The uptake of any ivermectin-based strategy by countries will depend on the presence of a clear WHO policy recommendation that is in turn supported by relevant evidence regarding efficacy and safety, as well as data on cost effectiveness, ethics, and community acceptance.
Role within WHO to assess the use of ivermectin for malaria
Once consensus on settings, comparators and outcome measures of new trials has been reached, evidence would likely be evaluated by the Malaria Policy Advisory Committee (MPAC) of the Global Malaria Programme at WHO. Given the geographic and disease overlaps, the interface between the malaria and NTD programmes will play an important role, and there are precedents for cross-WHO coordination to guide and evaluate product development and policy recommendations.
Refining the evidence needed for a WHO policy recommendation
It will be important to align the development of any ivermectin-based tool with the unique requirements of health systems of the endemic countries in which it would be used [31]. The type of evidence required during the WHO policy development process has been reviewed by Milstien et al. based on the introduction of malaria intermittent preventive treatment in infancy (IPTi) and four relatively recent vaccines as a case study for new malaria vaccines [32]. Their conclusions were used as guidance for the present section. The evidence needed for a policy recommendation can be classified in four main categories: efficacy, safety, feasibility and cost-effectiveness. Given the particular nature of an ivermectin-based tool to reduce malaria transmission, the category acceptability is also included here.
Key policy questions
Recommendation of ivermectin will be based on proven efficacy, safety, cost-effectiveness and feasibility for the geographies and populations where it would be used. Pivotal questions related to these four aspects are posed and answered below.
Efficacy
(i) Is there evidence of an acceptable level of reduction of morbidity and/or mortality in the target populations?
Using transmission-blocking vaccines as a proxy, “there is currently no clinical trial data available to determine the efficacy threshold that would be required to have a clinically beneficial impact on transmission and achieve elimination” [20]. What is considered an “acceptable” efficacy threshold for ivermectin must be estimated with the help of modelling and validated with empirical data during clinical trials? At a minimum, this must be statistically different than the referent (standard vector control and case management) in a well-designed, sufficiently-powered trial, but it should also be of public health relevance. Of note, given the mandate of providing population at risk with either LLINs or IRS any ivermectin MDA trial would be assessing the superiority of the combination which will require larger trial size. This incremental impact will be considered differently depending on the settings. Various epidemiological settings should be tested with priority given to pre-elimination settings where additional new tools are needed to cover the last mile to elimination.
(ii) Is the efficacy demonstrated in different malaria endemicity levels?
Different scenarios for the use of ivermectin to reduce malaria transmission have been considered [15], reflecting the variety of malaria endemicity conditions and elimination scenarios in which it will be used. It is possible that the dosage/dosing regimen combinations will need to be optimized to different scenarios. All scenarios cannot possibly be tested prior to recommendation, but a relevant strategy (dose and regimen) could be based on current approaches to MDA (3-day regimes) or, perhaps, an expansion of seasonal dosing schemes such as seasonal malaria chemoprophylaxis (SMC), although this last approach would require adaptation to include all ages rather than just children and drug–drug interaction studies with SMC drugs. An initial approach to the upper limit of ivermectin dose could be based on the cumulative dose recommended to patients with severe crusted scabies (up to seven 200 mcg/kg doses in a month) [33].
(iii) Should the use of an endectocide other than ivermectin be considered?
Other existing endectocides tested as mosquitocidal drugs include eprinomectin, selamectin, moxidectin (all available as systemic insecticides for lifestock) [34], spinosad and nitenpyram (available as systemic insecticides for companion animals) [34] and fipronil (available as a spot-on for companion animals but used systemically under experimental conditions) [35].
Some of the advantages of these alternatives include:
-
Primarily the possibility of selecting preclinical candidates with considerably longer half-life.
-
Possibly reducing concerns about increasing selective pressure on onchocerciasis and soil-transmitted helminths by using the ones with different mode of action.
-
Some of the tested endectocides are effective against Aedes mosquitoes, which makes them attractive for the control of arboviruses. Ivermectin is not effective against Aedes mosquitoes at physiologically relevant concentrations.
Some disadvantages include:
-
Most alternatives are early in development, and thus their safety profile in humans would need to be established. Development of any of these drugs would require extensive toxicological and clinical testing both for safety and efficacy. This would be a longer and costly development pathway that could be pursued in parallel with ivermectin.
-
Unknown efficacy of new compounds on NTDs.
Safety
(i) Is the safety profile acceptable?
In the absence of Loa loa co-endemicity, MDA programmes for onchocerciasis control report no severe adverse reactions to ivermectin and their rate of moderate adverse reactions is ≤1.3% [36]. These include ocular irritation, pruritus, rash, pain (general, lymph nodes, headache and joints), dizziness, weakness, fever, ocular irritation, nausea and diarrhoea [36]. In individuals with a high Loa burden (above 30,000 mf/ml) there is risk of severe adverse event including fata encephalopathy. Such high worm loads are more normally associated with areas of high prevalence which are normally avoided by ivermectin MDA campaigns [37]. However novel screening tools may allow a precise exclusion based on individual risk [38].
(ii) Is there significant adverse impact on other malaria prevention and treatment strategies?
This could occur through interaction of ivermectin and anti-malarials and should be addressed during development, particularly with ACT and HIV/TB drugs by means of pharmacokinetic studies [19].
(iii) What is the safety profile in immunologically compromised groups, i.e. HIV-infected?
Ivermectin can be used to treat crusted scabies and strongyloidiasis in HIV-positive patients. During MDA, individuals are not stratified according to their serological status; only pregnant women, lactating women in the first week after birth, children <90 cm in height (approximately 15 kg) and the severely ill are systematically excluded [39]. The safety questions in high risk groups will be related to the new dose and dosing schemes proposed that are the same as the rest of the population.
Acceptability
(i) Would an “only” transmission-blocking intervention be acceptable?
The reduction in malaria transmission achieved through ivermectin would mostly derive from mosquito mortality [22], hence ivermectin should be seen as a new paradigm of vector control, as opposed to a transmission-blocking drug that would treat malaria and also decrease transmission [15]. Moreover, as currently envisioned, ivermectin is not a stand-alone tool, but rather a complementary vector control strategy to be added to the emerging elimination strategy. Finally, the use of ivermectin will provide personal benefit in terms of NTDs and ectoparasites. The caveat is animal studies that indicate a direct of effect of ivermectin on Plasmodium liver stages [6, 7]. This is preliminary, intriguing and needs to be better understood, in terms of mechanism and possible effect in humans.
(ii) Potential consequences of malaria ivermectin MDA for NTD programmes
Ivermectin is the drug of choice for the treatment of onchocerciasis. It is also the only drug used in campaigns aimed at eliminating onchocerciasis. In Africa alone, the overlap between onchocerciasis [40] and malaria endemicity [41] is practically 100% as shown in Fig. 1. An increase frequency in the administration of ivermectin (as could be expected if used for malaria) could shorten the time to interrupt transmission of onchocerciasis in certain settings [42] and has been previously advocated as a necessary measure in areas where interruption of transmission has not been achieved after 10 years of annual treatment [43]. If there is potential to shorten the time during which ivermectin donation is needed, this could have profound implications for the business model used today. Moreover, ivermectin has also been demonstrated, in a triple combination, to have remarkable potential impact on elimination lymphatic filariasis [44].
While the single dose used for each of these diseases is not sufficient for impact on malaria, distribution for malaria indication should suffice as a dose for either disease, so careful coordination between malaria and NTD communities would result in most efficient use of supply. Additionally, ivermectin has at least partial activity against several soil-transmitted helminths and ectoparasites, it is reasonable to expect benefit in this context in communities where an ivermectin-based tool for malaria is implemented [45].
This potential tool will optimally require collaboration between the malaria and NTD programmes, including joint research efforts. Two examples or effective collaboration could be:
-
Data sharing at programme level to optimize timing of ivermectin distribution for malaria and increase impact (dry vs rainy season) and avoid unnecessary duplication of NTD programmes.
-
Ivermectin distribution for malaria with the co-administrations [46, 47] needed as an NTD intervention.
There have been concerns about increasing selective pressure on soil-transmitted helminths and filariae with a wider use of ivermectin. There is limited data on this possibility. Previous reports of ivermectin-resistant Onchocerca [48] have been the subject of debate [49–52]. The drug has in fact been used for decades with excellent results in reducing NTD transmission. Additionally, if used in malaria elimination efforts the number of MDA rounds will be limited. There is previous positive experience on the impact of malaria interventions on NTD transmission such as the possibility to halt LF transmission by scaling up LLINs in Nigeria [53].
There is increasing interest in the potential use of Moxidectin for onchocerciasis [54], having a second drug available for onchocerciasis might help manage resistance concerns. However, given the similarities in molecular structure and mode of action [55] there is potential for co-resistance [56]. The lethal concentration 50 of moxidectin for Anopheles mosquitoes [34] is one order of magnitude above the Cmax reached using maximum moxidectin doses in humans [57]. In the meantime, ivermectin remains the sole drug for the control and elimination of onchocerciasis and an important pillar for the treatment of lymphatic filariasis.
An additional potential risk is diverting the drug supply away from NTD programmes. Yet this is also an opportunity. A novel indication for malaria would increase market and demand, which should serve as incentive for manufacturers to go through the WHO-PQ process.
(iv) What are the expected compliance and adherence? And how could they influence effectiveness?
Effectiveness will be directly related to coverage. Coverage in turn can be greatly influenced by compliance and adherence. Complex and prolonged dosing schemes can negatively impact both [58, 59]. This aspect should be evaluated early through appropriate acceptability studies and addressed by identifying the shortest regimen necessary to have significant impact on malaria transmission.
Use of resources
Thus far, more than 2.7 billion doses of ivermectin have been donated and used by involved countries in Africa, Asia and Latin America, administered to more than 80 million people annually and with no cost for commodities to the public sector. The business model of the Mectizan Donation Programme was expanded in 2010 with the commitment of several pharmaceutical companies, along with NGOs, government agencies and academia to sustain, extend and expand the programmes to ensure the necessary supply of donated drugs to help control and eliminate NTDs [60]. The implication of this business model for ivermectin supply for malaria remains to be worked out as there is no commitment for donating the drug for this purpose. New manufacturers are needed to ensure supply for malaria and NTD elimination, and the public sector will need to understand the cost of goods at scale, to best negotiate of supply and price for malaria programmes.
The WHO guidelines for cost-effectiveness analysis of vector control were issued in 1993 and are now archived [61]. Four basic questions are proposed here, the comments on each question reflect the available data at this time.
(i) What are the expected costs of protection per person?
The median financial costs of protecting one person for one year with core vector control interventions have been estimated in US$ 2.20 (0.88–9.54) for insecticide treated nets and US$ 6.70 (2.22–12.85) for indoor residual spraying [62]. The most important factor affecting cost of goods for drugs is the clinically effective dose in patients [25]. Ivermectin has the advantage of being effective at low doses (µg/kg), which can reduce costs in comparison with drugs needing doses in the grams range. In the context of the Mectizan Donation Programme, the value for donation of one tablet of ivermectin has been calculated at US$ 1.50 [63]. The purchase price in context of large scale purchase for public sector purchase for malaria MDA will likely be much lower. Ivermectin is off-patent since 1996 and apart from Merck, is available from several generic manufacturers [64], although none of these are yet prequalified by WHO.
The programmatic costs of MDA for onchocerciasis and lymphatic filariasis vary according to geography as well as the method chosen for distribution (passive, community-based, community-directed, national mobile teams) [65].
The fact that the efficacy of ivermectin is directly related to blood concentrations and their duration, its small dose per body weight and its lipophilic nature makes it a good candidate for single-dose, slow-release formulations that can be used to achieve longer term benefits and further reduce costs [23, 25], once development is completed. Once consensus is reached on candidate doses and formulations, packaging discussions should start early as they can greatly influence compliance, costs and programmatic suitability [66].
(ii) What financing discussions are needed?
There are important data gaps on what the cost of goods of ivermectin at scale for malaria would be. The upper boundary should be the calculated U.S. donation value for NTDs of US$ 1.5 per 3 mg tablet 1.50 [63], the real price however should be negotiated. The economic benefits of ivermectin distribution for onchocerciasis are partly based on a donated drug. The value of this donated drug may surpass the operational budgets of the control programmes and the economic benefits expected from them for the next 20 years [64]. Given the higher burden and economic costs related to malaria and expected price negotiations, this balance might be more positive, especially in the context of elimination.
An important economic discussion would be the possibility for any new ivermectin-based tool to be financed by the GFATM in case it is included in a country plan, and recommended by WHO. WHO prequalification of the new indication for malaria or any new formulation will be a prerequisite for policy recommendation and thus GFATM financing.
Supply
(i) Is the manufacture process scalable?
Ivermectin is semi-synthetic derivate of a bacterial bio product [67]. The manufacturing process is technically scalable. As the global demand increased, its production has been enhanced and purified by a number of methods [68, 69]. The current global production is above 150 tons of active pharmaceutical ingredient per year (estimate from the Argentinian Chamber of Veterinary Products, pers. comm.), most of it is for veterinary use. As guidance, only 2.24 tons per year are needed to treat 80 million people, the target of the Mectizan Donation Programme (assuming an average weight of 70 kg, at the 200 µg/kg-dose, twice a year); that is less than 1.5% of the current global production. Even a tenfold increase on the global demand for human use, due to its theoretical application in malaria control would represent less than 15% of the current production due to the co-endemicity of onchocerciasis and malaria in many regions. Note that malaria use would likely be phased in over time.
Here it is important to distinguish between the production of the API (which would be the main limiting step where there an increase in the global demand due to malaria use), and the manufacturer of final products. There are dozens of large scale API producers, mainly in china with some of them reporting an annual production above 50 tons (see Additional file 1). The API used for veterinary and human products can come from the same source but must fulfil different quality standards which might require additional purification steps. Although there are several hundred manufacturers of final product (see for example [70] for a list with more than 100 generic products and manufacturers only in India), the production output and technical capacity of these manufacturers to obtain WHO-PQ will play a key role on ultimately fulfilling the global demand.
(ii) Current and prospective global demand for NTDs
More than 200 million doses were donated for the control and eradication of onchocerciasis and lymphatic filariasis in 2015 [28]. The demand will vary according to the operational goals for onchocerciasis (control, elimination or eradication). One estimate is it could reach up to 2.63 billion treatments for the 2013–2045 period [71], but higher demand due to accelerated LF elimination with drug combo strategies can be expected [44, 47].
(iii) What would be the Go/No-Go criteria for the development of new formulations or novel dosing schemes?
The efficacy threshold is expected to be in direct relationship with the total dose and the area under the curve [15]. The safety and programmatic feasibility of schemes requiring high or multiple doses should be measured against the expected efficacy. Novel formulations could simplify the dosing schemes and increase compliance but would require R&D investment.