The Impact of Hotspot-Targeted Interventions on Malaria Transmission in Rachuonyo South District in the Western Kenyan Highlands: A Cluster-Randomized Controlled Trial

Abstract

Background: Malaria transmission is highly heterogeneous, generating malaria hotspots that can fuel malaria transmission across a wider area. Targeting hotspots may represent an efficacious strategy for reducing malaria transmission. We determined the impact of interventions targeted to serologically defined malaria hotspots on malaria transmission both inside hotspots and in surrounding communities.

Methods and findings: Twenty-seven serologically defined malaria hotspots were detected in a survey conducted from 24 June to 31 July 2011 that included 17,503 individuals from 3,213 compounds in a 100-km2 area in Rachuonyo South District, Kenya. In a cluster-randomized trial from 22 March to 15 April 2012, we randomly allocated five clusters to hotspot-targeted interventions with larviciding, distribution of long-lasting insecticide-treated nets, indoor residual spraying, and focal mass drug administration (2,082 individuals in 432 compounds); five control clusters received malaria control following Kenyan national policy (2,468 individuals in 512 compounds). Our primary outcome measure was parasite prevalence in evaluation zones up to 500 m outside hotspots, determined by nested PCR (nPCR) at baseline and 8 wk (16 June-6 July 2012) and 16 wk (21 August-10 September 2012) post-intervention by technicians blinded to the intervention arm. Secondary outcome measures were parasite prevalence inside hotpots, parasite prevalence in the evaluation zone as a function of distance from the hotspot boundary, Anopheles mosquito density, mosquito breeding site productivity, malaria incidence by passive case detection, and the safety and acceptability of the interventions. Intervention coverage exceeded 87% for all interventions. Hotspot-targeted interventions did not result in a change in nPCR parasite prevalence outside hotspot boundaries (p ≥ 0.187). We observed an average reduction in nPCR parasite prevalence of 10.2% (95% CI -1.3 to 21.7%) inside hotspots 8 wk post-intervention that was statistically significant after adjustment for covariates (p = 0.024), but not 16 wk post-intervention (p = 0.265). We observed no statistically significant trend in the effect of the intervention on nPCR parasite prevalence in the evaluation zone in relation to distance from the hotspot boundary 8 wk (p = 0.27) or 16 wk post-intervention (p = 0.75). Thirty-six patients with clinical malaria confirmed by rapid diagnostic test could be located to intervention or control clusters, with no apparent difference between the study arms. In intervention clusters we caught an average of 1.14 female anophelines inside hotspots and 0.47 in evaluation zones; in control clusters we caught an average of 0.90 female anophelines inside hotspots and 0.50 in evaluation zones, with no apparent difference between study arms. Our trial was not powered to detect subtle effects of hotspot-targeted interventions nor designed to detect effects of interventions over multiple transmission seasons.

Conclusions: Despite high coverage, the impact of interventions targeting malaria vectors and human infections on nPCR parasite prevalence was modest, transient, and restricted to the targeted hotspot areas. Our findings suggest that transmission may not primarily occur from hotspots to the surrounding areas and that areas with highly heterogeneous but widespread malaria transmission may currently benefit most from an untargeted community-wide approach. Hotspot-targeted approaches may have more validity in settings where human settlement is more nuclear.

Trial registration: ClinicalTrials.gov NCT01575613.

Conflict of interest statement

We acknowledge the donation of Permanet® 3.0 LLINs by Vestergaard Frandsen (Hanoi, Vietnam) and Bti Vectobac® by Valent BioSciences Corp (Libertyville, IL, US).

Figures

Fig 1. The study area in the highlands of western Kenya.

The study area comprised a 5 × 20 km rectangle in Rachuonyo South District, Nyanza Province.

Fig 2. Spatial variation in malaria antibody prevalence and nPCR parasite prevalence in the study area in Rachuonyo South District during a community survey conducted in June–July 2011.

(A) Distribution of sampled compounds and variations in altitude across the study site (contour interval = 25 m). (B) Combined seroprevalence (for AMA-1 or MSP-119) for individual 250 × 250 m sub-cells and the location of 27 significant hotspots derived from spatial scan analysis of compound-level data. (C) The ten hotspots that were selected for the cluster-randomized trail, presented with evaluation zones. (D) The 27 serological hotspot locations are overlaid on a map of nPCR-detected malaria parasite prevalence for compounds consisting of >3 individuals.

Fig 3. Malaria parasite prevalence by nPCR inside and outside serologically defined hotspots in the study area in Rachuonyo South District during a community survey conducted in June–July 2011.

nPCR-based parasite prevalence is plotted for individuals residing inside 27 serologically defined hotspots (black bars), 1–249 meters from the hotspot boundary (grey hatched bars), 250–500 m from the hotspot boundary (open hatched bars), and >500 meters from the hotspot boundary (open bars). Parasite prevalence by nPCR is shown per altitude band. Error bars indicate the upper limit of the 95% confidence interval; the p-value for the trend test is given, adjusting for correlations between observations from individuals living in the same compound. The number of individuals for whom samples were available for nPCR inside hotspot boundaries was 2,222 individuals (1,350–1,449 m), 2,494 (1,450–1,499 m), 1,348 (1,500–1,549 m), and 118 (1,550–1,650 m). The number of individuals for 1–249 m from hotspot boundaries was 698 (1,350–1,449 m), 1,248 (1,450–1,499 m), 1,113 (1,500–1,549 m), and 246 (1,550–1,650 m). The number of individuals for 250–500 m from hotspot boundaries was 544 (1,350–1,449 m), 681 (1,450–1,499 m), 661 (1,500–1,549 m), and 164 (1,550–1,650 m). The number of individuals for >500 m from hotspot boundaries was 544 (1,350–1,449 m), 176 (1,450–1,499 m), 405 (1,500–1,549 m), and 135 (1,550–1,650 m).

Fig 4. Overview of the cluster-randomized trial conducted in Rachuonyo South District in March–September 2012.

Clusters were serologically defined hotspots with a surrounding 500-m evaluation zone and were randomly allocated to the intervention (n = 5) or control arm (n = 5). Cross-sectional surveys were conducted at baseline (22 March–15 April 2012) and at 8 wk (16 June–6 July 2012) and 16 wk after the intervention (21 August–10 September 2012). In each survey, 25 compounds were randomly selected from within hotspots and 50 from the surrounding evaluation zone (25 compounds 1–249 m from the hotspot boundary and 25 compounds 250–500 m from the hotspot boundary). In intervention hotspots, data were available for all compounds at baseline and were therefore included in the analysis. If selected compounds were not inhabited or compound members were absent, the nearest non-selected inhabited compound was selected as a replacement. Compounds were not revisited before replacements were sought. fMDA, focal MDA.

Fig 5. Indoor densities of female anophelines by light trap in intervention and control clusters in Rachuonyo South District in March–August 2012.

Each symbol represents the number of female anophelines caught indoors by CDC light trap inside hotspots (filled circles) and in evaluation zones (open circles). Each trap night, four compounds were randomly selected within the hotspot and eight were selected in the evaluation zone per cluster. Findings are summarized for trapping rounds prior to roll-out of interventions in 22 March–30 April 2012 (one trapping night per compound) and post-intervention in 1 May–30 June 2012 (three trapping nights per compound) and 1 July–31 August 2012 (five trapping nights per compound). Findings are presented for three intervention clusters combined and for three control clusters combined.