Data-driven diagnostic reform in Ethiopia's transition to pLDH-based malaria rapid diagnostic tests amid rising pfhrp2/3 deletions: a multisite, facility-based, cross-sectional study.

BACKGROUND: Rising histidine-rich protein (HRP) 2 and 3 gene deletions undermine the reliability of HRP2-based rapid diagnostic tests (RDTs) that are widely used to detect Plasmodium falciparum. In such settings, lactate dehydrogenase (LDH)-based RDTs, targeting an alternative antigen, might provide a useful diagnostic option. We aimed to assess the performance of HRP2-based and LDH-based RDTs among febrile patients across six districts in Ethiopia representing diverse transmission intensities, varying pfhrp2/3 deletion frequencies, and differing levels of Plasmodium vivax co-endemicity. METHODS: This multisite, facility-based, cross-sectional study was based in health centres from six districts in Ethiopia: Abobo, Arba Minch Zuria, Fentale, Gondar Zuria, Metema, and Pawe. The study recruited patients aged at least 6 months who had a fever (axillary temperature ≥37·5°C) or a history of fever within the preceding 24 h. The diagnostic performance of SD Bioline HRP2/PvLDH (Bioline), BIOCREDIT PfLDH/PvLDH (BIOCREDIT Pf/Pv), and BIOCREDIT HRP2/PfLDH (BIOCREDIT Pf-dual) RDTs was assessed against microscopy and qPCR targeting 18S rRNA. Plasmodium antigens were quantified using multiplex Luminex, whereas pfhrp2 and pfhrp3 exon 2 deletions were detected using digital PCR. The primary outcome was diagnostic accuracy (sensitivity and specificity) of each RDT compared with microscopy and qPCR. RDT performance was analysed by species, parasite density, site, and pfhrp2/3 gene deletion status, ensuring evaluation across diverse transmission settings. FINDINGS: 1800 participants (median age 20 [IQR 12-30]; 56·4% male) were enrolled between Sept 2, 2021, and Feb 5, 2022. One dried blood spot sample was lost; qPCR confirmed infection in 1055 (58·6%) of 1799 participants, comprising 739 (70·1%) P falciparum, 166 (15·7%) P vivax, and 150 (14·2%) mixed infections. Pfhrp2/3 deletions varied, with double deletions reaching 52 (35·1%) of 148 infections in Gondar Zuria. Using microscopy as a reference, the solely HRP2-based Bioline RDT showed reduced P falciparum sensitivity (75·7% [95% CI 73·7-77·6]), particularly in deletion-affected areas. Excluding deletion cases significantly improved the sensitivity of HRP2-based Bioline RDT (p<0·001), giving 96·7% for wild type versus 7·8% in double deletions. The BIOCREDIT Pf-dual RDT showed a P falciparum sensitivity of 97·1% (95% CI 96·3-97·9), and BIOCREDIT Pf/Pv RDT showed a P falciparum sensitivity of 91·2% (95% CI 90·0-92·5) across all settings. For P vivax, the BIOCREDIT Pf/Pv showed improved sensitivity of 95·7% (95% CI 94·8-96·7), outperforming the Bioline test, which reached 89·9% (95% CI 88·5-91·3). The BIOCREDIT Pf-dual had the lowest detection threshold (90% probability) for P falciparum at 97 parasites per μL, outperforming BIOCREDIT Pf/Pv (133 parasites per μL) and Bioline (2218 parasites per μL). INTERPRETATION: Our findings highlight the high and regionally variable prevalence of pfhrp2/3 gene deletions, which compromise HRP2-based test performance. By integrating diagnostic, genetic, and antigen data, our findings revealed key diagnostic gaps and informed Ethiopia's national policy shift from HRP2-based RDTs to non-HRP2-based (LDH-based) RDTs and offer a model for other countries facing similar diagnostic challenges owing to pfhrp2/3 deletions. FUNDING: The Gates Foundation and Wellcome Trust.