HARRY STEELE-BODGER MEMORIAL SCHOLARSHIP 2012 & BVA OVERSEAS TRAVEL GRANT REPORT BY EMILY JEANES SRI LANKA 9 th July 11 th August 2012

Transcription

1 Introduction: HARRY STEELE-BODGER MEMORIAL SCHOLARSHIP 2012 & BVA OVERSEAS TRAVEL GRANT 2012 REPORT BY EMILY JEANES SRI LANKA 9 th July 11 th August 2012 Investigating the zoonotic risk posed by Toxocara canis infection of dogs in Sri Lanka. Sri Lanka is a developing country with a large dog population living alongside and in close contact with humans. In Sri Lanka, dogs are considered to have no financial value, therefore any healthcare provision for dogs is provided either because of their emotional value as pets, or because of public health considerations. Programs for reducing the public health risk posed by dogs are run by both the Sri Lankan government, and by Non-Government Organisations (NGOs); these focus almost exclusively on vaccinating to reduce the risk of rabies and sterilisation to reduce the stray population. This is in many ways a very sensible approach; since rabies is lethal if not treated quickly it clearly should be a high priority, and preventing overpopulation is a well-established method of reducing the risk from rabies (Reece and Chawla, 2006). However, there is very little thought given to other potential zoonoses that might be spread by dogs, or how the risk from these might be reduced. For example, published research has shown that approximately 20% of children in Sri Lanka are seropositive for Toxocara canis (Fernando et al, 2007). T. canis is the commonest parasite that causes pathology in humans (Beck, 1975). T. canis is a nematode whose natural host is the dog. Eggs are shed in the dog s faeces, and transmitted to new hosts by the faeco-oral route. If accidentally ingested by a human, the nematode can develop, and the developing larva migrates through the viscera before returning to the intestine to mature and reproduce. While the larval migration phase is usually asymptomatic, it can occasionally cause major problems. Aberrant migration of the larva to the eye (ocular larval migrans) can cause permanent blindness, and as the symptoms resemble ocular cancer it is not unknown for enucleation to be mistakenly used as treatment (Despommier, 2003). In other cases, the larvae migrate through the visceral organs, resulting in visceral larval migrans, a condition where inflammation due to the larva damages the viscera it travels through. This disease is severe, with diverse clinical signs including fever, abdominal pain or respiratory signs, depending on the organ the larva damages. Young children are most at risk of infection, as they often play in areas where dogs foul, sometimes eat earth and other non-food items (pica), and have a poor awareness of hygiene (Overgaauw, 1997). The high prevalence of antibodies to T. canis found by Fernando et al (2007) shows that around one in five Sri Lankan children have been infected with T. canis. Despite the high prevalence of T. canis infection in humans, suggesting that a high proportion of the dog population are shedding infectious eggs, I could find no research looking at the level of infection 1

2 in dogs. I therefore decided to investigate the zoonotic risk posed by T. canis infection in dogs. Since human studies have shown that toxocarosis is most prevalent in tropical regions, and in rural areas, I anticipated that the prevalence in Sri Lanka, a tropical predominantly rural country, would be very high. I hoped that my investigation would prove that the prevalence of T. canis shedding among Sri Lankan dogs is high, thus making it easier for veterinary workers to argue the case for routine anthelmintic treatment and to promote preventive measures e.g. good hygiene. I also hoped to identify any subsections of the dog population which shed more eggs than average, e.g. young puppies, as this would allow anthelmintic treatments to be targeted for maximal effect. I was going to be working with the Dogstar Foundation, a British run charity providing free veterinary healthcare to stray dogs and cats or to pets whose owners were unable to afford treatment. The Dogstar clinic is based in Randiniya, a small semi-rural village in the central province of Sri Lanka. I was planning on collecting faecal samples from dogs seen by the clinic. This would allow me both to collect data and also expand my veterinary knowledge of tropical diseases and improve my practical skills. Toxacara canis epidemiology: Fig. 1. Schematic to show the lifecycle of T. canis. From ( Z/Toxocariasis/body_Toxocariasis_page1.htm) As described above, T. canis is spread from dogs to humans by the faeco-oral route, with eggs present in the faeces causing infection after being ingested. However, this nematode has a complex 2

3 epidemiology, and can spread between dogs by several routes. The entire life cycle takes place within dogs; however other species in contact with dogs, e.g. humans, may be affected incidentally (Overgaauw, 1997). Once shed in the faeces, the eggs incubate in the soil, and become infective within 9-15 days under optimal temperature and moisture conditions. The optimal temperature is o C, making Sri Lanka (average temperature 27 O C in low lying areas ( an ideal climate for T. canis development (Parsons, 1987)). The infective eggs can survive in the soil for over a year (Deplazes et al, 2011). When ingested by the canine host, the eggs hatch in the duodenum, and undergo their first moult here. As L2 larvae (i.e. larvae that have moulted once), they migrate through the liver and lungs to the trachea and moult to become L3 larvae. They are then coughed up and swallowed, and so reach the stomach and small intestine. In the small intestine they undergo their final two moults and mature to adulthood. The adult worms then produce eggs continuously for up to 6 months, up to 20,000 eggs may be produced by one worm each day. These events only occur in puppies younger than approximately 5 weeks after that immune responses develop that prevent larval development past the second stage, i.e. the larval migration is halted in the somatic tissues (Gillespie, 1988). However, these dormant larvae can be reactivated during pregnancy after roughly 6 weeks of gestation. This is thought to be due to hormonal changes during pregnancy (Gillespie, 1988). The larvae then migrate trans-placentally to infect the foetus (Scothorn et al, 1965), or to the mammary glands where they are shed in the milk, thus infecting the pups orally (Zimmerman et al, 1985). The perinatal transmission from the bitch to her pups is almost 100% (Scothorn et al, 1965). Adult dogs have also been found to transmit Toxocara without showing symptoms, as they can be re-infected even if they have previously been infected trans-placentally (Farion et al, 2008) therefore they must also be seen as a potential zoonotic threat, even if the risk of transmission is less than that from juveniles (Deplazes et al, 2011) The Toxocara eggs are very sticky, and are known to stick to dog hair in considerable numbers. This means that petting dogs is another potential source of infection (Wolfe and Wright, 2003; Aydenizöz-Özkayhan et al, 2008; Roddie et al, 2008). However, it is thought that this poses a low risk of transmission to humans (Deplazes et al, 2011). Firstly, the majority of eggs found on a dog s coat have not yet developed enough to be infectious, and those that are embryonated are mostly unviable. Furthermore, they are very strongly adhered to the animal s hair, and so are hard to remove. It has been calculated that even if the coat was highly contaminated, several grams of hair would have to be ingested to cause an infection (Overgaauw, 1997a; Overgaauw and Von Knapen, 2000, 2004; Overgaauw et al, 2009; Keegan and Holland, 2010; Nagy et al, 2011) It is also possible for humans to be infected by ingestion of larvae in the tissue of paratenic hosts, i.e. in undercooked meat (Deplazes et al, 2011). Many species commonly farmed for their meat may act as paratenic hosts, including chickens, cattle and sheep. However, since the dog is the primary host, establishing the extent of Toxacara egg shedding in the dog population would indicate the risk of humans being infected secondarily by consuming infected meat, as well as primarily from direct ingestion of the egg. Furthermore, the majority of the population in the regions I was working rarely ate meat, so I do not believe this to be a major route of transmission. This is due to religious and socio-economic factors. The predominant religion was Buddhism, with Hinduism as the second most 3

4 common religion. Killing of animals is seen by many as a religious crime, therefore many people are vegetarian. Also, meat is too expensive to be a common food source for many people. Human Toxocarosis : Toxocarosis is one of the most common zoonotic diseases worldwide, and has been shown to be the most common human helminth infection in several developed countries (Hotez and Wilkins, 2009; Rubinsky-Elefant et al, 2010). Human toxocarosis is caused by infection with either T. canis, or the related roundworm T. cati. The natural lifecycle of these worms involves only dogs or cats (respectively) as the definitive hosts, however humans can act as incidental hosts for both species. Toxocara parasites cannot complete their lifecycle in humans; however the larval migration through the tissues can take place, causing various clinical signs (Rubinsky-Elefant et al, 2010). This was first recognised in the 1950 s. Wilder et al (1950) was the first to describe the condition ocular larval migrans, having identified nematode larvae as the cause of eosinophilic granuloma in the retinas of children. Two years later, Beaver et al (1952) named and described the condition visceral larval migrans, having discovered Toxocara larvae in the livers of infected children. Other, less serious conditions caused by Toxocara infection in humans have also been described, including common toxocarosis (Glickman et al, 1987), covert toxocarosis (Bass et al, 1983, 1987; Taylor et al, 1987, 1988), and other rarer respiratory, cutaneous and neurological conditions (Finsterer and Auer, 2007; Bede et al, 2008; Gavignet et al, 2008). Visceral larval migrans typically presents with pyrexia, hepatomegaly, abdominal pain, and lower repiratory signs, e.g. coughing or dyspnoea in 2-7 year old children(jacob et al, 1994).The most common organ to be affected is the liver, where eosinophilic granulomatous nodules form in the tracks of the migrating larvae (Musso et al, 2007). Cutaneously, the most common symptoms are pruritis, urticaria, eczema and vasculitis (Gavignet et al, 2008). Other rarer complications including myocarditis, neurological signs, arthritis and renal syndrome have also been described (Shetty and Aviles, 1999; Prunier et al, 2001; Rayes and Lambertucci, 2001; Finsterer and Auer, 2007) Ocular larval migrans is usually found in older children than those affected with visceral larval migrans, and seems to be most common in males (Taylor, 2001). It has also been diagnosed not infrequently in adults (Raistrick and Dean-Hart, 1975; Dean-Hart and Raistrick, 1977). Typically only one eye is affected; symptoms include decreased vision in the affected eye, strabismus and leukocoria an abnormal white reflection from the retina of the eye (Taylor, 2001; Rubinsky-Elefant et al, 2010). The eosinophilic granulomas formed in the eye have a similar appearance to a tumour, which can lead to enucleation being unnecessarily performed, although this is rare nowadays due to improved immunodiagnostics (Despommier, 2003). Common toxocarosis is found most frequently in adults, while covert toxocarosis presents in children (Bass et al, 1983, 1987; Glickman et al, 1987; Taylor et al, 1987, 1988). Common toxocarosis presents with chronic dyspneoa, weakness, urticaria, pruritis, and abdominal pain. Common laboratory findings include eosinophilia, elevated IgE, and high titres of Toxocara-specific antibodies (Glickman et al, 1987). Symptoms of covert toxocarosis include headache, pyrexia, behavioural abnormalities, sleep disturbances, respiratory signs, anorexia, abdominal pain, hepatomegaly, nausea and vomiting. Eosinophilia may or may not be present (Taylor et al, 1983, 1987). It is thought 4

5 that these may be the same condition, presenting in slightly different ways due to age related differences (Smith, 2009). Albendazole is currently the drug of choice for treatment of human toxocarosis, as it has kills Toxocara larvae in the visceral tissues. It appears to give better results than thiabendazole (Stürchler et al, 1989). It is impossible in visceral larval migrans to ascertain whether all the larvae have been destroyed, so treatment is continued until the clinical signs have abated. Since albendazole crosses the blood-brain barrier, it is effective in ocular larval migrans. Oral steroids are also necessary in ocular larval migrans to reduce the inflammatory response to the larvae, as this inflammatory response leads to further tissue damage. Covert and common toxocarosis are milder conditions, and will usually resolve without anthelmintic treatment (Rubinsky-Elefant et al, 2010). Despite its wide prevalence and the potential severity of the condition, the public awareness of toxocarosis is low, even in developed countries (Wells, 2007; Katagiri and Oliveira-Sequira, 2008). 5

6 Method: I chose to analyse the zoonotic risk posed by T. canis infection of dogs by analysing faecal samples for the presence of the infectious eggs. I chose this over other methods of assessing prevalence, i.e. by looking at seroprevalence, for several reasons: The presence of eggs shows that the dog in question is currently a risk to human health. Since dogs develop a degree of immunity to the parasite (Overgaauw, 1997) the presence of antibodies would merely show that the dog has been infected previously and probably at some point in its life was shedding eggs. However it may no longer be posing any zoonotic risk. Faecal analysis is a very quick, simple and cheap technique, and can be easily done in areas with limited laboratory facilities. The high level of seroprevalence reported in Sri Lanka indicates that the seroprevalence among dogs is almost certainly going to be very high. Looking at the seroprevalence in dogs would confirm that Sri Lankan dogs had a very high exposure to T. canis. However, it would provide no information on the zoonotic risk they pose to humans, which from a public health perspective is arguably the more important question. I would be analysing faecal samples firstly for the presence or absence of T. canis eggs, which would allow me to calculate the percentage of the dog population shedding infectious eggs in their faeces. I would also be quantifying the egg production using the McMaster technique (described below). This would allow me to identify which dogs were shedding most eggs, and so posed a greater danger to human health. The McMaster Technique: 1-3g of faeces was weighed out, and 14ml of sucrose solution (454g sucrose dissolved in 400ml water, Specific Gravity 1.275) added per gram of faeces. Three small stones were added and the mixture was shaken 30 times to break up the faeces. The resulting suspension was filtered to remove gross debris. The filtered solution was stirred and a sample was transferred to both chambers on a McMaster slide. The McMaster technique is a commonly used procedure to quantify worm egg counts. The slide consists of two layers of glass, and a faecal suspension of a known concentration ( e.g. here 1g faeces / 14ml solution) is put between the two layers. The slide has two areas of known volume delineated by a grid, this means that when the number of eggs within this grid is counted, this can be used to calculate the number of eggs per gram in the faeces. The slide was left for five minutes to allow the eggs to float to the top. The number of Toxacara eggs within the grid was counted on both sides, and the mean number of eggs was used to calculate the number of eggs per gram. The number of other worm eggs, e.g. strongyles was also calculated in the same manner. Faecal samples were obtained from dogs seen at the Dogstar clinic in Randiniya, and from dogs neutered by the Tsunami Animal People Alliance sterilisation program in Vavuniya. 6

7 Fig.2. Map of Sri Lanka, from The red dots mark the sites I was working, the northern one is Vavuniya, the southern one is Randiniya. I was hoping to be able to collect faecal samples that had been left along the paths and roads by the many dogs which roamed freely, which would enable me to obtain a larger number of samples from a wider group of animals. However, it turns out that when dogs are left entirely to their own devices, they toilet train themselves impeccably. In the month I used to collect samples, I only found two in accessible places. The general consensus from owners was that their dog toileted in the bushes somewhere. The figure below (Fig. 2.) shows a typical place where we assume the dogs toilet. I had to resort to obtaining faeces from the rectum of the more compliant patients. While this meant I could not obtain as many samples as I had hoped, it did however mean that I was certain which dog each sample had come from, and that there was no possibility for the results to be biased by accidentally obtaining multiple samples from one dog. 7

8 Fig. 3. A typical Sri Lankan garden where the dogs could roam I was also fortunate enough to spend a few days with the Tsunami Animal-People Alliance (TAPA), at their catch, neuter, release program in Vavuniya. As well as giving me an invaluable chance to practise my surgical skills, the enviable ability of these surgeons to neuter up to 30 animals in a day gave me a large number of dogs to collect faeces from, all conveniently unconscious! Vavuniya is a city in the north of Sri Lanka, and is still recovering from the effects of the civil war. The majority of dogs operated on were strays, and it was a safe bet that none of them had ever been wormed. This gave me the opportunity to compare the worm burdens of two population groups; the dogs from Vavuniya with no veterinary healthcare, and the dogs from Randiniya under the care of Dogstar which promotes routine anthelmintic use. T. canis and T. cati possible misidentification? T. canis is not the only nematode in the Toxocara family that can cause pathology in humans. Toxocara cati has been found to be another cause of Toxocarosis (Rubinsky-Elefant et al, 2010). T. cati has a very similar life-cycle to T. canis, the major difference being that the definitive host is the cat, not the dog. T. canis is generally regarded as the primary zoonotic pathogen, however, it is very hard to distinguish between the species so it is possible that a significant proportion of cases are caused by T. cati (Fisher et al, 2003). The evidence for whether T. cati is a significant human pathogen is mixed. On the one hand, a study in Iceland (where dogs are banned) showed that all adult Icelanders exposed to cats were seronegative for Toxocarosis (Woodruff et al, 1982). However, significant seroprevalence levels have been found in Islamic countries where dogs are avoided, and cats are the preferred pets (Smith and Noordin, 2006). It is very difficult to differentiate T. canis and 8

9 T. cati eggs under light microscopy (Uga et al, 2000). Since T. cati has been found in faeces from dogs (Fahrion et al, 2011), it is possible that a proportion of the eggs I identified as T. canis were in fact T. cati, particularly since there was a large semi-domesticated cat population living alongside the dogs and the humans. However, it is uncertain whether dogs can act as a definitive host for T. cati, attempts to experimentally infect dogs with T. cati eggs have failed (Fahrion et al, 2011). A possible explanation for the presence of T. cati eggs in dog faeces is coprophagy. Coprophagy is a common behaviour in dogs (Houpt, 1991), so it is possible that dogs ingest cat faeces, and the T. cati eggs pass straight through the digestive system and are consequently found in the faeces (Fahrion et al, 2011) It must therefore be considered a possibility that a proportion of the Toxocara eggs in the faeces were actually T. cati, not T. canis. It would be impossible to differentiate these without further investigation such as PCR identification (Uga et al, 2000). However, it is questionable whether identification of the exact Toxocara species is relevant. Since both species are zoonotic, both cause the same clinical syndromes in humans, both are spread in the same manner, and the same measures can be used to reduce the zoonotic risk from both species, I believe that attempting to distinguish the species would be a purely academic aim. A more important aim would be to attempt to reduce the risk to humans from either Toxocara species. 9

10 Results: Whole data analysis: Of the 75 faecal samples I obtained, 22.67% had T. canis eggs present. Quantification of the worm egg count using the McMaster technique found that the mean worm burden (of the samples where T. canis was present) was 1280 eggs / gram. However there was a wide range of Toxocara worm egg counts, the lowest positive count being 50 eggs / gram, and the highest eggs / gram. A histogram of the data (Fig.3.) shows that they are inversely distributed, so the mean will be skewed by a couple of high worm eggs counts, while the majority of worm egg counts will be much lower. Therefore, the median and upper and lower quartiles will provide a better description of the data spread. Frequency Distribution of T. canis Worm Egg Counts Worm Egg Count Fig. 4. Only positive samples were included in this histogram. There were 58 samples where no Toxocara eggs were found, and 2 samples were Toxocara eggs were present but could not be quantitated. The graph of the T. canis worm egg counts shows that the data are inversely distributed, i.e. the majority of dogs have low worm burdens, and a small proportion of the population have relatively high worm burdens. In two of the samples, T. canis eggs were clearly present, but outside of the grid used for quantitative calculation. Therefore, these samples were counted as positive for calculating the percentage of dogs shedding T. canis, but these eggs were not included in the quantitative worm egg counts. Comparison of dogs from Vavuniya (no known veterinary healthcare) and dogs from Randiniya (seen by Dogstar clinic): 28 samples were collected from Randiniya, and 47 from Vavuniya % (n=3) of faecal samples from Randiniya contained T. canis eggs, compared to 29.78% (n=14) of the samples from Vavuniya. Furthermore, when assessed quantitatively, the T. canis worm egg counts from Vavuniya appeared much higher than those from Randiniya (in each group one faeces sample had T. canis eggs present, 10

11 but outside the grid used for quantitative analysis, so these samples did not contribute to the worm egg count data): Randiniya (eggs/gram) Vavuniya (eggs/gram) Minimum Value Lower Quartile Median Upper Quartile Maximum Value Range Fig.5. Box and whiskers plot of the Toxocara egg counts at Randiniya and Vavuniya. The graph on the right presents the same data, but without the negative results (i.e. with no Toxocara eggs) for clarity. To see if there was a significant difference between the worm egg counts of these populations, I employed the Mann-Whitney U statistical test. While my data did not quite reach the 95% confidence interval (U = 800.5, z= -1.56, p=0.0594), it was very close, suggesting strongly that there is a difference between the two populations. I believe that if I had managed to obtain a larger sample size, it is very probable that a statistically significant result would have been seen. Comparison of dog ages: The results were stratified by the dogs age (adults in one group and juveniles and puppies in another group). Since many of the dogs were strays, and even for owned dogs records were not usually kept, the majority of the ages are approximate based on the animals physical characteristics and appearance. For these calculations results were not stratified based on the area the sample was taken in. Samples where the identity of the dog was not known, i.e. those collected off the roadside, were not included in this analysis. Of the 48 faecal samples taken from adult animals, 8 (16.67%) had T. canis eggs present in their faeces. In comparison, a much higher percentage (34.78%) of the juveniles and puppies had T. canis eggs present (8/23 samples). 11

12 Adults (eggs/gram) Juveniles and Puppies (eggs/gram) Minimum Value Lower Quartile Median Upper Quartile Maximum Value Range Fig.6. Box and whiskers plot of the Toxocara egg counts of juveniles and adults. The graph on the right presents the same data, but without the negative results (i.e. with no Toxocara eggs) for clarity. Again, there was not a statistically significant difference between the two groups when analysed using the Mann Whitney U Test (U=552, z=0.01, p=0.496). However, my data does show a tendency for juvenile dogs to have a higher number of Toxocara eggs in their faeces than adults. It is likely I would have shown this if I had managed to obtain a greater number of samples, since higher levels of Toxocara egg shedding in juveniles has been well described in the literature (Overgaauw, 1997). I was planning to compare stray and pet dogs to see if either of these groups posed a greater zoonotic risk to humans. However when in Sri Lanka I found that the definition of stray or pet was much more lax than in the UK, which made it hard to separate the dogs into meaningful groups. In the majority of cases, both pet dogs and stray dogs roam freely, and doubtless share the same social groups, scavenge in the same areas, and toilet in the same places. Many dogs are described as strays even though they have been living on a family s land for many years, and are fed daily, and sometimes had veterinary healthcare provided. Similarly, a dog may be owned even if it only turns up rarely and the people have very little interaction with it. Also, many stray dogs end up living in the Buddhist temples, were they are often well fed and looked after by the monks; even if they are not technically owned by the monks, the dogs have many or all of the benefits of ownership. Furthermore, almost all dogs in Sri Lanka are fed on human leftovers, predominantly curry and rice. Although there is an increasing market for specialist dog food as sold in the West, this is currently far too expensive for most Sri Lankans to buy. As both stray and pet dogs are free to scavenge as well, there is very little, if any difference in their diet. 12

13 Since there was no clear distinction between the husbandry of stray and pet dogs, I was unable to do any statistical analysis to compare these two groups. Total egg count results: As well as counting T. canis eggs, I also took note of the number of other parasite eggs that were present in the faecal samples. This allowed me to see if the patterns of Toxocara prevalence were mirrored in other species. Randiniya vs. Vavuniya Randiniya (total eggs/gram) Vavuniya (total eggs/gram) Minimum Value Lower Quartile Median Upper Quartile Maximum Value Range Fig. 7. Box and whiskers plot of the total worm egg counts at Randiniya and Vavuniya. The graph on the right presents the same data, but without the negative results (i.e. with no worm eggs) for clarity. I could find no statistically significant difference between the two populations using the Mann Whitney U test (U= 658, z=0.01, p=0.496); however the box plots suggest that the maximum egg counts are far higher in Vavuniya than in Randiniya. This suggests that there is little difference between the average faecal egg counts of these populations, but that there is a difference between the individuals with the highest faecal egg counts, i.e. those that contribute most to the worm burden of an area. Age: 13

14 Adults (total eggs/gram) Juveniles (total eggs/gram) Minimum Value Lower Quartile Median Upper Quartile Maximum Value Range Fig.8. Box and whiskers plot of the total worm egg counts of adults and juveniles. The graph on the right presents the same data, but without the negative results (i.e. with no worm eggs) for clarity. As with the Toxocara results, the results including other eggs in the quantitative analysis show a similar trend. According to the Mann Whitney U test, there is no statistically significant difference between the group (U=336, z=0.01, p=0.596), however the Box and Whiskers Plots show that there are individual juveniles with much higher faecal egg counts than any of the adults. Again, this suggests that juveniles make a greater contribution to the prevalence and intensity of worm egg shedding than adults. 14

15 Discussion: The data I have collected shows that a significant proportion (22.67%) of dogs in Sri Lanka are actively shedding T. canis eggs in their faeces. Since T. canis is known to cause pathology in humans, particularly young children, it seems wise that steps are taken to minimise the risk to human health. This can be done by both minimising the number of egg shed by dogs, and by minimising the likelihood of the eggs being ingested by humans. Possible strategies are discussed below. Previous studies have shown that the faecal flotation test is not very sensitive. A study by Lillis (1967) compared the faecal flotation results of 2737 dog faecal samples and 1480 cat faecal samples with autopsy results from 685 dogs and 318 cats from these groups. A sensitivity of 51% and a specificity of 100% were calculated for T. canis. The positive predictive value was calculated as 100%, and the negative predictive value as 88% (Lillis, 1967). Therefore it is possible that the actual proportion of dogs shedding Toxocara eggs in their faeces is up to double the number I found. This means that while my data shows 22.67% of the dogs I sampled were shedding Toxocara eggs, it is possible that up to 45.34% of these samples were actually positive. This means that the zoonotic risk of T. canis infection from dogs may be far higher than shown by faecal flotation tests. While there was not a statistically significant difference (p=0.0594) between the levels of eggs shedding between dogs seen at the Dogstar clinic in Randiniya, and those at Vavuniya with no veterinary care, the data does suggest that provision of veterinary care could reduce the number of eggs shed. This would subsequently reduce the risk of zoonotic disease. This is likely to be mainly due to the veterinary use of anthelmintics. Other possible factors include the larger proportion of neutered animals, meaning that T. canis larvae halted in the tissues of bitches will never have a chance to mature as the bitch cannot become pregnant, also the reduced number of puppies will decrease the overall amount of eggs shed as juveniles have been shown to shed T. canis eggs in the greatest quantities (Deplazes et al, 2011). It is also possible that the improved health of the dogs may also allow them to mount a more effective immune response against the parasite, hence reducing the number of eggs shed. My data also showed that puppies and juvenile dogs are more likely to be shedding Toxocara eggs than adults, and that the quantity of eggs in their faeces was also higher. This is an expected finding, and has been reported in the literature previously (O Lorcain, 1994). This is thought to be because older dogs are able to attenuate the migrating T. canis larvae in their tissues, thus preventing them reaching adulthood and reproducing (Gillespie, 1988). However, previous exposure to Toxocara larvae does not mean an adult dog will not shed eggs if it is infected a second time (Farion et al, 2008), so adult dogs still contribute to the total number of worm eggs shed in an area, even if juvenile dogs make a larger contribution. Recommendations to reduce the zoonotic risk from T. canis: Population reduction: Control of the size of the dog population is an established method of reducing the risk of zoonotic disease (Reece and Chawla, 2006). While most studies have focussed on reducing rabies transmission (Totton et al, 2011), it seems reasonable that the transmission of other zoonotic diseases would also be reduced once a smaller, healthier dog population is established. In developing countries population control is usually done by mass euthanasia or by catch-neuterrelease programs. Mass euthanasia has been shown to be ineffective in the long run, as the area is 15

16 soon repopulated by a new dog population. Mass euthanasia also carries animal welfare issues, as inhumane methods such as poisoned baits may be used (WHO, 2004; OIE Guidelines on Stray Dog Population Control, 2009; Dalla Villa et al, 2010). Also, for a dog population control program to be effective in the long term it has to have the support of the local people. Since the majority of the Sri Lankan population are Buddhist and Hindu, and consider the taking of life to be sinful, a euthanasia program would be highly controversial. A pilot study by Matibag et al (2009) looking at the knowledge and attitudes of rural Sri Lankans to rabies found that 60% of the population supported animal birth control methods of reducing the stray dog population, while only 24% were in favour of euthanasia as a control method. I can find no published reports of the effectiveness of catch-neuterrelease programs in Sri Lanka, however a comparable program in Jaipur in India reported a population decrease of 28%. Like the programs run by TAPA, Dogstar and other organisations in Sri Lanka, this program also vaccinated animals against rabies. The Jaipur program found that the rabies incidence in that area dropped to 0%, while the rabies incidence in neighbouring areas increased slightly, indicating that population reduction it is an effective means of controlling zoonotic disease (Reece and Chawla, 2006). Another study in Jodhpur, India, on the health of the dog population following an animal birth control program found that the average body condition score was higher in neutered animals after the program (before a low body condition score was the most common health problem, affecting 70% of animals)(totton et al, 2011). This might be because sterilisation reduces roaming behaviour and thus decreases calorie expenditure (O Farrell and Peachey, 1990) or because the slower metabolic rate caused by the removal of sex hormones encourages weight gain (Zoran, 2010). Since a low body condition score was the most common health problem before the program, a significant increase in body condition score following neutering indicates improved population health. However, no other health benefits to the dogs could be shown, and there appeared to be an increase in skin conditions in sterilised dogs (Totton et al, 2011). However, Totton et al were comparing individual scores of neutered and entire dogs in the short term after the neutering program. It is possible that in the future more benefits to the health of the dog population will be seen, e.g. a reduced population size will mean less competition for food leading to better nourished dogs, a more stable dog population may mean less territorial fights, and reduced pressure on dogs to move around to find food and other resources could mean they become more familiar with the local area and thus less likely to fall victim to road traffic. Furthermore, it is likely the local people will be more willing to care for the dogs once they know they are vaccinated against rabies. Also, the Dogstar foundation has found that people are happy to take on neutered females as pets, but are often unwilling to take in an entire female in case they have to care for the puppies (Personal communication). Therefore, the evidence shows that catch-neuter-release programs are effective in controlling the population size. This should have multiple benefits on reducing the zoonotic risk to humans from T. canis, as well as from other infectious diseases. Firstly, reducing the population size and increasing the adult proportion of the population should directly reduce the number of Toxocara eggs shed in the environment. Secondly, by encouraging responsible pet ownership, neutering programs might also increase the proportion of dogs that are being wormed effectively, again reducing the number of Toxocara eggs shed. Some studies have shown that neutered animals have a decreased prevalence of infection compared to entire animals (Visco et al, 1977; Lightner et al, 1978). The reason for this finding is unknown, but provides an additional argument in favour of mass neutering of dogs as a measure to reduce zoonotic disease. It is possible this is because neutered dogs are less likely to roam, and thus are less 16

17 likely to get infected. It is also possible that since the immune system is weakened in pregnancy, entire bitches are more likely to have patent infections, while the immune system of neutered animals is better able to arrest larval development and prevent patent infection. Another potential method of reducing the zoonotic risk would be to limit dogs ability to roam and forage, however this does pose both practical and ethical problems. Theoretically, if dogs could be confined to a house and garden, have a controlled diet and be supervised during exercise, they would be less likely to scavenge and indulge in coprophagia, and so would be less likely to be infected with T. canis. Also, their owners would be able to dispose of their faeces safely, thus preventing any worm eggs in the faeces from hatching and infecting other dogs or people. However, practically this would be very hard to enforce. Very few houses, particularly in rural areas, have enclosed gardens, so it is practically very hard to prevent dogs from roaming. Some owners do try and prevent their dogs from roaming by keeping them chained or in a small kennel. However this method poses serious ethical problems, as the dog is unable to exercise, or socialise. Furthermore, while this means the dog s faeces can be disposed of safely, it also means that the humans caring for the dog are more likely to contact dog faeces and so are more likely to be infected by T. canis. Therefore, I would not recommend preventing dogs from roaming as a measure to prevent T. canis infection. Veterinarians play an important role in preventing zoonotic disease, by educating pet owners on the diseases their pets can expose them to, and methods of controlling and reducing these risks (Deplazes et al, 2011). Education of people, and dog owners in particular, is important for other measures that reduce the zoonotic risk from T. canis. Owner education should be done in a balanced manner; while it is important to make owners aware of the possible zoonotic risks from their pets, pet ownerships has many benefits so they should not be given the belief that it is unsafe to keep their pet (Johnson, 2010; Lee et al, 2010). Toxocara eggs are highly resistant to environmental conditions, and can remain infectious for years. Therefore once an environment has been contaminated with them it is practically impossible to decontaminate it. The best ways to prevent infection is therefore to prevent the initial contamination of the environment, and to utilise good hygiene to prevent accidental ingestion of any eggs picked up (Overgaauw, 1997). Hand hygiene is a vitally important measure, reducing the risk of infection from T. canis, and from a plethora of other infectious diseases, both zoonotic and human. Children in particular, who are most at risk of toxocarosis, should be taught to wash their hands thoroughly before eating, and after handling animals or playing. Children should also be discouraged from geophagy and pica, since these habits increase the chances of them ingesting infected eggs from the soil. Since infective eggs from the soil can contaminate food such as vegetables, it is recommended that food is washed thoroughly before eating (Uga et al, 2009). Since Toxocarosis can also be caused by ingestion of paratenic larvae, i.e. by eating larvae within the meat of infected animals, it is recommended that meat is always thoroughly cooked (Deplazes et al, 2011). A very important measure to reduce the zoonotic risk from T. canis would be to implement routine worming programs. A recommended program involves deworming puppies at 2, 4, 6, and 8 weeks, 4 months, 6 months and 1 year of age, and thereafter twice annually (Overgaauw, 1997). Ivermectin for subcutaneous administration is readily available to veterinarians in Sri Lanka, and is also cheap. Ivermectin has been shown to be effective against both adult and larval stages of T. canis when 17

18 given subcutaneously at 200mg/kg. (Lower doses were less effective, although other parasites, e.g. whipworms were eliminated at lower dose rates) (Anderson and Roberson, 1982). This dose of ivermectin would also be effective against ectoparasites such as Sarcoptes Scabiei, a very common parasite in Sri Lanka. Pyrantel, another anthelmintic readily available in Sri Lanka, also has a proven efficacy against T. canis (Lindquist, 1975; Todd et al, 1975; Klein et al, 1978). Pyrantel is also useful against cestodes. Veterinarians could alternate between using ivermectin and pyrantel to help reduce the selection pressure for resistance to either anthelmintic, or chose the anthelmintic based on other clinical signs. If the patient is also suffering from ectoparasites then ivermectin would be the drug of choice, while if the patient is also thought to have cestodes then pyrantel would be indicated. The lower number of T. canis eggs shed by dogs seen at the Dogstar Foundation where regular anthelmintic use is encouraged suggests that regular worming can help reduce the number of eggs shed, and thus reduce the zoonotic risk. Worming of stray dogs would be beneficial in reducing the number of T. canis eggs shed, but hard to organise. Incorporation of worming as part of catch-neuter-release programs might have an effect for a while. However, for long terms benefits the treatment would have to be repeated, as it would be impossible to prevent reinfection of stray dogs. It might be possible to worm stray dogs by giving them anthelmintics disguised in food, but it would be difficult to ensure all dogs in a neighbourhood were treated, or that individual dogs didn t receive multiple doses. Potentially they could be sprayed with a coloured stock marker after they are given the wormer to prevent multiple treatments; however this could still be hard to coordinate. Probably a better option is fund more catch-neuterrelease programs, since the resulting reduction in the dog population would indirectly reduce the number of T. canis eggs being spread in the environment. One major problem with mass anthelmintic use would be the probable development of anthelmintic resistance. Most veterinary research into anthelmintic resistance has focused upon its effect on the livestock industry; however the same principles will apply to anthelmintic use in companion animals. If anthelmintic resistance becomes widespread in the nematode population, then we would lose a major weapon in our fight to control the nematode population (Williams, 1997; Waller, 1997). This would significantly reduce our ability to tackle zoonotic diseases such as toxocarosis (Overgaauw, 1997). Anthelmintic resistance could also compromise animal welfare, as it is possible widespread use of anthelmintics would lead to resistance among other parasites, e.g. Sarcoptes scabiei, leaving veterinarians unable to treat parasitic diseases effectively. To date, anthelmintic resistance has been little reported in small animal medicine (Kopp et al, 2008), however this may be due to a lack of surveillance rather than because resistance is not present (Sangster et al, 1999). Pyrantel resistance has been reported in Ancylostoma caninum in Australia, possibly as early as 1987 (Kopp et al, 2008). It is reasonable to assume that there is potential for anthelmintic resistance appearing among canine nematodes in Sri Lanka. Theoretically, if pet dogs are wormed regularly, but stray dogs are not, and the two populations are allowed to intermingle, the untreated stray dogs would create a worm population in refugia, i.e. not exposed to anthelmintics (van Wyk et al, 2002). This would mean that a substantial proportion of the T. canis population would not undergo selection pressure for antibiotic resistance, which hopefully would prevent anthelmintic resistance developing and becoming widespread. This strategy would also mean that the pet dogs which have most contact with people, and so probably pose a greater risk to human health have a reduced worm burden. Therefore I would recommend that 18

19 veterinarians encourage dog owners to worm their animals routinely, with particular focus on worming juvenile dogs and pregnant and lactating females who are most likely to be spreading the parasite. Conclusion: My data has shown that a large proportion of dogs in Sri Lanka are actively shedding the zoonotic parasite T. canis in their faeces. The close relationship and proximity of humans and dogs means that humans, particularly young children, are at high risk of infection from this parasite. I recommend that this risk is reduced as far as possible by educating people on the parasite, how it is transmitted, and the importance of good hygiene. I also recommend that all pet dogs and those with a primary human carer, particularly young puppies and pregnant and lactating bitches, are put on a routine anthelmintic program. Furthermore, I recommend catch-neuter-release programs as a method to control the stray dog population, and thus reduce the number of dogs which can transmit zoonotic diseases as well as improve the health of the dog population. Acknowledgements: This study was generously funded by the British Veterinary Association, through the Harry Steele- Bodger Memorial Scholarship and the Overseas Student Travel Grant. I also wish to acknowledge the help of the Dogstar Foundation, whose help and support was critical for this project. I am also very grateful to the Tsunami Animal People Alliance for allowing me to collect faecal samples from dogs in their capture-neuter-release program in Vavuniya. References: Aydenizöz-Özkayhan, M., Yagci, B.B. & Erat, S. 2008, "The investigation of Toxocara canis eggs in coats of different dog breeds as a potential transmission route in human toxocarosis", Veterinary parasitology, vol. 142, pp Bass, J.L., Mehta, K.A., Glickman, L.T. & Eppes, B.M. 1983, "Clinically inapparent Toxocara infection in children.", New England Journal of Medicine, vol. 308, pp Bass, J.L., Mehta, K.A., Glickman, L.T., Blocker, R. & Eppes, B.M. 1987, "Asymptomatic toxocariasis in children. A prospective study and treatment trial.", Clinical Pediatrics, vol. 26, pp Beaver, P.C., Snyder, C.H. & Carrera, G.M. 1952, "Chronic eosinophilia due to visceral larva migrans.", Pediatrics, vol. 9, pp Beck, A.M. 1975, "The public health implications of urban dogs.", American Journal of Public Health, vol. 65, no. 12, pp Bede, O., Sze na si, Z., Danka, J., Gyurkovits, K. & Nagy, D. 2008, "Toxocariasis associated with chronic cough in childhood: a longitudinal study in Hungary.", Journal of Helminthology, vol. 82, pp Dalla Villa, P., Kahn, S., Stuardo, L., Iannetti, L., Di Nardo, A. & Serpell, J.A. 2010, "Free-roaming dog control among OIE-member countries", Preventive veterinary medicine, vol. 97, no. 1, pp

20 Dean-Hart, J.C. & Raistrick, E.R. 1977, "Adult toxocariasis. Uniocular retinal lesions in the 20- to 50- year age group.", Transactions in Opthalmology, vol. 97, pp Deplazes, P., van Knapen, F., Schweiger, A. & Overgaauw, P.A.M. 2011, "Role of pet dogs and cats in the transmission of helminthic zoonoses in Europe, with a focus on echinococcosis and toxocarosis", Veterinary parasitology, vol. 182, no. 1, pp Despommier, D. 2003, "Toxocariasis: Clinical Aspects, Epidemiology, Medical Ecology, and Molecular Aspects", Clinical microbiology reviews, vol. 16, no. 2, pp Fahrion, A.S., Schnyder, M., Wichert, B. & Deplazes, P. 2011, "Toxocara eggs shed by dogs and cats and their molecular and morphometric species-specific identification: Is the finding of T. cati eggs shed by dogs of epidemiological relevance?", Veterinary parasitology, vol. 177, no. 1 2, pp Fahrion, A.S., Staebler, S. & Deplazes, P. 2008, "Patent Toxocara canis infections in previously exposed and in helminth-free dogs after infection with low numbers of embryonated eggs", Veterinary parasitology, vol. 152, no. 1 2, pp Fernando, S.D., Wickramasinghe, V.P., Kapilananda, G.M., Devasurendra, R.L., Amarasooriya, J.D. & Dayaratne, H.G. 2007, "Epidemiological aspects and risk factors of toxocariasis in a pediatric population in Sri Lanka.", Southeast Asian J Trop Med Public Health, vol. 38, no. 6, pp Finsterer, J. & Auer, H. 2007, "Neurotoxocarosis", Revista do Instituto de Medicina Tropical de Sao Paulo, vol. 49, pp Fisher, M. 2003, "Toxocara cati: an underestimated zoonotic agent", Trends in parasitology, vol. 19, pp Gavignet, B., Piarroux, R., Aubin, F., Millon, L. & Humbert, P. 2008, "Cutaneous manifestations of human toxocariasis", Journal of the American Academy of Dermatology, vol. 59, pp Gillespie, S.H. 1988, "The epidemiology of Toxocara canis.", Parasitology today, vol. 4, no. 6, pp Glickman, L.T. & Shofer, F.S. 1987, "Zoonotic visceral and ocular larva migrans.", Veterinary Clinics of North America: Small Animal Practice, vol. 17, pp Hotez, P.J. & Wilkins, P.P. 2009, "Toxocariasis: America s most common neglected infection of poverty and a helminthiasis of global importance?", PLoS Neglected Tropical Diseases, vol. 3, pp Houpt, K.A. 1991, "Feeding and drinking behavior problems", Veterinary Clinics of North America: Small Animal Practice, vol. 21, no. 2, pp Jacob, C.M., Pastorino, A.C., Peres, B.A., Mello, E.O., Okay, Y. & Oselka, G.W. 1994, "Clinical and laboratorial features of visceral toxocariasis in infancy.", Revista do Instituto de Medicina Tropical de Sao Paulo, vol. 36, pp Johnson, R.A. 2010, "Psychosocial and therapeutic aspects of human-animal interaction", Human- Animal Medicine: Clinical Approaches to Zoonoses, Toxicants and Other Shared Health Risks,, pp