Insecticide‐treated nets for preventing malaria

Joseph Pryce; Marty Richardson; Christian Lengeler

doi:10.1002/14651858.CD000363.pub3

. 2018 Nov 6;2018(11):CD000363. doi: 10.1002/14651858.CD000363.pub3

Insecticide‐treated nets for preventing malaria

Joseph Pryce ^1,^✉, Marty Richardson ¹, Christian Lengeler ²

Editor: Cochrane Infectious Diseases Group

PMCID: PMC6418392 PMID: 30398672

Abstract

Background

A previous version of this Cochrane Review identified that insecticide‐treated nets (ITNs) are effective at reducing child mortality, parasite prevalence, and uncomplicated and severe malaria episodes. Insecticide‐treated nets have since become a core intervention for malaria control and have contributed greatly to the dramatic decline in disease incidence and malaria‐related deaths seen since the turn of the millennium. However, this time period has also seen a rise in resistance to pyrethroids (the insecticide used in ITNs), raising questions over whether the evidence from trials conducted before resistance became widespread can be applied to estimate the impact of ITNs on malaria transmission today.

Objectives

The primary objective of this review was to assess the impact of ITNs on mortality and malaria morbidity, incorporating any evidence published since the previous update into new and existing analyses, and assessing the certainty of the resulting evidence using GRADE.

Search methods

We searched the Cochrane Infectious Diseases Group Specialized Register, the Cochrane Central Register of Controlled Trials (CENTRAL) published in the Cochrane Library, MEDLINE, Embase, LILACS, the World Health Organization (WHO) International Clinical Trials Registry Platform, ClinicalTrials.gov, and the ISRCTN registry for new trials published since 2004 and up to 18 April 2018.

Selection criteria

We included individual randomized controlled trials (RCTs) and cluster RCTs comparing bed nets or curtains treated with a synthetic pyrethroid insecticide at a minimum target impregnation dose recommended by the WHO with no nets or untreated nets.

Data collection and analysis

One review author assessed the identified trials for eligibility and risk of bias, and extracted data. We compared intervention and control data using risk ratios (RRs), rate ratios, and mean differences, and presented all results with their associated 95% confidence intervals (CIs). We assessed the certainty of evidence using the GRADE approach. We drew on evidence from a meta‐analysis of entomological outcomes stratified by insecticide resistance from 2014 to inform the GRADE assessments.

Main results

Our updated search identified three new trials. A total of 23 trials met the inclusion criteria, enrolling more than 275,793 adults and children. The included studies were conducted between 1987 and 2001.

ITN versus no nets

Insecticide‐treated nets reduce child mortality from all causes by 17% compared to no nets (rate ratio 0.83, 95% CI 0.77 to 0.89; 5 trials, 200,833 participants, high‐certainty evidence). This corresponds to a saving of 5.6 lives (95% CI 3.6 to 7.6) each year for every 1000 children protected with ITNs. Insecticide‐treated nets also reduce the incidence of uncomplicated episodes of Plasmodium falciparum malaria by almost a half (rate ratio 0.55, 95% CI 0.48 to 0.64; 5 trials, 35,551 participants, high‐certainty evidence) and probably reduce the incidence of uncomplicated episodes of Plasmodium vivax malaria (risk ratio (RR) 0.61, 95% CI 0.48 to 0.77; 2 trials, 10,967 participants, moderate‐certainty evidence).

Insecticide‐treated nets were also shown to reduce the prevalence of P falciparum malaria by 17% compared to no nets (RR 0.83, 95% CI 0.71 to 0.98; 6 trials, 18,809 participants, high‐certainty evidence) but may have little or no effect on the prevalence of P vivax malaria (RR 1.00, 95% CI 0.75 to 1.34; 2 trials, 10,967 participants, low‐certainty evidence). A 44% reduction in the incidence of severe malaria episodes was seen in the ITN group (rate ratio 0.56, 95% CI 0.38 to 0.82; 2 trials, 31,173 participants, high‐certainty evidence), as well as an increase in mean haemoglobin (expressed as mean packed cell volume) compared to the no‐net group (mean difference 1.29, 95% CI 0.42 to 2.16; 5 trials, 11,489 participants, high‐certainty evidence).

ITN versus untreated nets

Insecticide‐treated nets probably reduce child mortality from all causes by a third compared to untreated nets (rate ratio 0.67, 95% CI 0.36 to 1.23; 2 trials, 25,389 participants, moderate‐certainty evidence). This corresponds to a saving of 3.5 lives (95% CI ‐2.4 to 6.8) each year for every 1000 children protected with ITNs. Insecticide‐treated nets also reduce the incidence of uncomplicated P falciparum malaria episodes (rate ratio 0.58, 95% CI 0.44 to 0.78; 5 trials, 2036 participants, high‐certainty evidence) and may also reduce the incidence of uncomplicated P vixax malaria episodes (rate ratio 0.73, 95% CI 0.51 to 1.05; 3 trials, 1535 participants, low‐certainty evidence).

Use of an ITN probably reduces P falciparum prevalence by one‐tenth in comparison to use of untreated nets (RR 0.91, 95% CI 0.78 to 1.05; 3 trials, 2,259 participants, moderate‐certainty evidence). However, based on the current evidence it is unclear whether or not ITNs impact on P vivax prevalence (1 trial, 350 participants, very low certainty evidence) or mean packed cell volume (2 trials, 1,909 participants, low certainty evidence).

Authors' conclusions

Although there is some evidence that insecticide resistance frequency has some effects on mosquito mortality, it is unclear how quantitatively important this is. It appeared insufficient to downgrade the strong evidence of benefit on mortality and malaria illness from the trials conducted earlier

12 April 2019

Up to date

All studies incorporated from most recent search

All eligible published studies found in the last search (18 Apr, 2018) were included

Plain language summary

Insecticide‐treated nets for preventing malaria

What is the aim of this review?

Insecticide‐treated nets (ITNs) are a core intervention for malaria control. A previous version of this Cochrane Review showed they are very effective at reducing malaria‐related death and illness. Since the review was published, many areas affected by malaria have reported mosquito populations that are resistant to the insecticides used in ITNs. The aim of this review update was to evaluate the available evidence and find out whether ITNs continue to be effective at controlling the disease. Cochrane researchers collected and analysed relevant studies and assessed the overall certainty of the evidence.

What was studied in the review?

This review update summarized trials published since the previous review that evaluated the impact of ITNs on malaria‐related deaths and illness, compared to both no nets and untreated nets. After searching for relevant trials up to 18 April 2018, we identified three new randomized controlled trials (studies in which participants are assigned to a treatment group using a random method). In total, we included 23 trials, enrolling more than 275,000 adults and children, to evaluate the effectiveness of ITNs for reducing the burden of malaria. The included studies provided evidence of the impact of ITNs on infection from two types of malaria parasites, Plasmodium falciparum and Plasmodium vivax.

What are the main results of the review?

Twelve trials (nine in Africa, one in Cambodia, one in Myanmar, and one in Pakistan) assessed the impact of ITNs in comparison to no nets. From these trials, we concluded that ITNs reduce the child mortality from all causes, corresponding to a saving of 5.6 lives each year for every 1000 children protected with ITNs (high‐certainty evidence). ITNs also reduce the number of P falciparum cases per person per year and the proportion of people infected with P falciparum parasites (high‐certainty evidence). ITNs probably reduce the number of P vivax cases per person per year and may reduce the proportion of people infected with P vivax parasites (moderate‐certainty evidence).

Eleven trials (three in sub‐Saharan Africa, six in Latin America, and two in Thailand) assessed the impact of ITNs in comparison to untreated nets. From these trials, we concluded that ITNs probably reduce the child mortality from all causes, corresponding to a saving of 3.5 lives each year for every 1000 children protected with ITNs (moderate‐certainty evidence). ITNs also reduce the number of P falciparum cases per person per year (high‐certainty evidence), and probably reduce the proportion of people infected with P falciparum parasites (moderate‐certainty evidence). Whilst ITNs may also reduce the number of P vivax cases per person per year (low‐certainty evidence), it is unclear if the proportion of people infected with P vivax parasites is any lower in those using an ITN than those using an untreated net (very low certainty evidence).

In interpreting these results, we considered that there are a growing number of mosquito populations that have been shown to be able to survive exposure to the insecticides used in ITNs. However, it is currently unclear how quantitatively important this is, and this seems insufficient to downgrade the existing evidence of an effect of ITNs in preventing malaria‐related mortality and illness.

Key messages

ITNs, whether compared to no nets or to untreated nets, continue to be effective at reducing child mortality and malaria‐related illness in affected areas.

Summary of findings

Background

The 2004 Cochrane Review ‘Insecticide‐treated bed nets and curtains for preventing malaria’ demonstrated the effectiveness of insecticide‐treated nets (ITNs) for reducing malaria prevalence, morbidity, and mortality. Incorporating information from 22 randomized controlled trials (RCTs), the review found that ITNs reduced child mortality by 17%. In areas of stable malaria transmission, ITNs also reduced parasite prevalence by 13%, uncomplicated malaria episodes by 50%, and severe malaria by 45% compared to equivalent populations with no nets (Lengeler 2004). The World Health Organization (WHO) now recommends ITNs as a core intervention for malaria control.

Between 2010 and 2015, the estimated percentage of the at‐risk population sleeping under an ITN rose from 30% to 53%. During this time, disease incidence and malaria‐related deaths have fallen by 21% and 29%, respectively (WHO 2016). Additionally, parasite prevalence in endemic sub‐Saharan Africa decreased by 50% between 2001 and 2015, with 68% of this decline attributed to the use of ITNs (Bhatt 2015)

Emerging insecticide resistance poses a challenge to current malaria vector control methods. There are only four classes of insecticide in use for public health, with just two mechanisms of action. A lack of funding for research into new insecticides has meant that the most recently developed class is the pyrethroids, which were developed over 40 years ago (Ranson 2011). During this period, 27 countries have reported resistance to pyrethroids, and the number of susceptible Anopheles populations continues to decline (Ranson 2016). The effectiveness of ITNs is particularly at risk, as pyrethroids are the only class of insecticide considered safe for prolonged human contact and therefore appropriate for ITN use (Zaim 2000).

Insecticide resistance is commonly detected using laboratory‐based bioassays and experimental hut studies, but these do not necessarily indicate reduced ITN impact on real‐life clinical outcomes (Ranson 2016). It remains unclear whether the dramatic increases in ITN use and pyrethroid resistance in the years following Lengeler’s 2004 review have reduced the clinical effectiveness of ITNs. The purpose of this review update was therefore to identify, critically appraise, and summarize any trials published since the last edition of the review, incorporating modern methods for systematic reviews that allow combined analysis of cluster RCTs (cRCTs), and assessment of the certainty of the estimates of the effect of ITNs. We were able to draw on a systematic review of entomological outcomes in the presence of pyrethroid resistance from 2014 to help inform the GRADE assessments for indirectness (Strode 2014).

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Individual RCTs and cluster RCTs (cRCTs).

Types of participants

Children and adults living in malaria transmission settings.

We excluded trials examining only pregnant women, because these are reviewed elsewhere (Gamble 2006), and trials examining only soldiers or travellers, as these are not representative of the general population.

Types of interventions

Bed nets or curtains treated with a synthetic pyrethroid insecticide at a minimum target impregnation dose recommended by the WHO, which is as follows.

200 mg/m² permethrin or etofenprox.
30 mg/m² cyfluthrin.
20 mg/m² alpha‐cypermethrin.
10 mg/m² deltamethrin/lambda‐cyhalothrin.

No distinction was made between insecticide‐treated bed nets and door/window/eave/wall curtains.

Control populations were those provided with either no net or with an untreated net.

Types of outcome measures

Primary outcomes

Child mortality from all causes.

Secondary outcomes

Uncomplicated clinical episodes: measured using site‐specific definitions, including measured or reported fever, with or without parasitological confirmation. Measurements were usually done in the frame of prospective longitudinal studies, as a rate of episodes per unit of time (incidence). We also included trials using validated retrospective assessments in the frame of cross‐sectional surveys, providing a percentage of the population who had experienced an uncomplicated episode in a unit of time (cumulative incidence). When reported separately, P falciparum and P vivax episodes were analysed separately. We also included trials that reported the incidence of episodes of any Plasmodium species.
Parasite prevalence: parasite prevalence due to P falciparum and P vivax was obtained using the site‐specific method for estimating parasitaemia, usually thick or thin blood smears or both. When more than one survey was done, the reported prevalence result is the average prevalence of all the surveys.
Severe disease: measured using site‐specific definitions, which were based on the WHO guidelines, WHO 1990, and on Marsh 1995. The definition included P falciparum parasitaemia. Cerebral malaria was defined as coma or prostration and/or multiple seizures. The cut‐off for severe, life‐threatening anaemia was set at 5.1 g/L (WHO 1990).
Anaemia: expressed in mean packed cell volume (PCV), equivalent to the percentage haematocrit. Results given in grams per decilitre were converted with a standard factor of 3:1 so that 1 g/dL equals 3% PCV.
The outcome measures below were considered in the previous review (Lengeler 2004), but were not considered priority outcomes at the time of this update and were therefore not included. Appendix 1 details the full inclusion criteria for this previous update.
- High parasitaemia: measured using site‐specific definitions of high parasitaemia, provided the cut‐off value between high and low was determined prior to data analysis.
- Splenomegaly: measured in all trials using the Hackett scale.
- Anthropometric measures: standard anthropological measures (weight‐for‐age, height‐for‐age, weight‐for‐height, skinfold thickness, or mid‐upper arm circumference)

Search methods for identification of studies

The previous review, Lengeler 2004, used the search strategy outlined below to identify included studies.

The following databases were searched using the search terms and strategy described in Appendix 2: Cochrane Infectious Diseases Group Specialized Register (January 2003); Cochrane Central Register of Controlled Trials (CENTRAL), published in the Cochrane Library (Issue 1, 2003); MEDLINE (1966 to October 2003); Embase (1974 to November 2002); and LILACS (Latin American and Caribbean Health Science Information database) (1982 to January 2003).
The following foreign language tropical medicine journals were handsearched, covering the period from 1980 to 1997: Bulletin OCEAC, Bulletin de la Société de Pathologie Exotique, Médecine Tropicale, and Revista do Instituto de Medicina Tropical de Sao Paulo.
Researchers actively involved in the field of ITNs were contacted and asked about unpublished past or ongoing work.
The following agencies, which have funded ITN trials, were contacted: UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (TDR); International Development Research Centre (IDRC), Canada; the Department for International Development, UK; and the European Union Directorate‐General XII.
The following manufacturers of pyrethroids used for treating netting were contacted: AgrEvo (now part of Bayer), Bayer, Cyanamid, Mitsui, Sumitomo, and Zeneca (now part of Syngenta).
The following reviews were consulted: Abdulla 1995; Bermejo 1992; Carnevale 1991; Cattani 1997; Choi 1995; Curtis 1992; Molyneaux 1994; Rozendaal 1989; Sexton 1994; Voorham 1997; WHO 1989; Xu 1988; Yadav 1997; and Zimmerman 1997.
The following books on the subject of ITNs were consulted: Control of Disease Vectors in the Community (Curtis 1991), Malaria: Waiting for the Vaccine (Targett 1991), and Net Gain, a New Method for Preventing Malaria Deaths (Lengeler 1996).
The reference lists of all trials identified by the above methods were consulted.

We considered for this review update all studies identified using the strategy above. Detailed below is the additional search process we undertook to identify new studies conducted since 2003.

Electronic searches

We searched the following databases, using the search terms and strategy described in Appendix 2: Cochrane Infectious Diseases Group Specialized Register (2003 to 18 April 2018); Cochrane Central Register of Controlled Trials (Issue 4, 2018); MEDLINE (PubMed, 2003 to 18 April 2018); Embase (Ovid, January 2003 to 18 April 2018); and LILACS (Latin American and Caribbean Health Science Information database) (2003 to 18 April 2018). To identify any ongoing trials, we also searched the World Health Organization (WHO) International Clinical Trials Registry Platform (www.who.int/ictrp/search/en/; 18 April 2018), ClinicalTrials.gov (https://clinicaltrials.gov/; 18 April 2018) and the ISRCTN registry (www.isrctn.com/; 18 April 2018)

Searching other resources

We contacted organizations, including the WHO and the Centers for Disease Control and Prevention (CDC), for ongoing and unpublished trials. The reference lists of all trials identified by the above methods were also consulted.

Data collection and analysis

Selection of studies

One review author (JP) screened the titles and abstracts of articles identified by the literature searches for potential inclusion in the review. The full‐text articles of potentially relevant trials were assessed using an eligibility form based on the inclusion criteria. Multiple publications of the same trial were included only once. Excluded studies are listed together with their reasons for exclusion in the Characteristics of excluded studies table. We have illustrated the study selection process in a PRISMA diagram (see Figure 1).

Data extraction and management

One review author (JP) extracted information from each of the included studies (identified in both search processes) using pre‐piloted electronic data extraction forms. In the case of missing data in studies from the initial search, we contacted the original study authors or the author of the original review (CL) for confirmation. In case of missing data in newly identified studies, we contacted the original study authors for clarification.

We extracted data on the following.

Trial design: type of trial; length of follow‐up; method of participant selection; sample size; and method of blinding of participants and personnel. For cRCTs we also recorded the number of clusters randomized, the number of clusters analysed, and method of adjustment for clustering.
Participants: number of participants; inclusion/exclusion criteria.
Intervention: description of intervention (active ingredient, dose, retreatment times, type of net); description of control.
Outcomes: definition of outcomes; diagnostic method or surveillance method; passive or active case detection.
Other: study location; malaria endemicity, entomological inoculation rate (EIR), primary vector species; Plasmodium species.

For dichotomous outcomes, we extracted the number of participants who experienced each outcome and the total number of participants in each treatment group. Where trials conducted multiple cross‐sectional surveys during the intervention period, we took an average of the numerators and denominators across the total number of surveys. We selected this procedure in order to avoid inflating the denominator artificially by adding up the participants from repeated surveys. For count data outcomes, we extracted the number of outcomes in the treatment and control groups and the total person‐time at risk in each group, or the rate ratio and a measure of variance (for example, standard error). For continuous outcomes, we extracted the mean and a measure of variance (standard deviation).

We considered the impact of ITNs on the primary outcome of child mortality from all causes likely to be age‐dependent. In addition to extracting the total number of deaths in the total study population, where possible we extracted the number of deaths and total number of children within a high‐risk age group of 1 to 59 months. This allowed an estimate for each age group of the number of deaths that can be avoided through the provision of ITNs.

Assessment of risk of bias in included studies

We assessed the risk of bias for each study using the Cochrane ‘Risk of bias' tool (Higgins 2011). For each included cRCT, we also assessed the five additional criteria relating specifically to cRCTs listed in Section 16.3.2 of the Cochrane Handbook for Systematic Reviews of Interventions. We classified judgements of risk of bias as either low, high, or unclear risk of bias. We have summarized the results of the assessment in a ‘Risk of bias' summary figure.

Measures of treatment effect

We compared intervention and control data using rate ratios, risk ratios (RRs), and mean differences, and presented all results with their associated 95% confidence intervals (CIs).

Unit of analysis issues

If included cRCTs had not adjusted for clustering in the analysis, we adjusted the data before combining it. We adjusted data by multiplying the standard errors by the square root of the design effect (Higgins 2011), which is determined by the intracluster correlation coefficient (ICC). If the trial did not report the ICC value, we used the ICC from a similar trial that reported the same outcome (Smithuis 2013).

Dealing with missing data

In case of missing data, we applied available‐case analysis, only including data on the known results. The denominator was the total number of participants who had data recorded for the specific outcome. For outcomes with no missing data, we performed analyses on an intention‐to‐treat basis. We included all participants randomized to each group in the analyses and analysed participants in the group to which they were randomized.

Assessment of heterogeneity

We inspected forest plots for overlapping CIs and assessed statistical heterogeneity in each meta‐analysis using the I² statistic and Chi² test values. We considered I² statistic values between 30% and 60% indicative of moderate heterogeneity; between 50% and 90% substantial heterogeneity; and between 75% and 100% considerable heterogeneity. We considered a Chi² test statistic with a P value greater than 0.10 indicative of statistically significant heterogeneity. We explored clinical and methodological heterogeneity through consideration of the trial populations, methods, and interventions, and by visualization of trial results.

Assessment of reporting biases

We intended to investigate reporting biases (such as publication bias) by assessing funnel plot asymmetry (Harbord 2006). However, as each meta‐analysis included fewer than 10 trials, such an assessment was not possible.

Data synthesis

We analysed data using Review Manager 5 (RevMan 2014). As we detected no heterogeneity between the study types, we pooled data from both individual RCTs and cluster‐adjusted cRCTs in a meta‐analysis (Richardson 2016).

Based on the consideration of clinical, epidemiological, and methodological heterogeneity between the trials, we used a random‐effects model.

Certainty of the evidence

We assessed the certainty of the evidence using the GRADE approach (Guyatt 2011), rating each outcome as follows.

High: we are very confident that the true effect lies close to that of the estimate of the effect.
Moderate: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect.
Low: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

As all the included studies were RCTs, the evidence for each outcome started as high certainty, but could be downgraded if there were valid reasons to do so within the following five categories: risk of bias, imprecision, inconsistency, indirectness, and publication bias (Balshem 2011). We summarized the certainty of the evidence for each outcome in a 'Summary of findings’ table.

We drew on a review of entomological outcomes in the presence of insecticide resistance to inform our indirectness judgement in GRADE (Strode 2014).

Subgroup analysis and investigation of heterogeneity

We planned that if we detected substantial heterogeneity, we would perform a subgroup analysis of malaria transmission stability (stable malaria defined as an EIR of 1.0 and above, or unstable malaria defined as an EIR of less than 1.0). We additionally intended to subgroup cRCTs and individual RCTs. However, we detected substantial heterogeneity in only one meta‐analysis, and as all of the included studies were cRCTs conducted in unstable malaria areas, the subgroup analyses would not have provided any insight into the heterogeneity.

Sensitivity analysis

We intended to perform a sensitivity analysis on the primary outcome to determine the effect of exclusion of trials judged to have a serious risk of bias, but we identified no such studies.

Results

Description of studies

Results of the search

Our search of the databases identified a total of 333 new records. We considered 20 articles for full‐text screening following title and abstract screening. From these, we identified three articles, reporting three new trials, that met our inclusion criteria, and five new articles relating to trials included in the previous update. The search also returned five articles that were included and referenced in the previous review. The remaining seven trials were excluded.

We also screened the full texts of the 22 trials included in the previous version of the review against the inclusion criteria of the review update. Of these, we identified 20 trials for inclusion in the updated review. One record described four individual trials, conducted in separate regions of Latin America (Kroeger 1995 (Colombia); Kroeger 1995 (Ecuador); Kroeger 1995 (Peru Amazon); Kroeger 1995 (Peru Coast)). The study selection process is shown in Figure 1.

Included studies

Trial design and location

Of the 23 RCTs meeting the inclusion criteria, two were individually randomized. The remaining 21 trials were cRCTs. In 15 trials, the unit of randomization was the village or larger administration unit, while six trials used households as the unit of randomization. The two individual RCTs were analysed on an intention‐to‐treat basis.

Twelve trials were conducted in sub‐Saharan Africa (Burkina Faso, Cameroon, Gambia (2), Ghana, Ivory Coast, Kenya (3), Madagascar, Sierra Leone, and Tanzania). Six trials were conducted in Latin America (Colombia, Ecuador, Nicaragua, Peru (2), and Venezuela). Four trials were conducted in the Greater Mekong subregion (Cambodia, Myanmar, and Thailand (2)), and one trial was conducted in Pakistan.

The three trials new to this update were cRCTs conducted in Cambodia, Myanmar, and Venezuela.

Participants

Eleven trials included the whole population of selected areas (typically in low‐endemicity areas), while 12 trials restricted participation to specific age groups (typically children in high‐endemicity areas). Two studies were conducted specifically in displaced‐persons camps (Luxemburger 1994; Rowland 1996), and one study was restricted solely to migrant workers in the area (Kamol‐Ratanakul 1992).

Intervention

The trials examined the impact of insecticide‐treated bed nets (n = 19), treated hammock nets (n = 2), or treated curtains (n = 2). Additionally, one trial compared treated nets, treated curtains, and no bed nets or curtains (Sexton 1990).

In some trials the intervention consisted of treating existing nets with an insecticide (‘treatment of nets'), while in other trials the investigators provided treated mosquito nets or curtains to the population (‘treated nets' and ‘treated curtains').

Most nets or curtains were treated with permethrin (200 mg/m² (n = 3), 500 mg/m² (n = 9), or 1000 mg/m² (n = 1)). The remaining nets or curtains were treated with lambda‐cyhalothrin (10 to 30 mg/m²; n = 5) or deltamethrin (25 mg/m²; n = 4), while one study used lambda‐cyhalothrin (10 mg/m²) for the first year and permethrin (500 mg/m²) for the second year (Kroeger 1995 (Peru Coast)).

Approximately half of the trials used untreated nets as a control (n = 11), while the remaining trials used no net or curtain as a control (n = 12). The usage rate of the untreated nets was high (> 80%), except in one region in Peru, in which it was 63% (Kroeger 1995 (Peru Coast)), and in the Gambia (D'Alessandro 1995), in which it varied between 50% and 90% according to the area. No usage rate was provided for Rabarison 1995, Magris 2007, or Smithuis 2013.

Outcomes

Seven trials reported on our primary outcome of child mortality from all causes. Of these, six were conducted in highly malaria endemic areas in sub‐Saharan Africa, and one was in conducted in Myanmar (Smithuis 2013). Two studies reported on the incidence of severe malaria episodes. Other outcomes reported throughout the studies included the prevalence, incidence, and cumulative incidence of each of P falciparum, P vivax, and any Plasmodium species.

Excluded studies

Of the 20 full texts we screened from the literature search update, we excluded seven articles. Six corresponded to four trials that were not truly RCTs as they included only one cluster per arm. The seventh article reported a trial that had an inappropriate control group, as most participants in both groups were regularly using ITNs.

We also excluded two trials that were included in the previous version of the review after screening against this update's modified inclusion criteria. The intervention in one trial was bed nets that were not treated with insecticide (Snow 1988), and the second trial did not describe the measured outcomes clearly (Zaim 1998).

Further details are in the Characteristics of excluded studies table. Characteristics of the studies excluded after the previous literature search are described in Lengeler 2004.

Risk of bias in included studies

A detailed description of the ‘Risk of bias' assessments against the following criteria are provided in each included trial's ‘Risk of bias' table in the ‘Characteristics of included studies' section. A summary is provided in the ‘Risk of bias' summary figure (Figure 2).

‘Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Though each study described the distribution of clusters to intervention or control arms as random, in several instances the specific randomization and allocation concealment processes are not described. We considered such trials to have an unclear risk of bias. Randomization procedures, where described, typically involved a public lottery, or a computer‐generated randomization sequence.

Blinding

Due to the nature of the intervention, it is difficult to blind participants and study personnel to the allocated intervention group, and blinding was only conducted in five trials (Kamol‐Ratanakul 1992; Luxemburger 1994; Magris 2007; Rabarison 1995; Snow 1987). However, the outcomes evaluated here, that is infection, mortality, and morbidity from malaria, were considered unlikely to be affected by participant knowledge of intervention status. We therefore assumed each of the trials to be at low risk of performance bias.

The measurement of outcomes was also considered to be unaffected by intervention knowledge for mortality, severe malaria, and prevalence of malaria as collected from cross‐sectional surveys. However, we considered that outcomes through participants' self reporting of fever may be influenced by knowledge of the allocated intervention group. If self reported cases were confirmed by microscopy or a rapid diagnostic test, we considered the risk of detection bias to be unclear. In five trials, cases were recorded solely on the basis of self reporting, without further confirmation (Kroeger 1995 (Colombia); Kroeger 1995 (Ecuador); Kroeger 1995 (Peru Amazon); Kroeger 1995 (Peru Coast); Kroeger 1999). We considered these trials to have a high risk of detection bias.

Incomplete outcome data

We considered one trial to be at high risk of attrition bias, as the study participants included children aged 1 to 59 months, but in cross‐sectional surveys, only children aged 1 to 3 years were sampled (Phillips‐Howard 2003). We judged the risk of bias to be unclear for nine trials that insufficiently reported the total numbers randomized and reasons for attrition.

Selective reporting

All included trials had a low risk of reporting bias.

Other potential sources of bias

We considered three trials to have a high risk of bias due to significant imbalances between intervention and control groups at baseline for one or more reported outcomes (Fraser‐Hurt 1999; Kroeger 1995 (Colombia); Kroeger 1995 (Peru Amazon)). We judged trials that did not adequately report on baseline differences as at unclear risk of bias.

Effects of interventions

See: Table 1; Table 2

Summary of findings for the main comparison. Insecticide‐treated bed nets and curtains compared to no nets for preventing malaria.

Insecticide‐treated bed nets and curtains (ITNs) compared to no nets for preventing malaria
Patient or population: people of all ages living in malaria transmission settings  Setting: Burkina Faso 1996 (Halbluetzel 1996); Cameroon 1992 (Moyou‐Somo 1995); Cambodia 2002 (Sochantha 2006); Ghana 1995 (Binka 1996); Ivory Coast 2000 (Henry 2005); Kenya 1988 (Sexton 1990); Kenya 1995 (Nevill 1996) and 1998 (Phillips‐Howard 2003); Myanmar 1999 (Smithuis 2013); Sierra Leone 1993 (Marbiah 1998); Pakistan 1991 (Rowland 1996); Tanzania 1996 (Fraser‐Hurt 1999)  Intervention: ITNs  Comparison: no nets
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	Number of participants  (trials)	Certainty of the evidence  (GRADE)	Comments
Outcomes	Risk with no nets	Risk with ITNs	Relative effect  (95% CI)	Number of participants  (trials)	Certainty of the evidence  (GRADE)	Comments
Child mortality from all causes	Children of all ages		Rate ratio 0.83  (0.77 to 0.89)	200,833  (5 RCTs)	⊕⊕⊕⊕  HIGH^a	Insecticide‐treated bed nets and curtains reduce all‐cause child mortality compared to no nets.
	32.9 per 1000	27.3 per 1000  (25.3 to 29.3)
	Children aged 1 to 59 months
	37.8 per 1000	31.4 per 1000  (29.1 to 33.6)
Plasmodium falciparum uncomplicated episodes^g	178 per 1000	96 per 1000  (86 to 107)	Rate ratio 0.55  (0.48 to 0.64)	35,551  (5 RCTs)	⊕⊕⊕⊕  HIGH^a	Insecticide‐treated bed nets and curtains reduce the incidence of uncomplicated episodes of P falciparum malaria compared to no nets.
Plasmodium vivax uncomplicated episodes (cumulative incidence)	149 per 1000	91 per 1000  (71 to 114)	Risk ratio 0.61  (0.48 to 0.77)	10,967  (2 RCTs)	⊕⊕⊕⊝  MODERATE^a,b due to indirectness	Insecticide‐treated bed nets and curtains probably reduce the incidence of uncomplicated episodes of P vivax malaria compared to no nets.
Any Plasmodium spp. uncomplicated episodes	256 per 1000	128 per 1000  (72 to 231)	Rate ratio 0.50  (0.28 to 0.90)	8,395  (1 RCT)	⊕⊕⊝⊝  LOW^a,c,d due to indirectness	Insecticide‐treated bed nets and curtains may reduce the incidence of uncomplicated episodes of any Plasmodium species compared to no nets.
P falciparum prevalence	147 per 1000	122 per 1000  (102 to 144)	Risk ratio 0.83  (0.71 to 0.98)	18,809  (6 RCTs)	⊕⊕⊕⊕  HIGH^a	Insecticide‐treated bed nets and curtains reduce the prevalence of P falciparum malaria compared to no nets.
P vivax prevalence	130 per 1000	130 per 1000  (98 to 175)	Risk ratio 1.00  (0.75 to 1.34)	10,967  (2 RCTs)	⊕⊕⊝⊝  LOW^a,b,e due to indirectness and imprecision	Insecticide‐treated bed nets and curtains may have little or no effect on the prevalence of P vivax malaria compared to no nets.
Severe malaria episodes	15.1 per 1000	8.5 per 1000  (5.7 to 12.4)	Rate ratio 0.56  (0.38 to 0.82)	31,173  (2 RCTs)	⊕⊕⊕⊕  HIGH^a	Insecticide‐treated bed nets and curtains reduce the incidence of severe malaria episodes compared to no nets.
Anaemia (mean packed cell volume)	31.4	32.7 (31.8 to 33.6)	Mean difference 1.29 (0.42 to 2.16)	11,489 (5 RCTs)	⊕⊕⊕⊕  HIGH^a,f	Insecticide‐treated bed nets and curtains increase the mean packed cell volume compared to no nets.
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). Abbreviations: CI: confidence interval; RCT: randomized controlled trial; ITN: Insecticide‐treated net
GRADE Working Group grades of evidence  High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.  Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Date	Event	Description
6 November 2018	New citation required but conclusions have not changed	We excluded the outcomes of splenomegaly, high parasitaemia, and anthropometric measures, which were considered in the previous review (Lengeler 2004). In this update we extracted mortality data for children of all ages who participated in the included studies. However, we also calculated the baseline risk in the high‐risk age group by extracting the number of deaths and total number of children aged 1 to 59 months. From this we estimated the impact of ITNs on mortality in the high‐risk age group. For the secondary outcomes, summary risk and rate ratios were previously presented without CIs, as cluster‐adjusted CIs were not available for all trials. If an included cRCT did not adjust for clustering, we adjusted the data using an imputed design effect. The cRCTs were then meta‐analysed, and cluster‐adjusted CIs for each outcome were provided.
6 November 2018	New search has been performed	This is an update of the Lengeler 2004 review. We performed a literature search update, and included three articles, reporting three new trials, and five new articles relating to trials included in the previous update. We excluded two trials included in the previous version of the review. We assessed the certainty of the evidence using the GRADE approach. We assessed the risk of bias for each included study in this update using Cochrane's ‘Risk of bias' tool; for each included cRCT, we also assessed five additional criteria relating specifically to cRCTs.
18 August 2008	Amended	Converted to new review format with minor editing.
19 January 2004	New citation required but conclusions have not changed	Issue 2, 2004  This is a major update with a revision of the text, tables, and results.  ‐ An additional 16 trials have been identified and reviewed, of which 4 were included.  ‐ The sensitivity analysis (with group 2 trials) has been removed to clarify the main results.  ‐ The literature in all sections and especially background and discussion has been updated.  ‐ Overall mortality results have been entered with the reverse variance function in order to present confidence intervals adjusted for clustering.
12 January 2004	New search has been performed	Minor update.
23 October 2003	New search has been performed	New studies sought but none found.
21 January 2003	New search has been performed	New studies found and included or excluded.

Study designs	Individual randomized controlled trials (RCTs) and cluster RCTs
Participants	Children and adults living in rural and urban malarious areas Excluded: trials examining only pregnant women and trials examining only soldiers or travellers, because they are not representative of the general population
Interventions	Bed nets or curtains treated with a synthetic pyrethroid insecticide at a minimum target impregnation dose of: 200 mg/m² permethrin or etofenprox; 30 mg/m² cyfluthrin; 20 mg/m² alpha‐cypermethrin; 10 mg/m² deltamethrin/lambda‐cyhalothrin. No distinction was made between insecticide‐treated bed nets and door/window/eave/wall curtains, which were assumed to have approximately the same impact.
Comparators	Untreated net or no net
Outcomes	Child mortality from all causes: measured using protective efficacy and rate difference. Malaria‐specific child mortality: measured using "verbal autopsy" reports that fulfil standard clinical criteria for a probable malaria death (Snow 1992; Todd 1994). Severe disease: measured using site‐specific definitions based on World Health Organization guidelines, WHO 1990, and on Marsh 1995. The definition included Plasmodium falciparum parasitaemia. Cerebral malaria was defined as coma or prostration and/or multiple seizures. The cut‐off for severe, life‐threatening anaemia was set at 5.1 g/L (WHO 1990). Uncomplicated clinical episodes: measured using site‐specific definitions, including measured or reported fever, with or without parasitological confirmation. Measurements were usually done in the frame of prospective longitudinal studies, but we also considered trials using validated retrospective assessments in the frame of cross‐sectional surveys. In areas with entomological inoculation rates below 1 (unstable malaria), we considered P falciparum and P vivax episodes separately. Parasite prevalence: parasite prevalence due to P falciparum and P vivax was obtained using the site‐specific method for estimating parasitaemia, that is usually thick or thin blood smears or both. When more than one survey was done, the reported prevalence result is the average prevalence of all the surveys.

Search	Query
#16	Search (#11) AND #15
#15	Search (#12) OR #13 OR #14 Filters: Publication date from 2003/01/01
#14	Search "drug therapy" [Subheading]
#13	Search "Randomized Controlled Trial" [Publication Type] OR "Controlled Clinical Trial" [Publication Type]
#12	Search randomized or placebo or randomly or trial or groups Field: Title/Abstract
#11	Search (#9) AND #10 Filters: Publication date from 2003/01/01;
#10	Search (#6) OR #7 OR #8 Publication date from 2003/01/01;
#9	Search (#1) OR #2 OR #3 OR #4 OR #5 Filters: Publication date from 2003/01/01;
#8	Search Olyset* or PermaNet* or Veeralin Field: Title/Abstract
#7	Search bednet* or net* or ITN* or LLIN* or curtain* or "insecticide‐treated net" or "insecticide‐treated bednet" Field: Title/Abstract
#6	Search ("Insecticide‐Treated Bednets"[Mesh]) OR "Mosquito Nets"[Mesh]
#5	Search malaria Field: Title/Abstract
#4	Search "Plasmodium"[Mesh]
#3	Search "Malaria"[Mesh]
#2	Search mosquito Field: Title/Abstract
#1	Search "Anopheles"[Mesh]

Protocol section	Refreshed protocol
Background	We updated information in the background to reflect the changes in global malaria distribution and its control since the previous update. We included further information on insecticide resistance and the need to consider this when evaluating the effectiveness of ITNs.
Methods	The primary objective of the review, to assess the impact of ITNs on mortality and malarial illness, remains relevant. The original PICO remains relevant. In the previous review, mortality data was age‐standardized across each included study by extracting only the number of deaths and total number of participants within a high‐risk age group of 1 to 59 months. To avoid the exclusion of potentially useful information, we planned to extract mortality data for children of all ages who participated in the included studies. However, we also planned to calculate the baseline risk in the high‐risk age group by extracting the number of deaths and total number of children aged 1 to 59 months. From this we estimated the impact of ITNs on mortality in the high‐risk age group. For the secondary outcomes, summary risk and rate ratios were previously presented without CIs, as cluster‐adjusted CIs were not available for all trials. If an included cRCT did not adjust for clustering, we planned to adjust the data using an imputed design effect. The cRCTs were then meta‐analysed, and cluster‐adjusted CIs for each outcome were provided. We excluded the outcomes of splenomegaly, high parasitaemia, and anthropometric measures, which were considered in the previous review (Lengeler 2004), as we did not consider them priority outcomes at the time of this update. We updated our approach to assessing risk of bias, and used the standardized Cochrane's ‘Risk of bias' tool. For each included cRCT we also assessed five additional criteria relating specifically to cRCTs. We included all relevant trials in the meta‐analysis regardless of the local area's transmission intensity, and intended to subgroup between stable and unstable transmission only if we identified substantial heterogeneity. We assessed the certainty of the evidence using the GRADE approach.

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Child mortality from all causes	5	Rate Ratio (Random, 95% CI)	0.83 [0.77, 0.89]
2 P falciparum uncomplicated episodes	5	Rate Ratio (Random, 95% CI)	0.55 [0.48, 0.64]
3 P falciparum uncomplicated episodes (cumulative incidence)	2	Risk Ratio (Random, 95% CI)	0.44 [0.31, 0.62]
4 P vivax uncomplicated episodes (cumulative incidence)	2	Risk Ratio (Random, 95% CI)	0.61 [0.48, 0.77]
5 Any Plasmodium spp. uncomplicated episodes	1	Rate Ratio (Fixed, 95% CI)	0.50 [0.28, 0.90]
6 P falciparum prevalence	6	Risk Ratio (Random, 95% CI)	0.83 [0.71, 0.98]
7 P vivax prevalence	2	Risk Ratio (Random, 95% CI)	1.00 [0.75, 1.34]
8 Severe malaria episodes	2	Rate Ratio (Random, 95% CI)	0.56 [0.38, 0.82]
9 Anaemia (mean packed cell volume)	5	Mean Difference (Random, 95% CI)	1.29 [0.42, 2.16]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Child mortality from all causes	2	Rate Ratio (Random, 95% CI)	0.67 [0.36, 1.23]
2 P falciparum uncomplicated episodes	5	Rate Ratio (Random, 95% CI)	0.58 [0.44, 0.78]
3 P vivax uncomplicated episodes	3	Rate Ratio (Random, 95% CI)	0.73 [0.51, 1.05]
4 P vivax uncomplicated episodes (cumulative incidence)	3	Risk Ratio (Random, 95% CI)	0.59 [0.30, 1.18]
5 Any Plasmodium spp. uncomplicated episodes (cumulative incidence)	2	Risk Ratio (Random, 95% CI)	0.47 [0.17, 1.28]
6 P falciparum prevalence	3	Risk Ratio (Random, 95% CI)	0.91 [0.78, 1.05]
7 P vivax prevalence	1	Risk Ratio (Fixed, 95% CI)	0.68 [0.25, 1.85]
8 Any Plasmodium spp. prevalence	1	Risk Ratio (Fixed, 95% CI)	0.17 [0.05, 0.53]
9 Anaemia (mean packed cell volume)	2	Mean Difference (Random, 95% CI)	0.48 [‐0.54, 1.50]

Methods	Study design: cluster RCT (cRCT)  Unit of allocation: clusters of compounds (average 120 compounds and 1400 people/cluster)  Number of units: 48:48  Length of follow‐up: 2 years (July 1993 to June 1995).  Outcome assessment: mortality was monitored by village reporters in addition to demographic data collected every 3 months by rolling census.  Adjustment: confidence limits for the rate ratio were calculated taking into account the cluster randomization.
Participants	Number of participants: approximately 134,400  Inclusion criteria: children < 10 years
Interventions	Intervention: bed net  Insecticide and dosage: permethrin suspension (0.5 g/mL)  Retreatment: every 6 months  Usage: year 1: 97% July‐Dec, 65% Jan‐June; year 2: 72% July‐Dec, 50% Jan‐June  Control: no net
Outcomes	Outcomes measured: mortality rate
Notes	Study location: Kassena‐Nankana district, Ghana EIR: 100 to 1000 Malaria transmission: variable but high Main vectors: Anopheles gambiae s.s.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Open lotteries were conducted during 21 community meetings to randomly select the clusters that were to receive the impregnated bed nets.
Allocation concealment (selection bias)	Low risk	Open lottery conducted.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Not possible to blind participants and personnel, but this was not likely to introduce bias to the outcome of mortality
Blinding of outcome assessment (detection bias)   Self‐reported fever	Low risk	This outcome was not assessed.
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	Not possible to blind outcome assessors, but this was not likely to introduce bias when measuring the outcome of mortality
Incomplete outcome data (attrition bias)   All outcomes	Low risk	High estimated sensitivity to all deaths in the study area
Selective reporting (reporting bias)	Low risk	All expected outcomes reported.
Other bias	Low risk	Recruitment bias: low risk Baseline imbalance: the pre‐intervention mortality rates were comparable (23.0 and 23.5/1000 child years in the treated and control clusters respectively). Loss of clusters: none Incorrect analysis: adjusted Comparability with RCTs randomizing individuals: suitable due to cluster‐adjusted confidence intervals (CIs)

Methods	Study design: cRCT Unit of allocation: village (52 pairs of villages formed on the basis of size, after stratification by 5 geographical areas) Number of units: 58:52 Length of follow‐up: 12 months Dropout rate unknown, but immigration/emigration rates were low (< 5% per year). Mortality monitored by village reporters and yearly census. Morbidity surveys were conducted once, at the peak of the transmission season in October (n = 1520 in 50 villages). All surveys were community‐based.
Participants	Inclusion criteria: children aged 0 to 9 years and living in the area were eligible at the start, but the analysis was later restricted to children aged 1 to 59 months (n = 25,000). Exclusion criteria: no explicit exclusion criteria except absence of written consent
Interventions	Intervention: treatment of existing bed nets in the frame of a national programme; target dose 200 mg/m² permethrin; impregnation done by village health workers with the assistance of other community members and under the supervision of community health nurses; retreatment was not done during the 1‐year follow‐up period since the transmission season lasts only about 4 months. Control: untreated bed nets Usage rate around 70% in both intervention and control areas (varied between 50% and 90% according to the area).
Outcomes	Overall mortality (1 to 59 months) Prevalence of parasitaemia (any) Prevalence of high parasitaemia (> 5000 trophozoites/µL) Anaemia (mean packed cell volume) Prevalence of splenomegaly (1 to 5 Hackett) Impact on nutritional status (weight‐for‐age, weight‐for‐height)
Notes	Study location: 5 distinct areas spread over the whole of the Gambia (all rural areas) EIR: 1 to 10 Malaria endemicity: hyperendemic Baseline parasite rate in children 12 to 59 months: 39% Main vector: Anopheles gambiae s.l. P vivax malaria: very low; not taken into account for analysis Access to health care moderately easy.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Intervention allocation by public lottery (information provided by CL)
Allocation concealment (selection bias)	Low risk	Low risk given the above intervention allocation procedure
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	No blinding described, but the review authors judge that the outcomes of mortality and malaria infection were unlikely to be influenced by the lack of blinding to participants and personnel
Blinding of outcome assessment (detection bias)   Self‐reported fever	Low risk	This outcome was not assessed.
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	No blinding of outcome assessors described, but this was unlikely to influence the outcome measurement for mortality or parasite prevalence
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No missing outcome data
Selective reporting (reporting bias)	Low risk	All expected outcomes reported.
Other bias	Low risk	Recruitment bias: low risk Baseline imbalance: the pre‐intervention mortality rates were adjusted for in the analysis. Loss of clusters: none Incorrect analysis: adjusted Comparability with RCTs randomizing individuals: unclear

Methods	Study design: individual RCT  Unit of allocation: individual  Number of units: 122  Length of follow‐up: 6 months  Outcome assessment: monthly cross‐sectional surveys were conducted. Thick blood films were prepared at enrolment and at each survey.
Participants	Number of participants: 122  Inclusion criteria: children aged 5 to 24 months who were afebrile, not using a bed net, and not taking chloroquine
Interventions	Type of intervention: bed net (n = 61)  Insecticide and dosage: permethrin (500 mg/m²)  Retreatment: after 3 months and at the end of the trial  Usage: 97%  Control: no net (n = 61)
Outcomes	Outcomes measured: P falciparum prevalence
Notes	Study location: Kiberege, Kilombero District, southern Tanzania EIR: approximately 300 Malaria transmission: intense and perennial  Main vectors: Anopheles gambiae s.l. and Anopheles funestus
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Individuals allocated randomly, but the randomization process is not described.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participants and personnel were not blinded to intervention group, but the review authors judge that this was unlikely to impact on the outcome of prevalence.
Blinding of outcome assessment (detection bias)   Self‐reported fever	Low risk	This outcome is not assessed.
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	Outcome assessors not blinded, but all participants were surveyed using objective blood smear examination, so this was unlikely to introduce bias.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Missing outcome data very minimal and balanced in numbers across the intervention groups (1 from each).
Selective reporting (reporting bias)	Low risk	All expected outcomes reported.
Other bias	High risk	Recruitment bias: low risk Baseline imbalance: the pre‐intervention prevalence was substantially different between the intervention group (54.1%) and control group (65%).

Methods	Study design: cRCT Unit of allocation: groups of villages (8 pairs of "clusters" (on average 10 villages) formed on the basis of baseline mortality and geographic similarity) Number of units: 8:8 Length of follow‐up: 24 months Mortality was monitored by village reporters and yearly census. A cross‐sectional morbidity survey was conducted once, at the peak of the transmission season in September 1995 (n = 800 in 84 villages). All surveys were community‐based.
Participants	Number of participants: 16,540 Inclusion criteria: children aged 0 to 59 months living in the area (newborns were excluded from the analysis) Exclusion criteria: no explicit exclusion criteria except absence of written consent
Interventions	Intervention: permethrin‐treated curtains on windows, door, and eaves; target dose of 1000 mg/m²; every house used for sleeping in the intervention clusters fitted with the curtains and retreated every 6 months Control: no curtains
Outcomes	Overall mortality (1 to 59 months) Prevalence of parasitaemia (any) Prevalence of high parasitaemia (> 5000 trophozoites/µL) Anaemia (mean haemoglobin in g/dL)
Notes	Study location: Oubritenga Province, 30 km north of Ouagadougou, Burkina Faso, in a rural area EIR: 300 to 500 Malaria endemicity: holoendemic Baseline parasite rate in children aged 6 to 59 months: 85% Main vectors: Anopheles gambiae s.l. and Anopheles funestus P vivax malaria: 0% Dropout rate unknown, but immigration/emigration rates were low (2% per year). Access to health care considered poor.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Intervention allocation by public lottery (information provided by CL)
Allocation concealment (selection bias)	Low risk	Low risk given the above intervention allocation procedure
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Not possible to blind participants and personnel, but the review authors do not consider this likely to introduce bias for the outcome of mortality
Blinding of outcome assessment (detection bias)   Self‐reported fever	Low risk	This outcome was not assessed.
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	No blinding, but the review authors judge that the outcome and the outcome measurement were not likely to be influenced by lack of blinding
Incomplete outcome data (attrition bias)   All outcomes	Low risk	High estimated sensitivity to all deaths in the study area
Selective reporting (reporting bias)	Low risk	All expected outcomes reported.
Other bias	Low risk	Recruitment bias: low risk Baseline imbalance: the pre‐intervention mortality rates were adjusted for. Loss of clusters: none Incorrect analysis: adjusted Comparability with RCTs randomizing individuals: suitable with adjusted CIs

Methods	Study design: cRCT  Unit of allocation: village (allocation of 8 villages by paired randomization)  Number of units: 4:4  Length of follow‐up: 12 months  Outcome assessment: active case surveillance by repeated cross‐sectional surveys. Blood smear taken from every sick child, as assessed by nurse visit. Generalized estimating equation (GEE) applied for clustered data.
Participants	Number of participants: 426  Inclusion criteria: children aged 0 to 59 months
Interventions	Intervention: bed net (n = 210)  Insecticide and dosage: lambda‐cyhalothrin formulated as a capsule suspension (15 mg a.i./m²)  Retreatment: nets dipped again after 6 months.  Usage: average coverage rate with ITNs ranged from 76.7% to 84.0% in different villages.  Control: no net (n = 216)
Outcomes	Outcomes measured: cumulative incidence of uncomplicated episodes
Notes	Study location: Korhogo, in the North of Côte d’Ivoire. High knockdown resistance (kdr) resistance EIR: 55 Malaria transmission: baseline prevalence of 82% in 1997 before implementation Main vectors: Anopheles gambiae s.s. and Anopheles funestus. A gambiae was strongly resistant to pyrethroids with a kdr allelic frequency of around 90%. A funestus was still susceptible to these insecticides.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	It is stated that the village chosen from each matched pair to receive ITN was randomly selected, but no details provided.
Allocation concealment (selection bias)	Unclear risk	No information provided.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Not possible to blind participants and personnel, but this was not likely to introduce bias to the outcome of malaria infection
Blinding of outcome assessment (detection bias)   Self‐reported fever	Unclear risk	Nurse examining and recording cases of sickness at the home would have been aware of study group being visited. Unclear if technicians examining blood slides were blinded to the study group
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	These outcomes were not assessed.
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Missing outcome data were balanced across the intervention groups. Reasons for missing data given, but only as a total and not for each intervention group, possibly masking biases (58 total deaths). An imbalance between the groups may have introduced bias to the examined outcomes.
Selective reporting (reporting bias)	Low risk	All stated outcomes reported.
Other bias	Low risk	Recruitment bias: low risk Baseline imbalance: baseline figures comparable for prevalence (P = 0.35) and incidence (P = 0.36). Loss of clusters: none Incorrect analysis: adjusted  Comparability with RCTs randomizing individuals: suitable for comparison

Methods	Study design: cRCT Unit of allocation: households Number of units: 91 households at baseline (46 intervention households, 45 controls). 78 households at the end of the study (39 intervention households, 39 controls) Length of follow‐up (study dates): 2 distinct periods for a total of 13 months (February 1993 to July 1993, and October 1993 to June 1994) Outcome assessment: passive surveillance through medical consultations at Ankazobé (Institut Pasteur)
Participants	Number of participants: 501 at baseline (244 in the intervention group, 257 in the control group). 431 at the end of the study (208 in the intervention group, 223 in the control group)  Inclusion criteria: households in Ankazobé district II
Interventions	Intervention: curtain nets attached to the doors and windows of bedrooms  Insecticide and dosage: deltamethrin (25 mg/m²)  Net retreatment: done 3 times (specific dates not provided)  Usage: not described. Likely to be high as nets are fitted by study personnel and then left  Control: untreated nets attached to the doors and windows of bedrooms
Outcomes	Outcomes measured: number of clinical episodes (defined as axillary body temperature over 37.5 C° plus parasitaemia > 1500/ μL)
Notes	Study location: Ankazobé (Madagascar), at altitude of 1300 m EIR: < 10 Malaria transmission (perennial, seasonal, etc.): low seasonal transmission. “Stability index” described as <= 2.5. Main vector species: Anopheles funestus and Anopheles gambiae % P vivax cases: 0%
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Houses were drawn by lottery.
Allocation concealment (selection bias)	Low risk	Low risk of bias considering the above allocation procedure
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Single‐blind: participants were blinded to the impregnation status of installed curtains
Blinding of outcome assessment (detection bias)   Self‐reported fever	Low risk	Passive case detection would have depended on self reporting, but as participants were blinded to intervention this was unlikely to have introduced detection bias. Cases were registered only if confirmed positive for parasitaemia.
Blinding of outcome assessment (detection bias)   All other outcomes	Low risk	Other outcomes were not assessed.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Loss of households was explained as being mainly due to participants relocating following cyclones in February 1994. Loss was balanced between groups.
Selective reporting (reporting bias)	Low risk	Expected outcomes are reported, except for data for August 1993 to September 1993, which is likely due to limited cases outside of the rain season.
Other bias	Unclear risk	Recruitment bias: low risk Baseline imbalance: unclear risk. Baseline prevalence not reported. Loss of clusters: 13 clusters lost due to participants relocating following cyclones. These were balanced in numbers between groups. Incorrect analysis: no adjustment to CIs to account for clustering Comparability with RCTs randomizing individuals: not comparable unless clustering is adjusted for by review authors

Study	Reason for exclusion
Bhatt 2012	The study design is not appropriate as it grouped villages into just three clusters. A single cluster was randomised to each of the ITN, UTN or NN arms.
Sahu 2008	The study design is not appropriate as it grouped villages into just three clusters. A single cluster was randomised to each of the ITN, UTN or NN arms.
Sharma 2009	The study design is not appropriate as it grouped villages into just three clusters. A single cluster was randomised to each of the ITN, UTN or NN arms.
Snow 1988	Comparison was between untreated nets and no nets. No participants received an ITN.
Soleimani‐Ahmadi 2012	The study design is not appropriate as it grouped villages into just two clusters. A single cluster was randomised to each of the ITN and UTN arms.
Thang 2009	The control group is not appropriate. The study assessed the impact of introducing insecticide treated hammocks to forest workers, compared to those not receiving hammocks. However, the participants all lived in villages with a high coverage of ITNs (88.17% in the control arm), and therefore were unsuitable to act as a control.
Zaim 1998	Definition of measured outcome described as "incidence of malaria" is unclear. No other epidemiological outcomes were reported.

PERMALINK

Insecticide‐treated nets for preventing malaria

Joseph Pryce

Marty Richardson

Christian Lengeler

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

1.

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Certainty of the evidence

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

Included studies

Trial design and location

Participants

Intervention

Outcomes

Excluded studies

Risk of bias in included studies

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Insecticide‐treated bed nets and curtains compared to no nets for preventing malaria.

Summary of findings 2. Insecticide‐treated bed nets and curtains compared to untreated nets for preventing malaria.

Comparison 1: Insecticide‐treated nets versus no nets

Child mortality from all causes

1.1. Analysis.

Uncomplicated clinical episodes

1.2. Analysis.

1.3. Analysis.

1.4. Analysis.

1.5. Analysis.

Prevalence

1.6. Analysis.

1.7. Analysis.

Severe malaria episodes

1.8. Analysis.

Anaemia

1.9. Analysis.

Comparison 2: Insecticide‐treated nets versus untreated nets

Child mortality from all causes

2.1. Analysis.

Uncomplicated clinical episodes