Control of malaria in the Comoro Islands over the past century

In the Indian Ocean, malaria spread via human migration and the transportation of mosquitoes by hurricanes and boats. Most data from this period were obtained through administrative reports written by crews of the boats. A high level of mortality by fever slowed down the colonization of Madagascar [11, 12], and in 1643, one-third of the “le Saint-Laurent” boat crew died from fever [13]. At the end of the 19th century, a dramatic malaria epidemic occurred in Mauritius and Réunion [14], followed by Mayotte [15]. However, the real period of the introduction of malaria in each Comoro Island is not well established, except for the Great Comoro [3]. According to White and Bowen, Mayotte was renowned as a healthy place without fever until 1703. However, from 1792 to the beginning of the 19th century, intensive slave trade occurred between the African continent, Mayotte and Bourbon Island, and successive invasions of Malagasy people occurred. Plasmodium falciparum and its vector An. gambiae likely settled in the island during this period. Indeed, Lebeau, a surgeon working in Mayotte from 1840 to 1848, described new epidemics of discontinuous fevers experienced by a large portion of the inhabitants of the island. Daulle confirmed this description between 1852 and 1857. The epidemic, death rate and morbidity related to malaria among Mayotte’s population was exhaustively described by Blin from the colonial administration [16]. In 1918, the splenic index in schools ranged between 80 and 100% [10, 16, 17]. Among navigators, the insalubrity of Mayotte and Moheli was usually opposed to the relative healthiness of the anchorage to the north of Anjouan and to the absence of fever in the Great Comoro until 1920 [2].

In Moheli, malaria (intermittent fever) was present as early as the 15th century [ 3]. During the 18th century, the quality of the anchorage of Anjouan was improved, but malaria was introduced by the middle of the 19th century [18, 19] before spreading to the centre of the island at the beginning of the 20th century. Similarly, the Great Comoro was considered a healthy place without anopheles due to the lack of permanent water until a first deadly epidemic of malaria occurred in 1920. An extension of malaria transmission throughout the island was registered during the wet season of 1923 [20]. From 1922 to 1924, epidemics occurred in all areas under 500 m of altitude with a very high mortality rate among the immunologically naïve population. A quarter of the population died during this time. The beginning of the epidemic was related to the building of water cisterns, where An. gambiae bred [21]. Supported by the economic expansion following the First World War, associated with the trade of vanilla, cisterns were built by the local population. Malaria transmission became rapidly hyperendemic in the island [2, 3]. In 1925, its spreading was confirmed throughout the island, particularly in the South Western and Western coasts [1]. During this period, malaria epidemics also occurred in Mauritius (1866–1867) and La Réunion (1868) [22].

The Culicidae fauna in Comoros remained virtually unknown until the end of the Second World War. However, the high mortality due to malaria and the occurrence of lymphatic filariasis led researchers to conduct several studies. In 1941, Lavergue (cited by [5]) described the presence of four Anopheles species and one Aedes species in the archipelago. A few years later, studies on Bancroft filariasis in Moheli, Anjouan, and Mayotte [23, 24] led to the description of seven new species: Anopheles coustani, An. funestus, An. gambiae, Anopheles maculipalpis, Anopheles mascarensis and Anopheles pretoriensis [21]. This settlement of Anopheles appears to be recent. The An. mascarensis species originates from Madagascar whereas the other five species are common all over the African continent. Blanchy also suspected the present of An. merus and of Anopheles arabiensis [3, 5, 21, 25], which were imported into the archipelago from the East African coast. Cytogenetic studies also suggested that An. gambiae s.s. [26] could have been transported from Africa during hurricanes [21]. Likewise, An. funestus was likely transported from Africa because the population found in Moheli is genetically distinct from the Malagasy population.

Facing this appearance of malaria, the archipelago colonial administration attempted several control strategies. In Anjouan, at the beginning of the 20th century, channels were dug in the valleys of Bandacoua and Chicony to drain water. In 1924–1925 in the Great Comoro, a systematic mass drug administration of quinine was introduced to block an epidemic of malaria that caused more than 1000 deaths [20]. Stocks of quinine were provided for each village and given for free. Simultaneously, coconut oil was spread on puddles, and cisterns were closed and controlled.

Studies were intensified and conducted by the colonial Health Services of Madagascar and Dependences, in cooperation with ORSTOM (a French Institute for research overseas). In 1949, the parasitological index in patients at the hospital varied between 10 and 30% and in rural areas between 69 and 83%. They confirmed that P. falciparum was the only Plasmodium transmitted on the islands.

Between 1945 and 1957, the first large campaign targeting the elimination of malaria was organized in Mayotte and Moheli using quarterly indoor house spraying (HIS) of DDT and dieldrin. In 1956, these actions were extended to Anjouan, and in 1957 to the Great Comoro. In 1972, preliminary studies were also conducted in Moroni [27]. However, due campaigns against DTT, experimentation with malathion was performed in Moheli but were not extended due to the smell of the product [28].

In Mayotte, quarterly fenitrothion in-house sprayings were very efficient and almost reached the goal of eradicating malaria [6]. They were extended in 1972 to Anjouan and Mohéli with the association of DDT–fenitrothion. From 1973 to 1974, temephos was sprayed to fight against mosquito larvae in Mayotte and in the Great Comoro [8, 17, 29, 30]. These actions were only conducted for a short time and were poorly evaluated. Few data are available on the effectiveness of these sprayings, but in 1972 and 1976, the plasmodic index was still 36.5 and 25.5%, respectively [16], in these two islands. The side effect of these programmes was the first description in 1967 of DDT-resistant Culex quinquefasciatus by Chauvet, rapidly followed by resistance to dieldrin and propoxur (1987), which can be the first sign of the selection of mutated insects under treatment pressure. The presence of Culex quinquefasciatus and Aedes aegypti is associated with the transmission of Wuchereria bancrofti and of yellow fever throughout the archipelago [31], but the nuisance induced by these insects is helpful to promote the use of bed nets and curtains by the inhabitants [32].

After independence, activities to control the vectors slowed down from 1975 to 1986. This was due to the cost of control programmes facing financial difficulties of the new independent administration and to political instability. Simultaneously, resistance to chloroquine was first reported in 1980 and confirmed during a systematic survey in 1984 [ 33, 34]. Chloroquine resistance was confirmed in vitro in 1987 [35]. By 2001, 29% of the isolates were resistant to chloroquine, and resistance to sulfadoxine/pyrimethamine reached 14%, leading to a review of the national treatment strategy [36, 37].

In 1869, the introduction of malaria occurred suddenly in Réunion 3 years after it appeared in Mauritius. Early in the 20th century, the first measures of control included the control of wetlands and damps areas, and the use of petrol to destroy larvae. However, these actions were usually applied locally with poor effects on the overall epidemic. In 1914, an office was created to coordinate prophylaxis and disinfection, and in 1949, the first control programme for malaria was established using DDT [44]. After preliminary studies in Mauritius [45] and in Réunion in 1948 [46], in-house spraying (IHS) was organized between 1945 and 1952 [45, 47, 48] in Réunion, followed by partial spraying until 1962. These actions were very efficient, and in 1973, the interruption of transmission was declared [14]. Since 1973, the strategy of malaria control has been based on the passive detection of cases, the control of imported cases, and the destruction of larvae [47, 49]. In Mayotte, the control of malaria followed the same process. Before 2001, the strategy was based on vector control, the systematic treatment of clinically suspected malaria cases, and chemo-prophylaxis for pregnant women [50]. Since 2001, rapid diagnostic tests are systematically used to confirm the diagnosis.

Due to the economic situation of the country, many Comorian people migrated abroad from their country during the past 40 years. In France, the Comorian population is estimated at 200,000 with 40,000 Comorian inhabitants in Réunion. Marseille (France) hosts more than 50,000 Comorians and is thus the largest ‘Comorian town’ in the world [51]. These expatriates used to return to the archipelagos, either during school/university holidays or for local festivals, particularly during the traditional “great weddings” period from July to September (a Great wedding is a traditional wedding separate from administrative and religious weddings and is organized by middle-aged people who want to obtain an honorific status in the community). These people are often infected by malaria during these periods and return to Europe with parasites. They represent the major source of imported cases of malaria in Marseilles [52, 53]. The French National Reference Centre for Malaria located in Marseilles collects sufficient data to be considered a reference centre for Comoros as well [54, 55]. However, due to the proximity of Mayotte (and its international airport), the population from Great Comoro travels more than the populations from the other islands, and most of the Comorians in Marseilles originate from this island. Indeed, imported cases in Marseille represent mostly the parasite populations of the Great Comoro, which differ from those from Moheli, Anjouan and Mayotte. Previous studies conducted in Marseilles showed that these parasites presented a higher frequency of mutations associated with drug resistance than isolates collected locally. Genetic analysis confirmed the proximity between malaria isolates imported in Marseilles and those from Great Comoro [56, 57]. The extrapolation of data obtained on malaria in Marseilles to the entire archipelago could thus be hazardous, particularly concerning drug resistance.

After many years of studies, the government decided to strengthen malaria control and to guarantee universal access to interventions through a national plan. This plan was based on an extended distribution of long-lasting insecticide impregnated bed nets (LLINs) and on the implementation of indoor residual spraying (IRS). These actions began in Moheli in 2008–2010, and 340,000 bed nets were distributed, reaching a theoretical 92% coverage of the all households and a 60% rate of use. Larvivorous fishes were introduced in the cisterns in many places in the Great Comoro and in some localities in Anjouan and Moheli. The IRS project began after preliminary studies on the sensitivity of the mosquitoes to insecticides and based on the cartography of the sites to plan resources.

By the end of 2007, a new programme was proposed on the island of Moheli (Fig. 2), conducted by an academic/industry Chinese consortium. Moheli was chosen because of its small population and the low transmission rate of malaria. A treatment of the entire population based on artemisinin–piperaquine (Artequick, Artepharm Co., China) was proposed, followed by primaquine alone, without any association with vector control. This strategy and this product are controversial because they were not previously fully evaluated in terms of efficacy or of acceptance by the population. Without clear evaluation of the programme, this action was resumed in Anjouan at the end of 2012. More than 80% of the people from Moheli and Anjouan received a curative dose of ACT (artemisinin–piperaquine, Artequick). A spectacular positive result was obtained thus far with more than a 98% decrease in the prevalence of malaria in Moheli and in Ndzouani. This result convinced the Comorian authorities to extend the experimentation to the island of Ngazidza in October 2015. This one-shot strategy gave much less convincing results due to the larger population and higher rate of transmission. Indeed, the proportion of patients attending consultations for malaria decreased slowly after these actions from 50% in 2004 to 36% in 2010. Only 53 deaths were registered in 2010. Therefore, in the population, the mortality of children under 5 years decreased sharply from 3.75‰ in 2005 to 0.75‰ in 2010. In the same year, the asymptomatic carriage of parasites decreased to 10.6% in children in Great Comoro, 9.1% in Anjouan and 5.4% in Moheli. A total of 103,670 malaria cases (mild and severe) were reported by public and private health centres in children younger than 5 years old in Anjouan and Great Comoro. In 2011, preliminary results from parasitological surveys showed a 6.3% asymptomatic carriage rate of parasites in the town of Fomoboni and 1.7% in the sanitarian district of Wanani [38], with an overall prevalence of 8.9%. In 2016, the threshold of pre-elimination was reached [35]. A total complete lack of information distributed to the public induced a context of suspicion among the population. No long-term evaluation of the efficacy was planned.