Case Finding, Patient Interviews, and Exposure Assessment

CDC Legionella travel-associated case notification reports included name and address of travel accommodations, potential exposures to aerosolized water or spas, test used for diagnosis, and whether the patient survived. Cases were defined as illness in patients who met the surveillance case definition and experienced symptoms of Legionnaires’ disease within 10 days after visiting HSNP. When additional exposure information was needed, local health jurisdictions where the patient lived shared case interview forms with ADH and NPS-OPH (3 patients) or patients or next-of-kin were reinterviewed by ADH (2 patients). Additional case finding was attempted through Epidemic Information Exchange (CDC, https://emergency.cdc.gov/epix) notifications (sent August 23, 2018; January 3, July 11, August 5, and October 4, 2019), bathhouse guest notifications, and press releases. Patient data were saved on secure NPS-OPH and ADH network servers, and relevant Legionella exposure information obtained from patients or next-of-kin was entered into an Excel (Microsoft, https://www.microsoft.com) spreadsheet and shared between NPS-OPH and ADH by using the US National Park Service (NPS) secure file transfer service.

Environmental Assessment and Legionella Testing

An ADH environmental health specialist evaluated all accommodations and places visited in Hot Springs, Arkansas, listed by patients for Legionella risk. ADH and NPS-OPH staff inspected the HSNP hot spring water distribution system, including plumbing, reservoirs, and point-of-use sites in bathhouses and fountains. Environmental samples were collected for Legionella culture testing from the bathhouses, reservoirs, jug-filling stations, decorative fountains, or cooling towers on 6 occasions during October 2018–November 2019. Sampling sites during each round of testing were selected on the basis of epidemiologic information at the time and results of previous environmental tests. Laboratory A or laboratory B performed the testing; both laboratories were members of the Environmental Legionella Isolation Techniques Evaluation Program.

Ethics

This investigation was completed as part of routine outbreak response activities by ADH and NPS-OPH. These activities were further reviewed by CDC and determined to be not research.

Water System Description

Figure

Figure. Simplified flowchart and location of Hot Springs National Park water distribution system relative to bathhouses, hot spring water jug-filling stations (depicted by asterisks), and the city of Hot Springs, Arkansas,...

In HSNP, 2 bathhouses perform spa services with spring water (bathhouses A and B). An additional bathhouse was converted into a hotel that pipes spring water to guest rooms for baths (bathhouse C). Of the remaining 5 bathhouses, 1 is vacant, and the others house HSNP staff offices and a retail store (bathhouse D), a brewery restaurant (bathhouse E), a visitor center, and a cultural center. Except for the visitor center, all businesses operate as private entities and maintain a contract with HSNP to operate in park facilities. A hot spring water reservoir is located under a park administration building, which is also on bathhouse row (Figure). A hotel spa facility (facility X) and a medical center (facility Y) adjacent to HSNP also offer beauty or health treatments using hot spring water piped from the HSNP. A total of 3 hot water jug-filling fountains, 2 within HSNP and 1 outside the park in Hot Springs, enable visitors and residents to fill containers with hot spring water that meets US Environmental Protection Agency drinking water requirements because of the disinfectant properties of naturally high water temperatures. In addition, 4 decorative hot spring water fountains and 2 natural display hot water springs are inside park boundaries.

All 27 hot springs that supply HSNP are covered, sealed, and protected, and hot spring water is collected in a complex water distribution system (Figure). First, the hot spring water is piped into an ≈265,000 gallon reservoir. From the reservoir, the water is pumped to the bathhouses, an additional 100,000 gallon reservoir (not depicted in Figure), and a heat exchange system that cools the >135°F water to 75°F–95°F. This warm water is then pumped to a 400,000-gallon reservoir (later modified to 200,000-gallon capacity after the outbreak). Hot and warm water reservoirs are at a higher elevation than the bathhouses and provide constant water pressure. Bathhouse spring water temperature is adjusted by mixing warm and hot water, but water temperature cannot exceed 104°F in full body–contact recreational water (such as hot tubs or pools), according to NPS regulations (7). NPS regulations also stipulate that cooled hot spring water used in pools or for purposes more likely to result in inhalation of water droplets, such as individual-use jetted tubs, must be chemically disinfected (e.g., treated with chlorine or ozone).

Patient Demographics, Potential Exposures, and Diagnostic Testing

Case finding revealed no additional cases beyond the 5 patients reported by CDC who met the case definition. Symptom onset occurred during July 2018–August 2019 for the 5 travelers who received a Legionnaires’ disease diagnosis (Table 1). Patient age range at diagnosis was 61–72 years. Among the 4 patients who survived, all had underlying health or behavioral risk factors for Legionnaires’ disease, including chronic obstructive pulmonary disease, smoking, or daily use of a continuous positive airway pressure therapy machine. Next-of-kin reported no underlying health or behavioral risk factors for the patient who died.

Among the 5 patients with Legionnaires’ disease, all had visited HSNP; 2 stayed >1 nights at accommodation Z near HSNP (patients 2 and 3, not in the same room), and 3 patients received hot spring water spa services from bathhouse A (patients 1, 2, and 4). One patient (patient 5) did not enter bathhouse A or stay in accommodation Z. A common likely exposure source for the 2 patients who did not visit bathhouse A (patients 3 and 5) was not identified.

In 2 patients, Legionnaires’ disease was diagnosed by a PCR assay on lower respiratory specimens with no serogroup reported (patients 1 and 4); in 3 other patients the diagnosis was made by urinary antigen test, suggesting infection with Legionella pneumophila serogroup 1 (patients 2, 3, and 5) (8). Specimens from 1 person (patient 4) underwent both urinary antigen and sputum PCR testing, but only the sputum PCR test was positive. No patient specimens were submitted for culture testing, because samples were discarded by the time cases were reported to ADH and NPS-OPH.

Environmental Assessment and Testing

An environmental inspection of accommodation Z after the patients’ visits revealed that the building was plumbed to municipal water only, and faucet water within guest rooms was found to contain residual chlorine disinfectant at hot and cold temperatures. No specific risks for Legionella growth were identified in accommodation Z, and water was not submitted for Legionella culture testing. ADH and NPS-OPH inspected bathhouse A after each of the first 2 patients who visited the spa facility were reported. At that time, no specific risks for Legionella growth were identified. All faucets plumbed to municipal water contained residual chlorine disinfectant, and all faucets plumbed to spring water could reach temperatures >135°F. After receiving the travel-associated case notification report of the third patient who visited bathhouse A, an August 2019 inspection by ADH discovered that water ozonators intended for water disinfection on the individual-use microbubbling jetted soaking tubs were detached. The length of time ozonators had been detached is unknown. A final inspection in September 2019, prompted by a positive Legionella test from a poolside shower, found the shower to be plumbed to cooled hot spring water rather than municipal water.

All environmental samples collected during October 2018, January 2019, and July 2019, after each of the first 3 Legionnaires’ disease cases were reported, were sent to laboratory A and tested negative by culture for Legionella (Table 2). On September 18, 2019, additional environmental samples were collected after the fourth patient was reported (the third with a potential exposure in bathhouse A). This round of testing was the first to use laboratory B and the first to produce a positive result; the positive sample was collected from a hot spring water poolside shower in bathhouse A and was determined to be L. pneumophila serogroup 2–14 (Table 3). The positive Legionella culture prompted extensive testing throughout HSNP in October 2019, revealing L. pneumophila serogroup 2–14 in hot springs water in 1 additional site in bathhouse A, hot and cooled spring water from an infrequently used faucet in bathhouse C, and a hot spring water spigot in the spa area of facility X. Spring water temperatures were recorded only during the October and November 2019 round of testing. The positive sample from bathhouse A was taken from a hot water (134°F) pipe with stagnant or low water flow in the basement (Table 3). Additional parkwide testing in November 2019 produced 1 positive sample from the cooled spring water supply pipe in facility Y. Facility Y staff indicated that the cooled spring water supply line is infrequently used and is not routinely flushed. All samples from the HSNP hot and warm water reservoirs, main hot spring water line, and heat exchange system tested negative for Legionella. None of the L. pneumophila from HSNP’s water samples that tested positive belonged to serogroup 1.