Application of magnetic cytosmear for the estimation of Plasmodium falciparum gametocyte density and detection of asexual stages in asymptomatic children

This was a cross-sectional study conducted in primary school children selected from five schools in two wards in Bagamoyo district, Tanzania. The study was carried out during a high malaria transmission seasons from May to June and from October to December 2014. Prior to commencement of the study, sensitization meetings were held in selected communities and their respective schools. School-teachers helped to distribute consent forms to the participants and arrange meetings with the children’s parents and/or guardians.

A total of 500 school children aged 6–14 years (whose malaria status was unknown) were enrolled of which 303 consented and were recruited for study. Out of the 303 children, 40 were recruited from Kongo primary school in Yombo ward (approximately 25 km from the laboratory) and 136, 49, 43 and 35 children were recruited, respectively from Kidomole, Fukayosi, Mwavi and Msinune primary schools in Kiwangwa ward (approximately 45 km from the laboratory). All 303 children whose parents consented were screened for malaria using RDT and MDM on the spot at the school premises. At the end of screening, the remaining blood was taken to the laboratory for microscopy on the same day. Both symptomatic and asymptomatic malaria cases were referred and offered appropriate treatment at nearby health facilities, according to the national malaria treatment guidelines [23, 24]. The inclusion criteria were child’s asymptomatic status, age between 6–14 years and parent’s consent.

This study received ethical approval from Institutional Review Board (No.IHI/IRB/No: 34-2013) of Ifakara Health Institute and Medical Research Coordinating Committee of the National Institute for Medical Research No. (NIMR/HQ/R.8a/Vol. IX/1705). Regional, district and community authorities in the study area were contacted and granted approval of the study. Prior to participation, a written consent of each respondent was obtained based on Informed Consent Forms (ICFs) designed for the study. Each parent or guardian was given the form that explained risks, benefits and confidentiality that would have been taken. Additionally, the confidentiality of all participants was assured by using unique identity study numbers. The parents and/or guardians, who consented, signed the ICFs and returned back to the study team through head teachers of the schools under study.

Three millilitres of venous blood from each participant were collected into heparin vacutainer tubes (Greiner Bio-One GmbH, Kremsmuenster, Austria) for RDT, LM and MDM. The RDT (ICT Malaria Dual Cassette Test; ICT Diagnostics, Cape Town, South Africa) was performed in the field for quick malaria screening as described elsewhere [25]. A volume of 6 and 3 µL of blood were used for thick and thin blood smears immediately after collection respectively. Parasites were counted in accordance to World Health Organization (WHO) standards to detect and quantify parasites and their stages [26]. Fifty microliter of whole blood was diluted 10× with warm phosphate buffered saline (PBS) (37 °C) to final volume of 0.5 mL and used for the MDM. The tubes containing a mixture of blood sample and buffer were kept at 37 °C before running into the prototype. All samples and buffers were maintained at 37 °C.

As a quality control measure two microscopists examined slides independently. Slides with high discordance were reexamined by a third microscopist. Samples were considered negative if no parasites were detected in 100 high-power fields of Giemsa-stained thick blood smears. Both, asexual and sexual stages of the parasites were assessed in thick smears by comparing ratio of infected red blood cells to uninfected ones. Thin smears were examined for parasite quantification. Gametocytes and asexual stage parasites were counted against 500 and 200 white blood cells (WBCs), respectively and densities (parasite per microlitre) were estimated using a factor of 8000 leukocytes/µL.

Prior to deployment of MDM for field-testing, its performance was validated and proven using P. falciparum laboratory strains, NF54 and 3D7. In vitro cultivation of these two strains was done at Case Western Reserve University, USA. The asynchronous parasite culture was set up at 0.5 % parasitaemia, 4 % haematocrit and maintained using complete malaria culture media [7, 27]. To obtain gametocytes, the cultures were left undisturbed for 4 days to allow induction of gametocytes. The young gametocytes were left to mature by changing the media daily while maintaining 1 % haematocrit. Mature gametocytes developed 15 days later. From the parasite culture stock solution of 1:10, MDM working solution was prepared by 1:10 serial dilution using 1× phosphate buffered saline (PBS) to obtain a working concentration of 1: 1000 dilutions for easy cell examination. The slides were Giemsa-stained and quantified under the microscope by counting 2000 total red blood cells (RBCs) according to WHO guidelines and distinguishing the number of infected and uninfected cells to determine parasitaemia. The slide reading results were recorded on an Excel sheet for further analysis.

The MDM assay was performed as previously described by [6] with modifications. The principle of MDM operation is shown in Fig. 1. Prior to its being shipped to Tanzania from the Cleveland Clinic, Lerner Research Institute, USA, it was tested and validated with the laboratory P. falciparum strains, 3D7 and NF54 at Case Western Reserve University. In the course of those preliminary studies a protocol was developed suitable for the planned field operation, and was tested for the intra-erythrocytic parasite enrichment by the MDM assay. In brief, the MDM apparatus as described in [6] was modified by equipping it with a programmable syringe pump (PHD 2000 Programmable, Harvard Apparatus Inc.) for automatic sample aspiration and infusion (back to the sample container), resulting in a 1.5× increase on average of the time of blood sample exposure to the magnetic field compared to single pass operation, and thus increasing the likelihood for the magnetic cell capture. The device was designed to process five different samples in a parallel fashion (Fig. 2), at approximately 10⁸ cell/mL number concentration, volume 0.5 mL each, with the flow rates set at 0.026 mL/min for aspiration and 0.013 mL/min for infusion, resulting in 42 min of total processing time of five different samples from five different individuals. Importantly, MDM concentrates infected RBCs on one area on the slide to increase sensitivity and decrease the time to read slides compared to LM. The MDM apparatus was a laboratory prototype but its principle of operation (magnetic separation in microfluidics channels) makes it amenable to miniaturization for field operation.

Briefly, blood samples were diluted using PBS (pH = 7.2) in a ratio of 1:10 before pumping 500 µL of each test sample through the flow channels. Infected RBCs were collected and concentrated on a 0.13 mm-thick Mylar sheet (also known as polyester slide) by the magnet to ensure maximum late parasitic stage enrichment from five samples obtained from the five different flow channels. The blood samples for the MDM analysis were prepared and processed freshly in the field. Cells captured on the Mylar slide were fixed by 100 % methanol for 30 s and stained with 10 % Giemsa for 30 min. Once the slides dried, they were observed and analyzed under a light microscope at 100× magnification.

Data were entered and processed in Microsoft Access (Microsoft Corporation, Microsoft Way- Redmond, Washington) and analysed in STATA 11 (StataCorp, College Station, Texas). Data were entered in 2 × 2 tables and analysed for sensitivity (the probability that the assay will be positive when the infection is present) and specificity (the probability that the assay will be negative when the infection is absent) using the following formulae: where TN represents true negative, TP true positive, FN false negative and FP false positive. The agreement between LM, RDT, and MDM were determined using STATA 11. The data sets were compared pair-wise by selecting one of the methods as a reference standard for TN and TP determinations (LM v. MDM, and MDM v. RDT).

The slides were counted against 2000 total RBCs under the microscope. Late stage parasites, trophozoites, schizonts, and gametocytes were enriched by MDM compared to thin smear counts, schizonts being the most observed developmental stage of the parasite on MDM slides compared to thin smear (TS) slides (Figs. 3, 4). The average corresponding parasitaemia density on the MDM slide and thin smear was 164.6 parasites/µL and 26.5 parasites/µL, respectively. The gametocytes enrichment by MDM for NF54 and 3D7 P. falciparum strains compared to TS slides was at approximately 10-fold.

Three methods, LM, MDM and RDT were compared to detect P. falciparum infections in 303 school children. Amongst the P. falciparum positive individuals as detected by LM, RDT and MDM, 88.2, 93 and 89.8 %, respectively, had body temperature below 37.5 °C. However, body temperature had no statistical significant influence on the proportion of P. falciparum infections among the asymptomatic individuals.

The results are summarized in Tables 1, 2. Table 1 shows prevalence of P. falciparum for all stages by all three tests (LM, RDT and MDM) and gametocytes prevalence (available from LM and MDM only). The prevalence of P. falciparum infections by MDM, RDT and LM were 19.5 % (95 % CI 15.2–24.4), 17.2 % (95 % CI 13.1–21.9) and 7.6 % (95 % CI 4.9–11.2), respectively. The difference between MDM and LM in detecting P. falciparum infections was highly significant (p < 0.01), while there was no significant difference when MDM was compared with RDT (p = 0.3748). The MDM results were further analyzed by sensitivity and specificity (Eqs. 1, 2) against the method that showed the highest prevalence of infection. The results are shown in Table 1. Thus the LM sensitivity was relatively low 36.5 % (95 % CI 32.2–60.5) while that of RDT was better than that of LM, 62.5 % (95 % CI 49.5–71.8) compared to MDM. Interestingly, the MDM specificities with respect to LM and RDT to all forms of P. falciparum were high, 87.5 % (95 % CI 71.2–89.6 %) and 89.0 % (95 % CI 82.9–91.4), respectively.

The prevalence of P. falciparum gametocytes was 1.3 % (4/303 95 % CI 0.4–3.3) by LM but was considerably higher and equal to 4 % (12/303 95 % CI 2.1–6.8) by the MDM (Table 1). The difference between the two methods in detection of gametocytes was statistically significant (p = 0.008). The sensitivity of the MDM to gametocyte detection was low 26.7 %, (95 % CI 0.6–80.5) but its specificity was high, 96.5 % (95 % CI 93.7–99.1) indicating that both tests do not differ significantly in specificity of gametocyte detection (Table 2B).