Introduction

Malaria remains a major public health burden in tropical countries, specifically in Indonesia. The Indonesian government targets national malaria elimination in 2030, and in 2017, 72% of the population lived in malaria-free zones. This achievement is possible due to the implementation of artemisinin combination therapy after the discovery of chloroquine resistance1,2,3, malaria screening in high transmission areas, development of rapid response protocol to suppress transmission foci and implementation of evidence-based policy across all sectors of the government4. All of these efforts lead to a reduction of the national annual parasite index (API) towards 0.9 cases per 1000 population5. However, malaria cases have plateaued since 20144. In the eastern part of Indonesia, particularly Papua, contributes to nearly 92% of the national malaria cases, with heterogeneous distribution of malaria transmission across different regions, requiring tailored malaria elimination effort6. One of the districts in Papua, Mimika, had increasing malaria cases annually with the second-highest API reaching 440.65 cases per 1000 individuals at risk as reported in the National Malaria Surveillance System (www.sismal.malaria.id). To intensify malaria elimination efforts in Mimika, a malaria diagnostic strategy needs to be prioritized in the high endemicity setting, such as increasing the deployment of RDTs to complement microscopy.

Microscopy and RDTs have been widely used as malaria diagnostic tools in Indonesia, following the WHO guidelines7. Initially, microscopy was the gold standard for detecting malaria parasites and their species. However, malaria microscopy has highlighted challenges: it is time-consuming and requires skilled expertise to identify Plasmodium parasites and their species4. Hence, RDTs have been utilized across Indonesia as an alternative diagnostic tool since 20108. The ease-of-use and simple result interpretations further improved feasibility and accessibility of RDT, unlike microscopy. Previous studies assessing RDTs performance to detect P. falciparum in Papua has shown a similar sensitivity and specificity compared to the microscopy and qPCR9,10, indicating that RDTs can be reliably used as an alternative diagnostic tool. Since malaria incidence in Papua has been rising since 2015 up to 2020, with P. falciparum becoming dominant in 2019 cases6, development of novel RDTs is urgent to detect any missed P. falciparum infections due to the absence of certain antigens used in previously available RDT.

Over the past decades, RDTs have been continuously developed and evaluated to improve their specificity and sensitivity. The main antigen targets in RDTs are Plasmodium lactate dehydrogenase (pLDH), which captures pLDH produced by the four common human Plasmodium and P. falciparum histidine-rich protein 2 (PfHRP2) which specifically detect P. falciparum parasite11. However, both types of RDTs have limitations. pLDH-detecting RDTs is relatively low sensitive12,13 and heat-unstable14. Moreover, PfHRP2-based RDTs can detect the presence of PfHRP2 in patient blood after treatment for several weeks15,16,17 which can potentially elevate false positives. A homologous protein of PfHRP2, called PfHRP3, is also circulating in P. falciparum-infected patients and share similar epitopes that can be detected by PfHRP2-based RDTs. In recent times, many combo RDTs combined antibodies against pLDH and PfHRP2 in a single cassette to distinguish between P. falciparum and non-falciparum infection.

Despite their utility and improvement, increasing reports of pfhrp2/3 gene deletions threaten the use of PfHRP2-based RDTs. A high prevalence of pfhrp2 gene deletion has been reported in many countries in Latin America and Horn of Africa, such as Peru18, Eritrea19,20, and Djibouti21 causing high levels of false negative RDT results. Several countries in Asia have also reported circulation of these gene deletions, which include India, Cambodia and Myanmar measured at up to 25%, 2.4% and 9.4%, respectively22,23,24. In response to the threat, WHO has recommended that countries assess the prevalence of pfhrp2/3 deleted parasites causing false negative RDT results, and switch away from HRP2-based RDTs when the prevalence reaches a threshold ≥ 5%11. For example, countries with high prevalence of pfhrp2/3 gene deletions have been suggested to use alternative non-HRP2-based RDT, such as pLDH or aldolase-based18,19,21.

In Indonesia, there is a paucity of information regarding the surveillance of these mutant parasites. Only one study conducted by the MalariaGEN Pf project included Indonesian isolates and found 4% of pfhrp2 gene deletion25. While the study demonstrated the presence of gene-deleted parasites in Indonesia, their impact on RDT results was not examined. The relatively small sample size does not allow accurate assessment of the prevalence of gene deleted parasites and does not meet the minimal sample size requirement indicated by WHO26. In addition, those samples were collected seven years ago and did not reflect the situations at the time of the study. In 2021, the Ministry of Health of The Republic of Indonesia has included pfhrp2/3 surveillance as one of the priorities in malaria research to accelerate the elimination, especially in high-endemic settings4. Therefore, we aimed to contribute to this surveillance to determine the updated proportion of pfhrp2/3 gene deletions circulating in Timika, Mimika District, Papua Province alongside their genetic diversity.

Methods

Study design

Detection of pfhrp2/3 gene deletion was part of a cross-sectional study conducted over four months to assess RDTs performance in collaboration with the Foundation for Innovative New Diagnostics (FIND), Geneva. The study was conducted in Timika, Papua where three health facilities with the highest malaria incidences were selected as the enrollment sites. Patients seeking treatment for suspected-malaria symptoms were asked for their consent prior to the blood collection. The sample size was calculated from a larger study (not published) in collaboration with FIND, based on the requirement of the “Technical Specifications Series for submission to WHO Prequalification – Diagnostic Assessment TSS-3” document27. In 2021, the estimated malaria prevalence in Timika was 25%, therefore the sample size was targeted to include a minimum of 2180 participants.

Patient enrollment and sample collection

Any patient aged 5 years or older who presented at a study-selected health facility with a fever (infrared forehead temperature of ≥ 37.5 °C) or a history of fever within the last 48 h was eligible to participate in this study. Patients who showed symptoms of severe malaria and/or central nervous system illness were excluded. For biological sample collection, a 3 mL venous blood was drawn from each participant via syringe for malaria RDTs testing, sub-microscopic speciation by real-time PCR based on Mangold et al.28, and pfhrp2/3 genotyping. Additionally, finger prick blood samples were collected for microscopy examination.

Identification of malaria cases

A 5 μl aliquot of venous blood was used to detect malaria infection by RDT. In total, three commercial RDTs were evaluated: two PfHRP2-based RDTs (SD Bioline™ Malaria Ag Pf, #05FK50 and SD Bioline™ Malaria Ag Pf/Pv, #05FK80) and one PfHRP2/Pf pLDH/Pv pLDH-based RDT combining detection of P. falciparum and P. vivax infection in a single cassette test (SD Bioline™ Malaria Ag Pf/Pf/Pv, #05FK120). All RDTs were provided by FIND, Geneva, Switzerland and manufactured by Abbott Diagnostics Korea Inc.

As the national gold standard, blood smear was prepared using Giemsa staining and examined by a certified microscopist at the study site who was blinded to the RDT results. Also, a second reading was done by an expert microscopist who was blinded to the first reader and RDT results. Parasite identification and parasitemia were determined by microscopy using the Obare Method Calculator, developed by the Worldwide Antimalarial Resistance Network (WWARN), to assess the result agreement between readers29. Discrepancies between the two readers were resolved by a third reader to obtain the final species and parasitemia. For simplicity, parasitemia was classified as low parasitemia (< 100 parasites/µL), medium parasitemia (100–1000 parasites/µL), and high parasitemia (> 1000 parasites/µL).

Molecular approaches were performed to validate the microscopy result. The genomic DNA (gDNA) was extracted from 200μL of whole blood by QIAGEN QIAamp DNA Blood Mini Kit and eluted at 150 μl. Identification of the malaria cases was determined by real-time PCR using BRYT Green® Dye with cycling conditions as published previously28. Malaria-negative sample was defined by a cycling threshold (Ct) above 33. Meanwhile, the Plasmodium species from positive samples were determined by comparing their melt temperature against the positive controls. Any disagreement results between microscopic and real-time PCR were re-evaluated by conventional PCR30.

pfhrp2/3 genotyping and sequencing

Only mono-infection cases of P. falciparum confirmed by both microscopic and molecular tests were selected to examine the pfhrp2/3 gene deletion. A confirmatory PCR step by single copy genes, including glutamate-rich protein (glurp) and erythrocyte binding antigen 175 region IV (eba175-RIV), was performed to assess the DNA quality of the samples. A nested PCR of glurp gene was conducted as instructed by Brockman et al.31, whereas single-step PCR was done for eba175-RIV gene. The primer set of eba175-RIV was designed in-house with 5ʹ-CAT GAA GAG GTG AAA GAA CAT AC-3ʹ as forward sequence (EBA175-R4-F) and 5ʹ-CTT CAT CCA AGC TAC TAG AAG G-3ʹ as reverse sequence (EBA175-R4-R). A total volume of 25 μl for eba175-RIV PCR reaction was mixed, consisting of 1X PCR Buffer, 2.5 mM MgCl2, 0.1 mM dNTPs Mix, 2.4 μM Primer set, and 0.75 U DNA Taq Polymerase. The cycling condition of eba175-RIV was performed as follows: Pre-Denaturation at 94 °C for 10 min; Denaturation/Annealing/Extension at 94 °C/55 °C/70 °C for 50 secs/50 secs/1 min (30 cycles); and Holding step at 15 °C.

Successful amplification of the samples for both single-copy genes were further analyzed for pfhrp2/3 genotyping. For this study, exon 2 region of pfhrp2/3 genes was investigated by conventional PCR as described elsewhere32, with modification on performing only the second-nest PCR. The prevalence of pfhrp2/3 gene deletion was calculated and categorized into 4 outcomes, including single deletion of pfhrp2 (pfhrp2-/pfhrp3 +) and pfhrp3 (pfhrp2 + /pfhrp3-), double deletion (pfhrp2-/pfhrp3-), and no deletion (pfhrp2 + /pfhrp3 +).

Sanger sequencing was conducted on pfhrp2 + or pfhrp3 + samples with medium/high parasitemia, but negative by PfHRP2-RDTs. The hrp2/3 sequence from these samples were then compared to PfHRP2-RDTs positive samples and samples from other studies in order to investigate possible novel repeat sequences that might cause the false negative results. Samples were defined as PfHRP2-RDTs negative if all three PfHRP2-RDTs tested negative. Sequences from the pfhrp2/3 positive but PfHRP2-RDTs negative samples were aligned against pfhrp2/3 + and PfHRP2-RDT positive samples by the MUSCLE 5.1 alignment software and their repeat-type variations assessed according to previous publications32,33.

Results

Demography and baseline data

Between December 2022 and April 2023, a total of 2184 individuals were enrolled. Patient numbers varied across the three health facilities, with the highest number of samples collected from Wania (n = 889) followed by Timika (n = 804) and Timika Jaya (n = 491). The majority of P. falciparum mono-infection samples (78.69%) were derived from the adult participant group (19–59 years old; Table 1). The median age amongst participants was 29 years with 57% males and 43% females. Out of 582 P. falciparum mono infection cases, 389 (66.83%) samples had high parasitemia by microscopy (Table 1), while the median parasitemia was 2836 parasites/µL (IQR at 450.5–22,500 parasites/µL).

Table 1 Data characteristics across 582 mono-infection P. falciparum samples during the 4-month enrollment period alongside their parasite density level.
Full size table

P. falciparum mono-infection cases

A total of 2157 blood samples were collected. A total of 615 participants (28.51%) had P. falciparum mono-infection detected by microscopy. Prior to pfhrp2/3 genotyping, a real-time PCR was performed to confirm the Plasmodium species. The final number of P. falciparum mono-infections was 582 samples (26.98%) in agreement between microscopy and real-time PCR results (Fig. 1).

Fig. 1
figure 1

A flow diagram showing the number of samples collected and included in pfhrp2/3 genotyping. Amplification of single copy genes was conducted after species identification for quality control, reducing the sample size to be pfhrp2/3 genotyped at 566 samples. The bottom level of the flowchart outlines the prevalence of pfhrp2/3 gene deletion.

Full size image

pfhrp2/3 genotyping

Amplification of single-copy genes (i.e. glurp and eba175-RIV) was performed in 582 samples. Sixteen samples (2.75%) that failed to amplify either one or both single-copy genes, which indicate degraded DNA quality. Thus, a total of 566 samples (97.25%) were selected for amplification of exon 2 regions of pfhrp2/3 genes (Figs. 1 and 2).

Fig. 2
figure 2

The arrangement of repeat-type variations of pfhrp2/3 sequences found in all samples; (a) pfhrp2 sequence and (b) pfhrp3 sequence. Sample ID 2016 and 2014 were tested positive by PfHRP2-based RDTs (showed in italic font) whereas others were negative. The color-coded number illustrates the repeat sequence as described in Table S1, according to previous publications32,33. 3D7 acted as reference sequence in both pfhrp2 (NCBI Accession Number: NC_004329) and pfhrp3 (NCBI Accession Number: XM_002809080). Some repeat-types highlighted in yellow were a unique repeat that were not reported yet in other publications (see Table S2). However, the unique type 6 repeat found in sample 2016, 3063, and 2377 were reported previously34,35 (see Table S2).

Full size image

The prevalence of the pfhrp2/3 gene deletion identified in this study varied. Overall, pfhrp2 deletion was observed at 2.12% (95% CI 1.1–3.67%), whereas pfhrp3 deleted samples were 37.1% (95% CI 33.11–41.23%). The distribution of pfhrp2/3 gene deletion was also illustrated (using ggplot2 3.5.1) according to their parasite density and PfHRP2-RDTs results in Fig. 3. Only three samples were found to have single pfhrp2 gene deletion (0.53%, 95% CI 0.11–1.54%), with one PfHRP2-RDTs positive. In contrast, a high prevalence of pfhrp2 + /pfhrp3- was observed in 201 samples (35.51%, 95% CI 31.57–39.61%). Among single pfhrp3 gene deletion cases, five samples were PfHRP2-RDTs negative, with four having low parasitemia (Fig. 3). Lastly, a low proportion of pfhrp2-/pfhrp3- were detected in nine samples (1.59%, 95% CI 0.73–3%), with eight having positive PfHRP2-RDTs results (Fig. 3).

Fig. 3
figure 3

Parasite density of various pfhrp2/3 gene deletion status (n = 566) with blue and red representing positive and negative PfHRP2-RDTs status, respectively. Thresholds for low, medium, and high parasitemia are < 100 parasites/µL, 100–1000 parasites/µL, and > 1000 parasites/µL, respectively.

Full size image

We also investigated the parasite density of samples in regards to pfhrp2/3 deletion status and PfHRP2-RDTs result. On samples with non-deleted pfhrp2/3, the median [IQR] parasitemia was 80 [52–102] and 4784 [1144–26,875] parasites/μL with PfHRP2-RDTs negative and positive results, respectively. While on samples with deleted pfhrp2/3, the median [IQR] parasitemia was 32 [28–76] and 2808 [680–27,500] parasites/μL with PfHRP2-RDTs negative and positive results, respectively.

pfhrp2/3 repeat-sequence variations

We further investigated high-parasitemia samples with no pfhrp2/3 deletion but PfHRP2-RDTs negative by Sanger sequencing to confirm the presence and structure of these genes. We sequenced seven pfhrp2 positive samples, five of which were PfHRP2-RDTs negative, and four pfhrp3 positive samples, three of which were PfHRP2-RDTs negative.

The arrangement of repeats in pfhrp2/3 sequences for each sample is illustrated in Fig. 2. The length of the exon 2 region in pfhrp2 sequences ranged from 689 to 824 base pairs, whereas the length of the exon 2 region in pfhrp3 sequences ranged from 471 to 501 base pairs. The repeat-type sequences in pfhrp2 varied across the samples (Fig. 2a), yet similar arrangements were found in pfhrp3 (Fig. 2b). The majority of the samples in pfhrp2 and pfhrp3 shared the same repeat-type code as reported elsewhere32,33 (Table S1). However, unique repeat-type variants were also observed in some samples (highlighted in yellow in Fig. 2). Some of these unique types, including repeat types 6 and 7 (Table S2), have been observed in previous publications34,35.

Discussion

As a response to the emergence of pfhrp2/3 gene deletion reports, WHO has announced a standardized protocol for detecting the prevalence of pfhrp2 deletion. A threshold of 5% of false negative PfHRP2-based RDT results caused by gene deletions has been recommended as a trigger to change the malaria diagnostic strategy away from PfHRP2-based RDT11,26. For instance, a high proportion of pfhrp2 deletion in Peru, Eritrea and Djibouti have led these countries to switch to non-PfHRP2-based RDTs18,19,20,21. This highlights the importance of identifying the prevalence of pfhrp2/3 gene deletion to inform country malaria diagnosis policy.

In our study, we identified 615 P. falciparum mono-infection cases by microscopy. Further confirmation by PCR was performed to verify the species and ensure the DNA quality within the samples (i.e. any fragmented or degraded gDNA might negatively affect the pfhrp2/3 deletion results)36,37. After excluding samples that failed to amplify two single-copy genes, a total of 566 samples were selected for pfhrp2/3 genotyping.

We identified different proportions of pfhrp2/3 gene deletion across the population. In this study, pfhrp2 deletions were found at 2.12% (95% CI 1.1–3.67%), while a previous report from the MalariaGEN Pf project reported a deletion of 3.75% (95% CI 0.78–10.57%)38. This might be due to the higher sample size that we assessed, different sampling methods, or laboratory assays used to determine the pfhrp2 gene deletion39. Three of these pfhrp2 deleted samples tested negative in all three RDTs, with one of them having a parasitemia above 100 parasites/µL, indicating that the pfhrp2 gene deletion might be responsible for this false negative result21,40. Alternatively, this might be due to the lower sensitivity of the RDTs in detecting low level of parasitemia. In contrast, a high proportion of pfhrp3 deletions were observed in 210 out of 566 samples (37.1%, 95% CI 33.11–41.23%), higher than the 25% (95% CI 15.99–35.94%) pfhrp3 deletion prevalence reported by the MalariaGEN study in Indonesian isolates38. Similarly, this pattern, where the pfhrp3 deletion is higher than pfhrp2 deletion, was detected in several countries, including Northeast Tanzania and Peru18,41. Of the single pfhrp3- samples, only 5/201 (2.48%) samples were negative by RDTs. These samples had low parasitemia (< 100 parasites/µL by microscopy); hence, these false negative results are caused by low parasite densities42,43. However, this pattern of higher hrp3 deletion prevalence compared to hrp2 in different countries warrants a further investigation into the impact of hrp3 deletion towards the malaria elimination effort. Finally, pfhrp2/3 double deletion was detected at 9/566 (1.59%, 95% CI 0.72–3.00%) samples, compared to no identified double deletion in the MalariaGEN Pf study38. Nevertheless, 8/9 (88%) double-deleted samples tested positive by RDTs, suggesting that either a recent infection with pfhrp2 + parasite or multiple infections might explain this incidence44. However, we found that all of these patients have not had recent infections diagnosed over the past year according to the clinical records, indicating that this hypothesis is less likely, although they may have had asymptomatic infections in the recent past. Quantifying PfHRP2 in these samples can inform on the possibility of unreported P. falciparum infections that left traces of the antigen, and it can enable the assessment of the RDTs performance by comparing the antigen levels to samples with concordant results. Future studies can probe further into the relatedness of these parasites using identity-by-descent (IBD) methods, which might reveal P. falciparum populations that carry certain hrp2/3 deletion characteristics. However, this requires selection of appropriate markers to avoid introducing bias to the IBD estimates45.

To investigate possible causes of false negative RDT results for several samples with medium/high parasite densities, we assessed the repeat-type variations across the pfhrp2/3 exon 2 region with medium/high parasitemia samples (> 150 parasites/µL) that were undetected by PfHRP2-based RDTs. Of the five pfhrp2/3 samples sequenced, only one sample had a parasitemia below 250 parasites/µL. pfhrp2 exhibited greater sequence diversity compared to pfhrp3, consistent with global isolates44. We observed that all the samples contained type 2 and 7 repeats in their sequences. These two repeats have been reported to play a key role in RDT positivity32. However, the results show clearly that the cause of false negative PfHRP2-based RDTs results in these samples were not due to pfhrp2 variations. Some studies have shown that the diversity of pfhrp2/3 may not contribute to false negative results44,46. Additional genotyping of the hrp2/3 flanking genes might add more information in the characterization of the deletions found in this study. Another possibility that can cause false negative RDT results might be explained by poor quality or degradation of RDTs due to poor transport and storage conditions and anomalies (such as red background). However, we did not find such discrepancies within these samples. There might be a possibility that non-deleted and deleted parasites co-exist within individuals, where a lower number of non-deleted parasites were detected by PCR since it has a higher sensitivity compared to RDT. Using more sensitive deletion detection methods such as digital droplet PCR47,48 can confirm the pfhrp2/3 deletion status on suspected multiple clone cases.

We also investigated the relationship between parasitemia, pfhrp2/3 deletion, and PfHRP2-RDTs results by comparing the median parasitemia between the different pfhrp2/3 deletion and PfHRP2-RDTs result combination. We found that the median parasitemia for negative PfHRP2-RDTs results was 32 and 80 parasites/µL for non-deleted and deleted pfhrp2/3 samples, respectively. This might suggest that negative PfHRP2-RDT results are more likely to be caused by lower sensitivity of RDTs to low parasitemia samples in this study. When looking at samples with positive PfHRP2-RDT results, the median parasitemia was 2808 and 4784 parasites/µL for non-deleted and deleted pfhrp2/3 samples, respectively. We suspect that pfhrp2/3 deleted samples with higher parasitemia accommodate for the loss of either pfhrp2 or pfhrp3 by having a high enough antigen level to be detected by PfHRP2-RDTs. Future studies might assess the association between PfHRP2-RDT results, parasitemia, and pfhrp2/3 deletion by sampling from areas with a high pfhrp2/3 deletion prevalence setting, enabling high-precision estimates on the impact of pfhrp2/3 deletion.

During this study, some limitations were encountered. First, the PfHRP2-based RDTs that we used were not the RDTs used in the National Malaria Control Program. In addition, this study was originally designed to evaluate RDTs performance, where sampling was conducted in three sites instead of the minimum of ten health facilities according to the WHO guideline for hrp2/3 deletion. Future studies should also increase the range of sample sites to better represent the deletion prevalence in Papua province. Lastly, we did not perform any immunoassays to detect and quantify the PfHRP2/3 antigen circulating within the samples, unlike other studies22,39,49. This data may have been beneficial in selecting the samples for pfhrp2/3 according to the presence of PfHRP2/3 antigens, as applied by some studies39,49,50.

Conclusion

This is the first pfhrp2/3 genotyping study to assess PfHRP2-based RDT performance in Indonesia, providing the current pfhrp2/3 deletion prevalence in a high transmission setting, Timika, Papua. A low proportion of pfhrp2 deletion was observed in this symptomatic population. However, some of the RDT readings were not affected by the deletion of pfhrp2 exon 2, leading to cases of pfhrp2 cryptic absence. This indicates that the PfHRP2-based RDTs tested in this study remain suitable since the false negative RDT rate due to pfhrp2/3 deletion is below the WHO threshold for changing the diagnostic strategy (≥ 5%). Overall, the prevalence of single and double pfhrp2/3 deletions in this study highlights the importance and urgency to conduct further surveillance following the WHO survey protocol in order to cover hrp2/3 deletion prevalence in other regions in Indonesia, especially the Papua province. This also allows monitoring of the performance of PfHRP2-based RDTs used in Indonesia, providing further information on whether the current national RDT used remain reliable. These outcomes will inform the NMCP on hrp2/3 deletion prevalence in Indonesia, and guide the continuing development of the national malaria diagnostic policy.