Study area
This study was carried out in in four most malaria endemic health areas in Kisantu health zone (HZ) in Kongo Central (KC), one of 26 provinces of DRC. Kisantu is located in Lukaya district covering a surface area of 2400 km² and a population of 153,188 inhabitants. In this area, the malaria incidence rate, in 2017, was 163/1000 Cases per person per year (DHIS2, 2017). The insecticide susceptibility testing conducted in the closest national entomological monitoring site, Kimpese (in the same province). In 2016 showed that An. gambiae s.l. was resistant to permethrin with 17% mosquito mortality in WHO susceptibility tests (95% CI: 11–26) and 76% to deltamethrin mortality (95% CI: 67–83)14.
Study design and Study population
A mixed entomological and epidemiological study was carried out in 4 health areas of Kisantu health zone. The epidemiological study consisted of a stratified multi stage cross-sectional study conducted in households with children aged 0 to 10 years. Informed and signed consent was obtained from the head of household. In case of illiteracy, we used fingerprint (thumb) consent in the presence of an independent witness.
The stratification took into account the model used in the strategic plan for the fight against malaria. This model divides the country in 3 zones according to the parasitic infection: hypo-endemic (˂5%), holo-endemic (6–30%) and hyper-endemic zone (30%). The entomological study consisted of identifying species of anopheles and their susceptibility to some insecticide.
Epidemiologic survey
Sample size
The number of children (n) included in the study was calculated using the following formula described by Kish, (1965) for proportion15.
\(\frac{{Z}_{\propto -\frac{1}{2}}^{2} . p . q}{{d}^{2}}\ge Deffx n\)
Where p represents the estimated prevalence of malaria among children (p = 0.471), q= (1-p) the proportion of children under five years who get negative blood smear. (Demographic and Health Survey II, DRC)16; \({z}_{\propto -\frac{1}{2}}\) Z is the level of confidence based on the normal distribution. Z is 1.96 if the alpha is 0.05 is in our calculation. d is the precision degree that we assumed at 5%; the design effect (deff) was assumed to 2. The minimal sample size computed was 765 children.
Sampling
In the first stage, 4 of the 15 Health areas within Kisantu Health Zone with the highest malaria prevalence were selected based on the 2017 parasitological surveys (Database of DLM 2017). In the second stage, 30 villages were selected randomly from the list of all the 72 villages in the four health areas using the Excel function rand. In the third stage, we conducted a census to collect a list of all children in these 30 villages between age of 0 and 10 years included. Then a systematic random sampling was used to identify children aged from 0 to 10 years per village as illustrated in the Fig. 1.
Data collection
A questionnaire was administered to each household of the selected children. Information on demographic characteristics (sex, age, literacy, occupation, and marital status of the parents/guardian), socio-economic status (of the parents/guardian) and related variables were collected. To assess the socio-economic level of the households, observations were made on household characteristics (i.e., material used for walls, roofs, and floors).
A thick blood smear (TBS) and a thin blood smear were taken from selected children in the households following WHO procedures17. Thick and thin smears were stained with 2% Giemsa, then stored in filing boxes and read with microscopes at the Clinic of the University of Kinshasa. When the thick smear was positive, a thin smear was used for the species determination and to assess the parasite density of the sample. The thick and thin smears were independently read by two readers. A third assessment was performed in National Institute of Biomedical Research (NIRB) when a difference in TBS interpretation or a difference in parasite density in the thin blood smear of more than 15% was observed between the two readers. The result of the third reader which was the most experienced was taken into account. Plasmodium infection prevalence was calculated using the formula: total children with positive (Plasmodium infection) TBS / total children tested. The infection prevalence by village was used to group villages according to the DRC NMCP’s 2016–2020 strategic plans as: (i) hypo endemic ≤ 5%, (ii) meso -endemic 6–30% and (iii) hyper-endemic > 30%2.
Haemoglobin levels were measured on the spot in the field using a portable automated HemoCue®Hb 301 hemoglobinometer system according to the manufacturer instructions. Anaemia was classified as mild, moderate and severe when haemoglobin concentrations were 9.0-9.9g/dl, 7.0-8.9g/dl and 5.0-6.9g/dl respectively18.
The weight of children was measured using digital scales (Seca GmbH, Hamburg, and Germany) to the nearest 0.1 kg. The height was measured using a mobile measuring Seca mat (Seca GmbH, Hamburg, Germany) for children 0 to 99 centimetres and a recumbent length board for children of height > 99 centimetres to the nearest centimetre. The outcome variable was undernutrition indicated by wasting, stunting and underweight status in children ˃6 months of age. Stunting (height/age), wasting (weight/height) and underweight (weight/age) were defined as per WHO child growth standards. Children with Z scores less than two standard deviations (≤ 2 SD) below the median of the WHO child growth standards were classified as malnourished (stunted, wasted or underweight), minus 3 SD indicates severe stunting, wasting, and underweight19. Z scores of less than − 2 SD were considered indicative of stunting20.
Entomological Survey
Adults Anopheles was sampled from each of the same 30 villages in June 2018. We selected randomly six households per village and performed human landing catches (HLC) in three of the houses for one night per house to estimate the biting rate. In the 3 other selected households, indoor resting mosquitoes were collected from the walls, roof and furniture using a standard battery-powered Prokopack aspirators in the three remaining houses between 6:00 to 8:00 AM. A total of 180 households were included in this survey (90 HLC, 90 aspirations).
The collectors of the HLC only worked for 6 hours at a time, so four collectors per household per night were needed. Collectors were selected from amongst the local population; after written informed consent, they were enrolled and trained on the entomological activities. The collectors were referred to a health center and were treated free-of-charge if any illness occurs during the HLC activities and up to three weeks after collections.
The collected mosquitoes were morphologically identified21 and subsequently stored in tubes with silica gel for later detection of sporozoites (CSP ELISA) and molecular analysis of alleles at the Kdr 1014 locus. Based on the biting rates and the Plasmodium falciparum prevalence, the P. falciparum entomological inoculation rate (PfEIR) were calculated this rate is a measure of exposure to infectious mosquitoes. It is usually interpreted as the number of P. falciparum infective bites received by an individual during a night or season (aPfEIR)22.
The sex and physiological status (fed, unfed, semi-gravid, and gravid) of the mosquitoes collected by Prokopack were determined. Subsequently they were dried on silica gel for the CSP by ELISA and the alleles of kdr 1014 locus, East and West together using the new developed technique from Huynd L23. at CDC Atlanta further analysis.
Susceptibility Testing
Anopheles larvae and pupae were collected in all villages included in the study and brought to the laboratory where they were reared to adults. Adults Anopheles species aged 3–5 days were tested using WHO susceptibility tube tests against the diagnostic concentrations of deltamethrin 0.05%; permethrin 0.75%; alpha-cypermethrin 0.05%; DDT 4%; bendiocarb 0.1% and PBO 4% following WHO guidelines24. After 24 hours, mosquitoes were scored as dead or alive, identified morphologically and stored in tubes with silica gel for subsequent molecular analysis. We corrected observed mortality using Abbott’s formula when control mortality varied between 5 and 20%. We repeated experiments when control mortality was greater than 20%25.
Parasite Infection Detection And Molecular Analysis
An. gambiae s.l. and An. funestus s.l. heads and thoraces were separated from the abdomen and Plasmodium falciparum sporozoite antigen was detected using ELISA according to standard protocols26. We performed ELISA tests at the NIRB. We extracted DNA from legs and wings for PCRs to identify species and resistance mechanism (kdr-w). We used SINE PCR to identify the species of 200 mosquitoes morphologically identified as An. gambiae s.l.27 Kdr resistance alleles 1014F were also determined by PCR on the same 200 mosquitoes using the protocol described by Huynd LY and al23. The PCR assays were conducted at the Centers for Disease Control and Prevention (CDC) laboratory (Atlanta, GA, USA).
Data analysis
Stata version 13.0 was used for statistical analysis. Descriptive statistics were used to summarize the characteristics of the study population. Continuous variables were reported using means with standard deviations when variables were normally distributed if not, medians with ranges were reported. Categorical variables were summarized in frequencies or percentage.
For the epidemiological study, the prevalence was calculated using all children who had a positive thick drop, over the total study population.
We used χ2 test to search for the association between strata of malaria transmission and some characteristics of the households (sex, age, weight, size, temperature, and number of people living in the household, number of sleeping spaces, occupation and level of education of the head of household, house type and nutritional status). Statistically significant associations were thereafter assessed in a logistic regression analysis to determine factors independently associated with malaria transmission. The odds ratio (OR) with a corresponding 95% confidence interval was reported to quantify the strength of association. Significance was set at p-value of 0.05. Mapping of vector distribution and prevalence was also performed. Student’s t test was used to compare means of dead anopheles according to the insecticide.
The following formulas were used to calculate entomological indicators:
-
The sporozoite rate (SR) = (ELISA positive total number/ total number of Anopheles species tested) x 100
-
Human biting rate (HBR) = total number of Anopheles species collected by HLC during a specific period / total number of man-nights
-
Nightly entomological inoculation rate (EIR) = Nightly HBR x SR