Meta-Analysis of the Prevalence of Dirofilaria Immitis in Dogs from Grenada , West Indies

Cite this Article as: Camille-marie Coomansingh, Michael Yabsley, Nannette Wagner, Rhonda Pinckney, Muhammad I. Bhaiyat, Alfred Chikweto, Calum Macpherson and Paul J. Fields (2015), " Meta-Analysis of the Prevalence of Dirofilaria Immitis in Dogs from Grenada, West Indies", International Journal of Veterinary Medicine: Research & Reports, Vol. 2015 (2015), Article ID 429690, DOI: 10.5171/2015.429690 Research Article


Introduction
Dirofilaria immitis is the cause of heartworm disease in many domestic animal species but primarily in dogs.The parasite is restricted geographically mainly to regions with warm climatic conditions including Curacao, Trinidad, Brazil and Malaysia as 130 growing degree days (GDD) is required for L3 development which is demonstrated to take 8-10 days at 28-30oC, 11-12 days at 24oC, 16-20 days at 22oC and arrested development below 14oC (Cancrini and Gabrielli, 2007;Genchi et al., 2009) However with the increase of worldwide surface temperature of an average 0.74oC, temperate regions which were once considered free of D.immitis are now observing infections during the summer months (Vezzani et al., 2006 Human infections have resulted in adult worms being found in the heart and inferior vena cava (Takeuchi et al., 1981).More commonly observed symptoms from the somatic migration of third stage larvae leads to fever, cough and pneumonitis (Robinson et al., 1977;Theis, 2005) and also granuloma formation that calcify and form what are known as 'Coin Lesions', which can be misdiagnosed as cancerous (Monchy et al., 1993;Miliaras et al., 2010).
Due to the significant health risk to dogs and potentially to humans of Dirofilaria immitis, nine separate studies were conducted in dogs in Grenada over an eight-year period spanning 2002 to 2009.A meta-analysis combining the results of these studies provided a more precise estimate of the prevalence of heartworms in Grenadian dogs.(Martini et al., 1991).

Dogs testing positive in the 2007
Coomansingh study were found in all parishes in Grenada with no evidence of infection in animals less than 6 months of age (Figure 1).There was evidence that younger dogs had a higher prevalence of Dirofilaria immitis infection than older dogs, and particularly higher in St. John and St. Mark parishes (P<.05, using a Z-test for difference in proportions).No evidence was found of a difference in the proportion of positive tests between male and female animals (P>.50, using a Z-test for difference in proportions).All test results were confirmed by microscopic examination at 400x in the laboratory.

Results
Using meta-analysis to combine the results for the four studies on live dogs, 16 2).Similarly, combining the results for the five studies on necropsied dogs, 11.8% (95% confidence interval 7.9% to 16.3% ) were identified as positive for Dirofilaria immitis (Figure 3).Based on the meta-analysis results for the nine heartworm studies combined, the estimated overall proportion of dogs with Dirofilaria immitis infection from 2002 to 2009 was found to be 13.9% with a 95% confidence interval of 10.4% to 17.9% (Figure 4).The purpose of meta-analysis is to improve the precision of the estimate of a population parameter -in this study the prevalence of D. immitis in Grenadian dogs as measured by the proportion of positive cases.By combining the statistical evidence from multiple studies, the confidence interval around the prevalence estimate will become narrower than the confidence interval for a single study.In fact, the confidence interval will be narrower than for any one of the studies included in the analysis.
The narrower confidence interval thereby gives a more precise estimate of the prevalence of a disease.
We performed meta-analysis combining the evidence from nine studies each of which had measured the proportion of cases judged to be positive for D. immitis infection within a sample of dogs drawn from the population.

Conclusion
These results show that one in seven dogs in Grenada is likely to be infected with Dirofilaria immitis.Dogs should therefore be kept on year-round heartworm prophylaxis regimens as mosquitoes are common throughout the year.Long term integrated vector control program would be useful as Dirofilaria immitisinfections have been shown to be developing resistance to some drugs (Blagburn, 2011).Public health officials and veterinarians should collaborate in educating the general public, and the local physicians, about the potential health issues which can arise from heartworm infection in dogs and the need for prophylactic care due to the possibility of zoonotic transmission.This would need to be a vertical program with community-based participation, as studies in the USA suggest that only one third of veterinarians regularly discuss zoonotic parasites or methods of reducing risks with their clients (Allert, 1995).In Grenada, very few animal owners take their pets to the veterinarian and therefore public health education programs could assist in explaining the dangers of parasitic zoonoses and the benefits of deworming protocols for dogs.

Figure 1 :Figure 2 :
Figure 1: Proportion of Dogs Testing Positive for Dirofilaria immitis by Age and Parish in Grenada (Coomansingh, 2007)

Figure 3 :
Figure 3: Proportion of Necropsied Dogs Identified as Positive for Dirofilaria immitis in Five Studies in Grenada from 2005 to 2009 and Combined Using Meta-Analysis A meta-analysis was performed to combine the results of all nine studies in Grenada from 2002 to 2009.The Stuart-Ord inverse double arcsine square root transformation was used to stabilize the variance among the studies.Cochran's Q (P<.001) was used to test for heterogeneity and the I 2 statistic (73%, 95% confidence interval 37% to 85%) was used to test for inconsistency.These tests indicated the presence of statistical heterogeneity.The random effects model is used to synthesize heterogeneous research study results by determining a weighted average of the effect sizes of the studies.So, the random effects model was used to perform the meta-analysis in this study.The DerSimonian-Laird method was used for estimating the variance component in the model.In addition, the Begg-Mazumdar test with Kendall's tau (P=.180), the Egger test (P=.136) and the Harbord test (P=.561)were used to test for bias.These tests indicated no evidence of bias among the study results.