactivation of Cryptosporidium parvum in demand-free phosphate buffer determined by in vitro excystation and animal infectivity. J. Appl. & Environ. Microbiol. 59: 4203-4210.

Finch, G.R., et al. 1997. Effects of various disinfection methods on the inactivation of Cryptosporidium. AWWA Research Foundation Report. AWWA, Denver, CO.

Hall, T. et al. 1995. Cryptosporidium removal during water treatment using dissolved air flotation. Water Science and Tech. 31: 125-135.

Jacangelo, J. G. et al. 1995. Mechanism of Cryptosporidium, Giardia and MS2 virus removal by MF and UF. J. AWWA. 87:107-121.

Korich, D.G. et al. 1990. Effects of ozone, chlorine dioxide, chlorine, and monochloramine on Cryptosporidium oocyst viability Appl. Environ. Microbiol. 56: 1423.

Lorenzo-Lorenzo, M.J., et al. 1993. Effect of ultraviolet disinfection of drinking water on the viability of Cryptosporidium parvum oocysts. J. Parasitology 79: 67-70.

Malley, J.P. Jr., J.P. Shaw, and J.D. Ropp. 1996. Evaluation of the by-products produced by the treatment of groundwaters with ultraviolet radiation (UV) and post disinfection following irradiation. Denver, CO: AWWA and AWWARE

Nieminski, E.C. and J. E. Ongerth 1995. Removing Giardia and Cryptosporidium by conventional treatment and direct filtration. J. AWWA. 87:96-106.

Ongerth, J. and J.P. Pecoraro. 1995. Removing Cryptosporidium using multimedia filters. J. AWWA. 87:83-89.

Ongerth, J.E. and P. E. Hutton. 1997. DE filtration to remove Cryptosporidium. J. AWWA 89:39-46.

Plummer, Jeanine D. et al. 1995. Removing Cryptosporidium by dissolved-air flotation. J. AWWA. 87:85-95.

(*.) Editor's note: Because this paper was originally published in the NSF Proceedings of the Small Drinking Water and Wastewater Systems International Symposium and Technology Expo, the references are not consistent with the normal style format of the Journal of Environmental Health.