The Committee on Human Research, Publication and Ethics of the School of Medical Sciences, Kwame Nkrumah University of Science and Technology granted approval for the study (CHRPE/AP/580/19).

The study was cross-sectional. Participants were selected from patients who were suspected of having malaria using the inclusion and exclusion criteria of the study. Dunkwa-on-Offin Municipal Hospital (200 patients), Kyerewere Health Center (50 patients), and St. Mark Hospital (60 patients) were selected from the municipality. These study sites were chosen from among three hospitals, four health centers, three clinics, and nineteen CHPS Compounds. Patients who were included in the study were from the surrounding villages. The study sites were chosen because most people from the rural communities attend these hospitals.

Patients who showed symptoms of malaria and were asked by a physician to perform a malaria test were included in the study. Those who read and understood the participant information leaflet and gave their consent were included in the study. Guardians of children who were below the age of 18 gave consent for their wards. The research protocol was explained to patients who could not read before they were included in the study. Patients who were not asked to perform a malaria test were excluded from the study, and patients who did not consent to the study were excluded from the study. A total of 310 participants were included in the research.

The sample size was calculated using the sample size formula

where: n = sample size, Z = 1.96 standard score at 95% confidence interval, M = margin of error of 5.0%, and P = estimated prevalence of malaria (28%) [10] Therefore, n = 1.96² × 0.28(1 − 0.28)/0.05² = 310. A total of 310 participants were selected for the research. The sample size was divided among the study sites based on the size of the facility and the frequency with which patients from villages attend them. The information concerning the attendance of patients was obtained from the Records Department of the hospitals. (Figure 1)

Ethylenediaminetetraacetic acid (EDTA) tube was labeled with the participant’s initials, date, and time of collection. The site of blood collection was cleaned using 70% alcohol. 2 mL of venous blood was collected with a syringe and needle and then transferred into the labeled EDTA tube. The blood sample in the tube was well mixed. Microscopy and Rapid Diagnostic Test were done immediately after blood collection [11].

Frosted-end slides were cleaned and labeled with numbers and dates. Two thick and two thin films were prepared immediately after the samples were collected. For the thin film, 2 µL of blood was pipetted and placed on the slide. A clean spreader slide was held at a 45˚ angle, toward the blood on the specimen slide. The spreader was pushed forward smoothly and rapidly.

Figure 1. The map of Ghana showing the location of the study in the Upper Denkyira East Municipality. The map was developed by one of the authors with ESRI ArcMap 10.8 using data from the Ghana Open Data Initiative and the Open Street Map Foundation.

To prepare the thick film, 6 µL of blood was pipetted and placed on a specimen slide. The blood was spread using the corner of a different clean slide. The spread was made into a circle with a diameter of 1 - 2 cm. The smear was made such that it was easy to read through. Both the thin and thick films were allowed to dry properly. The thin film was fixed in absolute methanol before staining [11].

Firstly, 9 mL of buffered water with a pH of 7.2 was poured into a beaker. Giemsa stock solution was filtered through Whatman filter paper into a 25 mL container. Using a clean pipette, 1 mL of the filtered Giemsa solution was added to the buffered water. The freshly prepared Giemsa solution was used within a maximum of 15 minutes.

The slide was placed in the working Giemsa stain for about 45 - 60 minutes. The thin film was rinsed by dipping it in the Giemsa buffer about 3 - 4 times, while the thick smear was left in the buffer for about 5 minutes. The stained slides were dried upright in a rack.

The thick film was used to detect malaria parasites and mixed infections. The prepared slide (smear) was first screened with a low magnification (×10) to detect parasites and then switched to the 100× oil immersion objective lens. Slides were declared negative after observing over 100 fields, with each containing about 200 white blood cells (WBCs).

The thin film was used for the species identification of the parasites. The prepared thin film was observed under the microscope using low magnification (×10 objective lens) to further detect parasites. The slide was examined thoroughly using the 100× oil immersion objective lens to identify species. Plasmodium species were distinguished based on the morphological characteristics of a blood smear. These include: the size of infected red blood cells and the nature of trophozoites observed in the blood film [11]. Two independent certified malaria microscopists read both the thick and thin smears, and the results were compared to ensure validation. The first microscopist examined the blood smear (thick and thin films) and reported the presence or absence of malaria parasites, species identification, and parasite density. The second microscopist independently re-examined the same slide without knowing the results of the first reader and provided his own diagnosis. All discrepancies between the results of the two independent certified microscopists were settled by a third, more experienced microscopist (tie-breaker).

The parasite density determination was done in the thick smear. Two tally counters were used to count the number of parasites against the WBCs. The count was done until at least 200 WBCs were counted. The number of parasites/µL of blood was determined by finding the number of parasites in relation to the standard number of 8000 WBCs/µL.

8000 WBCs/µL is considered as the standard number of white blood cells per µL of blood [11].

Malaria tests were done using three different RDTs. They were CareStart Malaria Pf (HRP2) Ag RT, Aria Malaria Pf/Pan Ag Rapid Test, and Acro Biotech Malaria Pf/Pan Ag Rapid Test.

The cassette was labeled with the patient’s initials. 5 µl of blood was pipetted from the sample already collected into an EDTA tube. The 5 µl of blood collected was transferred into the “S” well. Two drops of the buffer solution were added to the “A” well. The result was read after 20 minutes. The test was declared invalid when a line did not appear at the “C” portion. Such results were repeated. A result was declared negative when a line appeared at the “C” portion only. A result was declared positive for P.falciparum (Pf) when two lines appeared, one at the “C” portion and one at the “T” portion. The sensitivity and specificity of the CareStart test Kit were 98% and 97.5%, respectively, from the manufacturer.

The cassette was labeled using the participant’s initials. 5 µl of blood was pipetted from the EDTA tube. Two drops of blood lysis buffer were added to the centre of the buffer well (B well). The result was read within 20 - 30 minutes. The result was declared as negative when only the “C” line was present. The result was declared Pf positive or reactive when only the Pf line developed in addition to the control “C” line. This indicated the presence of the Plasmodiumfalciparum histidine-rich protein-II (pHRP-II) antigen [11].

The result was read as Pf negative but positive for any of the other three Plasmodium species (Plasmodiumovale, Plasmodiummalariae, Plasmodiumvivax) when only the Pan line developed together with the control line (“C”). This showed the presence of Plasmodium lactate dehydrogenase (pLDH) antigen. When both the Pf and the Pan lines together with the control line (“C”) developed, the result was read as Pf positive and any of the other three Plasmodium species also positive. This showed the presence of both pLDH and pHRP-II antigens. The result was read as invalid when no control (C) line developed.

The sensitivity and specificity of the Aria Rapid test kit were 100% and 100% for the detection of Pf and 95% and 100% for the detection of Pan, respectively. The sensitivity of the Acro Biotech test kit was 96.5% and 90.3% for detecting Pf and Pan, respectively. The specificity for detecting both Pf and Pan was 99.4%. The above sensitivity and specificity of the Aria Rapid Test and Acro Biotech RDTs were reported as given by the manufacturer.

2.13.1. Extraction of Plasmodium DNA

The samples were packed out of the storage freezer (−20˚C) and allowed to thaw. DNA isolation was performed using the Quick DNA kits (Zymo Research, USA; https://www.zymoresearch.com) with the manufacturer’s protocol. The procedure used was as follows: Blood samples were mixed thoroughly by pipetting up and down, and 100 µL aliquots were transferred into a labelled sterile 1.5 ml microcentrifuge tube. After which, 1 mL of TE (1 × Tris-EDTA buffer) was added, vortexed at 3000 rpm for 30 seconds, and then incubated at 4˚C for 10 minutes. After the incubation, samples were centrifuged at 12,000 ×g for 5 minutes, and the supernatant was discarded. The sediment was then suspended in a mixture containing 300 µL Genomic Lysis Buffer and 5 µL Proteinase K and mixed completely by vortexing for 15 - 30 seconds. At this stage, the whole mixture was incubated at 56˚C overnight.

After the overnight incubation, the mixture was vortexed again for 15 - 30 seconds and incubated at 80˚C for 10 minutes for improved lysis and inactivation of the proteinase K enzyme. The mixture was then vortexed and transferred into a labeled Zymo-Spin^TM IIC column in a microcentrifuge tube, after which centrifugation was performed at 12,000 ×g for one minute. The flow-through was discarded along with the microcentrifuge tube. At this point, the Zymo-Spin^TM IIC column was placed in a new microcentrifuge tube to begin the washing process. An aliquot of 200 μl of DNA Pre-Wash Buffer and 500 μl of g-DNA Wash Buffer were used independently for the first and second washing of DNA, respectively, with the previous centrifugation conditions.

To remove traces of washing solutions, centrifugation was repeated to dry the spin column. The dried spin column was transferred into a clean labeled micro centrifuge tube for the elution of DNA bound to the silica membrane. Finally, 70 μl DNA Elution Buffer was added to the spin column, incubated for 15 minutes at room temperature, and then centrifuged at 3000 rpm for 30 seconds to collect the DNA. The eluted DNA was stored immediately at −20˚C until ready for the Nested PCR test [11].

2.13.2. The ssrRNA Gene Amplification from the Plasmodium DNA

The amplification of the ssRNA gene of the Plasmodiumsp. involved nested-PCR as described by Snounou and Singh (2002) [12]. The pair of genus-specific oligonucleotide primers, rPLU1 and rPLU5, targets a 1.6 - 1.7 kb (kilobase) fragment of the gene for the first run of PCR (Nest 1; Figure 2). Subsequently, to further confirm the amplification of the Nest 1 PCR product, a Nest 2 PCR was run with a pair of genus-specific primers (RPLU 3 and 4) to amplify 235 bp (base pair) of the Nest 1 PCR product. In addition, the speciation of the Plasmodium involved the use of species-specific primers; P.falciparum (RFAL 1 and 2), P.malariae (RMAL and 2), P.ovale (OVA1 and RPLU2) and P.vivax (VIV1 and 2) to amplify 206 bp, 145 bp, 226 bp and 120 bp of the gene, respectively, of the Nest 1 PCR product (Nest 2; Figure 2).

Figure 2. Schematic diagram showing the Nest 1 and Nest 2 Plasmodium primers, their sequences, and corresponding fragment sizes. Source: [12].

In terms of Plasmodiumspp parasitemia among the samples collected from the study centers, Kyekyewere Health Center recorded the highest, followed by Dunkwa Municipal Hospital, and St. Mark Hospital recorded the least. Figure 3 below shows the distribution of Plasmodium parasites in the study area.

Figure 3. Distribution of parasitemia by study site.

It was observed that the ability of the RDTs to detect positive results increases as the parasite density (Plasmodium parasitemia) increases. Ninety-seven positive results by microscopy were compared to the CareStart, Aria Rapid Test, and Acro Biotech RDTs. At a parasite density of ≤1000 parasites/µL, microscopy detected one positive, while the RDTs did not detect any positives. At a parasite density of 1001 - 5000 parasites/µL, CareStart, Aria Rapid Test, and Acro Biotech RDTs all recorded 24 (77.40%) positives respectively out of the 31 positives for microscopy. At 5001 - 10,000 parasites/µL, out of the 27 positives for microscopy, CareStart, Aria Rapid Test, and Acro Biotech RDTs all recorded 24 (88.90%) positives. Out of 38 positives detected by microscopy, at a parasite density > 10,000, all the RDTs detected 37 (97.37%) positives. Table 2 below shows parasite density versus positive malaria cases identified by microscopy and RDT.

Table 2. Parasite density versus malaria positivity identified by microscopy and RDT.

RDT positive with microscopy negative was considered a false positive and not included as a malaria-positive case for Table 2 only. The various RDTs were compared with microscopy at various parasite densities.

The figures below, thus, Figures 4-8, show Nested PCR products under UV light. The DNA bands of various sizes (bp) clearly show the presence of various Plasmodium species. P1, P2, P3, P4, P5, P6, P7, P8, P9, P8, P9, P10 and P11 show some of the patients’ samples used for the Nested PCR. PC and NC show positive and negative control, respectively. Figures 4-8 were taken by our team during the research.

Figure 4. Gel electrophoresis image of Plasmodium PCR product (Nest 1).

Figure 5. Gel electrophoresis image of Plasmodium falciparumNest 2 PCR product.

Figure 6. Gel electrophoresis image of Plasmodiumovale Nest 2 PCR product.

Figure 7. Gel electrophoresis image of Plasmodium vivax Nest 2 PCR product.

Figure 8. Gel electrophoresis image of Plasmodium malariae.

Speciation was done for all 310 samples using Nested PCR, RDT, and Microscopy. Mixed species infections were recorded by all diagnostic techniques. Figure 9 below shows Plasmodium species detected by Nested PCR, RDT, and Microscopy.

Figure 9. Distribution of Plasmodium species detected by PCR, RDT, and Microscopy.

When PCR was used as the “Standard Reference of comparison”, the sensitivity of CareStart, Aria Rapid Test, and Acro Biotech RDTs was 79.10%. Their corresponding specificities were all 99.43%. When the ROC analysis was performed, it was observed that CareStart, Aria Rapid Test, and Acro Biotech RDTs all had an AUC (Area under the ROC curve) value of 0.94. The kappa value of 0.57 obtained for all RDTs showed a modest agreement between PCR and the RDTs. A chi-square Test of Association indicated a significant difference between the various RDTs and PCR (p = 0.001). Microscopy had a sensitivity and specificity of 69.57% and 99.42%, respectively. Table 3 shows the diagnostic evaluation of RDTs and Microscopy using PCR as reference.

Table 3. Evaluating the performance of CareStart, aria rapid test, acro biotech RDTs, and microscopy with PCR as a reference.

Out of the 310 cases, 141 patients tested positive by microscopy or PCR. The percentages of positives for PCR, microscopy, and RDT were 98.6%, 68.9%, and 75.9%, respectively. Table 4 shows the comparison of the distribution of malaria-positive cases by diagnostic techniques.

Table 4. Distribution of Malaria and the diagnostic efficiency of the three diagnostic methods (N = 310).