Mapping Antimalarial Drug Resistance in Mozambique: a systematic review of Plasmodium falciparum Genetic Markers Post-ACT Implementation

1. Introduction

Malaria remains one of the most pressing public health challenges in many tropical and subtropical countries, impacting millions of lives across the region [1,2,3] Mozambique faces a substantial malaria challenge, being one of the countries with the highest number of cases and ranking fourth globally in terms of the malaria burden [4]. To combat malaria effectively, Mozambique must address both the vector and the administration of antimalarial drugs [5].

The ongoing fight against malaria has been complicated by the emergence and spread of Plasmodium falciparum resistance to antimalarial drugs which compromises the efficacy of treatment regimens and poses a significant threat to malaria control efforts [6,7,8]. Understanding the distribution of molecular markers of antimalarial resistance is essential for monitoring and managing drug resistance, revising treatment guidelines, and informing the development of new antimalarial drugs.

In Mozambique, as in other parts of sub-Saharan Africa, P. falciparum is the predominant malaria parasite. The region has seen various waves of drug resistance, particularly to chloroquine (CQ), sulfadoxine-pyrimethamine (SP), and more recently, to artemisinin-based combination therapies (ACTs) [1]. Molecular markers have been instrumental in tracking and understanding these resistance patterns [1,9].

Chloroquine was first line treatment for uncomplicated malaria in Mozambique for almost 50 years until 2004 when sulfadoxine-pyrimethamine (SP) and amodiaquine (AQ) were introduced due the emergence and spread of resistance to chloroquine [10]. The primary molecular marker associated with chloroquine resistance is the P. falciparum drug resistance transporter (pfcrt) gene, particularly the single nucleotide polymorphism (SNP) K76T [11]. Several studies have reported high frequencies of this mutation in Mozambique and neighboring countries such as Tanzania or Malawi reflecting the extensive spread of CQ resistance (CQ-R) in the region [12,13,14]. The haplotype defined by specific mutations at amino acid positions 72-76 of pfcrt - CVIET has been associated with CQ-R (in Africa) while the haplotype CVMNK is associated with CQ susceptibility (CQ-S) [15,16]. Following the decline in CQ efficacy, artesunate plus SP was introduced in Maputo Province between 2004 and 2006 as the mainstay for malaria treatment [10]. However, during this pilot study, molecular markers associated with SP resistance (SNPs in the genes dihydropteroate synthase - pfdhps and dihydrofolate reductase - pfdhfr) increased dramatically [10,17,18]. SP resistance is associated with the SNPs A16V, N51I, C59R, S108N, and I164L in the pfdhfr gene, which confer resistance to pyrimethamine, and I431V, S436A/F, A437G, K540E, A581G, and A613S/T in the pfdhps, which confer resistance to sulfadoxine [17,19]. Parasites with multiple SNPs in both pfdhfr and pfdhps were categorized as follows: a quadruple mutant (pfdhfr 51I + 59R + 108N and pfdhps 437G [IRNG]) was classified as "partially resistant"; a quintuple mutant (pfdhfr 51I + 59R + 108N and pfdhps 437G + 540E [IRNGE]) as "fully resistant"; and a sextuple mutant (pfdhfr 51I + 59R + 108N and pfdhps 437G + 540E + 581G or 613S/T [IRNGEG or IRNGES/T]) as "super resistant" [20].

These findings led to a change in the national malaria treatment policy in 2008 to the use of ACTs. In Mozambique, the recommended treatment for uncomplicated P. falciparum malaria are: artemether-lumefantrine (AMT-LUM) (first line of treatment since 2006 [21], artesunate-sulfadoxine and pyrimethamine (AS-SP), artesunate-amodiaquine (AS-AQ), artesunate-mefloquine (AS-MEF), dihydroartemisinin-piperaquine (DHA-PPQ) and artesunate-pyronaridine AS-PY [10,22]. In ACTs, artemisinin derivatives (short half-life; <6h) are combined with a long-acting antimalarial drugs like AQ, MEF, PPQ, LUM, and pyronaridine (PY) to treat uncomplicated malaria [22,23,24]. Regarding ACT partner drugs the primary genes associated with resistance are pfcrt, pfmdr1, pfpm2/3 and the above mentioned pfdhfr and pfdhps [25]. Multiple copies (or copy number variations, CNV) of the P. falciparum multidrug resistance 1 - pfmdr1 gene are an established marker for resistance to MEF (MEF-R) [26,27]. Additionally, SNPs in pfmdr1 have been linked to altered parasite tolerance or susceptibility to several antimalarial drugs, including quinine (QN), AQ, CQ, MEF, and lumefantrine (LUM) [28]. The key pfmdr1 SNPs associated with drug resistance include N86Y, Y184F, S1034C, and N1024D [29,30,31,32,33,34]. The N86Y mutation is related to increased CQ-R and increased sensitivity to MEF [35]. Parasites carrying pfmdr1 haplotype 86Y Y184 showed increased susceptibility to LUM and MEF [36]. The role of the pfmdr1 N86, 184F, and 1246D alleles, as well as pfmdr1 CNV, in P. falciparum response to AMT-LUM remains debated [37].

In recent years, resistance to ACTs has been reported in Southeast Asia in 2008 [25,38,39]. Recently, SNPs associated with artemisinins resistance in Africa [40,41,42] were identified. Resistance to artemisinin derivatives is characterized by delayed parasite clearance times and is linked to SNPs in the Kelch13 protein coded by the gene - pfk13. In particular, the F446I, N458Y, M476I, Y493H, R539T, I543T , P553L, R561H e C580Y, are currently considered validated molecular markers of drug resistance by WHO [38,39,41]. This study objective is to provide a comprehensive analysis of prevalence and distribution of the molecular markers of antimalarial resistance in Mozambique. By mapping the prevalence and distribution of these markers, this research aims to contribute to support the development of targeted interventions to maintain the effectiveness of malaria treatments in Mozambique.

2. Methods

2.1. Selection of Relevant Literature

This study was conducted according to recommendations of the Preferred Reporting Items for Systematic Reviews (PRISMA) [43,44]. Briefly, the search terms and criteria for the inclusion or exclusion were previously defined to be searched across various databases. After conducting the article search, the quality of the selected studies based on the inclusion criteria were assessed by two independent researchers. In cases of disagreement, a third researcher was consulted to solve the dispute. Following the selection of articles for inclusion in the study, a thorough analysis was conducted to extract the most important findings and conclusions. Subsequently, these data were organized and presented in tables or figures. The databases searched were Scopus, PubMed, Web of Science and, in addition to isolated searches for relevant articles found on Google Scholar. A total of 132 articles published from 2007 to July 2024 complied with the inclusion criteria in the title, keywords or summary. The align with the national rollout of ACTs in 2006 [21], and to capture the progress made in molecular monitoring of antimalarial resistance, a 17-year study period was chosen.

2.2. Eligibility Criteria of Studies Include in the Review

The inclusion criteria were all original articles addressing molecular marker of antimalarial drug resistance published in indexed journals (PubMed, Science direct, Scopus and Google scholar), using the keywords: ‘pfpm2/3 OR pfmdr1 OR pfk13 OR pfdhps OR pfdhfr OR pfcrt’ OR ‘copy number variation’, AND ‘Mozambique’.

2.3. Screening and Data Extraction

The articles selected for the study were exported to Microsoft Excel to remove duplicates. The selection of articles was carried out by reading the titles and abstracts and then the full text. The studies were systematized by authorship, year, sociodemographic data, sample size, allele or gene, amino acid, haplotype, type of mutation, CNV, respective prevalence, antimalarial drug and main conclusions.

3. Results and Discussion

3.1. Basic Characteristics of Included Studies

The search strategy identified 132 studies, from which 43 duplicates were removed. After screening titles and abstracts, 56 studies were excluded. Of the remaining 33 studies, 13 were excluded after full-text review, leaving 20 studies for inclusion in this review (Figure 1).

Children under 59 months of age and pregnant women comprise most targeted groups for all types of antimalarials, followed by children and adolescents up to 15 years. Few studies have focused on adult patients. The genes pfdhfr and pfdhps, associated with SP resistance, were identified on studies focusing on patients of all ages and sexes [45,46]. Regarding the gene pfk13, associated with artemisinin derivatives resistance, the study by Da Silva [47] included both children and adults of both sexes, while the study by Escobar [48] was focused on adult patients of both sexes.

Figure 1. - PRISMA diagram of the systematic review. Steps followed by this systematic review according to the PRISMA (“Preferred Reporting Items for Systematic Reviews and Meta-Analyses”) guidelines.

Figure 1. - PRISMA diagram of the systematic review. Steps followed by this systematic review according to the PRISMA (“Preferred Reporting Items for Systematic Reviews and Meta-Analyses”) guidelines.

The information about the 20 studies included in this review is summarized in Table 1, Figure 2 and detailed in the Supplementary material (Tables S1, S2, S2 S4 and S5). Most studies (17/20; 85%) were conducted in southern Mozambique, specifically in the provinces of Maputo, and Gaza (Table 1). It is important to emphasize that 40% (8/20) of the total articles included in this study address or three genes (pfcrt, pfdhfr and pfdhps, pfmdr1, pfk13 and CNVs pfpm2/pfpm3/pfmdr1) in different provinces [61,62,63,64,65,66,67,68], (detailed in Supplementary material).

A total of nine studies (45%) monitor pfcrt gene (associated with CQ-R) (Table 1), six in Maputo, two in Gaza and one for Tete, Zambezia, Cabo Delgado and Inhambane. pfdhr and pfdhps genes, associated with SP resistance, were found in 13 studies (65%), ten of these studies were conducted in Maputo (Table 1). Of these 13 studies, nine involved children, four pregnant women, and one adult (Table S2).

A total of nine studies were found addressing gene the pfmdr1(Table 1). Five of these focused in Maputo and two in Gaza, with five involving children and four pregnant women (Table S3). Five studies were founded examining the pfk13 gene (Table 1), including two involving adults and pregnant women, and three involving children, mainly in Maputo (Table S4). Finally, three studies investigated CNV were identified, all included pregnant women and children, with two in Maputo and one in Niassa, Manica, Sofala, Tete and Gaza respectively. This overview highlights a concentration of studies in the Maputo and Gaza provinces (in the south of the country; Figure 2) and a predominance of research involving children and pregnant women (detailed in supplementary material).

Malaria remains a significant public health concern in Mozambique, with Plasmodium falciparum being the predominant species responsible for the disease [49]. Understanding the genetic variants associated with drug resistance is crucial for developing effective treatment strategies and transmission control of the disease. This review reveals substantial regional variability in genetic mutations associated with malaria drug resistance in Mozambique (Figure 2). While data is more robust for Maputo and Gaza, significant gaps remain for other provinces, underscoring the need for further research to monitor genetic variations over time. For instance, research on pfcrt gene polymorphisms, which encode CQ-R, primarily focused on two southern provinces. The prevalence's of pfcrt gene were detected between 40 -84% patients [10,50,51]. In the past, these data were important for changing the Malaria treatment, which at the time was based on chloroquine, in line with what was happening in all malaria endemic countries [52]. In Gaza and Maputo, showed a moderate to high prevalence of the 76T pfcrt SNP (CQ-R) during the first years after the introduction ACTs [10,53]. However, more recent studies conducted in the same provinces, after the discontinuation of chloroquine (samples collected 2017-2019) have identified high prevalence of CQ-S P. falciparum genotypes [54,55,56]. This shift suggests a reduced selective pressure from CQ. Similar trends were observed in other sub-Saharan African countries, including Kenya, Malawi, Sierra Leone, Ghana, Angola, and Ivory Coast, where CQ-S P. falciparum genotypes have re-emerged [57,58,59,60].

3.2. Antimalarial Resistance Associated Polymorphisms

3.2.1. pfcrt

Majority of studies (87.5%) addressing pfcrt gene were conducted in the provinces of Maputo and Gaza (Table 1 and Table S1), with only one study addressing multiple provinces, namely Inhambane, Zambezia, Tete, and Cabo Delgado (see Table S1). The most recent evaluation of pfcrt CVIET haplotype is from 2024 and revealed a prevalence of 1.1% in Maputo and 9.2% in Gaza, 0% in Inhambane, Zambezia, Tete and Cabo Delgado (see Figure 2 and Table S1). The most prevalent haplotype was CVMNK (CR-susceptible), found in 92.2% of patients sampled in Cabo Delgado, Tete, Zambezia and Inhambane in 2021 [54]. CVIET (CR-resistant) was reported in 7.8% of patients sampled in Inhambane, Zambezia, Tete and Cabo Delgado [56], and 76T, found in 84% of adult patients of both genders in Maputo province in 2024 [50], 48.8% in children aged 2 to 3 months and 46.4% in pregnant women [70]. The gene pfcrt confers resistance to a wide range of quinoline and quinoline-like antimalarial drugs in P. falciparum, with local drug histories driving its evolution and, hence, the drug transport specificities. For example, the change in prescription practice from CQ to PPQ in Southeast Asia has resulted in pfcrt variants that carry an additional SNPs (H97Y, F145I, M343L, or G353V ), leading to PPQ resistance [77]. There is a notable gap in the current understanding of specific pfcrt SNPs in Mozambique, particularly the ones associated to PPQ-R in Southeast Asia. Evaluating these markers in Mozambique could provide essential information for updating malaria treatment guidelines and managing potential PPQ-R as drug policies shift in the country.

3.2.2. pfdhr and pfdhps

The prevalence of SP-resistance haplotype IRNGE was high in Maputo (95.1%) and Gaza (89.5% )(Figure 2). The main studied mutations occurred at amino acid positions 51, 59, 108, 164, 437, 540, and 58, either individually or in combination within the pfdhfr or pfdhps genes, resulting in multidrug resistance haplotypes (see Table S2). Taking into account the most recent results, in Maputo province, the most frequent haplotype was IRNGE, with 95.1% prevalence in samples collected in 2015-2019 [69], and with 94.2% in samples collected in 2016-2019 (Table S2) [78]. In Gaza, the sextuple IRNGEG haplotype was observed in 8% of the samples and the quintuple IRNGE in 55% [79]. Studies from the centra (Tete and Sofala) [56] and southern (Gaza and Maputo) [80] regions reported higher prevalence of SNPs in the pfdhfr or pfdhps in various combinations than northern (Cabo Delgado ) provinces [56] (Table S2).

In Mozambique, SP is used for intermittent preventive treatment in pregnancy (IPTp) has been linked to the accumulation of SP-resistant mutations in pfdhfr and pfdhps [10,51,63]. This may facilitate the selection of resistant parasites due to the repeated exposure to SP. Nevertheless, despite widespread SP resistance, studies indicate that administering three or more doses of SP to pregnant women may still confer a protective benefit against P. falciparum [78]. The geographical distribution of pfdhfr and pfdhps SNPs studies in Mozambique reveals uneven coverage across provinces, with a significant focus on the southern region, particularly Maputo (10; Table S2), with a limited number of studies conducted in other provinces (3 in Gaza [10,55,63], 2 in Sofala [56,74] and 1 in Cabo Delgado and Tete [56]; Table S2). The reminder 5 provinces do not have published information (Figure 2). In Maputo, high prevalence rates of mutations associated with SP-R were reported, such as 51I (36.6%-88%), in pfdhfr, 59R (52.4-91%), 108N (50.4% - 99.2%), 540E (7.9% - 94.9), 437G (42-96.2%) in pfdhps. The quintuple mutant – IRNGE was reported in multiple studies with high prevalence namely 94.2% in samples from 2106-2019 [78] and 95.1% in samples from 2015-2019 [69] (Table S2). These underscores substantial SP resistance in Mozambique and follow the trend of other African countries, such as Ghana and Nigeria [81,82].

3.2.3. pfmdr1

SNPs in pfmdr1 were studied in all provinces, except Niassa, Manica and Nampula (Figure 2 and Table S3). The latest prevalence of SNPs reported were N86 found in 93.1% in Cabo Delegado, 95.7% Gaza, 95.5% in Sofala and 98.8% in Maputo respectively; 184F reported in 41.7% in Tete, 43.2% in Sofala, 50.5% in Maputo and 53.58% in Inhambane respectively (Figure 2).

The pfmdr1 encodes a protein involved in drug transport within the parasite and plays a key role in susceptibility to the key antimalarials ACTs. Although mutations in pfmdr1 are not directly responsible for resistance to artemisinins, they influence the effectiveness of partner drugs, such as LUM or AQ, the haplotype NFD has been associated with higher susceptibility to these partner drugs [29,30,31,32,33,34], After Mozambique transitioned from chloroquine to ACTs for malaria control, the prevalence of pfmdr1 mutations changed, with the NFD haplotype (amino acids 86/184/1246) variant becoming more common [54,55]. The current data reveals a significant geographical gap in the country regarding studies on the pfmdr1 gene. Most research has been concentrated in Maputo (5) [53,55,69,70,75] and Gaza (3) [10,56,80]. With limited data from other provinces like Cabo Delgado and Tete (2) [54,56] or Zambezia, Inhambane (1 ) [54] and Sofala (1) [56]. This regional imbalance of studies leaves large parts of Mozambique underrepresented, especially in the northern and central provinces. For instance, no studies have been recorded in Nampula for pfmdr1 (or any other molecular marker), and in Sofala, the only study available is based on samples collected nearly a decade ago (2015) [56]. The most recent studies have identified an appreciable prevalence of mutations in pfmdr1 namely the SNPs N86 (98.8%), 184F (75.4%) in Maputo (samples collected in 2015-2019) [69] and the haplotype NFD in Inhambane 74.4%; Cabo Delgado 66.7%; Tete 11.0% or Zambezia 50.0% (samples collected in 2018) [54]. Similar trends have been observed in several other African countries [83,84,85,86].

3.2.4. pfk13

Polymorphisms of pfk13 associated with multidrug resistance in P. falciparum were investigated in five studies (Table S4). Most studies (75%) were conducted in 8 provinces, except Inhambane and Nampula. Only one study examined multiple provinces (Figure 2 and Table S4). Low frequency of pfk13 was observed in all provinces where studies were conducted (Figure 2). Maputo and Tete, with 4% each, was the province with the highest prevalence of pfk13 SNPs (Figure 2). Notable, findings included the synonymous mutation at codon 469 (TGC to TGT) in one sample and at codon 548 (GGC to GGT) in three samples from Zambezia province (Mopeia city [54]). Two studies identified for the SNPs: 494I [55] and 578S [55], both with 4% prevalence and both in samples from Maputo province [48], (Table S4). None of these two SNPs are currently considered validated molecular markers of drug resistance by WHO [38,39,41].

The prevalence of pfk13 SNPs varies by region; in 2019, it was 45.4% in Southeast Asia compared to a much lower prevalence of 7.6% in Africa [52]. In Mozambique, 8/10 provinces have evaluated the presence of pfk13 SNPs and none of the validated or candidate mutations were identified so far. Similar findings have been reported in other African countries like Gabon [87], Senegal [88], Kenya and Ethiopia [89], where low frequencies of pfk13 SNPs were observed. However, A578S, was detected in samples from Niassa and Tete provinces [54,90], as well as in Uganda and Gabon [91]. The identification of independent emergence of pfk13 SNPs (with partial resistance to ACTs) in the African region, especially in Rwanda and Uganda [92,93,94,95], highlights the importance of surveillance efforts to obtain genotypic data and map the extent of pfk13 SNPs throughout the WHO African Region [96]. The recent detection of SNPs M476I, P553L, R561H, P574L, and C580Y in Africa serves as an early warning signal [40,41,42,97].

3.2.5. Copy Number Variations in pfmdr1 and pfpm2/3

Figure 2 summarizes the latest prevalence rates and primary study provinces and Table S5 displays detailed data collected from various populations (children, adults, pregnant women) between 2015 and 2023. Only three studies were found investigating the prevalence of copy number variations (CNVs) in pfmdr1 and pfpm2 (Table 1). Two studies Brown et al., 2024 [76] and Gupta et al., 2018 [56], covered multiple provinces, while the third study (Gupta et al, 2020 [55]) focused solely on Maputo province (Table S5). Pfmdr1 CNV prevalence rates were as follows: 4.8% in the north (Niassa), between 1.1%, 2.3% in Tete, Manica and Sofala (centre) and 5.7% in the south (Maputo; Figure 2). Regarding plasmepsins (pfpm2 and pfpm3) CNVs, prevalence rates were higher in the southern provinces of Gaza and Maputo (3.4% for pfpm2 and 2.3% for pfpm3) compared to the northern and central provinces (Niassa, Tete, Manica, and Sofala), where the prevalence ranged from 1.1% to 1.6% for pfpm2 and 1.6% to 2.3% for pfpm3 (specifically 1.6% for pfpm3 in Niassa and 2.3% in Manica).

There are only 3 studies assessing CNVs of pfmdr1, pfpm2 and pfpm3, in Mozambique one assessing all three [76], and two assessing pfmdr1, pfpm2 [55,56]. These reveled prevalence rates ranging from 0-5.1% for pfmdr1, 1.1-3.4% for pfpm2 and 1.6-2.3% for pfpm3 [10,22,70]. Studies from Mozambique reveled much lower prevalence of pfmdr1 CNV than other African countries, namely Kenya (6.2%), Ghana (18%), Tanzania (10.2%), West Ethiopia (8.4%) or North of Ethiopia (54.14%) [83,98,99,100,101]. Observations from Mozambique on the other hand, are in line with studies from other African countries like Nigeria or Democratic Republic of Congo were increased CNV was not observed for pfmdr1 [102,103]. Regarding pfpm2 prevalence, the two studies recorded in Mozambique also reported a much lower prevalence than others from Africa (7.7% in Tanzania [98]; or 67.9% in Guinea Equatorial [104], but a comparable to (e.g.) to Liberia, or Uganda were increased copies of pfpm2 was not observed [105,106].

Copy number variation (CNV) has also been found to play a significant role in the development of antimalarial drug resistance. One copy of pfmdr1 is associated with slower clearance of parasites after PPQ treatment as compared to more copies of pfmdr1 [107], while having two copies of pfpm2 is associated with slower clearance [108,109], after PPQ treatment. This inverse selection pressure argues in favor of keeping these molecular markers under constant surveillance.

4. Conclusions

Although malaria is endemic throughout the country, the central and northern regions of Mozambique have the highest incidences, especially the provinces of Zambezia, Nampula and Cabo Delegado, which are most affected by the disease [110]. This situation may be associated with the fact that these are coastal provinces, with climatic conditions and socio-economic factors favorable to the proliferation of the malaria vector. However, most studies on monitoring molecular markers of resistance to antimalarials are concentrated in the southern region of the country.

Maputo Province has the highest number and more up-to-date studies conducted (17), followed by Gaza (4) and Tete, Sofala and Cabo Delgado (3). Other provinces such as Manica and Niassa (2), Zambezia and Inhambane (1) have limited studies, while no studies are reported from Nampula. This review highlights the concentration of research efforts primarily in Maputo, reflecting a potential need for further investigation to gather more recent data on these genetic markers in the underrepresented provinces.