Many sites in the Arabian Peninsula turned into “malaria-free” following intensive control efforts mounted over a relatively short period of time
. In contrast, in Yemen and southwest Saudi Arabia malaria remains resilient to control efforts. Some of the challenges facing the prospect of elimination, in these sites, is the introduction of drug resistance parasites via asymptomatic carriers. The present study examined the extent of genetic diversity and distribution of drug-resistant genes among of P. falciparum in some major foci in Yemen.
A high degree of diversity of non-drug-resistant genes (msp-2 and pfg377 and microsatellites in Chr 8) was seen among P. falciparum in Yemen. The He index ranged between 0.51 and 0.89, suggesting a high rate of transmission and a bigger parasite population size than anticipated in this region. Both allelic diversity and prevalence of mixed-genotype infection were high, 57% of the isolates harboured more than one allele at any of the examined loci, with a mean of 1.8 genotypes per infected individual. Assuming that each clone is readily transmissible to mosquito, the rate of cross-mating and subsequent recombination is expected to be high
. This agrees with the observed random association between loci and absence of geographical differentiation among P. falciparum in the three sites (Dhamar, Hodeidah and Taiz) (F
< 0.05). Such a pattern of parasite structure is similar to that seen among P. falciparum populations in Africa
[29, 30]. These findings imply presence of a large effective population size (Ne) of P. falciparum in Yemen, as there is a direct relationship between the level of diversity and Ne[31–33]. This represents a challenge to control efforts in the whole region, as parasites can readily migrate from Yemen into other sites in the Arabian Peninsula, such as Oman
 and Saudi Arabia
, where transmission has been interrupted. In addition to local foci, imported malaria, via numerous migrations from Africa/Asia, can enrich diversity of parasites in the Arabian Peninsula. Over the past two decades political instability has resulted in mass movement of people from the Horn of Africa (Somalia, Eritrea and Ethiopia) into Yemen and Saudi Arabia
. Travellers from malaria-endemic areas can carry long-lasting, transmissible, asymptomatic parasitaemia
[36, 37]. Multiple introductions of distinct genetic sources could in part explain the high level of parasite diversity. However, the lack of genetic differentiation between parasites in different sites suggests that the prevailing epidemiological and demographical factors in Yemen are favorable to parasite dispersal. Imported malaria can, therefore, jeopardize the current success of control in the region despite the presence of a strong public health infrastructure. For example, countries such as Oman experience regular epidemics due to imported malaria
. To sustain malaria control and achieve elimination in the peninsula efforts should be directed to foci with on-going transmission, such as Yemen and Saudi Arabia.
The distribution of drug resistant genotypes seen in the present study fits with the history of anti-malarial usage, and possible migration of parasites from Africa and Asia into Yemen. The emergence of CQR in Yemen in 1989 Yemen
[2, 7, 9, 38, 39], coincided with its appearance in close countries in East Africa
[12, 16, 40] and Asia, Saudi Arabia
, Iran and Pakistan
[3, 41–46]. The presence of two distinct pfcrt genotype, CVIET (89%) and SVMNT (4%) suggests that CQR in Yemen has evolved from at least two different origins. Globally, there are five genotypes of pfcrt CVIET, SVMNT, SVIET, CVMNT and CVTNT. The CVIET genotype is predominant in Africa
, while the SVMNT is common in Asian countries close to the Arabian Peninsula, such as Iran
 and Pakistan
[41, 42]. This agrees with the hypothesis of frequent gene flow of African/Asian parasites into Yemen, and suggests that the CQR genotype (CVIET) is probably introduced via Africa, while genotype SVMNT has come from Asia. Analysis of microsatellites around pfcrt will allow further investigation to the origin of pfcrt resistance genotypes
Similar to pfcrt, high prevalence of pfmdr1 mutations was seen in Yemen. Mutations in codon 86Y has been linked to CQR, while those at codons 184, 1034, 1042 and 1246 are related to mefloquine (MQ), amodiaquine (AQ), halofantrine (HF) and quinine resistance
[32, 50, 51]. In the present study, 86 (80%), 107 (99%) and 108 (100%) of the examined P. falciparum isolates carried alleles 86N and 184F and 1246D, respectively. Selection of pfmdr1 86N and 184F in recrudescent P. falciparum parasites following lumefantrine suggested a possible role of these alleles in the development of tolerance/resistance to lumefantrine
[37, 52]. Increased prevalence of pfmdr1 86N, 184F and 1246D haplotype was seen in Zanzibar after several years of extensive use of artesunate-amodiaquine (ASAQ)
. A recent study has confirmed that P. falciparum parasites with the pfmdr1 N86/184F/D1246 haplotype can withstand 15-fold higher lumefantrine blood concentrations compared to those with the 86Y/Y184/1246Y haplotype
. The relatively high prevalence of the wild-type allele 86 N (80%) in the present study cannot be explained by withdrawal of CQ in Yemen, as the drug is still in use
, and pfcrt76T remains at a prevalence of 89%. It is therefore, more likely that the pattern of pfmdr1 polymorphisms, seen in Yemen, is due to resurgence in use of artesunate-SP combination. In addition to codon 86 and 184, mutations at the codons 1034 and 1042 were also high in Yemen; nonetheless no mutation was seen at codon 1246. Further longitudinal surveillance of pfmdr1 mutations, coupled with in vivo testing, should be considered to examine the impact of these polymorphisms on efficacy of ACT in Yemen.
With regard to genes controlling P. falciparum response to SP, a previous study has detected the mutant allele dhfr-59R among 4 out of 99 isolates in Lahj governorate, southeast of Yemen, and no information was given on other codons 51, 108 and 164
. However, the present study revealed high prevalence of mutations at codons 51 and 108 of the dhfr, while no mutation was seen at codon 59. This inconsistency can be attributed to the probability that some of the samples examined by Mubjer et al.
 may have been collected from expatriates. If mutation 59R exists among local parasites in Yemen, it would have increase in frequency following recent shift to us of artesunate + sulphadoxine-pyrimethamine (SP) as first line therapy
. Epidemiological evidence suggests that dhfr mutations starts at codon 108, yielding low level resistance to pyrimethamine
, then the resistance increases with acquisition of extra mutations at codons 51 or 59 and 164
. Analysis of evolution pattern of dhfr genotypes shows that the high resistance genotype (51I, 59R, 108 N) accumulates mutations in sequence. The most common orders of mutation are 108 N, 59R, 51I or 108 N, 51I, 59R. These cumulative mutations in dhfr restore the parasite fitness and allow it to overcome the effect of the drug
. dhfr mutations appear earlier than dhps mutations in the development of resistance to SP combined effect
, and P. falciparum has to gain mutations in dhfr and dhps genes to develop resistance to SP, thus absence of dhps mutation in Yemen suggest that the SP will be effective for some years to come. Nonetheless, the occurrence of high level resistance dhfr and dhps genotypes linked to SP failure in areas close to Yemen, such as Iran and Pakistan (100%), the Horn of Africa; Sudan (72.7%) and Ethiopia (84.5%)
[60–64], necessitate the need for regular surveys to monitor possible emergence of these mutations into Yemen for timely change in drug policy.