Durable wall lining for malaria control in Liberia: results of a cluster randomized trial

Abstract

Background: Malaria control in Liberia depends upon universal coverage with pyrethroid-impregnated long-lasting insecticidal nets (LLINs). Despite regular mass distribution, LLIN coverage and usage is patchy. Pyrethroid resistance in malaria vectors may further reduce LLIN efficacy. Durable Wall Lining (DWL), a novel material treated with two non-pyrethroid class insecticides, was designed to be installed onto the surface of inner walls, and cover openings and ceiling surfaces of rural houses.

Objectives: AIM: To determine the malaria control efficacy of DWL.

Primary objective: To determine if DWL has an additional protective effect in an area of pyrethroid resistance.

Secondary objectives: To compare surface bio-availability of insecticides and entomological effectiveness over the study duration.

Design: A cluster randomized trial.

Participants: Children aged 2-59 months.

Control arm: 50 houses per 20 clusters, all of which received LLIN within the previous 12 months.

Active arm: 50 houses per 20 experimental clusters, all of which received LLINs with the previous 12 months, and had internal walls and ceilings lined with DWL.

Randomisation: Cluster villages were randomly allocated to control or active arms, and paired on 4 covariates.

Main outcome measures: PRIMARY MEASURE: Prevalence of infection with P. falciparum in children aged 2 to 59 months.

Secondary measure: Surface bioavailability and entomological effectiveness of DWL active ingredients.

Results: Plasmodium falciparum prevalence in active clusters after 12 months was 34.6% compared to 40.1% in control clusters (p = 0.052). The effect varied with elevation and was significant (RR = 1.3, p = 0.022) in 14 pairs of upland villages. It was not significant (RR = 1.3, p = 0.344) in 6 pairs of coastal villages. Pooled risk ratio (RR) was calculated in SAS (Cary, NC, USA) using the Cochran-Mantel-Haenszel (CMH) test for upland and coastal cluster pairs. DWL efficacy was sustained at almost 100% for 12 months.

Conclusions: Findings indicate that DWL is a scalable and effective malaria control intervention in stable transmission areas with pyrethroid-resistant vectors, where LLIN usage is difficult to achieve, and where local housing designs include large gable and eve openings. Trial registration ClinicalTrials.gov identifier: NCT02448745 (19 May 2015): https://clinicaltrials.gov/ct2/show/NCT02448745.

Keywords: Durable wall lining; LLINs; Malaria; Rural housing; Vectors.

Fig. 1

Flowchart of study design—Bomi County, Liberia. Initially, 42 clusters were identified, each housing approximately 50 children and with houses were spatially clustered. Participants were recruited in the 42 clusters for the baseline epidemiological survey. Two of the 42 clusters subsequently became insecure and had to be removed. The remaining 40 clusters were matched based on P. falciparum prevalence, population size, LLIN usage and district. Following random allocation to continued use of LLINs or use of LLINs plus DWL installation, 20 clusters were selected to receive DWL. Analysis was conducted for all 40 clusters, and also separately for 28 upland clusters, and for 12 coastal clusters, to determine any regional specific difference

Fig. 2

Map of the study area in Bomi County, Liberia. Control clusters are shown in grey, Active clusters are shown in red, pairing is shown with black lines. The division between upland and coastal regions is shown with dashed orange line. Administrative boundaries were used to select villages; no clusters were selected in the central administrative district. Inset: map of region, main map area shown as an orange rectangle

Fig. 3

Typical housing design in Bomi County, Liberia showing open gables and ceilings. Exterior photo shows the large gable allowing for cross-breeze and mosquito entry. Interior photo shows interior dividing walls extending approximately 2 m above the floor, an LLIN hung above the bed, and recycled LLINs sewn together to make a ceiling

Fig. 4

Photos of typical installation of DWL in houses in Bomi County, Liberia. Deltamethrin LLINs were present as an additional control measure in both arms of the study. Ceilings and windows were covered with DWL using wooden strips

Fig. 5

P. falciparum prevalence is balanced at baseline, then decreases in both arms during the study period. Bioavailability of insecticide in DWL decreases rapidly. A baseline balance in P. falciparum (Pf) prevalence between control and experimental (DWL installed) study arms in village clusters in Bomi County, Liberia. Boxplot represents mean, interquartile range, whiskers represent range, notch represents 95% confidence interval. B Change in proportion of P. falciparum (Pf) prevalence from baseline in April 2014 to April 2015 (12 months), and from baseline to the end of the study in April 2016 (24 months), in each cluster by study arm. Boxplot represents mean, interquartile range, whiskers represent range, notch represents 95% confidence interval. C Mortality of An. gambiae s.l. following exposure to fenpyroximate abamectin treated DWL, observed at 24 and 72 h post exposure in WHO cone bioassays. Bars represent mean of replicates, dots represent individual bioassays, error bars represent standard deviation. There were no important harms of unintended effects in any of the paired clusters