Evaluation of Chlorine Dioxide in Potable Water Systems for Legionella Control in a Acute Care Hosptial Environment

The Johns Hopkins Hospital in 1999 began to evaluate the safety and efficacy of chlorine dioxide for Legionella control in a potable water system in an acute care hospital environment. The evaluation includes chlorine dioxide, its by-products and their impact on Legionella and pathogenic bacteria, biofilm, medical and laboratory filtration systems, corrosion rates and the environment. The application of a delivery system was also evaluated to optimize the efficacy of the disinfectant while maintaining affordable installation, operational and maintenance costs. Installation and operation challenges are discussed and solutions presented. Extensive Legionella and pathogenic bacterial culturing and analysis performed during this study are presented. The data clearly shows the effect of chlorine dioxide on Legionella and pathogenic bacteria. Maryland Department of the Environment (MDE) permit requirements and the hospital’s approach to compliance and monitoring requirements as mandated by EPA are outlined in the paper. The paper also discusses the application of chlorine dioxide for remediation of potable water systems (cold and hot) contaminated with bacteria. The successful remediation process is described. INTRODUCTION In 1999, prior to the June 14, 2000 Report of the Maryland Scientific Working Group to Study Legionella in Water Systems in Healthcare Institutions, Hopkins assembled a Legionella task force. The goal of the task force was to develop and implement prevention and control measures to minimize the risk of nosocomial Legionella and to control and eliminate Legionella in the hospital potable water systems. The Legionella task force is comprised of staff from Facilities Engineering, Department of Medicine, Division of Infectious Diseases, Department of Pathology, Microbiology Laboratory, Hospital Epidemiology and Infection Control, Health Safety and Environment, Nursing, Hospital Administration, Public Affairs, Legal and Risk Management. An engineering team was assembled to research, explore and implement methods to control and monitor Legionella in the hospital potable water system. The engineering team is comprised of staff from The Johns Hopkins Hospital Facilities Engineering, Water Chemical Service, Inc. and Legionella Risk Management. The focus of this paper is on the work of the engineering team and their results. METHODS Disinfection Selection The hospital researched available methods to disinfect hot and cold potable water systems. A selection criteria was then developed to select a disinfection method. The hospital required that the disinfection method must be Environmental Protection Agency (EPA) approved for potable water disinfection. Additional information necessary to select a disinfection method included: impact on biofilm, residual effect, by-products, environmental and health effects, impact on equipment and piping, impact on dialysis and laboratory equipment impact on organoleptic properties of treated water. After extensive research of current methods for controlling Legionella the hospital selected chlorine dioxide (ClO2), which is approved by the EPA for use as a potable water disinfectant under CFR Part 141 – National Primary Drinking Water Regulations. ClO2 is a gas that can be generated chemically or electrolytically from a sodium chlorite solution. Sodium chlorite is approved by the EPA (EPA registration number 5382-43) as a precursor for generating ClO2 as a potable water disinfectant. ClO2 is a powerful oxidant and kills bacteria via oxidative disruption of cellular processes. 2 There are few reports on the use of chlorine dioxide to specifically remove Legionella from hospital water supplies despite the fact that it has been used for many years in industrial and municipal water systems. , , 4 5 6 Available limited published data and experience with ClO2 indicate it is effective. Additionally, there is little published data documenting any adverse health effects in humans associated with ClO2 and its by-products. After reviewing available data on currently available ClO2 systems, a new state of the art ClO2 generation technology system was selected. The system converts 25% sodium chlorite into nearly pure ClO2 by utilization of an electrochemical cassette oxidation process (Halox Inc, Bridgeport, CT). The ClO2 equipment and associated delivery system for this study was operated and maintained by the hospital’s contracted water treatment company. Study Site The hospital receives potable water from the local city municipality operated by the Baltimore City Public Works. BCPW is responsible for compliance with the EPA drinking water regulations. They are required to ensure that the contaminants do not exceed Maximum Contaminant Levels (MCL’s) where applicable. In addition to regulated MCL’s the EPA has published Maximum Contaminated Level Goals (MCLG’s) for Legionella as well as other contaminants. Municipalities are not required to meet EPA MCLG’s. BCPW currently utilizes chlorine; an EPA approved potable water disinfectant. The water provided meets all federal and state standards for potable water. A newly constructed building was selected as the study site to evaluate chlorine dioxide. The Weinberg Building of the Johns Hopkins Hospital is a 600,000 sq. foot, 154 bed facility that houses surgical and oncology patients, including bone marrow transplant patients and patients requiring hemodialysis. In addition to general patient care floors and an intensive care unit, the building also houses 16 operating rooms, and surgical pathology, laboratory and sterile processing facilities. When construction plans were made it was decided that a water treatment system would be installed. This decision was based on historic problems with Legionella in the hospital water system and of the patient population occupying the building who have a high risk factor for nosocomial Legionella infections. Two, six inch mains serve the building from the BCPW’s 40 psig municipality water main. Once in the building the water pressure is increased to 95 psig by two 400 gallon per minute booster pumps. The water main then splits and serves both the potable cold and hot water for the building. There are two semi-instantaneous hot water generators that provide 120 degree F hot water to the building. The building’s cold and hot water piping systems are predominantly constructed of copper pipe with extended branch lines off the water mains. The water distribution mains are looped on each floor and have separate by-pass piping with valves to decouple each wing of the building for future repairs, maintenance or renovations. Installation of the water system was completed, filled and chlorinated in April 2000; however, it was minimally utilized until the building was occupied in September 2000. When fully occupied, the building utilizes an average of 140,000 gallons of water per day, equating to 980,000 gallons of water per week. An average flow is 96 gpm with maximum water flow at 170 gpm. Water flow data is obtained utilizing ultrasonic and insertion flow meters. The hot and cold water piping distribution systems, connected fixtures and equipment were extensively investigated to compare pipe sizing with design verses actual flow rates. This was necessary to identify possible design and operational deficiencies such as “dead –legs”, oversize piping, etc. that could impact the effectiveness of the ClO2 system. Delivery System Minimal information was available related to installing and operating the ClO2 generator on a potable water system. It was therefore necessary for the hospital to test and evaluate several engineering options implemented to maximize the effectiveness of the disinfectant. The design of the ClO2 generator system installation was finalized in November 2000. Installation of two generators associated piping, electric and controls were completed early January 2001. Based on the design of the potable water system capacity and projected water usage, two ClO2 generators, each capable of generating a 550 mg/l solution, were installed at the point of entry of the potable water distribution system. Installation included booster pumps, inductors, flow meters, corrosion coupon racks, ORP (oxidation and reduction potential) monitor, ClO2 monitor, pressure gauges, computer monitoring system and associated electrical and piping connections. The ClO2 generator system was provided with 3 factory external safety alarms and multiple internal alarms. The three external alarms are chemical, leak, and flow. Additional safety alarms were added; ClO2 level and high limit, loss of electrical power to monitors and loss of electrical power to ClO2 generator system. Chlorine Dioxide, Chlorite, Chlorate Monitoring Prior to activation of the system, performance testing and monitoring protocols were established. Testing was essential to measure and compare pre and post startup of the ClO2 generator system. ClO2 and its disinfection by-products were monitored closely during pre-startup and post startup testing. Levels of ClO2 residuals were obtained and analyzed from both hot and cold-water samples from one site at the main, and on the 1, 4 and 5 floors, and randomly during Legionella sampling. ClO2 was measured with a wavelength specific spectrophotometer utilizing DPD glycine chemistry. ClO2 (disinfection by-products, chlorite and chlorate) were evaluated using both ion chromatography (EPA method 300) and amperometric titration methods adapted from the EPA Standard Methods. 9 When the ClO2 generator system was activated for continuous operation, levels of ClO2 were measured continuously with the ClO2 monitor and daily DPD test to ensure levels did not exceed 0.8 mg/l. ClO2 and chlorite were also measured throughout the building after continuous activation of the system. EPA standards for chlorate levels in potable water do not currently exist. Chlorate was not measured during continuous operation of the ClO2 system. Dialysis and Laboratory Filtration Equipment No data was available for ClO2 and its disinfectant byproducts on hemodialysis filtration equipment (carbon and reverse osmosis filtration) and laboratory filtration equipment (carbon and demineralizer filtration). Extensive testing was conducted to ensure performance of hemodialysis and laboratory equipment prior to continuous operation of the ClO2 generator system. ClO2 was introduced at various levels into the hemodialysis and laboratory filtration equipment. Chlorine, ClO2, chlorate and chlorite levels were measured at each stage of filtration. A water meter was utilized to record total water usage in gallons through the filtration systems. Carbon filters were utilized to remove oxidants and disinfectant by products. Water for the hemodialysis units is filtered through two external carbon tanks piped in series. The tanks are 10" x 35" with 0.75 cu. ft. of 12x40 mesh Granular Activated Carbon (Norit, acid/washed, low fines granular activated carbon, Norit Americas Inc, Atlanta, GA) #20 flint, underbedding (approx. 4 inches) with 14 inches freeboard. The tanks are design at a flow rate of 3.5 gpm (continuous)/ 5 gpm max., with an operating flow rate of 1 gpm. The water then passes through a portable reverse osmosis unit (Mediport P.B., Better Water, Inc. Smyrna, TN). Water through this unit is pre-filtered through a carbon cartridge. The cartridge is a 0.125 cu. ft. of 20x50 mesh granular activated carbon, acid washed, Minimum iodine #1000. The water is then post filtered by a 10 inch spun wound 5.0 micron sediment filter before passing through the reverse osmosis membrane. For this application, no historical data was available regarding carbon tank capacities. As a safety precaution, the carbon tanks for portable dialysis equipment were sized to achieve 10 minutes EBCT (empty bed contact time) as required for monochloramines. At the end of testing each day, each carbon tank was backwashed separately to prevent channeling. The tank was backwashed to drain for five-minutes to stir up the carbon. Then the tanks went through a five-minute rinse cycle to slowly reset the carbon, completing the backwashing of the tanks. Water for the lab filtration equipment passes through one carbon filter (Neu-Ion OA6 carbon tank, Neu-Ion Inc., Baltimore, MD) and two demineralizer filtration tanks for laboratory water pretreatment. The carbon bed is sized for two minute EBCT. The lab filtration system was not modified in any way before, during or after the study. No extra precautions or measures (such as backwashing) were taken prior to, during or after the study. Corrosion Monitoring Minimal information was available related to ClO2 and corrosion particularly in a hospital environment. Corrosion monitoring was performed using standard copper and mild steel coupons placed in bypass racks. Coupons were placed in the potable water supplied to the building from the city, prior to treatment, treated domestic cold water, and treated domestic hot water. The coupons placed upstream of the treatment system served as control. Coupons were also placed in hot and cold domestic water systems of other nearby buildings for baseline monitoring purposes. Legionella and Bacteria Monitoring The hospital tested and evaluated the water quality in the building over a 46 month period. Samples of both hot and cold water were taken from an average of 28 sites during multiple stages of the project and numerous distal sites throughout the building on a regular basis. Samples were obtained quarterly except during testing of equipment and water events, when the sampling frequency was increased. Samples were obtained to assess all patient-care floors in both clinical and non-clinical areas where faucets were used both frequently and infrequently. Samples were also obtained at sites that Legionella would most likely colonize, such as areas with “dead legs”, extended branch piping and areas with inadequate hot water return, etc. At each site, faucets were opened and allowed to run for 30 seconds before the sample was collected. Aerators were removed at some of the sites to evaluate the impact of ClO2 on the devices. Samples were also obtained from the two backflow prevention devices where potable water enters the building, the potable water main after ClO2 treatment, a dedicated coldwater faucet at the end of the water distribution system, each hot water generator and the hot water return main. For each sample collected, direct and concentrated cultures were performed. The direct culture consisted of plating 100μl of water directly onto three separate plates of selective media for Legionella. The three plates used contained buffered charcoal yeast extract (BCYE) with PAV (polymixin B, anisomycin, and vancomycin), BCYE with DGVP (dye, glycine, vancomycin and polymixin B) and BCYE Legionella selective agar (vancomycin, colistin, and anisomycin) (Becton Dickenson, Sparks, MD). The concentrated culture consisted of filtering 50 ml of the original sample through a polycarbonate filter (Whatman, VWR Scientific, West Chester, PA). The filter was then placed into 5ml of the original, unfiltered sample and vortexed. Next, 100μl aliquots were then plated onto each of the three plates. All plates were incubated in CO2 at 37° C within a moist chamber for 7 days. Colonies suggestive of Legionella were sub-cultured on blood agar and BCYE plates. Organisms that grew on BCYE but not on blood agar were identified as Legionella species and were then speciated using direct fluorescent antibody reagents (m-TECH, Alpharetta, GA) and the gas liquid chromatography SherlockTM Microbial Identification System (MIDI Inc., Newark, DE). Legionella and gram–negative culture data was evaluated. Legionalla data was evaluated at 10 org/ml (typical action level) and total positive Legionalla sites. Remediation Disinfection Methods Commonly used methods to remediate potable water systems in healthcare facilities were evaluated. Hyperchlorination and super heating of the potable water system were utilized and their impact on Legionalla and bacteria evaluated. No information was available utilizing ClO2 to remediate potable water systems. Therefore, it was necessary for the hospital to develop a ClO2 shock remediation treatment method and evaluate its impact on Legionalla and bacteria. Patient Surveillance for Legionella The hospital performs active clinical surveillance for Legionella infections. All bronchoalveolar lavage samples taken from in-patients are cultured, and those who have evidence of a lower respiratory tract infection are routinely cultured for Legionella. All cultures that grow Legionella and all positive urinary antigen tests are reviewed by infection control staff to determine if the case is nosocomial. Patients who have culture confirmed Legionella infections up to nine days after admission are considered “possible” nosocomial cases and any patient who has a confirmed infection more than nine days after admission is considered a “definite” nosocomial case. Any case in which the patient’s isolate matches an environmental sample by pulsed field gel electropheresis is considered a definite nosocomial case regardless of the incubation period. RESULTS Chlorine Dioxide, and Chlorite Levels Prior to January 2004, the EPA maximum residual disinfectant level goal (MRDLG) for ClO2 was 0.8 mg/l, and maximum contaminant level goal (MCLG) for chlorite was 1.0 mg/l. In January 2004 new EPA guidelines extended ClO2 and chlorite regulations to small municipalities. Current EPA maximum residual disinfectant level (MRDL) for ClO2 is 0.8 mg/l, and maximum contaminant level (MCL) for chlorite is 1.0 mg/l. Between January 2001 and June 2002, ClO2 and chlorite residuals were monitored throughout the building. The generator system had been set to maintain an average of 0.7 mg/l of ClO2 and maximum not to exceed of 0.8 mg/l. ClO2 in the potable water system. This was confirmed as a measured level downstream of the disinfectant induction point. ClO2 and chlorite residuals did not exceed the EPA maximum residual disinfectant level goal (MRDLG) of 0.8 mg/l and maximum contaminant level goal (MCLG) 1.0 mg/l respectively. Free chlorine residuals entering the building ranged between 0.71 mg/l and 1.28 mg/l. Total chlorine residuals ranged between 0.93 mg/l and 1.62 mg/l. ClO2 residuals in the mechanical equipment room downstream of induction ranged between 0.28 mg/l 0.79 mg/l. ClO2 and chlorite residuals in cold water averaged at the distal sites between 0.23 mg/l – 0.79 mg/l and 0.22 – 0.68 mg/l respectively. ClO2 residuals in hot water averaged at the distal sites between 0.1 mg/l – 0.2 mg/l. In July 2002, the generator system was adjusted to maintain an average ClO2 residual of 0.5 mg/l in the potable water and maximum ClO2 not to exceed of 0.8 mg/l. ClO2 residuals in the mechanical equipment room downstream of induction averaged between 0.28 mg/l 0.5 mg/l. ClO2 residuals in cold water averaged at the distal sites between 0.11 mg/l – 0.31 mg/l. ClO2 in hot water averaged at the distal sites between 0.05 mg/l – 0.2 mg/l. Legionella and Bacteria Evaluation of Legionella and bacteria was divided into three phases: • Phase I July 2000 until November 2000, pre and post hyper-chlorination, thermal remediation. • Phase II December 2000 until May 2001, ClO2 system installed, intermittent introduction of elevated residuals of ClO2 in the potable water system during testing of filtration equipment. • Phase III June 2001 until July 2004, continuous introduction of ClO2 residuals below a 0.8 mg/l in the potable water system. Legionella cultures in charts 3, 4, 5 and 6 in the appendixes indicate total positive Legionella sites and 10 org/ml (typical action level). The data below in Phase I through Phase III references total positive Legionella sites in percent. Phase I data collected between August 31, 2000 and October 18, 2000 indicated no significant reduction in the total gram-negative bacteria sites in both the hot and cold water. The number of total positive Legionella sites decreased. No significant change was noted in positive Legionella sites in the cold water. ∗∗∗ However, a reduction in positive Legionella sites in the hot water was observed. Phase II data collected between January 16, 2001 and February 15, 2001 indicated no detectable gram-negative bacteria in either the cold and hot water systems.∗ The total number of positive Legionella sites reduced slightly in the cold and hot water systems. Phase III data collected between June 5, 2001 and July 6, 2004 indicated an initial increase followed by ∗ See Chart 1 in Appendixes. ∗∗ See Chart 3 in Appendixes. ∗∗∗ See Chart 5 in Appendixes. ∗∗∗∗ See Chart 6 in Appendixes. ∗∗∗∗∗ See Chart 5 & 6 in Appendixes. a gradual decrease to non-detectable levels of Legionella and gram–negative bacteria.∗ These results were obtained even with significant spikes of Legionella and gram-negative positive sites occurring around the same period of several disruptions to the buildings potable water service. The first period of water disruptions occurred between September 2001 and October 2001. Brown water and sediment was introduced into the building potable water system. During this period, gram-negative sites increased significantly. No change of positive Legionella sites was observed. Remediation was not implemented. The second period of water disruptions occurred between October 2002 and January 2003. Brown water and sediment was introduced into the building including the loss of water pressure in the potable water service. January culture results indicated, an increase in gram-negative bacteria sites. Legionella was not detected in the cold water while Legionella positive sites increased in the hot water. Again, remediation was not implemented. The third period of water disruptions occurred between May 5, 2003 and October 9, 2003. Brown water and sediment was introduced into the buildings potable water system. Also, on September 19, 2003 the region experienced heavy rains associated with hurricane “Isabel” which may have impacted water quality in the region. Water samples obtained on September 24, 2003 and October 6, 2003 indicated a significant increase of gram-negative sites and positive Legionella sites in the potable hot water. Most importantly, were the positive cultures of L. pneumophila, which up until the September 24, 2003 samples had been L. anisa. Legionella was not detected in the cold water. Water samples obtained on October 10, 2003 after additional disruptions to the potable water service, indicated a further increase of positive Legionella sites in the potable hot water system. Positive cultures of L. pneumophila were detected in both ∗ See Charts 2 & 4 in Appendixes. ∗∗ See Chart 2 in Appendixes. ∗∗∗ See Charts 5 & 6 in Appendixes. ∗∗∗∗ See Charts 2 & 6 in Appendixes. ∗∗∗∗∗ See Chart 5 in Appendixes. ∗∗∗∗∗∗ See Chart 6 in Appendixes. the cold and hot water systems. Gram-negative bacteria in the potable cold and hot water systems also increased. In response to disruptions, the hospital developed and implemented a flush and ClO2 shock remediation treatment. Also included was the cleaning of faucet aerators, showerheads and back flow prevention devices. During the cleaning of the back flow prevention devices (which are located in the two six inch potable water mains entering the building) large amounts of sediment was found in the strainers (Figure 1).