Background

Insecticide-treated bednets (ITNs) are effective in preventing malaria where vectors primarily bite indoors and late at night, but their effectiveness is uncertain where vectors bite outdoors and earlier in the evening. We studied the effectiveness of ITNs following a mass distribution in Haiti from May to September, 2012, where the Anopheles albimanus vector bites primarily outdoors and often when people are awake.

Methods

In this case-control study, we enrolled febrile patients presenting to outpatient departments at 17 health facilities throughout Haiti from Sept 4, 2012, to Feb 27, 2014, who were tested with malaria rapid diagnostic tests (RDTs), and administered questionnaires on ITN use and other risk factors. Cases were defined by positive RDT and controls were febrile patients from the same clinic with a negative RDT. Our primary analysis retrospectively matched cases and controls by age, sex, location, and date, and used conditional logistic regression on the matched sample. A sensitivity analysis used propensity scores to match patients on ITN use propensity and analyse malaria among ITN users and non-users. Additional ITN bioefficacy and entomological data were collected.

Findings

We enrolled 9317 patients, including 378 (4%) RDT-positive cases. 1202 (13%) patients reported ITN use. Post-hoc matching of cases and controls yielded 362 cases and 1201 matched controls, 19% (333) of whom reported consistent campaign net use. After using propensity scores to match on consistent campaign ITN use, 2298 patients, including 138 (7%) RDT-positive cases, were included: 1149 consistent campaign ITN users and 1149 non-consistent campaign ITN users. Both analyses revealed that ITNs did not significantly protect against clinical malaria (odds ratio [OR]=0·95, 95% CI 0·68–1·32, p=0·745 for case-control analysis; OR=0·95, 95% CI 0·45–1·97, p=0·884 for propensity score analysis). ITN and entomological data indicated good ITN physical integrity and bioefficacy, and no permethrin resistance among local mosquitoes.

Interpretation

We found no evidence that mass ITN campaigns reduce clinical malaria in this observational study in Haiti; alternative malaria control strategies should be prioritised.

Funding

The Global Fund to Fight AIDS, Tuberculosis, and Malaria, and the US-based Centers for Disease Control and Prevention (CDC).

Insecticide-treated bednets (ITNs) are a cornerstone of malaria prevention; multiple rigorous studies in sub-Saharan Africa have shown ITNs to be effective in preventing malaria morbidity when used consistently. ^{1 , 2 and 3} ITNs take advantage of the indoor feeding (endophagic) and indoor resting (endophilic) behaviours of some Anopheles mosquitoes and work by repelling and killing or decreasing the life span of mosquitoes, as well as providing a physical barrier between mosquitoes and users. ^{4 and 5} In Africa, the primary malaria vectors are from the Anopheles gambiae complex and the Anopheles funestus group, which predominantly bite indoors and when people are sleeping. ⁶

Limited evidence exists on ITN effectiveness in Latin America and the Caribbean. Anopheles albimanus , one of the dominant malaria vectors in Latin America, ⁷ often bites outdoors (exophagic) and rests outdoors (exophilic). ⁷ Peak feeding times for A albimanus occur closer to sunset and generally earlier in the night than other Anopheles species, ^{8 , 9 and 10} although there is substantial geographic heterogeneity. ⁹ Findings from previous research on ITN effectiveness with A albimanus have been mixed. In a Guatemalan study from the early 1990s, ¹¹ both untreated and treated nets reduced incidence of malaria (by 47% and 57%, respectively) compared with the absence of nets. A 1992 Peruvian study ¹² found non-significant reductions in malaria incidence after introduction of ITNs. Nicaraguan trials ¹³ in 1996 reported that insecticide-treated materials significantly reduced community-level clinical malaria incidence, but only where community usage was greater than 16%. ¹³ Two studies ¹⁴ from Ecuador from 1989–90 and from 1991–92, respectively, found no significant difference in malaria incidence after ITN introduction. A more recent observational study from Brazil, where the predominant vector, Anopheles darlingi , is active throughout the day, found no effect of ITNs after a mass distribution in 2012. ¹⁵