Salmonella Virchow and Salmonella Weltevreden in a Random Survey of the Asian House Gecko,Hemidactylus frenatus,in Houses in Northern Australia

Abstract

Background:

Salmonella enterica serotype Virchow is the most common cause of invasive nontyphoid salmonellosis in North Queensland, particularly in infants, but the zoonotic source is unknown. This study aimed at determining (i) the prevalence of the introduced Asian house gecko, Hemidactylus frenatus, in houses in North Queensland and (ii) whether they were a potential source of Salmonella Virchow.

Methods:

Asian house geckos were collected in a random survey of houses in Townsville, North Queensland. Gut contents underwent microbiological analysis within 2 h of removal using both direct plating and enrichment broth methods. Any organism found to be a presumptive Salmonella spp. was then sent to a reference lab for confirmation of genus/species, serotyping, and phage typing if indicated.

Results:

One hundred Asian house geckos were collected from 57 houses. Geckos were present in 100% of houses surveyed, and prevalence of Salmonella in large intestinal contents was 7% (95% confidence interval 2, 12%). Three serotypes were found: Virchow (phage type 8), Weltevreden, and an untypable subspecies 1 serotype 11:-:1,7.

Conclusion:

Since Salmonella Virchow (phage type 8) is associated with invasive disease, the introduced Asian house gecko may play a significant role in the epidemiology of sporadic salmonellosis in places invaded by these peridomestic reptiles. These results justify more detailed epidemiological studies on the role of the Asian house gecko in sporadic salmonellosis and development of evidence-based strategies to decrease this potential zoonotic hazard.

Introduction

S almonella enterica serotype Virchow is the most common cause of invasive nontyphoid salmonellosis (NTS) in North Queensland, particularly in infants (Ashdown and Ryan 1990, Messer et al. 1997), with the average age of onset being 2.8 months (Messer et al. 1997). Although the source of Salmonella Virchow in these infants was not identified, either people in close contact with the infants or animals inside the home environment were the most likely sources. A review of Townsville General Hospital laboratory records (1978–1988) concluded that Salmonella Virchow was the predominant serotype in North Queensland (Ashdown and Ryan 1990); further, information from various sources demonstrated that more than 90% of Salmonella Virchow infections in humans in Australia occurred in Queensland (Ashdown and Ryan 1990). The predominance of Salmonella Virchow in Queensland has continued to the present day (Sullivan et al. 1998, OzFoodNet Working Group 2007), whereas other serotypes predominate in different regions of Australia (OzFoodNet Working Group 2007). Phage typing of Salmonella Virchow isolates demonstrated that phage type 8 (PT 8) was dominant and most commonly isolated from humans and chickens (Sullivan et al. 1998). Although a superficial interpretation would suggest chickens as a source of Salmonella Virchow PT 8 for humans, epidemiological data, particularly the geographic distribution, did not provide supportive evidence. In Hawaii, house geckos were implicated as the source of human cases of NTS (Helm 1981).

There have been 27 studies reporting Salmonella from Geckonidae but no studies on Hemidactylus frenatus (Murphy and Myers 1993, Cyriac and Wozniak 2000). The reported prevalence of Salmonella carriage in free-ranging geckos has been from 2% to 48% (mean, 17%; 95% confidence interval [95% CI 11.6, 23.2%]) (Murphy and Myers 1993). In its native environments throughout Southeast Asia, the Asian house gecko lives almost exclusively in buildings (Wilson 2006). Since coming to Australia in the mid 1960s, this non-native species has steadily increased with a population explosion observed in the last 5–10 years. Consequently, it now dwells in most suburban homes, shops, inner city buildings, and factories from Cooktown in northern Queensland to the Gold Coast in southeast Queensland (Wilson 2006). Since the Asian house gecko is a species that appears to be expanding in range, possibly assisted by the increasing global trend in urbanization, understanding its role in the epidemiology of salmonellosis in humans is of increasing importance.

This study aimed at determining the prevalence of the Asian house gecko, H. frenatus, in houses in Townsville and the prevalence of Salmonella carriage as a first step in discerning any potential epidemiological link between this species and human salmonellosis.

Methods

Sampling

Collection sites were human dwellings in five randomly selected Townsville/Thuringowa (19°15′S 146°49′E) suburbs. Each suburb in Townsville/Thuringowa was assigned a number between 1 and 100. A random number generator was then used to generate a list of 50 numbers. The first five suburbs (Annandale, Condon, Cranbrook, Mount Louisa, and Rasmussen) with an assigned number corresponding to one on the generated list were selected for specimen collections. Residential streets in each selected suburb were randomly selected, and 2 houses per street were chosen from a list of 10 randomly generated numbers (maximum number determined by the number of houses on the street). If the owner/lessee was not home or did not allow access, the next house on the list was selected for survey. Of the houses selected, 95% of occupants allowed geckos to be collected. Collections occurred between 7:00 and 9:00 P.M. to maximize the chance of someone being home and were conducted in all rooms of the selected dwellings except in bedrooms (for privacy reasons). The search was terminated when either a maximum of two geckos was collected or 30 min had passed to minimize the disturbance caused by collection. A total of 100 geckos was collected.

Geckos were caught by placing a small plastic container over them, sliding a thin piece of cardboard between the mouth of the container and the catch-surface, and then replacing the cardboard with the lid of the container. The lid, which contained breathing holes, was then secured using masking tape and labeled for identification purposes. Captured geckos were transported to a Physical Containment Level 2 laboratory for sacrifice and dissection at the first convenient opportunity the next day.

Geckos were euthanized with 0.25 mL of sodium pentobarbitone (Lethabarb^®; Virbac) intraperitoneally administered via a 26G needle (ANZCART 2001). The lower gut was aseptically removed, and culture was performed within 2 h of removal. Any feces found in the catch containers or eggs in the uterus were also tested.

Microbiological analysis

Initial microbiological analysis was done at the Tropical and Aquatic Animal Health Laboratory, Townsville, using both direct plating and enrichment broth methods. Each specimen (mean weight, 2.13 g; range, 0.7–2.83 g) was separately processed by first macerating it in nutrient broth (a minimal amount of nutrient broth was used to mash the sample) and then mashing it further to produce a more homogenous mixture. One drop (0.04 mL) of the mixture was directly pipetted onto each plate (a Brilliant Green Agar [BGA] and Lysine-Mannitol-Glycerol [LMG]), and into a bottle (3 mL) of buffered peptone water (BPW). Plates were streaked using standard microbiological techniques (Isenberg 1992) and then incubated for 24 h at 37°C along with the BPW. The LMG and BGA plates were examined for colonies characteristic of Salmonella growth: On BGA, Salmonella appear as small, translucent, red colonies; on LMG, Salmonella appear as small, red colonies with a black center. After incubation, a few drops of the inoculated BPW were pipetted into two bottles of Rappaport-Vassiliadis Broth (RV Broth); one was incubated for 24 h at 37°C, the other was incubated at 42°C. One drop from each bottle of RV Broth was then pipetted and streaked onto divided—to separate the 37°C sample from the 42°C sample—LMG and BGA plates. These plates were incubated for 24–48 h and then examined. A suspect colony from either the direct plating or enrichment broth method was streaked onto a Blood Agar plate and incubated for 24 h at 37°C to ensure the bacteria were more viable for the next phase of tests: the Oxoid Biochemical Identification System (O.B.I.S.), Salmonella Test (Oxoid), and the API 10S (bioMérieux). Before running the two aforementioned tests, a simple oxidase test was used as a screening tool to detect the characteristic oxidase-negative reaction exhibited by Salmonella species. The O.B.I.S. Salmonella Test was used to differentiate Salmonella from other Enterobacteriaceae that can grow on the same selective media and have a similar colonial appearance. These organisms include Citrobacter spp., which have pyroglutamyl aminopeptidase activity and Proteus spp., Morganella spp. and Providencia spp., which possess nitrophenylalanine deaminase activity. Consequently, a test result of pyroglutamyl aminopeptidase-negative and nitrophenylalanine deaminase -negative suggest a Salmonella spp. colony (Oxoid 2006). The suspect organism was then tested using an API 10S strip. Any organism found to be a presumptive Salmonella sp. by both tests was then sent to the Public Health Microbiology Lab at Queensland Health Scientific Services in Coopers Plains, Queensland, for confirmation of genus/species, serotyping, and phage typing if indicated.

Sample size

It was hypothesized that the prevalence of salmonellae carriage would be around 20% based on a previous survey of reptiles in Queensland (Thomas et al. 2001). Hence, a sample size of 100 was chosen to produce a 95% CI equal to the sample proportion ± 0.08 when the estimated proportion is 0.20 (20%) (calculated using NCSS Trial and PASS software program) (Hintze 2002). It was decided that the 100 house geckos would be collected from at least 50 households to minimize the likelihood of a higher than naturally occurring prevalence of Salmonella-positive geckos—one infected gecko in a given environment may increase the chance of other geckos being infected due to shared food sources and/or environment. The prevalence of Salmonella spp. together with 95% CI was calculated and adjusted for the cluster sampling approach (primary sampling unit: household). Analysis was conducted using the survey commands of Stata, release 8.

Ethics

This project was conducted under permit A1126 of James Cook University Animal Ethics Committee.

Results

One hundred geckos were collected from 57 dwellings (houses and units) throughout the five randomly selected suburbs of Townsville/Thuringowa, North Queensland. All surveyed dwellings (100%) had geckos, but two geckos could be caught in only 43 houses and one gecko in 14 houses. Salmonella spp. were isolated from the lower gut contents of seven geckos to yield a prevalence of 7% (7/100) (95% CI 2, 12%). Each Salmonella-positive gecko came from a separate dwelling, resulting in a prevalence of Salmonella-positive dwellings of 12% (7/57) (95% CI 4, 20%). Salmonella enterica Weltevreden was the most common serotype (5 of 7 positive samples = 71.4%), with one isolate being Salmonella enterica Virchow (14.3%) and the other being Salmonella subspecies 1 serotype 11:-:1, 7 (14.3%). The latter isolate was unable to be fully identified, as one of the first-phase antigens failed to express (i.e., antigens were present but did not react). The Salmonella Virchow isolate was PT 8.

Nine samples of feces (taken directly from the capture containers) and one sample of gecko eggs (excised during dissection) were negative for Salmonella. Two of the nine Salmonella-negative samples of feces collected were from geckos found to be carrying Salmonella spp. in their lower gut.

Discussion

Using a random survey, we found 100% of houses surveyed in Townsville to be inhabited by the Asian house gecko. This highlights that the majority of human residents in Townsville intimately live with reptiles, a largely ignored relationship with zoonotic implications. This study found Salmonella spp. in the Asian house gecko at a prevalence of 7% inside houses surveyed in tropical Australia. Twelve percent of surveyed houses in Townsville had a Salmonella-positive gecko, and this value would possibly increase with a larger number of geckos sampled per house. Although detecting salmonellae at this prevalence is not unexpected, the finding of a major regional zoonotic pathogen, Salmonella Virchow PT 8, in these reptiles cohabiting with humans is of concern.

This is the first study on salmonellae in house geckos in Australia and the first on the Asian house gecko, H. frenatus, globally. The reported prevalence of Salmonella carriage in free-ranging geckos has widely varied (2%–48%) (Murphy and Myers 1993). A wide range of serotypes has been reported from geckos, including Salmonella Weltevreden, but Salmonella Virchow has not been previously reported (Murphy and Myers 1993, Cyriac and Wozniak 2000). Opportunistic surveys of vertebrates in the Townsville region have found Salmonella Virchow in wallabies, kangaroos (Speare and Thomas 1988, Thomas et al. 2001), and the cane toad (Bufo marinus) (O'Shea et al. 1990).

The two serotypes identified in this study are of human health significance. Salmonella Virchow was first reported as a cause of invasive salmonellosis 40 years ago (Semple et al. 1968) and has since been identified as a more invasive serotype of S. enterica with a tendency to abscess formation (Nathwani et al. 1999, Banky et al. 2002) and higher attack rate in infants (Ashdown and Ryan 1990, Weinberger et al. 2004). Salmonella Virchow has been a predominant serotype in Queensland since the mid 1970s; however, it has been isolated with increasing frequency in the southern Australian states, eliciting the concern of human health officials (Bennett et al. 2003). A trend for the incidence of Salmonella Virchow to increase over time has also been detected in Israel (Weinberger and Keller 2005) with an accompanying increase in invasive NTS in children (Yagupsky et al. 2002). In North Queensland, it was the most common cause of invasive salmonellosis in infants from 1978 to 1988, accounting for 46% of septicemic cases of salmonellosis (Ashdown and Ryan 1990). Salmonella Virchow had a well-defined geographic distribution with 90% of Australian isolates over that period (1978-1988) occurring in Queensland and >78% of those infections coming from the coast north of Bundaberg (Ashdown and Ryan 1990). PT 8 is now the predominant PT of Salmonella Virchow across Australia (Sullivan et al. 1998, Bennett et al. 2003) and was the second most common isolate in Queensland with an incidence of 5.3/100,000 (OzFoodNet Working Group 2007). In Australia, Salmonella Virchow has been detected in chickens, constituting 4.8% or less of isolates from this species, but was not commonly found in other livestock or in raw meats (Heuzenroeder et al. 2004). In contrast, contaminated chicken carcasses appeared to be a common source of Salmonella Virchow in Israel (Solnik-Isaac et al. 2007) and were reported as a cause of outbreaks in the United Kingdom (Semple et al. 1968, Mani et al. 1974). The increasing detection of multiple antibiotic resistant isolates of Salmonella Virchow has caused concern in Israel and Europe (Simarro et al. 2000, Weinberger and Keller 2005, Solnik-Isaac et al. 2007) but appears not to have been detected in Australia. To date, the source of sporadic cases in North Queensland has not been investigated.

Salmonella Weltevreden has been previously isolated from geckos in Southeast Asia, Africa, and the Pacific (Bockemuhl and Moldenhauer 1970, Steel 1983, Oboegbulem and Iseghohimhen 1985, Murphy and Myers 1993). In Australia, Salmonella Weltevreden is not as commonly found in humans as Salmonella Virchow, but like Salmonella Virchow, it was found more frequently in Queensland than any other state or territory (National Enteric Pathogens Surveillance System 2006). Salmonella Weltevreden is one of the most prevalent serotypes in Southeast Asia. Studies in Thailand, Malaysia, and the Philippines reported it as the most common serotype isolated (Thong et al. 2002, Bangtrakulnonth et al. 2004). The Malaysian study showed Salmonella Weltevreden in a number of environmental sources, including well water, vegetables, beef, chicken products, and other raw meat (Thong et al. 2002). A study on the island of Okinawa, Japan, reported it as the second most common serotype isolated from humans (Kudaka et al. 2006), whereas in Vietnam it was the third most common serotype in humans and was also found in nonhuman hosts (cattle, pigs and chickens) (Vo et al. 2006).

This study, along with others conducted in several different countries and environs, demonstrates the potential of house geckos as reservoirs for Salmonella infection (Bockemuhl and Moldenhauer 1970, Helm 1981, Murphy and Myers 1993). The Asian house gecko may be a host of great significance in the spread of human salmonellosis, as it shares living space with humans and contaminates the immediate environment of houses with its feces. Gecko droppings tested for Salmonella in a Nigerian study showed that the bacteria can last up to four weeks in wet environments (e.g., tap water and wet sand), 6 weeks in direct contact with air, and as long as 8 weeks in dry sand (Otokunefor et al. 2003). The findings from the Nigerian study contrast with the fecal test results from this study: No viable Salmonella spp. were isolated from the nine fecal collections, although Salmonella was isolated from the gut contents of two of these same individuals. The fact we did not isolate Salmonella spp. from feces may have been due to dessicated specimens or that they were heavily contaminated with other organisms from the feces itself or organisms from the environment. A larger sample population of house geckos and their feces—collected in both the capture containers and house environment—may yield concordant results. Though no Salmonella spp. were isolated from gecko feces, gecko excrement is still considered the main route of contamination and transmission.

Contact with reptiles and amphibians was estimated to cause 6% of all sporadic NTS infections in the United States and 11% of all cases <21 years of age, responsible for ∼74,000 cases annually (Mermin et al. 2004). Recommendations regarding the prevention of transmission of Salmonella from reptiles to children were made in the 1990s (Warwick et al. 2001). Among them are recommendations that reptiles are not be kept in any households containing children under the age of 5 or immunocompromized persons; children under the age of 5 should avoid contact with all reptiles; and reptiles should not be kept in child-care centers. Since these are formal evidence-based guidelines and have resulted in enforcement in institutions that care for children, a disturbing question suggested by our study is whether children are at comparable or greater risk from uncontrolled Asian house geckos that share their living space in northern Australia and in other countries.

This study has identified a potential risk for NTS in northern Australia. More comprehensive studies are required to quantify this risk. These should include follow-up investigation of the houses of notified sporadic cases of human salmonellosis with molecular typing of isolates from geckos, other animals, and the environment to help identify sources. Consideration must also be given to developing evidence-based strategies to lessen the potential risk of salmonellosis from Asian house geckos. Since our survey was random, valid comparison is now possible in Townsville houses over time to look for variation by season and the effects of intervention strategies.

Footnotes

Acknowledgments

We thank Judy Forbes-Faulkner for laboratory assistance with Salmonella spp. isolations. Financial support for the study was provided by the Anton Breinl Centre for Public Health and Tropical Medicine (research quantum grant to Zak Callaway).

Disclosure Statement

No competing financial interests exist.

References

Ashdown

, Ryan

. Invasive disease due to Salmonella Virchow: a north Queensland problem. Med J Aust, 1990; 153:330–332334–335.

Australian and New Zealand Council for the Care of Animals in Research and Teaching (ANZCART). Euthanasia of Animals Used for Scientific Purposes, 2nd Reilly

. Department of Environmental Biology, Adelaide University, 2001.

Bangtrakulnonth

, Pornreongwon

, Pulsrikarn

, Sawanpanyalert

et al. Salmonella serovars from humans and other sources in Thailand, 1993–2003. Emerg Infect Dis, 2004; 10:131–136.

Banky

, Ostergaard

, Spelman

. Chronic relapsing Salmonella osteomyelitis in an immunocompetent patient: case report and literature review. J Infect, 2002; 44:44–47.

Bennett

, Dalton

, Beers-Deeble

, Milazzo

et al. Fresh garlic: a possible vehicle for Salmonella Virchow. Epidemiol Infect, 2003; 131:1041–1048.

Bockemuhl

, Moldenhauer

. The importance of house lizards (Geckonidae) for epidemiology of salmonellosis in South-East Asia. Z Tropenmed Parasitol, 1970; 21:307–312.

Cyriac

, Wozniak

. Infantile Salmonella meningitis associated with gecko-keeping. Commun Dis Public Health, 2000; 3:66–67.

Helm

. Role of Geckos in the Transmission of Salmonellosis in Hawaii with Emphasis on Sal. Weltevreden MPH Thesis. University of Hawaii, 1981.

Heuzenroeder

, Murray

, Davos

, Ross

. Salmonella typing and colonization of chickens by characterized Salmonella Sofia. Canberra: Rural Industries Research and Development Corporation, 2004.

10.

Hintze

. NCSS Trial and PASS 2002. Number Cruncher Statistical Systems: Kaysville, UT, 2001.

11.

Isenberg

. Clinical Microbiology Procedures Handbook, 1. Washington DC: American Society for Microbiology, 1992; 14.

12.

Kudaka

, Itokazu

, Taira

, Iwai

et al. Characterization of Salmonella Isolated in Okinawa, Japan. Jpn J Infect Dis, 2006; 59:15–19.

13.

Mani

, Brennand

, Mandal

. Invasive illness with Salmonella virchow infection. Brit Med J, 1974; 2:143–144.

14.

Mermin

, Hutwagner

, Vugia

, Shallow

et al. Emerging Infections Program FoodNet Working Group. Reptiles, amphibians, and human Salmonella infection: a population-based, case-control study. Clin Infect Dis, 2004; 38,Suppl 3:S253–S261.

15.

Messer

, Warnock

, Heazlewood

, Hanna

. Salmonella meningitis in children in far north Queensland. J Paediatr Child Health, 1997; 33:535–538.

16.

Murphy

, Myers

. A review of Salmonella infections in reptiles with particular reference to Gekkonidae. Amphibia-Reptilia, 1993; 14:357–371.

17.

Nathwani

, Morris

, Laing

RBS

, Smith

et al.

Salmonella virchow: abscess former amongst the contemporary invasive Salmonellae?

Scand J Infect Dis, 1999; 23:467–471.

18.

National Enteric Pathogens Surveillance System. Table 13. Top 25 Salmonella types identified in Australia, by state or territories, 1 April to 30 June 2006 [report on the Internet]. Australian Government: Department of Health and Ageing. 2006. www.health.gov.au/internet/wcms/publishing.nsf/Content/cda-cdi3003s.htmCited October 31, 2006.

19.

Oboegbulem

, Iseghohimhen

. Wall Geckos (Geckonidae) as reservoirs of Salmonellae in Nigeria: problems for epidemiology and public health. Int J Zoonoses, 1985; 12:228–232.

20.

O'Shea

, Speare

, Thomas

. Salmonellae from the cane toad, Bufo marinus. Aust Vet J, 1990; 67:310.

21.

Otokunefor

, Kindzeka

, Ibiteye

, Osuji

et al. Salmonella in gut and droppings of three pest lizards in Nigeria. World J Microbiol Biotech, 2003; 19:545–548.

22.

Oxoid Limited. Oxoid Biochemical Identification System (O.B.I.S.) Salmonella. 2006. www.oxoid.com/UK/blue/prod_detail/prod_detail.asp?pr=ID0570&org=124&c=UK&lang=ENCited May 24, 2008.

23.

OzFoodNet Working Group. Burden and causes of foodborne disease in Australia: Annual report of the OzFoodNet network 2006. Commun Dis Intell, 2007; 31:345–365.

24.

Semple

, Turner

, Lowry

DMO

. Outbreak of food-poisoning caused by Salmonella virchow in spit-roasted chicken. Br Med J, 1968; 4:801–803.

25.

Simarro

, Navarro

, Ruiz

, Miro

et al. Salmonella enterica Serovar Virchow with CTX-M-Like ß–lactamase in Spain. J Clin Microbiol, 2000; 38:4676–4678.

26.

Solnik-Isaac

, Weinberger

, Tabak

, Ben-David

et al. Quinolone resistance of Salmonella enterica Serovar Virchow isolates from humans and poultry in Israel: evidence for clonal expansion. J Clin Microbiol, 2007; 45:2575–2579.

27.

Speare

, Thomas

. Orphaned kangaroo and wallaby joeys as a potential zoonotic source of Salmonella. Med J Aust, 1988; 148:619–623.

28.

Steel

. World epidemiology of salmonellosis. Int J Zoonoses, 1983; 10:45–52.

29.

Sullivan

, Ward

, Rowe

, Woolcock

et al. Phage types of Australian isolates of Salmonella enterica subsp. enterica serovar Virchow. Lett Appl Microbiol, 1998; 27:216–218.

30.

Thomas

, Forbes-Faulkner

, Speare

, Murray

. Salmonelliasis in wildlife from Queensland. J Wildl Dis, 2001; 37:229–238.

31.

Thong

, Goh

, Radu

, Noorzaleha

et al. Genetic diversity of clinical and environmental strains of Salmonella enterica serotype Weltevreden isolated in Malaysia. J Clin Microbiol, 2002; 40:2498–2503.

32.

ATT

, Duijkeren

, Fluit

, Heck

MEOC

et al. Distribution of Salmonella enterica Serovars from humans, livestock and meat in Vietnam and the dominance of Salmonella Typhimurium Phage Type 90. Vet Microbiol, 2006; 113:153–158.

33.

Warwick

, Lambris

AJL

, Westwood

, Steedman

. Reptile-associated salmonellosis. J R Soc Med, 2001; 94:124–128.

34.

Weinberger

, Andorn

, Agmon

, Cohen

et al. Blood invasiveness of Salmonella enterica as a function of age and serotype. Epidemiol Infect, 2004; 132:1023–1028.

35.

Weinberger

, Keller

. Recent trends in the epidemiology of non-typhoid Salmonella and antimicrobial resistance: the Israeli experience and worldwide review. Curr Opin Infect Dis, 2005; 18:513–521.

36.

Wilson

. Asian House Geckos: fact sheet. Queensland Museum, Brisbane. 2006. www.qm.qld.gov.au/inquiry/factsheets/Geckos_20070605.pdfCited May 19, 2008.

37.

Yagupsky

, Maimon

, Dagan

. Increasing incidence of non-typhi Salmonella bacteremia among children living in southern Israel. Int J Infect Dis, 2002; 6:94–97.