Community-Effectiveness of Temephos for Dengue Vector Control: A Systematic Literature Review

The application of the organophosphate larvicide temephos to water storage containers is one of the most commonly employed dengue vector control methods. This systematic literature review is to the knowledge of the authors the first that aims to assess the community-effectiveness of temephos in controlling both vectors and dengue transmission when delivered either as a single intervention or in combination with other interventions. A comprehensive literature search of 6 databases was performed (PubMed, WHOLIS, GIFT, CDSR, EMBASE, Wiley), grey literature and cross references were also screened for relevant studies. Data were extracted and methodological quality of the studies was assessed independently by two reviewers. 27 studies were included in this systematic review (11 single intervention studies and 16 combined intervention studies). All 11 single intervention studies showed consistently that using temephos led to a reduction in entomological indices. Although 11 of the 16 combined intervention studies showed that temephos application together with other chemical vector control methods also reduced entomological indices, this was either not sustained over time or–as in the five remaining studies—failed to reduce the immature stages. The community-effectiveness of temephos was found to be dependent on factors such as quality of delivery, water turnover rate, type of water, and environmental factors such as organic debris, temperature and exposure to sunlight. Timing of temephos deployment and its need for reapplication, along with behavioural factors such as the reluctance of its application to drinking water, and operational aspects such as cost, supplies, time and labour were further limitations identified in this review. In conclusion, when applied as a single intervention, temephos was found to be effective at suppressing entomological indices, however, the same effect has not been observed when temephos was applied in combination with other interventions. There is no evidence to suggest that temephos use is associated with reductions in dengue transmission.


Introduction
Vector control remains the only available intervention to prevent and control the transmission of dengue [1].Various vector control strategies aiming at controlling the principal vector of dengue, Aedes aegypti, are currently used with the intention of preventing the occurrence of dengue, or controlling outbreaks. These vector control measures often include the application of chemical or biological agents for the control of immature and adult mosquito stages, or environmental control methods that target mosquito breeding sites [2]. These vector control measures can be applied as single interventions or in combination. [3]. However, the efficacy and community-effectiveness of vector control strategies in terms of reductions in dengue transmission remain unclear, as previous systematic reviews have reported regarding the application of single intervention methods such as peridomestic space spraying and the use of Bacillus thuringiensis israelensis [4,5].
One of the most commonly employed methods for dengue vector controlis the use of the organophosphorous compound temephos (commercial name Abate) as a larvicide. Its use has been documented since 1965 [6] in ponds, marshes and swamps at a dosage of 0.1-0.5 kg/ha for vector control in general, although fewer studies exist in relation to Ae. aegypti. Per the WHO Pesticides Evaluation Scheme, temephos can be used safely in potable water when the dosage does not exceed 56-112 g/ha (5.6-11,2 mg/m 2 ) or 1 mg/l [7]. Moreover, the WHO hazard classification of temephos is "U", meaning it is unlikely to cause acute hazard under conditions of normal usage [8].Temephos is a widely preferred tool for several reasons, including its ease and simplicity of application, selective killing of mosquito larvae and its long lasting effect when compared to traditional oil application methods [6].Temephos is commercially available in standardised preparations such as emulsifiable concentrates, dilute solutions, dusts and granules, including slow release formulations. It can be applied in different ways depending on the site and rate of application required. It can be delivered by hand or by injection through drip system devices or power sprayers. Temephos sand granules can be applied to household water storage containers of varying capacity by using a calibrated plastic spoon in order to administer a consistent dosage of 1ppm(1 ppm = 10 −6 = 1 parts per million = 0,0001%) [9].
Temephos has widely been considered a cornerstone for controlling immature forms of Ae. aegypti yet while its efficacy has been demonstrated under laboratory conditions, comparable levels of efficacy are not necessarily replicated under field conditions [10].For the purpose of this article, efficacy under field conditions, including the use in ongoing control programmes, is referred to as community-effectiveness.
This systematic literature review assesses the community-effectiveness of temephos for controlling dengue vectors and dengue disease transmission when delivered in the field as a single intervention or in combination with other interventions.

Methods
This systematic literature review follows the reporting guidelines described in the PRISMA statement [11].A comprehensive systematic literature search protocol was developed and agreed on by the authors consisting of six databases (PubMed, WHOLIS, GIFT, CDSR, EMBASE, Wiley).An additional review of grey literature,-screening the reference lists of the included publications as well as asking stakeholders about relevant literature-was performed, including theses, unpublished data, and other reports. The search was conducted until 15 June 2013. All sources fulfilling the predefined inclusion criteria and exclusion criteria were cross checked for additional references and these were included if again fulfilling the inclusion and exclusion criteria.
No language restrictions were applied and abstracts of publications in languages other than English were translated. The search included all studies irrespective of the year of publication.
The literature search strategy was based on four categories1) dengue vectors (Aedes aegypti or Aedes albopictus),2) vector control intervention (temephos),3) dengue disease, and 4) dengue prevention and control. The search was conducted using the appropriate Medical Subject Heading (MeSH) terms followed by the Boolean operator "OR" for terms within each of the 4 categories, "AND" between categoriescombined with "free text" terms. The terms used for the 'vector' category included Aedes aegypti and Aedes albopictus. The search terms for the 'vector control intervention' category included insect control, vector control, mosquito control, larvicides, temephos, temefos and Abate. For the 'dengue disease' category, the terms dengue, dengue fever, DF, dengue hemorrhagic (haemorrhagic) fever, DHF, dengue shock syndrome and DSS were used. These terms were used in different combinations together with the terms prevention and control in order to broaden the search.
All titles and abstracts of potentially relevant articles were initially screened for the relevance of the research question and irrelevant articles as well as duplicates were excluded. The remaining articles that fulfilled the inclusion criteria were assessed, extracted and analysed using the full text of the study. The processes were conducted by two independent reviewers in consensus.
The inclusion criteria for this review were as follows: (i) Studies or programmes conducted with aim to prevent/control dengue; (ii) Studies with quantitative outcomes such as Breteau Index (BI), Container Index (CI), House Index (HI), larval mortality indicated by pupal skins, average number of positive containers per house, pupal index, indoor resting density, ovitrap index or dengue incidence; (iii) Community-effectiveness studies; iv) Peer reviewed studies with the study designs such as Randomised Control Trials (RCT), Cluster-Randomised Controlled Trials (CRCT), Non Randomised Control Trials (NRCT), Before and After studies, Studies with an Intervention and a Control area(Intervention studies); (v) Studies where temephos was used as a single intervention or in combination with other interventions.
Exclusion Criteria were: (i) Studies based in laboratory or semi-field settings; (ii) Studies where temephos was not used alone or as part of a combination intervention; (iii) Studies without clearly specified outcome measures; (iv) Cross-sectional studies, case series, reports, letters, newspaper articles, lectures, conference reports or abstracts.
After applying exclusion criteria, all included articles were categorised as either single or combination intervention studies and tabulated in an evidence table and stratified by type of study. Information on the study setting, objectives, design/sample size and study period as well as outcome measurements, main results and conclusions has been extracted.
Although considering the study quality by reflecting the study types as weighting tool for the discussion, no articles were excluded because of quality, taking the relative scarcity of relevant material into consideration.
To ensure the quality of this systematic review, the tool for assessment of systematic reviews (AMSTAR) was used [12].

Results
The systematic literature search generated 18,439 potentially relevant citations. After screening by title and abstract, application of the inclusion criteria and removal of duplicates, 54 studies were retrieved for full text evaluation. After the final application of the inclusion and exclusion criteria, 20 articles were included and 7 further articles were added from the cross references and grey literature for a final total of 27included studies (Fig 1).

General study characteristics
Of the27 studies, 14 were conducted in the Americas, 10 in South East Asia, one in Europe and two in the Western Pacific. Nine of the South East Asian studies were conducted in Thailand (Three by one researcher:Y.H.Bang).Even though no language restrictions were applied, all 27 articles retrieved were in English. All studies were publishedbetween1971 and 2012, of which seven studies were published between 2000 and 2012.

Categorisation of selected studies
Of the 27 studies , 11 studies used temephos as a single intervention (Group A), while 15 studies used temephos in combination with other vector control methods (Group B). Only one study [21] tested temephos both alone and in combination, and this study has been included in the combination intervention group. The studies are described in detail according to the two groups (Tables 1 and 2).
Outcome measures. In most studies, the classical immature entomological indices such as the House Index [16,22,23], Container Index [13,15,16,18,19,22,23]and Breteau Index [16,19,22,23] were used to measure the outcomes. Pupae/Person Index [13,23] and adult mosquito density [16,18,19] were also measured. Reduction of breeding sites [14] and positive  ovitraps were also measured [14,19,20,22,23]. Mean number of (total, late and early stage) Aedes larvae [13,17,19,20], % prevalence of Aedes [20]and mean % reduction of Ae.albopictus [21] were some of the other outcome measures that were analysed. Other findings such as quantity of temephos, labour and time required for its application were analysed only in study number [19]. Factors affecting the residual efficacy of temephos (such as temperature, organic debris, exposure to sunlight, water use patterns, maintenance of water storage containers) and level of acceptance or resistance to use of temephos among household residents were analysed only in one study [15], and expenditure for temephos application was calculated only in study number [19]. Study number [13] was the only study that analysed the association between larval mortality and factors such as the mode of application of temephos using spoons or zip-lock bags, different water types, effects of sunlight and type of containers used for water storage.
Interventions. The 11 studies had different temephos application periodicity, with four studies [14,16,20,23] conducted according to the dry and wet seasons while study number [23] applied temephos as a pre-seasonal focal treatment. Of the 11 included studies, five had used temephos in the form of sand granules [13,17,18,19,22], while study number [15] experimented with temephos zeolite formulations. Four of the studies [14,16,20,23] did not mention the type of formulation and only mentioned the dose (as parts per million (10− 6 or ppm), and in study number [21] different dosages of temephos were compared. In most of these studies, temephos was applied at a dose of 1 ppm based on the size of the containers [18,19]. However, in study number [23], temephos was applied based on the actual volume of water present at the time of inspection rather than the total holding volume of the containers. The mode of temephos application differed across studies and included such diverse methods as the use of permeable zip-lock bags [13], plastic spoons [19], granules placed in a well-perforated plastic micro-centrifuge tube that allowed the granules to be removed during monitoring procedures [20], corncob grit granules [21] and the application of wettable powder to walls [22]. In studies where temephos was compared to other insecticides [17,20], it was often compared to Spinosad [17,20] and Bti [20].
Sources of water and the container material also differed across studies. Study number [13] compared the effects of temephos in rain water, piped water or both and also its effects in fibrocement or plastic containers. Studies number [18,19] applied temephos in water which was not used for the purpose of drinking.
Community-effectiveness. All 11 studies reported a post-intervention reduction in the immature stages when compared to their respective control group. It was observed that the treated sources were free of larvae for a variable period of time depending on the season of application [16,20,23], number of applications [14,19], dosage of temephos [21], procedure of control [19], and method of application with respect to the source of water and container for storage of water [13].
For example, it was observed in study number [16] that the Aedes population was effectively controlled for 4-8 weeks in the dry season with no larvae and 6-12 weeks in wet season. In study number [20], there was an absolute inhibition of mosquito development for 6-9 weeks in the dry season and 9 weeks in the wet season. However, study [24] showed a pre-seasonal focal treatment before the rains reduced indices for 9-11 weeks, after which they returned to pretreatment levels.
When considering the number of applications, study [14] reported that as the number of applications increased to five, it led to the elimination of the immature population. Study [19] found that a total of four applications could be effective at suppressing entomological indices, but concluded that two treatments would most likely be effective if applied just before the rainy season and repeated again within 2 months.
Variable effectiveness of temephos according to dosage was reported in study number [21], when it was applied at half the maximum allowable dose (11.21kg AI/ha)it provided100% protection for 5 months, with excellent control at rates as low as 0.56kg AI/ha(100% reduction sustained for 8 weeks). Study number [17] reported 100% effectiveness for 90 days when temephos was applied at the dose of 20g/200 litres of water.
Study number [19] observed that cyclic mass treatment with temephos required a greater amount of larvicide, while the targeting of treatment to positive containers required a greater number of man-hours. The effective control period was 2.5 to 5 months with an average of 3 months. The study concluded that a combined method of cyclic mass treatment at 3 month intervals with simultaneous treatment of new habitats between the mass treatments would offer the best control.
Factors such as temperature, organic debris, and ultraviolet light from sunlight that possibly degraded the active compound were investigated in study numbered [15]. These factors coupled with rapid water use, draining, and refilling of containers shortened the residual effectiveness of temephos [15]. This study also pointed out that other factors such as regular supply and budget allocation for the purchase of temephos are major factors affecting the success of a programme. Likewise, in study number [13], the median duration of the temephos effect was two to three weeks, with the duration of the residuality affected by water management practices (water type and water turnover), which were found to be significantly associated with reduced larval mortality. This same study found that temephos applied inside zip-lock bags provided longer residual activity as compared to spoon based applications [13].
Study number [22] illustrated the community-effectiveness of temephos by eliminating Ae. aegypti from an island by applying temephos intensely as both a residual insecticide on walls of houses as well as the treatment of breeding sites. Through this combination of temephos treatments, Ae. aegypti was eliminated within 15 months of the implementation of the campaign.
Acceptability. Very few studies evaluated the acceptability of temephos applications [13,15]. As per study [13], acceptability of householders for temephos application via zip-lock bags was higher than spoon based application since it was easy to apply and inexpensive to implement. In study [15], 89% of the participants refused to use temephos sand granules due to the unpleasant odour and increase in water turbidity and perceived safety risk. No studies analysed the operational challenges associated with temephos application or the undesirable consequences caused by it.
Interventions. Of the four Intervention Control Studies [25,28,30,32], no interventions were applied in the control group of one study [28]. Ultra low volume (ULV) pyrethroid thermal fogging [25], Bti and Methoprene-S [30] and 4% Malathion fog [32] were applied in the control groups of the other three Intervention Control studies. Among the three Cluster Randomised Trials [29,33,39], two studies [29,39] used temephos in combination with other interventions as a control group for a different intervention under study. Health education with source reduction was applied in the control group of the third cluster randomised trial [33].
Sample size was highly variable, with the number of containers or premises varying from 235 [24] to the coverage of entire islands [30,31,35].Follow up periods ranged from 6months [28] to 9 years [34] post intervention. Only one study [25] used a parameter for dengue transmission (DHF), measuring in communities the basic knowledge of dengue in relation to previous occurrence of DHF cases.
All combination intervention studies except for two [24,25]used temephos at a dose of 1ppm. All studies used temephos sand granules except for two studies [24,35],which used temephos as emulsifiable concentrate. Only in study [30] temephos had been provided routinely before the intervention. In study [38],temephos was supplied for a nominal fee to households. In study [36],temephos was partially substituted with methoprene and in study [34], temephos was replaced with Bti because of its unpleasant odour. In study [27], the focal treatment with temephos was interrupted due to a dengue outbreak and supplemented with ULV spraying.
Most studies [24,25,26,27,29,33,34,35,36,37,38,39]used different community members/ groups either as a target group for education or for the delivery of the intervention. Study [39] formed a community working group exclusively for the implementation of the control activities in order to analyse the importance of community participation.
For example, the effect of temephos on larvae in study [27] lasted 6 weeks, with infestations in treated containers detected after 6 weeks of treatment, whereas study [28]showed an effect of 6 months, with an observed increase of adult mosquito population 6-7 weeks post-treatment. In study [32], the effect on larvae lasted 8 weeks, study [35] showed a variation between 2-8 weeks, and for study [38] the effect waned after 3 months.
Four studies stratified their analyses according to the individual effect of each intervention, showing that temephos was the most effective intervention alone [32] or in combination with source reduction [24,30] and ULV spraying [27].
From those reporting positive results, however study [28], while reporting successful outcomes in the short term, concluded that long term success depended on political commitment and community participation. Study [29] showed that factors such as high water turnover and temephos resistance in local vector populations contributed to decreased community-effectiveness. Another limiting factor was the limited knowledge of people about the interventions [25,38].
In the group of studies reporting shortcomings these included problems related to surveillance and coverage [27],inadequate source reduction and larvicidal application [30], low acceptability of larvicides [36] and lack of resources and manpower [31].Other challenges included the limited residuality of temephos [27] and overflowing of water in containers resulting in consequent dilution of the active ingredient [33].
One important challenge raised regarding the use of temephos in combination interventions described in both groups was the false sense of security that can arise, leading the population to believe that temephos application alone is sufficient to prevent dengue which resulted in foregoing activities such as source reduction and other environmental management [24,33].
Acceptability. Acceptability of temephos was not assessed in all studies but was found to be low in three studies [34,36,37] due to unpleasant odour. For this reason, study [34] replaced temephos with Bti as an intervention, and in study [36],temephos was partially substituted with methoprene. Also, the application of temephos was found to be labour intensive in study [24].

Limitations of the study
Studies on the efficacy of temephos were not assessed and only its community-effectiveness was analysed. However, temephos is used routinely in many parts of the world as a part of dengue vector control activities, and hence this systematic review focused on evidence for its community-effectiveness in the manners in which it has been applied routinely. Although we used a broad search strategy we could have missed potentially relevant studies. We also assume some publication bias towards studies demonstrating a positive effect.
The variability of the outcome measures encountered and the different larval and pupal indices used also limit the comparability of the studies, especially when relating outcomes to dengue transmission [40]. Moreover, very few studies monitored the changes in reported cases of dengue. Among the combined intervention studies, only four studies clearly stated which interventions were found to be effective, while the rest of the studies failed to disaggregate the data.

General discussion
Overall, a diverse picture emerged with regards to the community-effectiveness of temephos reported in the 27 reviewed studies. Whilst the single intervention studies showed consistently that using temephos led to a reduction of entomological indices, this did not always hold true when applying temephos in combination with other interventions. For the latter group, three studies [30,31,33] clearly stated that temephos application along with other chemical vector control methods failed to reduce the larval and pupal indices, and while a further 10 studies reported a post-intervention reduction in immature mosquito stages, the results were either not sustainable over time or the coverage was not complete. The reasons for this can be manifold, and this has important implications for dengue control programmes and raises further questions: were the reasons operational, since when applying combined interventions the focus may perhaps shift from quality to quantity, or are there yet unknown interactions between the different interventions? Or was it simply because of limited resources? A further implication for dengue control programmes is the unknown epidemiological impact of temephos-based interventions, only one study linked basic knowledge of dengue to a reduction of DHF [25].
The effectiveness of temephos interventions depends on many factors related to the quality of delivery and maintenance of the intervention, including water turnover rate and type of water, as well as environmental factors such as organic debris, temperature and exposure to sunlight. This suggests that quality control and the suitability of application sites are key to the ultimate success of the intervention. However, the reviewed studies reporting the residual effect of temephos showed a duration between two and three months, confirming similar findings upon which operational guidelines have been based [9]. This recommended periodicity could have practical implications, however, as frequent re-treatment led to reluctance of temephos use in one of the reviewed studies [22]. Another study [19] suggested that a combined method of cyclic mass treatment at three month intervals with targeted treatment of new habitats between the mass treatments would offer optimal control.
The timing of the use of temephos is another factor to consider: dengue epidemics tend to occur in warm, humid and rainy seasons, favouring the growth of mosquito populations [41]. Hence, the timing of temephos applications can influence the efficacy of dengue vector control, as demonstrated by studies [14,18,23]which suggested that temephos application at the beginning of the rainy season was most likely to control the occurrence of an epidemic, while two studies [16,20]reported the successful control of Aedes species during both the dry and wet seasons.
The method of temephos application recommended by WHO is the use of calibrated plastic spoons to apply the appropriate doses to water-holding containers [9]. However, this review has shown that many different formulations and methods of temephos application can be successfully used, including zip-lock bags [13]which had longer residual effects and were cheap and easy to apply. The application of temephos using spoons [13,18]was reported to create an unpleasant odour, taste and turbidity-a disadvantage not seen with temephos zeolite formulations [15]. Using corncorb grit granules with a blower [21]was cheaper, required less manpower and was effective due to its dispersion. In the Ae. Aegypti elimination programme carried out on an island, many temephos delivery systems were successfully applied, such as emulsion paint on walls, perifocal application of untreated areas, larviciding of water containers [22]. In one of the reviewed studies [21], temephos was also found to be effective when it was applied at half the recommended dosage.
Coverage of potential breeding sites is a recurrent issue: inaccessibility of breeding sites for temephos treatment can be important in cases such as leaf axils, brick or rock holes, miscellaneous containers such as cans, bottles, tiers, flower vases etc. Another aspect of coverage is the question of distance between the treated household and untreated areas. In a study related to Ae. aegypti dispersal, Reiter et al. [42] showed that maintaining a treated barrier zone of 50-100 meters around the house of a dengue case is unlikely to be effective, as Ae. aegypti can fly much further to oviposit. This was evident in one of the studies [16]where there-infestation of a treated area occurred despite the maintenance of a barrier zone of 87 houses.
Operational aspects, such as cost, supplies, time and labour efforts are further issues limiting the potential effectiveness of temephos application [19,29],as is low community acceptability (such as reluctance to use temephos in drinking water) [15,24,34,36,37]. These factors are very important considering that for any community based programme to be successful, it needs to be widely accepted by the people with whom the intervention is being carried out [43], and that programmes need to reflect the "felt needs" of the people in order to maintain their interest, motivation and long-term engagement [44].
No conclusive evidence was attained regarding which intervention was the most effective, nor was it possible to come to a conclusion regarding which combinations of interventions formed the best package in terms of effectiveness, feasibility, cost and sustainability. For the group of studies failing to show effectiveness of temephos combined with other interventions, the failure was attributed to false complacency arising from the perception that temephos was sufficient to control the vector, neglect of source reduction activities [22] and low acceptability of temephos application in potable water [36]. This highlights the importance of community participation, as has been proposed with "Communication-for-Behavioural-Impact" (COMBI) efforts [45].The findings of this study on community participation were similar to those of a systematic review conducted by Heintze et al. [46],which suggested that community-based control strategies implemented together with other interventions were able to reduce classical Aedes larval indices, but they were unable to disentangle the effect of different interventions and community participation.
The role of entomological surveillance is another important factor: WHO recommends the implementation of an integrated dengue surveillance and outbreak preparedness system [47]. The importance of implementation and maintenance of a surveillance programme was highlighted in study [22], which reported the importation of viable Ae.aegypti eggs through a crate of household articles and through trade delivered via the ports.
In conclusion, this review presents more questions than answers regarding the communityeffectiveness of temephos for dengue vector control. While there is little doubt concerning the effectiveness of temephos in controlling Aedes breeding sites, the same level of effectiveness was not clear from the studies using temephos combined with other interventions. No conclusive evidence has been shown regarding the impact of temephos interventions on dengue transmission.
This review highlights that although temephos is one of the most widely used interventions against dengue vectors worldwide, its effectiveness can vary greatly. The lack of data relating reductions in entomological indices to reductions in dengue transmission remains a significant knowledge gap in the area of dengue epidemiology and vector control efficacy. Integrated surveillance activities may need to be implemented along with vector control interventions, in order to address this knowledge gap.