Study area
The ITN durability study was performed in the Cove, Zagnanado, and Ouinhi Districts (COZO), of the Zou department of central Benin (7°11′N 1°59′E) between March 2020 and November 2023. The study was nested in a cRCT evaluating the efficacy of dual active ingredient nets for the control of clinical malaria compared to a pyrethroid-only net [16, 17]. The study area consisted of 123 villages with approximately 54,000 households and a population size of 220,000 inhabitants. It has two rainy seasons (May-July and September-November) though malaria transmission occurs year-round. A baseline survey performed in 2019 prior to the cRCT showed a high malaria infection prevalence of 43.5% despite high ITN use (96%) [18]. The vector population consists of both Anopheles gambiae and Anopheles coluzzii and is characterized by a high intensity of resistance to pyrethroids and susceptibility to chlorfenapyr and pyriproxyfen [19]. The main economic activities carried out by the population are agriculture, fishing, hunting, trade, and hospitality.
Study arms
Three different ITN types consisting of the 3 arms of the cRCT were evaluated in the durability study; two dual active ingredient nets (Interceptor® G2 and Royal Guard®) were compared to a standard pyrethroid-only ITN (Interceptor®). A description of the specifications of the three ITN types is provided below:
-
Interceptor® G2 (BASF AGRO B.V.), is a WHO-prequalified 100-denier, polyester ITN coated with a mixture of alpha-cypermethrin and chlorfenapyr at target concentrations of 100 mg/m2 (± 25%) and 200 mg/m2 (± 25%) respectively. The bursting strength of the fabric is ≥ 405 kPa while the mass per unit area is 40 g/m² (± 10%) for 100 denier yarn.
-
Royal Guard® (Disease Control Technologies, LLC, USA), is a WHO-prequalified 150-denier, polyethylene ITN incorporated with a mixture of alpha-cypermethrin and pyriproxyfen at target concentrations of 225 mg/m2 (± 25%) and 225 mg/m2 (± 25%) respectively. The bursting strength of the fabric is ≥ 450 kPa while the mass per unit area is 45 g/m² (± 10%) for 150 denier yarn.
-
Interceptor® (BASF AGRO B.V.), is a WHO-prequalified 100-denier, polyester ITN coated with alpha-cypermethrin at target concentrations of 200 mg/m2 (± 25%). The bursting strength of the fabric is ≥ 405 kPa while the mass per unit area is 40 g/m² (± 10%) for 100 denier yarn. Interceptor® is the control arm of the trial against which Interceptor® G2 and Royal Guard® are compared.
Study design
The study design has been described in detail previously [17]. Participating households were blinded to the type of ITN they received while field data collectors were blinded to the allocation of households to study arms. Ten out of 20 clusters of each study arm of the cRCT were randomly selected for the durability assessment (Fig. 1). Following ITN distribution, households were visited in these clusters to recruit and prospectively follow 2 separate cohorts of ITNs per arm (Cohorts 1 and 2) every 6–12 months to monitor their physical and insecticidal durability over 36 months. Cohort 1 consisted of a total of 2,428 nets from 348–391 randomly selected households in 5 clusters per study arm. These nets were followed for ITN survivorship and fabric integrity at 6-, 12-, 24-, 30- and 36 months post ITN distribution. Cohort 2 consisted of 1860 nets per study arm from 593 households from all 10 randomly selected clusters (~ 60 households per cluster) per study arm that were withdrawn (and replaced with new nets) for assessment of insecticidal activity in laboratory bioassays and experimental hut trials, the content of each active ingredient and for other studies. Cohort 1 and Cohort 2 nets received separate identification numbers and were sampled from different households to prevent the risk of destructive sampling of nets intended for ITN survivorship studies (cohort 1).
Household sensitisation and marking of study nets
ITNs were recruited for the durability study in the first month after ITN distribution from selected clusters of each arm of the cRCT using the ITN distribution list of the trial. Field workers visited selected households for each cohort of ITNs to mark study nets available in them. Using a wash-resistant marker, a unique code was written on the study net to indicate the study arm, cluster ID, household ID and the cohort to which it belonged. During this visit, the study team explained the durability study to the householders in their local language and assisted them in hanging up their nets if needed. Only nets that were hung up and in use at the start of the trial were included. The GPS coordinates and characteristics of each household included in the durability assessment were also recorded.
ITN attrition and fabric integrity (cohort 1)
Households in cohort 1 were visited at 6-, 12-, 24- and 36-months post ITN distribution and nets were assessed for attrition and fabric integrity at each time point. Attrition was assessed by recording the physical presence/absence of each net that was recruited in the household. Where a net was not found, the householder was asked about the reason for the loss of the net e.g given away, sold, stolen, worn out and disposed of etc. Attrition was classified into 3 categories; a) due to physical damage (wear and tear), b) due to the net being removed (given away, stolen, sold or used in another location) and c) due to the net being used for other purposes (repurposed). Nets that had never been used were recorded and excluded from the analysis. At each time point, all available cohort 1 ITNs of each type were assessed for fabric integrity (hole index) and condition. A minimum of 250 nets were assessed and to ensure this number, cohort 2 nets were also assessed for fabric integrity when available cohort 1 nets were insufficient. The survey team inspected the nets outside in broad daylight to determine the hole index using a portable frame over which the net was draped during the inspection. The nets were returned to the family right after the inspection.
The number of holes and hole sizes on each ITN panel was measured using hole assessment sheets and classified into 4 sizes (size 1: 0.5-2cm, Size 2: 2–10 cm, Size 3: 10-25cm, and size 4: >25cm). Physical integrity was measured as the proportionate hole index (pHI) which was calculated as pHI = (1 × number of size 1 holes) + (23 × number of size 2 holes) + (196 × number of size 3 holes) + (576 × number of size 4 holes). Nets were then categorized based on recommended cut-off points for pHI into “good” condition (pHI 0–64), “acceptable” condition (pHI 65–642) and “torn” (pHI 643+) defined in WHO guidelines [12].
To determine the proportion of study nets that were in active use at each time point, we recorded the proportion of nets that were found hanging over sleeping places during each survey. Previous studies in Benin have indicated that householders usually have spare new ITNs from previous campaigns or routine distribution channels in their possession (unpublished data) which may affect the frequency with which the study nets are discarded. At 24 months, 362 participating households that had lost their study nets were further surveyed to identify the number, type and sources of non-study nets that they had replaced their study nets with. Householders were also administered a short questionnaire to determine what ITN fabric texture type (polyester vs. polyethylene) they preferred and to provide reasons why. They were shown examples of the two ITN fabric types and allowed to touch them before providing their responses.
ITN bioefficacy and chemical analysis (cohort 2)
For each ITN type a total of 120 nets belonging to cohort 2 were randomly selected and destructively sampled from households at 6, 12, 24 and 36-months post-distribution for laboratory and chemical analysis (Table 1). They were replaced with new nets of the same type at each sampling point. Thirty nets of each ITN type withdrawn at each timepoint were subjected to laboratory bioassays (WHO cone bioassays and tunnel tests where necessary) to monitor entomological efficacy against mosquito vectors and to chemical analysis to monitor changes in alphacypermethrin, chlorfenapyr and pyriproxyfen content. Four to five net pieces measuring 30x30cm cut from each ITN were tested in laboratory bioassays. A similar number of net pieces were obtained from adjacent positions to those for laboratory bioassays and preserved at 40 C for chemical analysis.
Table 1
Number of Interceptor® G2, Royal Guard® and Interceptor® ITNs tested for bioefficacy and chemical content
Months post ITN distribution | No. of ITNs tested in lab bioassays | No. of pieces per ITN | Total number of pieces for bioassays | No. of pieces for chemical analysisa |
0 | 30 | 5 | 150 | 150 |
12 | 30 | 4 | 120 | 120 |
24 | 30 | 4 | 120 | 120 |
36 | 30 | 4 | 200 | 120 |
Total | 120 | | 590 | 510 |
a Net pieces preserved for chemical analysis were obtained from adjacent positions on the same ITNs cut for bioassays.
Mosquito strains for laboratory bioassays
Cone bioassays and tunnel tests were performed to monitor the bioefficacy of each AI in each ITN brand using laboratory-reared susceptible and pyrethroid-resistant strains of Anopheles gambiae s.l.. All strains were maintained at CREC/LSHTM insectary in Cotonou, Benin. The characteristic of each strain is described below.
-
An. gambiae sensu stricto Kisumu, an insecticide-susceptible reference strain originating from the Kisumu area in Kenya and was colonised at the CREC/LSHTM insectary.
-
An. coluzzii Akron, a pyrethroid and carbamate resistant strain originating from Akron (9°19′N2°18′E), Southern Benin, and maintained at CREC/LSHTM insectary. Resistance is mediated by target site mutations (L1014F kdr and Ace-1R) and overexpressed cytochrome P450 enzymes [20].
WHO susceptibility bioassays were performed during each round of bioassays to confirm the susceptibility status of each mosquito strain to the insecticides used on the nets. At each time point, approximately 100 mosquitoes of each strain were exposed in batches of 25 in WHO bottle bioassays treated with alpha-cypermethrin (0.05%), chlorfenapyr (100µg) and pyriproxyfen (100µg). For alpha-cypermethrin and chlorfenapyr unfed 2–5 days-old mosquitoes were exposed for 1 hour and mortality was recorded after 24h for alpha-cypemethrin and 72h after exposure to chlorfenapyr. For pyriproxyfen, blood-fed 5–8 days old mosquitoes were exposed for 1 hour in treated bottles and survivors were dissected 3 days later to measure the proportional reduction in fertility relative to the control unexposed mosquitoes as described previously [21]. The bioassays were performed at a temperature of 27°C ± 2°C and a relative humidity of 75% ± 10%.
Laboratory bioassay methods
In line with the existing WHO guidelines [6, 12], ITN bioefficacy at each timepoint was expressed in terms of the proportion of ITNs passing pre-determined efficacy criteria for each active ingredient in cone bioassays or tunnel tests. Table S1 summarises the methods and strains used to assess the efficacy of each active ingredient in each ITN type. All bioassays were performed at 27 ± 2°C and 75% ± 10% relative humidity.
Bioefficacy of alphacypermethrin in Interceptor® and Royal Guard®
Following WHO guidelines for testing the durability of the bioefficacy of pyrethroid-treated ITNs, the efficacy of alpha-cypermethrin in Interceptor® and Royal Guard® was monitored in 3-minute cone bioassays using unfed 3–5 days old mosquitoes of the susceptible An gambiae Kisumu strain. At each time point, 40–50 mosquitoes were exposed to each ITN in replicates of ~ 5 mosquitoes per cone and 2 cones per ITN piece. ITNs that failed to achieve WHO efficacy criteria in these cone bioassays (pooled knock-down ≥ 95% or pooled 24hrs mortality ≥ 80%) were subjected to tunnel tests. For each ITN that failed in cone bioassays, only one ITN piece (with cone bioassay mortality that is closest to the mean) was tested in tunnels. Approximately 100 unfed 5–8 days old susceptible An gambiae Kisumu were exposed to each ITN piece in the tunnel tests and efficacy of alpha-cypermethrin in Interceptor® and Royal Guard® was measured in terms of mortality after 24hrs (≥ 80%) or blood-feeding inhibition (≥ 90%).
Bioefficacy of pyriproxyfen in Royal Guard®
To monitor the bioefficacy of pyriproxyfen in Royal Guard, blood-fed 5–8 days old mosquitoes of the pyrethroid-resistant An coluzzii Akron strain were exposed for 3 minutes in cone bioassays and assessed for the impact on ovary development using dissection. Dissections were performed 72 hrs after exposure under a microscope and mosquitoes were classified as fertile or infertile following previously described SOPs [22]. A total of 80–100 blood-fed mosquitoes were tested against each Royal Guard® ITN in replicates of 5 mosquitoes per cone and 4 cones per ITN piece and efficacy was measured for each ITN in terms of the proportional reduction in fertility relative to control unexposed mosquitoes. A cut off of ≥ 50% reduction in fertility relative to the untreated control nets was applied to determine pass rates of Royal Guard® nets for pyriproxyfen bioefficacy. Only tests for which at least 30% of unexposed mosquitoes are found fertile were considered valid.
Bioefficacy of chlorfenapyr in Interceptor® G2
The bioefficacy of chlorfenapyr in Interceptor® G2 was assessed in tunnel tests using the pyrethroid-resistant Akron strain. To improve the efficiency of the tunnel test bioassays, only 1 ITN piece per whole Interceptor® G2 ITN was tested in tunnels. For each tunnel, a total of ~ 100 unfed 5–8 days old pyrethroid resistant An coluzzii Akron were exposed overnight and efficacy of Interceptor® G2 in tunnels was measured in terms of mosquito mortality after 72hrs (≥ 80%) or blood-feeding inhibition (≥ 90%).
Chemical analysis methods
Net pieces preserved for chemical analysis at 0, 12, 24 and 36 months (510 pieces per ITN type) were wrapped in Aluminum foil and stored at 40C (+/-20C) and afterwards shipped to Centre Walloon de Recherches Agronomiques (CRA-W), Belgium, for detection of fabric weight and AI content. The methods used for AI extraction have been described elsewhere [14, 23]. Gas Chromatography with Flame Ionisation Detection (GC-FID) was used to determine the content of each AI. ITN pieces from the same net were pooled to provide a single chemical content reading per AI for each whole ITN per net type sampled at each time point.
Data management
Household data collected during the census and ITN follow-up surveys were captured on electronic forms using smartphones installed with OpenDataKit (ODK) Collect while entomological data was recorded on data entry forms and double-entered pre-designed Excel databases. Throughout the study, electronic data was stored encrypted while paper forms were locked up in secured cabinets and were available only to study investigators and data management staff by passwords and keys. All personal data was anonymized using a unique identifier number for each participant and household to ensure confidentiality. At the end of the study, all electronic files and data entry forms have been respectively stored on the server and archive of the CREC-LSHTM GLP-certified Facility for 10 years.
Statistical analysis
Data from the surveys at 6, 12, 24, 30 and 36 months was used to calculate attrition, functional survival and median survival time. Attrition was estimated at each time point as the proportion of study nets that were not found in households either due to physical damage, removal or repurposing relative to the number of nets distributed as a baseline after removing nets that were lost to follow-up. Functional survival was calculated at each survey time point as the proportion of nets in serviceable condition after excluding nets that were given away, sold or stolen. Median survival was calculated as the linear extrapolation of the survival values from either the 12- or 24-month rounds (whichever was < 85% survival) and the 36 months round to the y = 50% line, using the following formula:
$$\:Tm=t1+\left(t2-t1\right)*(p1-50)/(p1-p2)$$
where tm is the median survival time, t1 and t2 are the first and second timepoints in years and p1 and p2 are the proportions surviving to the first and second timepoints respectively in percent. Confidence intervals (CI) for this estimate were calculated by projecting the 95% CI from the survival estimates in the same way as described above. Hazard ratios (HRs) for the difference in functional survival and 95% confidence intervals were predicted using Cox proportional regression models. A chi-squared test was used to assess the proportion of nets of each ITN type passing the WHO criteria for alphacypermethrin, chlorfenapyr and pyriproxyfen bioefficacy based on combined cone and tunnel tests.
Ethical considerations
Ethical approval
was obtained from the ethics review boards of the Ministry of Health in Benin (N°6/30/MS/DC/DRFMT/CNERS/SA), the institutional review board of the London School of Hygiene and Tropical Medicine (N°16,237), and the WHO Research Ethics Review Committee (ERC.0003153). The cRCT is registered on clinicaltrials.gov (NCT03931473) and the study protocol was published before the start of the trial [16]. Heads of households selected for the durability study gave informed consent prior to their participation. The consent form was written in French and explained to them in their local language. Where the individual was unable to read or write, their fingerprint was taken, and a signature obtained from a witness to the informed consent procedure. All personal data was anonymised prior to data processing. Approval for use of guinea pigs for tunnel tests was obtained from the LSHTM Animal Welfare Ethics Review Board (Ref: 2020-01). The methods described in this paper were carried out following relevant guidelines and regulations.