Regional variation in sulfadoxine-pyrimethamine resistance genotypes and haplotypes of Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase genes in Western Kenya

Abstract

Background

Surveillance of molecular markers associated with antimalarial resistance in Plasmodium falciparum is critical for tracking the emergence, evolution, and spread of resistant malaria parasites in the population for timely and effective interventions. As shifting use of sulfadoxine-pyrimethamine (SP) in Kenya constitutes a differential selection pressure, this study compared resistance genotypes and haplotypes in P. falciparum isolates from endemic and epidemic regions of western Kenya.

Methods

A cross-sectional design was employed to collect blood samples from febrile patients residing in Ahero in Kisumu County, an endemic region, and Marani in Kisii County, an epidemic region. Molecular markers for antifolate resistance, dihydrofolate reductase (Pfdhfr) and dihydrofolate synthetase (Pfdhps), were genotyped for selected samples (N = 112) from Kisumu (n = 60) and Kisii (n = 52). Subsequent analysis was conducted for sequence polymorphisms, mutation frequency and haplotype prevalence.

Results

Genotyping of SP resistance markers identified 436H (28.8%), 437G (99%), and 540E (97.1%) in Pfdhps and 51I (100%), 59R (97.3%), and 108N (100%) in Pfdhfr as mutations that presented high frequency, with low multiplicity of infection (MOI) in both Kisumu (0.3196) and Kisii (0.2738). The double mutant SGEAA (70.18% in Kisumu vs. 51.06% in Kisii) and triple mutant HGEAA (26.31% vs. 44.68%) in Pfdhps, along with the triple mutant IRNI (86.67% vs. 98.08%) in Pfdhfr, exhibited significant regional differences in prevalence (p < 0.05). The Pfdhps-Pfdhfr haplotype analysis revealed a high prevalence of the quintuple mutant SGEAA-IRNI (57.89% in Kisumu vs. 48.94% in Kisii; p > 0.05) and a significantly higher prevalence of the sextuple mutant HGEAA-IRNI in Kisii compared to Kisumu (44.68% vs. 16.31%; p < 0.05). Comparatively, given that Pfdhps-516F and Pfdhfr-164L were the dominant alleles in Kisumu, while the Pfdhps-436H allele was dominant in Kisii, along with HGEAA and HGEAA-IRNI haplotypes (p < 0.05), they highlight regional variation in SP resistance genotypes and haplotypes.

Conclusions

This study demonstrates that the fully resistant double Pfdhps-SGEAA and triple Pfdhfr-IRNI haplotypes have approached fixation in both endemic and epidemic regions, while the dominance of the 164L allele in the endemic region signals the emergence of super-resistance. These findings suggest a review of the therapeutic efficacy of SP and continuous surveillance of resistance due to the presence of fully resistant haplotype (SGEAA-IRNI) and super-resistant haplotype (F/HGEAA-IRNL) in the population of P. falciparum strains in western Kenya.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12936-025-05570-9.

Background

Despite the continued implementation of numerous control and prevention interventions, malaria persists as one of the leading causes of morbidity and mortality globally, with the highest burden in the sub-Saharan African region, including Kenya. In 2023, malaria accounted for 263 million cases and 597 thousand deaths globally, corresponding to an incidence rate of 60.4 per 1000 people at risk and a mortality rate of 13.7 per 100,000 people, with the WHO region comprising about 94% of the malaria burden [1]. In the same period, Kenya reported 3.2 million cases and 11,500 deaths, putting at risk over 70% of the population, particularly those in endemic regions, epidemic-prone highlands, and seasonal transmission zones [1].

Plasmodium falciparum is dominant in Africa, accounting for about 99% of estimated cases and deaths in Kenya and the sub-Saharan African region [2, 3]. Current interventions, namely, vector control, case management, vaccination, and chemoprevention, have significantly contributed to control and prevention efforts. However, the emergence of parasite and vector resistance to these interventions presents a new challenge in effectively controlling and preventing malaria among populations.

Drug resistance is a significant factor that reduces the efficacy of antimalarial drugs due to the mutation and evolution of Plasmodium parasites. Therapeutic efficacy studies have demonstrated that the use of sulfadoxine-pyrimethamine (SP) has reduced therapeutic efficacy due to parasite resistance in Kenya, Africa, and across the world [3, 4]. Historically, the Kenya Ministry of Health approved the use of SP as the second-line treatment (1983–1997), first-line treatment (1998–2007), intermittent preventive therapy (IPTp) across the country (2001–2008), and IPTp in endemic regions only (2009 to present) [2, 5–7]. These trends in SP usage contribute to the differential drug pressure and varied evolution of Plasmodium parasites within the population. In light of the rising resistance, the World Health Organization (WHO) recommends regular and timely surveillance of resistant parasites through therapeutic efficacy protocols to identify treatment failures and related molecular markers [3, 8]. Continuous monitoring of resistant parasites is integral to malaria surveillance to capture their evolution and spread in the population.

To effectively control and prevent malaria, the WHO has validated molecular markers linked to the use of antifolate drugs to assist in identifying resistant strains within the population [3]. Molecular markers in P. falciparum genes, dihydropteroate synthase (Pfdhps) and dihydrofolate reductase (Pfdhfr), which are responsible for the synthesis of nucleic acids, confer resistance to SP [9]. Validated mutations in Pfdhps (436A/F, 437G, 540E, 581G, and 613 T/S) confer resistance to sulfadoxine, while those in Pfdhfr (51I, 59R, 108N, and 164L) confer resistance to pyrimethamine, with proguanil having additional A16V and S108T mutation [3, 9]. The combinations and permutations of these molecular markers in the population determine the degree of resistance to SP. Recent studies in western Kenya have shown that double mutation in Pfdhps (437G and 540E) and triple mutations in Pfdhfr (51I, 59R, 108N) are approaching the fixation state with a low prevalence of 436A/F/H and 164L mutations [6, 10]. Given that the use of SP as an IPTp in Kenya continues to show increasing trends of resistance, surveillance of associated molecular markers is necessary.

This study compared the prevalence, diversity, and selection of mutations associated with SP resistance in the Pfdhps and Pfdhfr genes from P. falciparum isolates from the Lake endemic region of Kisumu, and the highland epidemic-prone region of Kisii in western Kenya. Specifically, comparative genotyping analysis of Pfdhps mutations (436F/H, 437G, 540E, 581G, and 613 T/S) and Pfdhfr mutations (51I, 59R, 108N, and 164L) was performed to determine their prevalence and distribution in the population between the two study regions. This study sought to demonstrate the regional variation in resistant genotypes and haplotypes associated with SP resistance in the endemic region of Kisumu and the highland epidemic-prone region of Kisii.

Methods

Study sites

The study health facilities for sample collection were Ahero sub-County Hospital (ASCH), Eramba Health Centre (EHC), and Marani Sub-County Hospital (MSCH) (Fig. 1). ASCH (0° 10′ 24.60″ S, 34° 55′ 24.96″ E) in Kisumu County, Muhoroni sub-County, has an elevation of about 1100 m above sea level (asl), representing the Lake endemic region. In contrast, EHC (0° 32′ 57.84″ S, 34° 46′ 51.24″ E) and MSCH (0° 34′ 32.39″ S, 34° 48′ 06.08″ E) in Kisii County, Marani sub-County, represent the highland epidemic-prone region with elevations of approximately 1700 m asl. With a prevalence of about 20%, the Lake endemic region has a significantly higher malaria burden compared to the highland epidemic-prone region, with a prevalence of about 1% [5]. Homa Bay County is situated between Kisumu and Kisii Counties and acts as a buffer zone between the two regions with different epidemiological zones. These two study regions have experienced varying usage of SP for IPTp since 2009, when the government limited its implementation to the lake and coast endemic regions only [2]. Regional comparative analysis of resistance in P. falciparum parasites in the population would shed light on the role transmission rates and use of SP for IPTp. A regional comparative analysis of resistance in P. falciparum parasites can provide insights into how transmission intensity, the use of SP for IPTp, and other contextual factors influence antifolate resistance patterns in the population.

Fig. 1.

The map displays the study health facilities in western Kenya, including Ahero Sub-County Hospital (ASCH) in Kisumu County, the Lake endemic region, and Eramba Health Centre (EHC), along with Marani Sub-County Hospital (MSCH) in Kisii County, which is the highland epidemic-prone region

Study design

An analytical cross-sectional research design was employed to conduct a comparative molecular surveillance of drug resistance markers for antifolates in the Pfdhfr and Pfdhps genes in P. falciparum isolates from the Kisumu and Kisii regions. The speed and cost-effectiveness of this research design make it suitable for surveillance activities involving common diseases [11]. Additionally, this research design is specifically appropriate for assessing the prevalence of resistant malaria parasites in the population [12]. The two study regions were selected for comparative molecular surveillance based on their differences in epidemiological zones and malaria transmission rates. The Malaria Indicator Survey classifies Kisumu County as part of the Lake endemic region, with the prevalence rate of about 20%, and Kisii County as the highland epidemic-prone region with the prevalence rate of approximately 1% [2]. The differences in epidemiological zones allow for a comparative analysis of the population’s emergence, evolution, and spread of resistant malaria parasites.

Participants and sample collection

A total of 1818 febrile patients across all the age groups who sought medical attention at Ahero sub-County Hospital in Kisumu County (n = 472) and Marani sub-County Hospital (n = 808) and Eramba Healthcare Centre (n = 538) in Kisii County (n = 1346) from June through August 2024 were recruited for the study through the administration of written informed consent for malaria diagnosis. A multistage sampling strategy [12] was employed to select malaria-positive samples based on the inclusion and exclusion criteria. The inclusion criteria entailed a positive malaria diagnosis by microscopy and provision of informed consent or assent, while the exclusion criteria included non-residency, recent travel history, RT-PCR cycle threshold (Ct) > 32, and failure to amplify Plasmodium DNA (Fig. 2). Qualified medical laboratory technicians collected about 200 µl of blood samples from individuals through a finger-prick. From the blood sample, Giemsa-stained thick and thin blood smears for microscopy and dried blood spot (DBS) on Whatman™ Blood Stain Cards (GE Healthcare, Cardiff, UK) for molecular analysis were prepared. The DBS samples were air-dried, placed in Ziploc bags containing silica beads, transported in a cooler box, and stored at − 4 °C. Blood smear and DBS samples were transported to the sub-Saharan-Africa International Centre of Excellence for Malaria Research (ICEMR) laboratory for storage and processing. Blood smear samples obtained from hospitals were re-examined by a qualified microscopist to validate malaria diagnosis and quantify parasitaemia, calculated as the number of parasites per 8000 leukocytes per microliter of blood. DBS samples were subsequently used for molecular analysis and amplification of target genes.

Fig. 2.

The flowchart showing the selection of febrile patients and processing of samples from Kisumu and Kisii Counties

Molecular detection of parasites

Genomic deoxyribonucleic acid (DNA) was extracted from malaria-positive DBS samples using the optimized protocol of the saponin-Chelex method [13] and stored at − 20 °C. A serial dilution of a known concentration of DNA obtained from a cultured P. falciparum strain (NF54) was used to generate a standard curve for estimating P. falciparum DNA concentration based on the cycle threshold [14]. The genomic DNA samples were further detected for P. falciparum infections using real-time/quantitative polymerase chain reaction (RT/qPCR) (Thermo Fisher Scientific, Carlsbad, CA) with species-specific probes and primers targeting the acid terminal sequence of the var gene (PfvarATS) (PF3D7_0617400) [15]. The qPCR setup had a final reaction volume of 12 µL containing 0.5 µL of probes (2 µM) 0.4 µL of reverse primers (10 µM), and 0.4 µL of forward primers (10 µM), which are specific to P. falciparum, 0.6 µL of PerfeCTa qPCR ToughMix (2X) (QuantaBio, Beverly, MA), 2 µL of extracted genomic DNA, and 0.1µL of nuclease-free water. The thermal cycling conditions for qPCR were initial incubation at 50 °C for 2 min, and then 45 cycles at 95 °C for 2 min, 95 °C for 3 s, and 58 °C for 30 s [16]. The genomic DNA samples with cycle threshold values ≤ 32 were selected because they contained sufficient parasitaemia (> 1000 parasites/µL) for Pfdhfr and Pfdhps genotyping.

Target amplification and sequencing

PCR amplification and optimization of Pfdhfr and Pfdhps genes were performed using the Bio-Rad T100 Thermal Cycler^® (Bio-Rad Laboratories, Inc., Hercules, CA, USA) with specific primer sets [17], a 25 µL reaction mixture, and optimized thermal cycling conditions (Additional file 1: Table S1). Amplicons (5 µL) were resolved via gel electrophoresis on a 1.5% agarose gel stained with 0.2 µL of RedSafe™ (New England Biolabs, Ipswich, MA, USA). Electrophoresis was conducted at 120 V and 400 mA alongside a 100 bp DNA ladder (New England Biolabs, Ipswich, MA, USA). Gel visualization was performed under blue light using the SmartDoc™ Imaging System (Accuris, Edison, NJ, USA) to confirm the expected amplicon sizes of Pfdhfr (594 bp) and Pfdhps (720 bp). Amplicons were purified using QIAquick PCR Purification Kit^® (Qiagen, Hilden, Germany) and sequenced in both the forward and reverse directions using the Sanger method (GENEWIZ, Inc., La Jolla, CA). Plasmodium falciparum DNA from the 3D7 strain was included as a positive control, while a no-template reaction mixture served as a negative control.

Data analysis

ChromasPro^® software version 2.2.0 (Technelysium Pty Ltd, South Brisbane, Australia) was used to process raw sequencing data (ab1 files) by trimming low-quality ends, generating contigs, aligning with references, editing sequences, and exporting them as a multi-FASTA file [18]. The generated contigs were aligned with respective references of P. falciparum 3D7 strain obtained from PlasmoDB for the Pfdhfr gene (PF3D7_0417200) and the Pfdhps gene (PF3D7_0810800) [19]. Additionally, chromatograms were visually inspected to check the quality of the base calls and identify mixed genotypes as double peaks with a minor peak covering ≥ 30% of the major peak. Multiple Sequence Comparison by Log-Expectation (MUSCLE) in Molecular Evolutionary Genetics Analysis (MEGA 12) was used in the alignment of nucleotide and translated amino acid sequences to differentiate synonymous and non-synonymous mutations [20]. The Expectation-Maximization (EM) algorithm was used to calculate maximum-likelihood estimates of haplotype frequencies, haplotype prevalence, and multiplicity of infection (MOI) of SNPs in Pfdhfr at codons 51, 59, 108, 164, and Pfdhps at codons 436, 437, 540, 581, and 613 [21]. The Chi-square test, Fisher’s exact test, t-test, and Mann–Whitney U test were used to assess differences in demographic and sample characteristics between the Kisumu and Kisii regions. Moreover, bootstrap test (10,000 iterations) was used to compare difference in MOI and prevalence of haplotypes between the two study regions based on the alpha level of p < 0.05 [22]. Nucleotide sequences for Pfdhfr and Pfdhps were deposited in GenBank under accession numbers PV775415-PV775526 and PV775527-PV775630, respectively.

Results

Demographic and sample characteristics

Malaria positivity, gender, age, body temperature, parasitaemia, P. falciparum DNA concentration, and multiplicity of infection (MOI) were evaluated as demographic and sample characteristics of the study participants (Table 1). Microscopy screening of 1818 febrile patients revealed a significantly higher malaria positivity rate in Kisumu (18.64%, 88/472) compared to Kisii (6.76%, 91/1,346) (χ² (1) = 54.26, p < 0.001). Subsequently, a total of 112 P. falciparum isolates were selected for genotyping analysis of Pfdhfr and Pfdhps genes from participants residing in the Kisumu region (n = 60) and the Kisii region (n = 52) based on the inclusion and exclusion criteria (Fig. 2). Comparisons of gender distribution, median age, mean body temperature, median parasitaemia, mean P. falciparum DNA concentration, and multiplicity of infection (MOI) did not show significant differences between the two study regions (p > 0.05).

Table 1.

Comparison of demographic and sample characteristics of febrile patients from Kisumu and Kisii counties

Characteristics		Kisumu county	Kisii county	Test statistic	P-Value
Malaria positivity (%) (N = 1818)		18.64% (n = 88/472)	6.76% (n = 91/1346)	χ2 (1) = 54.26	< 0.001
Sub-samples (N = 112)		n = 60	n = 52	–	–
Gender	male (%)	26 (43.3%)	23 (44.2%)	χ2(1) = 0.01	0.924
	female (%)	34 (56.7%)	29 (55.8%)
Age (years), median [IQR]		16.5 [9–24]	14.5 [1.5–27.5]	U = 1740.5	0.292
Body temperature (°C), mean ± SD		37.6 ± 0.5	37.7 ± 0.4	t(110) = 1.035	0.354
Parasitaemia (parasites/µL), median [IQR]		4,600 [1410–7790]	5,160 [380–7940]	U = 1,740.5	0.735
P. falciparum DNA concentration (ng/µL), mean ± SD		26.70 ± 1.31	25.81 ± 1.46	t(110) = 3.415	0.163
Multiplicity of infection (MOI)		0.3196	0.2738	Bootstrap test	0.494

Frequencies of Pfdhfr and Pfdhps mutations

Plasmodium falciparum isolates from the Kisumu and Kisii regions were genotyped for resistance markers in Pfdhps associated with SP at codons 436, 437, 540, 581, and 613, as summarized in (Table 2). At codon 436 of Pfdhps, the S436 wild-type allele was predominant (56.7%, 59/104), whereas mutant alleles of 436H and 436F were 28.8% (30/104) and 8.7% (9/104), respectively. Additionally, 5.8% (6/104) of the isolates had mixed genotypes of wild-type and mutant alleles (S436H). The Kisumu region had a significantly higher frequency of 436H mutants (40.4%, 19/47) than the Kisii region (19.3%, 11/57) (p < 0.05), with 7% (4/57) and 4.3% (2/47) of mixed genotypes (S436H), respectively. At codon 437 of Pfdhps, the 437G mutant allele was detected in most isolates (99%, 103/104) of both Kisii (98.2%, 56/57) and Kisumu (100%, 47/47) without any mixed genotypes. In the Kisumu region, an unvalidated 516F mutation in Pfdhps was observed at a significantly high frequency (8.5%, 5/57; p < 0.05), along with a mixed L516F genotype. The 540E mutant allele was predominantly present in isolates (97.1%, 101/104) with a single wild-type allele (1%) and two mixed genotypes (1.9%). Comparatively, the distributions of the wild-type, mutant, and mixed genotypes were 1.8% (1), 96.4% (56), and 1.8% (1) in Kisumu and 0%, 95.7% (46) and 2.1% (1) in Kisii, indicating that resistance alleles are dominant in both regions. The A581G mixed genotype (4.3%) was observed in Kisii, while A613T/S allele had no mutant or mixed genotypes.

Table 2.

Frequencies of single-nucleotide mutations in the Pfdhfr and Pfdhps genes conferring resistance to sulfadoxine-pyrimethamine in P. falciparum isolates from western Kenya

				Pfdhps resistance markers					Pfdhfr resistance markers
Region (n)	Type of allele	1306 T  >  C 1307 C >  A S436H	1307C  >  T S436F	1310G > C A437G	1548A  >  A L516F	1618A  >  G K540E	1742C > G A851G	1837G  >  A A613T/S	152A  >  T N51I	175 T  >  C C59R	323G  >  A S108N	490A  >  T I164L
		n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)	n (%)
Kisumu	Wild	37 (64.9)		1 (1.8)	51 (89.5)	1 (1.8)	57 (100)	57 (100)	0 (0)	3 (5)	0 (0)	52 (86.6)
	Mutant	11 (19.3)	5 (8.8)	56 (98.2)	5 (8.8)	55 (96.4)	0 (0)	0 (0)	60 (100)	57(95.0)	60 (100)	7 (11.7)
	Mixed	4 (7)	0 (0)	0 (0)	1 (1.8)	1 (1.8)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)	1 (1.7)
Kisii	Wild	22 (46.8)		0 (0)	47 (100)	0 (0)	45 (95.7)	47 (100)	0 (0)	0 (0)	0 (0)	51 (98.1)
	Mutant	19 (40.4)	4 (8.5)	47 (100)	0	46 (97.9)	0 (0)	0 (0)	52 (100)	52 (100)	52 (100)	1 (1.9)
	Mixed	2 (4.3)	0 (0)	0 (0)	0	1 (2.1)	2 (4.3)	0 (0)	0 (0)	0 (0)	0 (0)	0 (0)
Total	Wild	59 (56.7)		1 (1)	98 (94.2)	1 (1.0)	102 (98.1)	104 (100)	0 (0)	3 (2.7)	0 (0)	103 (92.0)
	Mutant	30 (28.8)	9 (8.7)	103 (99)	5 (4.8)	101 (97.1)	0 (0)	0 (0)	112 (100)	109 (97.3)	112(100)	8 (7.1)
	Mixed	6 (5.8)	(0)	0 (0)	1 (1)	2 (1.9)	2 (1.9)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.9)
	P-Value	0.032	0.500	0.548	0.046	0.573	–	–	–	0.150	–	0.048

Furthermore, P. falciparum isolates from the Kisumu and Kisii regions were genotyped for resistance markers in Pfdhfr associated with sulfadoxine-pyrimethamine at codons 51, 59, 108, and 164, as summarized in (Table 1). The 51I and 108N mutant alleles of Pfdhfr were present in all the isolates from the Kisumu and Kisii regions (100%). However, the 59R mutant allele was predominant (97.3%, 109/112) with a low frequency of wild-type alleles in the Kisumu region (5%, 3/60). A significant difference in the distribution of resistance alleles occurred with a higher mutant frequency (164L) in the Kisumu region (11.7%, 1/60) than the Kisii region (1.9%, 1/52) (p < 0.05). A mixed genotype of I164L allele was also observed in Kisumu (1.7%, 1/52), but it was lacking in the Kisii region.

Prevalence of the Pfdhfr and Pfdhps haplotypes

The analysis of haplotypes was performed for five validated genotypes (S436H/F, A437G, K540E, and A581G) and four genotypes (N51I, C59R, S108N, and I164L) in the Pfdhps and Pfdhfr genes, respectively (Table 3). For Pfdhps, the double SGEAA mutant was a significantly more prevalent haplotype in Kisumu (70.18%) than in Kisii (51.06%) (p < 0.05). In contrast, the triple HGEAA mutant had a higher prevalence in Kisii (44.68%) than in Kisumu (26.31%). FGEAA is another triple mutant that exhibited almost the same prevalence in Kisumu (8.77%) and Kisii (8.51%). A single isolate exhibited a wild haplotype (SAKAA) in Kisumu (1.75%), but it was lacking in the Kisii region. For Pfdhfr, the triple IRNI is the most prevalent haplotype with a higher frequency in Kisii (98.08%) than in Kisumu (86.67%). Comparatively, the frequency of quadruple IRNL was higher in Kisumu (10%) than in Kisii (1.92%), while a double ICNI mutant (1.67%) and triple ICNL (3.33%) existed in the Kisumu region only.

Table 3.

Frequency and prevalence estimates of the Pfdhps and Pfdhfr haplotypes in P. falciparum isolates from western Kenya

			Frequencya (%)		Prevalenceb (%)			P-Value
Gene	Type	Haplotypes	Kisumu	Kisii	Kisumu	Kisii	Average
Pfdhps	Wild	SAKAA	1.49	3.73	1.75	0	0.88	0.740
	Single	SGKAA	1.49	1.86	1.75	2.13	1.94	0.986
	Double	SGEAA	66.38	47.54	70.18	51.06	60.62	0.044
		SGEGA	0	3.73	0	4.26	2.13	0.238
	Triple	HGEAA	23.14	39.38	26.31	44.68	33.99	0.048
		FGEAA	7.51	7.49	8.77	8.51	8.64	0.870
Pfdhfr	Wild	NCSI	0	0	0	0	0	–
	Double	ICNI	1.56	0	1.67	0	0.84	0.736
	Triple	ICNL	3.13	0	3.33	0	1.67	0.230
		IRNI	85.89	98.08	86.67	98.08	92.38	0.010
	Quadruple	IRNL	9.43	1.92	10.00	1.92	5.96	0.213
Pfdhps-Pfdhfr	Wild	SAKAA/NCSI	0	0	0	0	0	–
	Triple	SAKAA-IRNI	1.51	0	1.75	0	0.88	0.740
	Quadruple	SGEAA-ICNI	1.51	0	1.75	0	0.88	0.732
		SGKAA-IRNI	0	1.87	0	2.13	1.07	0.724
	Quintuple	SGEAA-ICNL	3.02	0	3.51	0	1.75	0.214
		SGEAA-IRNI	53.95	45.53	57.89	48.94	53.43	0.366
		SGKAA-IRNL	1.51	0	1.75	0	0.88	0.694
	Sextuple	SGEAA-IRNL	6.07	1.87	7.02	2.13	4.58	0.264
		SGEGA-IRNI	0	3.74	0	4.26	2.13	0.256
		FGEAA-IRNI	6.07	7.52	7.02	8.51	7.77	0.606
		HGEAA-IRNI	23.38	39.47	26.31	44.68	35.50	0.030
	Septuple	FGEAA-IRNL	1.51	0	1.75	0	0.88	0.722
		HGEAA-IRNL	1.51	0	1.75	0	0.88	0.742

^aFrequency is the proportion of parasites with a given haplotype, as estimated by the maximum likelihood estimation (MLE)

^bPrevalence is the proportion of infected individuals with at least one haplotype while adjusting for the confounding effect of mixed infections (MOI) using MLE

In the analysis of combined Pfdhps-Pfdhfr haplotypes, the quintuple SGEAA-IRNI mutant was the most dominant haplotype, with a higher prevalence in Kisumu (57.89%) than in Kisii (48.94%). The sextuple HGEAA-IRNI mutant was the second most prevalent haplotype, with a higher frequency in Kisii (26.31%) than in Kisumu (44.68%). A septuple FGEAA-IRNL haplotype was present in the Kisumu region only (1.9%, 1/57) with a triple SAKAA-IRNI haplotype (1.75%), quadruple SGEAA-ICNI haplotype (1.75%), and quintuple SGEAA-ICNL haplotype (3.51%). Sextuple haplotypes that occurred at low frequencies were SGEAA-IRNL (7.02% vs. 2.13%) and FGEAA-IRNI (7.02% vs. 8.51%) in the Kisumu and Kisii regions, respectively.

Regional clustering of genotypes and haplotypes

Hierarchical heatmap (Fig. 3) clustered resistance genotypes and haplotypes to SP in the Pfdhps and Pfdhfr genes into low, moderately low, moderately high, and high prevalence levels. At moderately low prevalence, 436H genotypes, along with triple HGEAA, sextuple HGEAA-IRNI, and quadruple IRNI haplotypes, are significantly more dominant in Kisii than in Kisumu. In contrast, the 164L and 516F genotypes, along with the quadruple IRNL and sextuple SGEAA-IRNL haplotypes, are more dominant in Kisumu than in Kisii at a low prevalence. At a moderately high prevalence, the double SGEAA haplotype was significantly more dominant in Kisumu than in Kisii. At high prevalence, 51I, 59R, and 108N in Pfdhfr and 537G and 540E in Pfdhps have reached or are approaching the state of fixation in the population, with the triple IRNI haplotype being more dominant in Kisii than in Kisumu.

Fig. 3.

Hierarchical heatmap illustrates the differential distribution of genotypes (A) and haplotypes (B) of resistance markers associated with sulfadoxine-pyrimethamine in the Pfdhps and Pfdhfr genes from P. falciparum isolates from the Kisumu and Kisii regions. The prevalence of genotypes and haplotypes varies from 0 to 100%, with significant differences in the proportions of resistance markers between the two regions indicated with an asterisk (*p < 0.05)

Discussion

This study compared the prevalence and selection of resistance molecular markers in P. falciparum isolates (N = 112) associated with the use of SP in endemic and epidemic populations with dissimilar transmission rates in western Kenya. Clearly, gender, age, parasitaemia, P. falciparum DNA concentration, temperature, and MOI did not vary in the two populations; however, the positivity rate was significantly greater in Kisumu at 18.64% than in Kisii at 6.76% (p < 0.05). The variation in positivity rates can be attributed to factors like transmission intensity, vector ecology, and drug resistance, alongside other elements linked to the prevalence rate of about 20% in the Lake endemic areas compared with 1% in the highland epidemic-prone region [2]. To comply with WHO guidelines, Kenya’s Ministry of Health sanctioned the use of SP as the second-line treatment from 1983 to 1997, as the first-line treatment from 1998 to 2007, mandated a minimum of three doses for IPTp nationwide from 2001 to 2008, and recommended at least three doses of IPTp in endemic areas only from 2009 to the present. [2, 5, 7]. The use of WHO-validated molecular markers of SP resistance and the reporting of newly emerging alleles could provide critical information for the molecular surveillance of drug resistance and the evolution of parasites [3]. Therefore, this study hypothesizes that differential drug pressure resulting from varying SP use contributes to the emergence, evolution, and spread of resistant P. falciparum parasites in the two study regions.

Sulfadoxine resistance markers in Pfdhps

Genotyping analysis of sulfadoxine resistance markers in Pfdhps revealed four non-synonymous mutations at codon 436H/F (37.5%), 437G (99%), 516F (4.8%), 540E (97.1%), and mixed A581G genotype (4.3%) in western Kenya. These findings show that the study isolates had four of the five validated mutations in Pfdhps, namely, 436A/F, 437G, 540E, 581G, and 613 T/S [3]. At codon position 436, the mutant alleles detected were 436H (28.8%), 436F (8.7%), and 436H mixed genotypes (5.8%), accounting for 43.3% of the mutant alleles in the population. The 436H mutation was first reported in western Kenya in isolates collected from pregnant women between 2008 and 2012, with a prevalence of 28% [23], and was later detected among patients in Siaya County with a prevalence of 14% [24]. A recent study conducted in Siaya County among children documented an increase in the prevalence of the 436H mutation to 28%, indicating a rising trend in the population [6]. A more recent study conducted in Bungoma County reported a prevalence of 21.6% for the emergent 436H mutant in western Kenya [25]. Additionally, the detection of the validated 436F allele at a prevalence of 8.7%, with no significant difference observed between the two regions, implies a continued presence of this rare allele in western Kenya, as reported in previous studies [23, 26, 27]. The near-fixation of 437G (99%) and 540E (97.1%) mutants in both study regions highlights the high level of sulfadoxine resistance in western Kenya and across the country, as reported by recent studies [4, 6, 10, 25]. The identified L516F mutation (8.8%) in the Kisumu region was reported previously in an efficacy trial of artesunate-sulfadoxine-pyrimethamine (AS-SP) in eastern Sudan as a novel mutation [28]. Apparently, this same mutation is emerging in the Kisumu region due to continued drug pressure on malaria parasites.

Pyrimethamine resistance markers in Pfdhfr

Genotyping analysis of pyrimethamine resistance markers in Pfdhfr revealed the presence of four non-synonymous mutations at codons 51I (100%), 59R (97.3%), 108N (100%), and 164L (7.1%). These mutations are well-documented and validated as markers for pyrimethamine in the malaria parasite population worldwide [3]. Other consensus studies in western Kenya documented an increasing and near-fixation prevalence of 51I, 59R, and 108N mutations, including among children under 12 months in Siaya County [6], school children in Kisumu and Kakamega counties [17], school children in 15 counties across the country [10], and residents of Bungoma County [25]. A study conducted in coastal Kenya among pregnant women reported a similar trend of high prevalence of these triple 51I, 59R, and 108N mutants at rates approaching the fixation level [29]. The complete or near-fixation of the 51I, 59R, and 108N mutations indicates sustained drug pressure for the evolution and spread of these resistant malaria parasites among populations in both endemic and epidemic-prone regions. These findings are similar to those of a study conducted on P. falciparum isolates imported to China from 11 countries in the West Africa region, which demonstrated an increasing prevalence of the Pfdhfr 51I, 59R, and 108N alleles from about 90% in 2012 to the fixation level in 2022 [30]. Evidently, 164L is an emerging mutation in Kenya and the larger East African region because it accounts for the occurrence of super-resistant parasites to pyrimethamine in combination with 51I, 59R, and 108N mutations in their fixation state [6, 17, 31]. The increase in the prevalence of these mutants raises a serious concern as it results in treatment failure of SP, an implication of policy change on its use as IPTp.

Haplotypes for sulfadoxine-pyrimethamine resistance

Haplotype analysis of validated molecular markers in Pfdhps, Pfdhfr, and their combinations revealed the extent of SP-resistant P. falciparum parasites in the population. In Pfdhps, the wild-type haplotype was rare (~ 1%), while the double SGEAA haplotype was the most dominant (60.62%), followed by the emergent triple HGEAA (33.99%) and FGEAA (8.64%) alleles, a trend demonstrated in a study among children in Siaya [6]. In Pfdhfr, the haplotype analysis established that wild NCSI (0%), double ICNI (0.84%), triple ICNL (1.67%), and quadruple IRNL (5.96%) alleles are rare in the population, whereas the triple IRNI (92.38%) is the most dominant allele, which is approaching the fixation state in the population. Recent studies conducted on P. falciparum isolates from Siaya County in western Kenya [6], Msambweni County in the coastal region of Kenya [29], and imported cases from West Africa to China [30] indicated a similarly high prevalence of the triple IRNI allele approaching the fixation state in the population as observed in this study. In the analysis of Pfdhps-Pfdhfr combined alleles, the quintuple SGEAA-IRNI haplotype (53.43%) and the sextuple HGEAA-IRNI haplotype (35.50%) highlighted the increasing trends of SP resistance in the population as demonstrated by previous studies conducted [6, 10, 17, 32]. Genotype analysis of field isolates from 30 countries in Africa demonstrated that the quintuple SGEAA-IRNI haplotype is predominant in East Africa (58.1%), Southern Africa (89.3%), and the Horn of Africa (86.4%) [33]. Evidently, the sextuples SGEAA-IRNL (4.8%) and FGEAA-IRNI (7.7%) appear to be emerging haplotypes owing to the increasing proportions of mutants in the segregation sites at S436F/H and I164L alleles [6, 31]. Other combined haplotypes, such as triple SAKAA-IRNI, quadruple SGEAA-ICNI, quintuple SGEAA-ICNL, and septuple FGEAA-IRNL alleles, are rare in the population, as their respective prevalence rates across the 15 counties in Kenya include low-frequency mutations [10]. The presence of sextuple and septuple haplotypes in western Kenya underscores the need for continuous molecular surveillance of P. falciparum, as they signal the emergence of super-resistant strains, reduced therapeutic efficacy of SP for IPTp, and the necessity for effective alternative treatments.

Variation in prevalence, diversity, and selection dynamics

Comparative analysis of genotypes and haplotypes revealed significant variation in their prevalence, diversity, and selection. Particularly, 436F/H mutations together with the associated triple HGEAA and sextuple HGEAA-IRNI haplotypes, as well as 516F and 164L mutations, exhibited significant variation between the Kisii and Kisumu regions. Among the five validated Pfdhps mutations, the 436H mutation was significantly more prevalent in Kisii than in Kisumu (p < 0.05). The prevalence of the 436H mutation has increased in western Kenya since its initial detection in a 2008–2009 study among pregnant women at 3.8% [23], and in subsequent studies in Siaya, which reported 14% [24], 28% [6], and 34% [34]. The 516F mutation, previously reported in Sudan, a region where artesunate-sulfadoxine-pyrimethamine (AS-SP) is a first-line antimalarial drug [28], suggests that drug pressure may account for its occurrence in the Kisumu region, where pregnant women receive at least four doses of SP as IPTp. Furthermore, the significantly higher prevalence of the validated 164L mutation in Kisumu than in Kisii highlights the emergence of super-resistant parasites carrying the septuple FGEAA-IRNL haplotype, as indicated in a study conducted in 15 counties, predominantly in endemic regions [10]. The regional variations in genotype and haplotype distributions point to differences in drug pressure between the two study regions.

Limitations of the study

While this study offered significant insights into antifolate resistance in both epidemic and endemic regions of western Kenya, it also has some limitations. One limitation is the small sample size, which may reduce the external validity of the findings and limit their generalizability to P. falciparum populations in the study regions. The small sample sizes reduces the likelihood of detecting rare mutations present at frequencies of 1, 2, and 5% in the population by approximately 60, 38, and 7%, respectively [12]. However, rigorous and analytical procedures for sampling, molecular processing, sequencing, quality control, and data analysis ensured the internal validity and reliability of the observed genotypes and haplotypes [12]. Another limitation is that the cross-sectional nature of the study does not establish the temporal sequence of drug exposure and resistance phenotypes needed for inferring causal relationships. To mitigate this limitation, this study used WHO-validated molecular markers of SP resistance in the Pfdhps and Pfdhfr genes, which have been identified through in vitro, and in vivo therapeutic efficacy studies [3]. Despite the limitations of this study, the identification of a high prevalence of validated resistance markers for SP emphasizes the need for ongoing molecular surveillance, the creation of effective alternative treatments, and the optimal use of existing therapeutic regimens.

Conclusion

Genotyping of resistance markers in the Pfdhps and Pfdhfr genes associated with SP treatment revealed regional variation in the prevalence and selection of genotypes and haplotypes between the Lake endemic and highland epidemic-prone areas. The fixation state of 437G and 540E mutations (double SGEAA mutant) in Pfdhps and N51I, C59R, and S108N mutations (triple IRNI mutant) in Pfdhfr indicates sustained drug pressure and strong positive selection of these resistance markers in the two study regions. However, comparative analysis of molecular markers in Pfdhps showed a higher prevalence and selection of the 436H mutant and its associated triple HGEAA and sextuple HGEA-IRNI haplotypes in the Kisii region. In contrast, genotyping also indicated a higher prevalence and selection of Pfdhfr-164L and Pfdhps-516F in Kisumu. Therefore, further investigation is necessary to establish the selection drivers for 436H, 516F, and 164L genotypes and associated sextuple HGEAA-IRNI and triple HGEAA haplotypes, differentially observed in the two study regions. Additionally, the study advocates for ongoing monitoring of sextuple and septuple haplotypes in western Kenya since they indicate the emergence, evolution, and propagation of highly resistant P. falciparum strains, which diminish the therapeutic effectiveness of SP for IPTp, underscoring the need for the development and use of viable alternative treatments.

Abbreviations

ASCH

Ahero sub-county hospital

DNA

Deoxyribonucleic acid

DBS

Dried blood spot

EHC

Eramba Health Centre

ICEMR

International Centre of Excellence for Malaria Research

IPTp

Intermittent preventive treatment in pregnancy

MEGA

Molecular Evolutionary Genetics Analysis

MSCH

Marani sub-County Hospital

MUSCLE

Multiple Sequence Comparison by Log-Expectation

MUSERC

Maseno University Scientific and Ethics Review Committee

MOI

Multiplicity of infection

NACOSTI

National Commission for Science, Technology and Innovation

Pfdhps

Plasmodium falciparum dihydrofolate synthetase

Pfdhfr

Plasmodium falciparum genes, dihydrofolate reductase

RT/qPCR

Real-time/quantitative polymerase chain reaction

PfvarATS

Plasmodium falciparum var gene acid terminal sequence

SNP

Single nucleotide polymorphism

SP

Sulfadoxine pyrimethamine

WHO

World Health Organization

Author contributions

JB, CO, DZ, GZ, HA, JG, M-CL, and GY conceived and designed the study. JB and JO were responsible for sample collection and initial laboratory processing under the supervision of HA and JG. DZ and CW conducted additional sample processing and amplicon sequencing. JB, DZ, and M-CL performed data analysis and visualization. JB drafted the original manuscript, and all authors reviewed, revised, and approved the final version.

Funding

This study was supported by grants U19AI129326 and D43TW001505 from the National Institutes of Health.

Data availability

The dataset is available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was granted by the Maseno University Scientific and Ethics Review Committee (MUSERC) (Reference Numbers: MUSERC/01397/24 & MUERC/00660/19), and a research license was issued by the National Commission for Science, Technology and Innovation (NACOSTI) (Reference Number: 219416). In compliance with research ethics, the study sought written informed consent from adult participants and guardians in the case of children, permitted them to volunteer in their participation, and assured them of the confidentiality and privacy of their collected data.

Consent for publication

All authors consented to this publication.

Competing interests

The authors declare no competing interests.