Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
doi:10.1111/j.1365-3156.2009.02396.x
Systematic Review
Summary
objective To systematically review the effectiveness of biological, chemical and educational dengue
fever prevention programs on the reduction of entomologic indicators.
methods Searches of PubMed, GoogleScholar, CabDirect databases and reference lists yielded over
1000 articles containing mosquito abatement interventions. Inclusion criteria were: Vector control
programs targeting Aedes aegypti and Aedes albopictus mosquitoes; Studies providing pre- and post-test
data. Intervention effectiveness was assessed using Mullas formula to determine percent reductions for
all studies with control groups. Twenty-one studies were reviewed.
results Twelve dependent variables were presented, however, the Breteau, House and Container
indices were the primary measurement tools for monitoring larval populations. Behavioural methods
consisting of educational campaigns and maintaining water containers to reduce the mosquito population were applied in eight studies. Eight studies involved the use of biological methods such as
predatory organisms or bacteria. Finally, eight studies used chemical control techniques including
insecticide sprays, larvicides, insecticide-treated materials, and cleaning water of containers with
household chemicals with three studies using a combination of intervention techniques. Post-intervention reduction in entomologic indices ranged from 100% to an increase of 13.9% from baseline.
conclsuion Little evidence exists to support the efficacy of mosquito abatement programs owing to
poor study designs and lack of congruent entomologic indices. Creation of a standard entomological
index, use of clustered and randomized-controlled trials, and testing the generalizability of proven
methods are recommended for future research.
keywords dengue, Aedes aegypti, vector control, review
Introduction
Dengue fever is the most rapidly advancing vector-borne
disease in the world (WHO 2004; Kroeger et al. 2006;
Farrar et al. 2007). In 2002, the WHO estimated that
2.5 billion people worldwide were at risk of contracting
the disease (WHO 2002). Today, there are an estimated
50 million new infections annually (World Health Organization 2002; Farrar et al. 2007). Although dengue fever
is not often fatal in healthy adults, the attack rate can be
as high as 8090% during epidemics (World Health
Organization 2002). In the past, the cyclical nature of the
disease allowed roughly 1040 years to pass between
epidemics; however, the interval has dramatically
decreased due to a variety of social, economic and
political factors (Gubler & Clark 1995). Despite efforts
to control dengue, the incidence rate continues to climb
1542
Methods
Inclusion criteria
Searches in PubMed, Science Direct, Google Scholar, and
Cab Direct databases were conducted from July 2007 to
January 2008 using keywords such as dengue control,
Aedes aegypti, Aedes albopictus, mosquito control,
dengue interventions, Breteau index (BI), dengue fever
and dengue vector control to identify potential articles
and reference lists were used to locate other resources.
Figure 1 illustrates article inclusion process. Articles were
included if an intervention was implemented and quantitative pre post-data on larval or adult indices was
reported. Studies measuring mosquitoes other than
A. aegypti or A. albopictus were excluded. Study designs
were organised into cluster randomized controlled trials
(CRT) in which units were grouped into neighbourhoods
or villages and were randomly assigned to receive the
condition; randomized control trials (RCT), individuals or
Excluded (n = 1007)
Studies describing
vector behaviour,
beliefs, treatment, etc.
Included: (n = 231)
Aedes
aegypti and albopictus
control studies
Excluded (n = 177)
Cross-sectional studies
Included: (n = 54)
Intervention with
pre and post data
Excluded (n = 33)
Statistics not included
Included: (n = 21)
Larval or adult mosquito
indices
1543
units were randomly assigned to the condition; nonrandomized control trials (CT) in which intervention and
control groups were not assigned through a random
process; and, interrupted time series designs (ITS) in which
there was no control group but multiple data points were
collected before and after the intervention. Unlike the
selection criteria set forth by Erlanger and colleagues,
cross-sectional studies were not included, nor were there
geographic restrictions (Erlanger et al. 2008). Over 1000
articles were reviewed and 21 met the inclusion criteria.
Entomological indices
Among the most commonly used entomologic indices used
were the Breteau index (BI), House index (HI) and
Container index (CI). The BI was defined as the number of
positive containers per 100 houses (Lloyd et al. 1992;
Kroeger et al. 2006). The HI was defined as the number of
positive houses per 100 houses [Pan American Health
Organization (PAHO) 1994; WHO 1997]. The CI was
defined as the number of containers with immature stages
per 100 containers with water (PAHO 1994; Kroeger et al.
2006). Additional variables used to measure larval populations were the Larva Density index (mean number of
larva per container), the Ovitrap index (number of mosquito traps with eggs divided by the total number of traps
multiplied by 100; Pai et al. 2006), the Adult Density index
(number of adult mosquitoes per number of houses
surveyed; Nam et al. 2005), the Indoor Landing Density
(number of female mosquitoes caught per man-hour; Mani
et al. 2005), and the Pupae per Person index (number of
pupae collected per human population in a sector; Kroeger
et al. 2006). Serological data was not reported as the
primary dependent variable, it was given as an ancillary
measurement. Percent reductions of A. aegypti mosquitoes
were calculated using Mullas formula (Mulla et al. 1971):
100(C1 T1 T2 C2) 100.
This formula corrects for natural increases or decreases
occurring in the control group that may have similarly
affected the treatment group over time. The inclusion of
studies and statistics were independently assessed by two
authors and differences were resolved through discussion.
Results
Campaigns to increase knowledge or change human
behaviour were applied in eight studies with three
involving the screening, cleaning or disposal of water
containers. Biological interventions involving copepods or
other predatory organisms comprised another eight of the
21 studies. Although not a true biological control,
bacteria (Bacillus thuringiensis) was also included in this
1544
category due to its organic nature. Chemical interventions were also commonly implemented. Of the eight
studies within this category, five measured the effectiveness of insecticide sprays and or chemical larvicides.
Two studies measured the effectiveness of insecticidetreated materials and one used an insecticide-treated strip
in an ovitrap. Finally, three studies contained behavioural
and biological or chemical components, and therefore are
listed in both categories. Table 1 provides further information on each publications geographical location,
intervention type, study design, sample size and outcome
measures.
Study design
Three studies were CRTs, two studies were RCTs, 14
studies were CTs and two studies were ITSs. Most studies
used either clustered or non-clustered designs that compared the intervention and control groups. Three studies
employed CRTs which are considered the gold standard
design for dengue fever (Lloyd et al. 1992; Perich et al.
2003; Kroeger et al. 2006). Two studies were of questionable quality and did not contain control groups (Gould
et al. 1971; Pai et al. 2006) and one study contained only
one baseline measurement (Pai et al. 2006). The presentation of results varied greatly with only two studies
presenting confidence intervals (Espinoza-Gomez et al.
2002; Kroeger et al. 2006) and eight studies presenting
P-values (Lloyd et al. 1992; Phan-Ural et al. 1995;
Fernandez et al. 1998; Umniyati & Umayah 2000; EspinozaGomez et al. 2002; Perich et al. 2003; Mahilum et al. 2005;
Kroeger et al. 2006).
The BI was reported in eight studies, the HI was reported
in five studies and the CI was reported in seven studies.
Pupae per person was only reported in one study (Kroeger
et al. 2006).
Intervention techniques
Approximately one-third of the studies reviewed contained
educational or behaviour-altering elements. Percent
reductions using Mullas formula were calculated for five
studies (Table 2). The known monitoring durations ranged
from 18 weeks to 15 months with an average reduction of
41.6% (range 4.087.6%). Although the rates of infestation were relatively low to start, the most significant
decrease in A. aegypti populations occurred in Cuba where
a 12-month social mobilization campaign to eliminate
water containers was implemented (Sanchez et al. 2005).
Alternatively, at 15 months post-intervention, only a 4%
reduction was found for an intervention designed to teach
individuals a specific water container cleaning method
Reference
Setting
Intervention type
Borjas et al.
(1993)
El Progreso, Honduras
Dieng et al.
(2002)
Nagasaki, Japan
Espinoza-Gomez
et al. (2002)
Fernandez et al.
(1998)
El Progreso, Honduras
Focks et al.
(1987)
GorrochoteguiEscalante
et al. (1998)
Monterrey, Mexico
Biological: Mesocyclops
longisetus copepods
Gould et al.
(1971)
Chemical: ground-applied
malathion insecticide and
Abate larvicide
Kay et al.
(2002)
Kroeger et al.
(2006)
Lardeux et al.
(2002)
French Polynesia
Lloyd et al.
(1992)
Madarieta et al.
(1999)
Merida, Mexico
Behavioural: door-to-door
educational campaign
Chemical: permethrin
treated curtains
Mahilum et al.
(2005)
Biological: larvicide
Bacillus thuringiensis
Mani et al.
(2005)
CT; n = 13
neighbourhoods
(five controls)
CT; n = 42 ovitraps
(20 control) N = 7 adult
traps (two controls)
RCT; T1:n = 64 peri-domestic
drums 36 with routine water
use, 18 with no water use
(10 controls) T2:n = 28 tires
(10 controls) T3:n = 68
flower vases at cemetery
(23 controls)
ITS; n = 1108 indoor
containers n = 1318
outdoor containers
(no control)
CT; n = 2 urban communes
(one control) n = 8 rural
communes (three controls)
100 houses per commune
CRT; T1:n = 18 clusters of
houses, 1095 houses (nine
clusters controls) T2:n = 18
clusters of houses, 1122
houses (nine clusters controls)
ITS, no data on sample size
given; no control
information given
CRT; n = 12 four-block
communities (six controls)
CT; n = 130 households in
two barangays (65
households in one
barangay control)
CT; n = 11 breeding sites
(containers selected for
control)
CT; n = 3 residential
communities, 216260
homes in multistory
buildings per community
(one control community)
Outcome
Measurement
Cumulative Larva
Density Index
for all containers
Larva Density
Index
BI and KAP
CI, Washbasin
Infestation
Index
Adult mosquitoes
per trap per hour
CI
CI
BI and serological
testing
BI and HI
BI, HI and CI
Larva Density
Index and
serological testing
BI and Landing
Density
1545
Table 1 (Continued)
Reference
Setting
Intervention type
Outcome
Measurement
Nam et al.
(2005)
Pai et al.
(2006)
Three districts of
Kaohsiung, in
South Taiwan
Areia Branca and
Nilopolis,
Brazil
ITS; n = 90 households in
three districts (no control)
Perich et al.
(2003)
Phan-Ural et al.
(1995)
Sanchez et al.
(2005)
Amphoe Khlung,
Thailand
Havana City, Cuba
Sulaiman et al.
(2002)
Umniyati and
Umayah (2000)
Kuala Lumpur,
Malaysia
Yogyakarta,
Indonesia
Biological: larvicide
Bacillus thuringiensis
Behavioural: social
mobilization and
education
Chemical: cypermethrin and
cyfluthrin insecticide spray
Behavioural: emptying,
washing and removing water
containers with mosquito larva
CT, non-randomized controlled trial; CRT, clustered randomized control trial; ITS, interrupted time series; RCT, randomized controlled
trial; T, treatment group; n, sample size.
Table 2 Summary of results of behavioural interventions
Duration
Mullas
percent
reduction
unknown
73.1
15 months
4.0
6 months
22.9
12 months
87.6
1.7
28.2
18 weeks
33.1
34.4
54.1
29.9
26.7
18 weeks
28.6
35.2
29.3
Reference
Index
Baseline
measure
Espinoza-Gomez
et al. (2002)
Fernandez et al.
(1998)
BI
BI
CI
125.5
102.2
26.6
38.2
115.5
19.7
CI
31.4
24.2
BI
BI
HI
126.0
113.9
3.7
129.0
151.3
0.6
HI
OI
1.3
48.5
OI
OI
OI
Lloyd et al.
(1992)
Sanchez et al.
(2005)
Umniyati and
Umayah (2000)
Last
measure
T, treatment group; C, control group; BI, Breteau index; CI, Container index; HI, House index; OI, Ovitrap index.
Baseline
measure
Last
measure
Duration
Mullas
percent
reduction
2 months
100.0
3
3
3
3
months
months
months
months
100.0
100.0
100.0
100.0
21 months
92.0
21 months
()
Reference
Index
LD
20.0
0.0
LD
LD
LD
LD
LD
LD
BI
42.0
0.1
1.6
3.7
0.1
14.7
57.0
51.5
0.0
0.0
0.0
0.0
16.7
3.0
BI
BI
53.0
0.0
35.0
0.0
BI
BI
BI
BI
BI
AD
32.0
10.0
15.0
23.0
25.0
1.2
0.0
1.0
2.0
0.0
30.0
0.0
AD
AD
AD
BI
BI
0.1
0.1
0.7
429.8
344.5
0.0
0.0
0.4
52.0
167.1
Phan-Ural et al.
(1995)
21
21
21
21
months
months
months
months
100.0
91.7
88.9
100.0
3 years
100.0
3 years
3 years
100.0
100.0
4 months
75.1
T, treatment group; C, control group; BI, Breteau index; CI, Container index; LD, Larva density index; AD, Adult density index.
*Contain both biological and behavioural (integrated) interventions.
Reference
Index
Espinoza-Gomez
et al. (2002)*
BI
BI
BI
MT
MT
BI
110.8
53.0
102.2
12.3
9.3
60.0
102.2
32.6
115.5
11.9
7.9
7.0
Unknown
9.9
BI
BI
113.0
38.0
12.0
11.0
Unknown
42.1
BI
BI
BI
BI
BI
BI
BI
CI
CI
CI
CI
34.0
131.0
98.5
76.7
40.0
50.0
55.0
6.1
4.8
10.3
13.2
17.0
46.0
38.8
55.0
42.5
37.5
47.2
5.1
10.0
3.6
17.6
5 months
10.9
14 days
32.5
14 days
12.7
4 months
59.9
4 months
73.8
Last
measure
Mullas
percent
reduction
Baseline
measure
Duration
unknown
unknown
7 days
18.4
45.6
13.9
T, treatment group; C, control group; BI, Breteau index; CI, Container index; MT, adult mosquitoes trap per 24 h.
*Contain both chemical and behavioural (integrated) interventions.
Data is given using the Container index, yet the description matches that of the Breteau index: proportion of containers positive per house.
Insecticide sprays
Treated materials
Adult
mosquito
Source reduction
Water removal
Ovitraps
Ova
Human
Education
Repellent
Adult
mosquito
Turtles
Larva and
Copepods
pupa
Fish
Larvicide
Chemical interventions
It is widely accepted that insecticides are effective at
reducing mosquito populations; moreover, the results can
be seen and felt immediately (Phuanukoonnon et al. 2006).
The use of chemical larvicides effectively reduces mosquito
populations; however, no articles included in this study
measured the effectiveness of larvicides alone. The results
of Mullas percent reductions were the most diverse for
chemical interventions, compared to all other categories. In
some instances, larval populations decreased dramatically
(73.8%), while some populations increased (13.9%). This
finding illustrates the frustration of having limited study
Limitations
The studies were limited by the lack of congruency and
inherent weaknesses in the entomological measurements.
The BI, the HI and CI under-estimate the mosquito population by failing to describe larva densities (Service 1976;
Tun-Lin et al. 1996; Espinoza-Gomez et al. 2002). The
recent trend in dengue research has turned away from the use
of larval indices in favour of pupal indices because of the high
correlation between pupae and adult mosquitoes. Being able
to easily identify pupae of a particular species and to count
the absolute number of individuals makes this index
1549
1550
Corresponding Author K. Ballenger-Browning, C o John Elder, Graduate School of Public Health, San Diego State University, 9245
Sky Park Court #221, San Diego, CA 92123, USA. Tel.: +1 6195328777; Fax: + 6195326030; E-mail: kara.browning@med.navy.mil
1551