The data show a global and significant increase of the drug resistance associated molecular markers frequencies. This increase concerns several (9 out of 15) polymorphic molecular markers linked to different metabolic pathways. The study covered a relatively short-time period of four years, showing a mean doubling of frequencies of mutations highly linked to resistance, such as pfcrt76 and pfhdfr108. The dynamics highlighted here appear to be fast, global and intense.
These data add arguments to the debate about the links between transmission and resistance [23
]. On the one hand, Talisuna et al
in Uganda [24
] have shown higher levels of resistance to chloroquine and SP in zones of higher transmission intensity. On the other hand, history shows that chloroquine resistance emerged first in low transmission zones and that antifolate resistance has increased more rapidly in low transmission areas [25
]. Theoretical models support the view that the spread of resistance in low transmission area is favoured either by the high selfing-rate [23
] or by intra-host competition [24
]. Others models [26
] have shown that the transmission rate may not be the main force driving the resistance spread, which may rather be led by drug pressure or immunity.
However, field data are scarce. Here, the mutations tended to be more frequent in the village with the shorter malaria season: Banizoumbou. This was verified for the pfcrtK76T mutation. It has to be noted that the two villages share very comparable social patterns (distance from main road, health facilities, ethnicity, millet-based economy), but differ in their water bodies – ie Anopheles breeding sites – duration times. This corroborates the hypothesis that the resistance could spread more easily in low-transmission zones. The field data presented here correspond to one of the most restricted spatial scale ever described for this effect: these two villages are only 30 km distant. History has shown at the worldwide-scale that resistance emerged from low transmission area (Amazonia, South East Asia), and this may be verified at a very local scale. The same tendency for pfdhfr and others markers seems to give a global value to the processes involved in the spread of resistance.
The frequencies of mutation tended to be higher among the younger parasite carriers. However, this relation was significant only for the dhps436
double mutation. As the immunity or the drug pressure may be involved, the study was not designed to answer these issues. The global prevalence of pfcrtK76T
, 32.4%, was lower than in many published works performed in Africa [27
], but not so far from regional data [3
]. In comparison with local data this prevalence was different from the one observed in hospitalized urban population recorded in 2003: Ibrahim et al
found respective prevalence of pfcrtK76T
mutations of 45.4% and 61.9% [18
]. Later in 2005 [19
] among patients attending rural health centres, the prevalence was closer to the 2005 values: 50.8% for pfcrt76
and 57.7% for dhfr108
. However, the present work was targeting asymptomatic carriers in a rural zone, not patients attending to clinics or hospitals, and the drug pressure appears to be different in those different populations.
Recent works have studied the ATPase6
gene and specially the pfATPase6S769N
mutation as a candidate marker for artemisinin resistance [11
]. For first time, several field strains could be screened for five SNPs of PfATPase6
gene. The pfATPase6S769N
mutation was not found. However, the pfATPaseA623E
mutation was found in 4.7% of samples. The more precise sequence data for this gene are presented elsewhere [29
The respective factors driving the emergence and spread of malaria drug resistance are not well known. Drug pressure, clonality of infections, pharmacokinetics, human migrations, immunity, fitness of parasites and transmission levels have all been suspected [24
]. One interesting point of this work is the role of the intervention to block transmission. Two millions and three hundred thousands long-lasting insecticide-impregnated bed nets were distributed in December 2005. The bed nets reduce the lifespan of the vectors and the contact between human host and vector. This decreases the malaria transmission. Alifrangis et al
have observed a significant decrease of molecular resistance to pryrimethamine after impregnated bed nets in Tanzania [32
]. On the other hand, a controled trial of the impact of impregnated bed nets in Ivory Coast found no significant difference in pfcrt, pfdhfr
mutations prevalences between treated villages and negative controls [33
]. Diallo et al
in Burkina Faso observed that impregnated curtains are not associated with an increase of molecular resistance or therapeutic failure in children under five years of age [35
A decrease or stabilization of the molecular resistance has been shown between 2005 and 2006 for pfcrtK76T
(Figure ), when the trend is to a global increase during the 2003–2006 period. During this period, given the parasite carriage among children between two and five years of age, the malaria transmission in the two villages has decreased significantly (chi-square p < 0.0001), whether due or not to the impregnated bed nets distribution, from 56.5% (n = 255) in 2005 to 37.1% (n = 237) in 2006. What was observed in molecular markers then could be related either to a decrease of drug pressure or to a direct effect of transmission decrease on mutation carrying strains. A decrease of drug pressure cannot account for the pfdhfrS108N
dynamics, when the SP is not often used in the study zone and rarely as the first-line treatment [18
]. Beside this drug pressure hypothesis, the decrease of resistant strains may be related to:
• An absolute – i.e. in absence of drug pressure – weaker fitness of resistant strains in competition with wild strains during the transmission season: this is compatible with the observed higher frequency of resistant strains in low transmission area, where the competition between strains related to the chance of transmission events, is lower.
• A best fitness of mutation carrying strains, when drugs are delivered during the malaria season, with a relative lower fitness of wild strains, especially at the end of the transmission season when the drug pressure is maxima. This effect is cumulative during the dry season, when there is no transmission and thus no competition with wild strains. So the resistance could favour the longer parasite carriage during the dry season, but not the competition with wild strains as the new transmission season begins. This theoretical view is in accordance of the observed decrease after bed net distribution by personal protection and decrease of drug pressure.
• Whatever the processes involved, and taking in account the short duration time of the malaria transmission period in Sahel, the selective advantage kept all along the dry season by mutation carrying strains seems to be cumulative from year to year and tends to a global increase.