This cross sectional study shows an 11.9% overall prevalence of HIV primary drug resistance with a 7.5% NNRTI-related resistance in naïve pregnant women.
The recent HIV drug resistance report, released by WHO in November 2012 , reports the available data on the estimated prevalence of HIVTDR between 2003 and 2010 in 72 surveyed areas. WHO recommends a minimum-resource method to assess HIVTDR in resource-limited countries where transmitted HIV drug resistance is likely to be seen first (such as in urban areas). Of the 72 surveys, 52 (72.2%) had a low prevalence of resistance to all drug classes and 20 (27.8%) had a moderate prevalence classification of resistance to ≥1 antiretroviral drug class). The first reports from Tanzania in 2005 showed a 4% and 9% NNRTI resistance in naïve populations from the Kagera-Kilimanjaro regions and Dar-es-Salaam urban area, respectively [13, 14]. More recently, in 2011, a 14.7% HIVDR prevalence in a naïve population was reported in Mwanza . Authors combined drug-resistance prevalence data obtained from both peripheral blood mononuclear cells (PBMC) and plasma, whereas only 4 samples (9.5%) resulted positive at bulk sequencing assay from plasma . The NNRTI prevalence rates observed in these surveys are slightly higher than those observed in previous reports in other eastern and western African countries.
Reports showed the effect of a proper timing for introduction of ART, as a proxy for the amount of circulating drug-resistance HIV-1 strains at the population level, and level of primary resistance . The overall sample-weighted drug resistance prevalence was 5.6% (139 of 2436; 95% CI 4.6–6.7), ranging from 1.1% (two of 176; 0.0–2.7) in Pretoria, South Africa, to 12.3% (22 of 179; 7.5–17.1) in Kampala, Uganda .
Regarding PI mutations, no major resistance mutations were observed in our study. Nevertheless, the 76.1% prevalence of the L89M protease polymorphism raises concern. Some amino acid polymorphisms occur at sites that have been associated with drug resistance in the B-subtype virus . The L89M mutation increases the catalytic efficiency and vitality of the HIV-1 protease gene in the presence of other protease mutations in non-B African viral subtypes  and can determine a low accumulation of primary protease mutations in non-B subtypes [18, 19]. These findings suggest that in addition to the primary drug-related mutations already described in B clades, particular attention should be paid to some natural polymorphisms in the therapeutic management of patients infected by HIV-1 non-B subtypes.Our study aimed to determine the prevalence of transmitted HIV drug-resistance mutations among untreated patients and also provided novel data on the HIV-1 variants that circulate in Tanzania. We confirmed previous results that reported a high genetic diversity in the number of co-circulating variants with the predominance of A clade (53.7%), and a high prevalence of the A1 subtype (47%). Different from a 2004 report that described a low detection of drug resistance in A subtype compared to D subtype , we reported that the primary NNRTI drug-related mutations were all clustered in the A1 subtype. This variant was described as one of the most prevalent variants among young adults in Tanzania, Dar-es-Salaam  and adults in the Kilimangiaro, Kagera and Mwanza regions [13, 15]. Thus, the implementation of a surveillance study on the molecular epidemiology of different HIV strains appears strictly complementary to the data obtained from prevalence studies of drug-related mutations using bulk sequencing.
Several limitations were encountered in our study. Despite systematic screening of previous ART use, the mean number of pregnancies is three and the risk of unreported ART use in the AMANI cohort may be significant. Nevertheless, most HIVDR mutations were reported in parous women; the AMANI study was conducted in a single area; a single population was targeted, and the estimates of drug resistance have wide CIs. However, despite all these considerations, the 11.9% prevalence of drug resistance in a naïve population where the first-line antiretroviral regimen is still based on a NNRTI-based HAART raises concern. Further work should be done to determine if resistance is a consequence of short-term exposure during pregnancy or if in fact these individuals had already accessed ARVs. Together, this information could be used to guide the development of ART policy guidelines in Tanzania. Against this background, the increasing rates of antiretroviral resistance in adults and children from low-income settings represent a potential threat and urgent actions are needed. First, human and resource efforts should be doubled to deploy proven effective preventive methods. Second, early and sustained ART use for preventing MTCT must be fully embraced  and the recent 2012 WHO programmatic update on HIV PMTC transmission  should be strongly supported to provide the option B-plus, the use of a life-long triple therapy for the pregnant women.