Effectiveness of vector control methods for the control of cutaneous and visceral leishmaniasis: A meta-review

Elimination of visceral leishmaniasis (VL) in Southeast Asia and global control of cutaneous leishmaniasis (CL) and VL are priorities of the World Health Organization (WHO). But is the existing evidence good enough for public health recommendations? This meta-review summarises the available and new evidence for vector control with the aims of establishing what is known about the value of vector control for the control of CL and VL, establishing gaps in knowledge, and particularly focusing on key recommendations for further scientific work. This meta-review follows the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) criteria, including (1) systematic reviews and meta-analyses (SRs/MAs) for (2) vector control methods and strategies and (3) for the control of CL and/or VL. Nine SRs/MAs were included, with different research questions and inclusion/exclusion criteria. The methods analysed for vector control can be broadly classified into (1) indoor residual spraying (IRS); (2) insecticide-treated nets (ITNs; including insecticide-impregnated bednets); (3) insecticide-treated curtains (ITCs; including insecticide-treated house screening); (4) insecticide-treated bedsheets (ITSs) and insecticide-treated fabrics (ITFs; including insecticide-treated clothing) and (5) durable wall lining (treated with insecticides) and other environmental measures to protect the house; (6) control of the reservoir host; and (7) strengthening vector control operations through health education. The existing SRs/MAs include a large variation of different primary studies, even for the same specific research sub-question. Also, the SRs/MAs are outdated, using available information until earlier than 2018 only. Assessing the quality of the SRs/MAs, there is a considerable degree of variation. It is therefore very difficult to summarise the results of the available SRs/MAs, with contradictory results for both vector indices and—if available—human transmission data. Conclusions of this meta-review are that (1) existing SRs/MAs and their results make policy recommendations for evidence-based vector control difficult; (2) further work is needed to establish efficacy and community effectiveness of key vector control methods with specific SRs and MAs (3) including vector and human transmission parameters; and (4) attempting to conclude with recommendations in different transmission scenarios.

Introduction Elimination of visceral leishmaniasis (VL) in Southeast Asia and global control of cutaneous leishmaniasis (CL) and VL are priorities of the World Health Organizations' Department of Control of Neglected Tropical Diseases (WHO NTD). From the 5 elimination strategies for VL, vector control is one of the core ones against both forms of leishmaniasis [1]. However, the effectiveness of vector control for reduction of transmission of both VL and CL is repeatedly under discussion. This has been assessed in numerous reports, including WHO reports, e.g., WHO (2010) [2], but also in comprehensive reviews.
In this sense, Picado and colleagues (2012) [3] presented a review of studies for VL published in the period 2005 to 2010 on the efficacy of different tools to control Phlebotomus argentipes. The review indicates that "the current indoor residual spraying (IRS) and novel vector control methods mainly insecticide-treated nets (ITN) have low effectiveness for several reasons. Efforts to improve quality of IRS operations and further research on alternative and integrated vector control methods need to be promoted to reach the VL elimination target by 2015." However, the review stops short of recommending particular interventions and/or combinations of interventions. In a recent field trial using cluster randomised design, partially unexplored options were also tested for sand fly control [4] in order to strengthen the campaign for elimination, the deadline of which was extended from 2015 to 2020 [5]-no definite recommendations are available at this stage.
Similarly, Kassi and colleagues (2008) [6] conclude in a review for CL that ". . . it can be seen that many effective interventions exist. Considering the multitude of factors involved in transmission of CL and the various effective control measures tried and tested by investigators, an interdisciplinary approach involving more than one of the above interventions would make sense." As for VL, the review is not highlighting a particular vector control method or combinations of interventions for CL as most effective and/or recommendable.
However, CL and particularly VL are major public health problems in many countries, WHO reports that "8 countries and territories are endemic for leishmaniasis in 2018. This includes 68 countries that are endemic for both VL and CL, 9 countries that are endemic for VL only and 21 countries that are endemic for CL only" [7]. For VL, the incidence has become very low in Southeast Asian countries, which is related to the Regional Visceral Leishmaniasis Elimination Initiative [8], but in the post-elimination era, it will be a major challenge to sustain elimination, and investments are required for the remaining areas of scientific uncertainty [9]. With this in mind, WHO is making efforts to implement policies and strategies to reduce the disease burden of CL and VL-but is the existing evidence good enough for public health recommendations?
This meta-review follows up on these efforts and aims at facilitating further decision-making processes in 2020, with a process of systematically summarising the available and new evidence for vector control for CL and VL, focusing on existing systematic reviews and metaanalyses (SRs/MAs) and following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) criteria [10]. The key objective is to develop a meta-review including all quality SRs/MAs for the control of with the aims of establishing what is known about the value of vector control for the control of CL and VL, establishing gaps in knowledge, and particularly focusing on key recommendations for further scientific work.

Search strategy, databases, and search terms
This meta-review follows the PRISMA criteria [10]. Literature searches and analysis were developed and carried out through May 31, 2020. Data were extracted from Cochrane Database for Systematic Reviews, Lilacs, PubMed, Wholis, and Google Scholar (the latter was screened for the first 200 hits, since the database is sorted by relevance, 200 hits was established as a suitable number, with no relevant hits encountered towards the end of the search). All full-text assessed articles have been manually searched for additional articles in the reference list. As for grey literature, relevant global guidelines for CL and VL have been identified on WHO Iris: Guidelines were included if published after 2007, when the WHO guidelines committee assumed their work [11] and assuming that global guidelines are using SRs for their recommendations. Relevant SRs have been included.
The inclusion criteria were (1) SRs/MAs for (2) vector control methods and strategies and (3) for the control of CL and/or VL.
Excluded were SRs for clinical picture and treatment, diagnosis, and surveillance. No restrictions were applied regarding year of publication, geographic area, or publication language; however, the included SRs/MAs needed to be indexed in English, as the search was performed in English. 1. Scientific method of the study: "Systematic review" and/or "Meta-analysis" 2. Disease: "Leishmaniasis" (MeSH term, where applicable), considering that this will always include "Cutaneous leishmaniasis" and "Visceral leishmaniasis." Searches were performed with combinations of the first 2 categories. The search was intentionally kept very broad, and a third category was used with the results obtained from the searches: Articles were retained if they were mentioning the 3. Intervention: "Vector control", and its variations of different methods.
Selection of relevant articles was based on study title and abstract and, where needed, by full-text assessment (see Fig 1). The selection and categories of the studies included were initially based on 3 broad categories:

Data extraction and analysis
The above described data extraction matrix was further developed by including relevant variables such as type of disease, vector control forms, and regions included. All relevant data were extracted into the developed matrix. Evidence tables and recommendations for vector control have been developed and graded for level of evidence as well as strength of effectiveness followed by a gap analysis for further SRs/MAs. The gap analysis included geographical criteria, key vector control interventions, elapsed time since last searches, and quality criteria.

Descriptive results
Results of searches. A total of 2,971 initial hits were retrieved on the 6 included databases. After assessing by title and abstract, 177 articles were further screened. Moreover, 129 articles were fully assessed, removing 48 duplicates. In total, 29 SRs/MAs remained, of which only 9 dealt with vector control in the broader sense (1. targeting directly the vector; 2. targeting the vector though interventions using the reservoir host; and 3. education related to vector control). All other SRs were concerned about nonvector control issues, e.g., epidemiology of infections, both in humans (6) and in reservoir host (6), nonvector reservoir host control (2) socioeconomic status of cases (3), characteristics of the vector (1); and others (2) (see Fig 1). A total of 4 SRs/MAs were retrieved on Google Scholar, 3 on PubMed, 1 each on Lilacs, and WHO Iris. In the case of WHO Iris, this database should have WHO guidelines included, and these should have SRs and MAs included for the development of guidelines (see Table 1).
Time and geographical clustering of SRs/MAs. All included SRs/MAs were published between 2010 and 2018. All SRs/MAs were published by different groups of authors and focusing mostly on the global situation of both CL and VL.   [12] and de Sousa and colleagues (2015) [13] focused on Latin America only.
Methods applied by the SRs/MAs. All SRs followed the PRISMA guidelines [10], only 1 article was not labelled an SR, but followed predefined criteria, as prescribed by PRISMA [14]. Only 1 article included an MA [15], not surprisingly, since the studies included in the SRs are very heterogenous, with different interventions and outcome measures.
Focus of the SRs/MAs. The SRs/MAs included focused on CL and VL from a different viewpoint, assessing vector control from topics that range solely on CL and VL to reservoir host control for VL and health education (see Table 2).
This overview shows that comparability between the existing SRs/MAs is difficult, since most of the SRs and MAs have a different focus.
Also, it shows that the initially conceptualised categories for classification-as specified in the Methods section (1. focus on CL vector control, with 1 single intervention only; 2. VL vector control, with 1 single intervention only; and 3. combinations of vector control interventions, either with a focus on VL or CL, or combined) cannot be applied, since the research questions differ (see Table 3).

Type of studies included in the SRs/MAs
The SRs/MAs included different types of primary studies, if categorising with the hierarchy of evidence (https://consumers.cochrane.org/levels-evidence). Some SRs/MAs used randomised controlled trials (RCTs) and/or cluster randomised controlled trials (cRCTs) only [14,[16][17][18], and others used studies with lower level of evidence as well [12,13,15,19,20]. There was some overlap of the included studies, particularly for the RCTs (see Table 3). However, Table 3 clearly indicates that the different SRs/MAs used substantially different studies, with a total of 88 articles included in 9 SRs/MAs.

Type of outcome measures
Human disease parameters were mostly not captured in the studies included in the SRs/MAs. However, one SR focused on studies with human disease parameters [14], and another SR [15] assessed studies that reported, beside entomological parameters, clinical data focusing on CL and VL incidence in intervention and control groups.
Entomological parameters applied varied considerably. In some studies, the entomological parameters applied were of vector density or mean number of phlebotomine sand flies (per house per night) between intervention and control groups either using IRS, ITNs, or insecticide-treated curtains (ITCs) [12,[15][16][17][18]20]. Other entomological parameters were identified such as sand fly landing rates and vector mortality rates [12,20].
SRs that included studies on reservoir control applied canine VL incidence and prevalence parameters [12,16,19].

Quality analysis
Quality analysis showed that, when measuring against the PRISMA criteria [10], several SRs/ MAs meet most or all quality criteria, above 24 or all of 27 criteria [15,18,19]. Further three SRs meet between 20 and 23 criteria [13,17,20]. And 3 SRs met only 16 or below criteria.

VL IRS on VL:
The studies that evaluated IRS and fogging around the houses, a significant decrease of sand fly abundance was observed, with a residual effect of indoor spraying, that usual lasted 3 months. The specific construction style of the houses lowered the external validity of the study.

ITNS on VL:
One study reported the barrier effect, the 24-hour mortality rate and the human landing rates of Lutzomiya longipalpis in households using deltamethrinimpregnated bednets compared others using untreated bednets. The study described a 39% increase in barrier capacity of the impregnated bednets, 80% reduction in sand fly landing rates on humans and 98% increase in the 24-hour sand fly mortality rates. Despite the limitations (a small number of observations, a short period of exposure and without the measure of the residual effect), this intervention should be explored further because it could bring an additional benefit in areas where malaria is also endemic.

Multiple interventions on VL:
Studies that reported the impact of combined intervention, usual consisted in the combination of human VL case treatment, culling of seropositive dogs and insecticide spraying, one study reported the disappearance of human symptomatic cases after 15 years of application of this strategy. Another observed that the intensity of the application of control measures correlated with human VL incidence, the coverage of canine surveys, the number of canine surveyed, and the number of buildings submitted to insecticide spraying. Also, in combined interventions, the results of one study indicated a positive effect of canine removal on incidence of leishmanial infection in men but surprisingly, the combination of dog culling plus outdoor spraying of peridomestic animal shelters failed to demonstrate any effect. Another study reported that although a lower incidence was observed in the groups submitted to combined interventions and that reduction was more intense after 2 years, the study failed to detect statistically significant differences. Another study reported the effect of application of a 65% permethrin spoton formulation on canine VL infection and sand fly abundance. A decrease of canine VL prevalence was observed in the intervention area compared with increased prevalence in the control area. No effect was observed on sand fly population.

Reservoir control on VL:
The studies that reported dog screening and culling, usually screened by ELISA and IFAT and reported a reduction of seroconversion rate in dogs but without a significative difference. It was also observed that by using method ELISA, the reduction of canine seroprevalence was higher,

ITNS on CL:
The remaining 5 net studies all report a statistically significant reduction of cases of CL by using ITNs. A general criticism regarding these five is that, except one, the others all used selfreporting as the method of diagnosis which is at high risk of reporting bias.

ITCS on CL:
One

Analysis of results
Crosscutting results. The methods analysed in the SRs/MAs for controlling vectors to prevent transmission of the parasite causing either CL or VL have been found to be similar. The same method or technique has been named slightly differently by groups of researchers in different settings. The methods included IRS, the use of ITNs (including insecticide-impregnated bednets), ITCs (including insecticide-treated house screening), insecticide-treated bedsheets (ITSs), and insecticide-treated fabrics (ITFs) (including insecticide-treated clothing) and durable wall lining (treated with insecticides) and other measures to protect houses. In addition, environmental modifications (EVMs), control of the reservoir host, and strengthening vector control operations through health education were evaluated. Many of the studies included in the SRs/MAs evaluated several interventions at the same time.
This main results section presents the results therefore as  [17] and   [14] reported mixed results: Only 2 cRCTs from South America were included by González and colleagues (2015), with substantial reductions of vectors, without evidence for long-term effects. An additional cRCT from Afghanistan reported a reduction of human CL cases over 15 months.   [14], however, included 4 studies, limiting the studies to those presenting human measurements, and presenting mixed results, with 2 higher quality studies showing no efficacy of IRS, whereas 2 lower quality studies showed some efficacy in reduction of cases.

ITSs and ITFs (including insecticide treated clothing)
Asilian  analysed 3 studies testing ITNs versus no intervention or untreated nets on CL. One of 3 cRCTs in Iran evaluated the effect of ITNs on vector density, with a statistically significant reduction, however not measuring duration of effect. Two cRCTs from Afghanistan and Iran measured incidence of CL, with a marked reduction of CL cases over 15 months. Wilson and colleagues (2014) [15] with its focus on ITNs, included 6 studies evaluating the efficacy of ITNs against CL, allowing also for an MA. Random effect MA indicated a partial effect of 77%. However, studies assessing the efficacy of ITNs reported mixed results in terms of effect on vector density, ranging from a relative increase of 49% to a relative reduction of 96%. Stockdale and Newton (2013) [14] reported on 5 studies with a statistically significant reduction of cases of CL, however underlining that for 4 of the included studies, self-reporting was used for case definition.
ITCs. Also, ITCs were evaluated by 3 SRs/MAs [14,15,17]. Reduction of vector indices varied; however, reduction of human transmission indices was positive: González and colleagues (2015) [17] included 1 cRCT, although no statistically significant differences in the mean number of vectors were reported, the incidence of clinical cases of CL were 0/1,351 (0%) in the intervention group and 142/1,587 (9%) in the control group. Wilson and colleagues (2014) [15] reported a high percentage reduction in vector density of 54%, 87%, and 98% for the included studies; however, the 98% reduction was observed in a study that was deemed to be of low quality.   [14] reported only for one of 4 studies using ITCs a statistically significant decrease of CL cases.
ITSs and ITFs. ITSs/ITFs were again evaluated by 3 SRs/MAs [14,15,17], with overall positive effects on human transmission. González and colleagues (2015) [17] analysed ITS (bedsheets) versus no intervention on CL with a cRCT from Afghanistan, with substantially fewer cases over 15 months in the intervention households across all age groups. Two RCTs evaluated the effect of impregnating soldiers' uniforms with permethrin on the incidence of CL. The trials were small and underpowered to confidently detect or exclude effects. However, in one study, the incidence in the control group was 18/143 over 12 weeks (12%), and just 4/143 (3%) in soldiers with impregnated uniforms (risk ratio (RR) 0.22, 95% CI 0.08 to 0.64). Stockdale and Newton (2013) [14] included 4 studies with ITFs, and three reported a statistically significant decrease in numbers of human cases of CL between intervention and control groups.
EVM. Only Stockdale and Newton (2013) [14] analysed EVM against CL; none of the outcomes of the included studies were measured against human transmission indicators.
Results for VL only. Vector control methods analysed in the articles were essentially the same as for CL, and the same categories were used for the analysis.
No trials evaluated the effects of IRS on VL incidence, as stated by González and colleagues (2015) [17]. However, one trial assessed the effect on seroconversion in a VL endemic area in Brazil and found no statistically significant difference in seroconversion over 18 months postintervention.   [12] however reported evaluating IRS and fogging around the houses, a significant decrease of sand fly abundance, with a residual effect of indoor spraying, which usually lasted 3 months which was influenced mainly by house construction style.   [14] included 2 studies with no trend of effect: one evaluated VL in children under 12 years in 3 control and 3 intervention areas, with no difference in infections rates in all the evaluated areas. The other study used random allocation of 4 intervened areas in order to study VL in the population living in each area; the study observed no difference between the intervention and control areas.
ITNs. ITNs were analysed by González and colleagues (2015) [17],   [12], and Wilson and colleagues (2014) [15]. González and colleagues (2015) [17] reported on ITNs versus no intervention or untreated nets on VL. Two of the 3 included cRCTs in Asia evaluated the effect of ITNs on vector density: In Bangladesh, there was a substantial reduction in vector density in the ITN areas for 12 months post-intervention, but in a multicentre trial in Asia, the overall difference between intervention and control sites was not statistically significant. One additional cRCT in India reported a statistically significant reduction in male P. argentipes in areas with ITNs compared to untreated nets, but no difference in female P. argentipes or other vectors.
One cRCT evaluated the effect of ITNs on VL in India and Nepal. The overall risk of VL during the 30 months follow-up was 37/9,829 (0.38%) in the intervention group and 40/9,981 (0.40%) in the control group. In the same trial, there was also no significant difference in the risk of seroconversion in those who had negative results at baseline [17].
Wilson and colleagues (2014) [15] also compared the efficacy of ITNs against VL, including 3 studies. Similarly, one study did not show a significant effect on incident Leishmania donovani infections or incident cases of VL. However, in India and Nepal, the same study did appear to show an effect on vector density with a relative reduction in the mean number of female P. argentipes. Two studies conducted in Sudan and Bangladesh, India, and Nepal demonstrated a 100% and 35% reduction in vector density, respectively.
Romero and Boelaert (2010) [12] only included 1 study, with ITNs and VL: The study described a 39% increase in barrier capacity of the deltamethrin-impregnated bednets, 80% reduction in sand fly landing rates on humans, and 98% increase in the 24-hour sandfly mortality rates. The study had many limitations (a small number of observations, a short period of exposure, and without the measure of the residual effect).
Stockdale and Newton (2013) [14] reported as well on CL and VL in relation to ITN, with no clear effect reported (see the relevant section below).
ITCs. For ITCs and VL, no studies were included in the SRs/MAs. ITSs and ITFs. Only González and colleagues (2015) [17] included studies on ITSs: one cRCT in areas of Brazil with VL evaluated the effects of treated sheets near the chicken shed, with short-term reductions in geometric mean phlebotomine sand flies per trap after the intervention, which only differed statistically from control sheds at week 12 post-intervention.
EVM. González and colleagues (2015) [17] included 2 cRCTs, comparing EVM versus no intervention: Neither trial found evidence of statistically significant reductions in phlebotomine sand flies compared to no intervention up to 12 months follow-up.
EVM has been further analysed in the context of both CL and VL (see the relevant section below).
Results for CL and VL. The included SRs/MAs also compared studies with information on both CL and VL.
IRS. González and colleagues (2015) [17] included studies on IRS versus no intervention, with 2 included cRCTs, reporting substantial reductions in vectors at the intervention sites. Calderon-Anyosa and colleagues (2018) [20] reported on 8 intervention studies describing housing characteristics, sand fly density captured by light traps (5/8), and sand fly mortality by wall bioassay (3/8). Of the 8 studies, 4 evaluated the effect of insecticidal spray on different wall materials. One evaluated sand fly density and reported differences associated with housing quality and vector densities; in some cases, the number was higher even after insecticide thermal fogging.
From the remaining studies that evaluated wall bioassay mortality, one evaluated fogging on cement wall versus oil-painted wall, finding no significant differences in sand fly mortality at 7 or 125 days after fogging, whereas it was significantly higher in oil-painted wall at 69 days. The other study matched houses according to their structure and were randomly assigned to spray treatment or control, finding that sand fly mortality decreased progressively on wood and cement surfaces after 63 days compared with a more rapid decrease on mud and straw walls. The third study evaluated spray on the external and internal surfaces of 3 types of walls, finding that mortality rates were similar, whatever the type of wall, since the fourth month.
ITNs. ITNs were analysed again by Stockdale and Newton (2013) [14]: For human reservoir control, 7 studies using ITNs measured a human-specific outcome. Two studies used deltamethrin-impregnated bednets. Neither group reported any difference in cases of CL or VL between the treated nets and either untreated nets or existing intervention.
ITSs and ITFs. No studies on ITSs/ITFs were included. EVM. EVM was included by Horstick and Runge-Ranzinger (2018) [18], concluding that modifications to the structure of houses (e.g., wall plastering) had no impact on the control of vectors. However, protection of the house and its surroundings might affect the transmission of several diseases.
Calderon-Anyosa and colleagues (2018) [20] described housing characteristics and risk for presence of vectors and disease: Mud walls with cracks and holes, damp, and dark houses were risk factors for transmission of leishmaniasis. These characteristics create favourable conditions for sand fly breeding and resting as sand flies prefer humidity, warmth, and protection from sunlight during the day. A total of 18/23 studies found significant association between housing characteristics (e.g., walls, roof, floors, or windows) and leishmaniasis infection or sand fly density. Moreover, 16/18 studies found an association between leishmaniasis and wall type. A total of 15/16 studies found an association with clinical leishmaniasis: 5/15 with CL cases and 10/15 with VL cases. In addition, 4/8 intervention studies evaluated housing characteristics and home improvement against sand fly density captured by light traps. One experimental study evaluated the characteristics of chicken sheds against sand fly densities and found a significantly higher number of sand flies in open sheds. The 3 remaining studies evaluated the effect of plastering and closing crevices against sand fly densities: One study found no significant difference in sand fly density, whereas the other two found a decrease in sand fly density after the intervention.
Results for controlling the reservoir host. One SR [19] focused exclusively on the control of the reservoir host assessing the following studies: the use of insecticide-treated dog collars (4 studies including non-RCTs and a matched-cluster RCT) and a combination of dog collars and spot-on insecticides treatments (1 non-RCT). There was a statistically significant protective effect of collars, measured by the overall proportion of dogs infected with Leishmania infantum. Use of spot-on insecticides treatments (3 studies, including non-RCTs and RCTs): there was a statistically significant protective effect for the overall proportion of dogs infected with L. infantum. Three studies (all RCTs) evaluated prophylactic medications: 2 studies for domperidone liquid solution and 1 study for allopurinol capsules. There was a statistically significant protective effect for prophylactic medication with domperidone for the overall proportion of dogs infected with L. infantum, but not for allopurinol.
Results for control through education. Furthermore, education was the focus of one SR, in South America [13]. Five studies evaluated the influence of educational material showing an improvement or reinforcing the importance of educational activities to improve access to knowledge by the population.
One study showed the actions of local social representations as effective instruments of information and prevention of leishmaniasis. Also, including guidance to the public on the use of screens and mosquito nets impregnated with insecticide was evaluated. One study found that although 94% of participants knew leishmaniasis as a skin disease, with ulcers or blemishes, only 35% associated the disease with the bite of an infected "mosquito," and only 10% used the appropriate drug treatment.

Education
It is difficult to assess the effect of education on transmission with the included studies; however, a relation between knowledge of disease transmission and protective behaviour is assumed.

Discussion of key results
Vector control for CL and VL has been targeted using different tools in different settings at different times, as shown with the multiple studies included in the SRs/MAs. The methods used for controlling vectors to prevent transmission of the parasite causing either CL or VL have been found to be similar, including IRS, ITNs-mostly insecticide-impregnated bednets, ITCs, ITSs and ITFs, and durable wall lining (treated with insecticides) and other environmental measures to protect houses.
The key results are very difficult to interpret with a lot of contradictory messages, and it is difficult to describe a clear trend. The SRs/MAs that we included in our meta-review identified gaps in control measures, especially in relation to their evaluation rather than implementation. Given the SRs/MAs could only include the original research findings available up to the time point of their literature search, these reviews could also omit more recent investigations on CL and VL vector control. This is more likely to happen especially for the recent past when the control measures were strengthened with more technical cooperation between the countries, along with revised VL elimination targets for South Asia and the region [5]. Also, new vector control techniques could have been missed in the SRs/MAs. For other regions, because of the zoonotic nature of the disease, especially for VL, the reviews and investigations were split based on their focus on either humans or animals. This made the gap analysis from our meta-review constrained as we had less focus on vector control in animal reservoirs.
The gaps in the research findings identified through our meta-review have been discussed below under themes focusing on diseases as well as vector control methods, techniques, or tools. Also, there are overarching issues around control of the vectors irrespective of what disease they are causing-CL or VL. We combine below those overarching issues and individual disease specific considerations while discussing gaps in findings.
In majority of the studies included in different SRs/MAs, human disease was not considered as the primary outcome of interest. This clearly left a gap in understanding the associations between different vector control measures and their eventual impact in reducing burden of CL or VL in humans. Moreover, in the studies which used human disease as their outcome, case identification often relied on clinical symptoms, patient reporting, antibody detection, and clinical cure. Parasite detection through their visualisation was rarely performed, which could have resulted in misclassifications of disease condition, especially for VL which could mimic other endemic conditions in the study areas and regions [21,22]. Depending on the

Control of reservoir host
Insecticide-treated collars, spot-on insecticides, and medication with domperidone seem to have positive effects on reservoir host infection. In addition, dog culling as a control intervention was found to be consistently ineffective.
nature of bias, this could result in underestimation or overestimation of the associations between vector control measures and the occurrence of the disease. Some studies also lacked adequate power to detect a true association. All these could have affected the internal validity of the studies performed by different groups of researchers in different settings and regions. Generalisability of the study findings to all settings was also difficult because of this as well as due to the zoonotic and anthroponotic divide of the nature of VL by regions.
Moreover, research methodological variations (interventional versus observational) and weaknesses also made the studies less similar and difficult to summary effect measure estimation. The methodological weaknesses identified by the meta-review include issues around randomisation, blinding, inadequate sample size, participant adherence to the interventions, varying follow-up period, lack of adjustment of possible confounders through multivariable analysis, lack of adjustment of clustering, and poor description of trial designs. This is also reflected in the relative variation of the quality assessment of the included SRs/MAs. Implementation of intervention was also found problematic, for example with IRS not occurring at the same time for all study areas under the intervention arm with equal frequencies and different insecticides were used in different studies, especially in South Asia. Variation in insecticide susceptibility of the sand fly vector could also be problematic [23]. Also, control groups were poorly defined in some studies, whereas some studies had intervention and control arms not comparable due to their varying background VL prevalence.
The vector nature including species and their preferences for hosts [24] as well as resting and feeding indoor (endophagic and endophilic) or outdoor (exophagic and exophilic) needed to be accounted for in a more systematic way to better understand the impact of different interventions. The flying pattern of sand flies could also have been investigated more, especially when looking at the association between house structure and leishmaniasis. Data on intervention studies on house structure and leishmaniasis were also scarce, although more recent observational data have become available [25].
The studies considered in different SRs/MAs were mostly done in controlled environments without takings contexts into account. Further studies were needed to assess the implementation or operational aspect of vector control measures [9]. This might have been done in the recent past, especially in South Asia when they were reaching closer to the elimination target. But given no SR was conducted recently with a single focus on leishmaniasis, those studies might have been missed. Also, since no SR/MA assessed single vector control method only, reviewing each single vector control method including the new ones is warranted to identify more robust evidence on their efficacy and community effectiveness [26].

Discussion of level of evidence
Policy recommendations should be evidence based [11] and following a systematic approach of weighing and grading available evidence, including a process of expert consensus.
In the context of CL and VL, there is a wealth of available studies, primary studies, and including summary evidence, one of the key results of this meta-review. The existing SRs/MAs include a large variation of different studies, related to the different research question of each individual SR/MA, and the different inclusion/exclusion criteria.
However, as presented in the gap analysis of the results section, it is very difficult to summarise the results of the available SRs/MAs, with the shortcomings described above and it is difficult to recommend with the currently existing SRs/MAs particular vector control methods, or combinations of vector control methods.
A process should be initialised to systematically assess all available evidence for efficacy and community effectiveness of vector control in the context of both CL and VL.
One of the options, mentioned above, would be to assess each single vector control method with a specific SR, and if possible MA. This concept has the advantage to include more studies, on different levels of hierarchy, and to assess-if the information is available-different levels of transmission scenarios.

EVMs.
Technically, it is recommendable to assess these methods for the 2 diseases separately. Further specific SRs could be 6. vector control methods of the reservoir host in the context of VL; 7. strengthening vector control operations through health education; and 8. implementation of vector control programmes taking local and regional contexts into account.
With this process, it is expected that a better policy recommendation can be formulated, following a discussion of the results of the SRs/MAs to be developed, in expert consensus.

Conclusions
This meta-review has answered 3 key objectives: 1. Establishing what is known about the value of vector control for the control of CL and VL: Unfortunately, a clear trend for efficacy and community effectiveness of the different vector control methods for CL and VL is difficult to assess by the existing SRs/MAs. This is mostly due to the different research questions and studies included in each SR/MA. Considering this fact, it is not easy to formulate evidence-based recommendations for vector control methods for CL and VL.
2. Establishing gaps in knowledge: Clearly, there is a wealth of primary studies available to assess vector control for both CL and VL. Further specific gaps for primary research may emerge through a more thorough analysis of each vector control methods. Additionally, there is a gap of systematic assessment of each vector control method.
3. Key recommendations for further scientific work: To improve policy recommendations, one of the key elements for further scientific work is a systematic analysis of each individual vector control methods, e.g., IRS, ITNs, ITCs, ITFs, and EVM, for CL and VL separately, including vector and human transmission parameters and attempting to conclude with recommendations in different transmission scenarios. It may be of interest to conduct SRs/ MAs on reservoir host control, education, and programme implementation in support of vector control operations.