Insecticide-impregnated bed nets for malaria control: a review of the field trials.

Update/Le point Insecticide-impregnated bed nets for malaria control: a review of the field trials* A. Bermejo1 & H. Veeken2 Insecticide-impregnated bed nets act as a physical barrier to repel and kill mosquitos. Community inter- vention trials suggest that these nets are effective in preventing malaria-related mortality and morbid- ity-but not malaria infection-in areas of low and moderate transmission; the results from areas of high transmission are not so encouraging. Comparison of the results from these trials and their interpretation are difficult because of variations in the epidemiology of malaria and several methodological flaws. Problems such as defining appropriate health indicators, monitoring bed-net usage, introducing bed nets randomly, selecting adequate controls, performing statistical analysis, and comparing bed nets with other available interventions are considered. Further community intervention trials are needed, paying attention to the methods and to assessment of their impact on malaria. Malaria is by far the most important insect-borne disease, 100 million persons being infected each year (1). Since most malaria-transmitting mosquitos bite indoors at night, insecticide-impregnated bed nets might provide useful protection. The assessment of community interventions which aim at malaria control is fraught with difficulties. Variations in the epidemiology of malaria and the use of diverse clini- cal and entomological measures make it difficult to compare and interpret the results. Consequently, the value of impregnated bed nets is still a matter of controversy (2). This review of field trials that assessed the potential of impregnated bed nets and curtains for malaria control summarizes their find- ings and highlights the methodological problems and areas where more research is needed. Mode of action Pyrethroid-impregnated bed nets and curtains reduce man-vector contact by acting as a physical barrier and by repelling mosquitos and driving them out of houses (3, 4). Impregnation improves the effective- * A French translation of this article will appear in a later issue of the Bulletin. 1 Current post: Delegate of the Spanish Red Cross in Bolivia. Requests for reprints should be sent to Dr Alvaro Bermejo, San Cayetano 3 40 D, 28005 Madrid, Spain. 2 Public Health Consultant. Reprint No. 5278 ness of a torn bed net (5) and prevents mosquitos from feeding on a limb which may touch the bed net during the night (6). Community-wide distribution of impregnated bed nets, besides providing individual protection to the persons sleeping under them, can have a profound effect on the transmission of mala- ria ("mass effect") by killing mosquitos and thus reducing the mosquito population. Such a reduction in mosquito density, which has been seen in several studies (7-11), would benefit the entire community, even people who do not use nets. Field trials Results The degree to which impregnated bed nets protect against malaria depends on local circumstances. The reduction in the number of infective bites is greater in areas where mosquitos bite indoors, late at night, and where they are not exclusively anthropophilic (12). These characteristics may be modified by the intervention; a significant trend towards outdoor biting, earlier in the evening, has been observed (10, 11) but the evidence is inconclusive. Entomological studies show that impregnated bed nets reduce man-vector contact and diminish mosquito popula- tions. To assess if this is of any health benefit, their impact on malaria morbidity and mortality has been measured. Bulletin of the World Health Organization, 70 (3): 293-296 (1992) © World Health Organization 1992 293 A. Bermejo & H. Veeken In some studies (13-15) impregnated bed nets were distributed to some individuals, with others in the same community as controls. However, in most field trials impregnated nets (7-9, 16-19) or curtains (20) were distributed to entire communities in order to include the "mass effect" in the assessment of their health impact. Table 1 summarizes the impact on malaria morbidity and mortality found in the lat- ter series of trials. The Table includes those studies which provide enough information to evaluate the method. The results showed consistent decline in the pre- valence of malaria infection, but parasite density and malaria incidence were significantly reduced in most trials. These results suggest that the widespread im- pregnation of bed nets or curtains is more effective in reducing the number of 'heavy' infections that result in clinical episodes than in preventing infec- tion. The only trial that measured malaria mortality showed that the protection was greater against malaria-related mortality than malaria morbidity (17). The reasons for this are not fully understood; it could be due to a reduction in the average sporozoite load from infective mosquitos (8) or to protection against super-infection, which would account for the severity of the episode in areas of high transmission. Further research is needed to clarify the relationship between sporozoite inoculum (size, number and frequency) and infection, disease, and mortality (21). In areas of high transmission, protection against clinical episodes and malaria death may be the best outcome since the harmful effects of malaria are reduced without affecting the acquisition of immun- ity to the disease (22). Table 1: Summary of the indicators that measured the health impact of impregnated nets In eight trials (7-9, 16-20) in which nets were distributed to entire communities References of studies whose results were: Significant Indicator (P <0.05) Not significant Spleen rate 8 20 Prevalence of malaria 19a, (16a) b 7, 8, 18, 19 a, infection 20, (16 a) Parasite density 7, 20, (16 a) (168) Incidence of clinical 7, 8, 16a, 18c malaria 20, (9) Child mortality 17 a Significant in some age groups or communities but not in others. b References in parentheses indicate that significance tests were not done, the results showing a marked or no marked decline. c Incidence of malaria infection. Methodological problems Randomized controlled trials are the standard way of evaluating new therapeutic agents and procedures at the individual level (23). Extending this method to interventions at community level is difficult, particu- larly when trying to measure the health impact of malaria control interventions. Several methodologi- cal flaws hamper the drawing of definite conclusions from the trials. There are discussed below. (i) Health indicators. The most commonly used indicator of health impact is the incidence of clinical malaria. Neither the definition of a clinical attack of malaria nor methods of case detection are standard- ized. One trial (8) actively looked for people with fever (> 37.5 °C) and > 5000 parasites/ml, while another (7) used self-referred fever cases (> 38 OC) and > 10000 parasites/ml. Given that in areas of holoendemic malaria healthy parasite carriers are common and that impregnated bed nets provide bet- ter protection against 'heavy' infections, differences between the results of some of the trials could partly be accounted for by the different way they looked for cases (12). The use of child mortality as an indicator of health impact is most promising. However, it is confounded by the indirect effect of malaria on mor- tality and by competing causes of death. A study from Guyana (24) suggests that malaria has an indi- rect impact on mortality from acute and chronic re- spiratory diseases and on neonatal mortality (through the health of the mother). This fact can explain why the overall reduction in mortality found in Gambia (17) was greater than what could be expected from the prevention of deaths directly attributable to mala- ria. The Garki project illustrates the problem of com- peting risks. Before the intervention, Plasmodium falciparum was the immediate cause of many of the infant deaths, but the control of malaria prevented only a small number of deaths because of increase in mortality from other causes (25). This kind of com- petition limits the possibility of extrapolation from the short-term effects of intervention trials that aim at the removal of a single cause of death. (ii) Bed-net usage. Distribution of bed nets or curtains does not guarantee that they will be used. Failure to record whether the bed nets were actually used (7-9, 18, 20) is remarkable; in fact, in one study (18) the lack of impact on malaria morbidity was attributed, in a follow-up study (26), to a failure to use the bed nets. Malaria is markedly seasonal and has an uneven age distribution, which most trials have taken into consideration; but none of them has taken into account that the use of bed nets and cur- tains may also be seasonal and age-dependent. Future trials must include this issue. WHO Bulletin OMS. Vol 70 1992294 Impregnated bed nets for malaria control (iii) Randomization. The use of random alloca- tion ensures that confounding factors are equally likely to be allocated to the intervention or compari- son groups. Randomization is specially desirable when variables such as child mortality are being measured, as it is very difficult to control for all the possible confounders. In most studies (7, 9, 16-20) the intervention was not introduced randomly and they are thus liable to bias. However, the logistical difficulties of introducing community interventions on a random basis should not be overlooked. A true double-blind trial is not a practical option: the reduction in insect nuisance is likely to be so substantial that neither the study population nor the field staff can remain unaware of which com- munities had the treated bed nets. However, the blood slides and questionnaires referring to cause of death should be coded before they are read, so as to reduce possible bias at this stage. None of the trials mention whether this was done. (iv) Controls. Malaria transmission varies unpredictably from year to year and from village to village. Because of this annual variation the use of historical controls is inadequate and contemporary controls are required. In three trials (9, 18, 20), the baseline data gathered prior to the intervention were insufficient to establish the comparability of the control and intervention communities. In two of the trials (8, 17), baseline data showed the two groups to be different from the outset. It is doubtful whether these controls truly estimate what would occur in the intervention communities if impregnated bed nets had no impact-particularly as these nets had not been randomly allocated. (v) Sample size. When assessing the potential of community-wide distribution of impregnated bed nets for malaria control, each individual or house- hold is not an independent unit. The sampling unit is the community or village and analysis should be based on the differences between communities (23). Significance tests that assume independence between each person surveyed should not be used.a Several studies (7, 9, 20) compared one intervention commu- nity with one control community, thus drawing their conclusions from a one-to-one comparison. This is analogous to treating one person and using another as a control (27). The required sample size depends on the health impact that is measured, the desired power of the study, and the study period. For example, a two-year follow-up of about 40 commu- nities of 1000 people would be required to have a a World Health Organization. The use of impregnated bednets and other materials for vector-borne disease control. Unpub- lished document WHONBC/89.981, 1989. 90% chance of detecting a reduction of 30% in child mortality. The required sample size can be reduced by matching communities with respect to baseline characteristics. Communities should be separated by at least one kilometre in order to have separate mos- quito populations. (vi) Comparison with other available measures. There have only been two field trials comparing DDT house-spraying with pyrethroid-impregnated bed nets (16, 22). Both suggest that bed nets are more effective than spraying, but more comparative studies are needed before spraying can be replaced by impregnated bed nets. Conclusions Our object was not to criticize particular trials but to summarize their results and highlight the problems that challenge the assessment of the potential of insecticide-impregnated bed nets for malaria control. These problems are particularly important as they will also have to be confronted by malaria vaccine trials. Impregnated bed nets provide good personal protection against malaria by reducing exposure to mosquitos. Their potential for malaria control depends on factors such as the level of transmission, mosquito characteristics, the immune-status and sleeping habits of the population, and financial constraints. Since these factors are area-specific, caution is necessary when extrapolating the results of one area to another. The evidence suggests that impregnated bed nets are effective in preventing malaria-related mortality and morbidity-but not malaria infection-in areas of low and moderate transmission; the results from areas of high endemi- city are not so encouraging. Further trials are clearly needed and much attention needs to be given to the methods of community-intervention trials which aim to assess the malaria impact. Trials should include several communities, far enough apart to have sepa- rate mosquito populations. Intervention and control communities should be randomly selected and their comparability established prior to the intervention. The acceptability and actual use of the nets should be monitored systematically. The concepts of malaria infection, clinical malaria, and malaria- related death need to be clarified and their relation with sporozoite inoculum investigated. Acknowledgements We thank Pedro Alonso, Maurice King and Andrew Long for their comments on earlier drafts of this paper. WHO Bulletin OMS. Vol 70 1992 295 A. Bermelo & H. Veeken References 1. WHO Technical Report Series No. 735, 1986 (WHO Expert Committee on Malaria: eighteenth report). Geneva, p. 9. 2. Anonymous. Drawing the curtain on malaria. Lancet, 1:1515 (1991). 3. Snow, R.W. et al. Observations on Anopheles gam- biae Giles s.l. (Diptera: Culicidae) during a trial of permethrin-treated bed nets in the Gambia. Bulletin of entomological research, 77: 279-286 (1987). 4. Majorl, G. et al. Efficacy of permethrin-impregnated curtains for malaria vector control. Medical and veterinary entomology, 1: 185-192 (1987). 5. Lines, J.D. et al. Experimental hut trials of permethrin-impregnated mosquito nets and eave curtains against malaria vectors in Tanzania. Medical and veterinary entomology, 1: 37-51 (1987). 6. Hossain, M.I. & Curtis, C.F. Permethrin-impreg- nated bednets: behavioural and killing effects on mosquitoes. Medical and veterinary entomology, 3: 367-376 (1989). 7. Carnevale, P. et al. La lutte contre le paludisme par des moustiquaires impr6gnees de pyrethrinoides au Burkina Faso. Bulletin de la Societ6 de Pathologie exotique, 81: 832-846 (1988). 8. Snow, R.W. et al. Permethrin-treated bed nets (mosquito nets) prevent malaria in Gambian chil- dren. Transactions of the Royal Society of Tropical Medicine and Hygiene, 82: 838-842 (1988). 9. Li Zuzi et al.Trial of deltamethrin-impregnated bed nets for the control of malaria transmitted by Anopheles sinensis and Anopheles anthropophagus. American journal of tropical medicine and hygiene, 40: 356-359 (1989). 10. Charlwood, J.D. & Graves, P.M. The effect of per- methrin-impregnated bed nets on a population of Anopheles farauti in coastal Papua New Guinea. Medical and veterinary entomology, 1: 319-327 (1987). 11. Magesa, S.M. et al. Trial of pyrethroid-impregnated bed nets in an area of Tanzania holoendemic for malaria. Part 2. Effects on the malaria vector popu- lation. Acta tropica, 49: 97-108 (1991). 12. Rozendaal, J.A. & Curtis, C.F. Recent research on impregnated mosquito nets. Journal of the American Mosquito Control Association, 5: 500-507 (1989). 13. Snow, R.W. et al. A trial of permethrin-treated bed nets in the prevention of malaria in Gambian chil- dren. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81: 563-567 (1987). 14. Sexton, J.D. et al. Permethrin-impregnated curtains and bed-nets prevent malaria in Western Kenya. American journal of tropical medicine and hygiene, 43, 11-18 (1990). 15. Ranque, P. et al. Etude experimentale sur l'utilisa- tion de moustiquaires impregnees de deltamethrine dans la lutte contre le paludisme. Parassitologia, 26: 261-268 (1984). 16. Njunwa, K.J. et al. Trial of pyrethroid-impregnated bed nets in an area of Tanzania holoendemic for malaria. Part 1. Operational methods and accept- ability. Acta tropica, 49: 87-96 (1991). 17. Alonso, P.L. et al. The effect of insecticide-treated bed nets on mortality of Gambian children. Lancet, 1: 1499-1502 (1991). 18. Hii, J.L.K. et al. The influence of permethrin- impregnated bednets and mass drug adminis- tration on the incidence of Plasmodium falciparum malaria in children in Sabah, Malaysia. Medical and veterinary entomology, 1: 397-407 (1987). 19. Graves, P.M. et al. Reduction in incidence and pre- valence of Plasmodium falciparum in under-5-year- old children by permethrin impregnation of mosquito nets. Bulletin of the World Health Organisation, 65 869-877 (1987). 20. Procacci, P.G. et al. Permethrin-impregnated cur- tains in malaria control. Transactions of the Royal Society of Tropical Medicine and Hygiene, 85: 181-185 (1991). 21. Molineaux, L. The impact of parasitic diseases and their control, with emphasis on malaria and Africa. In: Vallin, J. & Lopez, A., ed. Health policy, social policy and mortality prospects. Liege, Ordina Editors, 1985, pp. 13-44. 22. Curtis, C.F. et al. Impregnated bed nets and cur- tains against malaria mosquitos. In: Curtis, C.F., ed. Appropriate technology in vector control. Boca Raton, Florida, CRC Press, 1990, pp. 5-46. 23. Kirkwood, B.R. & Morrow, R.H. Community-based intervention trials. Journal of biosocial science, S10: 79-86 (1989). 24. Giglioli, G. Changes in the pattern of mortality fol- lowing the eradication of hyperendemic malaria from a highly susceptible community. Bulletin of the World Health Organization, 46: 181-202 (1972). 25. Molineaux, L. & Gramiccia, G. The Garki project: research on the epidemiology and control of malaria in the Sudan savanna of West Africa. Geneva, World Health Organization, 1980. 26. Leake, D.W. & Hii, J.L.K. Giving bed nets "fair" tests in field trials against malaria: a case from Sabah, East Malaysia. Southeast Asian journal of tropical medicine and public health, 20: 379-384 (1989). 27. Blum, D. & Feachem, R.G. Measuring the impact of water supply and sanitation investments on diar- rhoeal diseases: problems of methodology. International journal of epidemiology, 12: 357-365 (1983). 296 WHO Bulletin OMS. Vol 70 1992