Seasonal distribution and environmental parameters associated with Brugia pahangi and Dirofilaria immitis in naturally infected dogs in Bangkok and vicinity, Thailand

Dirofilaria immitis and Brugia pahangi are vector-borne parasites found in dogs and cats, including Thailand. In order to evaluate the effects of season and environmental parameters on the prevalence of these parasites, this retrospective study was conducted in 2019. A total of 79,506 canine blood samples were examined. B. pahangi was found in 0.55% of samples (438/79,506; 95% CI 0.50–0.61) while D. immitis was detected in 0.43% (345/79,506; 95% CI 0.39–0.48). One-way ANOVA found no effect of seasonal conditions on prevalence. For B. pahangi, the parameters rainfall, relative humidity and sunshine hours showed associations with p ≤ 0.20 and were included in multiple logistic regressions resulting in adjusted odds ratios of 0.53, 1.31 and 0.55, respectively. For D. immitis, only average temperature showed p ≤ 0.20, resulting in an odds ratio of 0.42. In conclusion, Thailand has environmental parameters that do not change very much during the year, so they might not affect the prevalence of two filarial nematodes. However, the threat of B. pahangi and D. immitis should not be ignored, especially in subtropical regions where their vectors are abundant. Both owners and veterinarians should be concerned about filarial prevention and control of D. immitis and B. pahangi.

Dirofilaria immitis and Brugia pahangi are vector-borne parasites in dogs and cats that have zoonotic potential and are common in tropical, subtropical and some temperate regions of the world, including Thailand 1 . D. immitis is well known as a causative agent of heartworm disease in dogs and cats 2,3 . Microfilariae of the dog heartworm D. immitis present a subperiodicity without clear nocturnal or diurnal peaks, where a wave pattern is apparent but microfilaria do not completely disappear from the peripheral blood. However, the physiological periodicity is unknown 4 . It is also known as an occasional cause of pulmonary dirofilariasis in humans 5 . Lymphatic filariasis, a neglected tropical disease that is caused by filarial nematodes in the genus Brugia, affects approximately 80 countries around the world, particularly in dogs, cats and humans 6 .
Mosquitoes in the genera Mansonia, Armigeres and Aedes are potential vectors of dirofilariasis and lymphatic filariasis 7 . The tropical atmosphere is the most suitable for these mosquito vectors to survive. Increasing global temperatures and humidity are advantageous to the spread of mosquitoes and are also enhancing the effectiveness of pathogen transmission through mosquito-borne diseases such as dengue, malaria and lymphatic filariasis 8 . For instance, there is some evidence indicating that the prevalence of mosquito-borne disease in South America and South-East Asia relates to the El-Niño phenomenon. The relationship between El-Niño and the increasing risk of these diseases can be attributed to the rise in global temperature [9][10][11] .
Adults of D. immitis reside in the pulmonary artery and can induce endothelial damage. Some cases may develop into canine eosinophilic pulmonary granulomatosis caused by the infiltration of eosinophils 12 15 . Experimentally infected dogs showed abscesses in the adipose connective tissue around the popliteal node and nerve-cell tumours near the sciatic nerve. In one dog, lymphatic ducts were dilated distal to the popliteal node 16 . Dogs demonstrated a range of clinical signs, including episodic lymphadenopathy, lymphangitis, and limb oedema similar to the clinical signs reported in humans 17 .
Only a few studies performed in South-East Asian countries have reported the incidence and the population of animals affected by dirofilariasis and lymphatic filariasis [18][19][20] , and no studies have been designed to monitor the environmental factors affecting the distribution of these diseases. This study evaluated the association between seasonal and environmental factors related to the prevalence of D. immitis and B. pahangi infections in domestic dogs in Bangkok, Thailand and its vicinity.

Results
Brugia pahangi was found in 0.55% (438/79,506; 95% CI 0.50-0.61) and D. immitis was detected in 0.43% (345/79,506; 95% CI 0.39-0.48) of 79,506 samples tested during January to December 2019. The monthly detection rate of B. pahangi and D. immitis is shown in Table 1. The prevalence of B. pahangi infection was higher than that of D. immitis in all months.
The prevalence by season is shown in Fig. 1. The year was divided into three seasons; winter (November, December, January and February), summer (March, April, May and June) and the rainy season (July, August, September and October). The association between season and prevalence was not significant for either B. pahangi (p = 0.23) or D. immitis (p = 0.09).
The environmental parameters related to season, including rainfall, relative humidity, average temperature and sunshine duration, were selected based on vector and parasite biology. These parameters along with infection rates of B. pahangi and D. immitis are shown in Fig. 2.
Environmental factors, including rainfall, relative humidity, average temperature and sunshine duration, were analysed to reveal associations between infection rates and each factor using Pearson's correlation. The results revealed no significant relation for any parameter, as shown in Table 2; hence, no parameter had a significant effect and linear regression was not performed.
Since continuous data analysis using Pearson's correlation could not describe the association, categorical data analysis was performed by transforming the data using the average of each parameter as the cutoff: 20.38 mm rainfall, 29.78 °C average temperature, 5.52 h sunshine duration and 71.35% relative humidity, and the prevalence of B. pahangi 0.55% and D. immitis 0.43%. Categorical analyses including crude odds ratio and 95% CI are shown in Tables 3 and 4. The parameters showing p ≤ 0.20 were included in multiple logistic regression, which was used to calculate the adjusted odds ratio.
The results revealed that only average temperature showed an association with D. immitis infection (p ≤ 0.20). However, three parameters were included in the logistic regression in order to estimate associations between environmental parameters and B. pahangi infection. The multicollinearity between relative humidity and rainfall was checked using the Chi-square test and the results revealed no multicollinearity between these parameters. Unfortunately, no parameter was significant using logistic regression. The adjusted odds ratio (OR) for each parameter associated with B. pahangi infection was as follows: rainfall 0.53, sunshine duration 1.31 and relative humidity 0.55. www.nature.com/scientificreports/

Discussion
The current study presents the prevalence of B. pahangi and D. immitis circulating in owned dogs in Bangkok and its vicinity in 2019. A large number of the samples were collected from diseased dogs, which may have influenced the prevalence recorded herein. On the other hand, no antigen detection studies were carried out to detect amicrofilaemic animals, and only those that had detectable microfilariae in blood were counted. A limitation of this study was that samples were obtained from a clinical laboratory, and thus we do not have any details about the infected dogs. In this study, the monthly prevalence of the two filarial nematodes shown in Table 1 revealed that B. pahangi was more common than D. immitis. According to veterinarians, owners protect their pets from heartworm disease by using commercial products comprising extra-label ivermectin and its derivatives to control ticks, and these measures have been reducing the prevalence of D. immitis. In 2003, Nithiuthai reported that 10.2% of samples contained microfilaria of D. immitis (n = 83,476) in Bangkok during 1999-2001 21 . Unfortunately, given the lack of a prevention programme for B. pahangi, there is no medication to prevent B. pahangi infection; one of the important outcomes of this study was that the prevalence of canine lymphatic filariasis is higher than canine heartworm disease. Our analysis showed that prevalence was not influenced by the season. Interestingly, D. immitis was more prevalent than B. pahangi in weeks 17, 34 and 41. Thailand is in South-East Asia, which has a tropical savanna climate under the influence of the South Asian monsoon system. Some environmental parameters vary little throughout the year, including average temperature and humidity, thus making it difficult to associate these parameters with filarial nematode infection. There were no significant correlations between infection rate and these environmental parameters. However, sunshine duration seemed to show some correlation with B. pahangi infection rates (p = 0.19), as did relative humidity with that of D. immitis (p = 0.19). It was assumed that the longer daylight during summer affects some endocrine mechanisms in the dog, which may stimulate the female filarial worms to produce greater numbers of microfilariae 4 . Univariate analysis of B. pahangi prevalence revealed that three parameters showed p-values of ≤ 0.20: rainfall, relative humidity and sunshine duration, and so these were included in the multivariable logistic regression. Unfortunately, no parameter was significant in this regression; however, sunshine duration showed a positive adjusted OR of 1.31, which related to correlation. Vectors of most importance for filarial nematodes include Aedes spp. Culex spp. and Anopheles spp. and these are considered as potential vectors for D. immitis 22,23 , whereas Armigeres spp. and Manosonia spp. are considered as vectors for B. pahangi 24 . The OR-values of rainfall associated with B. pahangi and D. immitis prevalence were interesting: that for B. pahangi was 0.41, indicating that the rains are a protective factor, by washing floating water plants out  www.nature.com/scientificreports/ of swamp ponds and leading to a lack of suitable places for the life cycle of Mansonia spp. 25 , one of its potential vectors. On the other hand, the OR-value for D. immitis was 1.26, indicating that rain is a risk factor, due to clean water being suitable for Aedes spp. 23 , one of its potential vectors. Brugia pahangi and D. immitis are not the only two filarial nematodes reported in Thailand; other filaria include Brugia malayi 26 and Dirofilaria repens 27 . In this study, D. repens was found in one sample in December. Microfilaria of D. repens can be differentiated morphologically as they have two nuclei in the cephalic space. Molecular diagnosis is an alternative method for identification of D. immitis and D. repens.

Conclusions
In conclusion, environmental parameters in Thailand do not change much during the year, so they might not affect the prevalence of B. pahangi and D. immitis. The prevalence of these two filarial nematodes should not be ignored, and owners and veterinarians should be educated in the prevention and control of filarial nematodes in order to decrease the prevalence of these neglected canine vector-borne diseases.

Methods
Study design and sample collection. This retrospective study was conducted between January and December 2019. The results of blood examination were provided by the Vet Central Lab, which collected samples from private veterinary clinics and animal hospitals around Bangkok and its vicinity. All environmental parameters, including rainfall, relative humidity, average temperature and sunshine duration were obtained from the information service of the Thai Meteorological Department, Ministry of Digital Economy and Society. The data were the average from four stations in Bangkok: the Queen Sirikit National Convention Center, Bangkok Port, Thai Meteorological Department Bang Na and Don Mueang International Airport.

Blood examination and parasite identification.
A total of 79,506 EDTA-anticoagulated blood samples were collected from owned dogs and submitted to the Vet Central Lab. Buffy-coat thin blood smears were performed and stained with Wright-Giemsa stain. The positive microfilariae were examined by light microscopy.
Unsheathed and sheathed microfilaria were tested for acid phosphatase activity to identify species as D. immitis 28 and B. pahangi 28 , respectively. Statistical analysis. The prevalence of filarial worms was demonstrated using descriptive statistics with a 95% confidence interval (95% CI). The association between filarial worm infection and environmental parameters was analysed as continuous data using Pearson's correlation. Since one week was considered as a replicate, each parameter was represented as the average per week, except rainfall, which combined all data to represent one week. If these results were not acceptable, the infection rate and other continuous data were transformed using the average of each parameter or infection rate as the cutoff and analysed using the crude odds ratio at 95% CI. The parameters showing p ≤ 0.20 were checked for multicollinearity and included in multiple logistic regression. Statistical analysis used R software version 3.5.3.