Application of the UNAIDS Incidence Patterns Model to Determine the Distribution of New HIV Infection in Lagos State, Nigeria.

BACKGROUND: Identifying patterns in the distribution of new HIV infections in the population is critical for HIV programmatic interventions. This study aimed to determine the distribution of New HIV infection by applying the incidence patterns mathematical model to data from Lagos state. METHODS: The incidence patterns model (IPM) software is a mathematical model developed by UNAIDS to estimate the demographic and epidemic patterns of HIV infections. This model was adapted in Lagos state to predict the distribution of new HIV infections among specified risk groups in the next 12 months. RESULTS: The IPM predicted a total HIV incidence of 37 cases per 100 000 individuals (3979 new infections) will occur among the 15 to 49 subpopulations. The results also showed that sero-concordant HIV-negative couples with external partners (29%), female sex workers (26%), men-having-sex-with-men (18%), and previously married females (6%) accounted for the majority of the estimated new HIV infections. Overall, key populations constitute almost half (48%) of the estimated number of new HIV infections. CONCLUSION: The study helped to identify the population groups contributing significantly to new HIV infections. Therefore, priority interventions should be focused on these groups.

HIV-1 drug-resistance mutations among newly diagnosed patients before scaling-up programmes in Burkina Faso and Cameroon.

We analysed whether mutations associated with resistance to antiretroviral (ARV) drugs circulate among treatment-naive HIV-1-infected individuals at a period when these drugs started to become more widely available in Africa. Overall, major resistance mutations in the pol gene, as defined by the International AIDS Society Resistance Testing-USA panel, were observed in 16 treatment-naive individuals. Eight of the 97 patients tested in Burkina Faso bore mutations conferring resistance to one drug class of ARV drugs: two to nucleoside reverse transcriptase inhibitors (NRTIs; M41L [n = 1], M41L+T69S [n = 1]), four to non-NRTIs (NNRTIs; V106A/V [n = 1] and V1081 [n = 3]) and two to protease inhibitors (PIs; L33F [n = 2]). In Cameroon, resistance mutations were identified in 8 of 102 patients: three to PIs (M461/L [n = 2], L33F [n = 1]), three to NRTIs (T69N/T [n = 1], M184V [n = 1], A62V [n = 1]) and two to NNRTIs (P236L [n = 1], V1081 [n = 1]). It is important to note that not all genotypic drug-resistance algorithms give similar interpretations to the observed mutations. Population surveillance for ARV drug resistance is required and should be included in all implementation programmes.

Functional domains of the human immunodeficiency virus type 1 Nef protein are conserved among different clades in Cameroon.

The Nef protein of human immunodeficiency virus type 1 (HIV-1) has multiple functional domains, is immunogenic, and contains several cytotoxic T lymphocyte (CTL)-targeted epitopes. Several defined subfunctions of Nef are important for the pathogenesis of HIV-1 infection. In this study, we present the genetic diversity of the nef gene of 55 newly derived HIV-1 sequences obtained from Cameroonian patients. Four genetic subtypes and three circulating recombinant forms (CRFs) were identified: subtypes A (11%), G (7.3%), D (5.4%), F1 (1.8%), F2 (5.4%), CRF01_AE (5.4%), CRF02_AG (58.2%), and CRF11_cpx (1.8%). Two isolates clustered distinctly from the known HIV-1 genetic subtypes in nef and were designated as unclassified. Interestingly, the majority of all functional domains including the myristoylation signal, CD4 binding motif, beta turn motif, and the phosphorylation sites were well conserved in our cohort. Putative CTL-epitopic domains of the central portion of Nef were also well conserved, whereas those at the C-term were not. Our study demonstrated that despite high genetic diversity observed in the nef gene, most described functional domains and CTL epitopes were well conserved among Cameroonian HIV-1 subtypes. These findings could be used for the development of antiretroviral-acting therapeutics and anti-HIV-1 vaccines.

HIV genetic diversity in Cameroon: possible public health importance.

To monitor the evolving molecular epidemiology and genetic diversity of HIV in a country where many distinct strains cocirculate, we performed genetic analyses on sequences from 75 HIV-1-infected Cameroonians: 74 were group M and 1 was group O. Of the group M sequences, 74 were classified into the following env gp41 subtypes or recombinant forms: CRF02 (n = 54), CRF09 (n = 2), CRF13 (n = 2), A (n = 5), CRF11 (n = 4), CRF06 (n = 1), G (n = 2), F2 (n = 2), and E (n = 1, CRF01), and 1 was a JG recombinant. Comparison of phylogenies for 70 matched gp41 and protease sequences showed inconsistent classifications for 18 (26%) strains. Our data show that recombination is rampant in Cameroon with recombinant viruses continuing to recombine, adding to the complexity of circulating HIV strains. This expanding genetic diversity raises public health concerns for the ability of diagnostic assays to detect these unique HIV mosaic variants and for the development of broadly effective HIV vaccines.

Field assessment of generic antiretroviral drugs: a prospective cohort study in Cameroon.

OBJECTIVE: To assess the effectiveness of generic anti-retroviral drugs in terms of survival and virological and immunological responses, as well as their tolerability and the emergence of viral resistance. METHODS: A total of 109 HIV-1-infected patients were enrolled in a prospective cohort study in Yaoundé, Cameroon. Available generic drugs were a fixed-dose combination (FDC) of zidovudine (ZDV) and lamivudine (3TC), an FDC of 3TC, stavudine (d4T) and nevirapine (NVP), and individual formulations of ZDV, 3TC and NVP. RESULTS: At baseline, the median CD4 cell count was 150/mm3 [interquartile range (IQR) 61-223] and median viral load was 5.4 log10 copies/ml (IQR 4.8-5.6); 78% of patients received ZDV/3TC/NVP and 22% received 3TC/d4T/NVP. Median follow-up was 16 months (IQR 11-23). The survival probability was high (0.92 at 12 months); plasma viral load declined by a median of 3.3 log10 copies/ml and 86.9% of the intention-to-treat population had viral load <400 copies/ml at 12 months; CD4 count had increased by a median of 106 cells/mm3 at 12 months; drug resistance rarely emerged (incidence rate 3.2 per 100 person-years); and the treatments were reasonably well-tolerated (incidence rate of severe adverse effects 7.8 per 100 person-years). CONCLUSION: Together with previous pharmacological and clinical studies, this prospective study suggests that these generic antiretroviral drugs can be used in developing countries.

Effectiveness and safety of a generic fixed-dose combination of nevirapine, stavudine, and lamivudine in HIV-1-infected adults in Cameroon: open-label multicentre trial.

BACKGROUND: Generic fixed-dose combinations have been prequalified by WHO to treat HIV-infected patients in resource-limited countries. Despite their widespread use they are, however, not yet recommended by some of the major donor agencies owing to scarcity of clinical data on effectiveness, safety, and quality. We aimed to assess these issues for one of the most frequently prescribed treatments in Africa, a generic fixed-dose combination of nevirapine, stavudine, and lamivudine. METHODS: 60 patients were followed in an open-label, 24-week multicentre trial in Cameroon. All patients received one tablet of the fixed-dose combination drug twice daily. The primary outcome measure was the proportion of patients with viral load less than 400 copies per mL at the end of the study period, in an intention-to-treat analysis. FINDINGS: At baseline, 92% of patients (n=55) had AIDS; median CD4 count was 118 cells per microL (IQR 78-167) and median plasma HIV-1 RNA was 104?736 copies per mL (40804-243787). The proportion of patients with undetectable viral load (<400 copies per mL) after 24 weeks of treatment was 80% (95% CI 68-89). Median (IQR) change in viral load was -3.1 log10 copies per mL (-2.5 to -3.6) and in CD4 count 83 cells per microL (40-178). The probability of remaining alive or free of new AIDS-defining events was 0.85 (95% CI 0.73-0.92). Frequency of disease progression was 32.0 (95% CI 16.6-61.5), severe adverse effects 17.8 (7.4-42.7), and genotypic resistance mutations 7.1 (1.8-28.4) per 100 person-years. Mean reported adherence rate was 99%. Median drug concentrations in tablets were 96% of expected values for nevirapine, 89% for stavudine, and 99% for lamivudine. INTERPRETATION: Our findings lend support to use and funding of a generic fixed-dose combination of nevirapine, stavudine, and lamivudine as first-line antiretroviral treatment in developing countries.

Enfuvirtide binding domain is highly conserved in non-B HIV type 1 strains from Cameroon, West Central Africa.

Recently T-20 or enfuvirtide, the first drug of a new class of antiretrovirals targeting the entry stage of the virus life cycle, has been clinically approved. Enfuvirtide is a peptide derived from the HR2 region of the transmembrane glycoprotein from the HXB2 HIV-1 subtype B prototype strain that binds to the HR1 region. Drug resistance seems to occur in the HR1 region between amino acids 36 and 45. We examined to what extent this region is conserved in 184 non-B strains from Cameroon: 132 (71.7%) CRF02-AG, 14 (7.6%) subtype A, 11 (5.9%) F2, 9 (4.8%) subtype D, 8 (4.3%) subtype G, 4 (2.1%) CRF01-AE, 4 (2.1%) CRF11-cpx, and 2 (1.1%) CRF06-cpx. Among the 184 strains studied, no amino acid mutation was found in the highly conserved three amino acid motif at codons 36-38 (GIV) that are important determinants of viral susceptibility to enfuvirtide. Other common substitutions like Q40H and N42T were also absent. The N42S polymorphism was present in 148 (80.4%) strains. Analysis of the HR2 domain, from which the peptide is derived, indicated a much greater genetic variability as compared to HR1.

Nucleotide sequence analysis of the accessory genes of HIV-1 group O isolates.

Human immunodeficiency virus type 1 group O strains have been described as highly divergent, compared with the majority of the viruses classified in group M. To study the diversity and genetic characteristics of group O, we have sequenced the accessory genes of 7 isolates. Analysis of the deduced amino acid sequences for Vif, Vpr, Tat, Vpu, and Rev indicate that most of the functional domains of these proteins, as described for group M viruses, are highly conserved and retained among all the group O strains we have characterized. The only difference concerns the Vif phosphorylation sites, which are absent in all of the group O isolated we have sequence with the exception of two isolates in which only one phosphorylation site was conserved. These sites, present in nearly all of the group M isolates, play a critical role in the regulation of viral replication and infectivity. As described for group M isolates, the vpu gene is the one with the highest diversity among group O viruses. Phylogenetic analysis of these sequences suggests that group O viruses could be differentiated into at least four different clusters.

Genotypic and phenotypic analysis of HIV type 1 primary isolates from western Cameroon.

In this study we report the molecular and biological characteristics of 19 HIV-1 primary isolates obtained in April 1999 from 47 HIV-1-infected individuals living mainly in western Cameroon. Discontinuous portions of gag, pol, and env were amplified by polymerase chain reaction and directly sequenced. Phylogenetic analysis of these sequences showed that all were of HIV-1 group M with the following genotypes: A(gag)/A(pol)/A(env) (n = 4), A(gag)/AG(pol)/AG(env) (n = 2), AG(gag)/A(pol)/AG(env) (n = 1), AG(gag)/U(pol)/AG(env) (n = 1), AG(gag)/AG(pol)/AG(env) (n = 6), G(gag)/G(pol)/G(env) (n = 3), F2(gag)/F2(pol)/F2(env) (n = 1), and a novel A(gag)/J(pro/rt)/A(int)/U(env) complex recombinant (n = 1). This A/J/U recombinant shared the same gag-pol cross-over point with known CRF02.AG viruses and 99CMBD6, an AG recombinant from our panel of isolates. The biological phenotype of most of the isolates correlated with the clinical status of the patient. Six isolates were syncytium inducing (SI) on MT-2 cells whereas 13 isolates were of the non-syncytium-inducing phenotype (NSI). Coreceptor usage by these isolates determined on GHOST cells correlated with their biological phenotype, as all SI isolates used CXCR4 and all NSI isolates used CCR5. Our results show a high predominance of subtype A (mainly CRF02.AG-like viruses) in western Cameroon and fewer HIV-1 subtypes compared with other parts of Cameroon. Genetic variability was, however, not reflected in the biological characteristics of the isolates. The presence of a novel A/J/U complex recombinant from this region further emphasizes the role of recombination in the global evolution of HIV.

Protease mutations in HIV-1 non-B strains infecting drug-naive villagers in Cameroon.

To describe the presence of protease inhibitor (PI) resistance-associated mutations and subtype distribution in drug-naive villagers of six provinces of Cameroon, we sequenced the protease (PR) gene (297 bp) of 128 viruses. Secondary PI resistance-associated mutations were identified at five sites: L10I/V (16%), K20R (8%), M36I (98%), L63P (13%), and V77I (6%). No primary mutation in the PR was identified. Of the 128 specimens analyzed, subtypes A (11%), C(2%), D (6%), F2 (3%), G (6%), H (0.8%), J (6%), and CRF02_AG (60%) were identified. The mutations identified were not characteristic to any particular subtype. The absence of primary mutations, in addition to the few secondary mutations, gives good perspectives for PI treatment interventions in these rural areas.

Defining human immunodeficiency virus (HIV) type 1 immunotypes with six human monoclonal antibodies.

Studies of HIV-1 immunological relatedness have revealed that genetic diversity does not parallel antigenic diversity and have recently shown that HIV-1 strains from different geographic regions from around the world can be grouped into a small number of immunologically defined groups (immunotypes). Previously, the binding patterns of 28 monoclonal antibodies (mAbs) (specific for V3 and C5 of gp120 and cluster I of gp41) with 26 HIV-1 virions obtained globally were determined in a virus binding assay. Analysis of the binding patterns of these 728 mAb/virus combinations now reveals that a particular subset containing six of the 28 mAbs can correctly immunotype 24 of the 26 isolates (92%) into three immunotypes. Like the original panel of mAbs, the subset of six mAbs identified was directed against epitopes in the V3 and C5 regions of gp 120 as well as cluster I of gp41. The binding patterns ("profiles") of these six mAbs with 24 additional HIV-1 virions from Cameroon confirmed that epitopes in V3 and C5 of gp120 and cluster I of gp41 are well exposed on these viruses. Multivariate analysis of the binding patterns of these six mAbs with all 50 viruses (26 obtained globally and 24 obtained from Cameroon) indicates that the viruses from Cameroon have binding profiles similar to viruses from the rest of the world and can be classified into the same three immunotypes that were previously described. This study suggests that a vaccine against HIV-1 need not be based on geographic origin of the virus or on clade, but may better be based on antigenic properties that classify the plethora of different HIV-1 viruses into immunologically defined groups.

Evaluation of HIV type 1 group O isolates: identification of five phylogenetic clusters.

HIV-1 group O strains have a level of genetic diversity similar to that of strains in group M; however, group O has not been readily classified into genetic subtypes. Phylogenetic classification of group O has been hindered by the limited sequence information available. To facilitate phylogenetic analysis, we sequenced the gag p24 (693 nt), pol p32 (864 nt), and env gp160 (approximately 2700 nt) genes from 39 group O-infected specimens. These specimens include 32 plasma samples collected in Cameroon between 1996 and 1999, 2 specimens collected in the United States, and 5 infections previously isolated in Equatorial Guinea. Phylogenetic analysis of HIV-1 group O sequences resulted in the identification of five clusters that are maintained across gag, pol, and env, generally supported by high bootstrap values, and approximately equidistant from each other. In addition to the group O clusters, several isolates branch independently and are equidistant from the other group O isolates. Cluster I comprises greater than 50% of the group O isolates and is a diverse set of isolates that is subdivided into subclusters. The average intra-, sub-, and intercluster distances for group O are similar to the corresponding distances for group M subtypes. The five group O clusters have characteristics similar to those of group M subtypes. Thus the data presented may form the basis for classification of group O into subtypes. However, full-length genomes representing each group O cluster will be required to formalize a group O subtype classification.

Near full-length genomes of 15 HIV type 1 group O isolates.

Human immunodeficiency virus type 1 (HIV-1) is classified into three distinct groups; M (major), N (non-M/non-O), and O (outlier). Group M strains are further subclassified into subtypes, subsubtypes, and circulating recombinant forms (CRF). While the level of genetic diversity within group O is similar to that between group M subtypes, group O has not been classified into subtypes. A previous study, based on the phylogenetic analyses of the gag p24, pol p32, and env gp160 sequences from 39 group O isolates, laid the foundation for the classification of group O subtypes. Five phylogenetic clusters, I-V, were identified that have characteristics analogous to group M subtypes. However, a complete phylogenetic analysis and classification of group O requires the availability of at least two full-length and one partial genomes for each group O phylogenetic cluster. In this study, 15 group O isolates were selected for full genome sequencing. Phylogenetic analysis of the 15 sequences with eight additional group O genomes supports the classification of three group O subtypes (I-III) and the potential existence of one CRF (IV) and at least one additional subtype (V). The group O subtypes are equidistant to each other and lack subsegments of other subtypes. The intra- and intersubtype genetic distances for group O are similar in magnitude to the corresponding distances for group M subtypes. Intersubtype recombination was identified in three of the 23 (13%) group O genomes. Formal classification of group O subtypes should be forthcoming pending the analysis of additional group O genomes and agreement of the HIV nomenclature committee.

Genetic and biological properties of HIV type 1 isolates prevalent in villagers of the Cameroon equatorial rain forests and grass fields: further evidence of broad HIV type 1 genetic diversity.

To understand the evolution of HIV-1, the genetic and biological characteristics of viruses that infect persons living in regions in which the virus has been evolving for several decades must be studied. Thus, we investigated teh genetic subtypes, coreceptor usage, and syncytium-inducing ability of viruses in 47 HIV-1-infected blood samples from individuals living in rural villages in the equatorial rain forest and grass field regions in Cameroon. Heteroduplex mobility analysis (HMA) of gag (part of p24 and p7) and env (C2V5) or sequence and phylogenetic analysis of gag (part of p24 and p7), pol (protease), and env (C2V5), revealed a broad HIV-1 group M genetic diversity. Subtype analysis revealed genetic evidence of seven subtypes (A, C, D, F, G, H, and J) and three circulating recombinant froms (CRFs) (CRF01_AE, CRF02_AG, and CRF11_cpx). Only 15 (32%) of the 47 samples analyzed revealed a concordant subtype in all three genes (gag, pol, and env), while discordant subtypes and CRFs were identified for the remaining 32 (68%) samples. Two patterns of HIV-1 diversity could be discerned in two provinces. While more concordant subtypes in gag, pol, and env genes were identified in villages of South province (10 of 13, 77%), the HIV-1 diversity in the West province was characterized by intersubtype recombinants (63%). Five new intersubtype recombinants were identified including Agag Jpol Genv, Ggag Upol Aenv, AGgag Jpol Aenv, Agag AGpol Henv, and Cgag AGpol AGenv. All of the 40 viruses tested used the R5 coreceptor, of which four also used the X4 coreceptor. Four viruses were able to induce syncytia in MT-2 cells, however, syncytium induction did not correlate with coreceptor usage. This study further reveals the complexity of HIV-1 infection in rural Cameron and points to the future of the global epidemic, which may be characterized by more genetically diverse viruses.

A randomized, double-blind, placebo-controlled safety and acceptability study of two Invisible Condom formulations in women from Cameroon.

BACKGROUND: The objectives of this clinical trial were to evaluate the safety, tolerance and acceptability of two gel formulations of the Invisible Condom: (i) the polymer alone and (ii) the polymer-containing sodium lauryl sulfate (SLS) compared to placebo when applied intravaginally with our unique applicator in sexually abstinent and active woman volunteers. STUDY DESIGN: A randomized, doubled-blind, placebo-controlled study in healthy women from Yaoundé, Cameroon. Two hundred sixty women were randomized into three gel arms: (a) gel alone, (b) gel plus SLS and (c) placebo gel. Thirty-seven sexually abstinent women applied gel intravaginally once a day for 14 days, while 75, 74 and 74 sexually active women applied gel intravaginally once, twice or three times daily for 14 days, respectively. RESULTS: Retention rate was high at 85% and 221 women applied the two products and the placebo for a total of 6005 times. Nugent score, H(2)O(2)-producing lactobacilli and vaginal pH were stable throughout the study and were not affected by the study products. Colposcopy showed neither genital ulceration nor mucosal lesions. No study product-related serious adverse events were reported. The majority of reported adverse events were mild or moderate and largely similar in all 3 arms. Satisfaction questionnaire showed that the gel formulations and applicator were generally comfortable and acceptable. CONCLUSION: The Invisible Condom formulations and applicator were found to be comfortable, well tolerated and acceptable when applied intravaginally once, twice or thrice daily for 14 days. Thus, expanded safety evaluation is warranted.

Determinants of retention in care in an antiretroviral therapy (ART) program in urban Cameroon, 2003-2005.

BACKGROUND: Retention in long-term antiretroviral therapy (ART) program remains a major challenge for effective management of HIV infected people in sub-Saharan Africa. Highly Active Antiretroviral Therapy (ART) discontinuation raises concerns about drug resistance and could negate much of the benefit sought by ART programs. METHODS: Based on existing patient records, we assessed determinants of retention in HIV care among HIV patients enrolled in an urban ART at two urban hospitals in Cameroon. Extended Cox regression procedures were used to identify significant predictors of retention in HIV care. RESULTS: Of 455 patients, 314 (69%) were women, median (IQR) age and baseline CD4 cell count were respectively 36 years (30 - 43) and 110 cells/µL (39 - 177). Forty patients (9%) had active tuberculosis (TB) at enrollment. After a median (IQR) follow-up of 18 months (10-18), 346 (75%) were still in care, 8 (2%) were known dead, and 101 (22%) were lost to follow-up (LFU). Severe immunosuppression (CD4 cell count ≤ 50 cells/µL) at baseline (aHR 2.3; 95% CI 1.4 - 3.7) and active tuberculosis upon enrollment (aHR 1.8; 95% CI 1.0 - 3.6) were independent predictors of cohort losses to follow-up within the first 6 months after HAART initiation. CONCLUSION: These data suggest that three-quarter of HIV patients initiated on HAART remained in care and on HAART by 18 months; however, those with compromised immunologic status at treatment initiation, and those co-infected with TB were at increased risk for being lost to follow-up within the first 6 months on treatment.

The prevalence of diverse HIV-1 strains was stable in Cameroonian blood donors from 1996 to 2004.

OBJECTIVE: The HIV epidemic in Cameroon is characterized by a high level of strain diversity despite a relatively low prevalence of infection. In this study, HIV strains infecting blood donors in Cameroon were characterized to determine the prevalence of subtypes and intersubtype recombinants and if strain prevalence was changing over time. METHODS: From 1996 through 2004, 676 HIV-infected blood donations were collected at blood banks in Douala and Yaoundé, Cameroon. A subset of the HIV-1 group M strains (n = 574) were classified based on phylogenetic analysis of viral sequences from the gag p24, pol integrase, and env gp41 regions. RESULTS: HIV-1 group M accounted for 97.3% (n = 658) of infections, whereas group O was present in 2.2% (n = 15) and HIV-2 in 0.4% (n = 3). Within the group M infections, 14 subtypes and circulating recombinant forms (CRFs) and unique recombinant forms (URFs) were identified. Overall, CRFO2_AG accounted for 58.2% of infections, URFs 14.8%, and levels of subtypes, A, B, C, D, F2, and G, and CRFs, 01, 06, 09, 11, 13, 22, and 37, varied from 0.2% to 6.1%. Evaluation of HIV strains present in the donor population over this 9-year period showed no substantial changes in the proportion of infections caused by each subtype and CRF, the percentage of intersubtype recombinants, or the strain composition of the URFs. CONCLUSIONS: HIV-1 strain diversity in Cameroon did not significantly change, suggesting a mature and relatively stable epidemic.

Evaluation of FDA licensed HIV assays using plasma from Cameroonian blood donors.

Several diagnostic assays for the detection of HIV infection have been approved and licensed by the FDA for blood donor screening. However, the performance of these assays is unknown when testing genetically divergent blood specimens. To evaluate the performance of these assays with diverse HIV strains, we chose to study specimens collected from blood donors in Cameroon where genetic diversity and recombinant variants are prevalent. In this study, we tested 240 human plasma specimens collected from two blood centers in Cameroon. These samples were screened initially in Cameroon for antibody to HIV using a rapid assay. We also performed sequencing to determine subtype. Our evaluation has demonstrated that HIV infection in most HIV plasma samples could be detected by most of the US FDA licensed diagnostic assays. With the exception of a few specimens, HIV-1 p24 antigen was not detected in any of the samples. In addition, some nucleic acid tests (NAT) assays were not able to detect a few serologic reactive samples and all new variants including some CRF02_AG variants.

Antiretroviral drug resistance and routine therapy, Cameroon.

Among 128 patients routinely receiving highly active antiretroviral therapy in an HIV/AIDS outpatient clinic in Cameroon, 16.4% had drug resistance after a median of 10 months. Of these, 12.5% had resistance to nucleoside reverse transcriptase inhibitors (NRTIs), 10.2% to non-NRTIs, and 2.3% to protease inhibitors.