Investigation of Spa Pools Associated with Lung Disorders Caused by Mycobacterium avium Complex in Immunocompetent Adults

Abstract

Three cases of Mycobacterium avium complex-related lung disorders were associated with two poorly maintained spa pools by genotypic investigations. Inadequate disinfection of the two spas had reduced the load of environmental bacteria to less than 1 CFU/ml but allowed levels of M. avium complex of 4.3 × 10⁴ and 4.5 × 10³ CFU/ml. Persistence of the disease-associated genotype was demonstrated in one spa pool for over 5 months until repeated treatments with greater than 10 mg of chlorine per liter for 1-h intervals eliminated M. avium complex from the spa pool. A fourth case of Mycobacterium avium complex-related lung disease was associated epidemiologically but not genotypically with another spa pool that had had no maintenance undertaken. This spa pool contained low numbers of mycobacteria by smear and was culture positive for M. avium complex, and the nonmycobacterial organism count was 5.2 × 10⁶ CFU/ml. Public awareness about the proper maintenance of private (residential) spa pools must be promoted by health departments in partnership with spa pool retailers.

Mycobacterium species are ubiquitous in the environment and are found worldwide (7, 9, 14, 21, 30). They have a predilection for a variety of natural waters, including lakes, rivers, and streams, and have also been detected in water supply systems (1, 6, 7, 13, 14, 15, 30). In a study of eight water supply distribution systems across the United States, the average number of M. avium in biofilms was 0.3 CFU per cm² and the number of M. intracellulare was 600 CFU per cm² for all surfaces (13). Mycobacteria not only survive but may persist and increase in number, and waters that have traversed such systems yield mycobacteria. Human beings are regularly exposed to these waters, which represent a potential source for infection.

End uses of supplied water are myriad but include industry (machine coolant fluids), hospitals, commercial buildings, residential garden watering, residential drinking and cooking, hot water systems, baths and showers, swimming pools, spa pools or hot tubs, ice machines, aquariums, and miscellaneous purposes such as foot bath whirlpools (1, 6, 7, 12, 13, 14, 26, 29, 31). Recirculating hot water systems in buildings, spas, and hot tubs have yielded high numbers of M. avium. Indeed, hot water systems may have higher numbers of M. avium than the source water (10). Swimming pools have also yielded M. avium (12), and long-term exposure to aerosols has caused granulomatous pneumonitis in lifeguards (25).

A number of reports have demonstrated an association of M. avium complex (MAC) in spa pools with lung disorders in humans (5, 11, 16, 17, 18, 20). However, most did not study the relationship between clinical and environmental strains except for two studies that used restriction fragment length polymorphism and multilocus enzyme electrophoresis to demonstrate a genotypic link between MAC isolates from the patient and the spa pool (17, 20). The present study attempted to define the source, burden, and persistence of MAC in spa pools associated with four cases of lung disorders and examined the molecular epidemiology of MAC isolates obtained from clinical and environmental sampling. This is the first study to present genotypic evidence of the persistence of a MAC strain in a spa pool for over 5 months despite standard decontamination procedures.

MATERIALS AND METHODS

Case reports. (i) Cases 1 and 2.

A previously healthy 32-year-old male smoker developed dyspnea and cough associated initially with myalgia and arthralgia in late November 2001. During a 3-week working trip to India in December, he became increasingly fatigued, and upon returning, he underwent a chest x-ray which was reported as normal. A bronchial lavage was smear negative for acid-fast bacilli (AFB) and fungi but culture positive for M. avium. No other pathogens were isolated. Transbronchial biopsy specimens showed scattered nonnecrotizing granulomata with associated minor focal interstitial inflammation and intra-alveolar organization. There was no evidence of underlying immunosuppression. A diagnosis of MAC-associated hypersensitivity pneumonitis was made, and treatment with prednisolone and anti-MAC therapy was initiated. He made a full recovery over a 6-month period. Further history at the start of treatment established that he used a home outdoor spa pool daily and was responsible for its maintenance. Its use was discontinued as part of his treatment. The spa water came from both a rainwater tank and the mains system.

When the possible association with spa pool use for case 1 was revealed, his partner, a 36-year-old woman, presented with a similar clinical history, starting with a flu-like illness in September 2001. She used the spa pool almost as frequently as her partner. Lavage specimens were smear negative for AFB and fungi but culture positive for M. avium. Transbronchial biopsy specimens showed well-formed nonnecrotizing granulomata without AFB. Again, the diagnosis was MAC-associated hypersensitivity pneumonitis. She was treated with anti-MAC antibiotics and refused oral steroids.

(ii) Case 3.

A 41-year-old male was referred to a respiratory physician because of a persistent, productive cough over an 8-month period and an unexpected finding of multiple sputum specimens smear negative but culture positive for MAC. He had purchased a spa pool (supplied by rainwater only) some 3 years earlier, which was used on a daily basis. The results of a chest high-resolution computed tomography were normal, as were lung function studies. A presumptive diagnosis of MAC-associated transient airway hyperreactivity was made, and the patient was advised to avoid using the spa pool. His clinical response was excellent, and the only medication used was an aerosolized steroid for a short period. Serial sputum samples have remained culture negative for MAC. His wife and son rarely used the spa pool, and both have remained well.

(iii) Case 4.

A 39-year-old male presented with a history of persistent cough and dyspnea for several months following a flu-like illness with intermittent night sweats and episodes of blood-streaked sputum. Coincidentally, he had moved into a new house with an outdoor spa pool supplied by rainwater some 2 years earlier. He had a history of childhood asthma and bilateral spontaneous pneumothorax in 1995, requiring bilateral apical blebectomy and pleurodesis. At presentation, his chest x-ray showed a left apical cavity, present since 1995 but increased in size and wall thickness. Cultures of a computed tomography-guided fine-needle aspirate of the cavity grew Aspergillus fumigatus and MAC. No fungal elements or AFB were seen in the biopsy specimen. He was treated for M. avium disease, showing significant symptomatic improvement within several months, and treatment was completed without complication.

Water and environmental sampling.

Spa and rainwater tank water samples were collected into sterile 500-ml plastic bottles containing 35 mg of sodium thiosulfate; 400 ml of each sample was centrifuged at 4,420 × g for 25 min at 4°C. The remaining 15 to 30 ml of resuspended sediment was recentrifuged at 3,000 × g for 15 min, leaving a final volume of 4 ml. Large foam swabs (Enviroswabs; Techra International, New South Wales, Australia) were used to collect environmental samples from spa filters and fittings. Swabs were wrung out by hand into 20 ml of sterile, filtered distilled water, and the washings were centrifuged at 3,000 × g for 15 min, leaving a final volume of approximately 4 ml.

Sample processing and culture.

Samples were mixed with an equal volume of 2% sodium hydroxide and 0.5% N-acetyl cysteine, thoroughly vortexed, and incubated at room temperature for 60 min. The mixture was neutralized to pH 7 and concentrated by centrifugation at 3,000 × g for 15 min, leaving a final volume of approximately 4 ml. A smear was prepared from each sample, which was then cultured into Bactec 12B vials containing polymyxin B, amphotericin B, nalidixic acid, trimethoprim, and azlocillin (Becton Dickinson Instrument Systems, Sparks, Md.) and onto ferric ammonium citrate- and pyruvate-supplemented Löwenstein-Jensen medium (Oxoid, South Australia), and incubated at 30 and 35°C. Smears of positive cultures were stained by the Ziehl-Neelsen method. Isolates were identified with a combination of M. avium, M. intracellulare, and M. avium complex commercial DNA probes (Accuprobe; GenProbe, San Diego, Calif.). The concentration of MAC in spa water was determined by inoculating 100 μl of serial 100-fold dilutions from the processed spa water sample onto Middlebrook agar (Pathcentre). Colonies were manually enumerated with a plate microscope. Patient isolates stored at −70°C were resuscitated and subcultured to a Bactec 12B vial and, once the growth index reached 999, subcultured onto Middlebrook agar.

Molecular analysis.

For all patients and their environmental MAC isolates, individual colonies were collected from the Middlebrook agar plates, subcultured to Dubos broth (Oxoid), and used for genotyping. Pulsed-field gel electrophoresis (PFGE) was performed as described previously (23) once organisms had been inactivated in 80% (vol/vol) ethanol for 1 h at room temperature. Gel patterns were assessed visually. PCR amplification of genomic sequences located between the repetitive elements IS1245 and IS1311 was also used to examine the relatedness of MAC isolates from clinical and environmental sources (24). The method was essentially that described by Picardeau and Vincent (24) with the following modifications: 5 μl of a 1:10 dilution of lysate was amplified at an annealing temperature of 50°C and 40 amplification cycles. Molecular size markers were Geneworks (Adelaide, Australia) and New England Biolabs (Beverly, Mass.) 100-bp standards. Gel patterns were visualized and photographed with UV illumination, and band patterns were assessed visually (23). Comparison of band patterns was conducted only within single gels. However, the gel patterns were found to be reproducible, based on multiple repeat PCRs.

RESULTS

Cases 1 and 2.

Smears prepared from concentrates of the spa pool and filter swab contained >100 acid-fast bacilli per high-power field (HPF) (Fig. 1A and B). Based on serial dilutions of the spa water concentrate, there were approximately 4.3 × 10⁴ CFU of MAC per ml (Table 1). The concentrate of a spa water sample collected 1 month after the initial visit contained <1 AFB/HPF, but a swab of the recently disinfected filter contained >100 AFB/HPF. At that time, fragments of scale collected from pump water were composed almost entirely of mycobacteria (Fig. 2A and B). The spa pool was then drained, the pool was thoroughly cleaned, new filters were installed, and the spa was treated with a proprietary persulfate-based biocide.

FIG. 1.

Spa pool of cases 1 and 2. (A) Ziehl-Neelsen-stained smear prepared from the spa pool filter. (B) Ziehl-Neelsen-stained 100-fold concentrate of spa water.

TABLE 1.

Presence of mycobacteria in environmental samples

Case no.	Smear of concentrate (AFB/HPF)			Culture of spa water (CFU/ml)
	Spa water	Filter	Rainwater	MAC	Environmental organisms
1 and 2	>100	>100	Negative	4.3 × 104	<1
3	>100	>100	Negative	4.5 × 103	<1
4	<1	<1	<1	Positivea	5.2 × 106b

^aQuantitation not attempted.

^bIncluded 1.2 × 10² CFU of P. aeruginosa per ml.

FIG. 2.

Spa pool of cases 1 and 2. (A) Particles of scale found in the piping of the spa pool. (B) Ziehl-Neelsen-stained crush preparation of scale, demonstrating the presence of large numbers of acid-fast bacilli. Magnification, ×500.

The PFGE patterns for the M. avium isolates from cases 1 and 2 were indistinguishable from each other and matched the patterns obtained from the spa water and rainwater tank (Fig. 3). Five single-colony picks each were taken from the spa water samples, and PFGE pattern variations were observed. The M. avium genotype associated with disease persisted in the spa water for the next 5 months until eliminated by repeated supershocking treatment with greater than 10 mg of chlorine per liter for 1-h intervals (Fig. 4). Spa water samples and swabs of filters collected 3 and 4 months after supershocking were smear and culture negative for mycobacteria. Ongoing supershock treatment is conducted each month.

FIG. 3.

Pulsed-field gel electrophoresis gel containing clinical and environmental isolates from cases 1 and 2. Lane M, Staphylococcus aureus NTCT 8325. Lane 1, clinical isolate from case 1. Lane 2, clinical isolate from case 2. Lane 3, M. avium isolate from a spa water sample collected in February 2002. Lanes 4 and 5, M. avium isolates from a rainwater tank sample collected in February 2002. Lanes 6 and 7, clinical M. avium isolates from unrelated cases.

FIG. 4.

PCR typing gel containing clinical and environmental isolates from cases 1 and 2. Lane M1, 100-bp DNA ladder molecular size markers (New England Biolabs). Lane 1, clinical isolate from case 1. Lane 2, clinical isolate from case 2. Lane 3, M. avium isolate from a spa water sample collected in February 2002. Lanes 4 to 7, M. avium isolates from a spa water sample collected in July 2002. Lane 8, clinical M. avium isolate from an unrelated case. Lanes 9 and 10, reagent controls. Lane M2, 100-bp DNA ladder molecular size markers (Geneworks).

Case 3.

The spa water concentrate and swab of the spa filter both contained >100 AFB/HPF, and the spa water contained approximately 4.5 × 10³ CFU of MAC per ml (Table 1). Rainwater was used to supply water to the spa pool; a concentrate contained no AFB but did grow MAC after 4 weeks of incubation. The PCR genotypes of nine single-colony picks from the patient and two of five spa water single-colony picks were indistinguishable from each other, but multiple genotypes were observed in spa water isolates (Fig. 5). Multiple PCR runs confirmed the stability of major bands, with faint bands (e.g., Fig. 5, lane 4) no longer being present when repeated under higher-stringency conditions (data not shown).

FIG. 5.

PCR typing gel containing clinical and environmental isolates from case 3. Lane M1, 100-bp DNA ladder molecular size markers (New England Biolabs). Lane 1, clinical isolate from case 3. Lanes 2 to 4, M. avium isolates collected from spa water. Lane 5, clinical M. avium isolate from an unrelated case. Lanes 6 and 7, reagent controls. Lane M2, 100-bp DNA ladder molecular size markers (Geneworks).

Case 4.

A sample of spa pool water was visibly turbid and found to contain approximately 5.2 × 10⁶ CFU of mixed environmental organisms per ml, including 1.2 × 10² CFU of Pseudomonas aeruginosa per ml (Table 1). The spa water concentrate contained <1 AFB/HPF and grew MAC after 6 days of incubation. Concentrates of water collected from rainwater tanks used in part to supply water to the spa pool contained <1 AFB/HPF, and cultures grew MAC after 3 weeks of incubation. Although the spa water grew M. avium, the patient's clinical isolate was negative with chemiluminescent probes for M. avium and M. intracellulare but positive for the MAC probe, suggesting an X cluster MAC. The PCR genotypes of five single-colony picks each from the MAC isolates from the patient, spa water, and rainwater tanks were unrelated, and multiple genotypes were observed in spa water and rainwater (data not shown).

DISCUSSION

Spa pools have the potential to contain higher levels of contamination than swimming pools because of their reduced water capacity, greater bather load-to-water volume ratio, higher operating temperature, vigorous aeration of the water body, and elevated organic contaminant loading (27). The warm, aerated water provides an ideal environment for the rapid growth of microorganisms. In Australia, spa pools are frequently sited outdoors and are therefore exposed to environmental factors such as dust and sunlight which may predispose to contamination by environmental organisms and reduce disinfection efficacy.

The M. avium isolates recovered from cases 1 and 2, rainwater, and spa pool water were indistinguishable by PFGE and PCR genotyping. MAC strains with the same genotype persisted in the spa water over 5 months until the spa pool was decontaminated by repeated supershock dosing with greater than 10 mg of chlorine per liter for a minimum of 60 min, followed by draining, cleaning, and filter replacement. The persistence of a single MAC genotype in a spa pool has not been reported previously. However, the repeated isolation of a single M. avium genotype from a hospital hot water system over a period of 18 months demonstrated that persistent MAC colonization of water systems can occur (29).

The population dynamics of MAC in spa pools is unknown. This study has documented the persistence for more than 5 months of a disease-causing strain. The ability of this study and previous reports to match genotypically clinical and environmental isolates collected from patients who have been symptomatic for months before their spa pools are investigated also argues that offending MAC strains can persist in spa pools. Despite multiple attempts to clean the spa pool of cases 1 and 2, M. avium remained. A new filter installed after thorough cleaning was sampled 4 weeks later and found to contain 10 to 100 AFB/HPF, compared with <1 AFB/HPF in a spa water concentrate. Therefore, the filter may be an important site where organisms may multiply and act as a reservoir for reseeding of the spa pool.

In addition to the disease-related strains, multiple genotypes were observed for M. avium recovered from the concentrate of spa pool water of each case, suggesting that multiple MAC strains may inhabit a contaminated spa pool. Rainwater was the sole source of water for the spa pools of cases 3 and 4, while rainwater was used in conjunction with mains water for the spa pool used by cases 1 and 2. Rainwater has a low ion content and low alkalinity, creating a poor buffering capacity and thus making it more difficult to control pH than with mains water (8). Factors associated with higher mycobacterial numbers in raw water are high raw water turbidity and higher levels of assimilatable organic carbon and biodegradable organic carbon levels (13). Rainwater tanks require regular maintenance and cleaning (8).

This study has genotypically linked three cases of MAC lung disease with two spa pools. A previous report described the use of restriction fragment gel electrophoresis to compare M. avium isolates from three patients and their hot tub and found that all isolates had an identical pattern (20). Another study used restriction fragment gel electrophoresis to compare isolates from the patient and hot tub, and although it was similar to the patient isolate, the hot tub isolate had three additional bands, suggesting that the isolates were not related (17). Transmission of MAC from spa pools is likely related to (i) the infectiousness of the environmental source, (ii) the duration and intensity of exposure, and (iii) host susceptibility. Air bubbles rising through water collect particulates, chemicals, and microorganisms, including mycobacteria, and burst upon reaching the water surface, resulting in droplets' being ejected into the air (2, 3, 4). Mycobacteria are highly hydrophobic, and this property promotes their adsorption to air bubbles (22). Aerosolization may produce a 1,000-fold increase in viable mycobacteria per ml of ejected droplet. Upon drying, droplets may shrink to a size suitable for entry into the alveolar spaces in the lung (22).

The frequency of spa pool use is a second risk factor for developing MAC-related disease. In the present study, case 1 used the spa pool for 1 to 2 h on most days, while the other cases used the spa pool on an almost daily basis. The relationship between the frequency of spa pool use and illness has been noted previously (11, 20). Cases 1 and 3 were also responsible for undertaking maintenance of the spa pools, and activities such as cleaning the filter may increase exposure to high concentrations of mycobacteria (11).

The failure to isolate a MAC genotype from spa water that was indistinguishable from the clinical strain in case 4 is likely multifactorial. The harsh protocols used to decontaminate clinical and environmental samples may reduce the number of viable mycobacteria by 50% or more (1). A combination of broth and solid media used to amplify mycobacteria from the samples was followed by plating onto Middlebrook agar to obtain single colonies; some mycobacteria may grow poorly on such media and be overgrown by other genotypes.

The spa pools reported in the present and previous studies were all improperly maintained, and the faults included one or more of the following: (i) not maintaining the correct pH range, (ii) inadequate disinfection, (iii) irregular changing of water, and (iv) irregular cleaning or changing of filters. In aqueous solutions, free chlorine exists primarily in two forms, hypochlorous acid (HOCL) and hypochlorite ions (OCL⁻). Importantly, HOCL is a much more effective disinfectant than OCL⁻. The relative proportions of each are dependent upon pH; at pH 6.0, 98% of free chlorine exists as HOCL and 83% at pH 7.0, but only 1% at pH 10.0. In the operating range for spa pools of pH 7.2 to 7.6, the percentage of free chlorine as HOCl varies from 63 to 39%, respectively. The disinfection efficiency of chlorine increases with temperature, as the reaction rate of HOCL with bacterial components increases (19).

M. avium is relatively resistant to commonly used methods to disinfect water, such as chlorine, chloramine, chlorine dioxide, and ozone (28). One study demonstrated that five M. avium strains had 580- to 2,300-fold-greater resistance to chlorine than Escherichia coli. Inadequate disinfection and the relative resistance of mycobacteria to chlorine may account for the high level of MAC and the low level (<1 CFU/ml) of other environmental organisms in two of the spa pools in the present study. Interestingly, water-grown M. avium strains had a 3.6- to 14.6-fold-higher resistance to chlorine than the same M. avium strains grown on laboratory media, although this finding may have been related to the slower growth rate for the water-grown organisms (15). This finding has also been noted for other environmental mycobacteria (19). Resistance to chlorine dioxide and ozone for M. avium was 100- and 50-fold greater, respectively, than that for the E. coli control (28).

Poorly maintained spa pools are conducive to the survival and amplification of Mycobacterium spp., notably MAC. Current South Australian regulations for public spa pools are (i) constant disinfection and minimum disinfectant residuals (in the case of chlorine, the minimum residual at 35 to 37°C is 4 mg/liter), (ii) maintenance of pH within a range of 7.2 to 7.6, (iii) filtering of the total volume at least every 30 min, (iv) replacement of at least 20% of pool water each day or complete draining every week, and (v) weekly cleaning (8). Guidelines for private spa pool owners are similar, although water changes may be reduced to the replacement of 10 to 15% of total volume per week or total replacement every 1 to 2 months due to the likelihood of lower usage (www.dhs.sa.gov.au/pehs/topics/topic-spa-pools).

The rising number of homes in Australia, the United States, and other Western countries with a spa pool suggests that further cases of spa pool-related MAC lung disorders are likely. Physicians, clinical laboratories, and environmental health officers must be aware of this clinical entity, and public awareness about the importance of proper maintenance of private (residential) spa pools must be promoted by health departments in partnership with spa pool retailers. The key to control is preventative maintenance and continuous disinfection and filtration.