The proposed study is a four-arm randomized control trial (RCT) in 48 villages in the Lakes region in Tanzania comparing the relative effectiveness of 4 vector control interventions for reducing malaria transmission and controlling vector populations in an area where An gambiae s.s is pyrethroid and carbamate resistant: 1/ a standard long lasting insecticidal net (LLIN), 2/ a LLIN which incorporates a piperonyl butoxide (PBO) synergist, 3/ a long lasting indoor residual spray (IRS) formulation used in conjunction with standard pyrethroid LLIN or 4/ the long lasting indoor residual spray (IRS) formulation used in conjunction with the LLIN which incorporates a PBO synergist.

The trial will provide epidemiological, entomological, economic and social evidence of impact, as the investigators shall be measuring the reductions in malaria prevalence and malaria transmission rates EIR, and changes in the frequency of resistance, mosquito species ratios and economic cost effectiveness. The proposed trial will demonstrate whether the novel LLIN and long lasting IRS formulation will be more effective for controlling An.gambiae s.s. and reducing malaria prevalence than current practice with the conventional LLIN. There is great interest in conducting this trial. Alternative vector control products are limited and most new insecticides are not suitable for use on LLINs or as IRS.

JUSTIFICATION OF THE STUDY The district of Muleba has been subjected to IRS with pyrethroid (lambdacyhalothrin) every year for 5 years under the President Malaria Initiative (PMI) programme. The malaria vector Anopheles gambiae s.s. is predominant in Muleba region and shows high level resistance to pyrethroid insecticides [14]. In World Health Organisation (WHO) resistance tests with the pyrethroid used for IRS, An gambiae s.s. showed levels of mortality ranging from 8% to 40%, indicating a high frequency of resistance across the entire district. Mortality to permethrin, the pyrethroid used in the LLIN distributed during the UCC (Olyset Net) was only 22%. Entomological studies by the present authors showed that the pyrethroid resistance mechanism kdr is almost 100% in the Muleba An.gambiae s.s. population and elevated levels of mixed function oxidases and non-specific esterases are involved in the detoxification of pyrethroids. The carbamate bendiocarb was selected by PMI as the new insecticide of choice for IRS in the region. Following two rounds of bendiocarb IRS in January and May 2012, resistance is fast emerging. WHO resistance tests with bendiocarb showed mosquito survival rates rising from 18% in November 2011 to 23% in May 2012 and to 69% in November 2012 [14]. The only insecticide to which An.gambiae shows full susceptibility (100% mortality) is the organophosphate pirimiphos methyl, the proposed new IRS intervention.

No malaria control trial with either Olyset Plus or Actellic CS is currently being proposed elsewhere. Olyset Plus seems ideally suited - judging by the CDC bioassay test with permethrin and the synergist PBO and WHO cone bioassay tests with Olyset Plus - to provide protection against the local pyrethroid resistant An gambiae. IRS with Actellic CS seems ideally suited to provide long term transmission control. A trial of both interventions is essential to restore confidence in vector control. By itself Olyset Plus may be sufficient to provide personal protection if used every night but it may not be sufficient to provide transmission control in the entire community. An effective IRS is probably the more efficient way to reduce transmission quickly but IRS campaigns cannot be sustained year after year because of the expense, nor is this desirable because of the risk of resistance. The combination of LLIN and IRS may be required to drive transmission down to low levels whereafter the Olyset Plus should be sufficient to keep prevalence and transmission low.

2 RESEARCH PLAN 2.1 Main Objective The main objective of this research is to determine the relative effectiveness of 4 vector control interventions for reducing malaria transmission and controlling vector populations in an area where An gambiae s.s is pyrethroid and carbamate resistant: 1/ a standard long lasting insecticidal net (LLIN), 2/ Long lasting insecticidal nets that incorporate the synergist a PBO (Olyset Plus), 3/ Long lasting formulation of the OP insecticide pirimiphos-methyl (Actellic CS) for indoor residual spraying used in conjunction with standard pyrethroid LLIN or 4/ the long lasting indoor residual spray (IRS) formulation used in conjunction with the LLIN which incorporates a PBO synergist.

2.2 Specific Objectives

• To determine whether malaria transmitted by pyrethroid resistant mosquitoes can be controlled using the latest innovative preventive tools, Olyset Plus, Actellic CS IRS and Combination of both products.
• To demonstrate by means of a community randomised trial in NW Tanzania the protective efficacy of each of these interventions on the outcomes measures
• To define a strategy for using and establish in what situations it is possible to deploy just one of the interventions, both interventions, and when to switch from one type of intervention to another.
• To determine the cost-effectiveness of the interventions alone or in combination to facilitate decisions on resource allocation.
• To demonstrate whether combination interventions can serve as a model insecticide resistance management strategy to delay the future selection of resistance.
• To provide international malaria control agencies with evidence to enable them to revise or redefine their malaria control strategy in the face of growing resistance and maintain focus on malaria elimination.

2.4 Interventions to be trialed Olyset plus (Sumitomo Chemicals) is a LLIN containing the pyrethroid permethrin and the synergist piperonyl butoxide PBO as the active ingredients (AI).Nets combining insecticide plus PBO are able to control mosquitoes whose resistance is based on oxidative metabolism. Olyset net is a long established LLIN that contains permethrin but no synergist. Both LLINs are approved by WHOPES and registered in Tanzania. In experimental hut trials in West Africa, Olyset Plus killed twice as many resistant mosquitoes and reduced biting rates by one third as compared to Olyset. On exposure to the two types of LLIN for 3 minutes none of the An. gambiae mosquitoes collected from Muleba was killed by the conventional Olyset but all were killed by Olyset Plus. This indicates the type of resistance occurring in Muleba can be synergised by the PBO on Olyset Plus. This new LLIN has the potential to restore control of malaria transmitted by resistant mosquito populations.

The proposed IRS intervention, Actellic CS (Syngenta), is a microcapsule formulation of the organophosphate insecticide, pirimiphos-methyl. This product has been the subject of a 3 year public-private partnership between the company, LSHTM, the Bill and Melinda Gate (BMG) Foundation and the Innovative Vector Control Consortium (IVCC). Microencapsulation greatly extends the compound's residual life when applied to interior walls and ceilings. Parallel trials in Benin against pyrethroid-resistant mosquitoes killed more than 80% of An. gambiae for up to 10 months, which is an unprecedented level of effectiveness. Other products tested at the same time included bendiocarb, the carbamate used in our last cluster randomised trial in Muleba, and lambdacyhalothrin, the pyrethroid used by PMI for IRS in Muleba during the previous 5 years. Each of these was vastly inferior or short lived (lasting 1-2 months), making pirimiphos-methyl CS the best candidate to replace carbamate or pyrethroid IRS.

Combined LLIN and IRS. In practice the IRS will not be applied alone but either with a conventional LLIN or LLIN with PBO synergist. An example of the potential impact of a conventional LLIN and P-methyl CS combination intervention was shown in an experimental hut trial in Burkina Faso [15]. In this country with moderate pyrethroid resistance, the pyrethoid LLIN alone killed approximately 50% of the mosquitoes and reduced biting (blood-feeding) from 80% to 24%. The P-methyl IRS alone killed 100% of mosquitoes but not before 40% had succeeded in blood feeding. It was the combination of LLIN and sprayed wall which had the largest impact, killing 100% of mosquitoes and preventing biting by 93%, thus illustrating the potential benefit of the combination intervention.

Combined LLIN+synergist and IRS. This is anticipated to provide the greatest reduction in malaria transmission. For the second year of the intervention the IRS monitoring will continue to observe whether the LLIN+synergist can sustain the reduction in transmission by itself once transmission has been brought low by the combination.

2.5 Methods 2.5.1 Study Area The trial will be carried out in the district of Muleba (10, 70'S, 310, 60'E), on the western border of Lake Victoria, Tanzania. Muleba covers an area of 3,550 km2. The annual rainfall ranges from 850 mm to 1500 mm (Igabiro weather station 2006 to 2012).

The district has been subjected to blanket IRS with pyrethroid lambdacyhalothrin (ICON 10CS, Syngenta, Basel, Switzerland) every year from 2006 to 2011 under PMI program. Targeted spraying with bendiocarb (400 mg/m2 of Ficam 80% wettable powder, Bayer) have been done in 2012 & 2013.

As part of the National Malaria Control strategy, net coverage has increased since 2005 through net distribution campaigns initially targeting pregnant women and children under five in 2009 (63% of households provided with nets) and then targeting the entire population in 2011. After the universal coverage campaign distribution ownership increase to 91% and ITN/LLIN usage in all residents had risen to 56%.

An evaluation by LSHTM in 2011-2012, supported by PMI, recorded an overall malaria prevalence of 23% (95% CI: 17-29) in children under 15 years. Based on the previous study mosquitoes peaked in April-May and Nov-December during the two rainy seasons.

The malaria vector Anopheles gambiae s.s. is predominant in Muleba region (80%) and shows high level resistance to pyrethroid insecticides and DDT (mortality no higher than 40%.). with Following two rounds of bendiocarb IRS in January and May 2012, resistance is fast emerging. WHO resistance tests with bendiocarb showed mosquito survival rates rising from 18% in November 2011 to 23% in May 2012 and to 69% in November 2012 (Protopopoff unpublished). The kdr-eastern variant was present in homozygous form in 97% of An.gambiae s.s. but was absent in An.arabiensis.

2.5.2 Study Design: The trial will be conducted over 30 months with 1 month of preparation, 9 months baseline before the intervention start, and 20 months of intervention year. IRS will be applied at the beginning of the intervention period in February-March 2015 before the increase of mosquito density indication the start of the long transmission period. The two types of LLIN shall be distributed according to cluster randomisation at the same period of the first round of IRS to ensure that at least 80% of the households will have one (or more) LLIN for every two people.

The proposed trial will be a four arms, single blinded, cluster randomized factorial design comparing

1. High coverage (>80% access) of conventional Olyset LLIN
2. High coverage (>80% access) of Olyset Plus LLIN
3. High coverage (>80% access) of conventional Olyset LLIN and high coverage IRS (>80% of the household sprayed) with pirimiphos methyl CS
4. High coverage (>80% access) of Olyset Plus LLIN and high coverage IRS (>80% of the household sprayed) with pirimiphos methyl CS Randomization of the 48 Muleba clusters into four arms in a factorial design to test for differences in effect between Olyset LLIN and Olyset Plus LLIN and to test for further effect due to IRS with pirimiphos methyl versus no IRS against a background of LLIN use.

The 2 x 2 factorial design is strong because it allows comparison of the 2 types of nets, of a long lasting IRS treatment versus no IRS, and of the two interventions together.

2.5.3 Sample Size Estimations The sample size is based on the principles cluster randomised trials [16] using mean prevalence of infection in each study arm as the primary outcome. It was assumed that prevalence of infection in the reference arm has a mean of 20% with a coefficient of variation between clusters of 0.3, based on recent surveys in the study area, and that 80% power and 5% significance would be required. Therefore, with 48 clusters (4 arms of 12 clusters each) the study would be able to detect a 28% lower prevalence (prevalence ratio 0.78) between the Olyset Plus relative to the standard Olyset LLINs arms and between IRS with pirimiphos methyl relative to no IRS (24 clusters per arm for each of the two comparisons), with a sample of 80 children per cluster. Similarly, a 40% lower prevalence (prevalence ratio 0.6) could be detected between any individual arm and another, for example Olyset Plus + IRS vs Olyset Plus (12 clusters each).

Inclusion Criteria:

For the Household

• Having a children from 6 months to 14 years old in the household
• Provide written consent For the children
• Having between 6 months to 14 years
• Permanent residence in a selected household

Exclusion Criteria:

• Children severely ill