The diversity of Anopheles blood feeding patterns suggests different malaria protection strategies in different localities

Introduction

Malaria is a disease that is transmitted by female Anopheles vectors. Generally, malaria control is achieved by mass deployment of insecticide-treated bed nets (ITNs) and indoor residual spraying (IRS) and thus such strategy has affected the biological activity of the vector. It has been shown that the distribution of ITNs is responsible for a reduction of 68% in malaria burden in sub-Saharan Africa¹. This control method has been widely distributed and a dramatic increase in use has resulted in the mass utilization of ITNs in many countries^2,3. Additionally, personal protection (i.e. repellents) has been found to be effective against mosquito bites and its use has led to a reduction in malaria infection^4–6. However, frequent daily application is required in order to ensure its effectiveness^7–9.

The efficacy of such protection strategies may be problematic as mosquito behavioral activities differ significantly between locations, as observed in Africa, where the vectors exhibit behavioral plasticity^10–15. The shifting behavior of the Anopheles vector is a factor that contributes to reduced ITN effectiveness. The behavioral changes of Anopheles mosquitoes are in the form of shifts to exophagic behavior^14,15 and biting time modification¹⁶. Several findings indicate the ineffectiveness of repellent against malaria infection. The limitations of repellent seem to be related to daily adherence and compliance^17,18 and disproportional utilization¹⁹. This issue may be due to the assessment of mosquito protection agents mostly focused on human factor rather than the impact that such types of protection have on the biological property of the vector.

The most effective method against Anopheles biting varies between sites and is dependent on the biting activity of the vector. In Uganda, intensive use of ITNs has been suggested due to the biting pattern of Anopheles gambiae, with biting mostly occurring late at night, during the time the human population is asleep²⁰. In contrast, bed nets may not provide proper protection against the same Anopheles species in Burkina Faso due to an early evening biting time²¹. Limited studies have investigated Anopheles biting patterns in the Indonesian archipelago²². Thus, our study aimed to specifically address the information gap of Anopheles biological properties in Indonesia.

Methods

Results

The parasitological assessment found a total of 211 cases of malaria in both localities²⁹. Only Plasmodium vivax was found in Jambi, responsible for 71 malaria cases. Participants from Jambi were 60.6% male (43) and 39.4% female (28) with a mean age of 15.5 years, ranging from one to 59 years. In Sumba, three types of Plasmodium were successfully detected during ACD. From a total of 140 malaria cases in Sumba, 92 (65.7%) were Plasmodium falciparum, 43 (30.7%) were Plasmodium vivax, and 5 (3.6%) were Plasmodium malariae. Participants from Sumba were 58.6% male (82) and 41.4% female (58) with a mean age of 10.9 years, ranging from one to 53 years. The calculated APIs of the two study sites were 3.56 and 15.96, respectively (Table 1). The API result of this study is different to the national health report of the Ministry of Health, Indonesia. The API is up to 2.95-25.4-fold higher at the sub-district level, found in this report, than at the provincial level, as stated in the report.

A total of 2,435 Anopheles mosquitoes were successfully collected from 216 houses and 216 catchers at the two locations (108 houses and catchers at each study site)²⁹. There was a statistical difference in the total number of Anopheles mosquitoes caught between Jambi and Sumba (P value= <0.0001, U = 5938). Jambi had mosquito abundance of 71 and Sumba had 2,364. Four Anopheles species were successfully collected in Jambi, namely Anopheles balabacensis, Anopheles barbirostris, Anopheles maculatus and Anopheles sinensis. An. balabacensis, which belongs to Leucosphyrus group, had the highest abundance, as shown with its relative abundance of 78.87 and HLR of 0.52 per person per night, followed by An. maculatus (relative abundance: 18.31 and HLR: 0.12 per person per night), An. barbirostris (relative abundance: 1.41 and HLR: 0.01 per person per night) and An. sinensis (relative abundance: 1.41 and HLR: 0.01 per person per night). In contrast, the dominant Anopheles species in Sumba were Anopheles aconitus and Anopheles sundaicus, with a relative abundance of 40.02 and 58.50 and HLR of 8.76 and 12.81 per person per night, respectively. The other minor species found were An. barbirostris (relative abundance: 0.09 and HLR: 0.02), Anopheles farauti (relative abundance: 0.04 and HLR: 0.01), Anopheles leucosphyrus (relative abundance: 0.04 and HLR: 0.01), An. maculatus (relative abundance: 1.06 and HLR: 0.23), Anopheles subpictus (relative abundance: 0.17 and HLR: 0.04) and Anopheles vagus (relative abundance: 0.09 and HLR: 0.02) (Table 2).

There was a difference in Anopheles biting time between Jambi and Sumba (Figure 1 and Figure 2). An. balabacensis from Jambi has a peak in biting time during early evening (6 pm), which decreases substantially until midnight, while An. maculatus showed an irregular biting time pattern. On the other hand, there is a similar trend in biting time between An. aconistus and An. sundaicus collected from Sumba; it gradually increased until its peak biting time between 21.00-22.00 and 01.00-02.00; then, it decreased progressively until 05.00-06.00. Additionally, an irregular biting time pattern has also been observed for An. maculatus from Sumba.

Figure 1. Biting time pattern of Anopheles balabacensis and An. maculatus collected from Jambi (mean +/- SD).

HLC, human landing catch.

Figure 2. Anopheles aconitus, An. maculatus and An. sundaicus biting times in Sumba (mean +/- SD).

HLC, human landing catch.

To investigate the biting preference of Anopheles mosquito, an indoor and outdoor comparison was carried out (Figure 3 and Figure 4). There was a statistically significant finding for the biting preference of An. balabacensis from Jambi; the number of collected mosquitoes from outdoor was higher than that of indoor collection (P value = 0.0004, U = 10634). No statistical difference was observed for An. maculatus (P value = 0.1163, U = 5614). A similar pattern was found for An. aconitus (P value = 0.3481, U = 36423), An. maculatus (P value = 0.6623, U = 7202) and An. sundaicus (P value = 0.1466, U = 38622), where there was no difference between indoor and outdoor collection, suggesting that undertaking an indoor or outdoor activity carries the same risk of getting mosquito bites.

Figure 3. Indoor and outdoor biting preference of Anopheles balabacensis (left) and An. maculatus (right) in Jambi (mean +/- SD).

HLC, human landing catch.

Figure 4. Mean number of Anopheles aconitus (left), An. maculatus (center) and An. sundaicus (right) indoors and outdoors in Sumba (mean +/- SD).

HLC, human landing catch.

To investigate the difference in mosquito biting times between Jambi and Sumba, a multiple comparison analysis of pooled mosquito sample data was carried out (Table 3 and Figure 5). Based on the mosquito biting time in Jambi, the number of bites during the early evening (18.00-19.00) was statistically different from other biting times, from 21.00-22.00 to 05.00-06.00 (P value= <0.00001, H = 76.32). Additionally, the number of bites at 19.00-20.00 was statistically different from the number at 24.00-01.00 and 05.00-06.00 (P value= 0.0435, H = 76.32). In Sumba, the number of bites during the early evening at 18.00-19.00 was statistically different from the other biting times, except for 19.00-20.00, 04.00-05.00 and 05.00-06.00 (P value= <0.0001-0.0069). In addition, the number of bites at 19.00-20.00 differed from the number at 21.00-22.00, 22.00-23.00 and 23.00-24.00 (P value= 0.0088-0.0168, H = 9052): 21.00-22.00 differed from 04.00-05.00 and 05.00-06.00 (P value= 0.0030-0.0052, H = 9052); 22.00-23.00 differed from 04.00-05.00 and 05.00-06.00 (P value= 0.0030-0.0050, H = 9052); 23.00-24.00 differed from 04.00-05.00 and 05.00-06.00 (P value= 0.0059-0.0099, H = 9052);and 24.00-01.00 differed from 05.00-06.00 (P value= 0.0347, H = 9052). These results indicate that in Jambi, the peak biting time is during early evening at 18.00-20.00. In Sumba, the mosquitoes started feeding and feeding gradually intensified during the early evening (18.00-21.00), the intensity of the mosquitoes was stable until 02.00 and then the mosquito biting intensity declined during the early morning.

Figure 5. Mean number of Anopheles mosquito at different biting times in Jambi (left) and Sumba (right).

HLC, human landing catch.

Discussion

According to the Malaria Atlas Project³⁰, for API <0.1, Plasmodium falciparum and Plasmodium vivax distributions are similar across the Indonesian archipelago. Plasmodium falciparum is more stable in distribution, where each part of Indonesian archipelago has the same pattern of low to moderate API. Meanwhile, Plasmodium vivax is more intense in the eastern part of Indonesia and unstably distributed in the western part of Indonesia. However, only Plasmodium vivax was found in Jambi, and more diverse Plasmodium species have been observed in Sumba, suggesting a different diversity of Plasmodium species distribution in the two localities. A discrepancy was also found in the calculated API between this study and the basic health report by the Ministry of Health of Indonesia, which might be explained by the different ways of presenting the data. The national health report²³ used the provincial population and the larger the area, the larger the population involved in the calculation, as API is calculated by dividing the total cases and the total population. API at a sub-district level is often observed to vary from one district to another and variation between districts is observed at a provincial level^31,32.

There are 20 Anopheles species known to be vectors for malaria in Indonesia. In this study, four and eight species have been found in Jambi and Sumba, respectively. The student t-test suggested a different abundance in the number of Anopheles mosquitoes between the two sites. This difference is often explained by environmental conditions. A distinct sampling time may cause this difference in mosquito abundance; however, since rainfall anomalies have been observed in Indonesia, this may not be the case³³. Since the existence of Anopheles breeding sites depends on rainfall providing a sufficient water bodies for the mosquitoes to lay eggs, rainfall anomalies in Indonesia may lead to an irregular pattern of mosquito abundance across time and place in Indonesia. The limited number of water bodies or humidity conditions may affect the habitat and abundance of Anopheles mosquitoes in Jambi^34,35. The difference in the annual incidence rate of malaria infection may also reflect mosquito abundance in different endemic areas. However, no correlation may be found if the correlation of annual incidence rate and mosquito abundance takes into account the species of Plasmodium³⁶.

The main Anopheles vector and biting preference differs between Jambi and Sumba. An. balabencis, which belongs to leucosphyrus group, is the primary vector in Jambi, as determined from its highest relative abundance and HLR. Moreover, An. aconitus and An. sundaicus are the primary vectors in Sumba, along with other minor Anopheles species found. Only An. balabacensis in Jambi was found to be exophagic, as previously known from the biting preference of this peculiar species³⁷. An. maculatus has been found to be both endophagic or exophagic similar to the finding of Elyazar et al.³⁷. However, previous studies have found that An. aconitus has an irregular pattern of biting preference while An. sundaicus is mainly exophagic³⁷. This study found that there was no significant difference between the indoor and outdoor biting preference of An. aconitus and An. sundaicus, suggesting that these species can be both endophagic and exophagic.

Biting time is essential to understanding the underlying biological properties of mosquitoes and to avoid Anopheles bites to control malaria infection. The data obtained suggest different biting times of Anopheles in Jambi and Sumba. Early evening (18.00-20.00) is most likely to be the mosquito feeding time in Jambi, when most people are undertaking activities and are unprotected. However, in the late evening (21.00-02.00), more people in Sumba may get Anopheles bites, reflecting sleeping time, when Sumbanese people may be vulnerable to infection with malaria parasites. This suggests the importance of ITNs for evading malaria infection in Sumba. The biting time of Anopheles in Jambi is similar to that in Halmahera, Maluku Island²². However, the finding from Sumba Island is different from other parts of Indonesia, which shows a gradual increase or decrease in the number of Anopheles mosquitoes in accordance with its biting time²². Furthermore, the difference in mosquito biting activity in each location could be simply explained by its dominant species at each location. For example, the early biting Anopheles activity in Jambi is explained by its dominant species of Anopheles balabacensis that exhibit an early biting time. Limited studies have tried to describe mosquito biting patterns in relation to the selection of malaria control strategies^20,21. This finding strengthens the previous report that effective malaria prevention depends on local Anopheles vector biting behavior. Anopheles vectors in Jambi share the same behavior as those in Burkina Faso, where bed net protection may not be effective for preventing biting exposure as Anopheles species in the area are dominant in the early evening²¹. In contrast, similar to Uganda, intensive use of ITNs combined with indoor residual spraying is the most effective protection approach for Sumba Island for avoiding malaria infection²⁰. Interestingly, studies conducted in Solomon island suggested that Anopheles farauti has a similar pattern of early night and outdoor biting behavior^38–40. Although, these studies recommended that LLINs and IRS are still significantly effective in reducing transmission based on the feeding cycle of Anopheles farauti, which is far shorter than the Plasmodium falciparum or Plasmodium vivax extrinsic incubation period. However, in an area in which the feeding cycle of the vector is unknown, study will be challenging. Additionally, our study also suggests that a vector control implementation will need to consider the dominant vector species, as a different location may have a different predominant Anopheles species, as well as continuous monitoring of such assessment via sentinel sites⁴¹.

Biting preference has previously been known to have an underlying genetic background⁴². For instance, chromosome inversions of 2Rbc, 2Ra and 3Ra are associated with exophagic and endophagic behavior in some Anopheles species^43,44. However, genetic background may vary within the genus and among mosquitoes within the same species in different locations⁴⁵. The finding also suggests that differences in Anopheles biting time may be an effect of different genetic backgrounds. Further research might explore this aspect.

There are some limitations of the current study. There was no intervention included to measure the effectiveness of any type of protection in correlation with the different biting times in each study site. In further research, an intervention approach should be used to find the best protection strategy in locations that may have different Anopheles biting times. Additionally, our collection method was limited to three weeks observational research. A more prolonged study needs to be conducted to reflect yearly fluctuations in local Anopheles biting times.

Conclusion

In conclusion, this study suggests four important findings for public health control: (1) API may be significantly lower at the provincial level compared to the sub-district level and varied accordingly, suggesting that malaria foci may be maintained in a locality from a provincial level, especially in areas of low to moderate endemicity; (2) the importance of mosquito abundance information may reflect malaria incidence rate in a location^46,47; (3) all Anopheles species, except An. balabacnesis, can be both endophagic and exophagic, suggesting a comprehensive protection approach is required to avoid mosquito bites regardless of being indoors or outdoors; (4) biting time may suggest the use a different prevention approach in each area; for example, people in Jambi may need to use mosquito repellent during activities in the early evening, while ITNs combined with indoor residual spraying may need to be deployed to protect malaria infection during sleeping hours in Sumba.

Data availability