Overreliance on Plasmodium 18S rRNA gene for malaria molecular diagnosis—inferences from systematic review

Abstract

Background

Molecular diagnosis of malaria through nucleic acid-based amplification test is important to detect low-density, sub-microscopic and residual infections, as well as to prevent importations and re-establishment. Reliance on single/limited molecular targets could be detrimental as evidenced by false-negative PfHRP2-based RDTs, and the same may apply to PCR targets. No systematic exploration of the commonly used PCR targets has yet been documented.

Methods

A systematic search was made using a previously generated database through PubMed® and Google Scholar® and supplemented by additional searches. All studies that used PCR for detecting Plasmodium infections were included in this study. Further information was retrieved on molecular targets used and the type of PCR assay used. An independent search was also made to explore the identification/development of newer molecular targets.

Results

Almost all studies (93%) used 18S rRNA gene as a molecular target. Nested PCR alone (68%) was the most frequently used assay. Eighty-five percent of the studies that exploited the 18S rRNA gene target and nested PCR used the approach developed in 1993.

Conclusion

Overreliance on a solitary molecular target (18S rRNA gene) for many years might be a cause for concern. Research is needed to validate newer multi-copy targets in terms of limit of detection, robust reproducibility, reduced costs, and a possibility of multiplexing.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12936-025-05424-4.

Background

Since 2000, the global malaria burden has decreased significantly in the 83 malaria-endemic countries. Despite the aim of eliminating malaria by 2030, certain African countries still bear the major burden of malaria in the world [1]. Timely accurate diagnosis, specific treatment, and vector control are the cornerstones for malaria elimination. Rapid Diagnostic Tests (RDTs), microscopy and nucleic-acid amplification tests (NAAT; including PCR and loop-mediated isothermal amplification or LAMP) are the currently available diagnostic tools, in decreasing preference for field-deployability or practicability but increasing order of the tool’s parasite detection limit and costs, for malaria parasite detection in humans. Although microscopy is considered to be the ‘gold standard’ for malaria diagnosis in field settings (limit of detection; LOD 10–100 parasites/µL of whole blood), RDTs (with LOD of ~ 200 parasites/µL of whole blood) are rapidly replacing microscopy as the most commonly used malaria diagnostic at the point-of-care [2, 3]. This means that both RDT and microscopy miss the low-density parasite infections (LDIs; < 100 parasites/µL of whole blood) for which PCRs are preferred [4, 5]. However, some new ultra-sensitive RDTs based on PfHRP-2 detection are available that seem to have a tenfold lower PfHRP-2 LOD than standard RDTs, but the corresponding LOD in terms of parasites per microliter is unclear. In addition, considering qPCR as the gold standard, their sensitivity varies between 44 and 84% when tested on samples from different regions [6, 7]. Apart from higher LOD than that of PCRs, it has been observed from Indian data that the diagnostic accuracy of microscopy is particularly low for mixed-species infections [8], and currently used RDTs are threatened by the rising concerns over PfHRP2-dependent false negative diagnosis for Plasmodium falciparum [9].

Although the published LOD for PCR is < 5 parasites/µL of whole blood [10, 11], it varies depending upon multiple factors including the target used for amplification, the number of copies of the target, the specificity of the oligonucleotide primer, type, volume and quality of the sample and the PCR technique itself. Thus, the LOD of PCR varies accordingly: nested PCR (6 parasites/µL if using blood spots), multiplex PCR (0.2–5 parasites/µL), quantitative PCR (0.02–1.22 parasites/µL), LAMP (0.2–2 parasites/µL), Nucleic acid sequence-based amplification (0.01–0.1 parasites/µL) [12–19].

The molecular methods for the detection of Plasmodium were first developed in the 1990 s by Snounou and colleagues, in the year 1993, exploited the small subunit ribosomal RNA (ssrRNA, specifically 18S rRNA gene) to design species-specific oligonucleotides for P. falciparum, Plasmodium vivax, Plasmodium malariae and Plasmodium ovale [20]. In successive years, various research groups designed oligonucleotides using different loci of the 18S rRNA gene that were able to detect the above four Plasmodium species that commonly infect humans [14, 21, 22]. However, it appears that researchers have become more interested in the oligonucleotides generated by Snounou et al. [13] and nested-PCR has become a well-established molecular strategy to screen Plasmodium infections than any other method(s).

Although the cost of molecular detection of malaria is a major deterrent, currently for use at the point of care, research is underway for optimization of costs [23–25] and multiple ways have been suggested to make molecular tests cost-effective [26]. Since multiple countries across the globe have significantly reduced malaria burden and are nearing elimination, it is recommended that LDIs need to be timely detected and treated to achieve elimination at a faster pace and to timely manage importations and prevent re-establishment of malaria in areas that have eliminated malaria [27]. However, reliance on one molecular target detection of malaria parasites becomes an evolutionary threat for diagnosis as it has been speculated and demonstrated for the RDT target, HRP2, for detecting P. falciparum infections [9, 28]. No literature has been reported that has quantitatively evaluated the reliance of 18S rRNA target for malaria molecular diagnosis. Therefore, this study was undertaken to assess whether 18S rRNA is more commonly being used for malaria molecular diagnosis in research and to identify the availability of various other 18S rRNA-independent targets for the detection of Plasmodium spp. by PCR.

Methods

Studies considered for the current analysis were taken from a previously generated database to screen the burden of mixed Plasmodium infections regardless of time or country of data collection and the method used for diagnosis. The detailed method for creating the database is registered on PROSPERO (CRD42021234278) and is mentioned elsewhere [8] (modified & recreated here in Fig. 1). Briefly, the searches were made on PubMed® and Google Scholar® in the year 2020 and contain all studies retrieved from the above search engines that attempted to detect mixed species Plasmodium infections.

Fig. 1.

Screening and selection of studies for inclusion. The flow chart was modified & recreated for the current study from the one previously published [8], and the results were supplemented with a separate search made on PubMed® for articles published from 2021 to 2025

Out of the studies identified, all studies in which any kind of PCR was used to screen for Plasmodium infection were included for the current analysis. The studies that lacked information on the type of PCR and/or that did not use PCR for the detection of Plasmodium spp. were excluded. Two reviewers (ND & AS) independently screened the titles and abstracts of all the eligible studies and all discordant cases were resolved after discussing with the third reviewer (PKM). The included studies were examined for the target DNA sequence used to generate oligonucleotide primers for PCR (18S rRNA or others), type of PCR assay used (conventional or real time qPCR and nested or semi-nested within conventional PCR) and within qPCR, the type of method used to detect amplification (hydrolysis-probe or SYBR green). A diagnostic PCR for Plasmodium detection was considered'nested'if it consisted of two cycles of PCR with two sets of primers [20]. The PCR was considered'semi-nested'if one of the primers used in the second round was the same primer used in the first round of PCR [29]. Studies using nested conventional PCR on the 18S rRNA target either performed independently or with qPCR were further classified based on the approach used (Snounou’s versus others) and studies using the Snounou’s approach were further dissected to find out whether the PCR was done in one single reaction or was multiplexed for detection of > 1 parasite species. The data from individual studies were collected independently by two reviewers (ND and SS) and discrepancies, if any, were resolved after discussing with AS. The proportion (%) of each parameter from the included studies was estimated and displayed as pie-charts. If any clarification was required, the corresponding authors of the respective studies were contacted via email/phone call, and their responses were used to correct the data accordingly, ensuring comparability. Data that remained unclear despite email or phone confirmation were not included in this analysis.

Further, to supplement the above results (extracted from the previous database made in the year 2020), a separate search was made on PubMed® from the year 2021–2025 on 18-02-2025. The terms ‘Molecular OR PCR AND Malaria OR Plasmodium AND Diagnosis’ were searched in the ‘Title/Abstract’. Only primary research articles, including clinical studies/trials and evaluation, validation, and comparative studies, were considered for further analysis. In addition, studies that used alternatives to 18S rRNA as a PCR target were also searched separately and compiled.

Results

The current study was undertaken to gauge the usage of the 18S rRNA region of the Plasmodium parasite for its molecular diagnosis via PCR. After searching from a previously made database (containing 149 studies; Fig. 1), a list was compiled that comprised 79 studies that used PCR. Of those, 15 were eliminated since they lacked information on the type of PCR used or did not use PCR to diagnose Plasmodium infections (rather, they used PCR to screen specific mutations). Thus, this analysis included 64 studies from the previous database (Fig. 1). The supplementary search (2021–2025) yielded 924 studies (Fig. 1), 903 of which were excluded after applying filters (as specified in the methods), and the full text of 21 articles was eventually screened. Following full text screening, 09 articles were excluded as found either irrelevant or did not use PCR, leaving 12 articles for the current analysis. Therefore, including previous database and supplementary searches, a total of 76 studies were finally considered (Suppl Table 1).

The most widely used target for PCR-based diagnosis was 18S rRNA (93%; 71/76) alone, followed by mitochondrial targets (3%; 2/76), varATS (3%; 2/76), and EXP1 (1%; 1/76), depicted in Fig. 2A. Out of the studies that used 18S rRNA gene (n = 71) for Plasmodium diagnosis, 48 (68%) used nested conventional PCR alone whereas 14 (20%) used real time quantitative PCR (qPCR), 5 (7%) used both nested and real time qPCR, 3 (4%) used semi-nested PCR and 1 (1%) used other methods to screen Plasmodium (Fig. 2A).

Fig. 2.

Type of target and approach used for PCR assays. The figure A depicts the type of PCR targets used to screen Plasmodium, that include 18S rRNA, mitochondrial, and others by various studies extracted for the current analyses. Studies that used 18S rRNA target are further classified according to the amplification approach reported (nested, semi-nested, single-step, real-time qPCR, and combination of nested and real time qPCR; shown as a breakup of 18S rRNA number and percentage). B Shows the usage of Snounou’s approach over other approaches among the studies that used nested conventional PCR either alone or with qPCR for Plasmodium detection. Studies using Snounou’s approach are further classified into the way PCR was conducted in terms of single reaction or multiplexing. The figures in parentheses represent the number followed by the percentage

Of the most widely used nested PCR approach, 85% (45/53) of the studies (published between 1993 and 2025) used the same oligonucleotide (primer) pair that was developed in 1993 by Snounou and colleagues [20].

Since the oligonucleotides developed by Snounou and colleagues are the most widely used, it is very important to use them as recommended by the authors. The oligonucleotide pairs were recommended to be used separately for genus and each species detection. Because the same research group [13], observed that if a species is present at a burden 10²–10⁴ times lesser than the other species in a sample, it could not be detected in a multiplex PCR secondary reaction, however both the species were detected when performed individual reactions for each species. Studies that used Snounou’s approach were examined for the use of this approach in this context and it was found that 7% (3/45) performed multiplexing (that is known to be associated with compromised limit of detection) and 31% (14/45) did not mention anything about how the reaction was performed (Fig. 2B).

This report also found targets other than 18S rRNA that have been used by various studies and have more copies per genome as compared to 18S rRNA, thus providing a better (lower) limit of detection (Table 1).

Table 1.

Details of molecular targets opted to screen P. falciparum & P. vivax through various methods

Targets	Species	Copy number	Limit of detection	Amplification method	Cross-reactivity	References
msa-2	Pf	1	Not estimated	Conventional, nested	Not mentioned	[44]
Pvr47	Pv	14	0.1–10 Parasites/μL (culture & field samples)	Conventional, single-step	Not with human, Pf, Pm, Po, Pk	[42]
			0.66 Copies/µL (plasmid)	qPCR (probe-based)	Not with human, Pf, Pm, Pb	[43]
			15.6 Copies/µL (plasmid) and 3 parasites/µL (pooled patient blood)	LAMP	With Pk	[40]
MLS110	Pv	34	0.1 Parasite/µL	qPCR (SYBR Green)	Not with human, Pf	[45]
stevor	Pf	30–40	Not estimated	Conventional, nested	Not mentioned	[44]
Pfr364	Pf	41	0.1–10 Parasites/μL (culture & field samples)	Conventional, single-step	Not with human, Pv, Pm, Po, Pk	[42]
			3.27 Copies/µL	qPCR (probe-based)	Not with human, Pv, Pm, Pb	[43]
			27.3 Copies/µL (plasmid) and 3.7 parasites/µL (culture)	LAMP	Not mentioned	[40]
MLS152	Pf	44	0.1 Parasite/µL	qPCR (SYBR Green)	Not with human, Pv	[45]
varATS	Pf	59	0.06–0.15 Parasites/μL	qPCR (probe-based)	Not with human, non-falciparum Plasmodium	[39]
coxIIIa	Pf, Pv	20–150	0.6–2 Parasites/μL	Conventional, single-step	Not mentioned	[24]
coxI	Pv	20–150	23.9 Copies/µL (plasmid) and 2.4 parasites/µL (pooled patient blood)	LAMP	With Pk	[40]
			Not estimated	qPCR (probe-based)	With Pf	[41]
cytb	Pf, Pv	20–150	Not estimated	qPCR (probe-based)	Pvcytb cross-reacts with human, Pf, Ps, Pb. Pfcytb cross-reacts with human, Pv, Ps, or Pm/Pb	[41]
TARE-2	Pf	∼250	0.03–0.12 Parasites/μL	qPCR (SYBR Green)	Not with human, non-falciparum Plasmodium	[39]
PfEMP1	Pf	Several (number not given)	28.9 Copies/µL (plasmid) and 3.3 parasites/µL (culture)	LAMP	Not mentioned	[40]

Here, the different molecular targets were listed (along with their estimated limit of detection, amplification method and observed cross-reactivity with other species) based on their copy number in the parasite genome

cox cytochrome oxidase (mitochondrial), cytb cytochrome b (mitochondrial), EMP erythrocyte membrane protein, LAMP loop-mediated isothermal amplification, MLS152 & MLS110: intra and intergenic multiloci short stretch of 152 and 110 nucleotides, msa: merozoite surface antigen, Pb, P. brasilianum, Pf, P. falciparum, Pk: P. knowlesi, Pm P. malariae, Po P. ovale, Ps P. simium, Pv P. vivax, Pvr47 & Pfr364: non-coding sequences (functions of both are not known. Neither their sequence is annotated nor they encode any protein), qPCR quantitative real-time PCR, stevor subtelomeric variable open reading frame, TARE-2 telomere-associated repetitive element 2, varATS var gene acidic terminal sequence

^aPCR was performed without DNA isolation directly from dried blood spots. For rest, isolated DNA was used as the starting material

Discussion

Despite substantial progress in reducing the malaria burden and in the development of molecular methods for the diagnosis of Plasmodium infection, the concerns regarding the robust and reproducible detection of LDIs, sub-microscopic infections, and treatment of asymptomatic carriers need to be addressed to achieve and sustain malaria elimination. Although the complete genome sequence of all the major human-infecting Plasmodium species is available, it was felt that nested PCR assays based on a particular locus in the 18S rRNA gene (4–5 copies per genome of Plasmodium) have been the most popular method used despite being laborious and time consuming [20]. To avoid the overdependence on one particular molecular locus and target for detection and to improve the diagnostic accuracy, new molecular targets (including loci other than that suggested by Snounou and colleagues) with comparative or better sensitivity, limit of detection and reproducibility to the currently used locus in the 18S rRNA gene, and that can be used in multiplexing PCR assays are soon needed. The current study was hence undertaken to provide an account of the differential usage of 18S rRNA as PCR target and explore the availability of newer targets other than 18S rRNA.

This research has documented that almost 93% of the research articles published between 1993 and 2024 that used PCR for malaria molecular diagnosis used the 18S rRNA gene as the molecular target. Such a high dependence (overreliance) on a solitary molecular target for diagnosis might be a cause of concern as there is some evidence that shows that oligonucleotide binding site mutations generate the possibility of a false-negative PCR [30]. Due to anticipated slower mutation rates in the 18S subunit of ribosomal RNA gene, it has been an ideal target to make phylogenetic inferences and also for molecular diagnosis of Plasmodium [31, 32]. However, mutations could spontaneously occur and accumulate in the Plasmodium 18S rRNA gene [33]. Particularly for P. ovale, it has been documented in Vietnam that some infections may be missed by PCR done on parasite 18S rRNA genes due to deletions and mutations in the target DNA sequences [34] which must be taken into consideration, if the oligonucleotide pair is based on a widely used common target. Similar observations have been documented from other regions [35–37].

In an analogous explanation of a similar phenomenon, the downsides of dependence on a single diagnostic target for malaria have been demonstrated by the parasite’s PfHRP2 deletions, a diagnostic marker for P. falciparum used in the malaria RDTs. This has been identified as a potential threat for malaria elimination as RDTs are fast becoming the alternative ‘gold standard’ for malaria diagnosis and subsequent treatment, particularly in areas where MS is not possible, feasible, or available due to various reasons. Therefore, in such areas, parasites with PfHRP2/3 gene deletions have a potential survival (selection) advantage over parasites with intact PfHRP2 and are therefore expected to outgrow their counterpart [28]. Although malaria treatment is not dependent on PCR-based diagnosis and hence a similar selective advantage might not be a possibility with parasites bearing mutations in primer-binding regions of the 18S rRNA, this may become a threat in the future in sustaining zero malaria status and during prevention of re-establishment of malaria after elimination [38]. Also, natural mutations might occur anywhere across the Plasmodium genome, molecular diagnostic targets being no exception, over-reliance on a single molecular target for diagnosis could be falsely assuring in ruling out low-density and submicroscopic infections as the countries march towards elimination and thereafter enter the prevention of re-establishment phase. Thus, the molecular diagnosis for Plasmodium species infections should not be based on solitary targets but rather be complemented with alternative targets when needed.

However, attempts have been made to investigate additional molecular targets (having a higher copy number than 18S rRNA gene) for the screening of Plasmodium species (detailed in Table 1) utilizing differential diagnostic methods, but their performance must be evaluated simultaneously on a comparable platform. All these targets are well-known for having multiple copies, but they are associated with a few limitations, like their applicability for only limited species detection, cross-reactivity with other species, and lack of complete validation. For these reasons, 18S rRNA may have been a better choice for researchers.

In 2015, a qPCR assay was developed using telomere-associated repetitive element 2 (TARE-2 with ∼250 copies/genome) and the var gene acidic terminal sequence (varATS with 59 copies/genome) for the detection of P. falciparum, which was reported to be 10 × more sensitive than 18S rRNA gene-based assays [39]. The only limitation was that they have not been evaluated for cross-reactivity with other Plasmodium species.

Other than these two, Plasmodium mitochondrial genome having genes cytochrome b, cytochrome oxidase I and III was also exploited as an ideal target over the years for the detection of P. falciparum & P. vivax as these genes are present up to 20–150 copy numbers per genome [24, 40, 41]. However, the biggest limitation with these targets was non-specific amplification or cross-reactivity with either human DNA or other Plasmodium species. Therefore, these targets need to be further evaluated for the potential cause of cross-reactivity before developing diagnostic assays.

Moreover, diagnostic assays for low-density infections were developed using two non-coding sub-telomeric repeat sequences called Pvr47 (for P. vivax) and Pfr364 (for P. falciparum), which are found in 14 and 41 copies per genome, respectively [40, 42, 43]. However, the above targets exhibited equivalent sensitivity with the 18S rRNA gene for detecting low-density infections on clinical samples, despite having a copy number far higher than the latter gene, in contrast to experimentally created mixed P. falciparum-vivax infections, where they performed better. This demonstrates that the performance of diagnostic assays varies when tested in controlled laboratory settings versus clinical samples. However, as a general practice, the limit of detection of any diagnostic assay is only reported using a purified DNA sample (obtained from an in vitro culture) or a highly purified plasmid containing a specified target sequence. The limit of detection of any assay must also be reported on clinical samples containing various inhibitors & other factors to determine its applicability. Therefore, multiple copy target candidates have been published over the years, but they have not been fully validated.

The study’s findings may be limited by the scope of the database used to screen the studies for inclusion, as this systematic review was not done with the primary objective to include all published studies that used PCR for malaria molecular diagnosis. The comparative analyses of published reports that used different molecular targets (other than 18S rRNA gene) to screen P. falciparum & P. vivax also suffer from inherent limitations of bibliometric data that include a lack of sufficient information and a standardised way to compare the published indicators of diagnostic accuracy and limits of detection. Despite these limitations, this systematic study has been able to capture the central theme of overreliance on a particular locus within the 18S RNA gene and a need to explore other molecular loci for malaria diagnosis.

Conclusion

It was observed that a common approach based on the 18S rRNA gene was preferred over others for many years. Since the available genomic data on this parasite has increased over time, which resulted in identification of new molecular targets and qPCR has been popularized than ever before, more research is needed to validate those targets in terms of limit of detection, robust reproducibility in different laboratory settings, reduced costs and a possibility of multiplexing. However, in the interim, newer and non-18S rRNA-based targets may be used in tandem or as a supplementary molecular diagnostic method where PCR done on 18S rRNA target is negative, ambiguous, or non-conclusive.

Author contributions

N.D. Data acquisition, compilation, analysis and interpretation; Drafting, editing, revising & finalizing the manuscript; Illustrating figures and tables M. R. K. Data acquisition & compilation; Editing and revising the manuscript critically P. S. Data acquisition & compilation; Editing and revising the manuscript critically N. Z. Data acquisition & compilation; Editing and revising the manuscript critically S. S. Data analysis & interpretation; Editing, revising and finalizing the manuscript V. P. Data analysis & interpretation; Drafting, editing, revising and finalizing the manuscript P. K. M. Conceptualization; Data analysis and interpretation; Critically drafting, editing, revising and finalizing the manuscript A. S. Conceptualization; Data analysis and interpretation; Critically drafting, editing, revising and finalizing the manuscript.

Funding

This research received no dedicated or external funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.