Naegleria fowleri in Well Water

To the Editor: Naegleria fowleri, a protozoon found in hot springs and warm surface water, can cause primary amebic meningoencephalitis in humans. A survey of drinking water supply wells in Arizona determined that wells can be colonized and may be an unrecognized source of this organism that could present a human health risk. 
 
N. fowleri is a free-living ameboflagellate found in warm bodies of water such as ponds, irrigation ditches, lakes, coastal waters, and hot springs and can cause primary amebic meningoencephalitis. Humans come into contact with N. fowleri by swimming or bathing, particularly in surface waters. The ameba enters the nasal passages, penetrates the nasopharyngeal mucosa, and migrates to the olfactory nerves, eventually invading the brain through the cribriform plate (1). From 1995 to 2004, N. fowleri killed 23 persons in the United States (2), including 2 children in the Phoenix, Arizona, area in 2002, who had been exposed to well water but had not consumed it (3). There have been 6 documented deaths in 2007, all in warmer regions (Arizona, Texas, Florida) (4). 
 
Although N. fowleri’s presence in surface waters is well documented (5,6), no previous studies on its occurrence in wells have been conducted. We studied high-volume drinking water wells operated by municipal utilities or private water companies in the greater Phoenix and Tucson, Arizona, areas. Previous data from 500 wells in the region showed temperatures ranging from 13°C to 46°C. Typical well discharges ranged from hundreds to >3,780 L per minute. Well depths varied from 100 m to >300 m. 
 
Well water samples were collected by using 1-L sterile polyethylene bottles at or near the wellhead before disinfection by well owners or utilities (7). In phase 1, samples were collected after wells were flushed until the water was clear. During phase 2, samples were collected as water was turned on from spigots at or near wellheads (initial) and after a 3-borehole volume had flushed through the system (purged). Additional wells were sampled during this phase. Samples were tested for temperature, pH, turbidity, chlorine residual, conductance, coliforms, heterotrophic bacterial plate counts (HPC), and Escherichia coli following standard methods (7). 
 
To test for viable amebas, we spread aliquots on nonnutrient agar seeded with E. coli at 37°C (3,8). We placed scrapings from the advancing front of subsequent ameba plaques in distilled water to identify enflagellation (5); however, precise species identification was not possible. Live amebae were therefore harvested for PCR analysis to specifically identify N. fowleri. We chose PCR over the mouse pathogenicity test because other Naegleria species that are nonpathogenic in humans are lethal in mice (8). The genotype of isolates was not determined because all of the described genotypes found in the United States have been shown to be pathogenic in humans (9). 
 
To concentrate trophozoites/cysts, we gently agitated samples for 2 minutes and then centrifuged and filtered them through polyethylene filters (2-μm pore; Millipore, Bedford, MA, USA). A 10-μL volume of concentrate was used as a template for nested PCR (3,8) (triplicate tests were conducted immediately and after a 2-week 37°C incubation). Positive and negative PCR products were frozen at –80°C, coded to prevent bias, and shipped to Francine Marciano-Cabral at Virginia Commonwealth University for confirmation by cloning and sequencing (3). 
 
The general microbial quality of the wells was as follows: 73 (51%) had >500 HPC/mL; 8 (5.5%) were positive for coliforms; none were positive for E. coli. Oils used to lubricate well motors may result in the growth of HPC in well water (10). N. fowleri feeds on heterotrophic bacteria in water and could multiply in the well casing. This may explain N. fowleri’s colonization of wells. 
 
The recent association in Arizona between unchlorinated drinking water and the transmission of N. fowleri suggests that groundwater has been an unrecognized source of this organism. PCR detected N. fowleri DNA in 11 (7.7%) of 143 wells. Of 185 total samples, 30 (16.2%) tested positive for N. fowleri (Table). The organism was most often detected after the wells had been purged (17.9% purged vs. 10.0% initial samples), suggesting that N. fowleri was present in the aquifer or was released from the well casing or column during pumping. The wells testing positive for N. fowleri ranged in temperature from 21.9°C to 37.4°C (average 29.0°C; median 29.5°C). 
 
 
 
Table 
 
Naegleria fowleri in well water samples, Arizona 
 
 
 
The live trophozoite form was confirmed in only 1 well, though 11 of 143 wells tested positive according to PCR. This discrepancy may be due to the low occurrence of trophozoites in water or to differences in assay volumes for detection of live trophozoites (0.75 mL) versus PCR (30 mL equivalent unconcentrated volume). PCR is also more sensitive, capable of detecting 100 organisms/L in an unconcentrated sample (8); however, PCR did not determine if the amebas were infectious. Although PCR can determine the species by using primers for a specific gene sequence not found in other Naegleria species, it cannot determine the life stage (cyst/trophozoite). Trophozoites are believed to be the infectious form of the organism (1); nonetheless, cysts can be equally harmful because they may revert to trophozoites under optimal conditions (1). The surprisingly common occurrence of N. fowleri in drinking water wells suggests that groundwaters may be an unrecognized human health threat.

Most of the patients whose cases were diagnosed in Buenos Aires, including 5 who required hospitalization, were referred to Muñiz Hospital. Built a century ago, Muñiz Hospital comprises a number of independent pavilions surrounded by a spacious garden, where mosquitoes thrive, especially in summer. Thus, vectorborne infection in this case might have occurred either in Muñiz Hospital, in the Federal District, or in the southern city suburb, where the patient lives and works.
Until recently, dengue had not been suspected in patients with a fever living in the Buenos Aires area in the absence of a recent history of travel to an endemoepidemic area. Confi rmation of our case was evidence of local circulation of dengue virus. Thereafter, serum testing became recommended in Buenos Aires for acute febrile illness, among other dengue surveillance interventions in the area. More recently, epidemiologic surveillance of febrile illness has been strengthened countrywide upon the recent reporting of yellow fever cases in Argentina (8).
No circulation of dengue virus was reported in Buenos Aires during the fi rst 10 epidemiologic weeks of 2008. However, vector control measures should be strengthened to minimize the risk of infective persons triggering an epidemic of dengue or other fl avivirus disease.

Naegleria fowleri in Well Water
To the Editor: Naegleria fowleri, a protozoon found in hot springs and warm surface water, can cause primary amebic meningoencephalitis in humans. A survey of drinking water supply wells in Arizona determined that wells can be colonized and may be an unrecognized source of this organism that could present a human health risk.
N. fowleri is a free-living amebofl agellate found in warm bodies of water such as ponds, irrigation ditches, lakes, coastal waters, and hot springs and can cause primary amebic meningoencephalitis. Humans come into contact with N. fowleri by swimming or bathing, particularly in surface waters. The ameba enters the nasal passages, penetrates the nasopharyngeal mucosa, and migrates to the olfactory nerves, eventually invading the brain through the cribriform plate (1). From 1995 to 2004, N. fowleri killed 23 persons in the United States (2), includ- Although N. fowleri's presence in surface waters is well documented (5,6), no previous studies on its occurrence in wells have been conducted. We studied high-volume drinking water wells operated by municipal utilities or private water companies in the greater Phoenix and Tucson, Arizona, areas. Previous data from 500 wells in the region showed temperatures ranging from 13°C to 46°C. Typical well discharges ranged from hundreds to >3,780 L per minute. Well depths varied from 100 m to >300 m.
Well water samples were collected by using 1-L sterile polyethylene bottles at or near the wellhead before disinfection by well owners or utilities (7). In phase 1, samples were collected after wells were fl ushed until the water was clear. During phase 2, samples were collected as water was turned on from spigots at or near wellheads (initial) and after a 3-borehole volume had fl ushed through the system (purged). Additional wells were sampled during this phase. Samples were tested for temperature, pH, turbidity, chlorine residual, conductance, coliforms, heterotrophic bacterial plate counts (HPC), and Escherichia coli following standard methods (7).
To test for viable amebas, we spread aliquots on nonnutrient agar seeded with E. coli at 37°C (3,8). We placed scrapings from the advancing front of subsequent ameba plaques in distilled water to identify enfl agellation (5); however, precise species identification was not possible. Live amebae were therefore harvested for PCR analysis to specifi cally identify N. fowleri. We chose PCR over the mouse pathogenicity test because other Naegleria species that are nonpathogenic in humans are lethal in mice (8). The genotype of isolates was not determined because all of the described genotypes found in the United States have been shown to be pathogenic in humans (9).
To concentrate trophozoites/ cysts, we gently agitated samples for 2 minutes and then centrifuged and fi ltered them through polyethylene fi lters (2-μm pore; Millipore, Bedford, MA, USA). A 10-μL volume of concentrate was used as a template for nested PCR (3,8) (triplicate tests were conducted immediately and after a 2-week 37°C incubation). Positive and negative PCR products were frozen at -80°C, coded to prevent bias, and shipped to Francine Marciano-Cabral at Virginia Commonwealth University for confi rmation by cloning and sequencing (3).
The general microbial quality of the wells was as follows: 73 (51%) had >500 HPC/mL; 8 (5.5%) were positive for coliforms; none were positive for E. coli. Oils used to lubricate well motors may result in the growth of HPC in well water (10). N. fowleri feeds on heterotrophic bacteria in water and could multiply in the well casing. This may explain N. fowleri's colonization of wells.
The recent association in Arizona between unchlorinated drinking water and the transmission of N. fowleri suggests that groundwater has been an unrecognized source of this organism. PCR detected N. fowleri DNA in 11 (7.7%) of 143 wells. Of 185 total samples, 30 (16.2%) tested positive for N. fowleri (Table). The organism was most often detected after the wells had been purged (17.9% purged vs. 10.0% initial samples), suggesting that N. fowleri was present in the aquifer or was released from the well casing or column during pumping. The wells testing positive for N. fowleri ranged in temperature from 21.9°C to 37.4°C (average 29.0°C; median 29.5°C).
The live trophozoite form was confi rmed in only 1 well, though 11 of 143 wells tested positive according to PCR. This discrepancy may be due to the low occurrence of trophozoites in water or to differences in assay volumes for detection of live trophozoites (0.75 mL) versus PCR (30 mL equivalent unconcentrated volume). PCR is also more sensitive, capable of detecting 100 organisms/L in an unconcentrated sample (8); however, PCR did not determine if the amebas were infectious. Although PCR can determine the species by using primers for a specifi c gene sequence not found in other Naegleria species, it cannot determine the life stage (cyst/trophozoite). Trophozoites are believed to be the infectious form of the organism (1); nonetheless, cysts can be equally harmful because they may revert to trophozoites under optimal conditions (1). The surprisingly common occurrence of N. fowleri in drinking water wells suggests that groundwaters may be an unrecognized human health threat.