This study shows that, despite the change in national antimalarial policy, the prevalence of CQ-resistant and MF-resistant isolates in Franceville, Gabon, remains stable compared with the year 2000
. MDAQ resistance has not previously been investigated in Franceville. The level of MDAQ resistance found here is higher than that previously found in Libreville, Gabon, where 100% of isolates were susceptible to MDAQ in the year 2003
. Compared with previous data from southeast of Gabon, where only 5.4% of isolates were MDAQ-resistant
, we found an high prevalence of in vitro MDAQ resistance. The mean MDAQ IC50 found here is similar to that reported in Kampala, Uganda, between 2006 and 2008, after implementation of the ACT treatment policy
. The high level of MDAQ resistance found here could be explained by fact that amodiaquine is combined with artesunate in one of the recommended ACTs used in Gabon.
We found a high prevalence of isolates with reduced susceptibility to DHA. But, because of the brief action of DHA, this prevalence may have been overestimated. The decrease in DHA susceptibility may have arisen through the use of artemether or artesunate monotherapy to treat uncomplicated malaria, contrary to the recommendations of the Gabonese national program against malaria. Our results are in keeping with data from the capital, Libreville, where the prevalence of artemether resistance was 14%, with a range of 0.8 to 34.8 nM (mean IC50 5.0 nM)
. The frequency of isolates with reduced susceptibility to DHA found here is similar to that described in Lambarené, a town situated in central Gabon
. We found a strong correlation between decreased susceptibility to DHA and resistance to MF and CQ. These data are consistent with previous data showing that the selection of MF resistance leads to increased resistance to artemisinin in rodent model
, despite the fact that these drugs do not belong to the same class. Indeed, MF was used before implementation of artemisinin derivates. The correlation between MF and CQ resistance is also consistent with previous reports of cross-resistance
. In this case, CQ was introduced first, leading to the resistance to MF. This cross-resistance between these drugs warrants reinforced surveillance of antimalarial drug resistance in Gabon. Surprisingly, MDAQ resistance did not correlate with CQ resistance, despite the facts that MDAQ and CQ belong to the same drug family and that MDAQ and CQ antimalarial activities correlate with each other
. However, this lack of correlation might also be explained by a lack of statistical power in our study.
Seven isolates (11.3%) could not be cultured, possibly because of the presence of antimalarial drugs in the patients’ blood that was consistent with manufacture informations.
Previous reports indicate that the use of artemether-lumefantrine is associated with the selection of Pfmdr1-N86 wild-type parasites, which are tolerant of low LM concentrations
[15, 39–41]. The present data confirm those of our previous study showing an increase in Pfmdr1-N86 genotype field isolates in Franceville after implementation of the ACT treatment policy
. In comparison with the prevalence of N86 in 2004 (15.6%) and 2009 (31.3%), this gradual increase calls for reinforced surveillance of artemether-lumefantrine resistance.
A correlation between the Pfmdr1 Y86 and Y1246 mutations and both CQ and MDAQ resistances has been reported
[34, 42], but we did not find this correlation. Surprisingly, the Pfmdr1 N86Y polymorphism was not associated with decreased sensitivity in this study. This is, however, consistent with data from Lambarené (central Gabon), where Y86 was not found to be associated with CQ resistance in 1998
, before implementation of ACT: in this latter study the prevalence of N86 was 20.5% and the Y1246 mutation was not detected. Our data suggest the spread of the Y1246 mutation in Gabon. The 184F mutation was not analysed, as its impact on drug resistance is controversial
[14, 44], but this mutation has been identified as an independent marker of decreased lumefantrine susceptibility
 and it will be interesting to investigate it in our study region.
The Pfmdr1 copy number was not investigated here, as it has not been consistently linked to ACT treatment failure
[22, 31], and this is one of limitations of our study. Indeed, it has been shown that increased a Pfmdr1 copy number is associated with reduced in vitro susceptibility to mefloquine, lumefantrine and artesunate
While Pfcrt CVIET is the most prevalent haplotype in Africa, the SVMNT haplotype has recently migrated to Tanzania and Angola
[46, 47]. This haplotype, which has relatively little fitness cost, has been associated with emergence of AQ resistance
[48, 49]. Thus, it is crucial to monitor Pfcrt codons 72–76 in all countries in which AQ combination therapy is recommended. Because of human migration between Angola and Gabon, we investigated the Pfcrt haplotype 72–76 in our isolates: only haplotypes CVIET and CVIEK were detected. The high prevalence of CVIET (n=45/46) made it difficult to show a significant difference in the distribution of polymorphisms between resistant and sensitive isolates. This haplotype is associated with resistance against 4-aminoquinoline. The persistently high frequency of CVIET, despite the withdrawal of CQ in Gabon, could be due to use of amodiaquine as a partner in some ACTs. Recently, CVIET was also reported to be associated with reduced sensitivity to new 4-aminoquinolines such as piperaquine
Polymorphisms in the putative drug transporter PfATPase6 were not closely linked to in vitro drug sensitivity, in keeping with data from Cameroon, a neighboring country
. No mutations were found at codon 110. The E431K, A623E, and S769N polymorphisms were not analysed, because of their rarity in sub-Saharan Africa and their lack of impact on in vitro drug resistance