Systematic review of indoor residual spray efficacy and effectiveness against  Plasmodium falciparum  in Africa

There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

Abstract

Indoor residual spraying (IRS) is an important part of malaria control. There is a growing list of insecticide classes; pyrethroids remain the principal insecticide used in bednets but recently, novel non-pyrethroid IRS products, with contrasting impacts, have been introduced. There is an urgent need to better assess product efficacy to help decision makers choose effective and relevant tools for mosquito control. Here we use experimental hut trial data to characterise the entomological efficacy of widely-used, novel IRS insecticides. We quantify their impact against pyrethroid-resistant mosquitoes and use a Plasmodium falciparum transmission model to predict the public health impact of different IRS insecticides. We report that long-lasting IRS formulations substantially reduce malaria, though their benefit over cheaper, shorter-lived formulations depends on local factors including bednet use, seasonality, endemicity and pyrethroid resistance status of local mosquito populations. We provide a framework to help decision makers evaluate IRS product effectiveness.

Abstract

Indoor residual spraying is a commonly used method for mosquito, and malaria, control and there are a number of available insecticides that are available for this. Here, the authors evaluate the efficacy of widely-used and novel insecticides against pyrethroid-resistant mosquitoes.

Related collections

Most cited references 52

Record: found
Abstract: found
Article: found

Is Open Access

Reducing Plasmodium falciparum Malaria Transmission in Africa: A Model-Based Evaluation of Intervention Strategies

Jamie Griffin, T. Hollingsworth, Lucy C Okell … (2010)

Introduction Over the past five years, dramatic declines in malaria disease caused by Plasmodium falciparum have been reported across a range of settings within sub-Saharan Africa. These declines are associated with increased distribution of long-lasting insecticide-treated nets (LLINs) and with the switch from a failing drug regimen to artemisinin-based combination therapies (ACT) as first-line therapy [1]–[4]. Whilst this pattern of reducing disease is encouraging, there remain many countries within Africa that continue to have a high burden of disease and hence malaria remains a leading cause of mortality in children under five years of age [5]. Thus control of the disease, and ultimately elimination of the parasite in this continent, remain major public health goals. Eradication of malaria was attempted in the 1950s under the auspices of the World Health Organization-led Global Malaria Eradication Program (GMEP) [6]. Notably, Africa was not formally included in this program despite clear evidence of the large disease burden within the continent at that time. However, elimination campaigns were subsequently undertaken on a smaller scale within Africa, most prominently in two areas of moderate to high transmission in Nigeria (the Garki project [7],[8]) and on the Kenyan/Tanzanian border (the Pare-Taveta project [9]), but also periodically in areas of lower transmission including the Kenyan highlands [10] and the island of Madagascar [11]. These campaigns included frequent insecticide spraying of houses to reduce the vector populations and rounds of mass treatment to reduce the human infectious reservoir. Whilst substantial declines in infection and disease were observed in all of these campaigns, the control measures were not sufficient to eliminate the parasite on a short time scale, and failure to sustain control programs inevitably led to rebound of infection and disease in later years. This under-performance was perceived as a lack of success by past eradication attempts, which may in part be attributed to over-optimism about what could have been achieved with the tools then available [12]. Two years ago, following a renewed commitment to malaria control from donor organizations, the focus shifted again to malaria eradication as an ultimate goal. Previously, many countries had already intensified their own malaria control programs with much success in reducing both the burden of disease and ongoing transmission [1]–[4],[13],[14]. However, Africa poses the biggest challenge to a global eradication initiative, given the heterogeneous yet ubiquitous nature of P. falciparum transmission across much of the continent. Levels of transmission in Africa range from absent or low in many urban areas, through epidemic outbreaks in the highlands, to highly seasonal or perennial transmission in rural areas [15],[16]. This variable transmission pattern is further complicated by local variation in the major Anopheles vector populations that sustain transmission (principally An. gambiae s.l. and An. funestus, although approximately 70 relevant species have been identified worldwide [17]). Of the 47 countries within sub-Saharan Africa, the majority are currently classified by WHO/Roll-Back Malaria as being in the control stage and thus need to scale up interventions to sustain control and reduce the burden of disease via a reduction in transmission [18]. On the northern borders of the continent, transmission is already low, with Egypt and Algeria in the elimination phase and Morocco and Mauritius having interrupted local transmission. Similarly, in the southernmost countries, a sustained move towards local control and potentially elimination in border areas has been agreed upon via cooperation with neighbouring countries (the “elimination eight”) [19]. On the island of Zanzibar, a highly successful control program has reduced transmission to very low levels. However, a recent assessment of the feasibility of moving to elimination concluded that, whilst it is technically feasible to reduce local transmission to zero in this setting, the resources, both financial and operational, required to sustain elimination in the face of repeated reintroduction from mainland Africa make this a difficult prospect [20]. Compared to the past campaigns in the 1950s, additional tools are now available which, combined with sustained policy commitment, may make local elimination achievable in some settings and can aid control of disease by dramatically reducing malaria prevalence in countries with high rates of ongoing transmission. These include new LLINs, which have increased killing effects on the vectors compared to traditional nets and are more durable [21]–[25], and ACTs, which, through their gametocytocidal effect, can impact transmission from humans to vectors [26],[27]. In addition, a pre-erythrocytic malaria vaccine, RTS,S, has shown promising results in Phase II trials [28]–[34] and could soon contribute to elimination programs. National control agencies have varying levels of resources but can rarely implement all major control interventions at a given time. Understanding how to choose policy that is appropriate to the local setting is therefore key to effective control. Whilst the efficacies of most interventions have been individually evaluated in the field, the impact of different combinations of these is not clear. Field trials will be important to inform control policies but will be able to test only a few of the combinations of interventions in a limited number of settings. Mathematical models provide a tool with which to explore the expected impact of different interventions against malaria, both individually and in combination, on a range of program endpoints [26],[35]–[40]. Whilst simple models can provide important general insights, the heterogeneity in transmission intensity [41]–[43], the variability in vector species composition and associated bionomics [17],[35], and the seasonality in vector populations [44] are all important factors that affect the transmission potential of a site and the likely impact of intervention packages. We therefore constructed an individual-based simulation model which captures these key factors while remaining sufficiently mathematically tractable to enable the baseline model parameters to be rigorously fitted to data within a Bayesian framework. The model includes the suite of current tools most often employed by (or likely to be employed by) National Malaria Control Programs—namely, LLINs, IRS, ACTs in case treatment and in mass treatment campaigns, and a vaccine with characteristics similar to the RTS,S/AS01 vaccine now in Phase III trials. The principal aim of the modelling presented here is to explore the potential for current control measures to reduce parasite prevalence to a low level (defined here as below a threshold of 1% prevalence across all age groups detected through microscopy which represents a level below which surveillance would likely switch to case detection) as laid out in the control phase of the global elimination framework [45]. We illustrate our results by applying our model to six well-characterized transmission sites which represent the full range of transmission intensity–vector species combinations and seasonality patterns most commonly observed across Africa. Methods Simulation Model for Malaria Transmission Dynamics We developed a stochastic simulation model for P. falciparum transmission dynamics in which people are represented as individuals while vectors are represented as aggregated populations, stratified by species. The model builds on an earlier compartmental model which incorporates the acquisition and loss of immunity to disease and to detectable parasitaemia [46],[47], but is extended to incorporate infection-blocking immunity and heterogeneity in biting rates. Full technical details are given in Protocol S1 and the flow diagram is presented in Figure 1A. Briefly, individuals begin life susceptible (S) to infection but with partial maternal immunity determined by the level of immunity in women of childbearing age. Maternal immunity decays in the first six months of life, thereby increasing susceptibility to disease. Individuals become infected at a rate determined by the force of infection in the population (Λ), which is determined by the ratio of vectors to humans, the biting rate per mosquito on humans, the proportion of infectious mosquitoes in the vector population, and the person's level of anti-infection immunity. On infection, they pass through the liver (pre-patent) stage and then either develop clinical disease (with a probability φ determined by their current level of anti-disease immunity) or develop patent (detectable under microscopy) asymptomatic infection (1−φ). Those who develop clinical disease have a fixed probability (fT ) of being treated successfully (T), in which case they will clear infection and, depending on the drug, enter (with rate rT ) a period of prophylactic protection (P) before returning (rP ) to being susceptible to new infection. Those who fail treatment (1−fT ) are assumed to eventually clear disease (D) and become patently asymptomatic (A) with rate rD . From patent asymptomatic infection, individuals will eventually move to a sub-patent stage (U) which can be an important component of the infectious reservoir [48], at a rate (rA ) that depends on their current level of anti-parasite immunity. Sub-patent infection is eventually cleared (rU ) and individuals return to being fully susceptible. From all infected states, acquiring a new infection in the presence of an existing infection (superinfection) is possible. Rather than explicitly tracking mixed infections, we assume that the new infection dominates and thus individuals move to either the clinical disease or asymptomatic states dependent on their level of anti-disease immunity. Individuals become infectious to vectors, at differing rates, in the clinical disease, patent and sub-patent asymptomatic stages—the states that compose the human infectious reservoir (Figure 1D). Four types of human immunity are included and are modelled dynamically. Maternal immunity, which protects against clinical disease, is assumed to decay exponentially from birth. Anti-disease immunity, which reduces the probability of developing clinical symptoms on infection, and infection-blocking immunity, are both exposure-driven whilst anti-parasite immunity, in which individuals control parasite densities and thus leave the patent infection state more quickly, is assumed to develop with age, conditional on having been exposed. 10.1371/journal.pmed.1000324.g001 Figure 1 Transmission model; EIR, prevalence and seasonality; and infectious reservoir. (A) Flow diagram for the human component of the model. S, susceptible; T, treated clinical disease; D, untreated clinical disease; P, prophylaxis; A, asymptomatic patent infection; U, asymptomatic sub-patent infection. (B) The relationship between EIR and parasite prevalence in children under 15 y. Solid line: fitted relationship; filled circles: data representative of this age group; open circles: data from other age groups (mostly younger) used in the model fitting. (C) The relationship between transmission intensity characterized by EIR and seasonality, defined as the proportion of EIR over a single calendar year that occurs within the peak three months of transmission. The colours of the markers indicate the different transmission settings and the shapes the species. (D) The estimated age-specific infectious reservoir for the different transmission settings defined in (C), with the same colours as (C). This is defined as the product of the age-specific biting rate, age-specific prevalence states (T, D, A, and U), state-specific onward infectivity to mosquitoes and the size of the population at this age. Three Anopheles vector species (An. gambiae s.s., An. funestus, and An. arabiensis) are modelled explicitly as the predominant vectors in the transmission sites that we consider. Vectors begin susceptible and on taking an infectious bite move into a latent state. From this they become infectious to humans, with infectivity determined by their human blood index (HBI) and biting rate and are assumed never to recover before death. Vector density is assumed to follow a seasonal pattern as determined by fitting an appropriate functional form to entomological data from the areas considered (see Table 1 and Protocol S4). 10.1371/journal.pmed.1000324.t001 Table 1 Summary of the six malaria transmission settings considered here. Country Location Population Type of Transmission Reported Annual EIR (ibppy) Fitted Annual EIR (ibppy) Anopheles Species Relative Abundance Reference Cameroon Nkoteng Rural Moderate, perennial 94 81 72% An. funestus; 28% An. gambiae s.s. [97] Democratic Republic of Congo Kinkole Rural Moderate, perennial 48 43 Nearly 100% An. gambiae s.s. [98] Ghana Kassena-Nankana District Rural High, seasonal 630 586 60% An. gambiae s.s.; 40% An. funestus [99] Mozambique Matola, Maputo Coastal suburb of capital Moderate, perennial 28 46 42% An. arabiensis; 46% An. funestus (additional 12% An. coustani are not considered here) [100] Tanzania Matimbwa Rural High, seasonal 703 675 85% An. gambiae s.s.; 10% An. funestus; 5% An. arabiensis [101] Uganda Kjenjojo Kasiina Rural Low 7 3 65% An. gambiae s.s.; 35% An. funestus [102] Model Parameterization Model parameterization was undertaken in several stages. First, a literature search was undertaken to formulate prior distributions for all model parameters. Where there was no information in the literature, vague priors were used or parameters were fixed if they could not be identified from subsequent model fitting. The human model parameters were estimated by fitting the equilibrium model conditional on EIR using Bayesian Markov Chain Monte Carlo (MCMC) methods to data on the stationary distributions of parasite prevalence (by both microscopy and PCR) by age from 34 locations across a wide range of transmission intensities from Africa (see Protocol S3) and of clinical disease incidence from two settings in Senegal [49]. Site-specific prior distributions for EIR were used based on published data ([50] and Protocol S3). By fitting the model to these data we were able to characterize the relationship between EIR (ibppy, the number of infectious bites per person per year) and parasite prevalence (Figure 1B). The parameters determining the onward transmissibility of the human infectious stages (clinical disease, patent and sub-patent infection) to mosquitoes were obtained by model fitting to data from human feeding studies and the Garki project [7],[51]–[54]. These parameters combined with parasite prevalence determine the age profile of the infectious reservoir (Figure 1D) [55],[56]. Only age-targeted strategies are sensitive to this profile. Parameters for the vector model were taken from the literature. A full listing of model parameters, their prior and posterior medians, and literature sources are given in Table S3.1 in Protocol S3. To run the model in specific settings, data on vector species composition, their seasonal profile, and the intensity of transmission (EIR) were extracted from the literature (Table 1, Figure 1C, Figure 2, and Protocol S4). A functional form was fitted to monthly data on either EIR or vector density to enable a single seasonal driver input (emergence of vectors) into the model. Full details of the settings and the seasonal profile fitting are in Protocol S4. 10.1371/journal.pmed.1000324.g002 Figure 2 Fitted seasonal profile of EIR for the six transmission settings by vector species. The fitted seasonal profiles of EIR per day and fitted annual EIR were obtained by fitting a transformed sinusoidal function to reported time series of either EIR or mosquito densities in the settings (see Protocol S4). Grey, total; red, An. gambiae s.s.; blue, An. funestus; green, An. arabiensis. (A) Nkoteng, Cameroon; (B) Kinkole, DRC; (C) Kassena-Nankana District, Ghana; (D) Matola, Maputo, Mozambique; (E) Matimbwa, Tanzania; (F) Kjenjojo Kasiina, Uganda. Interventions The implementation of each intervention is described briefly below. Full mathematical details and tables of parameter values are provided in Protocols S2 and S3. Long-lasting insecticide-treated nets We adapted an existing model [36] to our individual-based framework. Nets are assumed to have four effects: direct killing of a mosquito that lands on them, repellency which results in a longer gonotrophic cycle and possible diversion to a non-human blood host, a direct protective effect for the individual sleeping under the net, and a reduction in transmission from infected individuals sleeping under the net to susceptible mosquitoes. The degree of indoor-biting (endophagic) behaviour for the different species is incorporated into the model when assessing the LLIN effect. These behaviours are assumed to remain constant throughout the intervention. Indoor residual spraying IRS was added to the LLIN model as an additional intervention which can kill mosquitoes as they rest within the house or repel them before they feed. In the model the repellency effect extends the duration of the gonotrophic cycle in the same way as the repellency effect of LLINs. For IRS the killing effect depends primarily on the indoor-resting (endophilic) nature of the species as well as its HBI. Simulations assumed a DDT-like insecticide with a half-life of 6 mo which acts by repelling and killing mosquitoes [57]. Switch to ACT as first-line treatment Effective treatment (i.e., treatment which fully clears infection) was assumed to be given to a proportion of those developing clinical disease. Treatment failures were not explicitly modelled but are assumed to follow the same infection path as untreated infections. The half-life of the drugs pre-ACT (where we assume sulphadoxine-pyrimethamine [SP] was first-line therapy) and following ACT introduction determine the period of prophylaxis. In addition, the gametocytocidal effect of ACTs was incorporated as a reduction in onward infectiousness as in a previous model, based on data from human-to-mosquito transmission experiments involving treated patients [27]. Mass drug administration We considered the impact of a mass screening and treatment approach (MSAT) using a single dose of an ACT. We assumed that a rapid diagnostic test (RDT) would have approximately the same sensitivity as microscopy and thus all those in the clinical disease or asymptomatic patent infection stages would receive the drug, but that the uninfected and sub-patent infected individuals would not. The ACT was assumed to clear any infection present and provide a period of prophylactic protection (25 d, corresponding to an artemisinin coupled with a drug such as SP). The coverage level refers to the number of individuals screened. Pre-erythrocytic vaccine A pre-erythroyctic vaccine was assumed to reduce the probability of transmission from mosquitoes to people. It remains unclear whether this lower exposure to infection will affect the development of anti-disease immunity. Here we assume that it does but that it has no effect on the development of anti-infection immunity. Individual vaccine efficacy was assumed to decay exponentially with a half-life of 3 y. The vaccine is delivered through the Expanded Program for Immunization (EPI) and given at ages 3–5 mo, or as a mass vaccination program across all ages every 3 y. Transmission Settings We considered the impact of these interventions, individually and in combination, in six different settings that characterize the spectrum of transmission patterns of P. falciparum across Africa. These settings range in transmission intensity from measured EIRs of approximately 5 to over 500, translating in our model to parasite prevalence in 2- to 10-year-olds of 14% to 85%. In 2007, 80% of Africa's population was estimated to reside in an area with parasite prevalence in 2- to 10-year-olds of >5% and 50% in an area with prevalence >40% [16]. These specific settings, summarized in Table 1, Figure 1C, and Figure 2, were chosen because of the large number of both entomological and clinical studies undertaken in these areas and to represent patterns of perennial/seasonal transmission of varying intensity and with different mixes of Anopheles species. We fitted the model to data from these settings, which we take as our baseline scenarios. For each scenario, we present the mean of ten simulation runs in a population of 10,000 individuals, which was sufficient to approximate the dynamics in a larger population. The population size was assumed to be static over time, with age structure based on data from Tanzania. After introducing infection, the model was run for 50 y to reach equilibrium representing the situation in the year 2000. Between 2000 and 2010 we increased the distribution of LLINs from a baseline of zero coverage to a maximum of 20% coverage [58] and implemented a switch to ACT as first-line therapy in the year 2000. Combinations of interventions were then introduced from 2010 onwards. Note that this does not necessarily reflect the true intervention programs in place in these settings in these years, and hence model outputs do not directly predict expected patterns in these settings; rather they give an indication of the likely effectiveness of the modelled intervention packages in different setting types. Intervention Package Scenarios Coverage here is defined as the proportion of individuals receiving an intervention (for LLINs ownership, for vaccination those receiving the vaccine, for IRS those that reside in houses where spraying occurs, and for MSAT the number of individuals screened). We separately consider the impact of adherence/usage for LLINs, which is assumed to decay over time. This proportion of people using LLINs is termed effective coverage. For IRS we assume no loss of adherence. For MSAT we assume that all those who are screened and positive on microscopy take the drug. Similarly, for the vaccine we assume that those offered it accept. Finally, for all interventions there is a decay in protective efficacy over time for those who have received and use the intervention. For LLINs this is due to wear-and-tear and loss of insecticidal effect. For IRS we model the loss of insecticidal effect. For vaccines we assume that efficacy declines through waning protection. Unless stated otherwise we assumed that IRS and MSAT were given at 80% coverage (the maximum achievable in well-managed control programs [59]) and the vaccine at 90% coverage (based on EPI distribution statistics). For the roll-out of LLINs we considered two realistic scenarios. In the first, distribution was increased gradually to a maximum of 80% within 5 y and a new net was distributed to individuals every 5 y. In the second, we assumed almost immediate distribution at 80% coverage, redistribution every 5 y, plus delivery of a net to 80% of newborn infants and an average of 0.75 adults for every infant who receives a net (Figure 3A). These coverage levels are similar to the targets set for 2010 for scaling up for impact in the Global Malaria Action Plan [18]. In both scenarios, we assume that LLIN use wanes over time so that effective coverage is lower. Here we assumed an exponential decay at a rate 0.2 per year so that after 5 y effective coverage is approximately 37% of the baseline level. We also considered the impact of a theoretical (unachievable) maximum of 100% coverage with LLINs coupled with no decay in usage over time (Figure 3A). Protective efficacy of the nets due to decaying insecticide efficacy and wear-and-tear was assumed to decay exponentially with a half-life of 2.64 y (Protocol S3 and [60]). We did not consider any decay in effective coverage of IRS as we assumed that coverage remained constant at each round (i.e., people do not refuse to have their house sprayed as the intervention goes on). The protective efficacy of DDT was assumed to decay exponentially with a half-life of 6 mo (Protocol S3 and [57]). Adherence to LLINs given receipt was assumed to be independent of IRS acceptance. 10.1371/journal.pmed.1000324.g003 Figure 3 Impact on parasite prevalence of LLINs alone. (A) Example of the ways in which coverage of LLINs is considered to increase in various model scenarios. Baseline (blue): our baseline scenario in which 80% coverage is achieved over five years but adherence also decays between net distribution rounds; rapid (brown): as baseline but with more rapid scale-up to 80% coverage; rapid, no drop-out (green): rapid scale-up to 80% coverage with no decay in adherence; 100% (yellow): 100% coverage with no decay in adherence (theoretical maximum effect). (B) Model-predicted impact on parasite prevalence over calendar time of four scenarios for LLIN scale-up combined with an earlier switch to ACT as first-line therapy in Kjenjojo Kasiina, Uganda. (C) Final parasite prevalence and (D) absolute reduction in parasite prevalence after 15 years of a sustained intervention program in the six transmission settings with the baseline scenario for LLIN distribution, the rapid scenario for LLIN distribution, and the rapid scenario with no loss of adherence for LLIN distribution. We undertook preliminary runs for IRS and MSAT to identify the optimal time of year for annual programs. The optimal time was defined on the basis of providing the maximum reduction in mean prevalence of parasitaemia across all age groups in year 10 of the intervention campaign. We found that in those settings which have a clear seasonal peak in the EIR, it is always optimal to spray just before the upward trend in EIR. In settings with less seasonality, there is less difference in impact, but spraying at the start of the main transmission season tends to remain optimal. In contrast, across most settings, the optimal time of year to mass treat in terms of reducing overall prevalence of asexual parasitaemia as an endpoint is at the beginning of the period of lowest EIR (also shown in [61]), which generally occurs approximately 2 mo after peak slide prevalence. For scenarios in which IRS and MSAT were undertaken every 6 mo, they were implemented at the optimal time of year as defined above, plus 6 mo thereafter. Effective coverage and protective efficacy do not alone determine intervention effectiveness, as they also depend on whether the same individuals receive multiple interventions or whether interventions are randomly distributed across the population. We therefore allowed correlations between repeat distribution for each individual intervention (where a correlation of 0 means that redistribution is completely random and of 1 that redistribution always occurs to those who had previously received the intervention). We also allowed correlations between receiving LLINs, IRS, and MSAT. Here a positive correlation means that individuals who receive one intervention are also more likely to receive the other (which could reflect access to interventions) whilst a negative correlation means that those who receive one intervention tend not to receive the other (which would reflect a propensity not to use multiple interventions). As our focus is to consider intervention packages aimed at reducing transmission, our primary outcome was the annual mean prevalence of asexual parasitaemia as measured by microscopy in the whole population up to 25 y following the start of the intervention program. We chose this rather than prevalence restricted to children as it enables us to correctly compare age-targeted interventions. We specifically do not focus on short-term “predictions” or timelines, as our sensitivity analysis shows that these are highly dependent on parameters relating to the loss of acquired immunity (which impact the fitted duration of infection). Currently these parameters are not well-estimated from the available data (see Section 5.2.1 in Protocol S5, and Box 1). Furthermore, time scales of impact will inevitably depend on the speed with which scale-up of interventions occurs and so cannot be reliably predicted without detailed assessment of local situations. Box 1. Uncertain Parameters Whilst models can be useful tools in setting realistic expectations for intervention programs, some key parameters in our current model are based on limited data. Further empirical work in these areas could improve future models. These include: The duration of natural infection and the extent to which super-infection prolongs this duration or increases infectivity; The rate of acquisition of immunity at different transmission intensities, and the rate of loss of immunity when transmission is reduced; The bionomics of the principal vector species and the impact of vector-targeted interventions on them; Detailed data on the speed with which coverage of interventions is scaled up, heterogeneity in coverage levels achieved, and the degree of adherence to the interventions over time. Software A user-friendly software package for reproducing the simulations presented here, as well as other potential combinations of the interventions included in this paper, is freely available to download from our Web site (http://www1.imperial.ac.uk/medicine/about/divisions/publichealth/ide/research_groups/malaria/). A short summary of the interface is provided in Protocol S6. Results Continued Scale-Up of LLINs Continued scale-up of LLINs from the baseline assumption of 20% coverage could potentially reduce transmission across all six transmission settings, given that the dominant vector species in these settings are primarily endophagic and their peak biting times coincide closely with normal sleeping hours [62] (provided changes in mosquito behaviour in response to the interventions are not dramatic). However, the magnitude of the effect will depend not only on the intensity of transmission in each setting but also how roll-out is achieved, the final level of coverage, adherence to LLIN use, and the decay in insecticide effectiveness over time. Figure 3A shows four potential scenarios for scale-up if nets are redistributed every 5 y. Theoretically, the greatest impact is achieved with rapid deployment, 100% coverage, and perfect adherence. However, even at this unrealistically high level, the efficacy will be less than its maximum due to decaying effectiveness of the insecticide. Even at the target coverage levels of 80%, with gradual roll-out and realistic adherence, effective coverage levels can, on average, be as low as 50% (Figure 3A). The additional decay in insecticide efficacy over time can result in protective coverage levels as low as 30%. This is even without the additional limitation of an interrupted supply chain, which is likely to reduce effective coverage further [63]. In the low-transmission setting of Kjenjojo Kasiina, Uganda, the basic reproduction number (R 0) is already close to 1 in the absence of additional interventions. Thus, parasite prevalence can be reduced to below the 1% threshold over a 15 y time horizon with LLIN use alone (Figure 3B). However, even in this relatively low-transmission setting, high levels of coverage and adherence are required. Furthermore, with decaying adherence in their use it is likely that transmission will be sustained, albeit at a low level. Furthermore, if LLINs have a lower killing effect than that assumed here, our model would predict sustained transmission in this setting (Figure S5.6 in Protocol S5). In contrast, in the moderate-transmission settings of Nkoteng (Cameroon), Kinkole (Democratic Republic of Congo or DRC), and Maputo (Mozambique), and in the high-transmission settings of Kassena Nankana District (KND) (Ghana), and Matimbwa (Tanzania), scale-up of LLINs alone does not reduce parasite prevalence to below 1%, even over longer time periods (Figure 3C). We can, however, expect to see dramatic declines in the first five years of the program followed by an increase to new endemic levels as levels of immunity in the population change (Figure S5.1 in Protocol S5). The time scale of this rebound is difficult to ascertain from current data due to uncertainty in the rate of loss of acquired immunity (see section 5.2.1 in Protocol S5, and Box 1). In high-transmission settings, with continued scale-up of LLINs to 80% coverage within five years, the absolute drop in prevalence is between 5% and 10%. If rapid scale-up occurs and adherence is sustained, drops in prevalence of 20%–25% can be expected (Figure 3D). However, despite the smallest relative impact occurring in the high-transmission settings, because most cases of infection and disease occur in these settings, the absolute impact in terms of numbers of infections averted will be greater. Thus, in terms of reduction in infections per net distributed, impact will be greatest in these high-transmission settings. Additional Use of IRS and MSAT Whilst continued scale-up of LLINs is predicted to reduce transmission substantially, under realistic assumptions about the level of coverage and adherence to LLIN use, additional tools will be necessary in many settings. In Kjenjojo Kasiina, Uganda, yearly rounds of IRS with DDT combined with continued scale-up of LLINs to 80% coverage is predicted to locally eliminate transmission (Figure 4A). Yearly rounds of MSAT as an alternative to IRS tend to have less impact although this would also achieve a reduction below the 1% parasite prevalence threshold. 10.1371/journal.pmed.1000324.g004 Figure 4 Impact of combining LLINs with IRS and MSAT. (A–F) Impact of intervention scenarios incorporating IRS and MSAT on parasite prevalence in the six transmission settings. All scenarios include the earlier switch to ACT as first-line therapy. “LLIN only” uses the baseline scale-up for coverage. All other scenarios include LLIN scale-up using the baseline scenario except where noted. (G and H) Final parasite prevalence and absolute reduction in prevalence after 15 years of a sustained intervention program in the six transmission settings with baseline scenario for LLIN distribution; baseline LLIN + yearly MSAT; baseline LLIN + yearly IRS; baseline LLIN + yearly MSAT + yearly IRS. In the moderate-transmission setting of Kinkole, DRC, more intensive rounds are required. Thus, in this setting, twice yearly IRS and MSAT are required to reduce parasite prevalence below the 1% threshold (Figure 4B). In contrast, in the slightly higher-transmission setting of Nkoteng, Cameroon, this is not sufficient in itself and additional faster scale-up of LLINs is needed to achieve this threshold (Figure 4D). In Maputo, Mozambique, in which transmission intensity as measured by EIR is similar to Kinkole, DRC and lower than Nkoteng, Cameroon, even these more intense programs are unable to reduce prevalence below the 1% threshold (Figure 4C). This is due to the high proportion of transmission that occurs via An. arabiensis in this setting, whose more exophilic behaviour reduces the impact of IRS on transmission. Assuming a lower degree of exophilic behaviour of this species compared to our baseline assumption, this conclusion continues to hold (section 5.2.2 in Protocol S5). In all three moderate-transmission settings, IRS with an insecticide similar to lambdacyhalothrin (which is less repellent and hence more lethal but has a shorter half-life than DDT) is predicted to have a lesser effect on transmission than DDT (Figure S5.7 in Protocol S5). In both high-transmission settings (KND, Ghana and Matimbwa, Tanzania), current tools are insufficient to reduce parasite prevalence below the 1% threshold (Figure 4E and 4F; see also higher levels of adherence and coverage in Figure S5.3 in Protocol S5, and higher frequency of MSAT in Figure S5.8 in Protocol S5). However, in both settings, an intense program involving rapid scale-up of LLINs with sustained adherence and twice-yearly rounds of MSAT and IRS could result in marked declines in prevalence from 60% to 10% in the population as a whole (Figure 4E and 4F). However, in these settings, the interventions would need to be sustained indefinitely to maintain this new endemic level. Yearly IRS and MSAT combined with 80% coverage of LLINs is predicted to reduce parasite prevalence after 15 y to below 10% in moderate transmission settings and below 25% in high-transmission settings (Figure 4G). Again, the absolute reduction will be greatest in the latter, with a 40%–50% drop in parasite prevalence in these settings (Figure 4H). Targeting and Overlap in Intervention Coverage LLIN distribution programs initially focused on young children as one of the high-risk groups for developing severe disease. However, as shown in Figure 5A and elsewhere [23],[24], this strategy is unlikely to have an additional impact on transmission, because the youngest children tend not to be major contributors to the infectious reservoir (Figure 1D). However, if limited coverage is achievable, substantially greater reductions in prevalence could be obtained if, for a given level of distribution, nets were targeted towards those living in the local foci of transmission which impact strongly on sustaining transmission [38],[42],[43],[64]. Thus in Kinkole, DRC, in a program of LLIN distribution with a low 20% coverage, if distribution is prioritised to those at highest risk we could expect a reduction in prevalence after 15 y of approximately 6% compared to a reduction of 3% if the same number of nets were distributed randomly. A similar picture emerges for MSAT programs (Figure 5B), although the effect of targeting is greater for LLINs because in addition to their direct protective effect, they kill mosquitoes in proportion to the rate at which the protected person would have been bitten. 10.1371/journal.pmed.1000324.g005 Figure 5 The effect of non-random distribution of interventions. (A and B) Parasite prevalence after 15 years of an intervention program as a function of the target coverage of (A) LLIN distribution and (B) MSAT for Kinkole, DRC. Blue: if the intervention is distributed randomly; green: if the intervention is preferentially distributed to the youngest children; red: if the intervention is preferentially distributed to those who are bitten most frequently (excluding age dependency in biting rates). (C and D) Parasite prevalence after 15 years of a single intervention program as a function of the frequency of the intervention and whether successive rounds are given randomly (green) or to the same people (purple) for Kinkole, DRC. (C) IRS; (D) MSAT. (E and F) Parasite prevalence in all individuals (red), in 2- to 10-year-olds (blue) and EIR (green) after 15 years of a combined intervention program as a function of the correlation in receipt of the two interventions for KND, Ghana. A correlation of 0 represents random distribution at each round, 1 represents those receiving one intervention also receive the other and −1 represents those receiving one intervention do not receive the other. (E) IRS and LLIN; (F) IRS and MSAT. For (E) and (F) there is 50% coverage per round for IRS and MSAT and the baseline scenario for LLINs. With any intervention, it is likely that the same individuals or villages will tend to access the intervention at each distribution round. Thus for example, if 80% coverage of LLINs is achieved, but at each redistribution the same 80% receive the intervention, then after three rounds of redistribution the percentage of the population ever receiving an LLIN is 80%. However, if this 80% coverage reflects random distribution, then after three rounds the percentage of the population ever receiving an LLIN is 100×(1−0.2×0.2×0.2) = 99.2%. Figure 5C and 5D shows the predicted effect of rounds of IRS and MSAT between these two extreme (systematic versus random coverage) scenarios. In both cases, assuming random distribution results in an overestimate of the effect of the intervention, and this difference increases the more frequently IRS or MSAT is undertaken. Thus, to optimize program effectiveness it is necessary to ensure that as wide a proportion of the target population is reached by the intervention. In addition to correlations between those who receive an individual intervention, there is likely to be overlap in those who are offered different interventions. This is likely to be most strongly correlated for IRS and LLINs, given the perception of these interventions as providing direct protection to the individual or household. A positive correlation will occur if the same individuals access the interventions. Under these scenarios, we can expect the least impact of the intervention program (Figure 5E). However, if uptake is negatively correlated, for example if those who are offered IRS and LLINs choose only to have one, for the same overall coverage levels of the individual interventions total population coverage is increased over and above naïve expectations assuming both are randomly distributed. This increased total coverage results in the largest reductions in transmission (Figure 5E). Similar effects are observed for IRS and MSAT, although again, this is not as pronounced as for LLINs given that there is less redundancy between IRS and MSAT than between two antivectorial measures (Figure 5F). Additional Impact of RTS,S/AS01 Vaccine In the low-transmission setting of Kjenjojo Kasiina, Uganda, RTS,S (when it becomes available) could further reduce transmission and thus negate the need for additional rounds of IRS to speed declines. As found by others [65],[66], vaccination at birth under the EPI is expected to have relatively little impact either with or without additional rounds of MSAT (Figure 6A). If mass vaccination every 3 y is undertaken as an alternative alongside the baseline scale-up of LLINs to 80% coverage, prevalence is predicted to fall to under 1%. 10.1371/journal.pmed.1000324.g006 Figure 6 Impact of additional vaccination on parasite prevalence in the different transmission settings. All runs assume the RTS,S vaccine is 50% efficacious and has a half-life of 3 years. PEV at EPI denotes the pre-erythrocytic vaccine being given through the Expanded Program on Immunization, whilst mass PEV denotes a mass vaccination campaign. All runs include LLINs. (A) PEV at EPI with or without additional MSAT in Kjenjojo, Uganda (B) Mass PEV with or without additional MSAT in Kjenjojo, Uganda (C to F) MSAT and IRS with mass PEV in: (C) Kinkole, DRC, (D) Maputo, Mozambique, (E) Nkoteng, Cameroon and (F) KND, Ghana. In the moderate transmission settings of Kinkole, DRC (Figure 6C), Maputo, Mozambique (Figure 6D), and Nkoteng, Cameroon (Figure 6E), continuation of programs incorporating IRS and MSAT in addition to LLIN distribution will be needed even if a vaccine is available. However, with a mass vaccination program prevalence in all three sites can be driven below 5%. In Maputo especially, where IRS is predicted to be less effective, an additional vaccination program has a noticeable further impact on prevalence. In both high transmission settings (KND, Ghana, Figure 6F; and Matimbwa, Tanzania, results not shown), mass vaccination results in modest reductions in prevalence. Across all transmission settings, a more efficacious vaccine with a longer duration of protection would further reduce transmission (section 5.2.6 in Protocol S5). Discussion If deployed in combination, current interventions can result in substantial declines in malaria prevalence across a wide range of transmission settings. Our results show that in areas with relatively low transmission (EIR<10 ibppy), increased distribution and use of LLINs, coupled with the switch to an effective ACT as first-line therapy, could reduce transmission to very low levels if high levels of coverage and adherence are achieved. Defining low-transmission areas as those where parasite prevalence in 2- to 10-year-olds is under 25%, approximately 20%–50% of individuals living in areas of stable risk of P. falciparum transmission in Africa live in such settings [16]. Additional use of IRS and/or MSAT in these settings would speed this reduction and also allow overall parasite prevalence to be reduced to <1% even if adherence to LLIN use is not perfect. These results agree with recent observations made in a very low transmission setting in Western Kenya, in which the parasite appears to have been eliminated in an area in which ACT and LLIN usage have been coupled with IRS rounds [67]. Large reductions have also been achieved in Zanzibar, where the preintervention parasite prevalence was 9% in children aged 0 to 5 y and 12.9% in children aged 6 to 14 y [13]. After a switch to ACT as first-line therapy and high coverage of both LLINs and IRS rounds from 2003, parasite prevalence in all age groups is now well below the 1% threshold. The challenge in such settings is to sustain interventions at a sufficient level to maintain effective control in the face of reintroduction from neighbouring areas via human migration and travel. In some moderate-transmission settings it is also possible to reduce parasite prevalence below the 1% threshold with existing tools. In our example settings, this could be achieved in Kinkole, DRC where the endemic EIR was 48 ibppy if an intensive program of twice-yearly IRS and MSAT were combined with increasing LLIN coverage to 80% levels. In the slightly higher transmission setting of Nkoteng, Cameroon (EIR = 96 ibppy), current tools could reduce transmission below the 1% threshold but in this case (perhaps unrealistically) high levels of adherence to LLIN use would also be needed. Thus the first phase of elimination programs is achievable in many areas in which the LLIN and IRS in combination are effective (that is, in areas with primarily endophilic vectors). Additional use of MSAT, to date not considered by many programs, has the potential to speed further declines in prevalence. We considered one area, Maputo, Mozambique, in which the high proportion of An. arabiensis (exhibiting a high degree of exophilic behaviour), made elimination more difficult. Whilst the scale of the declines that our model predicts are similar to those observed in an IRS-based campaign in that area (that commenced in 2000 using Bendiocarb rather than DDT [4]), this study also demonstrated a greater impact on the population of An. arabiensis compared to that resulting from our model. This may be because our estimate of the degree of exophilic behaviour is too high (see Protocol S3) or because mosquito behaviour changes both with season and with setting [35],[68]–[71] and requires further exploration. In high transmission intensity settings, current tools can be used to substantially reduce transmission and the associated disease burden, but are insufficient to drive prevalence below the pre-elimination threshold. This finding is not surprising given the high basic reproduction numbers previously estimated in large parts of sub-Saharan Africa [43]. Such outcomes have been observed in the Bioko Island control program where, with intensive ongoing interventions, parasite prevalence in 2- to 5-year-olds fell from 42% to 18% between 2004 and 2008 [14]. Similarly, in the 1970s Garki project in Nigeria, an area of moderate to high transmission (annual EIRs in the range 20–130 ibppy), substantial declines in prevalence were recorded but elimination was not achieved [7]. In these settings, additional new tools are likely to be required if pre-elimination targets are to be achieved. Whilst a detailed comparison of the range of potential tools under development is beyond the scope of this paper, there are two broad areas of innovation that merit further consideration. The first aims to target the mosquitoes that are not reached by current interventions, particularly those on whom indoor-targeted interventions are least successful. Notably, this includes major species such as An. arabiensis, which preferentially rest outdoors after feeding and may also obtain blood meals from non-human animals. These mosquitoes could be targeted in a number of ways, including additional interventions that are applied on non-human hosts [72], killing adult females feeding or resting outdoors [73]–[75], or at source in the larval habitat [76],[77]. Secondly, our results on the levels of human adherence required in high-transmission settings suggest that interventions that do not strongly depend on human participation are likely to be needed. The methods outlined above are examples of such approaches. Our results confirm findings by others that the bionomics of the local vector species, including the degree of exophagy, exophily, and zoophagy [35],[36],[78], can potentially be a strong determinant of intervention success. Current tools, in particular LLINs and IRS, are focused towards species with strong endophagic, endophilic, and anthropophagic tendencies. Further data on the degree of endophilic behaviour of the different Anopheles species, coupled with information on how these parameters may change in response to interventions (we assumed here that they remain fixed),are critically needed to understand the longer-term impact of IRS and LLINs on transmission. Historically, there is some evidence of species replacement following the introduction of IRS in three different geographical locations [79]. More recently, a shift in species relative abundance (though not replacement or increased density) has been observed in Western Kenya following high coverage of LLINs [80],[81]. In addition, mapping of vector species distribution and proportional composition [17] is critical to the ability to predict program success outside of the well-studied research areas. Behavioural aspects of intervention programs are characterized in multiple ways. For example, the WHO report bed-net coverage as the number of nets distributed per person at risk [5], whilst Malaria Indictor Surveys collect data on the proportion of households owning a net or sleeping under a net [82]. Our results demonstrate that patterns of coverage and effective coverage are an important determinant of intervention success and may be one reason why simple models of LLIN impact have tended to appear highly optimistic [35]–[38]. Furthermore, it is unrealistic to assume perfect and uniform adherence. Indeed, rates of sleeping under LLINs tends to be highest in young children, but lower in school-aged children [83], who are important contributors to the infectious reservoir (Figure 1). Furthermore, whilst we did not explicitly consider reduction in adherence or take-up of IRS, this is likely to occur after repeated rounds as perceived risk declines, and will reduce the impact of the intervention. Receipt of interventions is also an important consideration in assessing impact, particularly if coverage levels are low. It is well recognized that malaria transmission is highly focal with some individuals at much higher risk than others [42],[43],[64],[84],[85]. Our results confirm other models' findings [43],[64] that, by targeting interventions at areas of intense transmission, substantially greater reductions in transmission are possible than by distributing them randomly or by focusing distribution towards younger children. However, little attention has previously been paid to the heterogeneous distribution of interventions within such target populations. In general, the impact of an intervention will be lower if the same individuals in the target population continually receive and adhere to the intervention than if distribution fully covers the target population. Thus data on repeat uptake of interventions would be useful to determine true target population coverage levels. Furthermore, health systems will need to be strengthened and laboratory capacity put in place to allow rapid identification of these foci. In addition, overall coverage levels can potentially be enhanced through consideration of a wide range of different delivery mechanisms appropriate to the local setting [86]–[88]. One aspect with the potential to hinder elimination campaigns not considered here is the development of resistance—either to drugs, to the insecticides used to treat nets or for indoor residual spraying, or to vaccines—and the potential for alterations in the behaviour of the vector in response to the interventions. Resistance to DDT was a particular problem during the GMEP and is credited with being a major reason for the abandonment of the program. DDT resistance at varying levels has now been reported in over 50 anopheline species [89]; thus, to reduce the further emergence of resistance, elimination campaigns should aim to reduce transmission as rapidly as possible. The recent emergence of partial drug resistance to artemisinin in Cambodia [90] has further highlighted the need to guard against and reduce the emergence and spread of resistance, particularly as access to treatment is scaled up. Our model is necessarily a simplification of the more complex dynamics underlying malaria transmission and control, so numerical results should be interpreted more as providing intuitive insight into potential scenarios than as firm predictions of what might happen in a given setting. Furthermore, whilst we give an indication of impact over a 25-year time horizon (including graphs that track expected trends over this period), given the uncertainty in some of the key parameters, it is not possible to give short-term indications of impact or timelines. Precise, accurate prediction remains challenging for a number of reasons. First, the mean duration of asymptomatic infection, and the dynamics of acquisition and loss of immunity, are key parameters determining the speed of decline in parasite prevalence once transmission is reduced [47]. These are both poorly understood in semi-immune populations. These parameters also determine the time scale for which interventions would need to remain in place to ensure that a rebound in infection and disease does not occur. Current best estimates of model parameters suggest that this is likely to be decades rather than months or years, but further data are needed to refine these estimates. Second, there are multiple model structures that can reproduce important characteristics of malaria epidemiology such as the age patterns of infection prevalence across different transmission settings. Whilst we have invested substantial effort in developing a modern statistical framework to better choose between model structures and to estimate associated model parameters, there are limited data to distinguish some aspects of the model. In the current exercise, we have focused on fitting the human model cycle to a wide range of datasets. This will be extended in future applications to fit the full cycle using explicitly seasonal models to more detailed data from specific research sites. In addition, the individual intervention models have not to date been validated by fitting to specific trial data. This process is underway. Such fitting will enable the addition of uncertainty bounds to model output through sampling of parameter posterior densities [91]. If feasible, this could be extended to incorporate model uncertainty using a Bayesian methodology [92]. Third, in our current model we use a relatively simple vector cycle in which the vector population is driven by a constant birth rate. This may underestimate the additional impact of interventions that increase vector mortality and thus reduce population-level fecundity. Vector models which incorporate capacity constraints and behavioural change are a natural extension that may better represent competition for larval habitats [93]. However, to date, such models have not been adequately validated against weather measurements and entomological data and thus further work is required to obtain a model that can reproduce entomological patterns from multiple transmission settings. Last, our current model has been developed and parameterized to be applied to single locations. It thus considers isolated areas and does not address the focal and heterogeneous nature of transmission on a wider spatial scale or the connectedness of local populations. As such, the current model cannot be used to assess the risk of reintroduction of the parasite from outside areas, which has been shown to be a major challenge in ongoing control [94]. However, it is possible to extend this framework to a fully spatial continental-scale simulator. The major challenge here is not in developing the software tool but in parameterising the model across settings. Basic requirements of such a model, e.g. human population size in each area, are not well known across parts of Africa, although synthetic data derived from satellite observations can be used as a proxy [95],[96]. In addition, such models require local-level information on vector species, seasonality patterns, intensity of transmission, and human movements to enable assessment of the risks of transmission spatially. Despite these limitations, mathematical models based on the biology of the transmission cycle provide an appropriate tool for a range of stakeholders to explore the potential impact of current and future interventions on malaria transmission and disease burden in a systematic manner. Further development of the models and approaches outlined here can help to identify optimal policies for the range of stages of malaria elimination programs from the consolidation phase outlined here, through the pre-elimination and elimination phases, to sustained elimination. By considering current tools and exploring potential future interventions, models can help us to understand the limits of current strategies and evaluate the potential for future products to achieve the ultimate goal of global eradication. Supporting Information Alternative Language Abstract S1 Abstract translated into French by Emilie Pothin. (0.03 MB DOC) Click here for additional data file. Alternative Language Abstract S2 Abstract translated into Spanish by MGB. (0.03 MB DOC) Click here for additional data file. Alternative Language Abstract S3 Abstract translated into Dutch by TB. (0.03 MB DOC) Click here for additional data file. Alternative Language Abstract S4 Abstract translated into Portuguese by Dr. Jose Sousa-Figueiredo. (0.03 MB DOC) Click here for additional data file. Protocol S1 Transmission model. (0.37 MB DOC) Click here for additional data file. Protocol S2 Intervention models. (0.88 MB DOC) Click here for additional data file. Protocol S3 Bayesian model fitting and parameter values. (0.92 MB DOC) Click here for additional data file. Protocol S4 Seasonal patterns and transmission settings. (0.49 MB DOC) Click here for additional data file. Protocol S5 Additional results and sensitivity analyses. (1.26 MB DOC) Click here for additional data file. Protocol S6 User-friendly software for model runs. (0.17 MB DOC) Click here for additional data file.

0 comments Cited 283 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: found

Is Open Access

Outdoor host seeking behaviour of Anopheles gambiae mosquitoes following initiation of malaria vector control on Bioko Island, Equatorial Guinea

Michael Reddy, Hans Overgaard, Simon Abaga … (2011)

Background Indoor-based anti-vector interventions remain the preferred means of reducing risk of malaria transmission in malaria endemic areas around the world. Despite demonstrated success in reducing human-mosquito interactions, these methods are effective solely against endophilic vectors. It may be that outdoor locations serve as an important venue of host seeking by Anopheles gambiae sensu lato (s.l.) mosquitoes where indoor vector suppression measures are employed. This paper describes the host seeking activity of anopheline mosquito vectors in the Punta Europa region of Bioko Island, Equatorial Guinea. In this area, An. gambiae sensu stricto (s.s.) is the primary malaria vector. The goal of the paper is to evaluate the importance of An gambiae s.l. outdoor host seeking behaviour and discuss its implications for anti-vector interventions. Methods The venue and temporal characteristics of host seeking by anopheline vectors in a hyperendemic setting was evaluated using human landing collections conducted inside and outside homes in three villages during both the wet and dry seasons in 2007 and 2008. Additionally, five bi-monthly human landing collections were conducted throughout 2009. Collections were segregated hourly to provide a time distribution of host-seeking behaviour. Results Surprisingly high levels of outdoor biting by An. gambiae senso stricto and An. melas vectors were observed throughout the night, including during the early evening and morning hours when human hosts are often outdoors. As reported previously, An. gambiae s.s. is the primary malaria vector in the Punta Europa region, where it seeks hosts outdoors at least as much as it does indoors. Further, approximately 40% of An. gambiae s.l. are feeding at times when people are often outdoors, where they are not protected by IRS or LLINs. Repeated sampling over two consecutive dry-wet season cycles indicates that this result is independent of seasonality. Conclusions An. gambiae s.l. mosquitoes currently seek hosts in outdoor venues as much as indoors in the Punta Europa region of Bioko Island. This contrasts with an earlier pre-intervention observation of exclusive endophagy of An. gambiae in this region. In light of this finding, it is proposed that the long term indoor application of insecticides may have resulted in an adaptive shift toward outdoor host seeking in An. gambiae s.s. on Bioko Island.

0 comments Cited 199 times – based on 0 reviews      Review now

Bookmark

Record: found
Abstract: found
Article: not found

Preventing Childhood Malaria in Africa by Protecting Adults from Mosquitoes with Insecticide-Treated Nets

Gerry F. Killeen, Tom Smith, Heather M Ferguson … (2007)

Introduction The massive malaria burden in Africa merits particular attention as the world struggles to realize a better life for the poorest [1,2]. The Anopheles mosquitoes that act as vectors for human Plasmodium parasites must access sugar, blood, and aquatic oviposition sites to complete their life cycle and maintain parasite transmission. The availability of such ecological resources to mosquitoes has long been recognized as a crucial determinant of malaria transmission [3], but quantitative understanding of this process, as well as viable means to prevent it, remain poorly developed compared with other disease [4] and pest systems [5]. Recent theoretical work highlights the enormous influence of blood source and aquatic habitat availability in determining malaria transmission intensity, disease burden, and their responsiveness to various forms of control [6–12]. Here we apply field-parameterized kinetic models of mosquito host availability [11,13] to identify important shortcomings of current global targets for delivering insecticide treated nets (ITNs) [2,14,15], the most important vector control tool in Africa today. Not only does the model outline the limitations of existing strategies that emphasize targeting of vulnerable groups such as young children and pregnant women [16–18], it also indicates how complementary strategies to promote coverage of whole populations, including nonvulnerable adults and older children [19], will achieve greater and more equitable reduction of disease burden than otherwise would be possible. Insecticide-treated nets (ITNs) represent a practical and effective means to prevent malaria in Africa [20], so scaling up coverage to at least 80% use by young children and pregnant women by 2010 is a consensus target of the Millennium Development Goals (MDGs), the Roll Back Malaria Partnership, and the US President's Malaria Initiative [2,14,15]. Targeting individual protection to these vulnerable groups [16–18] is a well-founded and explicitly accepted priority of all three initiatives, because these groups bear the highest risk of morbidity and mortality from malaria. However, this strategy largely ignores the potentially greater community-wide benefits of broader population coverage [19], and no explicit resources, targets, or strategies have been proposed to achieve these benefits. ITNs can protect not only the individuals and households that use them, but also members of the surrounding community [19,21–26]. This is because they kill adult mosquitoes directly or force them to undertake longer, more hazardous foraging expeditions in search of vertebrate blood and aquatic habits [11]. Plasmodium falciparum, the malaria parasite responsible for the bulk of deaths in Africa, requires at least 8 d to develop from imbibed gametocytes into mature sporozoites within the salivary glands of the vector mosquito. This means that most malaria transmission is carried out by mosquitoes that are at least 10 d old and have taken several previous blood meals at intervals of 2–5 d [27,28]. By even modestly increasing mosquito mortality while they attempt to feed on humans, ITNs can greatly reduce the number of mosquitoes that survive repeated hazardous encounters with protected humans [11]. Also, the excito-repellent properties of ITNs can reduce the frequency with which mosquitoes successfully acquire blood, often diverting them to feed on other mammals that do not host the malaria parasite, resulting in greatly reduced prevalence of sporozoite infection [11]. This theoretical rationale is strongly supported by detailed observations from experimental hut studies [29–34] and from larger village-scale trials: ITNs have been clearly shown to reduce malaria risk among unprotected individuals by suppressing the density [35–37], survival [35–37], human blood indices [38,39], and feeding frequency [39] of malaria vector populations. Large reductions of transmission are required to appreciably reduce malaria burden in most of Africa [17,40], particularly in the longer term as exposure and immunity re-equilibrate [41]. ITNs can address this challenging need through direct personal protection and area-wide suppression of the malaria transmission intensity that benefits even nonusers. It has been suggested that such communal benefits can make large impacts on disease burden only if appreciable levels of coverage are achieved in the human population as a whole [11,12,19], but precise coverage targets for achieving this remain to be determined. So how much coverage is enough to protect individuals who do not use an ITN? Methods Overview Here we used recently developed kinetic models of mosquito behaviour and mortality [11,13] to answer this question by considering the impact of ITNs on human host availability and feeding hazards to mosquitoes, as well as the consequences of such changes for malaria transmission intensity. Protection was estimated in terms of protection against exposure to infectious mosquito bites, expressed as the relative change in the entomological inoculation rate (EIR). EIR is a proven epidemiological indicator of malaria transmission intensity and a key determinant of disease burden [17,40]. Two common but ecologically distinct African malaria transmission systems are considered. First, we modelled an Anopheles gambiae Giles or An. arabienis Patton (sibling species from the same species complex known as An. gambiae sensu lato) population with access to human blood only. Second, we considered An. arabiensis populations in the presence of abundant cattle, which can act as alternative blood sources. An. gambiae greatly prefers humans, but An. arabiensis will readily feed upon cattle [42,43], so populations of these species respond quite differently to increasing ITN coverage, with malaria transmission by the latter typically being lower to begin with but less sensitive to control with ITNs [11]. In both transmission systems we considered ITNs with properties typical of those evaluated in rigorous clinical trials [20] or those of emerging technologies with improved operational durability [44–47]. Note that coverage is expressed as the proportion of the total human population using an ITN each night, rather than in terms of ownership, because this value is the most direct indicator of both personal and communal protection. Figure 1 provides an overview of how mosquito behaviour and survival were modelled as a function of host availability, ITN properties, compliance, and coverage. The approach described is essentially a behaviourally explicit extension of existing vector biodemography [48] models, which predict epidemiologically relevant outcomes such as exposure to transmission (the biodemography–epidemiology model). The principles and utility of the biodemography–epidemiology models we have used [27,49,50], as well as several others that are based on similar assumptions [6,18,28,51], are well established. Notably, this family of models realistically assumes that mosquito behaviour cycles between host seeking, feeding, resting, oviposition-site seeking, oviposition, and back to host seeking again [51]. Similarly to recent analyses of the importance of oviposition [7,8,10] and host acquisition [11,12] processes, here we explicitly modelled the underlying behavioural events that determine the input parameters of these biodemographic processes (the behaviour–biodemography model). Detailed consideration of mosquito behaviour and mortality upon encounter with individual hosts (the individual-level submodel) allows simulation of the impact of ITNs upon the foraging requirements and risks for mosquito populations at the community level (the community-level submodel). This hierarchical approach links individual- and community-level submodels into an integrated behaviour–biodemography model, which drives the outcome of the biodemography–epidemiology model and allows the influence of ITNs upon malaria transmission intensity to be estimated in terms of EIR experienced by both users and nonusers [11,27,50]. Figure 1 A Schematic Outline of the Two-Tier Model Used for This Analysis, Adapted from Previous Detailed Descriptions A detailed model of mosquito behaviour and survival as a function of host availability, ITN properties, compliance, and coverage [11,13] was used to estimate the key biodemographic parameters that determine malaria transmission intensity (behaviour–biodemography model). This model allowed the influence of ITN usage upon malaria transmission intensity to be estimated (biodemography–epidemiology model) in terms of EIR experienced by both users and nonusers [11,27,50]. All terms and symbols are defined in detail elsewhere [11,27,50,52] and are summarized in Methods. The specific modelling approach described here is almost identical to our recent exploration of the optimal properties of ITNs as a function of local ecology [11], apart from subtle improvements in terms calculating mosquito diversion, mortality, and feeding probabilities per host encounter. It is also similar to and consistent with the approaches of others [6,12] but accounts for the fact that ITNs can act only during times of the night when they are actually in use, so that their overall protection is also influenced by subtle variations in the behavioural interactions between humans and mosquitoes [13]. This model has already been evaluated through improved iterations in terms of sensitivity to variations in the assumed parameter values for the insecticidal and excito-repellent properties of ITNs [11], the survival rate of mosquitoes while foraging for resources [11], the innate resource preferences of vector populations [11,50,52], and the availability of those resources, including oviposition sites [50] and alternative blood meal hosts [11,50]. While the analysis outlined here could be implemented with either of the recently developed (and perhaps more elegant) alternative models [6,12], this particular form captures all of the same processes without necessitating the mathematical subtleties of integration, differentiation, equilibrium analysis, or limits. While these are inherently valuable tools for mathematical modelling, they often constitute “black boxes” to nonmathematicians, including several authors of this article. We therefore chose a model that does not require mathematical complexities that might limit accessibility to some of the field biologists and epidemiologists for whom this analysis is most relevant. The model is presented as a downloadable spreadsheet (see Protocol S1) and has proven valuable for teaching the ecological basis of malaria epidemiology and control to students in both the developed and developing world. Modelling Mosquito Behaviour and Mortality at the Individual Level Here we describe a submodel of behavioural and mortality processes that occur at the level of individual mosquitoes seeking, encountering, attacking, and feeding upon individual blood hosts. Another important simplification to consider is that, like most deterministic malaria transmission models, our approach assumed a “malaria in a bottle” scenario in which populations of identical parasites, vectors, and hosts are mixed homogenously within an enclosed system [53]. One important corollary of this assumption is that well-established variations of vulnerability to malaria infection within human populations [16,17] or associated variations in attractiveness and availability to mosquitoes [9,54–56] are not explicitly modelled. As defined previously [52], the availability (a) of any host (j) of any species (s) is the product of the rate at which individual vectors encounter it (ɛs,j) and the probability that, once encountered, they will feed upon it (φs,j): Note that this kinetic definition of availability as a rate per unit time is consistent with applications of the same term to acquisition of oviposition sites [10], the term attraction rate for blood sources [6,57], and the terms feeding rate and oviposition rate for both resources [8,12]. We considered successful feeding as just one of three possible outcomes of a host encounter by a female vector, the other two being death while attempting to feed and diversion to seek another host (Figure 1). We considered this a two-stage process in which the vector first either attacks the encountered host or is diverted away and searches for another, the probabilities of which we denote as γ and Δ, respectively. This definition of diversion includes the combined effects of noncontact repellency and contact-mediated irritancy, often referred to as excito-repellency [58,59]. Considering mean values for hosts of any given species (s), the sum of these two probabilities is: We then considered the second stage of the blood acquisition process, namely feeding. Knowing the probabilities that the vector will either feed successfully (φs) or die in the attempt (μs) per attack (rather than per encounter) allowed us to calculate the probability of a successful feed per encounter: Specifically, the cases of cattle (c) and unprotected humans (h,u) were dealt with in a straightforward manner as follows, where Δu and μu represent a common parameter value for both types of host (Table 1): Table 1 Behavioural and Host Availability Input Parameters for Both Vector Species Personal protection measures such as bed nets, repellents, or domestic insecticide use were envisaged as three possible outcomes, the probabilities of which sum to 1: For a vector that would normally choose to feed upon an encountered unprotected human with a probability of φh,u, the presence of a net or other intervention is expected to influence this probability for protected humans (φh,p) as a function of the excess probability of diverting (Δp) and killing (μp) that vector (Figure 1). The combined baseline and net-induced probabilities of diversion (Δu + p ) or mortality (μu + p) were calculated as follows: and These parameters allowed us to calculate the feeding probability for a human who always uses and is protected by a net (φh,p): These equations are parameterized using data from experimental hut trials in which the human participants slept within the net throughout the period of data collection (Table 1). However, very few human beings spend their entire day asleep or using a net [13] so the true probability of feeding upon a typical net user ( ) is calculated by weighting φh,u and φh,p according to the proportion of normal exposure during which the host is actually covered (πi): Equations 5-7 differ slightly from those previously proposed [11], which treated diversion and killing as independent events, conditional on the host having and using a net. At low values of πi these changes relative to [11] make little difference, but the model described here is more realistic at high values of πi . Extrapolating Impacts of Insecticide-Treated Nets to the Community Level Given the above submodel for the interactions of mosquitoes with individual mammalian hosts, it was possible to extrapolate the likely large-area effects of these small-scale influences on entire vector populations and the human communities they feed upon. For any given number of cattle (Nc), unprotected humans (Nh,u), and protected humans (Nh,p), the mean seeking interval for vertebrate hosts (ηv) can be calculated as the reciprocal of total host availability (A) [52], using estimates of these feeding probabilities and their corresponding encounter rates, adapting Equation 1 from our original formulation [50]: where As refers to the total availability of all hosts of species s. In this case, the species or species categories considered were unprotected humans (h,u), protected humans (h,p), and cattle (c). Values for ac and ah,u (previously ah [50]) were estimated exactly as described previously [50] and ah,p was calculated as follows: where λp is the relative availability of protected versus unprotected hosts, estimated in terms of the ratio of their feeding probabilities: Foraging for resources is an intrinsically dangerous undertaking for mosquitoes, and it is commonly assumed that survival during these phases is lower than while resting in houses [6,60]. We adapted Equation 3 from our previous formulation [50] to estimate the survival rate per feeding cycle (Pf) as the product of the probability of surviving the eventual attack on a host that may be protected (Pγ) and the probabilities of surviving the gestation (g), oviposition site-seeking (ηo), and vertebrate host-seeking (ηv) intervals, with distinct daily survival probabilities for the resting (P), foraging for either oviposition sites or vertebrate hosts (Pov), and attacking (Pγ) phases: The mean probability of mosquitoes surviving their eventual chosen host attack (Pγ) was calculated assuming that the proportion of all attacks that end in death is the sum of the mortality probabilities for attacking protected and unprotected hosts, weighted according to the proportion of all encounters that will occur on such hosts. Assuming that protection does not affect encounter rates, and that these rates are proportional to availability when unprotected, we applied this weighting approach to estimate total attack-related mortality rate and consequent survival as follows: Similarly, the human blood index is calculated as the proportion of total host availability accounted for by humans [52], similarly to Equation 9: The EIR for protected and unprotected individuals was then calculated from the total number of infectious bites upon humans that occur in the population as a whole (β E) [27,49], the share of the total human availability represented by that group, and the population size of that group: where β is the mean number of infectious human bites each emerging mosquito takes in its lifetime and E is the emergence rate of mosquitoes [27]. Dividing Equation 16 by Equation 15, substituting with Equation 10, and rearranging also leads to an intuitively satisfactory solution, consistent with independently formulated models of personal protection [13]: Otherwise, we modelled malaria transmission exactly as previously described [50]. Note that this model has been adapted [11,50] from its original formulation [27] to account for superinfection of mosquitoes [28] and daily time increments to smooth the effects of changing host availability patterns on feeding cycle length [50]. For ease of comparison and interpretation, the impact of ITNs is presented in terms of the relative transmission intensity EIR C /EIR 0 at a given coverage level (C; note distinction from c, which denotes cattle hosts) as a result of personal and communal protection amongst users and nonusers: Baseline Mosquito Behaviour, Host Availability, and Survival Parameters The parameter definitions and values used to implement this analysis are summarized in Table 1. Namwawala, in the Kilombero Valley, southern Tanzania is the primary centre for parameterising our model because of the exceptionally detailed quantitative characterisation of malaria transmission and vector biodemography in this village and the surrounding area. This is a holoendemic village with intense seasonal transmission, stable high parasite prevalence in humans, and a heavy burden of clinical malaria [61–68]. At this site the bulk of transmission is mediated by An. gambiae sensu lato (of which the main species involved in transmission is An. arabiensis) and transmission intensity has been modelled with available field data [27,49]. As previously described [27,49], we based our estimate of human population size [62] approximately upon those reported for this particular village during the early 1990s. Nevertheless, we used a human population size of 1,000 and, where relevant, a bovine population of the same size so that the EIR experienced by users and nonusers could be easily calculated at net coverage levels approaching 0% and 100%. By setting coverage to 0.001 or 0.999, this model simulates a single user or nonuser in the population, respectively. Infectiousness of humans (κ) is set to 0.030, reflecting a more precise recent estimate [69] than was available previously [61,63]. In a typical holoendemic scenario, the infectiousness of the human population is thought to be largely insensitive to reductions in transmission intensity [69]. In the interests of making conservative and generalizable predictions, we assumed that increasing coverage with ITNs will not affect κ [69], even though reduction of κ is likely at EIR values below 10 infectious bites per person per year [56]. We set mean daily survival of the resting phase (P) at 0.90, reflecting a median value of daily survival at four well-characterised holoendemic sites [27] and estimated daily indoor survival for An. gambiae s.l. in Tanzania [70]. As previously described, the daily survival rate of mosquitoes while foraging for blood or oviposition sites (Pov) was set at 0.80, representing a median value of plausible field values [11]. The results of experimental hut studies [34] were combined with host-choice evaluations [71] and appropriate analytical models [50,52] to define the attack and mortality probabilities of An. arabiensis encountering cattle or humans: we set the probability that An. arabiensis will attack unprotected cattle or humans (γu), conditional upon encountering them, to be 0.90 and the chance that they will die in the attempt (μu) at 0.10. Using these parameters and Equation 3, we calculated that, for An. arabiensis, the overall feeding probability upon either cattle (φc) or unprotected humans (φh,u) would be 0.81, a value similar to previous estimates of approximately 0.80–0.85 for the feeding success of An. gambiae sensu lato on sleeping humans in Tanzania [34,62]. We also applied these same probabilities of attacking (γu), feeding (φh,u), and dying (μu) to An. gambiae sensu stricto encountering unprotected humans. The availabilities of unprotected humans and cattle were calculated for An. arabiensis using field measurements of the duration of the feeding cycle and were extended to An. gambiae s.s., accounting for the lower estimated relative availability of cattle (λc) to this mosquito species as previously described [52]. Note that λc is assumed to modify a c by affecting the encounter rate only, indicating that these mosquitoes can differentiate between preferred and nonpreferred hosts at long ranges [72–74]. In the case of An. arabiensis this assumption is consistent with the longer spatial range of attraction of cows relative to humans for zoophilic members of the An. gambiae complex [72–74]. Parameters Reflecting the Effects of Insecticide-Treated Bed Nets The parameter definitions and values describing the impacts of ITNs on vector behaviour and mortality at the level of individual interactions are listed in Table 1. The impacts of ITNs very much depend on their excito-repellent and insecticidal properties, which are most representatively evaluated using well-established experimental hut methodologies [59,75,76] that have been extensively applied to this particular intervention [29–34]. Furthermore, the interaction of these two properties, to yield varying levels of personal and communal protection, is complex and has crucial implications for ITN programmes across Africa [11]. Sensitivity analysis of models similar to those used in this paper [11] have previously been used to explore the influence that these properties might have upon the magnitude and equity of protection afforded by ITNs (Figure 2). In order to validate this slightly revised model (see Equations 4-8) and similarly investigate such interactions at ITN coverage levels that can be plausibly sustained, we examined usage data collected during routine socioeconomic status surveys of a long-standing demographic surveillance system in the Kilombero Valley, southern Tanzania, where social marketing programmes have been well established since 1997 [77,78]. Data from the annual ITN usage survey in 2004 were used because they overlap with detailed entomological surveys of malaria transmission (which will be reported elsewhere). These surveys of randomly sampled residents from across two rural districts indicate that 75% (11,982/16,086) net use was achieved although most of these nets were not effectively treated [79]. In this sensitivity analysis, we assumed that new long-lasting ITN technologies [44–47] will enable sustained coverage with nets that are effectively treated even under the most rigorous programmatic field conditions. Figure 2 The Simulated Protection ITNs Afford against Exposure to Malaria Transmission as a Function of Their Ability to Divert and Kill Host-Seeking Mosquitoes Protection is expressed as relative exposure to malaria transmission (EIR C /EIR o ) for individuals with (Equation 19) and without (Equation 18) nets is plotted as a function of their ability to divert (Δp) and kill (μp) mosquitoes attacking protected humans. To simulate the likely field properties of existing long-lasting insecticidal nets with a full range of insecticidal and excito-repellent properties, the parameters of this model reflecting increased mosquito mortality (μp) and diversion (Δp) were varied across a plausible range of 0–0.8. As described in the main text and previous publications, these results represent simulations in two distinctive scenarios: An. gambiae sensu lato in the absence of cattle (results for both sibling species are identical) and An. arabiensis in the presence of one head of cattle per person. The biodemographic parameters of the interacting vector and parasite are also exactly as described previously [11,13] with survival of foraging mosquitoes (Pov) set at 0.8 per day. Coverage levels of 75% net usage was assumed, consistent with the results of surveys in the Kilombero Valley, southern Tanzania (see Methods: Parameters Reflecting the Effects of Insecticide-Treated Bed Nets). Figure 2 shows that, for the comparatively zoophilic vector An. arabiensis, in the presence of alternative hosts, excito-repellency consistently enhances the benefits for both users and nonusers, regardless of the insecticidal properties of the net. Consistent with previous analyses using this model [11], this simulation suggests that nets that are purely excito-repellent and lack insecticidal properties could slightly increase exposure of nonusers to An. gambiae sensu lato by diverting mosquitoes to them where no alternative sources of blood are available. Thus, purely diversionary vector control strategies may indeed be ethically questionable, as was previously suggested [31,34,80,81]. Nevertheless, even modest insecticidal properties are expected to counterbalance this inequity and confer a useful communal reduction of EIR. While repellent properties do slightly reduce the benefits to nonusers exposed to anthropophagic vectors lacking an alternative host, this slight disadvantage is likely to be outweighed in practice by the advantage of improved personal protection for users: Excito-repellent properties and physical barriers add to the effectiveness of insecticides for personal protection because these two incentives constitute the major motivating force behind ITN uptake and use at the individual and subsequently the community level. It is also reassuring to note that the predictions and epidemiological implications of this slightly revised model are very similar to those reported for its previous iteration [11]. We therefore concluded that the simulations described in the main text should consider ITNs with both insecticidal and excito-repellent properties, consistent with those of products currently on the market that have been evaluated in a variety of settings and experimental designs. To simulate the likely properties of established ITNs under programmatic conditions, we conservatively assumed they will both divert and kill 40% more mosquitoes than an unprotected human (μp = 0.4 and Δp = 0.4). A net with such proper-ties would protect against 64% of indoor exposure (1 − [(1 − 0.4) × (1 − 0.4)] = 0.64), as measured in a typical experimental hut trial [46,76]. To explore the best possible future scenario for the development of highly durable ITNs [44–47] or regular retreatment services [82], we also simulated increasing co-verage with nets that divert and kill 80% more mosquitoes than with an unprotected human (μp = 0.8 and Δp = 0.8), providing 96% protection (1 − [(1 − 0.8) × (1 − 0.8)] = 0.96). The proportion of normal biting exposure that occurs while nets are actually in use (πi) has been estimated as 90% for A. gambiae in southern Tanzania [13], so we set πi to a value of 0.90. Results Figure 3 illustrates how increasing community-level protection of ITN nonusers and users alike combines with constant individual protection to reduce exposure to malaria. Regardless of vector species or the availability of alternative hosts, modestly effective conventional ITNs achieve much greater impact upon human exposure, even that of users, if approximately half or more of the whole human population is covered. While this principle has already been suggested by field trials [19] and two independently formulated models [11,12], here we have identified specific coverage thresholds at which communal protection becomes greater than or equal to individual personal protection. Where alternative hosts for vector mosquitoes are absent, 35% of the human population must sleep under regular ITNs to achieve equivalence of personal and communal protection mechanisms, resulting in major community-wide suppression of exposure. The same target is achieved at 55% coverage where alternative hosts such as cattle are present. Figure 3 Relative Exposure to Malaria Transmission (EIR C /EIR o ) as a Function of Increasing Coverage with Insecticide-Treated Nets We express coverage as the proportion of the total human population using an ITN each night, and protection as the proportional reduction of infectious bites to which a resident is exposed (see Methods). Individual protection afforded to users (thin solid line; Equation 20) and communal protection afforded to nonusers (thick dashed line; Equation 18), as well as their combined effect on users (thick solid line; Equation 19) are separately calculated [11,13]. Two distinct but common and broadly distributed ecological scenarios in Africa are considered: (1) An. gambiae or An. arabienis (sibling species of the same species complex known as An. gambiae sensu lato) populations in the absence of alternative blood sources and (2) vector populations dominated by An. arabiensis in the presence of abundant cattle as alternative hosts. Both scenarios are simulated with ITNs that have either standard or improved properties (See Methods). Grey shading represents an approximate absolute maximum for community-level coverage achievable by covering vulnerable under five years of age and pregnant population groups only with perfect targeting efficiency. Arrows extrapolate the thresholds at which communal and personal protection are equivalent. The insecticidal and excito-repellent properties of ITNs that define levels of personal protection also determine the extent of community-wide alleviation of exposure amongst users and nonusers alike [11], so improved ITN properties consistently result in improved overall impact. In our model, slightly higher usage rates were required to achieve equivalence of individual and communal effects, with thresholds of 40% and 64% coverage for vector populations with and without alternative hosts, respectively (Figure 3). While emerging ITN technologies with long-lasting insecticidal properties under programmatic conditions [44] would confer useful personal protection even at low coverage levels, personal protection was greatly enhanced by communal protection. At the 75% total population coverage recently achieved with largely untreated nets in southern Tanzania (Killeen et al., unpublished data), net users and nonusers are predicted to receive >98% and >90% protection, respectively, regardless of ecological scenario, if those nets were to be replaced with improved long-lasting insecticidal nets. Even for users of improved ITNs, this level of protection against African vector species is impossible without the contribution of community-level transmission suppression, because at least 10% of exposure occurs outdoors during times of the night when nets are not in use [13,83]. We conclude that modest coverage (thresholds of approximately 35%–65% use, depending on ecological scenario) of entire malaria-endemic populations, rather than just the most vulnerable minority, is needed to realize the full potential of ITNs, even with longer-lasting products or regular retreatment services [14,44]. This range of modelled thresholds is remarkably consistent with the figure of 50% suggested by large-scale field trials using approximately equivalent technology [19]. Discussion In addition to the direct impacts on vector populations explicitly modelled above, coverage of adults and older children is likely to have further benefits arising from subtleties of mosquito resource utilization that are often under-appreciated. Over 80% of human-to-mosquito transmission originates from adults and children over five years of age, because these groups constitute the bulk of the population and are more attractive to mosquitoes [56]. Where the entomological inoculation rate is fewer than ten infectious bites per person per year, the distributions of infectiousness [56,69], morbidity, and mortality will all shift into these older age groups, necessitating protection of all members of the population. Under such conditions, ITNs could suppress transmission not only through direct impacts on mosquito mortality, host choice, and feeding frequency [11], but also by limiting the prevalence, density, and infectiousness of malaria parasites in the human population [56]. An under-emphasized feature of communal protection is the enhancement of ITN programme equity, regardless of ecological scenario or ITN effectiveness: If the majority of people living in malaria-endemic Africa regularly used existing ITN technologies, nonusers would receive communal protection at least equivalent to using the only ITN in an otherwise unprotected population (Figure 3). This means that all children would equitably receive communal protection at least equivalent to the personal protection of an ITN, with users receiving multiplicative combined effects on exposure of both personal and communal benefits. While the wisdom of targeting interventions to protect at-risk individuals is based on solid scientific grounds [9,18,84] and is widely accepted [16], this approach should not preclude efforts to maximize communal protection through less selective delivery mechanisms. Targeting limited subsidies to maximize personal protection of the most vulnerable should remain a priority, but more equitable and effective suppression of risk for entire populations, including vulnerable groups, can be attained with quite modest coverage across all ages. Most field evaluations of ITNs have been conducted at reasonably high coverage levels [19], and all five mortality trials [21,85–88] that estimated that ITNs save 5.5 lives for every 1,000 children protected [20] covered large portions of entire communities rather than only the children themselves. The choice of ITN delivery strategy has proven contentious in recent years [89,90], but proponents of both market-based and public-sector approaches equally emphasize targeting strategies [9,16,84] to enhance equity and minimize leakage of subsidized ITNs beyond intended target groups [91–94]. While optimal targeting of finite subsidies is highly desirable, there are fundamental limitations to the impact that can be achieved: Even if resources were perfectly targeted, 80% coverage of pregnant women and children under five years of age could be accomplished with less than 20% coverage of the whole population, and even less of the total human host availability [11,56], as well as the infectious parasite reservoir [56,69]. Even if the ITN coverage targets of the MDGs were attained with flawless targeting efficiency, the substantial and equitable benefits of communal protection would not be achieved. Specifically, the target of 70% less exposure to transmission [13] would not be attained by the remaining minority of vulnerable individuals who are not covered and do not use an ITN, regardless of ecological scenario or ITN properties (Figure 3). We therefore highlight an important caveat to the following conclusion of the current Global Strategic Framework for ITN scaleup in Africa [95]: “In order to achieve maximum public health impact, ITN coverage needs to be maximized amongst those population groups that are most vulnerable to malaria infection and its consequences, primarily pregnant women and children under five years of age.” Specifically, we conclude that protecting the vulnerable can achieve maximum public health impact only if complemented by strategies that also achieve broad coverage of the population as a whole. In reality, the targets for coverage of vulnerable groups will not be reached without some leakage and inequity. Our analysis suggests that such concerns may be less of a problem than the targets themselves and may be minimized by extending coverage priorities to include all age groups. Fortunately, consensus is finally emerging that a range of approaches to ITN deployment merit investigation, development, and comparative evaluation at scales for which no precedent yet exists [95]. Note that this analysis supports the implementation of any of the diverse and rapidly emerging delivery strategies as long as high coverage with long-lasting ITNs is sustained across entire malaria-endemic populations on national scales. Perhaps the most important remaining question is: How can such population-wide coverage levels be affordably and cost-effectively sustained? Growing financial support for malaria control globally [14,15,95] may enable fully subsidized provision to entire populations [82] of the world's most impoverished, malaria-afflicted nations. Existing evidence, based largely on individual protection alone, indicates that ITNs are as cost-effective as childhood immunization [96], and future analyses should explicitly consider the additional benefits of communal protection. Implementing this goal may be relatively straightforward for programmes that are primarily subsidized and implemented through the public sector, such as recent successful initiatives associated with vaccination campaigns [91]. By comparison, social marketing approaches, including hybrid systems that deliver public subsidies through the private sector, may require more detailed consideration, particularly where cost sharing with the target population is substantial and biased toward the nonpregnant adults and older children who are key to communal protection. Although social marketing approaches to ITN distribution face substantial challenges [93,97,98], notable success in terms of coverage and impact have been reported in a variety of settings [94,99,100], including the KINET programme in Kilombero Valley, southern Tanzania where ITNs have been promoted and subsidized since 1996 [77,78]. Much of the essential experience generated by KINET was later integrated into the ITN promotion strategy of the National Malaria Control Programme of Tanzania, which supports private sector distribution through a voucher system that subsidizes purchase by vulnerable priority groups [101]. In the meantime, the preceding KINET pilot in Kilombero has achieved 75 % net use amongst randomly sampled residents of all ages (Killeen et al., unpublished data). It is particularly noteworthy that substantial levels of communal protection were achieved [102] (unpublished data) even though most of these nets were untreated or poorly treated at the time of evaluation [79] (unpublished data). Reassuringly, the model applied here approximately reproduces these patterns of communal protection using plausible parameter estimates for the net properties, vector behaviours, and host demographics of the area (unpublished data). We therefore recommend that the cost-effectiveness of such hybrid approaches be explicitly evaluated in terms of the complementary respective contributions of public-sector subsidies and cost-sharing by target populations to personal and communal protection. While appropriate engagement and sensitization of malaria-afflicted populations is essential to the success of any ITN promotion programme, this is likely to be especially true where cost-sharing by the target population will be needed to complement limited public subsidies. Such cost-sharing schemes may be the only affordable means to support full population coverage where available subsidies are inadequate. In such resource-limited circumstances, high levels of awareness, acceptance, and willingness to pay will be essential to enable concerted use of ITNs by adults and shared protection of all children within their communities. Overly confident extrapolation from mathematical models to set operational targets for malaria control has proved to be a grave mistake in the past [103]. A number of complications not captured by this model could emerge as ITN coverage increases, not least of which might be increased selection for insecticide resistance [104,105]. While we urge caution in interpreting the numerical results of our analysis, the phenomenon outlined is well established and has clear implications for malaria control in Africa and beyond [19]. In fact, the analysis presented here provides a generalizable rationale that strongly supports the conclusions of the most recent and meticulous evaluations of the community-level benefits of ITNs: “High coverage with ITNs will do more for public health in Africa than previously imagined” [19]. We therefore suggest that further field data, analyzed with appropriate theoretical models and cost-effectiveness frameworks, are required to verify and quantify the levels of communal protection afforded by increasing ITN use across Africa. International targets [2,14,15] should be amended to include thresholds for coverage of entire populations and monitored accordingly. By making life increasingly difficult for mosquitoes through programmes that promote ITN use by the majority of their human victims, it may be possible to protect the 15%–20% of children and pregnant women in African communities who would not otherwise be covered even if existing personal protection targets of the MDGs [2], the Roll Back Malaria Partnership [14], or the U.S. President's Malaria Initiative [15] were to be achieved. Supporting Information Protocol S1 Model Spreadsheet A Microsoft Excel spreadsheet version of all model simulations presented here is available to download. (1.1 MB XLS) Click here for additional data file.

0 comments Cited 191 times – based on 0 reviews      Review now

Bookmark

All references

Author and article information

Contributors

Ellie Sherrard-Smith: e.sherrard-smith@imperial.ac.uk

Journal

Journal ID (nlm-ta): Nat Commun

Journal ID (iso-abbrev): Nat Commun

Title: Nature Communications

Publisher: Nature Publishing Group UK (London )

ISSN (Electronic): 2041-1723

Publication date (Electronic): 26 November 2018

Publication date PMC-release: 26 November 2018

Publication date Collection: 2018

Volume: 9

Electronic Location Identifier: 4982

Affiliations

[1 ]ISNI 0000 0001 2113 8111, GRID grid.7445.2, MRC Centre for Global Infectious Disease Analysis, Department of Infectious Disease Epidemiology, , Imperial College London, ; Norfolk Place, London, W2 1PG UK

[2 ]ISNI 0000 0001 2171 1133, GRID grid.4868.2, School of Mathematical Sciences, , Queen Mary University of London, ; Mile End Road, London, E1 4NS UK

[3 ]ISNI 0000 0001 2097 0141, GRID grid.121334.6, Institut de Recherche pour le Développement (IRD), Maladies Infectieuses et Vecteurs, Ecologie, Génétique, Evolution et Contrôle (MIVEGEC), , University of Montpellier, ; 34393 Montpellier Cedex 5, France

[4 ]GRID grid.452477.7, Institut Pierre Richet, ; BP1500 Bouaké, Côte d’Ivoire

[5 ]ISNI 0000 0001 0382 0205, GRID grid.412037.3, Faculté des Sciences et Techniques, , Université d’Abomey-Calavi, ; Cotonou, Benin

[6 ]ISNI 0000 0004 0587 0574, GRID grid.416786.a, Epidemiology and Public Health Department, , Swiss Tropical and Public Health Institute, ; Socinstrasse 57, PO Box, 4002 Basel, Switzerland

[7 ]ISNI 0000 0004 1937 0642, GRID grid.6612.3, University of Basel, ; Petersplatz 1, 4001 Basel, Switzerland

[8 ]ISNI 0000 0000 9144 642X, GRID grid.414543.3, Ifakara Health Institute, ; Bagamoyo Research and Training Centre, Bagamoyo, Pwani, Tanzania

[9 ]GRID grid.452416.0, Innovative Vector Control Consortium, ; Pembroke Place, Liverpool, L3 5QA UK

[10 ]ISNI 0000 0001 0697 1172, GRID grid.462846.a, Centre Suisse de Recherches Scientifiques en Cote d’Ivoire, ; Abidjan 01, BP 1303 Abidjan, Côte d’Ivoire

[11 ]ISNI 0000 0004 0425 469X, GRID grid.8991.9, Department of Disease Control, Faculty of Infectious and Tropical Diseases, , London School of Hygiene and Tropical Medicine, ; Keppel Street, London, WC1E 7HT UK

[12 ]ISNI 0000 0004 0384 7952, GRID grid.417585.a, PMI VectorLink Project, , Abt Associates, ; 6130 Executive Boulevard, Rockville, MD 20852 USA

[13 ]GRID grid.473220.0, Centre de Recherche Entomologique de Cotonou (CREC), ; Cotonou, Benin

Author information

Pie Müller http://orcid.org/0000-0003-0181-5781

Thomas S. Churcher http://orcid.org/0000-0002-8442-0525

Article

Publisher ID: 7357

DOI: 10.1038/s41467-018-07357-w

PMC ID: 6255894

PubMed ID: 30478327

SO-VID: b4ab1b49-5c11-479c-9b21-e0ff415ab4d8

License:

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

History

Date received : 18 June 2018

Date accepted : 25 October 2018

Custom metadata

ScienceOpen disciplines: Uncategorized

Data availability:

ScienceOpen disciplines: Uncategorized

Comments

Comment on this article

scite_

Cited by 62

See all cited by

Most referenced authors 576

See all reference authors

Systematic review of indoor residual spray efficacy and effectiveness against Plasmodium falciparum in Africa

Read this article at

Abstract

Abstract

Related collections

Efficacy of Communicating Science

Most cited references 52

Reducing Plasmodium falciparum Malaria Transmission in Africa: A Model-Based Evaluation of Intervention Strategies

Outdoor host seeking behaviour of Anopheles gambiae mosquitoes following initiation of malaria vector control on Bioko Island, Equatorial Guinea

Preventing Childhood Malaria in Africa by Protecting Adults from Mosquitoes with Insecticide-Treated Nets

Author and article information

Contributors

Journal

Affiliations

Author information

Article

History

Categories

Custom metadata

Comments

Comment on this article

Similar content 281

Cited by 62

Most referenced authors 576