Study design
Two brands of LLINs: Olyset® and DawaPlus® 2.0 (Tsara® soft) that were stored for more than five years under recommended conditions were evaluated. The study was conducted in two stages. First, through a randomized double blinded, bio-efficacy evaluation of LLINs using standard WHO assays [8]. This was followed by partially randomized double blinded semi-field tests to compare the bio-efficacy of long stored LLINs against the new LLINs of the same brand using the Ifakara Ambient Chamber tests (I-ACT) [22]. Untreated Safi Net was used as a negative control in all tests to monitor the quality of the experiment.
Test facility
The experiments were performed at the Vector Control Product Testing Unit (VCPTU) of the Ifakara Health Institute located in Bagamoyo, Tanzania. (http://ihi.or.tz/static/media/Vector-Control-Product-Testing.e31c173f.pdf).
Test nets
Olyset® is a high-density mono-filament polyethylene LLIN, incorporated with 20 g/kg (±3 g/kg), 2% w/w of permethrin (corresponding to 1000 mg/m2). Olyset® is manufactured by A to Z Textile Mills Ltd, Arusha, Tanzania. Tsara soft formally known as DawaPlus® 2.0 is manufactured by NRS Moon Netting FZE [9]. Tsara® soft is a deltamethrin-coated LLIN. The target dose of deltamethrin coated on a knitted multi-filament polyester fiber is 2.0 g/kg ± 25% with 100-denier yarn (corresponding to 80 mg/m2 deltamethrin). Untreated Safi Net is made of polyester fibres, manufactured by A to Z Textile Mills Ltd, Arusha, Tanzania. All nets were double sized and coded by an independent IHI staff, to allow blinding of investigators and participants. All test nets have WHO-PQ listing [9].
Net Storage conditions
Olyset® and DawaPlus® 2.0 LLINs were stored for approximately five years in the Ifakara Health Institute (IHI-Bagamoyo) storage facility and are denoted in this study as long storage (LS) nets. All nets were received directly under similar conditions from the manufacturer and were manufactured shortly before shipping for the purpose of product evaluation. LS Olyset® LLINs with batch number L2605 were manufactured in May, 2013 and logged into the IHI-Bagamoyo storage facility on 4th June, 2013. LS DawaPlus® 2.0 LLINs were regular production manufactured in November, 2013 and were logged into the IHI-Bagamoyo storage facility on 4th December, 2013.
The new Olyset® LLINs were manufactured in 2017 with batch number 7X15BZS, and were logged into the IHI-Bagamoyo storage facility on 22nd December, 2018. The new DawaPlus® 2.0 LLINs were test series manufactured in March, 2018, with batch number 18SPL005, were shipped from the manufacturer on 15th May, 2018 and logged into the IHI-Bagamoyo storage facility on 1st June, 2018. All nets were stored and maintained at an average temperature of 29°C [25°C to 33.4°C] and 40% to 100% relative humidity (RH) in the IHI-Bagamoyo storage facility. Temperature was recorded and logged each afternoon at 14:00 hours which coincides with peak temperatures.
IHI-Bagamoyo LLINs storage facility
The LLIN storage used in this study was a shipping container that uses only passive cooling for the majority of the year. The container is raised above the ground and is situated under a second shade roof to reduce the radiant transfer of heat. It is installed with aluminium heat reflecting foil between the rafter and iron sheet of the storage for efficient cooling, also equipped with ventilation gaps (similar to the eaves of African houses) to allow air movement through the store. Electric ceiling fans are used only at the hottest times of the year irrespective of the temperature (Fig. 1).
The experiments were conducted from 25th January 2019 to July 2019. Olyset® LLINs had been stored for 5 years and 2 months while DawaPlus® 2.0 LLINs had been stored for 4 years and 8 months at the time of WHO cone assays and tunnel testing. Olyset® LLINs had been stored for 5 years and 8 months while DawaPlus® 2.0 LLINs had been stored for 5 years and 2 months at the time of I-ACT testing.
LLINs preparation and washing procedures for WHO assays
Eight LLINs (four nets of each brand) were randomly selected from the same product batch. LLINs were coded, cut into pieces (25cm x 25cm) and washed one, three, five, ten, fifteen and twenty times following WHO standard procedures for laboratory testing (phase I) [8] The washing interval of one day was used based on the reported regeneration time for both products [23].
LLINs preparation and washing procedures for I-ACT assays
Eight LLINs (two old and two new DawaPlus® 2.0; and two old and two new Olyset®) and two untreated nets were randomly selected from their product batches and coded. Two LLINs of each brand were washed twenty times as per WHO small-scale field trials (phase II) washing procedures, used as a standard procedure to simulate aging of nets under user conditions [8], while the other two were unwashed. All washed, unwashed and un-treated nets were deliberately holed (4cm by 4cm) six times, with one hole on each width and two holes on each length side, 75 cm from the top of the net as per WHO procedures [8]
Test systems
Mosquitoes used for the evaluations were Anopheles gambiae sensu stricto (s.s.) (Ifakara strain) fully susceptible to all classes of insecticides and Anopheles arabiensis (Kingani strain) strongly resistant to all pyrethroids including, deltamethrin and permethrin (<20% mortality with WHO discriminating doses, through metabolic P450 mechanism). In the WHO cone bioassays, nulliparous 3-5 days old female sugar-fed mosquitoes were used while in the tunnel test and I-ACT, nulliparous 5-8 days old female mosquitoes’ sugar-starved for eight hours were used. The VCPTU mosquito colonies are maintained at 27°C ± 5°C and 40% to 100% relative humidity with access to 10% sucrose ad libitum supplemented by membrane feeding using cow blood for the purposes of egg production following MR4 guidelines [24].
WHO assays
Cone bioassays were conducted on unwashed and “field-washed” long storage LLINs, with four LLINs of each brand tested per condition. Five mosquitoes were exposed for three minutes per cone on each net replicate. Long storage LLINs that failed to meet WHO cone bioassay threshold criteria (Table 1), were subjected to WHO tunnel test as per WHO guideline [8].
I-ACT assays
The I-ACT was used as an intermediate between laboratory and experimental hut tests [22]. One LLIN per condition (unwashed or washed twenty times) was tested as a confirmation of the WHO laboratory bioassay findings. Each LLIN or control was randomly assigned to one of the ten testing chambers of the I-ACT (Fig. 2). At 21:00 hours, volunteer-sleepers released thirty An. gambiae and thirty An. arabiensis, in each testing chamber. Mosquitoes were lightly dusted with fluorescent powder (SWADA, Cheshire, United Kingdom) to distinguish the strain as they are morphologically identical. At 06:00 AM, mosquitoes were collected into paper cups using a mouth aspirator. Mosquitoes were scored immediately after collection by strain and sorted into four categories: 1) dead blood-unfed, 2) dead blood-fed, 3) alive blood- unfed and 4) alive blood-fed. Mosquitoes were then held in the testing laboratory at 27°C ± 5°C and 40% to 100% relative humidity with access to 10% sugar solution. After 24-hours, the proportion of mosquitoes in each of the four categories was again scored using the above criteria. Following each night of the experiment, test nets were re-packed in their respective bags, chambers were cleaned and bed sheets were washed to prevent contamination. LLINs remained fixed to their respective chambers while sleeper volunteers rotated nightly for ten experimental nights so that each volunteer tested each net type once. This was done to account for difference between human attractiveness to mosquitoes that might affect the proportion of mosquitoes blood feeding. Acceptable mortality was ≤ 10% or ≥50% blood-feeding success in control [8] (Table 1: Fig. 2).
Data management and analysis
Data were recorded onto paper forms, double entered into Microsoft excel 2013 and cleaned prior to analysis. Data analysis was performed using STATA 13.1. Descriptive statistics were used for WHO cone and tunnel tests. For the I-ACT, both descriptive statistics and binomial logistic regression with mixed effects were conducted. The outcome measures were 24-hour mortality and blood-feeding inhibition. Model fit was tested using AIC [25]. For the model with mortality as the outcome, the best fitting model had treatment and volunteer as fixed effect and day as a random effect while best model with feeding success as the outcome had treatment as a fixed effect, with both volunteer and day as random effects.