Selection of pfdhfr/pfdhps alleles and declining artesunate/sulphadoxine-pyrimethamine efficacy against Plasmodium falciparum eight years after deployment in eastern Sudan

Malaria continues to affect over two million people worldwide with more than 600,000 reported deaths in 2010, most of these in sub-Saharan Africa [1]. There are increasing efforts to control, eliminate or eradicate malaria in many countries [2], however these are hindered by ineffective treatment and vector control strategies that cannot be sustained in economically disadvantageous countries.

Malaria in Sudan causes morbidities and mortality among all age groups and was responsible for more than 95,000 hospital admissions in 2011 [3]. Plasmodium falciparum causes over 95% of malaria infections and drug resistance to CQ and SP is well established [4‐7]. The wide-scale deployment of artemisinin combination therapy (ACT) across malaria-endemic zones is intended to counteract the emergence and spread of resistance [8]. Artesunate-sulphadoxine/pyrimethamine (AS/SP) and artemether-lumefantrine (AL) became the first- and second-line of treatment, respectively, for uncomplicated P. falciparum malaria in 2004 [9]. Recently, the efficacy of AS/SP has been questioned among practicing physicians with some patients returning with recurring infections. The in vivo efficacy of SP in eastern Sudan was reported at 82% [10], similar to Tanzania (86%) [11] and high levels of resistance to SP in East Africa [12, 13] may render the short half-life artesunate partner unprotected. Furthermore the emergence of slow clearing parasites following artesunate treatment in South East Asia [14, 15] has prompted closer monitoring of the efficacy of ACT and necessitated extensive research on molecular markers of resistance to artemisinins and partner drugs.

Specific haplotypes of the pfcrt gene at positions 72 – 76 (CVMNK; wild haplotype and CVIET; mutant Asian/African haplotype; mutant SVMNT South American haplotype) are well-established markers for resistance to CQ [16, 17]. In eastern Sudan the prevalence of the pfcrt-76T increased from 54% in the early 1990’s [18, 19] to more than 90% in the 2000’s [7, 20, 21]. However, the pfcrt-76K allele has returned to high prevalence in parts of Africa where CQ was discontinued, such as Malawi [22, 23], Kenya [24, 25], Senegal, [26] Tanzania [27, 28] and Mozambique [29].

Pfmdr1 haplotypes N86, 184F and 1246D have been selected by artemether-lumefantrine in eastern Sudan [30] as well as several parts of Africa [31‐34]. These haplotypes were also found to increase the risk of treatment failure with AL [35]. Amplification of pfmdr1 which is associated with decreased sensitivity to artemisinins in South East Asia [36, 37] has recently been observed in eastern Sudan [30] and in neighbouring east African countries [38, 39]. Therefore, pfmdr1 may continue to be under the influence of ACT currently employed in Sudan.

Resistance to pyrimethamine is due to a set of sequential mutations in the dihydrofolate reductase gene (dhfr) [40, 41]. The serine to asparagine at position 108 is essential with additional mutations at positions N51I, C59R and I164L leading to higher resistance levels [42, 43]. Mutations in the pfdhps gene (S436A, G437A, K540E, A581G and A613S) encoding the target enzyme for sulphadoxine are also predictors of treatment failure in vivo[44]. It has been recommended that SP should be discontinued in areas where the prevalence of the pfdhps-540E allele is more than 50% [12]. In eastern Sudan, the prevalence of the pfdhfr CI CN I haplotype at codons 50, 51, 59, 108 and 164 has been reported at 90% [20]. The pfdhps haplotype SGE AA (436/437/540/581/613) was higher than 80% in the current study area prior to the introduction of AS/SP [20]. These two alleles may be maintained in the population by the strong selection pressure imposed by SP as an ACT component.

In this study, the prevalence of molecular markers in four genes (pfcrt, pfmdr1, pfdhfr and pfdhps) associated with drug resistance following the wide scale deployment of AS/SP in eastern Sudan is investigated and an updated clinical efficacy of AS/SP is reported.

A total of 453 patients who came to the clinic with fever and reporting any malaria-like symptoms were screened for P. falciparum infection. Sixty-eight patients fulfilled the enrolment criteria and consented to participate in the study, sixty-three (92.6%) of whom completed 28-days of follow up. Patient age, weight, temperature, parasite density and gender are summarized in Table 1. Mean time to resolving clinical symptoms such as fever (axillary temperature above 37°C), malaise and nausea was 2.7 days (range 2 – 7 days) and was not a risk for failing treatment.

On day 1; 12.7% (8/63) patients were febrile and 20.6% (13/63) patients were parasitaemic. However, all patients cleared their parasitaemia by day 3 with one patient remaining febrile (temperature 38°C). Gametocytaemia was observed in 3.2% (2/63) of patients at recruitment. Adverse effects of the drugs such as nausea, headache, cough, dizziness, abdominal pain and diarrhoea were reported in 34.9% (22/63) of patients between day 19 and 28. In the intent-to-treat analysis where all patients who were enrolled were included in the analysis, the efficacy of AS/SP was 85.3%. While in the per-protocol analysis the efficacy was 90.5%. Two patients failed treatment at day 21, one failed treatment at day 26 and two failed treatment at day 28. The PCR corrected efficacy was 93.6% where 4/5 of treatment failures were classified as recrudescing parasites. Factors such as pre-treatment parasite density, axillary temperature and age did not differ significantly between those who failed treatment and those who were successfully treated.

Plasmodium falciparum carrying mutations associated with drug resistance may be less fit to survive in the absence of drug pressure. The continuous use of CQ monotherapy for decades has lead to almost saturation of the P. falciparum population in eastern Sudan carrying the mutant pfcrt allele CVIET [7]. Lower drug pressure is reflected by more wild type pfcrt where transmission is higher and perennial in central Sudan [53]. Moreover SP has been used intensively in patients where CQ failed. This has resulted in the high prevalence of the pfdhfr/pfdhps mutants (51I/108N/437G/540E) in the late 1990’s and early 2000’s [18, 20, 54]. The adoption of AS/SP in eastern Sudan in 2004 has lead to considerable decrease in malaria cases in the country [9, 55‐57]. The clinical efficacy of SP monotherapy in Sudan was approximately 68% prior to introduction of AS/SP [58]. Therefore, the wisdom of deploying this SP containing combination has been debated among the malaria research community in Sudan e.g. [59]. The base-line efficacy of AS/SP in eastern Sudan has been reported between 98 - 100% [56, 60‐62], while we have reported a base-line PCR corrected efficacy of 93% [57]. In the present study the in vivo per-protocol PCR corrected efficacy of AS/SP over 28 days is 90.5%. All patients cleared their microscopic parasitaemia by day 3. Approximately 10% of patients exhibited recrudescing parasites as confirmed by msp1 and msp2 genotyping during the study. It is important to emphasize the relatively lower transmission season in eastern Sudan compared to other parts of East Africa. Therefore, to definitely establish whether these post-treatment infections were true recrudescence or re-infections requires genotyping at more polymorphic loci.

A limitation to our data is the lack of parallel PCR data for days 3 and 14. Sub-microscopic parasites at these time points have been detected in previous studies from eastern Sudan [7, 30]. Detection of parasites on day 3 is an early warning of slow clearance by artemisinin derivatives observed in South East Asia [14, 15]. A decline in the efficacy of ACT e.g. AL and DHA/PIP has been reported from Kenya [63]. Sub-microscopic parasitaemia has previously been observed following AL treatment in eastern Sudan [30]. While an artemisinin resistant phenotype has not been reported in Africa [64], the history of anti-malarial resistance suggests that artemisinin resistance is likely to spread to East Africa despite the global efforts in its containment.

In the present study, the prevalence of molecular markers eight years after the adoption of ACT is compared to previous reports from the same study area of eastern Sudan. In this area IPTp is recommended for all pregnant women, as in other parts of Sudan.

The prevalence of a double mutant pfdhfr allele (51I/108N) has reached 100%, which clearly demonstrates the effect of drug pressure on this locus. The observation that contrasts with other African countries is the slow evolution of pfdhfr in this region where the double mutant 51–108 allele has widely spread while the wild type C59 has been maintained. However it is in agreement with a recent report from Yemen [65]. The 59R is associated with increased resistance to pyrimethamine, therefore absence of this mutation may be a good indicator for the in vivo efficacy of SP in this region. The pfdhfr 59R has been previously reported at low frequency in Khartoum, central Sudan in 1996/1997 [66] and Nuba mountains in western Sudan in 2003 [67]. However the 51I/108N haplotype has been observed at low frequencies in Kenya and Cameroon in the early 2000s where the pfdhfr-51I/59R/108N haplotype currently predominates [68, 69]. These studies employed microsatellite markers which demonstrated the divergence of origin of the double (51I/108N) and triple mutants within the same geographic location. An earlier study in eastern Sudan has shown that microsatellites around pfdhfr were stable over four consecutive seasons [18]. Thus, the pfdhfr allele 51I/108N may have been imported into this area earlier and has been maintained within a limited parasite population in a low transmission area of eastern Sudan. However, it would be of interest to examine pfdhfr sequences and microsatellite markers from other regions of Sudan with different transmission intensities to test this hypothesis.

Interestingly in this study, a small decrease in pfdhps-437G (Fisher’s exact p = 0.0134) is observed compared to previous reports [20]. The 581G allele has significantly increased to 33% and has previously been associated with SP failure in eastern Sudan [20] emphasizing the importance of close monitoring of the parasite population in this region. For the first time, a particular pfdhps allele, SGEGA, is significantly associated with treatment failure in patients receiving AS/SP for clinical malaria. However of 23 pretreatment infections harbouring SGEGA, only four recrudesced. The increase of SGEGA has been previously reported from Tanzania in cross sectional surveys following several years of SP use [27]. The potential of this haplotype as a single marker of drug efficacy for this anti-malarial combination should be further investigated.

As previously reported from other African countries [22, 24, 28], there is a significant decrease in CQR associated genotypes pfcrt-76T/pfmdr1-86Y. Wild type genotypes at these two loci have significantly recovered, which provides supporting evidence that CQ use has been successfully reduced [70]. Interestingly the multiple copy pfmdr1 genotype previously reported from eastern Sudan [30] was not observed here. This genotype is associated with resistance to mefloquine (MQ) in South East Asia [50] and may not spread in the absence of MQ pressure, which is the case in Sudan. In addition pfmdr1 amplification has failed to be established and spread in Africa, possibly due to lack of the use of MQ across the continent.

In Sudan SP is recommended for iPTp and the availability of SP as an over-the-counter drug maintains the pressure on the parasite population. The efficacy of AS/SP is reaching the current threshold for anti-malarial policy change of 90%. Continuous use of SP as a partner to artesunate may be the cause of declined AS/P efficacy. In addition discontinuation of SP in areas where the population prevalence of pfdhps-540E is above 50% is also recommended [12, 71]. Monitoring of SP resistance associated loci, and in particular the SGEGA pfdhps haplotype identified here, is required to provide informed treatment guidelines in the near future. In addition the emergence of slow parasite clearance in South East Asia warrants close monitoring of the efficacy of ACT and molecular markers associated with resistance to partner drugs.