Introduction
This intervention is delivered to all children aged 3–59 months in the community, regardless of malaria status, as a single dose of sulfadoxine–pyrimethamine and three daily doses of amodiaquine, monthly for up to 5 months during the short transmission season. Meta-analysis of clinical trial data provides an estimated mean decrease in clinical malaria episodes per child per year of 75% with seasonal malaria chemoprevention compared with placebo, and a modest beneficial effect on the prevalence of anaemia.
According to WHO, 13 countries in the African Sahel had active seasonal malaria chemoprevention programmes in 2021.
Evidence before this study
We sought detailed molecular studies of resistance to sulfadoxine–pyrimethamine and amodiaquine in Plasmodium falciparum in populations implementing seasonal malaria chemoprevention. We searched PubMed using the following text: (seasonal malaria chemoprevention) AND (Plasmodium falciparum) AND (molecular markers) AND (children) AND (drug resistance). No date or language restrictions were applied. The search yielded eight studies, of which five were conducted in one of the seven countries that participated in the Achieving Catalytic Expansion of Seasonal Malaria Chemoprevention in the Sahel project, which sought to remove barriers to the scale-up of seasonal malaria chemoprevention in seven countries in 2015 and 2016. These studies included genotype analyses of P falciparum isolates from 201 children with fever with a positive rapid diagnostic test in Niger; 394 men and four women with fever in Chad; and 1164 children younger than 5 years sampled cross-sectionally in the community over 3 years of seasonal malaria chemoprevention implementation in Mali. One study conducted k13 and mdr1 genotyping among 27 children PCR positive for P falciparum DNA receiving seasonal malaria chemoprevention in Burkina Faso. None of these results were available before the current study. Our earlier work assessing dhfr and dhps genotypes in 1000 PCR-positive samples from pregnant women and children with uncomplicated malaria in Nigeria was available and informed our study design. There were no population-level studies available that permitted systematic comparison of parasite genotypes under selective pressure from programmatic implementation of seasonal malaria chemoprevention in the Sahel.
Added value of this study
This study provides a comprehensive, high-throughput assessment of P falciparum genotype variation at all four parasite genes known to contribute to resistance to the seasonal malaria chemoprevention drugs (amodiaquine and sulfadoxine–pyrimethamine) across seven countries implementing seasonal malaria chemoprevention in the African Sahel at the outset of scale-up in 2015–16. Both the target age group of children younger than 5 years and older residents who would not have received the study drugs were sampled. This assessment was repeated 2 years later in 2017–18, using identical sampling and genotyping methodologies, permitting direct comparison between the two sample periods.
Implications of all the available evidence
This study, and previous smaller studies in the region, provide substantial evidence that seasonal malaria chemoprevention with sulfadoxine–pyrimethamine and amodiaquine is not currently under a serious threat from drug-resistant parasites in these implementation areas. However, the data indicate that continuing surveillance is needed to guard against future emergence of resistance to an extent that would threaten the effectiveness of seasonal malaria chemoprevention. Our data provide a comprehensive baseline in seven locations across the regions where seasonal malaria chemoprevention is being deployed at scale. The surveys could be repeated, using the same sampling and laboratory methods, to monitor the effect of seasonal malaria chemoprevention at scale on the frequencies of markers of resistance and to provide early warning of loss of effectiveness.
Variant haplotypes of crt at codons 72–76, encoding 72Cys-73Val-74Ile-75Glu-76Thr (CVIET) and 72Ser-73Val-74Met-75Asn-76Thr (SVMNT), and of mdr1 (encoding Tyr at codons 86, 184, and 1246 [YYY]) are associated with amodiaquine resistance in therapeutic studies.
A variety of point mutations in P falciparum dihydropteroate synthase (dhps) confer resistance to sulfadoxine and point mutations in dihydrofolate reductase (dhfr) confer resistance to pyrimethamine. In Africa, the combined haplotype GE-IRN, comprising mutations in both dhps (encoding 437Gly and 540Glu [GE]) and dhfr (51Ile, 59Arg, and 108Asn [IRN]), is known to be strongly associated with sulfadoxine–pyrimethamine resistance.
To date, this GE-IRN variant haplotype has been very rare in west Africa compared to east and southern Africa,
,
,
,
,
but evidence acquired since 2009 suggests that other haplotypes of dhps, in particular 431Val-436Ala-437Gly-540Lys-581Gly-613Ala (VAGKGA) and 431Val-436Ala-437Gly-540Lys-581Gly-613Ser (VAGKGS), are emerging in Nigeria and Cameroon.
,
Methods
Study sites
Seasonal malaria chemoprevention was implemented in each of the study sites at the time of the baseline survey, with the exception of Upper River Region of The Gambia, which had 1 year of seasonal malaria chemoprevention implementation before the 2016 survey. All seven countries deploy artemether–lumefantrine as the first-line therapeutic antimalarial drug.
Figure 1Proportion of participants with Plasmodium falciparum parasites detected by qPCR across seven Sahelian countries implementing seasonal malaria chemoprevention with sulfadoxine–pyrimethamine and amodiaquine (2016 and 2018)
Parasite prevalence estimates are shown for children younger than 5 years (A) and individuals aged 10–30 years (B), by country, from the 2016 and 2018 surveys. Error bars indicate the upper limit of the 95% CI for the proportions estimated from the binomial distribution.
Survey design
Survey methods
Blood sample preparation
Finger-prick blood from each participant was applied to filter paper (Whatman 3MM; ThermoFisher Scientific, Waltham, MA, USA) with a barcode attached. Each barcode was linked to a participant identification number, and the linkage list retained by field teams. Dried blood spot samples were attached to individual cardboard covers, assembled in batches of 50–100, and stored in a plastic bag containing silica gel. All dried blood spot samples were subsequently transported to the LSHTM laboratory.
Laboratory methods
Extracted DNA was stored at –20°C until use.
was used to simultaneously detect P falciparum parasites and genotype the P falciparum crt locus. Three dual-labelled probes designed to detect three crt genotypes at codons 72–76 (encoding Cys-Val-Met-Asn-Lys [CVMNK], CVIET, and SVMNT) were combined with a fourth dual-labelled probe (cy5 reporter) to detect an extraction control target, the human β-tubulin gene.
Laboratory isolates 3D7, Dd2, and 7G8 were positive controls for the CVMNK, CVIET, and SVMNT haplotypes, respectively. qPCR amplification was done as a qualitative assay in a single well to maximise throughput in the 72-well rotor of a Rotorgene Q thermal cycler (Qiagen, Hilden, Germany).
,
,
Polymorphisms were identified by direct sequencing of amplified products (BigDye Terminator v3.1 cycle sequencing kits and ABI 3730 sequencer; ThermoFisher Scientific) and data analysed using Geneious v10.1.3 (Biomatters, San Diego, CA, USA).
Statistical analysis
In each country and age group, we estimated the prevalence of dhfr mutations (the individual mutations 51Ile, 59Arg, and 108Asn, and the combined dhfr triple-mutant haplotype IRN); dhps mutations (431Val, 436Ala, 437Gly, 540Glu, 581Gly, and 613Ser, and combined dhps haplotypes, including VAGKGA and VAGKGS; the two-locus dhfr and dhps haplotype GE-IRN; crt mutations (74Ile, 75Glu, and 76Thr, and the CVIET haplotype); mdr1 mutations (86Tyr and 184Tyr, and the YY haplotype); and two-locus haplotype YY-CVIET, comprising mutations in crt (CVIET) and mdr1 (YY); and the combined haplotype YY-CVIET-GE-IRN. We defined a genotype as one or more mutations in a single codon associated with resistance, and a haplotype as a combination of at least one mutation in each of two or more codons of interest in one or more genes.
Role of the funding source
The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Results
In 2016, 5130 (17·5%) samples were qPCR-positive for P falciparum DNA, with infection detected in 2844 of 14 345 samples (19·8% [95% CI 19·2–20·5]) from children younger than 5 years and 2286 of 14 929 samples (15·3% [14·7–15·9]) from people aged 10–30 years (appendix p 3). In 2018, 2176 (7·6%) samples were positive, with infection detected in 801 of 14 019 (5·7% [5·3–6·1]) samples from children younger than 5 years and 1375 of 14 527 (9·5% [9·0–10·0]) samples from people aged 10–30 years (appendix p 3).
Table 1Prevalence of Plasmodium falciparum DNA in 54 907 dried blood spot samples collected across a west-to-east transect of seven countries in 2016 and 2018, in children younger than 5 years and in residents of the same compounds aged 10–30 years
Data are n/N (%), representing the number of qPCR positive tests out of the total number of samples for each age group and country.
,
,
CVIET showed varying prevalence across the seven countries, fluctuating slightly or decreasing among children younger than 5 years between 2016 and 2018 in three countries, but markedly increasing in prevalence in Burkina Faso, Niger, Nigeria, and Chad (table 2; appendix p 4). Only in Burkina Faso was a marked increase from 2016 to 2018 also observed in the older age group (table 2). These patterns were reflected in the seven-country combined unadjusted prevalence ratios (table 3).
Table 2Prevalence of genetic markers of sulfadoxine–pyrimethamine and amodiaquine resistance among samples PCR-positive for P falciparum DNA in 2016 and 2018 by country
Prevalence estimates incorporate survey weights as described in the Methods. crt CVIET=crt 72Cys-73Val-74Ile-75Glu-76Thr. YY=mdr1 86Tyr-184Tyr. YY-CVIET=mdr1 86Tyr-184Tyr and crt CVIET. IRN=dhfr 51Ile-59Arg-108Asn. GE=dhps 437Gly-540Glu. GE-IRN=dhfr 51Ile-59Arg-108Asn and dhps 437Gly-540Glu. VAGKGS=dhps 431Val-436Ala-437Gly-540Lys-581Gly-613Ser.
Table 3Overview of change in prevalence of selected markers and haplotypes, for all seven countries combined, from 2016 to 2018
Values in parentheses are 95% CIs. These analyses are not survey-weighted or adjusted for clustering. crt CVIET=crt 72Cys-73Val-74Ile-75Glu-76Thr. YY=mdr1 86Tyr-184Tyr. YY-CVIET=mdr1 86Tyr-184Tyr and crt CVIET. IRN=dhfr 51Ile-59Arg-108Asn. GE=dhps 437Gly-540Glu. GE-IRN=dhfr 51Ile-59Arg-108Asn and dhps 437Gly-540Glu. VAGKGS=dhps 431Val-436Ala-437Gly-540Lys-581Gly-613Ser.
fluctuated between 0% (zero of 53 isolates, adjusted) and 47·7% (64 of 134 isolates, adjusted) prevalence across all countries. Although mdr1 184Tyr was relatively common, the haplotype comprising mdr1 86Tyr and 184Tyr (mdr1 YY) occurred at a prevalence below 6% in all surveys across both years, and there was no evidence of an increase in prevalence after seasonal malaria chemoprevention scale-up (Table 2, Table 3).
Figure 2Baseline prevalence (2016) of resistance-associated haplotypes of concern in Plasmodium falciparum in both age groups combined across the African Sahel region
YY-CVIET=mdr1 86Tyr-184Tyr and crt 72Cys-73Val-74Ile-75Glu-76Thr. VAGKGS=dhps 431Val-436Ala-437Gly-540Lys-581Gly-613Ser. Graphs show prevalence with error bars indicating 95% CIs.
,
,
,
,
,
,
were each common in all seven countries, with overall baseline prevalence estimates of 87·8% (3834 of 4369 single genotype isolates) for Asn51Ile, 87·4% (3668 of 4199 single genotype isolates) for Cys59Arg, and 90·8% (3998 of 4403 single genotype isolates) for Ser108Asn (table 2; appendix p 4). The triple mutation (IRN) was the most common haplotype, with frequency ranging from 59·4% (319 of 601 isolates, adjusted) in participants under 5 years in Nigeria to 98·4% (48 of 51 isolates, adjusted) in the same age group in The Gambia at baseline in 2016. The wild-type Asn51-Cys59-Ser108 haplotype was relatively uncommon, ranging in frequency from 11·2% (70 of 704 isolates, adjusted, in Mali) to 0·8% (two of 184 isolates, adjusted, in Chad). No mutations were observed at dhfr codons 140 and 164. The prevalence of the IRN haplotype of dhfr appeared higher in most countries in both age groups in the 2018 survey compared with the 2016 survey (prevalence ratio 1·42 [95% CI 0·58–3·46] in children younger than 5 years and 4·01 [1·63–9·83] in those aged 10–30 years, pinteraction=0·115).
,
Although there was evidence of an increase in prevalence of dhps 540Glu in the 10–30 years age group only, in total 21 of the 2165 evaluable isolates from all ages in 2018 harboured this substitution, representing 0·97% of participants. The dhps mutant Ile431Val, mostly encountered in the more easterly countries, occurred as eight different haplotypes of which VAGKGS and VAGKAA were the most common (appendix pp 8–9). VAGKGS and VAGKAA were observed at highest frequencies in Chad, particularly in 2018 (VAGKGS 11·8% [16 of 136 isolates], VAGKAA 5·2% [seven of 136 isolates]). It remains unknown what effect on sulfadoxine–pyrimethamine efficacy, for therapy or chemoprevention, results from these haplotypes that combine variants at codons 431, 437, 581, and 613 of dhps. In the 2016 dataset, 28 (88%) of 32 isolates carrying the dhps VAGKGS haplotype also carried the dhfr triple mutation IRN, compared with 55% of isolates with the wild-type ISAKAA haplotype in dhps, a relative risk of 1·58 (95% CI 1·36–1·85). The 431Val variant was not combined with the 540Glu variant in any of the 7306 isolates analysed.
Discussion
,
,
,
,
,
,
,
,
,
,
This study reports on two large-scale surveys of qPCR-confirmed P falciparum carriage and molecular markers of resistance to sulfadoxine–pyrimethamine and amodiaquine across seven countries in sub-Saharan Africa before (in 2016) and after (in 2018) implementation of seasonal malaria chemoprevention through Achieving Catalytic Expansion of Seasonal Malaria Chemoprevention in the Sahel (ACCESS-SMC), a programme that sought to remove barriers to the scale-up of seasonal malaria chemoprevention in seven countries in 2015 and 2016. Our findings indicate a substantial decrease in parasite carriage from 2016 to 2018 among children younger than 5 years, the recipients of seasonal malaria chemoprevention, in all six of the countries that had no previous seasonal malaria chemoprevention deployment in the study areas. In The Gambia, where seasonal malaria chemoprevention was already in place before 2015, parasite carriage remained low and stable in this age group. This finding confirms a substantial parasitological benefit, and is consistent with the evidence that, when implemented at scale in The Gambia, Guinea, Mali, Burkina Faso, Niger, Nigeria, and Chad, seasonal malaria chemoprevention was 88·2% effective at preventing clinical malaria within 28 days of administration of seasonal malaria chemoprevention with sulfadoxine–pyrimethamine and amodiaquine.
,
The SVMNT allele of crt, which mediates amodiaquine resistance in Asia and the Americas, was not detected.
Similarly, although the dhfr IRN haplotype at codons 51, 59, and 108 was very common across all study sites, the high-level resistance allele with an additional mutation at codon 164 was not observed. The dhps locus was the most variable, with highly complex patterns of amino acid substitution at more than six positions, generating 24 distinct haplotypes. We described six novel variant positions in dhps: Ile141Met, Gly425Asp, Ile451Met, Ile466Val, Ile470Thr, and Asp575Ala. These rare, emerging mutations have, as yet, no known effect on sulfadoxine susceptibility. However, the key substitution Lys540Glu, which appears to be a principal mediator of sulfadoxine–pyrimethamine treatment efficacy,
remained almost absent apart from in Guinea, and was never combined with Ala581Gly, known to be a highly resistant form in east Africa, suggesting no immediate threat to sulfadoxine–pyrimethamine effectiveness. Of most concern was the emergence in four countries of dhps haplotypes that comprise some combination of the Ile431Val, Ala437Gly, and Ala581Gly mutations, either as VAGKGS or VAGKAA, both previously identified in parasites originating from Nigeria.
,
Limitations to the generalisability of our study include uncertainty as to any possible confounding effect of different transmission intensities across the region on parasite carriage rates, and also the relatively short time interval to capture the emergence of evidence of genetic selection due to seasonal malaria chemoprevention in the parasite population. Furthermore, there is a lack of evidence concerning the effect of the emerging VAGKGS and VAGKAA haplotypes on sulfadoxine–pyrimethamine effectiveness, either for therapy or chemoprevention, making the impact of our findings unclear. Finally, as our data are now 4 years old, repeat surveillance of these haplotypes is urgently needed as seasonal malaria chemoprevention implementation continues throughout the Sahel.
Data published in 2021 from the region broadly concur with our findings.
,
However, a potential threat that the dhps VAGKGS haplotype might increase in prevalence, moving westwards across the Sahel, remains. Focused cohort studies to ascertain the effect of these variants on seasonal malaria chemoprevention efficacy are required.
In conclusion, our analysis of 57 666 blood samples collected across seven countries in the Sahel region of Africa before and 2 years after the implementation of seasonal malaria chemoprevention at scale provides an important overview of the parasitological effect of the intervention. We found strong evidence of a reduction in P falciparum carriage rates in the age group receiving the intervention, but no evidence of an imminent threat from parasite genomes harbouring multilocus mutations conferring multidrug resistance. Some genotypes of concern were identified, particularly at the dhps locus, and continued monitoring is essential to maintain and protect the effectiveness of seasonal malaria chemoprevention in the long term.
KBB, MC, PS, SS, AD, CSM, JLN, LR, DM, J-BO, IZ, KB, SC, KL, AD, IS, IL, PM, and CJS designed the study. KBB, JM, and CJS wrote the protocols. IZ, J-BO, KB, SC, KL, AD, IS, IL, AD, HK, DD, HM, SO, and TE supervised the field surveys. AT, KG, CS, TB, FK, and ML conducted the field surveys. KBB, JM, JN, RM, AT, KG, and SC did the laboratory analyses. KBB, RM, MC, PS, SS, PM, and CJS analysed the data. KBB, PM, and CJS wrote the first draft of the manuscript. All authors reviewed the manuscript. KBB, PM, and CJS accessed and verified all the data. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
