Mapping lymphatic filariasis in Loa loa endemic health districts naïve for ivermectin mass administration and situated in the forested zone of Cameroon

The study was carried out in 31 health districts, located in 4 Regions in the forest zone of Cameroon. Thirteen health districts were surveyed in the East, 10 health districts were in the Center Region, four health districts each were in the South and Littoral Regions respectively. A total of 124 communities were involved, 4 from each of the 31 health districts. (Fig. 1; Supplementary material Figures S1 and S2). These communities are located in the forest area of Cameroon which are breeding sites of different filarial transmitting vectors and MDA has not been introduced in these communities. The survey was conducted from July 2016 to January 2017.

This study was a cross-sectional community-based survey. Following mass sensitization on the importance of the study, a complete census of all households was carried out in each community and 50 households selected using simple random sampling. A minimum of 2 participants were randomly selected from each of the selected households.

In each community, at least 125 participants both male and female of greater than 5 years, who had been resident in the community for at least 5 years, were recruited and screened during the day for the presence of W. bancrofti using the Alere Filarial Test strip (FTS, Alere, Scarborough, ME, USA). All eligible participants in the community who willingly consented or accented (if below 21 years) were recruited for the study. Sociodemographic factors were collected using a structured questionnaire. We used the algorithm shown in Fig. 2 to detect the presence of W. bancrofti. Blood samples for FTS testing, thick blood film (TBF) and dry blood spot (DBS) for real time polymerase chain reaction (qPCR) were also obtained. Night blood samples (10:00 pm to 12.30 am) was collected from FTS positive individuals for microscopy to detect W. bancrofti microfilariae, and for qPCR.

Antigen testing was performed with FTS according to the manufacturer’s instructions. In brief, 75 μL of finger prick blood collected from eligible participants using non-heparinized microcapillary tubes (soda lime glass, Modulohm A/S Herlev, Denmark) was tested using FTS and the test was allowed to run for 10 min before being read. The result was recorded on the data record sheet. Individuals who tested either negative or positive on the FTS were informed of the result and those who were positive asked to return at night between 10:00 PM – 12:30 AM to have additional blood taken for microscopic evaluation of W. bancrofti mf. It was necessary to have the blood taken at night due to the nocturnal periodicity of W. bancrofti mf in the blood. Quality controls for the FTS strip were conducted daily before the exercise, using positive and negative controls from the manufacturer, to ascertain standards and kits performance throughout the study.

Standardized 50 μL of blood was collected with a non-heparinized microcapillary tube to identify mf of W. bancrofti (night blood of FTS positive individuals, between 10:00 pm to 12.30 am) or mf of L. loa (day blood, 8 am to 4 pm). In brief, the collected 50 μL of blood was placed on the centre of a clean slide, and spread repeatedly in a circular area of about 1.5 cm using the microcapillary tube. They were air dried and packaged for staining at the base. All blood smears were stained with 10% Giemsa within 24 h. The stained smears were examined using a light microscope at 10X objective lens (or using 40X objectives lens), for blood dwelling mf. Any mf present were identified based on the the size and presence or absence of a sheath, quantified and recorded.

DNA was extracted from the DBS, using the QIAGEN DNeasy kit (QIAGEN, Valencia, CA) following the manufacturer’s instructions. Briefly, one DBS of a participant was cut out and placed in a 2 ml microtube, covered with 270 μL of ATL buffer, incubated at 85 °C for 10 minutes, then at 56 °C for 1 hour after addition of 30 μL of proteinase K. Addition of 300 μL of Al buffer to the digested suspension brought up the lysate volume to a total of 600 μL. From this step, the lysates were heat-treated (100 °C), prior to the DNA purification, to denature the genomic DNA to make the DNA target sequence more accessible to the primers. A volume of ethanol equivalent to half the volume of lysate sample was added before loading the mixture onto a DNeasy spin column. Depending on the volume loaded, additional centrifugations were performed to pass all the solution through the column. After washing, (twice with AW1 buffer, once with AW2 buffer), purified DNA was eluted in 200 μL of AE buffer. The qPCR assays for W. bancrofti were performed using 2 μL of DNA and the W. bancrofti-specific long DNA repeat (LDR) primers as described by Rao et al. in 2006 [18]. For the detection of W. bancrofti the “long DNA repeat” (LDR) was used as a target. The nucleotide sequences for the forward and reverse primers were LDR 1, 5′ ATTTTGATCATCTGGGAAGGTTAATA 3′ and LDR 2, CGACTGTCTAATCCATTCAGAGTGA3 and the sequence for the probe was /56FAM/ATCTGCCCA/ZEN/TAGAAATAACTACGGTGGAT CTCTG/3IABkFQ. All assays were performed in duplicate using kappa probes master mix kit (Kappa Biosystems, Wilmington, MA) with 20 pmol of each primer (LDR1 and LDR2) and 6 pmol of LDR probe per well in the final volume of 20 μL and the fast PCR programmed automatically (95 °C temperature 40 cycles for 20 s, 60 °C for 1 s and 72 °C for 20 s) in a Step-One-Plus PCR system (Applied Biosystems, Foster City). The primers for L. loa PCR were LLMF72, 5’CGGAAGACTCAACGTCAGAAATCA3’ and 5’AGGAACGCTGATGGTGATGT3 and for the probe was /56FAM/CCAACAGCC/Zen/TGCTTT/31ABkFQ. qPCR were performed to identify DNA from L.loa mf as described by Fink et al. [19]. These entire tests were performed with 1 μL of extracted DNA (representing 0.1427 μL of whole blood).

Data generated either in the field or laboratory were collected using smart phones and uploaded to a center server coded with a password at the end of each day, to prevent accidents or the phones getting bad. The data were compiled and managed using Epi Info version 7.2 (Center for Disease Control and Prevention, Atlanta, GA) and Microsoft Excel 2013. It was checked for missing values and redundancy. All missing values were registered as missing.

After cleaning, the data was re- uploaded to the main server and hard copies in DVDs were also made. All hardware and software carrying the research data can only be accessed by the research team. Security password and server cupboard had been set up to protect the data at the University of Buea/REFOTDE, Buea, Cameroon.

Data were analyzed using the IBM SPSS Statistics (version 20.0, IBM, Armonk, NY, US). Some descriptive graphs were drawn with MS Excel 2010 and Thematic analysis were performed using the ArcGIS software (version 10.2, ESRI Inc. Redlands, CA, US), to draw the LF map in the 31 health districts and other maps in the analysis (Using GPS coordinates from each study community). The differences across age and gender were tested using fishers statistics and the 95% confidence interval (CI) computed. Chi-square, Mann-Whitney and Kruskal-Wallis tests were also used to compare LF and L. Loa prevalence level, the Geometric mean index of infection between regions and HDs, sex and age groups respectively. To assess the relationship between FTS positivity and L. loa Mf loads, the geometric mean intensity (GMI) of Mf counts was calculated as follows: GMI = (∑log(x + 1)/n),Where X = the number of Mf per ml of blood in Mf positive individuals, n = the number of mf positive individuals. Spearman rho correlation analysis was carried out between FTS positivity rate and TBF data (L.loa Mf load) at the regional level to find out if there was any relationship. Odds ratios were calculated to quantify the risk of an individual harboring L. loa Mf and testing positive to FTS compared to the risk for an amicrofilaraemic individual. The value was used to quantify the risk or the relation established by the correlation statistics. Logistic regression analysis was used to determine the predictors of FTS positivity. Gender, sex and quantity of L. loa Mf load were cross tabulated with FTS positivity.

The base map of the global administrative areas was downloaded from the Natural Earth (https://​www.​naturalearthdata​.​com/​) [20]. All maps were produced using ArcGIS Desktop v10.5 (Environmental Systems Research Institute Inc., Redlands CA, USA).

Trained medical personnel recruited for this study, examined all the participants for lymphedema. All males were examined for signs of the limb lymphedema and hydrocele and female for limb lymphedema. The lymphedema cases also underwent night blood collection for circulatory filarial antigen test and thick blood film microscopy.

In all, 14,446 participants took part in this survey and underwent the FTS test. Of these, 49.99% (7221/14446) were males with a mean age of 31.8 ± 19.8 years and 50.01% (7225/14,446) were females, with a mean age of 32.1 ± 20.1 years. 4151 (28.7%) persons tested were less than or equal to 15 years and 10,295 (71.3%) persons tested were > 15 years (Supplementary material Table S1, Supplementary material Fig. S3).

Out of a total of 14,446 individuals screened, 233 tested positive with the FTS (Table 1). The overall prevalence of positive FTS was 1.6% (233/14446), ranging from 0.0% in Abo and Bonassama HDs (Littoral Region) to 8.2% in Ayos HD (Center Region) (Fig. 3 and Supplementary material Table S1). Among the 31 HDs surveyed, 18 HDs had a prevalence of FTS positive rates equaled to or greater than 1% (Supplementary material Table S3). The males had a significantly higher FTS positivity rate of 2.0% (141/7221) compared to the females 1.3% (92/7225) (Fischer’s statistics, p < 0.05). The FTS positivity between the age groups was also statistically significant, higher in adults (15 years and above) with 2.1% (218/10295) compared to 0.4% (15/4151) in children (< 15 years) (Supplementary material Table S2).

In the Center Region, out of the 4475 participants that took part in the study, 2.4% (109/4475) were FTS positive (Fig. 3, Supplementary material Table S7). The prevalence in this region ranged from 0.5% (2/422) in Ngog Mapubi HD to 8.2% (37/452) in Ayos HD. The prevalence was higher in males 2.9% (62/2158), than in females 2.0% (47/2317), but the difference was not significant (p < 0.066). There was a significantly more positive FTS strips in adults 3.0% (101/3312) compared to children 0.7% (8/1163) (p < 0.001). (Supplementary material Table S3).

In the East region, a total of 6444 participants took part in the study with an overall prevalence of 1.4% (87/6444). The prevalence ranged from 0.2% (1/534) in Garoua Boulaye to 2.5% (12/481) in Lomie (Supplementary material Table S3). The 1.8% (12/481) prevalence among adults was significantly higher when compared to 0.3% (6/2005) in Children (p < 0.001). The distribution in females was lower than that of males, 1.1% (34/3147) and 1.6% (53/3297) respectively (p < 0.067) but not statistically significant.

A total of 1789 participated in the Littoral, with a general prevalence of 0.30% (5/1789), the FTS positivity ranged from 0.00% in Abo and Bonassama to 0.8% (3/397) in Dibombari. Here adults (> 15 years) were more FTS positive than children (< 15 years), with 0.40% (5/1261) and 0.00% (0/528) respectively but the difference was not significant (p < 0.334). Though not significant, males were more positive 0.40% (3/853) than females 0.20% (2/936) (p < 0.674) (Supplementary material Table S3).

In the South Region, a total of 1738 took part in the study recording a prevalence of 1.80% (32/1738). FTS positivity ranged from 0.60% (3/470) in Kribi to 2.7% (7/264) in Lolodorf. Males were more positive to FTS than females with 2.50% (23/913) and 1.10% (9/825) respectively (p < 0.031). Children were significantly less positive 0.20% (1/455) compared to adults 2.40% (31/1283) (p < 0.001) (Supplementary material Table S3).

A total of one hundred and sixty three (163) lymphedema cases were diagnosed in the entire study population, giving a prevalence rate of 1.14% (Supplementary material Table S1 and Figure S4). None of them tested positive with FTS for W. bancrofti antigen. Mbang health district in the East region had the highest number of lymphedema cases 36/503 (7.2%) followed by Ayos in the Center region 20/452 (4.4%). Both males 1,1% (82/7221) and females 1.1% (81/7225) were affected almost equally with lymphedema.

L loa mf prevalence for the 31 health districts is shown on Supplementary material Tables 1 & 8. The overall prevalence of L loa mf from diurnal blood was 12.5% (1805/14442). The prevalence ranged from 2.7% (50/1789) in the Littoral to 16.6% (745/4475) in the Center region (Supplementary material Table S8). A high mf prevalence was also recorded in the East region 12.3% (793/6442). The prevalence among the males of 14.8% (1071/7221) was significantly higher than that in females of 10.2% (734/7225) (p < 0.001) (Supplementary material Table S2). Between the two age groups, adults with 15.8% (1624/10295) were significantly more infected than children with 4.4% (181/4151) (p < 0.001). The thematic map showing L loa mf distribution across the 31 health districts is shown on Fig. 4.

We had 233 participants that were FTS positive. Microscopic results were available for all 233 for day time TBF and just 188 for night time TBF (Table 1). No W. bancrofti mf was found on any of these slides (day or night TBF). But, L. loa mf were present in 147/233 (63.1%) day blood TBF and in 109/188 (58%) night blood TBF (Supplementary material Table S9). The total mean mf load of L. loa (mf/ml) was 11,343 during the day and 2402 at night. The average L. loa mf GMI was 311 mf/ml during the day and 30.9 mf/ml during the night.

There was a positive correlation between the proportion of positive FTS and the prevalence of L. loa mf (Supplementary material Figure S5 and S7). In FTS positive individuals across the four regions, FTS positivity was significantly related to L loa mf load (p < 0.001) (Supplementary material Table S6), as the amount of L loa mf load increased, the probability of being FTS positive also increased, indicating that harboring L loa in the blood was a good predictor for having a positive FTS results.

The odd ratios of a person in any of the four regions with high L loa mf load of 8001–20,000 mf/ml being detected CFA positive by FTS was 53.2 times (OR = 53.2; 95% CI: 35.2–80.4; p < 0.001), compared to an individual in these regions with lower L loa mf load of 1–8000 mf/ml whose odds ratio to be detected CFA positive by FTS is just 6.1 (OR = 6.1; 95% CI: 4.4–8.6; p < 0.001). The OR becomes extremely high when mf load was > 30,000 mf/ml (OR = 151.9; 95% CI: 85.4–270.2; p < 0.001) (Supplementary material Table S4).

The logistic model also showed that gender is not associated with FTS positivity and L. loa prevalence (p < 0.076) but there was a contrast across age groups, adults had a significant risk of being FTS positive with high L. loa loads within these four regions with an OR of 3.395 times more than children (OR = 3.395; 95% CI: 1.942–5.935; p < 0.0001) (Supplementary material Table S10).

In Supplementary material Figure S6 a plot of positive FTS and L loa, showed that every increase in L loa Mf prevalence is followed by an increased in FTS positivity. The concordance between FTS positivity and that of Thick blood film (TBF) prevalence of L. loa in the surveyed Communities is in shown in Fig. 5.

The detection of W. bancrofti DNA by qPCR was done for 182 of the 229 night blood DBS collected from the FTS positive individuals. All tested negative for LDR DNA sequence, specific to W. bancrofti. However, 104 of the 182 samples (57.1%) were positive for L. loa specific LLMF72 DNA sequence (Table 2 and Supplementary material Table S11).

There was no significant difference between the qPCR results and those obtained from microscopy (Wilcoxon p value p < 0.999). The qPCR and microscopy results were plotted to depict how these tests performed in each HD with Supplementary material Figure S8.