• View in gallery

    Cohort selection procedure and exclusion criteria. ACD = active case detection.

  • View in gallery

    Monthly incidence of cases of Plasmodium falciparum malaria. Months 1, 2, etc. refer to January, February, etc. The two years of follow-up were combined. py = person-year; Reg. = regular.

  • 1

    Phan VT, Tuy TQ, Hung LX, Ngo LV, 1998. Epidemiological characteristics of malaria in Vietnam 1992–1997. http://www.ec-malaria.org/mcis/book1_21.htm.

  • 2

    Schuftan C, 2000. A story to be shared: the successful fight against malaria in Vietnam. WHO WPRO and the global Roll Back Malaria Program. http://www.afronets.org

  • 3

    Tuy TC, 1998. Malaria situation in Vietnam January–June 1998. http://www.ec-malaria.org/mcis/book1_21.htm.

  • 4

    Verle P, Tuy TC, Lieu TT, Kongs A, Coosemans M, 1998. New challenges for malaria control in northern Vietnam. Res Rev Parasitol 58 :169–174.

    • Search Google Scholar
    • Export Citation
  • 5

    Cong LD, 1999. Results of malaria control in Vietnam 1992–1997 & malaria control plan for 1998–2000. http://www.ec-malaria.org/mcsi/lib.nsf/issue1

  • 6

    Phan VT, 1998. Epidémiologie du Paludisme et Lutte Antipaludique au Vietnam. Hanoi: Editions Medicales Vietnam.

  • 7

    Van Bortel W, Harbach RE, Trung HD, Roelants P, Backeljau T, Coosemans M, 2001. Confirmation of Anopheles varuna in vietnam, previously misidentified and mistargeted as the malaria vector Anopheles minimus. Am J Trop Med Hyg 65 :729–732.

    • Search Google Scholar
    • Export Citation
  • 8

    Manguin S, Kengne P, Sonnier L, Harbach RE, Baimai V, Trung HD, Coosemans M, 2002. SCAR markers and multiplex PCR-based identification of isomorphic species in the Anopheles dirus complex in southeast Asia. Med Vet Entomol 16 :46–54.

    • Search Google Scholar
    • Export Citation
  • 9

    Coosemans M, 2001. Identification and Characterization of Malaria Vectors in Southeast Asia: A Prerequisite for Appropriate Vector Control. Final report (2001). Antwerp: International Cooperation with Developing Countries (INCO-DC).

  • 10

    Hung LQ, De Vries PJ, Giao PT, Nam NV, Binh TQ, Chong MT, Quoc NTTA, Thanh T.N, Hung LN, Kager PA, 2002. Control of malaria: a successful experience from Vietnam. Bull World Health Organ 80 :660–666.

    • Search Google Scholar
    • Export Citation
  • 11

    Demedts P, Wery M, 1985. Indirect immunofluorescence using differentiating antigens in detection of imported malaria. Ann Soc Belg Med Trop 65 (Suppl 2):89–96.

    • Search Google Scholar
    • Export Citation
  • 12

    Zeger SL, Liang KY, 1986. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42 :121–130.

  • 13

    Smith T, Schellenberg JA, Hayes R, 1994. Attributable fraction estimates and case definitions for malaria in endemic areas. Stat Med 13 :2345–2358.

    • Search Google Scholar
    • Export Citation
  • 14

    Forest Malaria in Southeast Asia, 1991. Proceedings of an Informal Consultative Meeting WHO/MRC. Sharma VP, Kondrashin AV, eds. New Delhi.

  • 15

    Kondrashin AV, Jung RK, Akiyama J, 1991. Ecological aspects of forest malaria in Southeast Asia. Sharma VP, Kondrashin AV, eds. Forest Malaria in Southeast Asia. New Delhi: 1–28.

  • 16

    Armrattana A, Kaewkantha S, Kamboonruang C, Webber RH, 1993. Villager activities and malaria illness in a Thai-Myanmar border area. Trans R SocTrop Med Hyg 87 :340–348.

    • Search Google Scholar
    • Export Citation
  • 17

    Butraporn P, Sornmani S, Hungsapruek T, 1986. Social, behavioural, housing factors and their interactive effects associated with malaria occurrence in east Thailand. Southeast Asian J Trop Med Public Health 17 :386–392.

    • Search Google Scholar
    • Export Citation
  • 18

    Lansang MA, Belizario VY, Bustos MD, Saul A, Aguirre A, 1997. Risk factors for infection with malaria in a low endemic community in Bataan, the Philippines. Acta Trop 63 :257–265.

    • Search Google Scholar
    • Export Citation
  • 19

    Luxemburger C, Thwai KL, White NJ, Webster HK, Kyle DE, Maelankirri L, Chongsuphajaisiddhi T, Nosten F, 1996. The epidemiology of malaria in a Karen population on the western border of Thailand. Trans R Soc Trop Med Hyg 90 :105–111.

    • Search Google Scholar
    • Export Citation
  • 20

    Rahman WA, Adanan CR, Abu-Hassan A, 1998. A study on some aspects of the epidemiology of malaria in an endemic district in northern Peninsular Malaysia near Thailand border. Southeast Asian J Trop Med Public Health 29 :537–540.

    • Search Google Scholar
    • Export Citation
  • 21

    Singhasivanon P, Thimasarn K, Yimsamran S, Linthicum K, Nualchawee K, Dawreang D, Kongrod S, Premmanisakul N, Maneeboonyang W, Salazar N, 1999. Malaria in tree crop plantations in south-eastern and western provinces of Thailand. Southeast Asian J Trop Med Public Health 30 :399–404.

    • Search Google Scholar
    • Export Citation
  • 22

    Somboon P, Aramrattana A, Lines J, Webber R, 1998. Entomological and epidemiological investigations of malaria transmission in relation to population movements in forest areas of north-west Thailand. Southeast Asian J Trop Med Public Health 29 :3–9.

    • Search Google Scholar
    • Export Citation
  • 23

    Alonso PL, Lindsay SW, Armstrong JR, Conteh M, Hill AG, David PH, Fegan G, de-Francisco A, Hall AJ, Shenton FC, Cham K, Greenwood BM, 1991. The effect of insecticide-treated bed nets on mortality of Gambian children. Lancet 337 :1499–1502.

    • Search Google Scholar
    • Export Citation
  • 24

    D’Alessandro U, Olaleye BO, McGuire W, Langerock P, Bennett S, Aikins MK, Thomson MC, Cham MK, Cham BA, Greenwood BM, 1995. Mortality and morbidity from malaria in Gambian children after introduction of an impregnated bednet programme. Lancet 345 :479–483.

    • Search Google Scholar
    • Export Citation
  • 25

    Nevill CG, Some ES, Mung’ala VO, Mutemi W, New L, Marsh K, Lengeler C, Snow RW, 1996. Insecticide-treated bednets reduce mortality and severe morbidity from malaria among children on the Kenyan coast. Trop Med Int Health 1 :139–146.

    • Search Google Scholar
    • Export Citation
  • 26

    Clarke SE, Bogh C, Brown RC, Pinder M, Walraven GE, Lindsay SW, 2001. Do untreated bednets protect against malaria? Trans R Soc Trop Med Hyg 95 :457–462.

    • Search Google Scholar
    • Export Citation
  • 27

    Genton B, Hii J, Al-Yaman F, Paru R, Beck H-P, Ginney M, Dagoro H, Lewis D, Alpers MP, 1994. The use of untreated bednets and malaria infection, morbidity and immunity. Ann Trop Med Parasitol 88 :263–270.

    • Search Google Scholar
    • Export Citation

 

 

 

 

 

FOREST MALARIA IN VIETNAM: A CHALLENGE FOR CONTROL

View More View Less
  • 1 Institute of Tropical Medicine Prince Leopold, Antwerp, Belgium; National Institute for Malariology, Parasitology and Entomology, Hanoi, Vietnam; Institute for Malariology, Parasitology and Entomology, Ho Chi Minh City, Vietnam; Provincial Malaria Station, Phan Thiet, Binh Thuan Province, Vietnam

Forest malaria is a complex but common phenomenon occurring in southeast Asia. We studied its epidemiology through a prospective community-based study in central Vietnam. A total of 585 individuals were followed for two years by active case detection and biannual cross-sectional surveys. The prevalence of antibodies to Plasmodium falciparum was constantly about 20% across surveys and the incidence rate of clinical episodes of P. falciparum malaria was 0.11/person-year. Multivariate analysis showed that regular forest activity was the main risk factor for clinical malaria and malaria infections. Untreated bed nets had a significant protective effect (60%), except for people regularly sleeping in the forest. The population-attributable fraction for regular forest activity was estimated to be 53%. Our results confirm the major role played by forest activity on the malaria burden in this area and provide the basis for targeting control activities to forest workers. New interventions based on insecticide-treated materials need to be urgently evaluated.

INTRODUCTION

After 10 years of intensive control efforts, malaria morbidity and mortality in Vietnam have decreased by 60% and 97%, respectively.1–3 Although malaria has disappeared in many areas,4 it is still endemic in the central and southern provinces of the Tay Nguyen highlands. These forested mountainous areas, where agriculture and woodcutting are highly profitable, attract many migrant workers from non-endemic provinces and from border areas of Laos and Cambodia where malaria is endemic.5,6 Such population movements contribute to maintain malaria transmission in central Vietnam and to the re-introduction of parasites in places where they had disappeared. This is a highly variable situation where target groups are not well defined and consequently where malaria control measures are difficult to implement. Moreover, the main vector, Anopheles dirus A, is highly exophagic and exophilic and interventions based on residual indoor spraying or insecticide treated bed nets have had a limited impact.7,8 In this context, it is extremely important to identify and quantify the effects of factors involved in malaria transmission. We carried out a longitudinal study in four sentinel sites aiming at describing the epidemiology and the transmission dynamics of malaria in relation to local entomologic and ecologic parameters.9 We report here the epidemiologic results of a two-year follow-up in one sentinel site located in a forested area of central southern Vietnam. This is one of the few community-based studies on malaria morbidity in Vietnam.10

MATERIALS AND METHODS

Study population.

The study was carried out between September 1999 and August 2001 in Village 3, Suoi Kiet commune, a hilly and forested area situated in Binh Thuan province (in the southern part of the central coast of Vietnam). The dry season is from January to April, with the coldest months in January and February and the rainy season is from May to December. The annual average temperature is 26.4°C and the average humidity is 82%. Malaria transmission is perennial with two peaks: June and October. Anopheles dirus A is the main vector and several potential secondary vectors such as An. aconitus, An. maculatus, and An. sinensis are present.7,8 Village 3 has approximately 1,000 inhabitants, all belonging to the Kinh ethnic group. Most people are farmers growing cashew nut trees, manioc, and maize. The soil is dry and poor, and people have established fields in the forest directly surrounding the village (hut farmers). The most profitable but illegal activity is woodcutting. All the 104 households (total = 541 individuals), situated in the eastern half of the village where the mosquito collections were carried out were selected for the study. This section of the village was situated between the district road and the river and was representative of the different kind of house structures found in the entire village (bamboo/wooden or brick walls; palm leaf/ wooden or tiled roofs; mud or concrete floors). The forest was situated beyond the river approximately 1 km from the nearest houses. Only permanent residents were included in the study and a unique code number was given to each of them. Informed consent was obtained from all study participants, the village leaders, and the People’s Committee. Ethical approval for the study was obtained from the Ministry of Health of Vietnam and the National Institute for Malariology, Parasitology and Entomology in Hanoi.

Data collection.

A census of the study population was done at the start of the study and information on age, sex, occupation, forest activity, bed net availability, and previous vector control measures was collected. Forest activity was categorized into four groups: never, occasionally (less than once a week), regularly working (on average every week), and regularly working and sleeping in the forest (same as previous but also regularly sleeping in the forest). Bed net availability is an estimate of the average number of individuals sleeping under a net in each house (number of individuals sleeping in the house divided by the number of bed nets in use). This variable was arbitrarily categorized into two groups: inadequate = no bed nets or more than three persons sleeping under one net and adequate = 1–3 persons/net. In our study we consider bed nets as non-treated with insecticide since in this commune such activities ended in 1998, one year before the study started.

Active case detection (ACD) started in September 1999 after the census and the first cross-sectional survey, and consisted in weekly home visits. Information on malaria symptoms and treatments between visits, as well as body (axillary) temperature, was collected and reported on pre-coded standardized questionnaires. A blood slide (thick and thin films) was taken whenever the body temperature was ≥37.5°C and/ or there was a history of fever within the past 48 hours. Suspected malaria cases were treated presumptively by the hamlet health workers with a full course of artesunate. Treatment compliance was not checked. Reasons for any absence of more than one week were recorded and an updated census with population movements was kept. If an individual could not be followed-up for a total period of 12 weeks or more during the two-year study, he or she was considered not to be representative of the population at risk and was excluded from the analysis of ACD. Malaria attacks detected during the first week of follow-up were also excluded from the analysis because infection occurred before the follow-up.

All five cross-sectional surveys concerned all individuals living in the 104 households included in follow-up study. The first survey was done in September 1999; four additional surveys were carried out twice a year at the end of each transmission period. Each participant was administered a questionnaire on malaria symptoms and treatments and clinical examination measured body temperature and spleen size. Blood samples were taken for microscopic examination (thick and thin films), hematocrit (micro-capillary tubes), and identification of antibodies to malaria parasites (filter papers). Suspected malaria cases were treated presumptively.

Laboratory methods.

Microscopic examination.

Blood slides were stained with a 3% Giemsa solution for 45 minutes. The number of asexual forms per 200 white blood cells (WBCs) was counted and parasite densities were computed assuming a mean WBC count of 8,000/μL. A slide was defined as negative if no asexual forms were found after counting 1,000 WBCs. Slide reading was blinded and an external quality control done at the Institute of Tropical Medicine (Antwerp, Belgium) (all positive slides and 10% of the negative slides) found a 99% agreement at first reading, then 100% after re-reading by the two laboratories.

Hematocrit.

Blood in micro-capillary tubes was centrifuged and the packed cell volume was determined according to the principles of the Hawksley micro-hematocrit reader method, as recommended by the World Health Organization.

Indirect fluorescent antibody test (IFAT).

Filter papers (No. 3; Whatman, Kent, United Kingdom) were stored at −20°C. Total immunoglobulin titers for P. falciparum were measured by an IFAT that has been previously described.11 Plasmodium falciparum antigen was prepared from in vitro cultures of an isolate from a patient in southern Vietnam. Negative control serum was obtained by pooling the sera of five malaria-free individuals; positive control serum was obtained from five patients that had several malaria episodes. The serum dilutions studied ranged from 1:40 to 1:640. Slide reading was blinded and double reading for all samples found an agreement of more than 98% between two independent readers.

Case definition.

A clinical case of malaria was defined as fever (body temperature ≥37.5°C) and/or a history of fever in the past 48 hours, with a blood slide positive for Plasmodium asexual forms. The terms cases or episodes will be equally used is this paper. Recrudescence was defined as clinical P. falciparum malaria occurring within 28 days following the first episode. All recrudescences were excluded from the final analysis of ACD. A positive IFAT result for either current or recent infection with P. falciparum was defined as an IFAT titer ≥1:80. New infections determined by an IFAT (five surveys) were defined as all individuals whose titers became positive or, if previously positive, increased two-fold or remained ≥1:320 between two consecutive surveys.

Statistical analysis.

Data were analyzed with STATA 7.0 software (Stata Corp., College Station, TX). Incidence rates of P. falciparum malaria attacks were calculated for each group at risk in the cohort study. Person-years at risk were calculated by defining as one week at risk for malaria each individual record in which the subject was present and not currently under anti-malarial treatment. To assess the effect of different risk factors on new clinical episodes, we restricted the analysis to P. falciparum (excluding recrudescence) because it was impossible to distinguish new infections from relapses with P. vivax.

For cross-sectional surveys, P. falciparum seroprevalence, which is a cumulative prevalence of all recent and current infections, was calculated for each group at risk and was the outcome measure for the analysis of the overall risk of P. falciparum infections. Since repeated measurements on individuals were done both in longitudinal and cross-sectional studies, clustering was addressed by applying generalized estimating equations as previously described.12 A binomial family and its default link and exchangeable working correlation structure were used. The time variable (t) taken was either the week number (longitudinal study) or the time of survey (cross-sectional studies). Individuals were taken as the cluster units. Univariate and multivariate-adjusted odds ratios (ORs) (adjusting for age, sex, forest activity, and bed net availability) were calculated for the risk of new clinical episodes (ACD) and for the overall risk of P. falciparum infections. Within household clustering effect was investigated as well as all possible interactions of up to order two. The population-attributable fraction (PAF) for forest activity represents the proportion of P. falciparum infections in the population that would be avoided if the exposure (forest activity) were removed.13

RESULTS

A total of 573 individuals were included in the main analysis of the cohort study (ACD), (Figure 1). Two hundred eleven malaria clinical episodes were detected in 109 (19%) subjects, while 464 (81%) remained symptom-free during the two-year follow-up. The study population was very young, 50% of the individuals were less than 20 years old, and ethnically homogeneous (all but two were Kinh) (Table 1). Half of the active population (175 of 338) regularly worked in the forest, either farming or woodcutting, although many people had both activities. The population was stable because more than 80% were permanent residents. Emigrations were evenly distributed during the follow-up period. A total of 956.7 person-years at risk were investigated with a median follow-up of 98 weeks. Ten percent of the individuals were lost to follow-up due to emigration and 15 (25%) were regular forest workers.

Risk factors for P. falciparum malaria attacks determined by ACD.

Among the 211 malaria clinical episodes, 123 (58.3%) were due to P. falciparum, 85 (40.3%) to P. vivax, and 3 (1.4%) to P. malariae (single and/or mixed infections). The overall incidence rate of malaria cases was 0.22/person-year; the incidence rate of new clinical cases of P. falciparum malaria (excluding 21 recrudescences) was 0.11/person-year (0.13/person-year when including the 21 recrudescences). The incidence rate of clinical cases of P. falciparum malaria among individuals lost to follow-up was similar to that in the cohort (0.13/person-years). The results of the univariate analysis show that sex, age, profession, regular forest activity, bed net availability, and season were significant risk factors for new episodes of P. falciparum malaria (Table 2). Working regularly in the forest (especially woodcutting) was a significant risk factor (OR = 5.20, P < 0.001), although it was not different from that of people sleeping regularly in the forest (OR = 4.63, P < 0.001). Males and young adults (less than 40 years old) were significantly more at risk than females and children less than nine years old. Untreated bed nets had a significant protective effect (OR = 0.40, P < 0.001). The incidence rate among people lost to follow-up was 0.13/person-years.

Profession and forest activity were strongly correlated; thus, both of them could not be kept in the final model. Forest activity was chosen as the primary exposure in the multivariate analysis. Multivariate-adjusted ORs for effects of risk factors on clinical episodes of P. falciparum malaria are shown in Table 3. There was a significant interaction between bed net availability and regular work and sleeping in the forest. If people had bed nets at home, sleeping in the forest increased significantly the risk of clinical malaria attacks (OR = 8.09, P < 0.001), whereas such a risk did not change for those without nets (OR = 1.72, P = 0.25). People with regular forest activity, but who did not usually sleep in the forest, were also at higher risk for clinical malaria (OR = 3.85, P < 0.001) but this was independent of bed net availability at home (no interaction). Conversely, untreated nets had a significant protective effect, except for people sleeping in the forest. Such risk factors with corresponding ORs and significance levels remained similar after including the 21 clinical attacks classified as recrudescences. The monthly incidence rate of clinical cases of P. falciparum malaria are shown in Figure 2; the two years were combined because there was no significant difference between them (P = 0.27). The overall incidence rate shows that malaria transmission is perennial with two small peaks in June and October. After adjusting for regular forest activity, the monthly incidence rate was much higher in individuals with regular forest activity and the seasonal effect much more pronounced than in people without regular forest activity.

Risk factors for overall P. falciparum infections determined by IFAT (five surveys).

Parasite and spleen rates, particularly in children 2–9 years old, were constantly less than 10% and decreased with time. No severe cases of anemia were detected (Table 4). The prevalence of clinical malaria was very low, ranging from 0.5% to 1.8% across surveys, and most (42 of 60) of the infections detected by microscopy were asymptomatic. The prevalence of antibodies to P. falciparum was approximately 20%, ranging from 18% to 24% across surveys. The incidence rate of new P. falciparum infections determined by the IFAT, was approximately two times higher (0.22/person-year) than that estimated by ACD. Most of the clinical episodes missed by ACD occurred in people with regular forest activity. The risk factors for malaria infection identified by serology were the same (except for season) than those identified by ACD (Table 5). Multivariate-adjusted ORs are shown in Table 6. The two former categories of regular forest activity (sleeping and not sleeping in the forest) were merged into one (ORs and 95% confidence intervals were identical) and the effect of regular forest activity was compared with the baseline of never going to the forest. A significant interaction was present between regular forest activity and sex because the former was significantly associated with the risk of infection in men (OR = 10.76, P < 0.001), but not in women (OR = 1.56, P = 0.25). Conversely, sex was an important risk factor because women were significantly less at risk of malaria infections than men (OR = 0.18, P < 0.001), but only in those with regular forest activity. The summary OR for the effect of occasional forest activity was 2.85 (P = 0.006) regardless of sex (no interaction).

Age, bed net availability, and survey time remained significant risk factors for malaria infections after adjusting for the combined effect of forest activity and sex. The odds of infection increased significantly with age and there was a significant protective effect of untreated nets (OR = 0.38, P < 0.001). In the risk factor analysis for all P. falciparum infections, the estimated PAF was 53% for regular forest activity and 60% when considering any forest activities (including occasional).

DISCUSSION

Malaria and forest activity is a well-known association in southeast Asia,14,15 and many studies carried out in this part of the world have reported forest activity as a strong risk factor for malaria.16–22 We report here one of the few prospective community-based studies on malaria morbidity in Vietnam. The combination of ACD and regular cross-sectional surveys gives a dynamic picture of malaria transmission and an estimation of the importance of several risk factors, such as forest activity, for both clinical malaria and malaria infection. According to the malariometric indices estimated by the five surveys, this area can be classified as hypoendemic because parasite and spleen rates in children 2–9 years old were consistently less than 10%. Malaria transmission within the village is low according to the malaria incidence rates in people without regular forest activity. This observation is consistent with the estimated entomologic inoculation rate of approximately one infective bite/person-year due to the main malaria vector (An. dirus A).7

We arbitrarily defined clinical attacks occurring within 28 days after treatment as recrudescences because we wanted to specifically estimate risk factors for new malaria infections. A total of 21 recrudescences were identified, indicating an artesunate failure rate of 17%. If one considers the duration of the treatment and the fact that it was not supervised, this failure rate might be due to low compliance.

Regular forest activity (woodcutting and agriculture), a common activity in Village 3, was a strong and significant risk factor for malaria and increased by 4-fold and 10-fold the odds of disease and infection, respectively. Our estimation is unlikely to be biased because the information on forest activity was collected independently of malaria symptoms and diagnosis (microscopic examination and IFAT reading) was blind to the subject’s identity and exposure. Moreover, only 10% of the people under surveillance were lost to follow-up, and the malaria incidence rate in this group was similar to the one estimated in the whole cohort.

In the cohort study, regular forest activity (without sleeping in the forest) was a significant risk factor for clinical malaria and this effect was independent of bed net availability at home. Nevertheless, regularly sleeping in the forest increased eight times the risk of clinical malaria, but only if people had bed nets at home. Such a finding can be explained only by admitting the protective efficacy of untreated nets and consequently a non-negligible (even if low) level of malaria transmission within the village. People not having bed nets are exposed to the risk of clinical malaria in the village and sleeping in the forest does not increase significantly such a risk. Conversely, people with bed nets are protected when sleeping at home but are exposed to infection when sleeping in the forest where they usually do not use nets. Thus, bed nets are not adequate for preventing malaria in forest workers.

The ACD was based on weekly home visits and might have missed some clinical cases. Antibody titers to P. falciparum measured from samples collected during the five surveys can estimate the cumulative frequency of all infections. The comparison between these two methods can estimate the amount of malaria cases missed by ACD. The incidence rate estimated by the evolution of antibody titers is approximately double that estimated by ACD. Since most of the missed infections occurred in people with regular forest activity, they might have some level of immunity resulting in asymptomatic or mild clinical episodes that would be self-treated and not reported to the community health workers. Self-treatment is a current practice in Vietnam, and anti-malarial drugs can be bought in any shop. If one considers the efforts and the resources involved in the set up of a performing ACD system, serologic surveys might be more cost-effective for the surveillance of malaria morbidity in Vietnam or in any other place with low transmission. Other studies carried out in areas of low endemicity, such as The Philippines, also suggested the usefulness of serology for the surveillance of malaria morbidity and the identification of focal areas at risk.18

The multivariate analysis carried out with the serologic results of the five surveys specifically examined risk factors for malaria infections, and confirmed the strong and significant effect of both forest activity and bed net protection. Moreover, there was a significant interaction between regular forest activity and sex. Regular forest activity was a significant risk factor for men, but not for women. This could be explained by different exposure to mosquito bites because women, even if they performed regular forest activity, do not go into the forest as often as men and do not stay as long because they also perform additional domestic tasks. Women also remain well covered while men usually work stripped to the waist and in shorts, making them more exposed to mosquito bites. Among people who never worked in the forest, no sex difference was found.

In the analysis of risk factors for P. falciparum infections, the PAF for regular forest activity estimated that half of the malaria infections in the village could be avoided by removing this factor. However, even if this estimate needs to be interpreted with caution, given the underlying assumptions in the PAF calculation,13 it shows the contribution of forest activity in the overall malaria burden.

Our study shows that untreated nets have a significant protective effect on the risk of clinical malaria and malaria infection. The ORs estimated by ACD and cross-sectional surveys were similar and highly significant, showing a protective effect of approximately 60%. The protective effect of insecticide-treated bed nets has been frequently reported,23–25 but there are few reports on the impact of untreated bed nets.26,27 We did not measure individual bed net use in our study; instead we used the bed net coverage per household as a proxy measure assuming that people are more likely to use them if they are available in their house. This is likely to underestimate the true individual protection as a result of an effect dilution among all people sleeping in the household (under and not under net). At the same time, unmeasured confounding factors such as socioeconomic status might have overestimated such protection. However, the magnitude of the protection was consistent between the different measurement methods (ACD and cross-sectional surveys). Therefore, our estimation of the protective effect of untreated bed nets is probably close to the true value.

Age in itself was not a significant risk factor for clinical malaria after adjusting for forest activity and bed net availability. Nevertheless, the odds of having antibodies to P. falciparum (i.e., the odds of infection) significantly increased by age groups even after adjusting for all the other risk factors. This suggests that outdoor night activities could be an important risk factor for malaria infection inside the village since the main vector, An. dirus A, is highly exophagic and exophilic. Human landing catches in Village 3 were found to be four times more frequent outdoors than indoors.7 Children go to sleep earlier and usually under bed nets when available, while adults stay outside and often sleep in hammocks in front of the house during the hot season.

In conclusion, most of the malaria burden in Village 3 is attributable to forest activity. Transmission inside the village is low but nevertheless high enough to observe a significant protective effect of untreated bed nets. Moreover, transmission occurs also outdoors because the main vector is highly exophagic and exophilic. The role of the numerous secondary vectors collected in Village 3 has yet to be defined. In this situation, control measures should focus on the protection of forest workers who constitute an important reservoir of parasites, and the impact of new interventions such as insecticide-treated materials (hammocks, sheets, etc.) specifically targeted to them should be tested in field trials. Ecologic and sociologic factors, movements within and towards the forests, and migration are important features in the dynamics of forest malaria transmission resulting in an epidemiologic mosaic with large site-specific variations. In this context, epidemiologic surveillance based on serologic follow-up can be a cost-effective tool for monitoring malaria morbidity, identifying site-specific risk factors, and evaluating new interventions to control forest malaria.

Table 1

Baseline characteristics of the study population

Total number of individuals573
* No = no bed nets or more than three persons per net; yes = 1–3 persons/net.
Total person-years956.7
Median follow-up time, weeks (range)98 (12–99)
    Median date of entry, week (range)1 (1–82)
    Median date of exit, week (range)100 (14–100)
Males/females (ratio)285/288 (1.0)
Median age at entry, years (range)19 (0–78)
Age (years) categories, no. (%)
    <116 (2.8)
    1–9154 (26.9)
    10–19117 (20.4)
    20–39190 (33.2)
    40–5969 (12.0)
    ≥6027 (4.7)
Ethnic group, no. (%)
    Kinh571 (99.7)
    Chinese2 (0.3)
Bed net availability,* no. (%)
    No186 (32.4)
    Yes370 (64.6)
Missing data17 (3.0)
Forest activity, no. (%)
    Never342 (59.7)
    Occasionally56 (9.8)
    Regular work (not sleeping in forest)129 (22.5)
    Regular work and sleeping in forest46 (8.0)
Professional categories, no. (%)
    None (children, students)235 (41.0)
    Woodcutters119 (20.8)
    Farmers53 (9.2)
    Housewives, retired78 (13.6)
    Others (business, officer, driver, worker, etc)88 (15.4)
Status, no. (%)
    Residents469 (81.9)
    Emigrants58 (10.1)
    Immigrants28 (4.9)
    Several movements (immigrate, emigrate)18 (3.1)
Table 2

Risk factors for Plasmodium falciparum clinical episodes determined in active case detection by univariate analysis*

Risk factorsRateCases/person-yearORP95% CI
* OR = odds ratio; CI = confidence interval.
Overall0.11102/956.7
Sex
    Male0.1571/466.81
    Female0.0631/489.90.420.001(0.25–0.69)
Age, years
    0–90.0410/251.31
    10–190.1327/212.42.890.01(1.28–6.53)
    20–390.1651/3183.740.002(1.65–8.48)
    ≥400.0814/175.11.930.21(0.69–5.42)
Profession
    None0.0623/400.01
    Woodcutters0.2956/192.25.05<0.001(2.70–9.44)
    Hut farmers0.1413/95.72.420.07(0.92–6.38)
    Housewives, retired0.045/123.50.750.58(0.27–2.07)
    Others0.035/145.30.630.36(0.23–1.71)
Forest activity
    Never0.0527/578.41
    Occasionally0.076/85.71.560.33(0.63–3.88)
    Regular work (not sleeping in forest)0.2451/211.45.20<0.001(2.81–9.64)
    Regular work and sleeping in forest0.2218/81.24.63<0.001(2.46–8.72)
Bed net availability
    No0.1852/295.61
    Yes0.0744/636.90.40<0.001(0.24–0.66)
    Missing data0.256/24.2
Season
    Dry: Jan–Apr0.0621/331.01
    Rainy: May–Aug0.1445/323.62.180.006(1.25–3.81)
    Sep–Dec0.1236/302.01.880.020(1.11–3.19)
Table 3

Risk factors for clinical Plasmodium falciparum cased determined by multivariate-adjusted analysis*

Risk factorsORP95% CI
* using generalized estimating equation. OR = odds ratio; CI = confidence interval. An interaction term was included between regular work and sleep in forest and bed net protection. For the interaction term, OR = 4.70, P = 0.01, by Wald test.
† OR for regular work and sleep in forest (compared with never/occasional) according to bed net availability.
‡ OR for bed net availability (compared with non-availability) according to forest activity.
Forest activity
    Never/occasional1
    Regular work (not sleeping in forest)3.85<0.001(2.15–6.89)
Effect of regular work and sleeping in forest*†
    In people not using bed nets1.720.25(0.68–4.39)
    In people using bed nets8.09<0.001(3.96–16.54)
Effect of bed nets according to forest activity‡
    Never/occasional0.310.004(0.14–0.68)
    Regular work, not sleeping in forest0.380.025(0.16–0.89)
    Regular work and sleeping in forest1.470.39(0.61–3.53)
Season
    Dry1
    Rainy2.010.009(1.19–3.39)
Table 4

Participation rates and malariometric indices in the five cross-sectional surveys*

ParametersSurvey 1 Sep 1999Survey 2 Jan 2000Survey 3 Aug 2000Survey 4 Feb 2001Survey 5 Aug 2001
* PCV = packed cell volume; ND = not done.
† N = total number of individuals with an indirect fluorescent antibody test sample taken.
Number of participants338429398412389
Survey coverage, %62.582.177.081.176.4
Mean PCV. (SD)ND40.5 (4.6)39.6 (4.0)40.9 (4.1)40.9 (4.1)
Spleen rate, % (n)5.3 (18)2.1 (9)2.5 (10)3.4 (14)0.8 (3)
    In those 2–9 years old, % (n)6.0 (6)02.7 (3)0.9 (1)0
Parasite rate, % (n)5.0 (17)3.5 (15)3.8 (15)1.5 (6)1.8 (7)
    P. falciparum, no.139411
    P. vivax, no.461155
    P. malariae, no.00001
    In those 2–9 years old, % (n)7.0 (7)2.9 (3)1.8 (2)01.0 (1)
Symptomatic malaria, % (n)1.8 (6)0.7 (3)0.5 (2)0.5 (2)1.3 (5)
Seroprevalence, % (n/N†)21.0 (66/314)24.3 (103/424)22.0 (87/395)18.0 (70/390)20.7 (79/381)
Table 5

Risk factors for all Plasmodium falciparum infections (n = 405) determined by IFAT in five surveys by univariate analysis*

Risk factorsPrevalence (%)Cases n/NOR95% CIP
* IFAT = indirect fluorescent antibody test; OR = odds ratio; CI = confidence interval.
Survey
    121.02(66/314)1.30(1.02–1.67)0.033
    224.29(103/424)1.54(1.24–1.91)<0.001
    322.03(87/395)1.33(1.07–1.66)0.009
    417.95(70/390)1
    520.73(79/381)1.28(1.04–1.57)0.017
Sex
    Males30.25(258/853)1
    Females13.99(147/1051)0.36(0.25–0.52)<0.001
Age (years)
    0–91.88(10/532)1
    10–1912.04(55/457)1.99(0.80–4.94)0.14
    20–3933.45(193/577)10.25(5.31–19.79)<0.001
    ≥4043.49(147/338)16.54(8.34–32.83)<0.001
Forest activity
    Never8.21(75/913)1
    Occasionally24.00(36/150)2.83(1.44–5.56)0.003
    Regular (not sleeping)46.89(151/322)9.28(5.88–14.65)<0.001
    Regular work and sleeping in forest46.38(64/138)9.73(5.39–17.56)<0.001
Profession
    None (children, students)3.7132/8621
    Woodcutters51.84197/38030.92(17.43–54.87)<0.001
    Hut farmers43.3281/18720.27(10.54–39.00)<0.001
    Housewives, retired25.2161/2428.13(4.06–16.27)<0.001
    Others21.4650/2337.39(3.75–14.56)<0.001
Bed net availability
    No28.45(173/608)1
    Yes17.64(220/1247)0.58(0.40–0.84)0.005
    Missing data24.49(12/49)
Table 6

Risk factors for P. falciparum infections (n = 405) determined by IFAT in five surveys by multivariate-adjusted analysis*

Risk factorsOR95% CIP
* Using generalized estimating equation. An interaction was included between forest activity and sex. IFAT = indirect fluorescent antibody test; OR = odds ratio; CI = confidence interval.
† OR for the effect of regular forest activity (compared to never) separately in males and in females.
‡ OR for the effect sex (females compared to males) separately for different forest activity. For the interaction term, OR = 0.15, P < 0.001, by Wald test.
Forest activity
    Never1
    Occasionally2.85(1.36–5.99)0.006
Effect of regular forest activity according to sex†
    Males10.76(5.41–21.41)<0.001
    Females1.56(0.73–3.36)0.25
Effect of sex according to forest activity‡
    Never1.21(0.60–2.42)0.59
    Occasionally0.50(0.10–2.45)0.39
    Regular0.18(0.08–0.36)<0.001
Effect of age (years)
    0–191
    20–393.59(2.12–6.08)<0.001
    ≥409.18(5.10–16.50)<0.001
Bed net availability
    No1
    Yes0.38(0.23–0.58)<0.001
Effect of surveys
    Survey 1, 2, and 3 (first year)1
    Surveys 4 and 5 (second year)0.55(0.41–0.73)<0.001
Figure 1.
Figure 1.

Cohort selection procedure and exclusion criteria. ACD = active case detection.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 2; 10.4269/ajtmh.2004.70.110

Figure 2.
Figure 2.

Monthly incidence of cases of Plasmodium falciparum malaria. Months 1, 2, etc. refer to January, February, etc. The two years of follow-up were combined. py = person-year; Reg. = regular.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 2; 10.4269/ajtmh.2004.70.110

Authors’ addresses: Annette Erhart, Nico Speybroeck, Marc Coosemans, and Umberto D’Alessandro, Department of Parasitology, Prince Leopold Institute of Tropical Medicine, Nationalestraat 155, 2000 Antwerp. Belgium, Telephone: 32-3-247-63-08, Fax: 32-3-247-63-59, E-mail: aerhart@itg.be. Ngo D. Thang, Le X. Hung, Tran Q. Tuy, and Le D. Cong, National Institute of Malariology, Parasitology and Entomology, BC 10 200, Tu Liem District, Hanoi, Vietnam, Telephone: 84-4-854-30-34, Fax: 84-4-854-30-15, E-mail: nimpe @netnam.org.vn. Nguyen Q. Hung, National Institute of Malariology, Parasitology and Entomology, BC 10 200, Tu Liem District, Hanoi, Vietnam and Institute of Malariology, Parasitology and Entomology, 699 Trang Hung Dao Street, District 5, Ho Chi Minh City, Vietnam. Le V. Toi, Institute of Malariology, Parasitology and Entomology, 699 Trang Hung Dao Street, District 5, Ho Chi Minh City, Vietnam, Telephone: 84-8-835-31-17, Fax: 84-8-839-07-34, E-mail: vmcpimpehcm@saigonnet.vn.

Acknowledgments: We thank all the study participants in Village 3 and the two Hamlet Health Workers for their effective contribution to the present study. We also thank the staff of the Provincial Malaria Station of Binh Thuan Province, the Institute of Malariology, Parasitology and Entomology in Ho Chi Minh City and the National Institute of Malariology, Parasitology and Entomology in Hanoi for their full and constant support of this study.

Financial support: This work was supported by the Belgian Cooperation in the framework of the Institutional Collaboration between the Institute of Tropical Medicine in Antwerp and the National Institute of Malariology, Parasitology and Entomology in Hanoi.

REFERENCES

  • 1

    Phan VT, Tuy TQ, Hung LX, Ngo LV, 1998. Epidemiological characteristics of malaria in Vietnam 1992–1997. http://www.ec-malaria.org/mcis/book1_21.htm.

  • 2

    Schuftan C, 2000. A story to be shared: the successful fight against malaria in Vietnam. WHO WPRO and the global Roll Back Malaria Program. http://www.afronets.org

  • 3

    Tuy TC, 1998. Malaria situation in Vietnam January–June 1998. http://www.ec-malaria.org/mcis/book1_21.htm.

  • 4

    Verle P, Tuy TC, Lieu TT, Kongs A, Coosemans M, 1998. New challenges for malaria control in northern Vietnam. Res Rev Parasitol 58 :169–174.

    • Search Google Scholar
    • Export Citation
  • 5

    Cong LD, 1999. Results of malaria control in Vietnam 1992–1997 & malaria control plan for 1998–2000. http://www.ec-malaria.org/mcsi/lib.nsf/issue1

  • 6

    Phan VT, 1998. Epidémiologie du Paludisme et Lutte Antipaludique au Vietnam. Hanoi: Editions Medicales Vietnam.

  • 7

    Van Bortel W, Harbach RE, Trung HD, Roelants P, Backeljau T, Coosemans M, 2001. Confirmation of Anopheles varuna in vietnam, previously misidentified and mistargeted as the malaria vector Anopheles minimus. Am J Trop Med Hyg 65 :729–732.

    • Search Google Scholar
    • Export Citation
  • 8

    Manguin S, Kengne P, Sonnier L, Harbach RE, Baimai V, Trung HD, Coosemans M, 2002. SCAR markers and multiplex PCR-based identification of isomorphic species in the Anopheles dirus complex in southeast Asia. Med Vet Entomol 16 :46–54.

    • Search Google Scholar
    • Export Citation
  • 9

    Coosemans M, 2001. Identification and Characterization of Malaria Vectors in Southeast Asia: A Prerequisite for Appropriate Vector Control. Final report (2001). Antwerp: International Cooperation with Developing Countries (INCO-DC).

  • 10

    Hung LQ, De Vries PJ, Giao PT, Nam NV, Binh TQ, Chong MT, Quoc NTTA, Thanh T.N, Hung LN, Kager PA, 2002. Control of malaria: a successful experience from Vietnam. Bull World Health Organ 80 :660–666.

    • Search Google Scholar
    • Export Citation
  • 11

    Demedts P, Wery M, 1985. Indirect immunofluorescence using differentiating antigens in detection of imported malaria. Ann Soc Belg Med Trop 65 (Suppl 2):89–96.

    • Search Google Scholar
    • Export Citation
  • 12

    Zeger SL, Liang KY, 1986. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 42 :121–130.

  • 13

    Smith T, Schellenberg JA, Hayes R, 1994. Attributable fraction estimates and case definitions for malaria in endemic areas. Stat Med 13 :2345–2358.

    • Search Google Scholar
    • Export Citation
  • 14

    Forest Malaria in Southeast Asia, 1991. Proceedings of an Informal Consultative Meeting WHO/MRC. Sharma VP, Kondrashin AV, eds. New Delhi.

  • 15

    Kondrashin AV, Jung RK, Akiyama J, 1991. Ecological aspects of forest malaria in Southeast Asia. Sharma VP, Kondrashin AV, eds. Forest Malaria in Southeast Asia. New Delhi: 1–28.

  • 16

    Armrattana A, Kaewkantha S, Kamboonruang C, Webber RH, 1993. Villager activities and malaria illness in a Thai-Myanmar border area. Trans R SocTrop Med Hyg 87 :340–348.

    • Search Google Scholar
    • Export Citation
  • 17

    Butraporn P, Sornmani S, Hungsapruek T, 1986. Social, behavioural, housing factors and their interactive effects associated with malaria occurrence in east Thailand. Southeast Asian J Trop Med Public Health 17 :386–392.

    • Search Google Scholar
    • Export Citation
  • 18

    Lansang MA, Belizario VY, Bustos MD, Saul A, Aguirre A, 1997. Risk factors for infection with malaria in a low endemic community in Bataan, the Philippines. Acta Trop 63 :257–265.

    • Search Google Scholar
    • Export Citation
  • 19

    Luxemburger C, Thwai KL, White NJ, Webster HK, Kyle DE, Maelankirri L, Chongsuphajaisiddhi T, Nosten F, 1996. The epidemiology of malaria in a Karen population on the western border of Thailand. Trans R Soc Trop Med Hyg 90 :105–111.

    • Search Google Scholar
    • Export Citation
  • 20

    Rahman WA, Adanan CR, Abu-Hassan A, 1998. A study on some aspects of the epidemiology of malaria in an endemic district in northern Peninsular Malaysia near Thailand border. Southeast Asian J Trop Med Public Health 29 :537–540.

    • Search Google Scholar
    • Export Citation
  • 21

    Singhasivanon P, Thimasarn K, Yimsamran S, Linthicum K, Nualchawee K, Dawreang D, Kongrod S, Premmanisakul N, Maneeboonyang W, Salazar N, 1999. Malaria in tree crop plantations in south-eastern and western provinces of Thailand. Southeast Asian J Trop Med Public Health 30 :399–404.

    • Search Google Scholar
    • Export Citation
  • 22

    Somboon P, Aramrattana A, Lines J, Webber R, 1998. Entomological and epidemiological investigations of malaria transmission in relation to population movements in forest areas of north-west Thailand. Southeast Asian J Trop Med Public Health 29 :3–9.

    • Search Google Scholar
    • Export Citation
  • 23

    Alonso PL, Lindsay SW, Armstrong JR, Conteh M, Hill AG, David PH, Fegan G, de-Francisco A, Hall AJ, Shenton FC, Cham K, Greenwood BM, 1991. The effect of insecticide-treated bed nets on mortality of Gambian children. Lancet 337 :1499–1502.

    • Search Google Scholar
    • Export Citation
  • 24

    D’Alessandro U, Olaleye BO, McGuire W, Langerock P, Bennett S, Aikins MK, Thomson MC, Cham MK, Cham BA, Greenwood BM, 1995. Mortality and morbidity from malaria in Gambian children after introduction of an impregnated bednet programme. Lancet 345 :479–483.

    • Search Google Scholar
    • Export Citation
  • 25

    Nevill CG, Some ES, Mung’ala VO, Mutemi W, New L, Marsh K, Lengeler C, Snow RW, 1996. Insecticide-treated bednets reduce mortality and severe morbidity from malaria among children on the Kenyan coast. Trop Med Int Health 1 :139–146.

    • Search Google Scholar
    • Export Citation
  • 26

    Clarke SE, Bogh C, Brown RC, Pinder M, Walraven GE, Lindsay SW, 2001. Do untreated bednets protect against malaria? Trans R Soc Trop Med Hyg 95 :457–462.

    • Search Google Scholar
    • Export Citation
  • 27

    Genton B, Hii J, Al-Yaman F, Paru R, Beck H-P, Ginney M, Dagoro H, Lewis D, Alpers MP, 1994. The use of untreated bednets and malaria infection, morbidity and immunity. Ann Trop Med Parasitol 88 :263–270.

    • Search Google Scholar
    • Export Citation

Author Notes

Reprint requests: Annette Erhart, Epidemiology Unit, Department of Parasitology, Prince Leopold Institute of Tropical Medicine, Nationalestraat 155, 2000 Antwerp. Belgium.
Save