Epidemiology and control of canine leishmaniasis in Peru and Brazil
The aims of the work presented in this thesis were two-fold: [i) to investigate whether domestic dogs are important reservoir hosts of ACL in a L. [Viannia]-endemic area where domestic transmission to humans has been reported [Part 1]; and [ii] to identify the shortcomings of currently practiced leishmaniasis dog control programmes and to evaluate whether topical insecticides could be used to control canine leishmaniasis [Part 2]. There is a growing belief that dogs [Canis familiaris] are peridomestic reservoirs of American cutaneous leishmaniasis [ACL] , as numerous studies have reported high ACL infection rates in dogs. The work described here is the first longitudinal study of ACL in dogs, and was carried out in 18 Leishmania [Viannia] -endemic villages of the Department of Huanuco, Peru. Over three years [1997-1999] a total 1104 dogs were surveyed, 104 of which prospectively. A polymerase chain reaction [PCR)] protocol to identify L. [Viannia] parasites in dogs was developed [Chapter 2], and, together with serology [ELISA. and IFAT], used to calculate prevalence and incidence of disease [Chapter 4]. The data was used to estimate the basic reproduction number [Ro] [Chapter 4] of canine ACL and to demonstrate a correlation between canine and human ACL incidences in the study villages [Chapter 5]. Several findings presented in Chapters 2-5 suggest that dogs are reservoir hosts of [peri-]domestic L. [Viannia] transmission in Huanuco, Peru. First, whereas the prevalence (3.8%) and incidence of clinical disease [4.2%] are low, the prevalence [25.6%] and incidence [29.0% per year] of L. [Viannia] infection in dogs is comparatively high. Second, the average duration of infection [2.2. years] can be as long as the mean life expectancy of an infected dog [2.5 years]. Third, L. [Viannia] parasites in dogs do not remain localised at the site of inoculation but are able to disseminate to both viscera and mucosa. Fourth, the detection of L. [Viannia] parasites by PCR in the blood of a high proportion of both symptomatic [32%] and asymptomatic [7.5%] dogs suggests that infected dogs are potentially infectious to sandfly vectors. Fifth, after controlling for inter-village differences in transmission rates, household dog ownership was shown to be a significant risk factor for human ACL. The results presented here show that if dogs were the main ACL reservoir host and if L. [Viannia] transmission were homogeneous, a dog control strategy (e.g. culling, insecticide-treated dog collars) in the study villages in Huanuco would be very feasible and effective, because the control effort [i.e. coverage] to reduce Ro<1 would be comparatively small [as low as 47%]. Domestic dogs are established reservoir hosts of ZVL caused by L. infantum. Hence, one of the approaches to reduce the incidence of human ZVL is to target infected dogs. The findings of a comparative study testing topical insecticides and applications to protect dogs from sandfly bites are presented in Chapter 7. It was shown that whilst permethrin and fenthion pour-on lotions had a more immediate effect on sandfly biting rates and mortality, deltamehtrin-impregnated collars [DMC] had a more prolonged protective effect, with the survival rate of bloodfed sandflies reduced by up to 86% after two months of deltamethrin collar application. In the work described in Chapter 8 the effectiveness of DMC to control canine ZVL was tested in a matched cluster intervention trial. Possible constraints associated with a community-wide implementation of a dog collar ZVL control strategy were also investigated using mathematical modelling. Although ZVL incidence was 32% lower in collared as compared to uncollared dogs after five months collar application, the difference was not significant. However, DMC did significantly reduce the odds (by 50%) of dogs increasing their anti-Leishmania antibody titre. Whether topical insecticides such as DMC will be effective as leishmaniasis control tools will depend on several factors. Firstly, the strategy will be most effective in those endemic areas where domestic dogs are the main ZVL reservoir and the epidemiological significance of wild reservoirs or stray dogs contributing to ZVL transmission is negligible. Secondly, in order to achieve a significant epidemiological impact on ZVL transmission, high dog collar coverage rates are essential. This will not only require the rapid replacement of lost collars, but also the collaring of new dogs recruited into the population; where population turnover rates are high, maintaining high coverage rates will be a greater logistic challenge, which invariably applies to dog populations in tropical, ZVL-endemic countries. Ultimately, the decision to replace the dog culling strategy with community-wide application of DMC will depend on [i] the relative cost of the interventions; and [ii] the practical applicability of DMC in the field [e.g. the willingness of the community to apply DMC and the efficiency with which they replace collars which have detached]. Clearly, the implementation of DMC on dogs is more likely to have the consent of the population at risk than the highly unpopular dog culling policy that continues to be practised in some ZVL-endemic countries.