Background
During the past 20 years, many strains of
Plasmodium falciparum have become resistant to chloroquine and other anti-malarial drugs [
1]. One strategy for reducing malaria prevalence is the use of drugs in combination. Drug combinations help prevent the development of resistance to each component drug and reduce the overall transmission of malaria [
2]. In response to increasing chloroquine resistance, Senegal switched in 2004, to sulphadoxine-pyrimethamine with amodiaquine as the first-line therapy. In 2006, the Senegalese National Malaria Control Programme recommended artemisinin-based combination therapy (ACT) as the first-line treatment for uncomplicated malaria. The combination sulphadoxine-pyrimethamine and amodiaquine treatment was changed to artemether-lumefantrine and artesunate-amodiaquine. Since 2006, more than 1.5 million ACT-based treatments have been administered in Senegal [
3]. In 2006, the Senegalese National Malaria Control Programme also recommended testing for all suspected cases of malaria with the
P. falciparum histidine-rich protein 2 (PfHRP2)-based rapid diagnostic test (RDT). Since this time, ACT use has been restricted to confirmed malaria cases to reduce drug pressure. In 2009, 184,170 doses of ACT were dispensed in Senegal [
4].
Malaria is transmitted in Dakar and its surrounding suburbs with a spatial heterogeneity of the human biting rate, which ranged from 0.1 to 250 bites per person per night during the rainy season from 2007 to 2010 [
5,
6]. In 2008 to 2009, the human biting rate was 0.7 bites per person per night during the rainy season in Médina, a district of the south of Dakar [
5]. In 2008, the
P. falciparum prevalence varied from 0.9% to 7.4% in asymptomatic women and children in Dakar [
7]. Morbidity in public health facilities decreased from 17.9% in 2007 to 2.6% in 2008 in Dakar [
8].
Since the introduction of ACT, there have been very few reports on the level of resistance of P. falciparum to anti-malarial drugs. To determine whether parasite susceptibility has been affected by the new anti-malarial policies, an ex vivo susceptibility study was conducted on local isolates from Dakar obtained from the Centre de santé Elizabeth Diouf (Médina, Dakar, Senegal). The malaria isolates were assessed for susceptibility to chloroquine (CQ), quinine (QN), monodesethylamodiaquine (MDAQ), the active metabolite of amodiaquine, mefloquine (MQ), lumefantrine (LMF), dihydroartemisinin (DHA), the active metabolite of artemisinin derivatives and doxycycline (DOX).
In addition, the prevalence of genetic polymorphisms in genes associated with anti-malarial drug resistance was evaluated. The genes of interest included
P. falciparum CQ resistance transporter (
pfcrt) for CQ [
9],
P. falciparum dihydrofolate reductase (
pfdhfr) for pyrimethamine [
10],
P. falciparum dihydropteroate synthase (
pfdhps) for sulphadoxine [
11] and
P. falciparum multidrug resistance 1 (
pfmdr1) for mefloquine resistance [
12] and potentially for quinoline resistance [
13,
14].
The
pfcrt gene was first identified in 2000 [
9]. To date, at least 20 point mutations have been described [
9,
15,
16], but only one is the reference mutation (K76T), which is a marker of the CQ-resistant phenotype. This mutation is often associated with other mutations in the
pfcrt gene, whose role is not yet defined. The odds ratio (OR) for CQ failure associated with the 76T mutation was 2.1 (95% confidence interval: 1.5-3.0, meta-analysis of 13 studies) for a 14-day follow-up and 7.2 (95%CI: 4.5-11.5, meta-analysis of 12 studies) for a 28-day follow-up [
17]. However, the existence of CQ-susceptible strains associated with the 76T mutation suggests that other genes could be involved in the resistance to chloroquine.
The S108N mutation in the
pfdhfr gene is associated with resistance to anti-folate drugs [
18]. The OR for sulphadoxine-pyrimethamine failure associated with S108N was 3.5 (95%CI: 1.9-6.3, meta-analysis of 10 studies) for a 28-day follow-up [
17]. The additional mutations N51I, C59R or I164L increase the level of
in vitro resistance to anti-folate drugs and sulphadoxine-pyrimethamine. The OR values for codon 51 and 59 single mutants were 1.7 (95%CI: 1.0-3.0) and 1.9 (95%CI: 1.4-2.6), respectively [
17]. The triple mutation (51 + 59 + 108) increases the risk of
in vivo resistance to sulphadoxine-pyrimethamine by 4.3 (95%CI: 3.0-6.3, meta-analysis of 22 28-day studies) [
17].
Sulphones (dapsone) and sulphonamides (sulphadoxine) are inhibitors of the
P. falciparum DHPS [
19]. The mutations S436A, S436F, A437G and K540E are involved in resistance to sulphadoxine [
11]. The single mutation A437G and the double mutation A437G + K540E increase the risk of
in vivo resistance to sulphadoxine-pyrimethamine by 1.5 (95%CI: 1.0-2.4, meta-analysis of 12 studies) and 3.9 (95%CI: 2.6-5.8, meta-analysis of 10 studies), respectively [
17].
The quintuple mutant of
pfdhfr (codons 51 + 59 + 108) plus
pfdhps (codons 437 + 540) increases the risk of
in vivo resistance to sulphadoxine-pyrimethamine by 5.2 (95%CI: 3.2-8.8, meta-analysis of 3 studies) [
17].
Pfmdr1, which encodes a 162 kDa protein named
P. falciparum homologue of the P-glycoprotein (Pgh1), is located on chromosome 5. Field work has shown that the predictive value for CQ resistance and point mutations in the
pfmdr1 sequence resulting in amino acid changes varies depending on the geographic area [
20,
21]. Five point mutations have been described: N86Y, Y184F, S1034C, N1042D and D1246Y. Point mutations, most notably N86Y, have been associated with a decrease in the CQ susceptibility [
22]. However, in some of these epidemiological studies, the number of CQ-susceptible samples is too limited to provide a statistically meaningful analysis [
21,
23]. Using precautions, no relationship or only weak relationships are established between CQ resistance and mutations in
pfmdr1 in
P. falciparum[
24]. However, the risk of therapeutic failure with CQ is greater for patients harbouring the 86Y mutation, with an OR of 2.2 (95%CI: 1.6-3.1) for a 14-day follow-up and 1.8 (95%CI: 1.3-2.4) for a 28-day follow-up [
17]. The combination of
pfmdr1 86Y and
pfcrt 76T increases the risk of
in vivo resistance to CQ by 3.9 (95%CI: 2.6-5.8, meta-analysis of 5 studies) [
17].
In addition, the risk of therapeutic failure with amodiaquine is greater for patients harbouring the 86Y mutation with an OR of 5.4 (95%CI: 2.6-11.2, meta-analysis of six studies) [
17]. This mutation increases the risk of failure with amodiaquine plus sulphadoxine-pyrimethamine by 7.9 [
25].
It has been shown through heterologous expression that
pfmdr1 mutations at codons 1034 and 1042 abolish or reduce the level of resistance to mefloquine [
26]. Moreover, transfection with a wild-type
pfmdr1 allele at codons 1034, 1042 and 1246 confers mefloquine resistance to susceptible parasites [
27]. However, mutations at codons 1034, 1042 and 1246 of
Pfmdr1 in
P. falciparum isolates are not sufficient to explain variations in mefloquine susceptibility [
28]. Analyses of
P. falciparum isolates showed an association between mutation at codon 86 and an increase in susceptibility to mefloquine, halofantrine or artemisinin derivatives [
12,
29,
30].
Discussion
This report describes the evaluation of the ex vivo susceptibility of P. falciparum isolates, taken from patients in Dakar, to seven standard anti-malarial drugs and the prevalence of several molecular markers involved in anti-malarial drug resistance. The majority of the patients, recruited at the Centre de santé Elizabeth Diouf (Médina, Dakar) from August 2010 to January 2011, said that they did not leave Dakar and its surrounding suburbs during the month preceding their malaria attack.
The prevalence of isolates with reduced susceptibility to MQ remains high (62.1%) in Dakar, but relatively stable compared with the previous year (55%) [
40]. The level of
in vitro resistance to MQ has increased since previous studies conducted in Senegal. In Dakar, the percent of isolates with decreased susceptibility was 17% in 2001 [
41] and 13% in 2002 [
33]. In Dielmo and Ndiop (280 km south-east of Dakar), the prevalence of
in vitro resistance to MQ was 22% in 1995 [
42] and 15% in 1999 [
43,
44]. Prophylaxis failure with MQ has been previously described in Senegal [
45], and MQ is one of the three anti-malarial drugs recommended for travellers as an anti-malarial prophylaxis in Senegal. Clinical trials are in progress to evaluate the efficacy of MQ for intermittent preventive treatment of infants and pregnant women, while MQ is still used for the treatment of uncomplicated malaria in infants in Dakar. Nevertheless, MQ has been employed relatively infrequently in Africa compared to Asia. The combination of artesunate-mefloquine, which is administered to patients in Asia, is not yet used in Senegal. However, scientific data are not available for MQ monotherapy, and very few data are available on the
in vitro decreased susceptibility to MQ and its clinical implications in Africa. It is important to monitor the evolution of
P. falciparum susceptibility to MQ, to archive suspicious isolates and to correlate clinical outcomes with pharmacokinetic and phenotypic responses and with molecular markers.
As far back as 1988,
in vitro P. falciparum resistance to CQ was reported in Dakar, and reports of resistance in other regions of the country followed shortly [
46]. From 1991 to 1995, parasitological failures were observed in 21% of patients in Pikine and in 23% of patients in another region of Senegal [
47]. The
in vitro resistance to CQ increased from 1995 to 1999 in Dielmo with 32% resistance in 1995 [
48] compared to 49% resistance in 1996 [
49], 44% resistance in 1997 [
50] and 55% resistance in 1999 [
43]. In 2010, the prevalence of
in vitro resistance to CQ in Dakar was low and stable in comparison with the previous year (24.2% versus 22%) [
40]. These data are consistent with previous work on
in vitro resistance in Thies in 2007 (23% of isolates exhibiting CQ resistance) [
51]. The evolution of susceptibility to CQ is confirmed by evaluation of molecular markers of CQ resistance, and mutations in
pfcrt have been shown to be correlated with CQ resistance in different parts of the world [
52]. The prevalence of the
pfcrt 76T mutation has decreased since 2004 in Dakar. From 2000 to 2001 in Guediawaye, a suburb of Dakar, a prevalence of 92% of was observed for the 76T mutation in pregnant women with malaria [
53]. In Pikine, another suburb of Dakar, the prevalence of the 76T mutant was 79% in 2000 [
54], 63.9% in 2001 [
55] and 59.5% in 2004 [
56]. In 2002, the prevalence of the
pfcrt 76T mutation was 65% in patients hospitalised for malaria at the Hôpital Principal de Dakar [
33]. From 2001 to 2002, the prevalence of the
pfcrt 76T mutation was 75.8% in pregnant women taking chloroquine prophylaxis in Thiadiaye (84 km southeast of Dakar) [
57]. In this study, the
pfcrt 76T mutation was identified in 43.6% of the patients recruited from August 2010 to January 2011. These data are consistent with previous works on CQ molecular resistance in Dakar in 2009 (37.8%) [
58], in central Senegal (Mbour, Fatick and Bambey) in 2009 (29.3%) and in 2010 (25.1%); and in south Senegal (Tambacounda, Velingara and Saraya) with 2010 (34.5%) and in 2011 (28.8%) [
59]. However, in Pikine, the prevalence of the
pfcrt 76T mutation ranged from 64% to 79% before CQ withdrawal (2000 to 2003) [
54,
55,
60], the prevalence then decreased to 47-60% [
56,
60] while amodiaquine plus pyrimethamine-sulphadoxine was the first-line treatment (2004–2005); this prevalence has increased slightly to 59% since ACT has been implemented (2006 to 2009) [
60].
This decrease in CQ resistance parallels the withdrawal of CQ treatment and the introduction of ACT in 2002 in Senegal. However, in 2003, CQ was still being administered to patients. The prevalence of CQ in the urine ranged from 14.5% to 47.5% in two- to nine-year-old children from northern Senegal and from 9.0% to 21.4% in children from southern Senegal [
61]. In 2006, the Senegalese National Malaria Control Programme recommended ACT as the first-line treatment for uncomplicated malaria, in this same year, Senegal reported 10.6% chloroquine use and 9.7% ACT use [
62]. In Dakar in 2006, CQ represented 5.1% of the anti-malarial drugs used in children [
63] and 3.5% in 2009 [
64]. Since 2006, more than 1.5 million ACT treatments have been administered in Senegal [
3], and 184,170 doses of ACT were dispensed in 2009 [
4]. A reduction in CQ resistance was also reported in Malawi after the withdrawal of CQ treatment [
65]. This observation prompted an
in vivo CQ study in Malawi five years later, in which CQ was found to be 99% effective [
66]. The rapid dissemination of CQ resistance in Dielmo, despite strictly controlled anti-malarial drug use, argues against the re-introduction of CQ, at least in mono-therapy, in places where the resistant allele has dropped to very low levels following the discontinuation of CQ treatment [
67]. Despite the regain of CQ susceptibility, any reintroduction would likely result in a rapid re-emergence of resistant strains.
The prevalence of isolates with
in vitro reduced susceptibility to MDAQ slightly increased in 2010, 11.8% versus 6% in 2009 [
40]. The prevalence of reduced susceptibility to MDAQ was 0% in 1996 and 1999 in Dielmo [
44,
48], while it was 5% in Mlomp (Casamance, south-western Senegal) in 2004 [
68].
The prevalence of the
pfmdr1 mutations 86Y, 184F and 1246Y were 16.2%, 50.0% and 1.6%, respectively. In 2000 and 2001, prevalence rates of 31% and 30.6%, respectively, were observed for
pfmdr1 86Y in Pikine [
54,
55], and a prevalence rate of 17.2% was observed in Dakar in 2009 [
58]. However, the role of polymorphism in
pfmdr1 is still debated. The
pfmdr1 86Y mutation has been shown to be associated with
in vivo resistance to amodiaquine in recrudescence after monotherapy with amodiaquine [
69] or after combination therapy with artesunate-amodiaquine [
70]. The
pfmdr1 1246Y mutation has also been found to be associated with
in vitro resistance to amodiaquine [
71] and with recrudescent infection after treatment with amodiaquine or amodiaquine-artesunate [
70,
72]. In a meta-analysis, the
pfmdr1 86Y mutation was found to be associated with amodiaquine failure, with an odds ratio of 5.4 [
17]. Based on this hypothesis, the 16.2% prevalence of
pfmdr1 of 86Y predicts that 16.2% of isolates would be resistant to amodiaquine in 2010 in Senegal. The resistance to amodiaquine has remained low even after the introduction of artesunate-amodiaquine in 2006 in Senegal. A study in Dakar and Mlomp from 1996 to 1998 showed that monotherapy with amodiaquine remained effective for treating uncomplicated malaria in areas where CQ resistance was prevalent [
73]. The artesunate-amodiaquine–associated cure rates were > 99.3% in Mlomp and Keur-Socé when administered either as a single daily dose or as two daily doses [
74]. The fixed-dose combination of artesunate-amodiaquine (ASAQ) exhibits a cure rate > 98.5% [
75]. However, ACT efficacy and resistance must be monitored because clinical failures, or at least extended parasite clearance times, have been described in Cambodia [
76,
77]. In this context, it is important to implement
in vitro and
in vivo surveillance programmes, such as those championed by the Worldwide Antimalarial Resistance Network [
78,
79].
No isolate exhibited reduced
in vitro susceptibility to DHA. This result is consistent with previous studies that did not show any parasites resistant to artesunate [
33,
43,
44]. However, Agnamey
et al. reported that 3% - 23% of isolates had IC
50 values greater than 15 nM in Mlomp between 2000 and 2004 [
73]. High IC
50 values can also be found for artemisinin, with an IC
50 > 30 nM in Dakar [
51] and artesunate with an IC
50 > 45 nM [
41].
The other ACT first-line treatment for uncomplicated
P. falciparum malaria in Senegal is the combination of artemether-lumefantrine. Only 2.9% of the isolates presented reduced susceptibility to LMF, and this prevalence did not rise in Senegal after the introduction of ACT. In 1996, 6% of isolates from Dielmo were resistant to LMF
in vitro[
80]; and 1% of the isolates were resistant to LMF in 2009 in Dakar [
40].
In 2009, 7% of isolates showed low reduced susceptibility to QN, which is in accordance with previous studies in Dakar [
33,
40,
41] and Dielmo [
42‐
44]. Isolates with a high IC
50 to QN were already identified in 1984 [
81]. Even in areas where QN efficacy remains good, such as sub-Saharan Africa, the susceptibility of individual
P. falciparum isolates to QN has varied widely. The IC
50s for isolates collected in Senegal were 31 to 765 nM in 1984 (Thies and Kaolack) [
81], 5 to 932 nM in 1996 (Dielmo) [
49] and 6 to 1291 nM in 2009 (Dakar) [
40]. The wide range in QN susceptibility and recent evidence for QN treatment failure seen across Africa [
82,
83] or in Senegal in a patient who spent two months in Dielmo in 2007 [
84] suggest that the evolution of parasites with reduced susceptibility may contribute to QN decreased efficacy. QN used in combination with DOX in the treatment of severe malaria in Dakar.
A prevalence of 10.3% of isolates with
in vitro reduced susceptibility to DOX was observed in Dakar in 2010, which is similar to the prevalence observed in 2009 (12%) [
40]. The mean IC
50 of DOX was similar to those estimated in Dielmo in 1998 [
85,
86]. The slow activity of DOX
in vitro has a delayed effect upon growth and requires the prolonged incubation of parasites [
85,
87]. However, the standard 42 h test is still used to monitor DOX
in vitro susceptibility.
The
pfdhfr 108N mutation has been shown to be correlated with
in vitro and
in vivo resistance to pyrimethamine [
10,
17]. The OR for sulphadoxine-pyrimethamine failure associated with Ser108Asn has been shown to be 3.5 (95%CI: 1.9-6.3, meta-analysis of 10 studies) for a 28-day follow-up [
17]. The additional mutations N51I, C59R or I164L increase the level of
in vitro resistance to anti-folate drugs and sulphadoxine-pyrimethamine. The OR values for single mutants of codons 51 and 59 are 1.7 (95%CI: 1.0-3.0) and 1.9 (95%CI: 1.4-2.6), respectively [
17]. In 2010, the prevalence of
pfdhfr 108N was 81.9% in patients with malaria who were treated at the Centre de santé Elizabeth Diouf, which is similar to the results of those treated in 2009 at the Hôpital Principal de Dakar (82.4%) [
58]. The triple mutation (51 + 59 + 108) increases the risk of
in vivo resistance to sulphadoxine-pyrimethamine by 4.3 (95%CI: 3.0-6.3, meta-analysis of 22 28-day studies). Isolates carrying a combination of three mutations (108N, 51I and 59R) were associated with high-level pyrimethamine resistance and represented 73.6% of isolates. In 2002, in Dakar, the prevalence of
pfdhfr 108N was 65%, and triple mutants were identified in 50% of the isolates [
33]. In 2003, the prevalence of mutations in
pfdhfr codon 108 was 78% in Pikine, and the prevalence of the triple mutant was 61% [
88]. Additionally, in 2007, in the rural area Keur Soce, triple mutants were identified in 67% of patients treated with sulphadoxine-pyrimethamine combined with amodiaquine [
89].
The
pfdhps 437G mutation has been shown to be correlated with
in vitro and
in vivo resistance to sulphadoxine [
11,
17]. The single mutation Ala437Gly and the double mutation Ala437Gly + Lys540Glu increase the risk of
in vivo resistance to sulphadoxine-pyrimethamine by 1.5 (95%CI: 1.0-2.4, meta-analysis of 12 studies) and 3.9 (95%CI: 2.6-5.8, meta-analysis of 10 studies), respectively [
17]. In 2010, the prevalence of the
pfdhps 437G mutation was 54.4% in patients with malaria in Dakar. Only one isolate (0.6%) carried the double mutation (437G and 540E) that is associated with high-level sulphadoxine resistance. The mutation of codon 613 (A613S) (1.2%) is very rare in Africa. In 2002, in Dakar, only 20% of isolates harboured the
pfdhps 437G mutation [
33]. In 2003, the mutation rate in
pfdhps codon 437G was 40% in Pikine [
88]. Several studies from 2006 to 2008 in Senegal showed that the prevalence of
pfdhps 437G significantly increased after intermittent preventive treatment of infants with sulphadoxine-pyrimethamine [
89,
90]. Given the prevalence of the triple and quadruple mutants in the population of Dakar (73.6% for the
pfdhfr 108N, 51I and 59R triple mutant and 36.7% for the quadruple mutant
pfdhfr 108N, 51I and 59R and
pfdhps 437G), the use of sulphadoxine-pyrimethamine as an intermittent preventive treatment (IPT) must be monitored. Encouragingly, only one quintuple mutant,
Pfdhfr 108N, 51I and 59R and
Pfdhps 437G and 540E, which is associated with high-level sulphadoxine-pyrimethamine resistance has been identified to date.
IPT with anti-malarial drugs given to all children and pregnant women once per month during the transmission season can provide a high degree of protection against malaria. Seasonal IPT with sulphadoxine-pyrimethamine and one dose of artesunate resulted in a 90% reduction in the incidence of clinical malaria in Senegal [
6]. The combination of sulphadoxine-pyrimethamine and amodiaquine was more effective than the combination of sulphadoxine-pyrimethamine and artesunate or the combination of amodiaquine and artesunate in preventing malaria [
91]. During IPT with sulphadoxine-pyrimethamine and piperaquine, only 3.4% of the treated children developed malaria [
89]. However, the single use of sulphadoxine-pyrimethamine as seasonal IPT is inadvisable; for instance, sulphadoxine-pyrimethamine must be used in combination with amodiaquine, artesunate or piperaquine [
89,
92].
The introduction of ACT in 2002 in Senegal did not induce a decrease in P. falciparum susceptibility to individual drug components, such as DHA, MDAQ and LMF. However, the prevalence of P. falciparum isolates with reduced drug susceptibility to MQ increased, and clinical failures with QN have been reported in Senegal. Additionally, in Senegal, isolates with high IC50 values for artemisinin derivatives. Since 2004, the prevalence of chloroquine resistance has decreased, but the data argue against the re-introduction of chloroquine for mono-therapy in places where the resistant allele has dropped to very low levels following discontinuation of chloroquine treatment. The prevalence of isolates resistant to pyrimethamine is high (81.9%), with 73.6% of parasites exhibiting high-level pyrimethamine resistance. The prevalence of isolates resistant to sulphadoxine was 54.4%. Susceptibility to anti-malarial drugs remains stable between 2009 and 2010 in Dakar. Nevertheless, an intensive surveillance of the susceptibility of P. falciparum to anti-malarial drugs in vitro must be conducted in Senegal. In addition, maximising the efficacy and longevity of ACT as a tool to control malaria will critically depend on pursuing intensive research into identifying in vitro markers as well as implementing ex vivo and in vivo surveillance programmes. In this context, there is a need to identify molecular markers that predict ACT resistance, which can provide an active surveillance method to monitor temporal trends in parasite susceptibility.