Background

Global malaria control efforts are stalled. There were 249 million cases in 2022, more than 90% in Africa, where Plasmodium falciparum is the most prevalent species [1]. Progress toward malaria elimination is contingent on having to prompt, affordable, and sensitive diagnosis.

The World Health Organization (WHO) recommends parasitological confirmation to receive anti-malarial treatment [2]. Although light microscopy remains the gold standard, rapid diagnostic tests (RDT) targeting P. falciparum histidine-rich protein 2 (PfHRP2) protein and its cross-reactive histidine-rich protein 3 (PfHRP3) protein are widely used in sub-Saharan African countries due to their affordability, easy performance, and interpretation [2]. These proteins are abundant in parasite blood stages, but their function is unclear [3]. RDTs were introduced in Equatorial Guinea (West-Central Africa) in 2012 but have been more widely used in Bioko Island, where malaria control efforts have been scaled up, leading to lower prevalence. In contrast, there are less intense control measurements in the Continental region, leading to higher malaria prevalence [4].

PfHRP2 and PfHRP3 are encoded by subtelomeric genes located in chromosomes 8 and 13, respectively. Such areas are prone to deletions. Indeed, reports of RDTs false negatives associated with pfhrp2 and pfhrp3 gene deletions (pfhrp2Del and pfhrp3Del, respectively) are increasing worldwide, threatening the use of RDTs [3]. pfhrp2Del and pfhrp3Del have been reported in Central African countries, including Equatorial Guinea [5,6,7,8]. Single deletion (pfhrp2Del or pfhrp3Del) could lead to RDT positive diagnosis, as the remaining protein could be still detected [3]. Deletions of linked genes could be present upstream and downstream [9], reaching 40% of flanking deletions in Africa [10, 11]. However, patterns and extension of block deletions are still unclear [9, 12].

Clonal expansions of pfhrp2Del lineages have been reported in Peru [13], Ethiopia [14], Eritrea [15], and the China—Myanmar border [16]. This is interpreted as driven by selection due to RDT use and “test—treat” strategies. Contrastingly, there is no evidence of selection in pfhrp3Del [14]. Independent multiple origins have been reported for pfhrp2Del and pfhrp3Del12 [14, 15]. Although its impact is unclear [17], amino acidic sequence variability in the PfHRP2 and PfHRP3 proteins may also affect RDT sensibility [18, 19]. Thus, understanding the dynamic of pfhrp2Del and pfhrp3Del locally is critical to assessing RDT’s long-term efficacy.

This study describes and compares the changes in pfhrp2Del and pfhrp3Del prevalence before and after RDT introduction in Equatorial Guinea, considering two provinces with differences in their malaria control interventions. The origins and patterns of deletions, including flanking genes of pfhrp2 and pfhrp3, were explored using a comparative approach with other non-human malaria species. The evolution of exon 2 of pfhrp2 and pfhrp3 genes was also characterized using phylogenetic analysis.

Methods

Study site and sample collection

This study included samples separated by, at least, 18 years from two provinces of Equatorial Guinea: North Bioko, (Insular Region) and Litoral Province (Continental region) (Table 1; Fig. S1). Samples from North Bioko were collected from symptomatic children under 5 years old for 2001 and 2018 [20, 21], and samples from Litoral Province were from symptomatic children under 5 years old for 2001 samples [22], and from the asymptomatic general population for 2019 [22]. Whole blood was collected from finger pricks and preserved as dried blood spots (DBS) using Whatman 903TM paper. The DBS were stored in double zip-lock plastic bags with silica gel at − 20 °C, and while they were being used, they were stored at 4 °C.

Table 1 Sample information
Full size table

Ethical statement

Samples were taken from patients, or their parents or caregivers after they signed informed consent, guarded in the Ministry of Health and Social Welfare of Equatorial Guinea (MINSABS), and transferred to the National Centre for Tropical Medicine (CNMT-Spain) (C.0005278/C.0005279/C.0005485).

DNA extraction and detection of pfhrp2 Del and pfhrp3 Del

DNA was extracted from DBS using the Saponin/Chelex method [23]. Confirmation of P. falciparum infections where performed using the malaria PCR assay [23]. Independent amplification of two fragments, one including exon 1 and the beginning of exon 2 and the other including exon 2, of both pfhrp2 and pfhrp3 genes were performed by semi-nested or conventional PCR as previously described [23]. In the absence of amplification, the PCR was repeated two more times to confirm deletion. Plasmodium falciparum 3D7 clone (intact pfhrp2 and pfhrp3 genes) was used as a positive control, clone Dd2 (pfhrp2Del) as a negative control for the pfhrp2 gene amplification, and clone HB3 (pfhrp3Del) as a negative control for the pfhrp3 gene amplification. Finally, the amplification of single-copy genes pfdhps and pfdhfr was used as a control to ensure a sufficient amount/quality of DNA.

Flanking regions of pfhrp2 and pfhrp3 genes and data analysis

Genes flanking pfhrp2 (PF3D7_0831900 and PF3D7_0831700) and pfhrp3 (PF3D7_1372100 and PF3D7_1372400) genes were amplified according to Akinyi et al. [12]. Haplotypes of each sample were reconstructed by combining information on the pfhrp2 or pfhrp3 genes and their respective flanking genes. Comparisons between frequencies were calculated using an X2 or Fisher’s exact test depending on the number of samples included in each group. P-values of < 0.05 were considered statistically significant. All data were analysed using R v4.0.

Comparative study of pfhrp2 and pfhrp3 genes in subgenus Laverania

Pfhrp2 and pfhrp3 orthologous genes from Plasmodium species from subgenus Laverania were identified and downloaded using BLAST tool [24] in PlasmoDB [25]. Synteny of flanking pfhrp2 and pfhrp3 regions of the different species were analysed using JBrowse in PlasmoDB [25]. Available sequences of pfhrp2 and pfhrp3 genes were aligned using the ClustalX (v2.0.12), implemented in SeaView (v4.3.5) software with manual editing [26]. Phylogenetic relationships were inferred using both Bayesian method implemented in MrBayes [27] and a Maximum Likelihood method in IQ-TREE [28]. Then, phylogenetic trees were visualized and edited using the FigTree (v1.4.3) software [29].

Population genetic analyses

Genotyping was performed using fluorescently labelled PCR primers for amplification of seven P. falciparum standardized neutral microsatellite loci following previously published conditions [12, 30] (Table S1). Amplified fragments were multiplexed and separated on AB3730 XL DNA Analyzer (Applied Biosystems) by capillary electrophoresis. All the alleles for each locus with more than 200 relative fluorescence units (rfu) were scored using Geneious software. Peaks lower than one-third the height of the predominant peak were discarded [31]. Infections with more than one allele at any locus were considered multiple infections, and infections with only one allele per locus at all loci were considered single infections. No amplification was reported as missing data and not considered for defining the multiplicity of infection.

Nei’s index of genetic diversity (He), defined as He = [n/(n − 1)][1 − ∑] [32], was estimated for each locus and considering all for each group of samples using Haplotype Analysis software v1.04 [33]. He gives the average probability that a pair of alleles randomly selected from the population is different.

Population structure was analysed using a Bayesian model-based clustering approach- to infer the number of genetically related cluster or populations (K) based on the seven neutral microsatellite profiles obtained. Structure v.2.3.4 software was used with 430 samples (with only one or zero missing values) that have single infection or multiple infections with only one locus with more than one allele, so their haplotype can be faced, there were 56 and 147 from 2001 to 2018 respectively, from North Bioko, and 91 and 136 from 2001 to 2019 respectively from Litoral province. The admixture model was used in all analyses, allowing the presence of samples with ancestry in more than one of the K populations [34]. To assess the most likely K—number for the data, Delta K values were calculated using Structure Harvester v0.6.94, K value with higher delta K were then considered as the best clustering for the data [35]. Results for the most likely K—populations were interpreted using CLUMPP (Cluster Matching and Permutation Program) [36] and graphically displayed using distruct v1.1 [37].

Haplotype genealogies were estimated using an Minimum Spanning Tree—like structure inferred for the seven neutral microsatellites using the Global Optimal eBURST algorithm [38] implemented in PHYLOViZ software [39]. Haplotypes were compared according to province, year of collection, and pfhrp2Del or pfhrp3Del status.

Amino acid variation in pfhrp2 and pfhrp3 genes

Amplified products of exon 2 of pfhrp2 and pfhrp3 genes were purified using Ilustra Exoprostar 1-step (GE Healthcare Life Sciences) and sequenced using a standard dye terminator in an ABI PRISM 3730 XL Analyser (Big Dye Terminator v3.1 Cycle Sequencing kit) as was mentioned before. All samples were sequenced in both directions, using forward and reverse primers (6 pmol/μL) and sequences from samples with single infections were overlapped using Bio Edit Sequence Alignment Editor Software v7.1.3.0. Final sequences were then translated into amino acids. Analysis of the frequency and number of occurrences of each amino acid repeat [18, 40] or epitope [19] was carried out with a custom code using Python 3.9 Software [23]. That analysis was repeated for available sequences of subgenus Laverania.

Results

Frequency of pfhrp2 Del and pfhrp3 Del exon 2 per locality

First, pfhrp2Del and pfhrp3Del frequencies of the total 566 samples were analysed independently for each province, focusing on temporal changes, and then both provinces were compared to understand geographical differences. pfdhps and pfdhfr genes for all samples were successfully amplified.

In North Bioko, pfhrp2Del and pfhrp3Del frequencies increased from 1999 to 2018 (Table S2). Deletion of exon 2 of both genes reached more than 5% in 2018 (Fig. 1a). However, according to X2–test, the increase was only significant for pfhrp3Del (p–value < 0.05). Parasites conserving both genes were the most common over time, and the frequency of double–deletions or pfhrp2Del + pfhrp3Del was maintained low (Fig. 1b; Table S3).

Fig. 1
figure 1

Frequencies of pfhrp2Del and pfhrp3Del per sampled year and location. a The dotted lines are used to show the increase of the pfhrp2Del and pfhrp3Del per locality. b Frequency of combination of the presence and/or pfhrp2Del + pfhrp3Del

Full size image

Significant increases of pfhrp2Del and pfhrp3Del were observed in Litoral region, using X2–test (Fig. 1a). Deletion frequency of exon 1–2 and 2 of both genes reached 15% in 2019 (Table S2). Although, conservation of both genes was the most common genotype over the years, frequencies of single pfhrp2Del and double—(pfhrp2Del + pfhrp3Del) significantly increased in 2019 (Fig. 1b; Table S3).

Notably, before RDT introduction, the pfhrp2Del frequency was similar in North Bioko and Litoral provinces whereas pfhrp3Del was higher in the Litoral Province (Table S2). Although in both provinces pfhrp2Del and pfhrp3Del increased after introducing RDTs, their increase in Litoral province was higher in pfhrp2Del and pfhrp3Del genes (Fig. 1). Regarding exon 2—deletion combinations, the frequencies were also higher in Litoral than in Bioko province, where single pfhrp2Del and double pfhrp2Del + pfhrp3Del markedly increased, but only the increase in the single pfhrp2Del was significant (Table S3). Single pfhrp3Del was maintained at a relatively high frequency before and after RDT introduction (Fig. 1).

Haplotypes flanking pfhrp2 Del and pfhrp3 Del .

A higher frequency of deletions was found in downstream flanking pfhrp2 and pfhrp3 genes than in upstream. These downstream deletions increased over time reaching frequencies over 15% (Fig. 2; Table S4). Different haplotypes combining pfhrp2Del and pfhrp3Del and their flanking genes suggested that the deletions occur independently, being the haplotype with the three genes deleted the least common (Fig. 2).

Fig. 2
figure 2

Deletion patterns and combinations in pfhrp2 and pfhrp3 genes and their flanking regions per year and locality. Each row represents one sample. Each sample has the information of deletion or presence of an upstream gene, exon 1 and exon 2 of pfhrp2 or pfhrp3, and of the downstream gene. Green means the presence of the genes labelled above, and red means gene deletion. a Deletion patterns of pfhrp2 gene and its flaking genes. b Deletion patterns in pfhrp3 gene and its flanking genes

Full size image

High diversity of haplotypes was also found analysing the synteny of these regions among available annotated P. falciparum sequences and Plasmodium spp. from subgenus Laverania (Fig. 3a, b). For thepfhrp2 block, all the species and strains, except Dd2 (with pfhrp2 deleted), conserved the hsp70, PHISTa (pseudogene), hrp2, and stevor in that order, whereas with different gene lengths. However, immediately downstream of pfhrp2, they presented different genes from PHIST family in different copy—number (Fig. 3a). Similarly, for pfhrp3 block, all the P. falciparum strains, except HB3 (with pfhrp3 deleted), conserved PHISTb, hrp3, acyl-CoA, and stevor genes (generally as a pseudogene), but between hrp3 and Acyl-Coa there were different number of PHIST family genes or pseudogenes (Fig. 3b). Interestingly, pfhrp3 block of the other Laverania species have more than one Acyl-CoA gene copies followed by two PHIST family genes without stevor gene.

Fig. 3
figure 3

Haplotype diversity and phylogenetic relationships of hrp2 and hrp3 genes of Plasmodium species in the subgenus Laverania. a, b Haplotype diversity of hrp2 and hrp3 genes respectively, including P. falciparum and other Plasmodium spp. from subgenus Laverania with their region available. Full coloured squares represent genes, and white squares represent pseudogenes. Code before each row is the identification of P. falciparum strains and the other Laverania species (BILC: Plasmodium billcollinsi; GAB: Plasmodium gaboni; PRAE: Plasmodium praefalciparum; REI: Plasmodium reichenowi; ALD: Plasmodium adleri). When there is more than one code before the row means that all strains share the same genotype. c, d Phylogenetic relationship of hrp2 and hrp3 genes respectively. Posterior probabilities/bootstrap supports are shown for each node. *Indicate inconsistencies between the Bayesian and Maximum Likelihood methods

Full size image

Figure 3c, d show the phylogenetic relationships among the Laverania species for hrp2 and hrp3 genes, respectively. It is worth noticing that Plasmodium praefalciparum fell within P. falciparum clade in both cases, without clear separation from P. falciparum hrp2 and hrp3 genes. Additionally, these two phylogenetic trees do not resemble the phylogenetic relationships that have been reported before using other nuclear genes or mtDNA genomes [41].

Population genetic analyses

More than 50% of samples genotyped had infections with more than one allele in at least one locus (Table S5). Litoral province presented a higher frequency of multiple infections than North Bioko, the difference between provinces was significant for 2001 (p-value = 0.00895; X2 = 6.8313), but not for 2018–2019 (p-value = 0.6648; X2 = 0.1878). Both regions had opposite temporal trends in the multiplicity of infection, decreasing over time in Litoral province and increasing in North Bioko. However, such temporal differences were not statistically significant (Table S5). All seven neutral microsatellite markers were polymorphic for all samples, ranging in allele number from 1 to 20. Very high heterozygosity was found in all the groups (Table 2).

Table 2 Microsatellite fragment size ranges, number of alleles, and heterozygosity (He) per locus, year, and locality
Full size table

Population structure and origin of deletions

The analysis of genealogical relationships between the samples from North Bioko and Litoral province showed multiple origins of deletions (Fig. 4), for that analysis there were included 430 samples (Table S6). North Bioko province populations from 2001 (in red) and 2018 (in blue) were clearly separated conforming two different clusters (Fig. 4a). Contrastingly, there were different and dispersed origins for pfhrp2Del and pfhrp3Del, there was not a particular cluster of genotypes with deletions in the minimum spanning tree (Fig. 4b, c). However, some pfhrp2Del (in red) conformed small groups with more than two isolates. In the case of Litoral province isolates from 2001 and 2019, although some temporal diversification could be observed, the network showed associated isolates from the two populations (Fig. 4d). Also, the network representing gene deletions showed multiple different origins for pfhrp2Del and pfhrp3Del (Fig. 4e, f). However, there was a group where 2019—isolates with deletions were grouped with an old non-deleted isolate from 2001 (Fig. 4e, f; purple squares).

Fig. 4
figure 4

Minimum spanning tree of microsatellite allelic data by goeBURST showing genetic relatedness of P. falciparum isolated from Bioko North and Litoral province. Each genotype is represented by a circle, and the colour represents the different characteristic. a Tree for all samples from Bioko North province (2001 and 2018). b Tree showing the pfhrp2Del for Bioko North province. c Tree showing the pfhrp3Del for Bioko North province. d Tree for all sample from Litoral province (2001 and 2019). f Tree showing the pfhrp2Del for Litoral province. g Tree showing the pfhrp3Del for Litoral province. Purple squares remarks one group of samples that suggest the potential beginning of clonal expansion of parasites with deletions from an ancient sample without deletions

Full size image

Population structure and genetic relatedness between the two provinces

Using Structure v.2.3.4, only two clusters were identified for P. falciparum samples from Bioko North and Litoral provinces (k = 2, Fig. 5). The only cluster obtained for North Bioko Province samples from 2001 (in orange) was maintained in 2018, however, a second and more predominant cluster (in blue) appeared in 2018. Something to highlight is that the cluster detected in 2018 (in blue) was the main cluster in Litoral province population for both years. Importantly, there were no specific clusters linked to the pfhrp2Del and pfhrp3Del gene genotypes. Indeed, different, and independent origins for pfhrp2Del and pfhrp3Del deletions were obtained in this study (Fig. S2). However, there are small clusters of 2–5 isolates with single pfhrp2Del/pfhrp3Del or double deletions.

Fig. 5
figure 5

Bayesian cluster analysis of P. falciparum samples from Bioko North and Litoral provinces divided per year and pfhrp2Del and pfhrp3Del haplotypes (k = 2). Each column represents one sample. Each square below the graph represents number of samples included in each population; O = population represented in orange; B = population represented in blue; O/B = mixed population with percentage of each O and B population between 30 and 70%

Full size image

Amino acid variation in pfhrp2 and pfhrp3 genes

There were successfully analysed 144 sequences of pfhpr2 and 85 sequences of pfhrp3 (Table S6). There were no differences in amino acid repeats in the low complexity regions of both pfhrp2 and pfhrp3 genes between different years and localities (Table S7; S8). In the case of pfhrp2, amino acid repeats 2 (AHHAHHAAD) and 7 (AHHAAD) were the most common, with repeat 2 being the most abundant after RDT introduction, ranging from 1 to 10 or more repetitions per sequence. Although repeats 2 and 7 were present in 100% of samples in most years included in this investigation, North Bioko sequences from 1999 show repeat 19 as the most common (AHHAA, ranging from 0 to 24), whereas 2001 Litoral Province have type 4 (AHH, ranging from 0 to 31) as the most common. Plasmodium praefalciparum shows types 2 and 7 as the most common; however, type 4 was the most common in the other species from subgenus Laverania, except in P. adleri (Table S9). Regarding amino acid repeats found in the pfhrp3 gene, repeats 7 (AHHAAD) and 16 (AHHAAN) were the most common ones, being in 100% of sequences for all years except 2019. But repeat 7 appeared only one or two times per sequence, while repeat 16 appeared up to 15 times per sequence. In other species, this pattern changes, except P. praefalciparum, which also has type 16 as the most common repeat (Table S10).

Similarly, the frequency of epitopes was maintained over time. Specifically, 3A4 and C1-13 appeared in all samples, except in 1999, with a median frequency above 10 times per sequence (Table S9). These two epitopes were also the most common in P. praefalciparum, but not in other species (Table S12). In pfhrp3 gene sequences, only epitopes 3A4 and 1E1-A9 appeared in lower frequency (< 1 per sequence).

Discussion

This study shows a significant increase of pfhrp2Del and pfhrp3Del frequencies in Equatorial Guinea after RDT introduction, being also significant the differences between geographical areas. In addition, multiple origins of deletions were evidenced, probably due to the described instability of the subtelomeric regions [42] as suggested by sequencing and comparative phylogenetic analyses done in this study, without clear clusters of parasites with pfhrp2Del or pfhrp3Del. Notable, there were no significant differences in amino acid sequences of exon 2 of pfhrp2 and pfhrp3 before and after RDT introduction.

The reported high frequency of deletions post-RDT agrees with those from close West and Central African countries [3]. Nevertheless, this contrasts with the lower frequencies reported in bordering countries, such as Cameroon [6, 7] and Gabon [8], whose frequencies are closer to the deletion frequencies found in Equatorial Guinea among samples collected previously to RDT introduction. The high pre–RDT pfhrp2Del and pfhrp3Del and the observed temporal increase, considering also the intermediate published pfhrp2Del prevalence (5.3%) in Litoral province [5], establish Equatorial Guinea as a high-risk country for RDTs losing their efficacy.

One model of pfhrp2Del selection proposes that in areas with more than 1% starting frequency, the spread of deletions could be significantly faster compared with areas with < 1% starting frequency [43]. Indeed, it is expected a stronger selection pressure driven by an extended RDT use and properly implementation of “test—treat strategies” [44]. However, these results differ from such predictions. North Bioko province showed a lower frequency of deletions, but RDT is more widely used. Given the lower transmission in North Bioko, there could be complex interactions between inbreeding, selection, and the local epidemiology [44,45,46] affecting the frequencies of deletions. Regarding differences between island and continental regions, human mobility evidenced by parasite migration, even replacement, in population structure analysis, could explain the observed pattern. One hypothesis is that parasites with deletions could be first selected on the Island and then spread in the continental region.

It is worth noting that there could be bias introduced by the patient’s age. Samples from 2019, with significantly higher deletion frequency, were from adults, while the rest were from children. A high percentage of deletions is expected in adults, because they usually have less multiplicity of infection (MOI), which has been associated with the likelihood of detecting deletions [47]. According to methodological limitations, multiclonal infections could underestimate the frequency of deletions because clones with or without deletions in the same infection cannot be separated.

The results provide further that the subtelomeric location of these genes makes it difficult to assess the probability of deletion occurrence and that such events occur multiple times independently of RDT use. Such complexities are evidenced by the variability of these regions, composed by pfhrp2 and pfhrp3 and their flanking genes before RDT introduction, and their evolutionary histories when compared with non–human malaria Plasmodium spp. that are part of the P. falciparum clade. Add to this the possibility that inbreeding during low transmission seasons could facilitate the expansion of deletions even in the absence of selection by the use of RDTs.

The results disagree with previous reports that established that complete deletions of pfhrp2 and pfhrp3 genes are more common than partial deletions [30]. However, they agree with the higher frequency of deletions in downstream flanking genes for pfhrp2 and upstream for pfhrp3 [14], that have been related to their potential fitness impact [14], but the putative deletions fitness costs remain unsolved. A study suggested fitness cost for parasites with deletion [48] whereas another reported no fitness cost [49]. Perhaps such differences could be explained because one study used HB3 and Dd2 strains [49] with wider deletion blocks, while the other only had pfhrp2 and pfhrp3 deleted [49]. Thus, differences could be related to flanking genes [14].

It can be speculated that the rapid evolution of pfhrp2 and pfhrp3 genes, also evidenced by the phylogenetic analyses, together with the high reactivity and the lack of strong evidence of fitness cost for the deletions, suggest that HRP2 and HRP3 proteins could elicit a non- protective immune response. However, it is not clear if there is a specific immune response against pfHRP2 [50].

Multiple and independent origins of deletions [12] were identified, in agreement with previous reports. Although such lineages could spread as clusters of deletions [16], no evidence was found in this study. Nevertheless, complex dynamics involving multiple origins, the spread of deletions driven by selection, and demography emphasize the need for molecular surveillance to address the impact of deletions in Equatorial Guinea's malaria control efforts [46]. It seems that predictive models that consider such complexities are still needed. Moreover, population structure analysis was performed using only a subset of samples, the ones with simple infections or with multiallelic only for one locus. So, analysis using other techniques that include all samples, especially the multiclonal ones, could give insightful information.

The absence of temporal changes in amino acid diversity supports that this diversity did not have a direct effect on RDT sensitivity [17]. Indeed, the most frequent repeats and epitopes in Equatorial Guinea were in agreement with previous literature [23, 40], so RDTs based on them are recommended.

A problem in pfhrp2Del and pfhrp3Del studies is the differences in study designs. In this study, samples from different population groups and without data about RDT results were used, so the impact of these deletions on RDT efficacy was not assessed. However, to our knowledge, this is the first study on pfhrp2Del and pfhrp3Del dynamics in Africa that includes pre-RDT samples and comparisons with other Plasmodium spp. to understand the evolutionary dynamics of these genes.

Conclusions

Although there were differences, pfhrp2Del and pfhrp3Del were in high frequency in both Equatorial Guinea provinces. Deletion frequencies significantly increased after RDT introduction, however, they were present before RDT use. pfhrp2 and pfhrp3 genes and their flanking genes are in a highly unstable genome region where deletions had multiple origins, but clonal expansions of genotypes with deletions were not found in this study. High genetic diversity of pfhrp2 and pfhrp3 genes was observed for P. falciparum and other Laverania species, providing additional evidence of its rapid evolution. The relative frequency of amino acid repeats was maintained over the years and locations in P. falciparum. Thus, RDTs are likely to remain effective if the genes encoding the proteins are present. Overall, molecular surveillance of pfhrp2Del and pfhrp3Del is necessary to support decisions leading to mitigating the threat that they could pose.