Evidence of suppression of onchocerciasis transmission in the Venezuelan Amazonian focus

Background The World Health Organization (WHO) has set goals for onchocerciasis elimination in Latin America by 2015. Most of the six previously endemic countries are attaining this goal by implementing twice a year (and in some foci, quarterly) mass ivermectin (Mectizan®) distribution. Elimination of transmission has been verified in Colombia, Ecuador and Mexico. Challenges remain in the Amazonian focus straddling Venezuela and Brazil, where the disease affects the hard-to-reach Yanomami indigenous population. We provide evidence of suppression of Onchocerca volvulus transmission by Simulium guianense s.l. in 16 previously hyperendemic Yanomami communities in southern Venezuela after 15 years of 6-monthly and 5 years of 3-monthly mass ivermectin treatment. Methods Baseline and monitoring and evaluation parasitological, ophthalmological, entomological and serological surveys were conducted in selected sentinel and extra-sentinel communities of the focus throughout the implementation of the programme. Results From 2010 to 2012–2015, clinico-parasitological surveys indicate a substantial decrease in skin microfilarial prevalence and intensity of infection; accompanied by no evidence (or very low prevalence and intensity) of ocular microfilariae in the examined population. Of a total of 51,341 S. guianense flies tested by PCR none had L3 infection (heads only). Prevalence of infective flies and seasonal transmission potentials in 2012–2013 were, respectively, under 1 % and 20 L3/person/transmission season. Serology in children aged 1–10 years demonstrated that although 26 out of 396 (7 %) individuals still had Ov-16 antibodies, only 4/218 (2 %) seropositives were aged 1–5 years. Conclusions We report evidence of recent transmission and morbidity suppression in some communities of the focus representing 75 % of the Yanomami population and 70 % of all known communities. We conclude that onchocerciasis transmission could be feasibly interrupted in the Venezuelan Amazonian focus. Electronic supplementary material The online version of this article (doi:10.1186/s13071-016-1313-z) contains supplementary material, which is available to authorized users.


Background
Onchocerciasis is a chronic and cumulative skin and ocular disease caused by infection with the filarial nematode Onchocerca volvulus Leuckart and transmitted among humans through the bites of blackfly species of the genus Simulium Latreille. The embryonic stages of the parasite (microfilariae, mf ) migrate through the skin and cause severe itching, skin disease, and ocular lesions, with the severity of the clinical manifestations depending on the length of exposure to blackfly bites and the density of mf in the skin [1,2]. Visual loss and blindness can result from exposure to heavy parasite loads in the human host over time [2]. Since blackflies breed in fast flowing rivers, the disease is also known as 'river blindness' , although in the Americas it is called Robles' Disease after Rodolfo Robles, who described it one hundred years ago in Guatemala [3].
In the Americas, the infection was formerly prevalent in 13 endemic foci distributed in 6 countries (Brazil, Colombia, Ecuador, Guatemala, Mexico and Venezuela), where 565,232 persons were considered at risk of infection [4]. In Venezuela, there were three onchocerciasis foci (north-eastern, north-central, and southern), of which, the latter remains as the only persistent focus of the infection in the country [4,5]. The southern focus comprises endemic areas in the rainforest of the Upper Orinoco, Upper Siapa and Upper Caura River basins (in the Venezuelan Guayana), affecting the Yanomami indigenous group and extending beyond the border with Brazil to join the Yanomami Brazilian area to form the onchocerciasis Amazonian focus [6]. This is the largest focus by area in Latin America, affecting 26,715 people [4], of which 13,231 (49.5 %) are in the Venezuelan part of the focus. The southern focus shows an epidemiological spatial gradient, including areas of high transmission intensity with substantial levels of cutaneous and ocular morbidity observed prior to the start of the elimination programme. In the hyperendemic communities of the focus, skin disease was highly prevalent, with 24 % of the population affected by lichenified onchodermatitis and 10 % suffering from skin atrophy [7]. The pre-treatment prevalence of onchocercal nodules (onchocercomata), especially on the head, was 29 %, reaching 51 % in some communities (e.g., in Orinoquito). Presence of lymphatic lesions-including hanging groin-previously described in Africa was also reported [8]. Similarly, ocular pathology-up to 50 % prevalence of punctate keratitis, mainly due to the presence of mf in the cornea (MFC) and up to 75 % prevalence of mf in the anterior chamber of the eye (MFAC)-was a major clinical manifestation attributable to onchocerciasis. In some hyperendemic communities of the Parima area, the prevalence of any onchocerciasis-associated ocular lesions was greater than 50 %, reaching up to 70 % in those individuals aged ≥40 years. The prevalence of irreversible ocular lesions such as sclerosing keratitis (cumulative inflammatory lesions in the cornea that do not regress but cause progression to eye damage and irrecoverable loss of vision) reached up to 17 % in the Orinoquito area. Bilateral blindness due to onchocerciasis was observed in 0.45 % of the general population [8].
Simulium guianense sensu lato (s.l.) Wise, S. incrustatum Lutz, and S. oyapockense s.l. Floch and Abonnenc are the main vectors in the Southern Focus of Venezuela, with the former species being the most competent for O. volvulus and the predominant humanbiting blackfly in most of the hyperendemic areas of the focus [9,10]. Simulium incrustatum, with a lower vector competence than S. guianense s.l., contributes to the transmission of onchocerciasis in mesoendemic and some hyperendemic areas, whereas S. oyapockense is the main vector in hypoendemic communities with low intensity of transmission [10][11][12].
The strategy adopted by the Onchocerciasis Elimination Program for the Americas (OEPA) since its commencement in 1993 has included elimination of new (ocular) morbidity caused by O. volvulus, and interruption of transmission by 6-monthly mass administration of ivermectin (Mectizan®, donated by Merck & Co Inc), delivered by mobile teams with a therapeutic coverage ≥85 % of eligible population in all the endemic communities of the region, including hypoendemic areas [13,14]. (Given that, on average, approximately 15 % of the population are commonly not eligible for ivermectin treatment, this translates into a therapeutic coverage ≥70 % of the total population.) Ivermectin kills the mf and temporarily inhibits their release by gravid adult female worms [15], as well as killing adult worms after several years of mass treatment given at 6-monthly intervals [16,17]. More recently, 3-monthly treatments have been introduced in some communities in Mexico and Venezuela [6,18], given the results of clinical trials conducted in Guatemala [19] and Africa [20]. The OEPA strategy has led to the elimination of incident cases of ocular disease and the interruption of transmission in 11 of the formerly 13 endemic foci [4]; the two remaining foci being the Venezuelan and the Brazilian parts of the Amazonian focus.
The present work reports on the progress towards onchocerciasis elimination in southern Venezuela, according to the protocols proposed by the World Health Organization (WHO), which include in-depth parasitological, entomological and serological surveys, as well as guidance on operational thresholds [21]. Specifically, we report evidence of recent suppression (as defined in [21]) of O. volvulus transmission by S. guianense s.l. in 16 sentinel and extra-sentinel (previously hyperendemic) communities localized in different geographical areas of the southern Venezuelan focus after 15 years of 6monthly and 5 years of 3-monthly mass drug administration (MDA) of ivermectin.

Ethics approval and consent to participate
The parasitological, clinical, entomological and serological studies received ethical clearance from the Ethics Review Committee of CAICET (as part of the Ministerio del Poder Popular para la Salud-Venezuelan Ministry of Health). All the participants signed an informed consent form before undergoing any examination, testing, or agreeing to help as human attractants for entomological collections. Additionally, there was active participation of Yanomami volunteers (and their organization "Horonami") in the identification of new communities and distribution of ivermectin across the focus.

Study area and study population
Onchocerciasis transmission in southern Venezuela occurs in the lowlands (0-500 m above sea level, asl) and uplands (500-1200 m asl) of the Upper Orinoco, Upper Siapa and Upper Caura River basins (in the Amazonas and Bolivar States), which are part of the ancient Guayana Shield of northern South America, the oldest (3600 million years) region of the world (Fig. 1). Additional file 1 describes in detail the geographical and environmental characteristics of the focus pertinent to the transmission of onchocerciasis (see Text S1. Geographical and environmental characteristics of the Venezuelan part of the Amazonian onchocerciasis focus). The main onchocerciasis transmission seasons (regardless of the Simulium vector species present), occur during the dry to rainy (February-April) and rainy to dry (September-November) transitions [11]. The Yanomami indigenous group is the human population afflicted by Fig. 1 Venezuelan part of the Amazonian onchocerciasis focus. The legend lists the 12 geographical areas of the focus coloured by baseline endemicity of Onchocerca volvulus infection, from lowest (light blue) in Ventuari to highest (dark red) in Orinoquito. The numbers indicate the 31 geographical sub-areas described in Table 1 onchocerciasis in the Amazonian focus, with more than 25,000 inhabitants and four distinct linguistic subgroups (Yanomami, Yanomam, Yanam, and Sanemá), of which the Yanomami is the most affected in Venezuela (Amazonas State), followed by the Sanemá (Bolivar State).
The Venezuelan part of the focus encompasses 12 geographical areas: Padamo; Ocamo; Mavaca; Platanal; Guaharibos; Orinoquito; Parima; Chalbaud; Ventuari; Uasadi; Caura, and Siapa. Within these 12 main areas, 31 geographical sub-areas have been described ( Fig. 1 and Table 1). Further details of the endemic communities, mapping and geographical information system can be found in Additional file 1 (see Text S2. Endemic communities, mapping and geographical information system).
The (mostly semi-nomadic) Yanomami at-risk population has been estimated as 13,231 people, living in 241 'shaponos' scattered deep in the forest, and practicing shifting cultivation, hunting, fishing and gathering of forest products [6]. Their scanty clothing leads to an almost continuous exposure to biting blackflies. Further details a The numbering of the geographical sub-areas corresponds to that indicated in the map of Fig. 1 b Endemicity levels defined as: hyperendemic, microfilarial prevalence ≥60 %; mesoendemic, mf prevalence = 20-59 %; hypoendemic, mf prevalence <20 % [24] on how the anthropological features of the Yanomami influence their exposure to onchocerciasis have been presented elsewhere [6,22,23].
Baseline endemicity and survey communities for monitoring and evaluation According to OEPA guidelines [24], those communities with a microfilarial prevalence lower than 20 % are classified as hypoendemic; those with prevalence between 20 and 59 % as mesoendemic, and those communities with a prevalence of 60 % or greater as hyperendemic. Table 1 presents the number of communities thus classified and the population at risk and eligible for ivermectin treatment in the 12 geographical areas and 31 geographical sub-areas of the Venezuelan part of the Amazonian focus. The population at risk, 13,231 people as of 2015, was calculated based upon regularly updated demographic censuses conducted by the mobile teams that distribute ivermectin treatment. The population eligible for mass administration of ivermectin, 11,456 (86.6 %) people, were those aged five years or older, excluding those weighing less than 15 Kg (or measuring less than 90 cm in height), pregnant women and those breastfeeding a child younger than one week old (representing 13.4 % of the population). The pre-treatment levels of endemicity across all geographical areas were very heterogeneous (Table 1), with some areas including communities that were all hyperendemic (e.g., Orinoquito, Parima, Chalbaud), and others where hypoendemic communities prevailed (e.g., Padamo, Mavaca) [25]. In other areas (Ocamo), a gradual increase with increasing altitude in the proportion of hyperendemic communities has been observed [6,7,22], with hypoendemic communities in the low reaches of the Ocamo river (Lower Ocamo, 5 communities), mesoendemic communities in the middle reaches of the Ocamo river (Middle Ocamo, 5 communities) and hyperendemic communities in the upper reaches of the Ocamo river (Upper Ocamo-Shitari, 17 communities and Upper Ocamo-Parima, 16 communities), as summarized in Table 1. Overall, of the 241 endemic communities, 167 (69.3 %) were hyperendemic and mostly localized in the Ocamo, Siapa, Guaharibos, Chalbaud, Parima, and Orinoquito areas (Fig. 1), where the predominant anthropophagic blackfly species is S. guianense s.l. in most of the localities [6,22].
The criteria for selection of sentinel and extra-sentinel communities for regular monitoring and evaluation activities, consisting of in-depth epidemiological evaluations included: a) hyperendemic status, b) relative ease of accessibility by the mobile teams that conduct such evaluations, c) existence of historical, baseline epidemiological data prior to wide-spread ivermectin distribution; d) illustrative of the simuliid species composition of the focus. Accordingly, eight communities were selected as sentinel communities, namely (omitting the suffix 'theri' that designates the place name for a Yanomami village, for simplicity):  (16) Pokoshiprare (721 m asl; 90 inhabitants) in the Parima-Shamatari sub-area within the Parima area, the latter two communities derived from Yoreshiana A and Yoreshiana B (see [26]).

History of mass ivermectin treatment in the Amazonian focus
In the Amazonian focus, annual ivermectin distribution commenced in 1993 only in a few communities, and with a low mean therapeutic coverage (of less than 60 % up to 2000; Fig. 2a). This period is henceforth referred to as 'pre-ivermectin MDA' , given the low geographical and therapeutic coverage that had been achieved. From 2000 onwards (period henceforth referred to as 'during ivermectin MDA'), the onchocerciasis elimination programme in Venezuela was drastically re-organised under OEPA's strategic plan and started 6-monthly ivermectin treatment, with steadily increasing coverage. The 85 % coverage goal (for each treatment round) was reached in 2006 throughout the focus and it has been sustained since then (Fig. 2a). Treatment frequency was further increased from twice to four times per year in 45 communities during 2009 and currently, this quarterly treatment regimen has been extended to 192 out of 241 (80 %) of the endemic communities in the focus, albeit with a slight decreasing trend in coverage for the second to the fourth quarterly rounds (Fig. 2b). This treatment approach was adopted to accelerate interruption of transmission and to accelerate the death of adult worms, especially in areas with very high vector biting density, in communities whose mf prevalence and intensity seemed to have reached a new (lower than baseline) pseudo-equilibrium, or in communities that had been recently identified and incorporated into the programme at later stages. The number of treatment rounds per geographical sub-area that attained a coverage ≥85 % during 1995-2015 was calculated dividing the total number of treatments per sub-area in each round by the eligible population for this period (Fig. 3). Although the quarterly treatment regimen is more difficult to sustain at a consistently high 85 % coverage for each round, particularly for the most remote communities and during some times of the year, the greater frequency of visits to each community has meant that at least two complete treatment rounds with a coverage ≥85 % are received annually by each at-risk community. Table S1 of Additional file 1 provides details, for the 31 geographical sub-areas of the focus, of the number of twice-yearly and quarterly ivermectin rounds achieving ≥85 % therapeutic coverage. (The criteria for transmission suppression as reported in this article are described in Additional file 1: Text S3. World Health Organization (WHO) criteria for onchocerciasis elimination.)

Parasitological and ophthalmological surveys
Parasitological surveys were carried out in the sentinel and extra-sentinel communities of the focus at baseline (1981), pre-ivermectin MDA (1995)(1996)(1997)(1998)2000, the latter just before the twice per year treatment), and during ivermectin (twice a year and subsequently quarterly) MDA (2001,2008,2013,2015). The skin snip method was used to determine the prevalence and intensity of O. volvulus mf. Two skin biopsies from the (right and left) iliac crests were taken from each examined individual with a 2-mm Holth corneoscleral punch, and incubated for 24 h in buffered saline solution; the emerging mf were counted under an inverted microscope and snips were weighed to express results as mf/mg [8,26]. When it was not possible to weigh the snips in the field, an average weight of 1.62 mg was used, based on data collected in [22,26].
Ophthalmological evaluations were carried out during similar periods (1981 for baseline; 1997-1998 and 2000 for pre-ivermectin MDA, and 2001, 2008, 2015, during ivermectin MDA). The prevalence of microfilariae in cornea (MFC) and/or in the anterior chamber (MFAC) was determined by an ophthalmologist experienced in conducting onchocerciasis ocular evaluations for OEPA (HM). Ocular examinations were conducted with a slitlamp in a darkened area after patients were asked to sit with their head between their legs for 5 min to allow any mf present in the anterior chamber to settle in a visible position in order to determine MFAC [24,27]. The criterion of [27] was followed of taking into account only non-inflammatory keratitis lesions, with evidence of the presence of live or dead mf in the cornea (punctate keratitis stages A and B), as an indicator of MFC. (According to [27], inflammatory punctate keratitis (stages C, D, and E) is neither specific nor a reliable indicator of onchocerciasis-associated ocular disease.) Baseline and pre-ivermectin MDA data were obtained on i) skin mf prevalence and ii) skin mf intensity as both arithmetic (AM) and Williams (WM) mean numbers of mf/mg (in those aged 5 years and above), iii) community microfilarial load or CMFL (geometric mean number of mf per skin snip (mf/ss) in those aged 20 years and above [28]), iv) MFC prevalence, and v) MFAC prevalence. Additional file 1 describes the calculation of the parasitological indices (mf prevalence, AM, WM, CMFL, MFC and MFAC) (see Text S4. Calculation of parasitological indices). Detailed entomological evaluations were carried out between 1982 and 2000 in two sentinel communities (Koyowë (=Coyowë) and Niayopë (=Niyayowë) from the hyperendemic geographical areas of Orinoquito and Parima, respectively. In these communities, the proportion of flies with O. volvulus L3 larvae and the mean number of L3 larvae per fly were evaluated using classical dissection methods, examining the abdomens, thoraces, and heads of flies [10]. The baseline transmission indices for S. guianense s.l. were calculated according to [29].

During ivermectin MDA
In a selection of sentinel (Hasupiwei, Pashopëka, Koyowë) and extra-sentinel (Arokofita) communities, and during several consecutive collection days (e.g., three to five days or until twelve days in some cases), host-seeking S. guianense s.l. females were collected throughout the high O. volvulus transmission seasons (January-March/February-April or September-November) by using collecting protocols adjusted to the local transmission conditions previously established in the area and known periods of highest biting activity by parous flies [11,30]. All the simuliid females that landed on two human attractants selected from each community, working simultaneously but positioned at a distance of at least 50 m apart, were caught with manual aspirators by a team of two collectors during the first 50 min of each hour, beginning at 09:00 h and ending at 17:50 h, with one break of two hours at midday (12:00-14:00 h), due to a lull in biting density by parous flies during such period [30], for a total of 7 h of collection each day. Flies were collected before procuring a blood meal. Attractants received ivermectin one week prior to participating in fly collection in order to minimise the possibility of the flies becoming infected with ingested mf (if not caught soon enough after landing on the attractants; which could give positive results in the PCR analyses described below). Whenever possible, the collection teams in each community were the same throughout the surveys to minimize variations resulting from individual differences in catching ability. In the field, all hourlycaught flies were anesthetized with chloroform vapour, identified to species, and counted by community, date, day and hour of the day. The number of collection days depended on the biting density in each community in order to reach a number of at least 6000 flies as described below. Polymerase chain reaction (PCR) using O. volvulusspecific DNA probes has been generally applied to examine pools of flies in the OEPA region [21]. Therefore, for each community, flies were combined into pools containing 200 flies per pool (smaller pools containing remaining flies were also analysed), and the heads and bodies were tested separately for O. volvulus using a species-specific PCR assay [31]  purification and parasite detection have been published elsewhere [32]. To maximize the efficiency of the screening process, insect body pools were analyzed first; if any of those pools were positive, all of the head pools were then analyzed, providing an estimate of the infectivity rate (the prevalence of flies carrying only L3 infective larvae in the head). The Pool Screen® software (Version 2.0; University of Alabama, Birmingham, AL) was used to estimate the proportion of positive head pools in the PCR assay and the associated 95 % confidence intervals (95 % CIs) [33]. This software employs a Bayesian method to calculate the probability of infection of an individual blackfly from the number of positive pools and the size of the pools is used to calculate the infectivity rate in the community. Since parity status (proportion of flies that have already laid a batch of eggs) is a very laborious parameter to determine routinely in the field, OEPA's entomological criterion for cessation of treatment and commencement of surveillance has been simplified to <1 infective fly per 2000 (0.05 %) flies tested (i.e., assuming that 50 % of flies are parous). To reach this operational threshold, it has been determined that the minimum sample size required to have enough power to detect a statistically significant prevalence of infective flies lower than 0.05 % (i.e. not included in the 95 % CI), given that no infective fly may be found, is at least 6000 flies per community [21,34]. Additional file 1 provides details of the calculation of the transmission indices (hourly biting rate (HBR), seasonal biting rate (SBR), seasonal transmission potential (STP) and annual transmission potential (ATP) (see Text S5. Calculation of transmission indices).

Serological evaluations
Serological evaluations were conducted only during the ivermectin MDA period with the aim of measuring the prevalence in samples of children of IgG4 antibodies to Ov-16-a recombinant O. volvulus antigen able to signal prepatent infections- [35,36]. ELISA Ov-16 testing is currently being used for serosurveys of children in areas where transmission is deemed to have been interrupted in the Americas [33,37,38]. The serological protocol was as follows [36]. Sterile procedures were used to prick the fingers of all participants and four to six drops of blood (80-120 μL) were absorbed onto Whatman No 2 filter paper. The filter paper blood samples were dried, separated by sheets of paper, and then bundled and stored in sealed plastic bags in a cooler until they were returned to the laboratory where they were stored at −20°C. Two 6-mm punches of blood-saturated filter paper were placed in a phosphate-buffered saline-Tween 0.05 % and bovine serum albumin 5 % buffer and eluted overnight at 4°C. The elution was then run in duplicate in a standard ELISA to detect IgG4 antibodies against the Ov-16 recombinant antigen. A 5-year cumulative incidence rate of <1 new case per 1000 susceptible children (0.1 %) is the level acceptable in the OEPA region provided that the appropriate population size is available [21]. Here, and following [34], the prevalence of Ov-16 antibodies is taken as equivalent to this cumulative incidence rate. Consequently, to calculate a prevalence statistically significantly lower than 0.1 % (i.e. with a 95 % CI not including this value), and assuming no positives, a sample size of at least 3000 children <10 years of age is required. In the Amazonian focus, given the small community sizes, it is difficult to find this number of children. Consequently, we aimed to show a prevalence of Ov-16 < 1 %, requiring a minimum sample size of 300 children. We tested 396 children aged <10 years throughout the 16 endemic (sentinel and extra-sentinel) communities included in this study during 2013.

Parasitology and ophthalmology
Baseline and pre-ivermectin MDA Pre-treatment data ( Table 2) showed high prevalence of microfilarial infection by skin biopsy in sentinel and extra-sentinel communities, ranging from 56 % in Pokoshiprare to 100 % in Waharafitha and Toumawei (in the last two only individuals aged ≥10 years were examined). In general, the levels of initial endemicity varied from hyperendemic to very highly hyperendemic or holoendemic, with 11 out of the 16 communities (69 %) having mf prevalence ≥80 %. The highest values of infection intensity were recorded in Toumawei (AM = 231.2 mf/mg; WM = 102.6 mf/mg; CMFL = 104.7 mf/ss). Regarding ocular onchocerciasis, the prevalence of MFC was lowest in Awei (18 %) and highest in Waharafitha (50 %), the latter also being the community with the highest prevalence of MFAC (75 %).

During ivermectin MDA
The results of the parasitological surveys conducted at various time points during ivermectin MDA are shown in Table 3 and, for a selection of communities, graphically in Fig. 4. To avoid parasitological and entomological evaluations being conducted too soon after the last treatment round-which would lead to erroneous conclusions about the effectiveness of the programme-the immediately prior round of treatment was suspended. Therefore, in those communities receiving 6-monthly treatment, epidemiological evaluations were conducted one year after the last treatment round. In those communities receiving 3-monthly treatment, the evaluation surveys were conducted 6 months after the last treatment round. The prevalence of O. volvulus mf in most of the communities declined markedly from pre-treatment levels  (Fig. 4a) and MFAC (Fig. 4b), with the prevalence of MFAC decreasing to zero in 5 communities.

Entomological evaluations
Baseline and pre-ivermectin MDA Table 4 presents data on biting and infectivity rates of S.    [26] as well as the relationship between the proportion of infective flies and the mean number of L3 larvae per fly (Additional file 1: Figure S4). Additional file 1:    (Table 6).

Discussion
In this paper we present a compendium of the parasitological, ophthalmological, entomological and serological data obtained in hyperendemic communities of the Venezuelan part of the Amazonian onchocerciasis focus since the original studies conducted in the Parima and Orinoquito areas in 1981 [8]. These studies, and those presented previosuly [7,22,25,26] provided the epidemiological baseline situation prior to the introduction of ivermectin MDA. Both the geographical and therapeutic coverage of annual administration were low when the programme first started in a few communities in 1993 (Koyowë, Kanoshewë, Niayopë). In 2000 a twice per year treatment strategy was adopted, and in 2009, treatment frequency was increased to four times per year (Fig. 2). Overall, skin microfilarial prevalence and intensity have declined substantially, with reductions in prevalence ranging from 58 % ( Table 4). This is despite these communities having received 35 rounds of ivermectin with a coverage ≥85 % of eligibles (~75 % of the total population) by 2015 (Fig. 2c). By contrast, communities located in the Parima sub-areas (Niayopë, Masiriki, Toumawei, Arokofita, Okiamo, Warapawë, Kakarama and Pokoshiprare) have experienced greater    Table 4).
In the OEPA region, absence or near absence of L3 larvae in the head of blackfly vectors (as measured by poolscreen PCR in samples of 6000-10,000 flies), a 99 % reduction in the intensity of transmission (as measured by seasonal transmission potentials), and the absence of detectable O. volvulus infection (by parasitological or immunological diagnostics) in children have been the WHO criteria followed to certify focal interruption of parasite transmission [21]. Prior to reaching this epidemiological status, the focus starts to show declining to very low or negative parasitological results in skin, eyes (indicators of reversible morbidity) and flies, suggesting that transmission of the infection has been suppressed by the treatment [21,34]. Here, we report 81 % reductions in fly infectivity and 97 % reductions in seasonal transmission potentials, with an overall prevalence of 7 % in Ov-16 seroprevalence among children aged up to 10 years and of 2 % among those under 5 years, providing evidence of suppression of O. volvulus transmission by the most competent vector of the focus, S. guianense s.l., in areas formerly hyperendemic to holoendemic. The dramatic decline in the seroprevalence among children and the lack of evidence of parasitevector contact suggest that four times per year treatment has successfully suppressed transmission in many of the endemic communities. It will now be necessary to maintain pressure on the parasite population, continuing to suppress transmission, until the existing adult female parasites are either killed or rendered sterile by repeated ivermectin treatments [20].
Ivermectin is expected to have a faster impact in those areas of the Amazonian focus with lower vector competence blackfly species (e.g. S. oyapockense s.l. and/or S. incrustatum [12,39], which sustain hypo-to    Numbering of sub-areas as in Table 1 and Fig. 1 mesoendemic transmission [22], or where vector biting rates are lower and perhaps closest to critical biting rates necessary to maintain endemic transmission (basic reproduction ratio, R 0 ≥ 1 [40]). These threshold biting rates do not only depend on vector competence for O. volvulus, but also on the human blood index (HBI, the proportion of vector blood meals of human origin) of the various blackfly species and populations therein. At present, this parameter is unknown for the simuliid species prevailing in the Amazonian focus, but field and theoretical studies on S. damnosum s.l. in West Africa indicate that the HBI can be highly variable [41] and possibly host and fly density dependent [42]. Given that the Venezuelan part of the Amazonian focus is sparsely inhabited by human populations, it is likely that the HBI is relatively low and threshold biting rates correspondingly high. Serological data (albeit with low sample sizes reflected in the 95 % CI shown in Modelling studies conducted with the EPIONCHO transmission model in African savannah settings [43] have suggested that switching to a twice per year treatment strategy during an ongoing annual treatment programme can substantially decrease (nearly by half in highly hyperendemic settings) the additional number of years required to reach the provisional, operational mf prevalence thresholds suggested by the African Programme for Onchocerciasis Control [44] to achieve focal elimination of onchocerciasis. In hyperendemic settings this switch can lead to cost savings. However, these provisional thresholds are not equivalent to transmission breakpoints. A recent comparison of the (stochastic) ONCHOSIM and (deterministic) EPIONCHO models highlights this difference, and indicates that more than 20 years of twice per year treatment (40 treatment rounds) would be required to drive the parasite population to elimination when the initial microfilarial prevalence is greater than 90 % and the coverage of treatment is between 65 and 80 % of the total population (80 to 95 % of eligible individuals) [45]. These projections appear compatible with the results seen in the Amazonian focus in those holoendemic communities in which vector biting rates are very high such as Koyowë. Quarterly treatments may help to accelerate progress to elimination in these communities by further reducing the amount of transmission that takes place between consecutive ivermectin rounds, increasing effective coverage, and exerting a macrofilaricidal effect [19,20]. Interestingly, the presence of live fertile worms in the host population, as well as of infected (mf positive) people is predicted for 2015 in Koyowë by simulations conducted (by DR) with the EUSIMON model (see [46] for a published precursor), a community with the highest number of treatment rounds (37). This model also predicts the occurrence of transmission and of seropositivity in children by 2015. However, the EUSIMON simulations conclude that the chance of recrudescence in this locality is low if ivermectin treatment were to be interrupted, provided there is no migration of infected people and/or flies into the assumed closed population.
More likely, in the Amazonian focus, and due to the Yanomami network of kinship ties, alliances and hostilities [6,22], onchocerciasis occurs in a network of interconnected nodes reminiscent of a metapopulation structure, the degree of connection depending both on vector and human movement-the latter being possibly more important and/or better documented. This spatial structure has important repercussions for onchocerciasis transmission and control, as some sub-areas/communities that by themselves may not be able to sustain endemic transmission, may receive an important and periodic influx of heavily infected people from highly endemic areas, making it possible for the infection to persist or be re-introduced. This potential exchange of parasites between otherwise different transmission zones by virtue of the Yanomami micro-and macro-movements [47], may also weaken potential barriers to gene flow. This may allow spread of onchocerciasis from currently non-controlled or less well controlled areas, calling for the development of spatiallyexplicit, patch transmission models and anthropologybased research avenues for onchocerciasis control in the Amazonian focus. The transport of Onchocerca parasites along the reticular nature of the Yanomami use of space may indeed be very diffuse, necessitating intensive treatment in all sub-areas; however, if particular networks could be identified as being responsible for most transmission, a more targeted approach could be beneficial towards the goal of achieving elimination.

Remaining challenges and directions for future work
The semi-nomadic characteristics of the human population, the remoteness of the Yanomami territory, the holoendemic status of some areas, and the ongoing identification of new endemic communities in the Venezuelan part of the Amazonian focus constitute the main challenges for the elimination of onchocerciasis in the Amazonian focus. From 2009, ivermectin treatment frequency has been increased to four times per year in 80 % of the hyperendemic communities in an attempt to hasten interruption of transmission in areas showing slow progress or in communities recently identified and incorporated into the programme at later stages. Sustaining a high geographical and therapeutic coverage for each treatment round is essential. Only the first round of the 2012-2014 quarterly distributions reached the ≥85 % goal. Besides, since the sentinel and extra-sentinel communities included in this study were partly selected because of relative facility of access, there is the possibility that more remote communities received a lower treatment coverage and/or frequency. Currently, there is a total of 72 communities identified as being remote (~30 % of the total), with a population of 3359 individuals (~47-50 individuals per shapono). This represents approximately 25 % of the total population. Most of these communities are receiving regular treatment, and many of them have received more than 8-12 treatment rounds (Pasumopë, Chalbaud, Mayo and Hashimu subareas). However, there are communities located in the Upper Siapa and Upper Ocamo-Parima sub-areas that are only accessible by helicopter (25-30 communities, with~1000-1500 individuals). This represents only 11 % of the total population and 12 % of all the communities under treatment. Treatment is delivered to these communities when helicopter support is available. Depending on their endemicity status and their connectedness with well-controlled areas, these less accessible communities could pose a risk of infection re-introduction. The use of high spatial-resolution satellite data to identify remote communities in the rainforest is a strategy currently used in an attempt to delineate the extent of the focus, the distribution of transmission zones [6], and the intensification of treatment efforts that will be required to achieve elimination over the entire focus. The prospect of deploying test-and-treat doxycycline treatment as a complementary macrofilaricidal therapy in communities with suppressed transmission has also been considered [48]. Finally, in May 2014, under the auspices of the WHO, Brazil and Venezuela signed a bi-lateral memorandum of understanding calling for a closely coordinated effort between both national programmes in order to reach the goal of onchocerciasis elimination from the Yanomami area. A first joint meeting was held in February 2015, and an action plan for 2015-2016 was agreed and is being implemented [49].

Conclusions
Our results contribute to the ongoing success of the OEPA strategy [4,5,14,33,37,38]. This success has spurred prospects of onchocerciasis elimination in Africa, particularly by increasing coverage and adopting a twice per year treatment strategy [50][51][52][53][54]. However, the OEPA experience may not be fully reproducible in African foci. In the OEPA region treatment has been delivered by mobile teams, who observe directly that ivermectin tablets are ingested by the individuals receiving treatment, circumventing the barriers to elimination posed by the occurrence of systematic non-compliers that affect African countries, particularly those with onchocerciasis-loiasis endemic areas [55]. Currently, no new cases of onchocerciasis-associated blindness have been reported in most of the OEPA region and ocular morbidity has been eliminated from eleven of the 13 previously endemic Latin American foci. Parasite transmission has been interrupted in these eleven foci (~96 % of the total population at risk, representing four of the six countries where the disease was formerly endemic), and elimination has been reached in 10 foci (represent-ing~78 % of the population at risk). In 2013, onchocerciasis transmission was declared eliminated in Colombia [56], and by 2014 Ecuador became the second Latin American country to attain this goal [4,33]. In Mexico this target has been reached by 2015 [49,57] and Guatemala [58] has filed its verification dossier in the WHO. The two countries lagging behind, Venezuela and Brazil, share the most difficult to approach and hard-toreach populations of the Amazonian focus, yet our results show that given sufficient commitment and determination by the control programme, and unwavering support by OEPA, it is possible to attain and sustain high levels of treatment coverage and increased frequency, attesting to the feasibility of suppressing and ultimately interrupting transmission in the last bastions of onchocerciasis in Latin America.

Additional file
Additional file 1: Detailed description of study area, criteria for elimination, calculation of infection indices and additional entomological data. Text S1. Geographical and environmental characteristics of the Venezuelan part of the Amazonian onchocerciasis focus. Text S2. Endemic communities, mapping and geographical information system. Text S3. World Health Organization (WHO) criteria for onchocerciasis elimination. Text S4. Calculation of parasitological indices. Text S5. Calculation of transmission indices. Table S1. Number of ivermectin rounds for twice yearly (6-monthly) and quarterly (3-monthly) treatments that reached a therapeutic coverage ≥85 %, by geographical sub-area in the Venezuelan part of the Amazonian focus. Figure S1. Monthly biting rates of Simulium guianense s.l. recorded in Orinoquito and Parima B during baseline (1982-2000) entomological evaluations. Figure S2. Infectivity rates and mean nos. of L3 larvae per fly for S. guianense s.l. in Orinoquito and Parima B during baseline (1982-2000) entomological evaluations. Figure S3. Monthly infective biting rates and monthly transmission potentials for S. guianense s.l. in Orinoquito and Parima B during baseline (1982-2000) entomological evaluations. Figure S4. Relationship between the proportion of infective flies and the mean number of infective larvae per fly in S. guianense s.l. at baseline. Table S2. Comparison of manual