Malaria is a major cause of morbidity and mortality in Africa with an estimated 360 million clinical attacks 1 and 1–2 million deaths annually 2. Malaria vector control relies on the use of effective insecticides, most commonly through indoor residual spraying (IRS) or insecticide-treated nets (ITN). The increased number of reports of insecticide resistant Anopheles species in Africa 3 is a threat to the success of insecticide-based malaria control programmes.

DDT was introduced for malaria control in 1943 4 and was widely acclaimed due to its impact on reducing morbidity and mortality in malaria naïve troops in endemic regions in World War 2. Due to its success, DDT was rapidly introduced into public health and malaria control campaigns, and was the main insecticide used in the WHO malaria eradication campaign carried out between 1955 to 1969 5. Insecticide resistance in the vector is one of the major reasons given for the failure of the WHO campaign 5, but there is little evidence to support this claim in Africa 36.

The use of DDT was reduced from the 1970's with the introduction of pyrethroids. Its use in agriculture ceased internationally in the 1980s and there have been various attempts to ban its use completely since then. The Stockholm Convention on Persistent Organic Pollutants 29 seeks to ban persistent pollutants, including outdoor use of DDT in agriculture. However, due to the beneficial effects of DDT for malaria control the treaty contains an amendment which specifically authorises indoor use of DDT for vector control, subject to certain safeguards.

Pyrethroids, although an excellent insecticide class for controlling malaria, are only available in formulations with an accredited residual life of up to four months, requiring 2–3 rounds of IRS per year in endemic regions, compared to the 1–2 rounds of spray with DDT 7. Pyrethroids remain the only class of insecticide available for the use on ITNs. As pyrethroid resistance has been selected, several control programmes, including Angola, South Africa, Mozambique and Zambia have reverted back to using DDT.

Early use of DDT also left the potential legacy of cross-resistance between DDT and pyrethroids through alterations in their common target site, the sodium channel 8, known as kdr resistance 9. Most reports of kdr resistance in Anopheles gambiae come from West Africa where the use of DDT in agriculture probably contributed to the original selection and extensive spread of this resistance mechanism 10. The only confirmed report in African Anopheles of kdr outside West Africa comes from Kenya, where a different mutation occurs changing the same amino acid residue in the sodium channel 11.

Whether kdr has an operational impact on ITNs has been tested in experimental field trials with conflicting results. An experimental hut trial in Côte d'Ivoire demonstrated a survival advantage for kdr resistant mosquitoes 12 and village randomized trials showed that ITNs continued to prevent malaria despite kdr resistance in the vector population. A recent study in Benin suggests that kdr is capable of undermining ITNs 13. There is a real need to scale these studies up into malaria control programmes. The impact of kdr on IRS was significant in the malaria control programme on Bioko Island, Equatorial Guinea, as monitored through relative vector density resulting in a change from pyrethroid to carbamate for IRS 14. Monitoring malaria cases in Kwa-Zulu Natal, South Africa, picked up the failure of pyrethroids in the IRS programme in the 1990s resluslting in DDT being reintroduced 15. There are no reports of insecticide resistance not affecting an IRS programme.

Understanding the intricacies of resistance mechanisms, cross-resistance and impact on vectors is a pre-requisite for those involved in the selection of insecticides for and maintenance of large scale vector control programmes.

The Lubombo Spatial Development Initiative (LSDI) includes the southern most provinces of Mozambique. The LSDI, with sustained vector control and the use of effective drug treatment for malaria, has reduced Plasmodium falciparum prevalence rates from 88-65% to 33-4% at the 16 sentinel sites in Maputo province where entomological monitoring was completed 16. The impact of the LSDI IRS programme reduced the numbers of vector species over time making mosquito collections increasingly difficult. It has been established that the impact of a successful IRS programme may eradicate Anopheles funestus from an area 1517. In central Mozambique, outside the LSDI area, control has primarily been through the distribution of pyrethroid impregnated bednets, although IRS with DDT is now being introduced in central provinces as part of the National Malaria Control Programme (NMCP).

The objective of this work is to determine if monitoring insecticide resistance is feasible for a malaria control programme and what the impact of monitoring will have on policy.

The location of sixteen collection sites in Mozambique from which An. funestus were collected between 2002–2006 are shown in Figure 1. A total of 4,162 one to three-day old F1 adult progeny were reared from the wild-caught adult An. funestus females collected, from these localities for subsequent bioassays and biochemical assays.

Six of the sixteen sites (Benfica, Boane, Catuane, Chokwe, Mafambisse and Moamba) were included in the original baseline study in 1999 23. In 2006, as mosquito numbers were limited from each family, priority was given to testing lambda-cyhalothrin and bendiocarb, as these were the insecticides in use in the malaria control programme. All sites were tested with lambda-cyhalothrin, eleven with bendiocarb, three with DDT and four with deltamethrin (Table 1). Low level resistance to bendiocarb was detected at Benfica, Boane, Chokwe, Mafambisse, Mahotas and Motaze. These sites also had resistance to lambda-cyhalothrin and deltamethrin. No resistance was detected to DDT.

Significant increase in pyrethroid resistance was detected in Benfica, Boane, Catuane, Chokwe and Moamba (p < 0.05). Other sites, e.g. Mahotas, also showed increases in pyrethroid resistance, although the significance of the rise is unknown as the sample sizes (n<30) were low. A significant decrease (P < 0.001) in pyrethroid resistance was recorded at Catuane, where baseline mortality was 72.7% which increased to 100% susceptibility in 2006.

Mosquitoes from all sites tested had significantly higher p450 levels compared to the Durban susceptible strain (Table 2). Increased p450 activity has already been suggested as the pyrethroid resistance mechanism in An. funestus from Mozambique 2324.

The original baseline survey showed low levels of carbamate resistance that were associated with low levels of an altered AChE 23. A high level of altered AChE resistance frequency was observed at all sites tested (Table 2). This mechanism is the probable cause of the low levels of carbamate resistance observed in bioassays.

No increased levels of esterase or GST activity were detected in An. funestus from any locality tested compared to the Durban susceptible strain.

The development of insecticide resistance is a potential threat to any insecticide-based malaria vector control programme. The number of insecticides and formulations recommended by the WHO Pesticide Evaluation Scheme (WHOPES) for IRS is severely limited 25. This arsenal may be further depleted by the lack of local country or regional insecticide registrations. To ensure that the insecticides used for IRS in Mozambique remain effective and their choice is evidence-based, an assessment of the resistance profile and potential resistance mechanisms within the targeted vector populations needs to be routinely monitored. Since the original baseline established in 1999 23, the resistance profile has been monitored in sixteen localities using WHO bioassays and in ten of these localities using biochemical assays to assess potential resistance mechanisms.

Previously the NMCP in Mozambique used DDT before a change in policy in 1993 when the pyrethroid lambda-cyhalothrin was introduced. In 1999 when the baseline survey was undertaken, both An. funestus and An. arabiensis were resistant to lambda- cyhalothrin 2326. The same resistance profile was detected in An. funestus in Kwa-Zulu Natal, South Africa, which borders southern Mozambique. The onset of measurable insecticide resistance selection correlated with a surge in malaria in that region 27. Pyrethroid resistance in An. funestus, in this region was correlated with increased titres of p450 2324. The detection of resistance prompted a change of insecticide in the LSDI programme, the carbamate bendiocarb replacing lambda-cyhalothrin during the 2000 spray season 16. Bendiocarb was then sprayed bi-annually until 2005, while resistance to the three insecticide classes registered in Mozambique (carbamates, pyrethroids and latterly DDT) were monitored in an attempt to establish a resistance management plan to ensure sustainability of the programme.

Low levels of bendiocarb resistance were detected in An. funestus in the original 1999 baseline 23. Resistance was still detectable by bioassay and associated with high frequencies of an altered AChE resistance mechanism in the 2002–06 collections, which is the likely cause of this resistance. This, coupled with the appearance of carbamate resistance in Mozambican An. arabiensis, the second malaria vector in the region (Coleman et al in press), and the high economic costs associated with bendiocarb use, prompted an operational change of insecticide in 2006 back to DDT. The levels of pyrethroid resistance still segregating in An. funestus at this point were considered too high to justify a switch back to pyrethroid treatment.

The decline in pyrethroid resistance at some sites suggests that with the correct resistance management strategy in place, pyrethroids may again play a role in southern Mozambique's malaria control programmes 28.

This work demonstrates that operationally incorporating monitoring of insecticide resistance in space and time is feasible. The ability to use this data to develop evidence-based insecticide resistance management strategies that tie in with good monitoring of the impact of vector control operations on disease is also essential for sustainable large scale insecticide-based malaria vector control. The monitoring effort reported here resulted in National Malaria Control Campaign policy changes from pyrethroid to carbamate 2326 to DDT over time. This informed decision making will ultimately have cost savings for the malaria control programmes in Mozambique. Continual monitoring will also allow for the establishment of insecticide resistance management programmes, protecting the limited number of insecticides available for malaria control.

SC carried out all the field work and laboratory work. MC completed the analysis and draft manuscript. JH corrected the manuscript and guided analysis. BS conceived the initial ideas with JH.