The prevalence of antibodies to HIV was 6.1%, with no significant difference between 2005 and 2008 or by age group. Although few data are available for Gabon, the prevalence of antibodies to HIV between 1986 and 1989 was about 2.0% in the general population of Libreville . More recent studies on populations at risk in Gabon showed a seroprevalence of 4.5% among manganese miners and 7.5% in the general population living close to Cameroon [7, 9].
Of note, migrant populations and pregnant women had higher seroprevalences than miners in Gabon; this finding might be due to the geographic location of sampling. The low seroprevalence among manganese miners might be due to containment of this semi-rural population, whereas migrant populations and pregnant women are more representative of the general urban population. Furthermore, the mining company educated workers about HIV transmission, perhaps resulting in a 'healthy worker effect'. Similar findings were made in Cameroon in 2004, with seroprevalences of 4.0% and 6.7% in rural and urban populations . In the Central African Republic in 2006, a prevalence of 6.2% was found in Bangui and 7.9% for the rest of country, with a seroprevalence almost twice as high among women than men .
Our results show that widely genetically diverse HIV-1 strains are circulating among untreated pregnant women in Gabon, including pure subtypes (A, D, F, G, H), a unique recombinant form (CRF02_AG) and complex recombinant forms (CRF06_cpx, CRF11_cpx). At least eight viral lineages have been observed, confirming the results of previous studies in central Africa [7–10]. As a number of strains remained unclassified, genetic diversity in Gabon is probably even greater, with complex mosaic compositions. A wide genetic diversity of circulating strains was also reported in the Democratic Republic of the Congo and Cameroon [20, 21]. We confirmed that CRF02_AG is still the most prevalent HIV-1 subtype throughout Gabon. This subtype was widespread in west Africa a few years ago and is now present in central Africa .
The geographic distribution of HIV subtypes is changing constantly, with the greatest genetic diversity of HIV-1 in sub-Saharan Africa . A preliminary study in the Democratic Republic of the Congo determined pure subtypes of two HIV strains, and further investigations revealed a combination of several subtypes . In Cameroon, unique recombinant forms have been found to be far more complex than previously described, indicating evolution of unique forms into complex combinations . Recombinant forms were reported to have greater replicative capacity in vitro than the pure subtypes from which they derived . In our study, the viral load of HIV-1 strains among untreated pregnant women was high, suggesting a potentially critical role of dual infections in vivo, as the attained HIV viral load could limit the frequency of inter-subtype recombination .
Vergne et al. reported resistance to ARVs among patients in Gabon followed up for HIV infection. No other studies in this country have addressed direct effects on the effectiveness of the most frequently used ARVs in HIV-1 treatment. As expected, high frequencies of polymorphism mutations were found in the P and RT coding regions, which do not affect the three classes of ARVs. Even in HIV-1 strains from untreated patients, mutations associated with susceptibility to PIs, NRTIs and non-NRTIs were observed. The L33F and M46I mutations reduce susceptibility to PIs, except for saquinavir/r; L210W confers low-level resistance to several NRTIs, except for lamivudine and emtricitabine; V118I induces low-level resistance to lamivudine and K101Q causes a low-level reduction in susceptibility to non-NRTIs. These findings confirm resistance to HIV drugs among untreated patients infected with non-B subtypes, even though these drugs were introduced in Gabon relatively recently.
Some ARVs resistance-associated mutations occur commonly without selective pressure at highly polymorphic positions; these polymorphic mutations should therefore not be considered in surveillance of transmitted resistance to ARVs , as their inclusion leads to inflated estimates. We found that 2.8% were major mutations for resistance to PIs (1.9%) and NRTIs (0.9%), providing little evidence for transmitted resistance to ARVs among untreated pregnant women in central Africa.
In the present study, we evaluated resistance in 107 samples obtained from strains in the active phase of viral replication, representing about the half of the HIV-positive women. As these samples contain strains from all regions of Gabon, they give an overall picture of natural ARVs resistance-associated mutations in this population. We cannot exclude the possibility that other ARVs resistance-associated mutations were present among strains from samples not evaluated here. Therefore, the true level of resistance might be underestimated, and epidemiological studies should be conducted in the general population each year to follow-up the circulating subtypes of the HIV strains and to evaluate the emergence of new ARVs resistance-associated mutations.
The emergence of strains with both minor and major mutations for one or several classes of ARVs poses a real problem, especially for the introduction of new ARVs or switch therapy for HIV-infected patients. Furthermore, the risk for resistance to ARVs treatment might be exacerbated by cumulated mutations. Although experts have dismissed the relevance of changes related to polymorphic mutations of non-B subtypes, low-to-moderate resistance to HIV drugs has been reported in neighboring central African countries in which high HIV diversity was found [28, 29].