Background
Malaria is one of the deadliest infectious diseases of humanity, which causes significant mortality and morbidity in the tropics, particularly in Africa [
1]. Malaria is a parasitic disease transmitted through the bite of an infectious female
Anopheles mosquito. Early diagnosis is very important for disease management and the effective treatment of malaria. Before the advent of malaria rapid diagnostic tests (RDTs), diagnosis was based on microscopy of thick blood smears, which is still the gold standard for malaria diagnosis. However, in a number of rural and semi urban settings where lack of equipment, trained personnel and electricity prevents this essential diagnosis, health practitioners diagnose malaria based solely on clinical evaluation of symptoms [
2]. RDTs offer a great potential for rapid immediate diagnosis of malaria infections, which has led to prompt and appropriate treatment of the disease, particularly in highly endemic rural settings [
3].
Presently there is a very large demand for malaria RDT kits, as the World Health Organization (WHO) has recommended its use and majority of National Malaria Control Programmes have accepted it as the first step in the diagnosis of malaria. Due to the importance of the results of this initial screen, the WHO has established two programmes, the Foundation for Innovative New Diagnostics (FIND) malaria RDT quality assurance programme and WHO-FIND malaria RDT lot testing programme whose main mandate are to ensure accurate diagnosis of malaria [
4]. Malaria RDT kits are designed to detect either
Plasmodium falciparum specifically or discriminately detect both
P. falciparum in addition to another human malaria parasite or indiscriminately detect all human malaria parasites [
4,
5]. The main antigens that malaria RDT kits detect are PfHRP2, parasite lactate dehydrogenase (pLDH), and parasite aldolase (pAldo). PfHRP-2 is a
P. falciparum specific antigen with the advantage of being highly abundant and heat stable however, the PfHRP-2 antigen remains in circulation for up to 4 weeks after the malaria parasites have cleared [
6,
7]. Some monoclonal antibodies directed against PfHRP-2 have been found to cross react with PfHRP-3, a structural homologue of PfHRP-2 [
8,
9]. Thus although PfHRP-2 based RDT kits have the highest sensitivities [
4], they also have high false positive rates. By 2015, 171 different malaria RDT products had been tested by the WHO. Forty-five of these products detect only
P. falciparum, ten detect
P. falciparum as a part of a mixed infection with other human malaria parasites, one is
Plasmodium vivax specific and 115 detect and distinguishing
P. falciparum from either
P. vivax mixed infections or mixed infections containing all the other human malaria parasites,
P. vivax,
P. ovale and
P.
malariae [
10].
The accuracy of malaria RDT results can be affected by test antibody stability, product design and quality as well as the transport and storage conditions of the kits and sample parasite density [
10]. Accurate diagnosis of malaria by PfHRP-2 RDT kits can be affected by the
pfhrp2 and or
pfhrp3 genotype of the parasite [
5,
10,
11], the amount of PfHRP-2 antigen produced by the parasite [
12,
13] as well as the longevity of PfHRP-2 antigen after parasite clearance. One major obstacle in the diagnosis of malaria by RDT, without additional confirmation of parasitaemia is false positive test results, which leads to the unnecessary administration of anti-malarial drugs when no malaria parasites are actually present in the patient. False positive RDT test results are frequently obtained immediately following an anti-malarial drug regimen, when parasites are cleared or densities very low, but the antigen remains in circulation weeks later [
14,
15].
In some facilities in Ghana, where microscopy is unavailable, malaria is treated based on RDT results. It is thus very important to monitor the accuracy of RDT results as well as identify factors that affect the diagnostic ability of malaria RDTs. So far the main studies conducted in Ghana have determined the sensitivity and specificity of different brands of malaria RDT kits, including the CareStart™ and Paracheck RDT Kit [
13,
16‐
18], in different cohorts of malaria This study systematically identifies and confirms the presence of
pfhrp2 deletant (
pfhrp2−) parasites as well determines the prevalence of
P. falciparum parasites with deletions in
pfhrp2 and
pfhrp3 (
pfhrp3−) in two communities in Ghana.
Methods
Ethics, consent and permissions
This study was approved by the Institutional Review Board (IRB) of the Noguchi Memorial Institute for Medical Research, University of Ghana. Prior to enrollment, the study was explained to all participants after which written informed consent was obtained. Parental consent was obtained from parents and guardians of all children in addition to child assent obtained from children between 12 and 17 years.
Study area and sample collection
Abura Dunkwa, also known as Abura, is the district capital for Abura-Asebu-Kwamankese district and the Cape Coast Metropolis of the Central Region with a rural population of 31,768 for children under 14 years of age [
19]. The Central Region is situated 165 km west of Accra (capital of Ghana). Malaria peak season coincides with the major rainy season between June through August. The community is a farming community.
Obom is in the Ga south municipality of the Greater Accra Region with a rural population of 22,368 for children under 14 years of age [
20]. Malaria is perennial although it increases during the peak rainy season from June to August. The community is a fishing community. In 2014, malaria was estimated by microscopy to account for 35 % of all out patient visits at the local Obom health centre.
The study utilized a total of 226 spent PfHRP-2 based RDT kits as well as 558 filter paper blood blots from consenting healthy children within the two study sites in 2015. Thick and thin blood smears as well as DBS samples were obtained from healthy school children as part of a monthly malaria-screening programme from February through May. In April, RDT was performed according to manufacturers instructions in addition to the DBS and blood smears. The spent RDT cassettes were stored at room temperature for a maximum of 1 week, after which their membranes were processed for DNA. Approximately 50 μl of finger-pricked blood was spotted on to filter paper to make the DBS and thick blood smears. DBS were kept in sealed plastic bags with a desiccant and stored at −20 °C for no longer than 2 weeks after which they were processed for genomic DNA (gDNA). The slides containing the thick and thin blood smears were air dried and stored in slides boxes.
A sample was defined as negative by microscopy when two independent microscopists confirmed the absence of P. falciparum parasites on a Giemsa-stained thick blood smear. A sample was considered PCR positive when P. falciparum parasite genotyping using standard WHO genotyping procedures yielded a product. A positive RDT result was referred to as RDT positivity results, while a sample was considered positive for P. falciparum by RDT when the positive test strip was confirmed by microscopy or PCR. RDT positivity is used frequently as an indication of malaria in some facilities in Ghana where microscopy is unavailable.
Microscopic estimation of malaria parasite
Thick and thin blood smears as well as dried filter paper blood spots (DBS) were each made from a drop (~50 μl) of finger-prick blood. The blood smears were processed and then stained with 10 % Giemsa for 15 min. The stained slides were subsequently air-dried and viewed under 100X oil immersion microscope. Two independent microscopists read the slides and parasitaemia was determined as the % of malaria parasite infected RBCs observed per 200 white blood cells (WBCs).
Genomic DNA was isolated from the membranes of the previously used PfHRP-2 RDT kits and dried filter paper blood spots (DBS) using either Tris–EDTA (TE) [
21] or chelex [
22]. Briefly, the RDT cassette was opened and portions between the filter paper through to the nitrocellulose membrane and some of the conjugated pad were cut and placed into a 1.5 ml microcentrifuge tube containing 200 μl TE; a separate scalpel was used for each RDT. Similarly, a 3 mm punch was used to punch two 3 mm
2 disks from each of the dried blood spot (DBS). Each sample pieces was put into a 1.5 ml microcentrifuge tube containing 200 µl TE. The sample tubes were heated at 97 °C for 15 min on a dry heating block, centrifuged at 10,000
g for 30 s after which the supernatant transferred into a 500 μl tube for storage at −20 °C. For the chelex extraction, 150 μl of 6 % chelex in PBS was added to the tube with the punched DBS disks. The tubes were then incubated at 95 °C for 30 min with intermittent mixing by vortexing followed by a quick centrifugation step. The samples were centrifuged at 6000
g for 6 min, after which 120 μl of the supernatant was transferred into a 500 µl tube for storage at −20 °C.
Plasmodium falciparum genotyping
The WHO malaria parasite genotyping protocol [
23] was followed with slight modifications. PCR reactions were carried out in 15 μl volumes for both the primary and nested reactions. Briefly, the 200 nM M2-0F and M2-0R primers were used to amplify 4 μl of gDNA using One Taq polymerase (NEB). The nested reaction was carried out using 1 μl of the primary PCR product with 200 nM each of the combination of S1Fw/N5rev for the 3D7 type alleles or S1Fw/M5rev for the FC27 type alleles. For GLURP, the G-F3 and G-F4 primer pair was used for the outer PCR reaction and the G-NF and G-F4 primer pair used for the nested inner reaction. All the PCR fragments and the digested products were viewed under UV after resolving on a 2 % agarose gel containing 0.5 μg/ml ethidium bromide. Samples were classified as positive by PCR genotyping if the MSP2 and or GLURP PCR yielded a product following gel electrophoresis.
PCR-based detection of pfhrp2 and pfhrp3 genes
The PCR amplification was adapted from Baker et al. [
23] with very minor modifications. Briefly, 2 μl of gDNA was used as a template in a 20 μl PCR reaction mixture that contained 200 mM of each primer and 1X AmpliTaq Gold
® Fast PCR Master Mix UP. The DNA was initially denatured at 96 °C for 10 min followed by 41 cycles of denaturation at 95 °C for 50 s, annealing at 55 °C for 50 s (
pfhrp2 gene) or 51 °C (
pfhrp3 gene) and extension at 68 °C for 1 min. The final extension was performed at 72 °C for 5 min then to 4 °C. Genomic DNA from Dd2.
(
pfhrp2−), HB3 (
pfhrp3−) and 3D7 (wild type) were used as controls for the PCR amplifications. The primers used in the amplification of the exon 2 regions of
pfhrp2 and
pfhrp3 were
pfhrp2-F1,
pfhrp2-F2 and
pfhrp2-R1 for
pfhrp2;
pfhrp3-F1,
pfhrp3-F2 and
pfhrp3-R1 for
pfhrp3, as previously listed [
23].
PCR amplification for all samples that gave a negative result for any primer set was repeated using twice the volume of gDNA as template. All PCR amplifications were either nested or semi nested.
Resolution of PCR amplicons by agarose gel electrophoresis
PCR products were separated by electrophoresis on a 2.0 % agarose gel stained with ethidium bromide in 1X TAE buffer. 10 µl of PCR amplicons were loaded onto the gel, which was run for 1 h at 100 V then observed under UV light. The resolved fragment sizes were determined by comparison with 0.5 μg/μl Gene Ruler 100 bp DNA ladder (Thermo Scientific) loaded on the same gel.
Data analysis
Crosstab descriptive analysis was performed using IBM SPSS Statistics (version 22). Microsoft Excel was used to draw the table and graphs.
Discussion
The recent recommendation for accurate classification of
pfhrp2− parasites calls for an initial microscopic evaluation of the parasites, followed by
Plasmodium species-specific PCR analysis, after which confirmation is carried out by
pfhrp2 specific gene amplification to determine the absence of the gene or antigen analysis using a second quality PfHRP-2 based RDT or PfHRP-2 based ELISA [
24]. Many studies have reported the presence of
pfhrp2− parasites in a number of South American countries [
25‐
28], one study analysed 68 isolates and did not find any deletions in the
pfhrp2 gene nor its flanking sequences but rather found 50 % of the isolates to have deletions in the
pfhrp3 gene and its flanking genes [
27]. Such variation between nearby countries raises the need for all malaria endemic countries to engage in nationwide
pfhrp2 surveillance.
In Ghana, RDTs are used for malaria diagnosis throughout the year, during both peak and off peak seasons. A common practice in a number of health facilities is to rule out malaria in patients that test negative with an RDT kit, without further confirmation. This makes accurate malaria diagnosis using RDT kits very essential for malaria control.
The few previous studies on the PfHRP-2 based RDT kits in Ghana have focused on determining the sensitivity and specificity of Pf-HRP2 RDT kits [
17,
29]. This study provides some preliminary evidence for the existence of
pfhrp2− parasites as well as determines how mutant parasites with deletions in one or both
pfhrp2 and
pfhrp3 influence the accuracy of malaria diagnosis by PfHRP-2 RDT in Ghana.
Parasite prevalence estimated by PCR genotyping of 70 % for February to May was almost twice what was estimated by microscopy of corresponding thick blood smears in Accra. In Cape Coast, parasite prevalence estimated by microscopy of 0.7 % was only a small fraction of that estimated by PCR (41.2 %) (Fig.
1a). This suggests that more sensitive diagnostic tools are needed to accurately diagnose malaria in settings with a high prevalence of sub microscopic parasites.
To estimate the true prevalence of pfhrp2− parasites, which includes double pfhrp2− and pfhrp3− (pfhrp2−/pfhrp3−) parasites; each gDNA sample was analyzed by PCR genotyping prior to pfhrp2 and pfhrp3 exon 2 PCR.
False positive PfHRP-2 RDT results are not uncommon in malaria endemic settings as the PfHRP2 antigen persists for weeks after parasite clearance; however, in Accra where
P. falciparum parasite prevalence was high, the PCR estimate of parasite prevalence was comparable to the RDT positive rate (Fig.
1b). Despite the similar diagnostic read out between PCR genotyping and RDT, PCR confirmed the presence of
P. falciparum in 94/114 of the RDT positive samples, suggesting a false positive rate of 17.5 % (20/114). Microscopic evaluation of the RDT samples increased the false positive rate to 60/114 (52.6 %) in Accra. PCR genotyping identified 26/38 negative branded RDT kits to be positive for
P. falciparum, out of these 26 samples, 12 were positive by microscopy.
In Cape Coast, the RDT positive rate was higher than parasite estimation by both microscopy and PCR, confirming PfHRP2 antigen persistence. There were only two positive microscopy slides over the entire 4 months (Table
2). The low prevalence and density of
P. falciparum parasites causes the persistence of PfHRP2 to become more evident.
False negative RDT results are obtained when parasite carriage is confirmed by either microscopy or PCR, however the RDT kit produces a negative test results. This can have severe consequences in malaria endemic settings where negative RDT kit results are not confirmed by any other diagnostic tests such as microscopy. The prevalence of false negative RDT results increased from 18/38 when the samples collected in Accra were confirmed by microscopy to 26/38 when confirmed by PCR (Table
2). This increase was due to PCR confirming more samples as parasite positive than microscopy. Twenty-three percent (6/26) of the false negative samples carried deletions in the
pfhrp2 gene (Table
3), which suggests other factors including low parasite density contributed more to the negative RDT diagnosis than deletions in
pfhrp2.
In Cape Coast, PCR genotyping confirmed the presence of 12 false negative RDT tests. All 12 samples were positive for
pfhrp2 by exon 2 PCR, confirming our observation that factors other than
pfhrp2 deletion, including the high prevalence of submicroscopic parasites accounted for the false negative RDT results (Table
3).
Persistence of PfHRP-2 antigen from a recent past infection could explain the false positive RDT test, however possible compensation of
pfhrp3 for the lack of
pfhrp2 could also contribute to the positive RDT results obtained in samples that lacked
pfhrp2. Eighty-eight percent of the samples from Accra that were
pfhrp2− but were positive by PfHRP2 RDT and PCR genotyping were
pfhrp3+ (Fig.
2). Although the sensitivity for the CareStart™ PfHRP-2 RDT has proven to be very high in relation to microscopy in the recent WHO screen [
4,
10] and between 100 and 96 % in Ghana [
16,
17]. During the off-peak malaria season, the prevalence of false negative tests was as high as 68.4 % by PCR and 52.6 % by microscopy. The specificity of the CareStart™ PfHRP-2 RDT has previously been found to be between 70 and 73 % in Ghana [
16,
17], however the false positive RDT results obtained in Accra were 17.5 % by PCR and 52.6 % by microscopy.
Double
pfhrp2−/
pfhrp3− parasites have been found to be as high as 25.7 % in some countries within the Amazon basin [
25,
28]. The prevalence of parasites with
pfhrp2−/
pfhrp3− was 28 % in samples obtained from Cape Coast over February through May (Fig.
3), however the subset of these samples that were analysed in April did not contain any double
pfhrp2−/
pfhrp3− parasite. The prevalence of double
pfhrp2−/
pfhrp3− parasites obtained in Accra over the months of February to May was 4.3 %, which was similar to 3 % that obtained in the samples collected in April.