Intervention Review

You have free access to this content

Mosquito larval source management for controlling malaria

  1. Lucy S Tusting1,
  2. Julie Thwing2,*,
  3. David Sinclair3,
  4. Ulrike Fillinger1,
  5. John Gimnig4,
  6. Kimberly E Bonner5,
  7. Christian Bottomley6,
  8. Steven W Lindsay1,7

Editorial Group: Cochrane Infectious Diseases Group

Published Online: 29 AUG 2013

Assessed as up-to-date: 24 OCT 2012

DOI: 10.1002/14651858.CD008923.pub2


How to Cite

Tusting LS, Thwing J, Sinclair D, Fillinger U, Gimnig J, Bonner KE, Bottomley C, Lindsay SW. Mosquito larval source management for controlling malaria. Cochrane Database of Systematic Reviews 2013, Issue 8. Art. No.: CD008923. DOI: 10.1002/14651858.CD008923.pub2.

Author Information

  1. 1

    London School of Hygiene and Tropical Medicine, Department of Disease Control, London, UK

  2. 2

    US Centers for Disease Control and Prevention (CDC), Strategic and Applied Science Unit, Malaria Branch, Atlanta, Georgia, USA

  3. 3

    Liverpool School of Tropical Medicine, Department of Clinical Sciences, Liverpool, UK

  4. 4

    US Centers for Disease Control and Prevention (CDC), Entomology Branch, Atlanta, Georgia, USA

  5. 5

    Princeton University, Woodrow Wilson School of Public and International Affairs, Princeton, New Jersey, USA

  6. 6

    London School of Hygiene and Tropical Medicine, MRC Tropical Epidemiology Group, London, UK

  7. 7

    Durham University, School of Biological and Biomedical Sciences, Durham, UK

*Julie Thwing, Strategic and Applied Science Unit, Malaria Branch, US Centers for Disease Control and Prevention (CDC), 4770 Buford Highway, NE, Mailstop F-22, Atlanta, Georgia, GA 30341, USA. juliethwing@gmail.com. fez3@cdc.gov.

Publication History

  1. Publication Status: New
  2. Published Online: 29 AUG 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Balfour 1936 GRC

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Town and rural areas

Cluster size: Population of towns: 1700; 1130; 830; 32,200; 31,550 individuals

Number of clusters in each arm: Intervention arm: two; control arm: three

Adjusted for clustering? No


ParticipantsAge: School children

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Parasite prevalence): 210, 112, 97, 853, 650 participants per survey

Secondary outcome sample size (Splenomegaly prevalence): 210, 112, 97, 853, 650 participants per survey


InterventionsIntervention: Habitat modification with larviciding

Details of the intervention:                                           

Habitat modification: Drainage and reclamation of marshland, straightening of rivers and construction of embankments                                                    

Larviciding: Larval habitats were treated with Paris Green (dosage not stated)

Frequency of application: Not stated

Duration of intervention period: 60 months

Who was responsible for LSM? The government

Co-interventions: Case management: treatment with quinine (coverage not stated)

Co-interventions equal in each arm? Not stated


Outcomes1. Parasite prevalence (measured with yearly cross-sectional surveys)

2. Splenomegaly prevalence (measured with yearly cross-sectional surveys)


NotesContinent: Europe

Country: Greece

Ecosystem: Coastal

Urban or rural: Urban and rural

Extensive or localized larval habitats: Localized

Primary larval habitats: Primarily man-made habitats

Transmission intensity: Low to moderate

Transmission season(s): May to October

Primary and secondary vector: An. elutus, An. superpictus

Primary malaria parasite: P. falciparum, P. vivax

Source of funding: Not stated


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
High riskImpossible to blind evaluators to intervention.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskReporting ceased from one clinic. Individual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskOutcome reporting complete.

Baseline characteristicsLow riskBaseline characteristics reported.

ContaminationUnclear riskNot stated how far apart the towns were.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Castro 2009 TZA

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Area of city (area around large drain)

Cluster size: Unclear

Number of clusters in each arm: Intervention arm: four; control arm: two

Adjusted for clustering?  No


ParticipantsAge: Any

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Parasite prevalence): 1162, 1513, 1991, 1793, 1711, 900 participants in the surveys


InterventionsIntervention: Habitat manipulation with larviciding

Details of the intervention:                                                    

Habitat manipulation: Drains in the city were cleared to increase the water flow and to reduce flooding in the rainy season. Minor repairs such as slab replacement were conducted                     

Larviciding: In half the intervention neighbourhoods, larval habitats were treated with larvicide by the Urban Malaria Control Progam (details not given)

Frequency of application: Not stated

Duration of intervention period: Not stated

Who was responsible for LSM? Drain clearance was initially conducted by a contractor with 90% of the workforce local. Intensive education of the local community led to community-led maintenance of drains. Larviciding was organized by the Urban Malaria Control Program.

Co-interventions: None. However ITNs are used in the study area (coverage not stated).

Co-interventions equal in each arm? Not stated


Outcomes1. Parasite prevalence (measured with six cross-sectional surveys (one every two months)


NotesContinent: Africa

Country: Tanzania

Ecosystem: Coastal

Urban or rural: Urban

Extensive or localized larval habitats: Localized

Primary larval habitats: Drains

Transmission intensity: Low to moderate

Transmission season(s): March to June, October to December

Primary and secondary vector: An. gambiae, An. funestus

Primary malaria parasite: P. falciparum

Source of funding: Japan International Cooperation Agency 


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskEnvironmental management sites purposefully chosen according to stated criteria.

Allocation concealment (selection bias)High riskSites purposefully selected.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskParasite prevalence assessed by blinded reading of blood slides collected from randomly selected participants.

Blinding of participants and personnel (performance bias)
All outcomes
High riskNo way to blind participants and personnel to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)High riskOutcomes reported as per methods, however little detail pertaining to the data is reported.

Baseline characteristicsUnclear riskStated to be similar, but not specified.

ContaminationHigh riskIn one EM cluster, drain not maintained; distances of clusters from one another not reported.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Coulibaly 2011 MLI

MethodsTrial design: Cluster-RCT

Type of cluster: Village

Cluster size: Not stated

Number of clusters in each arm: Three

Adjusted for clustering? No


ParticipantsAge: n/a

Sex: n/a

Co-morbidities and pregnancy: n/a

Primary outcome sample size (EIR): 12 sentinel houses per village

Secondary outcome sample size (Adult mosquito density (measures other than human biting rate)): 12 sentinel houses per village


InterventionsIntervention: Larviciding

Details of the intervention:                                   

Larviciding: Larval habitats were treated with Bti (Vectobac®, applied at 400g/ha using a sprayer) and Bs (VectoLex®, dosage not stated)

Frequency of application: Larviciding with Bti: weekly; larviciding with Bs: every two weeks

Duration of intervention period: 18 months

Who was responsible for LSM? Malaria Research and Training Center staff and selected members of the community were trained to conduct larviciding. The local community was educated about the importance of larviciding.

Co-interventions: IRS: two rounds of district-wide were conducted, covering all study villages in July to August 2008 and June to July 2009 (coverage not stated)

Co-interventions equal in each arm? Not stated


Outcomes1. EIR (measured with monthly pyrethrum spray collections in sentinel houses)

2. Adult mosquito density (measured with monthly pyrethrum spray collections in sentinel houses)


NotesContinent: Africa

Country: Mali

Ecosystem: Savannah

Urban or rural: Rural

Extensive or localized larval habitats: Localized

Primary larval habitats: Brick pits, ponds, tyre prints

Transmission intensity: High

Transmission season(s): June to October

Primary and secondary vector: An. gambiae

Primary malaria parasite: P. falciparum

Source of funding: Malaria Research and Training Center, University of Bamako; Research Triangle International; National Institues of Health; Centers for Disease Control; United States Agency for International Development; United States President's Malaria Initiative


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskVillages randomly assigned; however method of randomization not specified.

Allocation concealment (selection bias)Unclear riskNot described.

Blinding of outcome assessment (detection bias)
All outcomes
High riskImpossible to blind entomologic data collection.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskStated outcomes reported.

Baseline characteristicsUnclear riskBaseline characteristics not stated, though villages chosen from same health district.

ContaminationLow riskVillages a sufficient distance apart.

Incorrect analysisHigh riskNot adjusted for clustering.

Other biasLow riskLow risk of confounding.

Fillinger 2008 TZA

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Area of city (ward)

Cluster size: 0.96 to 15km2

Number of clusters in each arm: Intervention arm: three; control arm: 12

Adjusted for clustering?  No


ParticipantsAge: n/a

Sex: n/a

Co-morbidities and pregnancy: n/a

Primary outcome sample size (EIR): 67 sentinel sites


InterventionsIntervention: Larviciding alone

Details of the intervention:                              

Larviciding: Open (light-exposed) larval habitats were treated with Bti water-dispersible granules (VectoBac®, applied at 0.04g/m2 using knapsack sprayers), Bs water-dispersible granules (VectoLex®, applied at 0.2g/m2 using knapsack sprayers), Bti corn granule formulations (VectoBac®, applied at 1g/m2 by hand) and Bs corn granule formulations (VectoLex®, applied at 3g/m2 by hand). Closed habitats (the main larval habitat of Culex quinquefaciatus, a nuisance-biting mosquito) were treated with Bs corn cob granules (VectoLex®, applied at 1g/m2 by hand).

Frequency of application: Larviciding of open habitats: weekly; closed habitats: every three months.

Duration of intervention period: 15 months

Who was responsible for LSM? Open habitats were treated by modestly paid members of the community, Mosquito Contro CORPs, each of which was assigned to a specific area (mtaa). An additional team of CORPs was responsible for treating closed habitats. CORPs reported to the Ward Office.

Co-interventions: None. However ITNs were used in the study area (coverage not stated).

Co-interventions equal in each arm? Not stated


Outcomes1. EIR (measured with weekly CDC light trap catches and pyrethrum spray catches)

2. Adult mosquito density (human biting rate) (measured with weekly CDC light trap catches and pyrethrum spray catches)


NotesContinent: Africa

Country: Tanzania

Ecosystem: Coastal

Urban or rural: Urban

Extensive or localized larval habitats: Localized

Primary larval habitats: Man-made habitats exposed to sunlight

Transmission intensity: Low to moderate

Transmission season(s): March to June (primary), October to December (secondary)

Primary and secondary vector: An. gambiae s.s., An. arabiensis

Primary malaria parasite: P. falciparum

Source of funding: Swiss Tropical Institute, the United States Agency for International Development (Environmental Health Project, Dar es Salaam Mission and the United States President's Malaria Initiative), the Bill and Melinda Gates Foundation, Valent BioSciences Corporation, Wellcome Trust.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
High riskImpossible to blind entomologic data collection.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskComplete outcome reporting.

Baseline characteristicsLow riskBaseline mosquito densities reported.

ContaminationHigh riskControl and intervention clusters are adjacent.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Fillinger 2009 KEN

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Highland valley villages

Cluster size: Between 107 and 214 individuals in each group (2-4km sq)

Number of clusters in each arm: Three

Adjusted for clustering? No


ParticipantsAge: 6 months to 10 years

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Malaria incidence): 720 participants

Secondary outcome sample size (EIR): 10 sentinel sites per valley


InterventionsIntervention: Larviciding alone

Details of the intervention:                      

Larviciding: Larval habitats were treated with Bs water-dispersible and corn granules (VectoLex®) during months one to six, then Bti water-dispersible and corn granules (VectoBac®) during months seven to 19.

Frequency of application: Weekly

Duration of intervention period: 19 months

Who was responsible for LSM? Study staff

Co-interventions: ITNs (coverage: intervention arm: 25% to 51%; non-intervention arm: 24% to 51%).

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured by three cross-sectional surveys in the pre-intervention period, and three cross-sectional surveys in the post-intervention period, two months apart, using rapid malaria tests and microscopy)

2. EIR (measured by monthly indoor resting collection (pyrethrum spray collection) at sentinel sites)

3. Adult mosquito density (human biting rate) (measured by monthly indoor resting collection (pyrethrum spray collection) at sentinel sites)

4. Adult mosquito density (measures other than human biting rate (measured by monthly indoor resting collection (pyrethrum spray collection) at sentinel sites)


NotesContinent: Africa

Country: Kenya

Ecosystem: Highland

Urban or rural: Rural

Extensive or localized larval habitats: Localized and extensive

Primary larval habitats: Small streams, papyrus swamps

Transmission intensity: Moderate

Transmission season(s): April to June, November to January

Primary and secondary vector: An. gambiae s.l., An. funestus s.l.

Primary malaria parasite: P. falciparum

Source of funding: Environmental Health Project of the United States Agency for International Development


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskMalaria incidence determined by blinded reading of blood smears.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskComplete outcome reporting.

Baseline characteristicsLow riskBaseline characteristics reported and similar.

ContaminationLow riskClusters at least 1 km apart.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Geissbühler 2009 TZA

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Ward

Cluster size: Total study population of 4761

Number of clusters in each arm: Intervention arm: three; control arm: 12

Adjusted for clustering? No


ParticipantsAge: 0 to five years

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Parasite prevalence): 4450 participants

Secondary outcome sample size (EIR): 268 sentinel sites (4 sites in each of 67 mitaa)


InterventionsIntervention: Larviciding

Details of the intervention:                                                     

Larviciding: Open (light-exposed) larval habitats were treated with Bti water-dispersible granules (VectoBac®, applied at 0.04g/m2 using knapsack sprayers) and Bti corn granules (VectoBac®, applied at 1 g/m2 by hand). Closed habitats (the main larval habitat of Culex quinquefaciatus, a nuisance-biting mosquito) were treated with Bs corn cob granules (VectoLex®, applied at a dosage rate of 1 g/m2 by hand).

Frequency of application: Larviciding of open habitats: weekly; closed habitats: every three months.

Duration of intervention period: 12 months

Who was responsible for LSM? Open habitats were treated by modestly paid members of the community, Mosquito Contro CORPs, each of which was assigned to a specific area (mtaa). An additional team of CORPs was responsible for treating closed habitats. CORPs reported to the Ward Office.

Co-interventions: None. However ITNs were used in the study area. Coverage: Non-intervention area: 23.6% (year 1), 27.7% (year 2), 24.6% (year 3); Intervention area: 23.3% (year 1), 26.3% (year 2), 22.4% (year 3).

Co-interventions equal in each arm? Yes


Outcomes1. Parasite prevalence (measured with randomized, cluster-sampled household surveys in May to September 2004, November to July 2004, September 2005 to May 2006, July 2006 to March 2007, with parasite prevalence determined by microscopy).

2. EIR (measured by (1) human landing catch for 45 minutes of each hour from 6pm to 6am, at sentinel sites every four weeks, and (2) laboratory analysis of specimens for sporozoites).


NotesContinent: Africa

Country: Tanzania

Ecosystem: Coastal

Urban or rural: Urban

Extensive or localized larval habitats: Localized

Primary larval habitats: Man-made habitats exposed to sunlight

Transmission intensity: Low to moderate

Transmission season(s): July to September

Primary and secondary vector: An. gambiae s.l.

Primary malaria parasite: P. falciparum

Source of funding: Bill & Melinda Gates Foundation; Valent Biosciences Corporation; United States Centers for Disease Control and Prevention and United States Agency for International Development (Environmental Health Program, Dar es Salaam Mission and the President’s Malaria Initiative, all administered through Research Triangle International); Wellcome Trust.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskMalaria prevalence determined by blinded reading of blood smears of randomly chosen individuals.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)High riskAll household members tested, but results presented only for children aged 0 to five years.

Baseline characteristicsUnclear riskBaseline characteristics not specified.

ContaminationLow riskMost of control clusters > 1 km from intervention clusters.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Majambere 2010 GMB

MethodsTrial design: Randomized cross-over trial

Type of cluster: Area of land (zone)

Cluster size: Each zone was 12 x 8 km and was subdivided into three parallel 4 km wide bands perpendicular to the river. Study villages were recruited from the central band of each zone.

Number of clusters in each arm: Two

Adjusted for clustering? Yes, included as random effects.


ParticipantsAge: Six months to 10 years

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Malaria incidence): Zone 1: 496; Zone 2: 508; Zone 3: 525; Zone 4: 510

Secondary outcome sample size (EIR): 15 traps per zone, divided between the villages with one to three sentinel houses per village proportional to village size


InterventionsIntervention: Larviciding alone

Larviciding:

Larval habitats in areas of low vegetation coverage were treated with Bti water-dispersible granules (VectoBac® AM65-52, applied at 0.2kg/hectare using knapsack compression sprayers). Less accessible larval habitats in areas of high vegetation coverage were treated with Bti corn granules (VectoBac® AM65-52, applied at 5.0kg/hectare by hand from buckets or using motorized knapsack granule blowers).

Frequency of application: Weekly

Duration of intervention period: June to November 2006 (6 months), May to November 2007 (7 months)

Who was responsible for LSM? Field applicators were recruited from local communities and trained for one month before larviciding. Applicators were supervised by one field supervisor in each of the four study zones

Co-interventions: None. However ITNs were used in the study area (coverage: Zone 1: 27.6% (2006), 37.2% (2007); Zone 2: 6.1% (2006), 81.4% (2007); Zone 3: 38.3% (2006), 71.2% (2007); Zone 4: 34.3% (2006), 70.4% (2007).

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured with passive case detection by study nurses and government village health workers)

2. Parasite prevalence (measured with two cross-sectional surveys per year, one before and one after the main transmission season)

3. Splenomegaly prevalence (measured with two cross-sectional surveys per year, one before and one after the main transmission season)

4. EIR (measured using CDC light traps at 60 sentinel sites every two weeks).

5. Adult mosquito density (measures other than human biting rate) (measured using CDC light traps at 60 sentinel sites every two weeks).


NotesContinent: Africa

Country: The Gambia

Ecosystem: Savannah

Urban or rural: Rural

Extensive or localized larval habitats: Extensive

Primary larval habitats: Flood plains, rice paddy fields

Transmission intensity: High

Transmission season(s): July to October

Primary and secondary vector: An. gambiae

Primary malaria parasite: P. falciparum

Source of funding: National Institutes of Health


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskEach area served as its own control.

Allocation concealment (selection bias)Low riskEach area served as its own control.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskData collectors blinded to intervention status.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskOutcomes all reported as specified.

Baseline characteristicsLow riskEach area served as its own control.

ContaminationLow riskClusters bordered by 4 km zones.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasLow riskLow risk of confounding.

Samnotra 1980 IND

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Town

Cluster size: Intervention arm 92,000 individuals; control arm 5000 individuals

Number of clusters in each arm: One

Adjusted for clustering? n/a


ParticipantsAge: Any

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Malaria incidence): Intervention arm: 92,000; control arm: 5000           

Secondary outcome sample size (Adult mosquito density (measures other than human biting rate)): 80 sentinel sites


InterventionsIntervention: Habitat manipulation with larviciding

Details of the intervention:                                                           

Habitat manipulation: attempts to persuade householders to remove domestic water storage containers made with limited success    

Larviciding: Larval habitats (excluding stored domestic water) were treated with pirimiphos-methyl (applied at 12.5g active ingredient/ha, with knapsack sprayers)

Frequency of application: Weekly

Duration of intervention period: 15 months

Who was responsible for LSM? Study staff were responsible for larviciding. Attempts were made to persuade the local community to conduct habitat modification

Co-interventions: Case management (active case detection): presumptive treatment of all fever cases with chloroquine (coverage not stated)

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured with continuous community surveillance)

2. Parasite prevalence (measured with community surveys)

3. Adult mosquito density (measures other than human biting rate): (measured with weekly indoor resting collections using an aspirator, at sentinel sites. 16 of 80 sentinel sites visited each week day)


NotesContinent: Asia

Country: India

Ecosystem: Desert fringe

Urban or rural: Urban

Extensive or localized larval habitats: Localized

Primary larval habitats: Containers, wells, rainwater pools, canals, stagnant pools in drains

Transmission intensity: Low

Transmission season(s): May to September

Primary and secondary vector: An. culicifacies, An. stephensi

Primary malaria parasite: P. falciparum

Source of funding: Haryana State Health Authorities; Alkali and Chemical Corporation of India Ltd; ICI Plant Protection Division


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information given as to blinding of those seeing patients and reading blood slides.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Unclear riskOutcomes not specified.

Baseline characteristicsUnclear riskBaseline characteristics not stated; intervention town much larger than control town.

ContaminationLow risk8 km between control and intervention towns.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Santiago 1960 PHL

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Area of town (barrio)

Cluster size: 25,545 people (intervention cluster)

Number of clusters in each arm: One

Adjusted for clustering? No


ParticipantsAge: Two to 10 years

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Parasite prevalence): Intervention arm: 500; control arm: 200

Secondary outcome sample size (Adult mosquito density (measures other than human biting rate)): Not stated


InterventionsIntervention: Habitat manipulation alone

Details of the intervention:

Habitat manipulation: automatic siphons were constructed over two streams which were the main larval habitats. Water was flushed to control larvae over distances of 1073m and 2897m downstream, respectively. Existing siphons were repaired.

Frequency of application: Constant

Duration of intervention period: 12 months

Who was responsible for LSM? United Stated Public Health Service

Co-interventions: None

Co-interventions equal in each arm? n/a


Outcomes1. Parasite prevalence (measured with community-based cross-sectional surveys)

2. Splenomegaly prevalence (measured with community-based cross-sectional surveys)

3. Adult mosquito density (measures other than human biting rate) (sampled with human baited traps and carabao baited traps every two weeks)


NotesContinent: Asia

Country: Philippines

Ecosystem: Coastal

Urban or rural: Urban

Extensive or localized larval habitats: Localized

Primary larval habitats: Streams fed by a lake

Transmission intensity: High

Transmission season(s): Not stated

Primary and secondary vector: An. minimus flavirostris

Primary malaria parasite: P. falciparum

Source of funding: Malaria Eradication Project, San Pablo City


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskSampling method for periodic surveys not stated, though reportedly surveyed 50% to 80% of children per year.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskOutcomes reported as specified.

Baseline characteristicsUnclear riskClusters in same town, but no baseline characteristics specified. Only 6 months of pre-treatment entomological data were collected.

ContaminationLow riskClusters 8 km apart.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Sharma 2008 IND

MethodsTrial design: Controlled before-and-after trial

Type of cluster: Village

Cluster size: Intervention arm: 271 individual; control arms: 143 and 156 individuals

Number of clusters in each arm:  Intervention arm: one; control arm: two

Adjusted for clustering? No


ParticipantsAge: Any

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Malaria incidence): Total study population: 570

Secondary outcome sample size (Parasite prevalence): 40% households sampled in each of the three clusters (combined total population 570)


InterventionsIntervention: Habitat modification alone

Details of the intervention:                                     

Habitat modification: Construction of a small concrete dam 25m x 4m across the stream in the village to provide water for irrigation reduced the number of larval habitats in the village.

Frequency of application: n/a

Duration of intervention period: 23 months

Who was responsible for LSM? The district administration constructed the dam at the request of the village panchayat (governing body)

Co-interventions: None. However indoor residual spraying was conducted annually with DDT and a synthetic pyrethroid (coverage: 60% to 80%).

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured with weekly longitudinal surveillance and continuous passive case detection)

2. Parasite prevalence (measured with three cross-sectional surveys per year)


NotesContinent: Asia

Country: India

Ecosystem: Forest

Urban or rural: Rural

Extensive or localized larval habitats: Localized

Primary larval habitats: Streams (An. fluviatilis), stagnant pools, ditches,  irrigation channels (An. culicifacies)

Transmission intensity: Moderate

Transmission season(s): October to December

Primary and secondary vector: An. fluviatilis, An. culifacies

Primary malaria parasite: P. falciparum

Source of funding: Indian Council of Medical Research; Ministry of Health and Family Welfare, Government of India


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot randomly chosen.

Allocation concealment (selection bias)High riskNot randomly chosen.

Blinding of outcome assessment (detection bias)
All outcomes
High riskSurveillance personnel not blinded to intervention status.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Unclear riskOutcomes reported as specified.

Baseline characteristicsLow riskBaseline incidences reported and similar.

ContaminationLow riskControl and intervention villages 30 km apart.

Incorrect analysisUnclear riskCluster adjustment not applicable.

Other biasHigh riskHigh risk of confounding.

Shililu 2007 ERI

MethodsTrial design: Cluster-RCT

Type of cluster: Village

Cluster size: Not stated.

Number of clusters in each arm: Four

Adjusted for clustering? No


ParticipantsAge: n/a

Sex: n/a

Co-morbidities and pregnancy: n/a

Primary outcome sample size (Adult mosquito density (measures other than human biting rate)): 12 light traps per study village

Secondary outcome sample size: n/a


InterventionsIntervention: Habitat modification with larviciding

Details of the intervention:    

Habitat modification: Filling or drainage of rain pools, puddles at water supply points and stream bed pools

Larviciding: Larval habitats which could not be eliminated by habitat modification were treated in rotation with Bti granules (VectoBac®, applied at 11.2kg/ha using a granular spreader), Bs corn granules (VectoLex®, applied at 22.4kg/ha using a granular spreader) and temephos (Abate®, applied at 112 ml/ha using a liquid sprayer)

Frequency of application: Weekly

Duration of intervention period: 24 months

Who was responsible for LSM? Study staff; local community

Co-interventions: None. However ITNs and IRS were conducted as part of the national malaria control programme (coverage not stated).

Co-interventions equal in each arm? Not stated


Outcomes1. Adult mosquito density (measures other than human biting rate) (measured using CDC light traps from dusk to dawn (12 hours) 2 days per week for 24 months)


NotesContinent: Africa

Country: Eritrea

Ecosystem: Desert fringe, highland and lowland

Urban or rural: Rural

Extensive or localized larval habitats: Localized

Primary larval habitats: Stream bed pools, canals, drainage channels, wells, communal water supply points

Transmission intensity: Not stated

Transmission season(s): Short period of transmission coinciding with short rainy season

Primary and secondary vector:An. arabiensis

Primary malaria parasite: P. falciparum

Source of funding: United States Agency for International Development, Environmental Health Project, International Center of Insect Physiology and Ecology, National Institutes of Health


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskClusters randomly assigned; however method of randomization not stated.

Allocation concealment (selection bias)Unclear riskOne village randomly selected in each zone; however method of randomization not stated.

Blinding of outcome assessment (detection bias)
All outcomes
High riskSurveillance personnel not blinded to intervention status.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)Low riskOutcomes reported as specified.

Baseline characteristicsUnclear riskPairs of villages selected to be similar but baseline characteristics not reported.

ContaminationUnclear riskDistance of villages from one another not stated.

Incorrect analysisHigh riskNot adjusted for clustering.

Other biasLow riskLow risk of confounding.

Yapabandara 2001 LKA

MethodsTrial design: Cluster-RCT

Type of cluster: Village

Cluster size: Four villages of <500 people, four villages of 600-1100 people

Number of clusters in each arm: Four

Adjusted for clustering? No


ParticipantsAge: Any

Sex: Any

Co-morbidities and pregnancy: Not stated

Primary outcome sample size (Malaria incidence): 4566 (pre-intervention); 4659 (post-intervention)

Secondary outcome sample size (Parasite prevalence): 3351


InterventionsIntervention: Larviciding

Details of the intervention: 

Larviciding: Gem pits and riverbed and stream pools were treated with pyriproxyfen S-31183 granules (Adeal® 0.5%, applied at 2g/m3).

Frequency of application: December 1994, June to July 1995, November 1995

Duration of intervention period: 12 months

Who was responsible for LSM? Study staff

Co-interventions: Case management following whole community survey (coverage comprehensive).

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured by passive case detection)

2. Parasite prevalence (measured by cross-sectional surveys (two in pre-intervention year, two in post-intervention year)

3. Adult mosquito density (measures other than human biting rate) (measured by window exit trap collection, pyrethrum spray sheet, indoor human landing catch, cattle-baited hut collection, cattle-baited net trap collection at sentinel sites)


NotesContinent: Asia

Country: Sri Lanka

Ecosystem: Forest

Urban or rural: Rural

Extensive or localized larval habitats: Localized

Primary larval habitats: Abandoned gem mine pits

Transmission intensity: Moderate to high

Transmission season(s): October to December

Primary and secondary vector: An. culicifacies, An. subpictus Grassi

Primary malaria parasite: P. vivax

Source of funding: Sumitomo Corporation, United Nations Development Program, World Bank, WHO


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomly assigned, though method not stated.

Allocation concealment (selection bias)Unclear riskNot described.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskParasite prevalence determined by blinded reading of blood slides, but incidence in local clinics and blinding impossible.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)High riskSeveral methods of collection of entomologic data described, not all reported.

Baseline characteristicsUnclear riskCharacteristics not reported, but stratification and randomization were performed based on baseline data. Baseline data for 12 months pre-treatment is presented.

ContaminationLow riskAt least 1.5 km between villages.

Incorrect analysisHigh riskNot adjusted for clustering.

Other biasLow riskLow risk of confounding.

Yapabandara 2004 LKA

MethodsTrial design: Cluster-RCT

Type of cluster: Village

Cluster size: Each of the 12 villages was defined as a circle of 1.5km radius centred on a stream or irrigation canal

Number of clusters in each arm: Six

Adjusted for clustering? No


ParticipantsAge: Any

Sex: Any

Co-morbidities and pregnancy: Any

Primary outcome sample size (Malaria incidence): 15415 individuals

Secondary outcome sample size (Adult mosquito density (measures other than human biting rate)): Not stated


InterventionsIntervention: Larviciding alone

Details of the intervention:                    

Larviciding: Riverbed pools, streams, irrigation ditches, quarry pits and agricultural wells were treated with pyriproxyfen S-31183 0.5% granules (Sumilarv®, applied at 2g/m3 using a spoon).

Frequency of application: Two rounds of larviciding were conducted: July 2001 and December 2001.

Duration of intervention period: 12 months

Who was responsible for LSM? Study staff

Co-interventions: Larvivorous fish: Poecillia reticulata were added to drinking water wells. IRS was conducted as part of the national malaria control programme during November and June each year (coverage not stated).

Co-interventions equal in each arm? Yes


Outcomes1. Malaria incidence (measured by passive case detection at two field clinics and two clinics at outpatient departments at a hospital and dispensary)

2. Adult mosquito density (measures other than human biting rate) (measured using cattle-baited huts at sentinel sites)


NotesContinent: Asia

Country: Sri Lanka

Ecosystem: 'Dry zone'

Urban or rural: Rural

Extensive or localized larval habitats: Localized and extensive

Primary larval habitats: River bed pools, streams, irrigation ditches (dry season); rice paddies (rainy season)

Transmission intensity: Moderate

Transmission season(s): January to March

Primary and secondary vector: An. culifacies, An. subpictus

Primary malaria parasite: P. vivax

Source of funding: United Nations Development Program, World Bank, World Health Organization Special Program for Research and Training in Tropical Diseases


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomly assigned, though method not stated.

Allocation concealment (selection bias)Unclear riskNot described.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskParasite prevalence determined by blinded reading of blood slides, but incidence measured at local clinics and blinding impossible.

Blinding of participants and personnel (performance bias)
All outcomes
High riskImpossible to blind implementers or inhabitants to intervention.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIndividual patients not followed up therefore not possible to measure percentage loss to follow-up.

Selective reporting (reporting bias)High riskSeveral methods of collection of entomologic data described, not all reported.

Baseline characteristicsUnclear riskCharacteristics not reported, but stratification and randomization performed based on baseline data.

ContaminationUnclear riskDistance of villages from one another not specified.

Incorrect analysisHigh riskNot adjusted for clustering.

Other biasLow riskLow risk of confounding.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Anon (a)We could not obtain the full-text article.

Anon (b)We could not obtain the full-text article.

Anon (c)We could not obtain the full-text article.

Anon (d)We could not obtain the full-text article.

Baduilin 1931We could not obtain the full-text article.

Barbazan 1998No control.

Berti 1946We could not obtain the full-text article.

Bini 1925We could not obtain the full-text article.

Booker 1936We could not obtain the full-text article.

Castro 2000No control.

Castro 2002No control.

Cross 1933No control.

Curry 1935We could not obtain the full-text article.

Davis 1928We could not obtain the full-text article.

Dryenski 1936Study did not have one year of baseline data.

Dua 1991Uneven application of other malaria control interventions between control and intervention areas: weekly active surveillance and treatment of fever cases in intervention area, but not in controls.

Dua 1997Uneven application of other malaria control interventions between control and intervention areas: weekly active surveillance and treatment of fever cases in intervention area, but not in controls.

Elmendorff 1948No control.

Essed 1932No control.

Fillinger 2006No control.

Gallus 1970We could not obtain the full-text article.

Gammans 1926We could not obtain the full-text article.

Gladney 1968No control.

Guelmino 1928We could not obtain the full-text article.

Hackett 1925We could not obtain the full-text article.

Ivorro Canno 1975Uneven application of other malaria control interventions between control and intervention areas: chloroquine chemoprophylaxis applied in intervention village and not in control village.

Kinde-Gazard 2012Insufficient information reported to determine eligibility.

Kumar 1998No control.

Lee 2010No control.

Martini 1931We could not obtain the full-text article.

Mulligan 1982No control.

Murray 1984No control.

Okan 1949We could not obtain the full-text article.

Rodriguez Ocana 2003We could not obtain the full-text article.

Rojas 1987Uneven application of other malaria control interventions between control and intervention areas: indoor residual spraying with DDT every six to 10 months used in intervention area, but not in control.

Sharma 1989Uneven application of other malaria control interventions between control and intervention areas: weekly active surveillance and treatment in intervention area, as well as extensive use of larvivorous fish; control villages changed multiple times over the life of the study, compromising comparability.

Singh 1984No control.

Singh 1989Uneven application of other malaria control interventions between control and intervention areas: weekly active surveillance and treatment in intervention area, compared to bimonthly in control; DDT indoor residual spraying in control villages.

Stratman-Thomas 1937We could not obtain the full-text article.

Symes 1931Larval habitats differed between control and intervention sites at baseline.

Vittal 1982No control.

Williamson 1934We could not obtain the full-text article.

Xu 1992No control.

Yasuoka 2006Study did not have one year of baseline data.

Yohannes 2005Larval habitats differed between control and intervention sites at baseline.

 
Comparison 1. Habitat modification alone

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Malaria incidence1Rate Ratio (Fixed, 95% CI)Totals not selected

    1.1 Controlled before-and-after trials; pre-intervention
1Rate Ratio (Fixed, 95% CI)0.0 [0.0, 0.0]

    1.2 Controlled before-and-after trials; post-intervention
1Rate Ratio (Fixed, 95% CI)0.0 [0.0, 0.0]

 2 Parasite prevalence1Risk Ratio (M-H, Fixed, 95% CI)Totals not selected

    2.1 Controlled before-and-after trials; pre-intervention
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    2.2 Controlled before-and-after trials; post-intervention
1Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 
Comparison 2. Habitat modification with larviciding

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Parasite prevalence1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 Controlled before-and-after trials; pre-intervention
11737Risk Ratio (M-H, Fixed, 95% CI)0.44 [0.30, 0.64]

    1.2 Controlled before-and-after trials; post-intervention
11538Risk Ratio (M-H, Fixed, 95% CI)0.25 [0.19, 0.34]

 2 Splenomegaly prevalence1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    2.1 Controlled before-and-after trials; pre-intervention
11737Risk Ratio (M-H, Fixed, 95% CI)0.58 [0.51, 0.66]

    2.2 Controlled before-and-after trials; post-intervention
11538Risk Ratio (M-H, Fixed, 95% CI)0.41 [0.36, 0.47]

 
Comparison 3. Habitat manipulation alone

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Parasite prevalence1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 Controlled before-and-after trials; pre-intervention
1847Risk Ratio (M-H, Fixed, 95% CI)1.37 [0.70, 2.68]

    1.2 Controlled before-and-after trials; post-intervention
1846Risk Ratio (M-H, Fixed, 95% CI)0.02 [0.00, 0.15]

 2 Splenomegaly prevalence1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    2.1 Controlled before-and-after trials; pre-intervention
1832Risk Ratio (M-H, Fixed, 95% CI)0.51 [0.31, 0.85]

    2.2 Controlled before-and-after trials; post-intervention
1846Risk Ratio (M-H, Fixed, 95% CI)0.02 [0.00, 0.17]

 
Comparison 4. Habitat manipulation with larviciding

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Malaria incidence1Rate Ratio (Fixed, 95% CI)Subtotals only

    1.1 Controlled before-and-after trials; pre-intervention
197000Rate Ratio (Fixed, 95% CI)1.14 [1.01, 1.28]

    1.2 Controlled before-and-after trials; post-intervention
197000Rate Ratio (Fixed, 95% CI)0.24 [0.22, 0.25]

 2 Parasite prevalence1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    2.1 Controlled before-and-after trials; pre-intervention
11887Risk Ratio (M-H, Fixed, 95% CI)1.44 [0.99, 2.11]

    2.2 Controlled before-and-after trials; post-intervention
12713Risk Ratio (M-H, Fixed, 95% CI)0.54 [0.45, 0.65]

 
Comparison 5. Larviciding alone

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Malaria incidence3Rate Ratio (Random, 95% CI)Subtotals only

    1.1 Cluster-RCTs; pre-intervention
219981Rate Ratio (Random, 95% CI)0.95 [0.84, 1.08]

    1.2 Cluster-RCTs; post-intervention
220124Rate Ratio (Random, 95% CI)0.26 [0.22, 0.31]

    1.3 Controlled before-and-after trials; pre-intervention
1400Rate Ratio (Random, 95% CI)1.28 [0.75, 2.20]

    1.4 Controlled before-and-after trials; post-intervention
1663Rate Ratio (Random, 95% CI)0.69 [0.33, 1.43]

 2 Malaria incidence (post-intervention) sensitivity analysis2Rate Ratio (Fixed, 95% CI)Subtotals only

    2.1 Not adjusted for clustering
2Rate Ratio (Fixed, 95% CI)0.26 [0.22, 0.30]

    2.2 Adjusted using ICC = 0.01
2Rate Ratio (Fixed, 95% CI)0.25 [0.16, 0.40]

    2.3 Adjusted using ICC = 0.1
2Rate Ratio (Fixed, 95% CI)0.25 [0.06, 0.98]

 3 Parasite prevalence2Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    3.1 Cluster-RCTs; pre-intervention
13351Risk Ratio (M-H, Fixed, 95% CI)1.02 [0.66, 1.56]

    3.2 Cluster-RCTs; post-intervention
12963Risk Ratio (M-H, Fixed, 95% CI)0.11 [0.05, 0.22]

    3.3 Controlled before-and-after trials; pre-intervention
12439Risk Ratio (M-H, Fixed, 95% CI)1.29 [1.04, 1.59]

    3.4 Controlled before-and-after trials; post-intervention
12374Risk Ratio (M-H, Fixed, 95% CI)0.60 [0.42, 0.87]

 4 Parasite prevalence (post-intervention) sensitivity analysis1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    4.1 Not adjusted for clustering
12963Risk Ratio (M-H, Fixed, 95% CI)0.11 [0.05, 0.22]

    4.2 Adjusted using ICC = 0.01
1631Risk Ratio (M-H, Fixed, 95% CI)0.13 [0.03, 0.56]

    4.3 Adjusted using ICC = 0.1
178Risk Ratio (M-H, Fixed, 95% CI)0.16 [0.01, 3.14]

 
Comparison 6. Larval source management versus control

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Malaria incidence5Rate Ratio (Random, 95% CI)Subtotals only

    1.1 Cluster-RCTs; pre-intervention
219981Rate Ratio (Random, 95% CI)0.95 [0.84, 1.08]

    1.2 Cluster-RCTs; post-intervention
220124Rate Ratio (Random, 95% CI)0.26 [0.22, 0.31]

    1.3 Controlled before-and-after trials; pre-intervention
397970Rate Ratio (Random, 95% CI)1.50 [0.89, 2.52]

    1.4 Controlled before-and-after trials; post-intervention
398233Rate Ratio (Random, 95% CI)0.51 [0.18, 1.44]

 2 Parasite prevalence6Risk Ratio (M-H, Random, 95% CI)Subtotals only

    2.1 Cluster-RCTs; pre-intervention
13351Risk Ratio (M-H, Random, 95% CI)1.02 [0.66, 1.56]

    2.2 Cluster-RCTs; post-intervention
12963Risk Ratio (M-H, Random, 95% CI)0.11 [0.05, 0.22]

    2.3 Controlled before-and-after trials; pre-intervention
57480Risk Ratio (M-H, Random, 95% CI)0.99 [0.65, 1.52]

    2.4 Controlled before-and-after trials; post-intervention
58041Risk Ratio (M-H, Random, 95% CI)0.32 [0.19, 0.55]

 3 Splenomegaly prevalence2Risk Ratio (M-H, Random, 95% CI)Subtotals only

    3.1 Controlled before-and-after trials; pre-intervention
22569Risk Ratio (M-H, Random, 95% CI)0.57 [0.50, 0.65]

    3.2 Controlled before-and-after trials; post-intervention
22384Risk Ratio (M-H, Random, 95% CI)0.11 [0.01, 2.10]

 
Summary of findings for the main comparison. LSM for controlling malaria

LSM for controlling malaria

Patient or population: People living in malaria endemic areas
Settings: Urban or rural settings in Africa, Asia and Europe
Intervention: LSM

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlLSM

Malaria incidence65 per 100017 per 1000
(14 to 20)
Rate Ratio 0.26
(0.22 to 0.31)
20124
(2 cluster-RCTs)
⊕⊕⊕⊝
moderate1,2,3,4
The 95% CI may be falsely narrow as trials did not adjust for cluster design.

232 per 1000118 per 1000
(42 to 334)
Rate Ratio 0.51
(0.18 to 1.44)
98233
(3 controlled before and after studies)
⊕⊝⊝⊝
very low5,6,7,8

Parasite prevalence44 per 10005 per 1000
(2 to 10)
Risk Ratio 0.11
(0.05 to 0.22)
2963
(1 cluster-RCT)
⊕⊕⊕⊝
moderate4,9,10
The 95% CI may be falsely narrow as the trial did not adjust for cluster design.

157 per 100050 per 1000
(30 to 86)
Risk Ratio 0.32
(0.19 to 0.55)
8041
(5 controlled before and after studies)
⊕⊕⊕⊝
moderate11,12,13,14,15

*The basis for the assumed risk (for example, the median control group risk across studies) is provided in the footnotes. The corresponding risk (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).
CI: Confidence interval; RR: Risk ratio.

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 Downgraded by 1 for serious risk of bias: Both studies were described as randomized but did not adequately describe a process to reduce the risk of selection bias.
2 No serious inconsistency: There was no statistical heterogeneity.
3 No serious indirectness: Both studies were conducted in rural Sri Lanka. The primary vectors were An. culicifacies and An. subpictus and the primary mosquito larval habitats were river bed pools, streams irrigation ditches and rice paddies (Yapabandara 2004 LKA), and abandoned gem mine pits (Yapabandara 2001 LKA). The intervention was larviciding with pyriproxyfen approximately every six months. Generalization of this result to wider settings is supported by the findings from the non-randomized studies.
4 No serious imprecision: Although these studies did not adjust for the cluster design, a sensitivity analysis adjusting this result for the cluster design suggested the result is likely to be both statistically significant and clinically important.
5 Downgraded by 1 for risk of bias: In two of these studies, there were important baseline differences in malaria incidence between groups. The incidence was higher in the intervention group pre-intervention and reduced to similar levels as the control group post-intervention.
6 Not downgraded for inconsistency: There was heterogeneity in this result which can be explained by baseline differences in two of the studies. However, there was a reduction in malaria incidence in the intervention groups in all three studies.
7 No serious indirectness: Sharma 2008 IND was conducted in rural India where the primary vectors were An. fluviatilis and An. culicifacies, the main larval habitats of which were streams, stagnant pools, ditches and irrigation channels. A dam was constructed across the stream, reducing the number of larval habitats in the intervention village. Fillinger 2009 KEN was conducted in highland villages in rural Kenya, where the major vectors were An. gambiae and An. funestus. The primary larval habitats were small streams and papyrus swamps, which were treated weekly with Bs for six months and then Bti for 13 months. Samnotra 1980 IND was conducted in a desert fringe area of urban India where the primary vectors were An. culicifacies and An. stephensi, the main larval habitats of which were containers, wells, canals and rainwater pools and drains. Larviciding with pirimiphos-methyl was conducted weekly for 15 months.
8 Downgraded by 1 for imprecision: The overall effect is not statistically significant but is difficult to interpret due to the baseline differences.
9 Downgraded by 1 for serious risk of bias: This study was described as randomized but did not adequately describe a process to reduce the risk of selection bias.
10 No serious indirectness: This single study was conducted in rural Sri Lanka where the primary larval habitats were abandoned gem mine pits and the findings may not be easily generalized elsewhere. However generalization of this result to wider settings is supported by the findings from the non-randomized studies.
11 No serious risk of bias: the risk of bias inherent in these non-randomized studies is already accounted for in the initial downgrading to 'low quality evidence'.
12 No serious inconsistency: All five studies showed a large benefit with LSM. The smallest effect was a 40% reduction in malaria prevalence which is still considered clinically important.
13 No serious indirectness: These five studies were in conducted in urban and rural settings in Greece, Tanzania, India and the Philippines. Mosquito larval habitats ranged from man-made habitats, containers and wells to rainwater pools, irrigation channels, ditches and streams, and interventions included dam construction, flushing of streams, straightening or lining of streams, drainage of marshland and larviciding.
14 No serious imprecision: All studies showed clinically important and statistically significant effects.
15 Upgraded by 1 as the effects seen were large. The two studies with smaller effects (Sharma 2008 IND; Fillinger 2009 KEN) had baseline differences which would lead to an underestimation of the true effect.
 
Table 1. Assessment of risk of bias

Risk of bias componentLowHighUnclear

Sequence generationRandom component in the sequence generation process is described.Non-random method is used.No or unclear information reported.

Allocation concealmentPatients and investigators could not foresee assignment.Patients and investigators could foresee assignment.No or unclear information reported.

Blinding (performance)Performance bias due to knowledge of the allocated interventions by participants and personnel during the studyNo evidence of performance bias due to knowledge of the allocated interventions by participants and personnel during the studyNo or unclear information reported.

Blinding (detection)Primary outcomes assessed blinded.Primary outcomes not assessed blinded.No or unclear information reported.

Incomplete outcome dataNo or low missing data, reason for missing data is unlikely to be related to the true outcome, or missing data is balanced across groups.High missing data, reason for missing data is likely to be related to the true outcome, or missing data is unbalanced across groups.No or unclear information reported.

Selective outcome reportingAll pre-specified outcomes are reported (expected or see protocol).Not all pre-specified outcomes are reported; or additional outcomes reported.No or unclear information reported.

Recruitment bias No change in size or number of clusters after randomization.Possible change in size or number of clusters after randomization.No or unclear information reported.

Baseline characteristicsIf baseline characteristics of the study and control areas are reported and similar.If there are differences between control and intervention areas.No or unclear information reported.

Contaminationit is unlikely that the control group received the intervention.It is likely that the control group received the intervention.No or unclear information reported.

Incorrect analysis

(Randomized studies only)
Randomized studies: clustering taken into account in analysis.Randomized studies: clustering not taken into account in analysis.Randomized studies: No or unclear information reported.

Other biases (confounding)Non-randomized studies: no evidence of confounding (selection bias)Non-randomized studies: evidence of confounding (selection bias)Non-randomized studies: no or unclear information reported.

 
Table 2. Summary of interventions and eco-epidemiological settings

InterventionStudy IDStudy designDetails of the interventionWho was responsible for LSM?EcosystemPrimary vectors (primary larval habitats)Malaria transmission intensity

Habitat modification aloneSharma 2008 INDControlled before-and-afterDam constructionCommunity, governmentForest; ruralAn. fluviatilis (streams), An. culicifacies (stagnant pools, ditches,  irrigation channels)Moderate

Habitat modification with larviciding

 
Shililu 2007 ERICluster-RCTLand filling and grading; drainage; larviciding with synthetic organic compounds and microbialsStudy staff, communityDesert fringe, highland and lowland; ruralAn. arabiensis (stream bed pools, canals, drainage channels, wells, communal water supply points)Not stated

Balfour 1936 GRCControlled before-and-afterStraightening, deepening and lining of natural streams; drainage; larviciding with Paris GreenGovernmentCoastal; urban and ruralAn. elutus; An. superpictus (primarily man-made habitats)Low to moderate

Habitat manipulation aloneSantiago 1960 PHLControlled before-and-afterControlling water levels and stream flushingCoastal; urbanAn. minimus flavirostris (streams fed by a lake)High

Habitat manipulation with larviciding

 
Castro 2009 TZAControlled before-and-afterClearing of aquatic vegetation and debris; larviciding with microbialsStudy staff, community, governmentCoastal; urbanAn. gambiae, An. funestus (drains)Low to moderate

Samnotra 1980 INDControlled before-and-afterRemoval of 'domestic' larval habitats; Larviciding with synthetic organic compoundsStudy staff, communityDesert fringe; urbanAn. culicifacies, An. stephensi (containers, wells, rainwater pools, canals, stagnant pools in drains)Low

Larviciding alone

 

 

 

 

 

 
Coulibaly 2011 MLICluster-RCTLarviciding with microbialsStudy staff, communitySavannah; ruralAn. gambiae (brick pits, ponds, tyre prints)High

Yapabandara 2001 LKACluster-RCTLarviciding with insect growth regulatorsStudy staff, communityForest; ruralAn. culicifacies, An. subpictus Grassi. (abandoned gem mine pits)Moderate to high

Yapabandara 2004 LKACluster-RCTLarviciding with insect growth regulatorsStudy staff'Dry zone'; ruralAn. culifacies, An. subpictus (river bed pools, streams, irrigation ditches (dry season); rice paddies (rainy season))Moderate

Fillinger 2008 TZAControlled before-and-afterLarviciding with microbialsStudy staff, communityCoastal; urbanAn. gambiae s.s., An. arabiensis (man-made habitats exposed to sunlight)Low to moderate

Fillinger 2009 KENControlled before-and-afterLarviciding with microbialsStudy staffHighland; ruralAn. gambiae s.l.,An. funestus s.l. (small streams, papyrus swamps)Moderate

Geissbühler 2009 TZAControlled before-and-afterLarviciding with microbialsStudy staff, communityCoastal; urbanAn. gambiae s.l. (man-made habitats exposed to sunlight)Low to moderate

Majambere 2010 GMBRandomized cross-overLarviciding with microbialsStudy staff, communitySavannah; ruralAn. gambiae (flood plains, rice paddy fields)High

 
Table 3. Summary of original data for Balfour 1936 GRC

Outcome

 

 
GroupParasite or splenomegaly prevalence

(total positive/total examined)

Pre-interventionPost-intervention


193019311932193319341935

Parasite prevalence

 
Control9.1%

(59/650)
23.9%

(164/685)
15.0%

(104/692)
21.9%

(147/670)
10.0%

(69/690)
18.0%

(123/682)

Treatment4.0%

(43/1087)
6.0%

(51/853)
9.0%

(75/837)
4.0%

(33/830)
1.0%

(8/834)
1.6%

(13/827)

Splenomegaly prevalence

 
Control46.0%

(299/650)
56.9%

(390/685)
43.1%

(298/692)
44.0%

(295/670)
35.9%

(248/690)
40.0%

(273/682)

Treatment26.5%

(288/1087)
23.4%

(200/853)
18.0%

(151/837)
13.0%

(108/830)
12.0%

(100/834)
7.0%

(58/827)

 
Table 4. Entomological data: Adult mosquito density (density measures other than human biting rate)

InterventionStudy IDStudy designMean adult mosquito density (95% CI)Percent reduction

(95% CI)1
Notes

Pre-interventionPost-intervention


ControlTreatmentControlTreatment

Habitat modification with larvicidingShililu 2007 ERICluster-RCT - -4.994.2315.2Mean number of female adult anophelines per night (light traps)

Habitat manipulation aloneSantiago 1960 PHLControlled before-and-after trial0.150.200.170.0291.2Mean number of adult anophelines per catching station (human-baited traps)

Habitat manipulation with larvicidingSamnotra 1980 INDControlled before-and-after trial22270269621390.3Mean number of adult anophelines per catching station (resting catch)

Larviciding aloneCoulibaly 2011 MLI (2009 data)Cluster-RCT - -2.271.4934.4-

Coulibaly 2011 MLI(2010 data)Cluster-RCT - -6.033.7537.8-

Yapabandara 2001 LKACluster-RCT16.8827.6322.133.3890.7Mean number of adult anophelines per man per night (partial night human landing catches) (An. culicifacies)

Yapabandara 2001 LKA2Cluster-RCT - - - - -Mean number of adult anophelines per man per night (all night human landing catches) (An. culicifacies)

Yapabandara 2004 LKACluster-RCT6.649.118.751.4488.0Mean resting density of adult anophelines (cattle baited huts)  (An. culicifacies)

Fillinger 2009 KENControlled before-and-after trial3.69 (2.25 to 6.06)3.49 (2.49 to 4.88)0.60 (0.45 to 0.79)0.08 (0.06 to 0.13)85.9

(68.3 to 93.7)
Mean number adult anophelines per house (pyrethrum spray catch)

 1 Where pre- and post-intervention data are reported: percent reduction is calculated by difference in differences method (see Methods); Where post-intervention data only are reported: percent reduction is calculated as: 1 - (mean density in treatment group/mean density in control group).
2 Paper states "Percentage change An. culicifacies density in treatment group before and after intervention was -58% (95% CI - 84% to + 5%)".
 
Table 5. Summary of additional results for Majambere 2010 GMB (clinical data)

OutcomeZoneIncidence or prevalence 

Rate or Risk Ratio

 

Control year

(2006)
Treatment year

(2007)
Treatment year

(2006)
Control year

(2007)

Malaria incidence1 1--70.9

(58.8 to 85.6)
7.2

(4.3 to 11.9)
9.85

(4.58 to 21.19)

230.3

(23.1 to 39.7)
17.0

(12.4 to 23.5)
--0.56

(0.31 to 1.02)

3--44.1

(35.2 to 55.2)
27.2

(20.9 to 35.4)
1.62

(1.01 to 2.61)

429.1

(22.1 to 38.4)
24.7

(18.8 to 32.3)
--0.85

(0.50 to 1.45)

Parasite prevalence2 1--41.0%

(163/398)
20.7%

(95/458)
1.97

(1.59 to 2.45)

212.2%

(54/443)
8.2%

(39/474)
--0.67

(0.46 to 1.00)

3--12.8%

(57/447)
10.4%

(47/452)
1.23

(0.85 to 1.76)

410.5%

(45/430)
22.3%

(105/472)
--2.13

(1.54 to 2.94)

Splenomegaly prevalence3 1--12.0%

(47/393)
7.7%

(35/456)
1.56

(1.03 to 2.36)

25.9%

(26/442)
6.2%

(12/471)
--0.43

(0.22 to 0.85)

3--6.5%

(29/447)
2.6%

(12/455)
2.46

(1.27 to 4.76)

45.8%

(25/434)
3.8%

(18/471)
--0.66

(0.37 to 1.20)

 1 Total cases (95% CI) per 100 person years at risk; rate ratio.
2 Parasite prevalence (total positive / total examined); risk ratio.
3 Splenomegaly prevalence (total positive / total examined); risk ratio.
 
Table 6. Summary of additional results for Majambere 2010 GMB (entomological data)

OutcomeZoneDensity or ratePercent reduction across all zones

(95% CI) 3

 

Pre-intervention year

(2005)
Post-intervention

Control year (2006)Treatment year (2007)Treatment year (2006)Control year (2007)

Adult mosquito density (measures other than human biting rate) 1 13 (0 to 7)--1 (0 to 3)2 (0 to 5)11.3 (-217.6 to 75.2)

219 (4 to 44)13 (6 to 26)13 (4 to 26)--

324 (6 to 78)--12 (4 to 31)34 (10 to 69)

411 (3 to 26)3 (1 to 11)9 (2 to 26)--

EIR 2 18.80--0.002.2417.6 (-376.1 to 85.7)

28.290.002.32--

316.55--5.8217.00

46.133.133.91--

 1 Median female An. gambiae / trap / night (interquartile range).
2 Seasonal EIR.
3 Overall percent reduction calculated using difference in differences method (see Data synthesis).
 
Table 7. Entomological data: EIR

InterventionStudy IDStudy designEIR (95% CI)Percent reduction

(95% CI)1
Notes

Pre-interventionPost-intervention


ControlTreatmentControlTreatment

Larviciding aloneCoulibaly 2011 MLI(2009 data)Cluster-RCT - -0.000.18Not estimableMonthly EIR

Coulibaly 2011 MLI(2010 data)Cluster-RCT - -2.920.4584.6Monthly EIR

Fillinger 2008 TZAControlled before-and-after trial1.05

(0.68 to 1.65)
0.81

(0.58 to 1.15)
1.06

(0.64 to 1.77)
0.56

(0.43 to 0.77)
31.5

(-59.4 to 70.6)
Annual EIR (An. gambiae)

Fillinger 2009 KENControlled before-and-after trial11.98

(7.39 to 19.40)
10.30

(7.20 to 14.95)
1.68

(1.16 to 2.42)
0.39

(0.19 to 0.79)
73.0

(22.0 to 90.7)
Annual EIR 

Geissbühler 2009 TZAControlled before-and-after trial1.44

(1.14 to 1.81)
1.18

(0.80 to 1.73)
1.24

(0.97 to 1.57)
0.80

(0.60 to 1.06)
21.3 (-42.3 to 56.4)Annual EIR 

 1Where pre- and post-intervention data are reported, percent reduction was calculated by difference in differences method (see Methods). Where post-intervention data only were reported, percent reduction was calculated as: 1 - (mean density in treatment group/mean density in control group).
 
Table 8. Entomological data: Adult mosquito density (human biting rate)

InterventionStudy IDStudy designHuman biting rate (95% CI)Percent reduction

(95% CI)1
Notes

Pre-interventionPost-intervention


ControlTreatmentControlTreatment

Larviciding aloneCoulibaly 2011 MLI(2009 data)Cluster-RCT -16.408.3749.0Mean number of bites per person per month

Coulibaly 2011 MLI (2010 data)Cluster-RCT - -41.4022.4345.8Mean number of bites per person per month

Fillinger 2008 TZAControlled before-and-after trial0.93 (0.60 to1.46)0.72 (0.51 to 1.02)0.94 (0.57 to 1.56)0.50

(0.38 to 0.68)
31.3

(-59.2 to 70.4)
Mean number of bites per person per year (An. gambiae)

Fillinger 2009 KENControlled before-and-after trial0.45 (0.28 to 0.73)0.39 (0.27 to 0.56)0.06 (0.04 to 0.09)0.014 (0.006 to 0.028)73.1

(20.3 to 90.9)
Mean number of blood fed female anophelines per person per sampling date

 1 Where pre- and post-intervention data were reported, percent reduction was calculated by difference in differences method (see Methods). Where post-intervention data only were reported, percent reduction was calculated as: 1 - (mean density in treatment group/mean density in control group).