Distribution and spread of pyrethroid and DDT resistance among the Anopheles gambiae complex in Tanzania.

The development of insecticide resistance is a threat to the control of malaria in Africa. We report the findings of a national survey carried out in Tanzania in 2011 to monitor the susceptibility of malaria vectors to pyrethroid, organophosphate, carbamate and DDT insecticides, and compare these findings with those identified in 2004 and 2010. Standard World Health Organization (WHO) methods were used to detect knock-down and mortality rates in wild female Anopheles gambiae s.l. (Diptera: Culicidae) collected from 14 sentinel districts. Diagnostic doses of the pyrethroids deltamethrin, lambdacyhalothrin and permethrin, the carbamate propoxur, the organophosphate fenitrothion and the organochlorine DDT were used. Anopheles gambiae s.l. was resistant to permethrin in Muleba, where a mortality rate of 11% [95% confidence interval (CI) 6-19%] was recorded, Muheza (mortality rate of 75%, 95% CI 66-83%), Moshi and Arumeru (mortality rates of 74% in both). Similarly, resistance was reported to lambdacyhalothrin in Muleba, Muheza, Moshi and Arumeru (mortality rates of 31-82%), and to deltamethrin in Muleba, Moshi and Muheza (mortality rates of 28-75%). Resistance to DDT was reported in Muleba. No resistance to the carbamate propoxur or the organophosphate fenitrothion was observed. Anopheles gambiae s.l. is becoming resistant to pyrethoids and DDT in several parts of Tanzania. This has coincided with the scaling up of vector control measures. Resistance may impair the effectiveness of these interventions and therefore demands close monitoring and the adoption of a resistance management strategy.

The performance of health laboratories and the quality of malaria diagnosis in six districts of Tanzania.

Early laboratory diagnosis is critical for the optimal management of human malaria, particularly following the introduction of relatively expensive, artemisinin-combination therapies (ACT). The performance of the laboratories and the quality of malaria diagnosis have recently been assessed in 36 healthcare facilities in six districts of Tanzania. Questionnaires, checklists and observations were used to collect information on the availability and functional status of equipment as well as on laboratory personnel and their performance in malaria diagnosis. Together, the surveyed facilities had 112 laboratory staff [almost half (41.1%) of whom were laboratory assistants] and 57 microscopes. Twenty-seven (75.0%) of the healthcare facilities included in the survey had only one functional microscope each. Only seven (12.3%) of the assessed microscopes had been serviced in the previous 2 years. Of the 38 microscopists who were assessed, 24 (63.2%) were re-using microscope slides, 29 (73.5%) were producing bloodsmears of low quality, and 30 (79.0%) were using Field's stain. Although the facility microscopists gave similar results to experienced research microscopists when reading bloodsmears prepared by the survey team, using high-quality reagents (kappa=0.769), they appeared far less competent when reading smears stained using the reagents from the study laboratories (kappa=0.265-0.489). The quality of malaria diagnosis at healthcare facilities in Tanzania, which is generally poor (largely because of inadequate supplies of consumables and the limited skills of laboratory staff in the preparation of bloodsmears), urgently needs to be improved if the utilization of ACT is to be sustainable.

Behavioural and insecticidal effects of organophosphate-, carbamate- and pyrethroid-treated mosquito nets against African malaria vectors.

Three insecticides - the pyrethroid deltamethrin, the carbamate carbosulfan and the organophosphate chlorpyrifos-methyl - were tested on mosquito nets in experimental huts to determine their potential for introduction as malaria control measures. Their behavioural effects and efficacy were examined in Anopheles gambiae Giles s.s. (Diptera: Culicidae) and Anopheles funestus Giles s.s. in Muheza, Tanzania, and in Anopheles arabiensis Patton and Culex quinquefasciatus Say in Moshi, Tanzania. A standardized dosage of 25 mg/m(2) plus high dosages of carbosulfan (50 mg/m(2), 100 mg/m(2) and 200 mg/m(2)) and chlorpyrifos-methyl (100 mg/m(2)) were used to compare the three types of insecticide. At 25 mg/m(2), the rank order of the insecticides for insecticide-induced mortality in wild An. gambiae and An. funestus was, respectively, carbosulfan (88%, 86%) > deltamethrin (79%, 78%) > chlorpyrifos-methyl (35%, 53%). The rank order of the insecticides for blood-feeding inhibition (reduction in the number of blood-fed mosquitoes compared with control) in wild An. gambiae and An. funestus was deltamethrin > chlorpyrifos-methyl > carbosulfan. Carbosulfan was particularly toxic to endophilic anophelines at 200 mg/m(2), killing 100% of An. gambiae and 98% of An. funestus that entered the huts. It was less effective against the more exophilic An. arabiensis (67% mortality) and carbamate-resistant Cx quinquefasciatus (36% mortality). Carbosulfan deterred anophelines from entering huts, but did not deter carbamate-resistant Cx quinquefasciatus. Deltamethrin reduced the proportion of insects engaged in blood-feeding, probably as a consequence of contact irritancy, whereas carbosulfan seemed to provide personal protection through deterred entry or perhaps a spatial repellent action. Any deployment of carbosulfan as an individual treatment on nets should be carried out on a large scale to reduce the risk of diverting mosquitoes to unprotected individuals. Chlorpyrifos-methyl was inferior to deltamethrin in terms of mortality and blood-feeding inhibition and would be better deployed on a net in combination with a pyrethroid to control insecticide-resistant mosquitoes.

Health laboratories in the Tanga region of Tanzania: the quality of diagnostic services for malaria and other communicable diseases.

Although critical for good case management and the monitoring of health interventions, the health-laboratory services in sub-Saharan Africa are grossly compromised by poor infrastructures and a lack of trained personnel, essential reagents and other supplies. The availability and quality of diagnostic services in 37 health laboratories in three districts of the Tanga region of Tanzania have recently been assessed. The results of the survey, which involved interviews with health workers, observations and a documentary review, revealed that malaria accounted for >50% of admissions and out-patient visits. Most (92%) of the laboratories were carrying out malaria diagnosis and 89% were measuring haemoglobin concentrations but only one (3%) was conducting culture and sensitivity tests, and those only on urine and pus samples. Only 14 (17%) of the 84 people found working in the visited laboratories were laboratory technologists with a diploma certificate or higher qualification. Sixteen (43%) of the study laboratories each had five or fewer types of equipment and only seven (19%) had more than 11 types each. Although 11 (30%) of the laboratories reported that they conducted internal quality control, none had standard operating procedures (SOP) on display or evidence of such quality assurance. Although malaria was the main health problem, diagnostic services for malaria and other diseases were inadequate and of poor quality because of the limited human resources, poor equipment and shortage of supplies. If the health services in Tanga are not to be overwhelmed by the progressively increasing burden of HIV/AIDS, malaria, tuberculosis and other emerging and re-emerging diseases, more funding and appropriate policies to improve the availability and quality of the area's diagnostic services will clearly be required.

Experimental hut evaluation of the pyrrole insecticide chlorfenapyr on bed nets for the control of Anopheles arabiensis and Culex quinquefasciatus.

OBJECTIVE: To determine the efficacy of chlorfenapyr against Anopheles arabiensis and Culex quinquefasciatus in East Africa and to identify effective dosages for net treatment in comparison with the commonly used pyrethroid deltamethrin. METHODS: Chlorfenapyr was evaluated on bed nets in experimental huts against A. arabiensis and C. quinquefasciatus in Northern Tanzania, at application rates of 100-500 mg/m(2). RESULTS: In experimental huts, mortality rates in A. arabiensis were high (46.0-63.9%) for all dosages of chlorfenapyr and were similar to that of deltamethrin-treated nets. Mortality rates in C. quinquefasciatus were higher for chlorfenapyr than for deltamethrin. Despite a reputation for being slow acting, >90% of insecticide-induced mortality in laboratory tunnel tests and experimental huts occurred within 24 h, and the speed of killing was no slower than for deltamethrin-treated nets. CONCLUSIONS: Chlorfenapyr induced low irritability and knockdown, which explains the relatively small reduction in blood-feeding rate. Combining chlorfenapyr with a more excito-repellent pyrethroid on bed nets for improved personal protection, control of pyrethroid-resistant mosquitoes and pyrethroid resistance management would be advantageous.

Co-infection with subclinical HIV and Wuchereria bancrofti, and the role of malaria and hookworms, in adult Tanzanians: infection intensities, CD4/CD8 counts and cytokine responses.

The association between HIV and Wuchereria bancrofti, and the role of malaria and hookworms, were analysed by comparing three groups of individuals with: (1) HIV (HIV+; n=16); (2) W. bancrofti (circulating filarial antigen (CFA)+; n=25); and (3) HIV and W. bancrofti (HIV+/CFA+; n=18). A slightly higher HIV load and lower CD4% was observed in the HIV+/CFA+ group compared with the HIV+ group, and a slightly higher W. bancrofti CFA intensity was observed in the CFA+ group compared with the HIV+/CFA+ group, but none of these differences were statistically significant. Specific and non-specific IL-4, IL-10, IFNgamma and TNF levels were measured. Only specific IL-4 was significantly higher in the CFA+ group compared with the HIV+/CFA+ group. Thus, there was no clear evidence for an interaction between HIV and W. bancrofti infection. A multiple linear regression model showed that the presence of CFA was strongly positively associated with specific TNF response and, similarly, that HIV-positive individuals had higher TNF responses than HIV-negative individuals. Interestingly, the CD4% and CD4/CD8 ratio were higher in HIV-positive individuals with hookworms than in those without hookworm co-infection. Malaria was not associated with any of the other infections, or with CD4/CD8 counts or cytokine responses.

Comparison of coverage with insecticide-treated nets in a Tanzanian town and villages where nets and insecticide are either marketed or provided free of charge.

BACKGROUND: There is much emphasis on social marketing as a means of scaling up coverage with insecticide-treated nets and the question has arisen whether nets provided free-of-charge will be looked after by householders. METHODS: Over several years questionnaires and surveys of usage and condition of nets were carried out throughout a town and 15 villages in north-east Tanzania, where nets and insecticide have to be purchased and in 24 other villages where over 15000 nets had been donated and annual re-treatment is provided free-of-charge. RESULTS: There was very high population coverage in the town but, in the villages where nets have to be purchased, only 9.3% of people used nets which were intact and/or had been insecticide-treated and could, therefore, provide protection. However, where nets had been provided free, over 90% of the nets were still present and were brought for re-treatment several years later. CONCLUSION: In this part of Tanzania, social marketing has performed well in a town but very poorly in villages. However, the study showed that people look after and bring for re-treatment nets which had been provided free-of-charge.

Cross-sectional relationship between HIV, lymphatic filariasis and other parasitic infections in adults in coastal northeastern Tanzania.

The relationship between HIV, lymphatic filariasis, malaria (Plasmodium falciparum) and intestinal helminths (Ascaris lumbricoides, Trichuris trichiura and hookworm) was assessed in a cross-sectional study conducted in 2002 among 907 adults in Tanga Region, Tanzania. Overall prevalences were 7.9% for HIV, 43.5% for Wuchereria bancrofti-specific circulating filarial antigen (CFA), 12.3% for P. falciparum, 1.2% for A. lumbricoides, 7.1% for T. trichiura and 75.7% for hookworm. Anaemia was assessed separately for males and females and was found to be more prevalent among females (58.8%) than males (34.8%). When sex and age were controlled for, there was a statistically significant positive association between HIV and W. bancrofti (CFA) infection and between malaria and HIV, but not between malaria and W. bancrofti (CFA) infection. Hookworm infection was positively associated with W. bancrofti (CFA) infection but, surprisingly, negatively associated with HIV. Infection with HIV and hookworms, but not malaria, was associated with a significant reduction in haemoglobin concentration. These associations are likely to reflect underlying mechanisms that need to be clarified to better understand the role of co-infections in HIV pathogenesis, and vice versa.

Insecticide-treated bed-nets for malaria mosquito control.

Pyrethroid-treated bed-nets act against late-night biting mosquitoes, like traps baited by the body odor of the occupant. The personal protective effect of treated nets is considerable, even if they are torn. However, some biting of the occupants does occur, as shown by matching microsatellite alleles in mosquito blood meals to those of net occupants. When whole communities were provided with treated nets, ovarian age grading showed that mosquito survival was reduced, and so was the number of sporozoite-positive mosquitoes in malarious communities. Thus, a high percentage of coverage of all members of malaria-endemic communities is considered to be the most effective way of providing protection for highly malaria-vulnerable children and pregnant women. Teams distributing nets or retreating them free of charge show high productivity, and we consider this the most cost-effective way to proceed. There is evidence for reduced anti-malaria antibody levels in children in communities where treated nets have long been used. However, overall benefits in reduced anemia and mortality are sustained. A high frequency of the kdr resistance gene has not prevented pyrethroid-treated nets from functioning, but it is important to develop alternative fabric treatments in case stronger forms of resistance emerge.

Tests of Olyset nets by bioassay and in experimental huts.

BACKGROUND: Olyset nets are a type of long-lasting insecticidal net made of polyethylene fibre with permethrin incorporated into it and with a 4 mm mesh size. BIOASSAYS: Olyset netting was wrapped around a wire frame and the mosquitoes were released inside for bioassays. There was significantly faster knockdown and higher percent mortality than bioassays with the netting attached to a WHO cone with a piece of cardboard on the other side of the net to prevent escapes through the large mesh. It is suggested that with the latter method some mosquitoes place their tarsae through the mesh on to the cardboard, thus avoiding insecticide contact. TRIALS IN EXPERIMENTAL HUTS: Four mm mesh nets were compared with conventional 1.5 mm mesh nets treated with permethrin. In further trials in huts Olyset nets which were either unwashed or five times washed, with or without subsequent heating, and a Olyset net which had been in domestic use for four years or a new Olyset net were compared with a net treated with bifenthrin. RESULTS & CONCLUSION: In all cases Anopheles biting on sleepers under the nets was reduced and Anopheles mortality was increased by the use of the insecticidal nets. No significant impact of washing or heating was detected and an Olyset net was as good as new after four years use, but did not cause as much mosquito mortality as bifenthrin treated nets.

Evaluation of the bio-efficacy of three brands of repellents against wild populations of anthropophilic mosquitoes.

Three commercial repellents marketed in Tanzania: Zero Bite (a blend of microcrystalline waxes, mineral oils, natural flavours, Olibanum oil, Eucalyptus oil, Geranium oil, Citronella oil and Isopropyl myristrate); X-pel (a petroleum jelly formulation containing diethyl toluamide (DEET) and dimethyl phthalate); No Bite (a spray formulation with diethyl toluamide, 2 methyl 2,4 pentondiol and pthalic ester acids) were tested and compared for their repellency effect against wild anthropophilic mosquito populations. Human forearms, feet and legs were treated with the repellent products. All repellents provided protection against wild populations of biting mosquitoes (mainly Culex quinquefasciatus and Aedes scatophagoides) with varying levels of efficacy. No Bite provided the best overall protection (98%) followed by X-pel (87%). Zero Bite gave the least protection (48%) against the two mosquito species. All products except No Bite displayed reduced efficacy after four hours of application. The results indicate that the two best products give satisfactory levels of personal protection against biting mosquitoes at least for the first five hours, following application, thus could provide complementary protection against mosquito bites particularly during the period when most people have not retired to bed where they may be protected by treated bednets.

Indoors man-biting mosquitoes and their implication on malaria transmission in Mpwapwa and Iringa Districts, Tanzania.

Entomological surveys were carried out in six villages at different altitudes in Mpwapwa and Iringa Districts in central Tanzania in March 2002. A total of 1291 mosquitoes were collected. Of these, 887 mosquitoes were collected by light traps and 404 by indoor pyrethrum spray catch technique. Seventy-nine percent (1026) were Anopheles gambiae s.l., 0.2% (N = 3) were An. funestus, and 20.3% (N = 262) were Culex quinquefasciatus. Other species including Cx cinereus, An. coustani and Aedes spp accounted for 0.5% of the mosquito population. In Iringa, more mosquitoes were collected by pyrethrum spray catch than light trapping technique. The light trap catch: spray catch ratio in Iringa and Mpwapwa was 1:1.15 and 2.5:1, respectively. Indoor pyrethrum spray catch gave an overall estimate of An. gambiae density of 8 and 0.6 mosquitoes per room in Iringa and Mpwapwa, respectively, whereas light trap collections gave an overall respective density of An. gambiae of 63.9 and 2.9 mosquitoes per room. The densities of house entering mosquitoes were found to range from 0 to 135 in Iringa and from 2.6 to 3.5 per room in Mpwapwa. An.funestus mosquitoes were collected in Iringa only. None of the dissected An. gambiae collected in the two districts was infected with malaria sporozoites. Despite low mosquito densities and absence of infective mosquitoes in our study, the two districts are malaria epidemic prone, thus a continuous surveillance is critical for a prompt response to any impending outbreak. Further longitudinal studies are required to determine the transmission potential of the malaria mosquitoes in the two districts.

Distinguishing Plasmodium falciparum treatment failures from re-infections by using polymerase chain reaction genotyping in a holoendemic area in northeastern Tanzania.

An in vivo drug sensitivity study was conducted in Magoda village in northeastern Tanzania to evaluate the usefulness of polymerase chain reaction (PCR)-based genotyping of Plasmodium falciparum parasites to distinguish between re-infection and treatment failure. The study tested P. falciparum susceptibility to a combination of sulfadoxine/pyrimethamine (Fansidar; F. Hoffmann La Roche, Basel, Switzerland). Blood samples were collected before treatment and on days 7, 14, or 28 post-treatment in 51 asymptomatic children, of which 26 could not clear parasitemia within seven days post-treatment. Among the remaining 25 children who had no detectable parasites on day 7, only five remained parasite negative up to day 28. Primary and recrudescent P. falciparum parasites were analyzed by PCR using family specific primers for merozoite surface protein-1 (MSP-1), MSP-2, and glutamate-rich protein (GLURP). All samples contained multiple P. falciparum infections. For all children with recrudescent P. falciparum, common alleles were detected in both the primary and recrudescent samples. However, in no child were the exact same alleles detected in both samples, indicating that probably at least some of the recrudescing parasites originated from new infections. The study demonstrates the general usefulness of PCR genotyping technique in distinguishing re-infections from true recrudescences following therapeutic drug treatment.

Observations on the periodicity of Plasmodium falciparum gametocytes in natural human infections.

The circadian periodicity of Plasmodium falciparum gametocytes in peripheral blood was analysed in a group of children from an holoendemic community of north-eastern Tanzania. No periodicity was observed with asexual stage parasites. Gametocytes were shown to display a diurnal subperiodic pattern with a periodicity index of 31. 8. Mathematical analysis of the data indicated that P. falciparum gametocytes tend to display periodicity with a peak (15:30-19:30 h) that do not coincide with the peak (00:30-03:30 h) biting activity of the local vector, Anopheles gambiae. We were thus able to show a P. falciparum gametocyte periodicity with a harmonic wave pattern, but its functional biological significance if any, is currently unknown.

Trial of pyrethroid impregnated bednets in an area of Tanzania holoendemic for malaria. Part 1. Operational methods and acceptability.

In five Tanzanian villages, nets impregnated with permethrin or lambdacyhalothrin were given out. The people received them enthusiastically and brought their nets for re-impregnation at six monthly intervals. Bioassays showed that the insecticidal power of permethrin impregnated nets remained adequate for six months unless the nets were washed. Nets with 30 mg lambdacyhalothrin/m2 retained high insecticidal power despite washing, but this dose caused temporary cold-like symptoms in those sleeping under freshly treated dry nets. Methods by which durable bednets might be made affordable by Tanzanian villagers are discussed.

Trial of pyrethroid impregnated bednets in an area of Tanzania holoendemic for malaria. Part 2. Effects on the malaria vector population.

The malaria vector population consisted mainly of Anopheles gambiae s.s. with a small contribution from An. funestus and An. rivulorum. The mosquitoes coming to bite in bedrooms were monitored with light traps set beside untreated bednets. When impregnated bednets were provided for all the other beds in a village the Anopheles populations declined but the Culex quinquefasciatus populations were unaffected. The survival of An. gambiae (as measured by the mean number of ovarian dilatations) and the sporozoite rate declined following introduction of the nets and the estimated sporozoite inoculation rates into people not under their nets declined by more than 90%. The net introductions caused sharp declines in the number of mosquitoes resting indoors, but the evidence was inconclusive regarding diversion to outdoor resting, animal biting, earlier biting or outdoor biting. DDT spraying greatly reduced the Anopheles populations.

Diversity of Plasmodium falciparum clones infecting children living in a holoendemic area in north-eastern Tanzania.

The diversity of Plasmodium falciparum clones and their role in progression from asymptomatic to symptomatic condition in children have been investigated. Attempts to identify whether particular parasite genotypes were associated with the development of clinical symptoms have been made. A cohort of 34 initially asymptomatic parasitaemic children aged 1-5 years were followed daily for 31 days. Clinical examinations were made each day for signs and symptoms of clinical malaria, followed by parasitological investigation. Nineteen children developed symptoms suggestive of clinical malaria during this period. Daily blood parasite samples from 13 children who developed clinical malaria symptoms and 7 who remained asymptomatic were genotyped by PCR-amplification of the polymorphic regions of the merozoite surface proteins 1 and 2 (MSP1 and MSP2) and the glutamate rich protein (GLURP) genes. Infections were found to be highly complex in both groups of children. Every isolate examined from both groups had a mixture of parasite clones. Daily changes were observed in both parasite density and genotypic pattern. The mean number of genotypes per individual was estimated at 4.9 and 2.7 for asymptomatic and symptomatic groups of children, respectively. Analysis of allele frequency distributions showed that these differed significantly for the MSP1 locus only.

Diversity, spatial and temporal abundance of Anopheles gambiae complex in the Rufiji River basin, south-eastern Tanzania.

The Anopheles gambiae complex contains the most efficient malaria vectors in the world. Identification of the species and the concomitant distribution are vital for effective malaria control. The objective of the study was to establish the diversity, spatial and seasonal abundance of malaria vectors in the Rufiji River Basin in Tanzania and the implications on malaria vector control. Mosquitoes were collected using CDC light-traps. Anopheles mosquitoes were identified by conventional keys. Anopheles gambiae complex were speciated using standard PCR method. Of the 562 specimens analysed by PCR 69% produced fragments equivalent to An. gambiae s.s. (390 bp), 23% equivalent to An. arabiensis (315 bp), and 7% as An. merus (464 bp). An. gambiae s.s. and An. merus were more abundant on the plateau than on the flood plain (Fisher's exact test, P < 0001), whereas An. arabiensis was equally abundant between the two sites (Fisher's exact test, P=1656). The density of the three sibling species of An. gambiae complex also varied with the seasons. An. gambiae s.s. and An. arabiensis were most predominant species during the start of the rainy season, but as season progresses, An. gambiae s.s. predominated. An. merus was only recorded during the short rainy season. In conclusion, both An. gambiae s.s., An. arabiensis and An. merus are malaria vectors in the Rufiji River basin and that An. merus is recorded for the first time in the south-eastern coast of Tanzania. These findings are important in the planning and implementation of malaria vector control activities in the Rufiji River basin, south-eastern Tanzania.

Community directed approach beyond ivermectin in Tanzania: a promising mechanism for the delivery of complex health interventions.

The Community Directed Intervention (CDI) is currently used for Ivermectin distribution for the treatment of onchocerciasis in Africa. This study was carried out to determine the extent to which the CDI process can be used for the delivery of other health interventions with different degrees of complexity. The study was conducted in five districts of Kilosa, Muheza, Lushoto, Korogwe and Ulanga in Tanzania and involved communities, health facility and district healthcare providers. Implementation of CDI across these health interventions involved addressing six major processes, namely, stakeholder processes, health system dynamics, engaging communities, empowering communities, engaging CDI implementers and broader system effects. Community and health systems changes were triggered, such that the inherent value of community involvement and empowerment could be internalized by communities and health workers, leading to a more receptive health system. The CDI process was accepted at the community levels as many were willing and ready to adopt the approach. Health workers at community levels were readily available and supportive of the process. Additionally, noted were the verified willingness and ability of community implementers to deliver multiple interventions; confirmed efficiency of CDI leading to cost savings at health systems level; increasing interest of the health system in CDI; interest of health workers in the process of integrated planning. However, there were factors that may have a negative influence on the CDI process. Drug and supply policy for CDI process was lacking at the national and district levels and the presence of parallel community-based programmes that provide financial incentives for community members to run them discouraged Community-directed distributors who in most cases are volunteers. In conclusion, the results have clearly and evidently demonstrated the potential of CDI approach for effectively and efficiently control of other diseases such as malaria, tuberculosis and childhood illnesses. The study has provided unique information on the feasibility and effectiveness of integrated delivery of interventions at the community level.

Patterns of malaria related mortality based on verbal autopsy in Muleba District, north-western Tanzania.

Reliable malaria related mortality data is important for planning appropriate interventions. However, there is scarce information on the pattern of malaria related mortality in epidemic prone districts of Tanzania. This study was carried out to determine malaria related mortality and establish its trend change over time in both epidemic and non-epidemic areas of Muleba District of north-western Tanzania. A verbal autopsy survey was conducted to obtain data on all deaths of individuals who died in six randomly selected villages from 1997 to 2006. Relatives of the deceased were interviewed using a standardized questionnaire. Communicable diseases accounted for about two thirds (61.9%) of deaths among > or =5 years individuals and 84.8% in < or =5 years. Non-communicable diseases accounted for 28.9% and 14.1% deaths in > or =5 years and < or =5 years, respectively. Malaria was the leading cause of deaths in all age groups (40.3%) and among children <5 years (73.8%). Infants accounted for about two third (64.5%) of all malaria related deaths in children <5 years. Peak of malaria proportional mortality was highest during malaria epidemics. Most of the malaria-related deaths in this group were among 1-12 months (64.5%) followed by 13-24 months (20.9%), and 25-59 months (14.8%). Cerebral malaria accounted for 18.9% (N=32) of death related to malaria in all age groups; 12.1% (17/141) were in under-five, 42.9% (6/14) were in 5-14 years and 64.3% (9/14) in 15-70 years old. More than half of malaria related deaths (61.0%) in <5 years children were associated with severe anaemia followed by diarrhoeal disease (24.1%), cerebral malaria (12.5%) and respiratory infection (8.5%) as common conditions. The majority of the deceased caretakers first sought treatment at health facilities within 24hr of the onset of illness. Significantly a higher proportion of caretakers of the underfives in the epidemic area sought treatment within 24hr than in non-epidemic area (39.3% vs. 18.5%; P = 0.0385). In conclusion, malaria accounts for majority of deaths in Muleba district, with substantial proportion being attributed to malaria epidemics.