The spread of molecular markers of artemisinin partial resistance and diagnostic evasion in Eritrea: a retrospective molecular epidemiology study.

Background: Eritrea was the first African country to discontinue the use of histidine rich protein 2 (HRP2)-detecting rapid diagnostic tests (RDTs) for malaria diagnosis following reports of a high prevalence of pfhrp2/3-deleted Plasmodium falciparum parasites causing false-negative results in the country. Eritrea was also the first African country to report partial artemisinin resistance due to the P falciparum kelch13 (pfk13) Arg622Ile mutation. We aimed to characterise the spatial distribution of pfk13 mutants and their interactions with pfhrp2/3 deletions in Eritrea and to assess the role of the use of HRP2-detecting RDTs and antimalarial (artesunate-amodiaquine) therapy in the spread of the two variants.

Methods: We conducted a retrospective molecular epidemiological analysis of pfk13 mutations and pfhrp2/3 deletions in existing P falciparum-infected blood samples collected as part of previous pfhrp2/3 deletion and severe malaria studies. Samples were collected in March, 2016 and between September, 2018, and January, 2020, from symptomatic patients seeking care at 15 health centres in four administration zones (Semenawi Keyih Bahri, Gash Barka, Anseba, and Debub) in Eritrea. A fragment spanning the propeller region of pfk13 was amplified from samples and sequenced using Sanger sequencing or targeted amplicon sequencing to identify genetic mutations. Deletions of pfhrp2/3 genes in samples were determined using multiplex quantitative PCR. Parasite haplotypes and genetic relatedness of parasite haplotypes were determined previously using microsatellite marker typing. The primary objective was to determine the prevalence of pfk13 mutations at health centres and administrative zones. The secondary objective was to investigate whether pfk13 mutants and pfhrp2/3 deleted parasites converge.

Findings: We sequenced 50 samples collected in March, 2016 from the Semenawi Keyih Bahri zone and identified no pfk13 mutations. By contrast, in 587 samples included in this study that were collected from health centres in Gash Barka, Anseba, and Debub in 2018-20, we detected five different single non-synonymous mutations: Glu605Lys, Arg622Ile, Asn657Lys, Lys658Glu, and Ser679Leu. The most prevalent mutation was pfk13 Arg622Ile, which was detected in samples collected from all nine health centres where more than five samples were available across all three administration zones, with an overall prevalence of 11·9% (70 of 587 samples; range 5·9-28·0%). We identified 22 unique pfk13 Arg622Ile mutant haplotypes among 26 samples tested, of which 13 (59·1%) were genetically related, whereas the remaining nine (40·9%) were not. The prevalence of pfk13 Arg622Ile was significantly higher in parasites with a single pfhrp3 deletion (46 [18·0%] of 255 samples) than in parasites without pfhrp2/3 deletions (ten [6·2%] of 161 samples; odds ratio 3·89 [95% CI 1·59-7·61]; p=0·0006) and with dual pfhrp2/3-deleted parasites (13 [9·0%] of 145; 2·23 [1·13-4·68]; p=0·018).

Interpretation: The geographical spread of the pfk13 Arg622Ile mutation might have initially resulted from the clonal expansion and spread of pfhrp2/3 deletions under the test-and-treat policy using HRP2-detecting RDTs. Subsequently, selective pressure from artemisinin combination therapy could have further facilitated the spread of both pfk13 Arg622Ile and pfhrp2/3 deletions. Continuous monitoring of trends in pfk13 and pfhrp2/3 variants is needed to inform effective malaria control and elimination strategies in Eritrea and other African countries.

Funding: US Department of Defense Armed Forces Health Surveillance Division, Global Emerging Infections Surveillance Branch (AFHSD/GEIS), and Wellcome Trust.