To the Editor:
The introduction of highly active antiretroviral therapy (HAART) for treatment of AIDS patients significantly decreases AIDS-associated morbidity and mortality.1 Resistance to antiretroviral (ARV) drugs may arise during treatment, however, and is considered the main cause of therapy failure.2 Although the major reason for the onset of drug resistance is poor adherence to treatment, acquisition of resistant strains of HIV also contributes to therapeutic failure.3-5
Brazil was the first country to provide free-of-cost ARV drugs for HIV therapy. This raised concerns about the selection and dissemination of drug-resistant HIV because of poor adherence and the extensive exposure to ARV drugs. Nevertheless, the Brazilian AIDS program is considered a success, and the available data show a low rate of HIV-1 drug resistance among drug-naive patients.6
To determine the prevalence of resistance mutations among recently diagnosed HIV-infected drug-naive patients, we evaluated plasma or HIV-1 RNA samples stored at the Retrovirology Laboratory of the Federal University of Bahia Hospital in Salvador, Brazil. The samples were obtained from the plasma of recently diagnosed patients from our AIDS clinics and from local blood banks after the year 2000. The institutional Ethics Review Board approved the study.
A total of 140 samples were tested for HIV resistance: 8 came from blood donors, and 132 came from our AIDS clinics (including 26 samples from vertically infected children). All samples were collected between 2000 and 2004 and were stored in a freezer at −80°C.
Resistant mutations were assessed in HIV-1 reverse transcriptase (RT) and protease (Pro) regions using the TruGene sequencing system (Visible Genetics, Toronto, Canada).
A multiple sequence alignment of the examined region and the related sequences in the GenBank/European Molecular Biology Laboratory (EMBL) database was performed with Clustal X software (available at http://ftp.ebi.oc.uk/pub/software/dos/clustalX) and further edited with the GeneDoc program (available at http://psc.edu/biomed/genedoc). The substitution model was selected with Modeltest 3.06 (available at http://darwin.uvigo.es/software/modeltest.html). Neighbor-joining (NJ) and maximum-likelihood (ML) phylogenetic analyses were performed using PAUP*4.0b10 (available at http://paup.csit.fsu.edu). The reliability of the NJ trees was evaluated by analyzing 1000 bootstrap replicates. A likelihood ratio test was used to calculate the statistical support for the branches (expressed in P values). The recombination analysis was performed using Simplot 3.5.1 (available at http://sray.med.som.jhml.edu/SCROftware/simplot).
To confirm the HIV-1 subtype assignment, we sequenced an env fragment of 34 samples that had enough remaining material for testing through reverse transcription of RNA samples using the enzyme Superscript III (Invitrogen, Carlsbad, CA). The resulting complementary DNA (cDNA) was subjected to nested polymerase chain reaction (PCR) amplification using V3 to V5 outer primers for the first round, with 3 options of primers for the second round. After sequencing the reaction, the samples were analyzed using the software Sequencing Analysis version 3.7 (Applied Biosystems, Foster City, CA) and Sequencher version 4.2 (available at http://sequencher.bio.indiana.edu).
We detected 39 different mutations associated with HIV resistance, totaling 307 amino acid substitutions. The most prevalent mutation was L63P (present in 48% of samples), whereas M36I was detected in 52 (39%) samples and L10V/I was found in 30 (23%) samples. Twenty-five (18.9%) samples presented at least 1 drug-resistant mutation: 15 samples (11.4%) had at least 1 mutation related to resistance to nonnucleoside reverse transcriptase inhibitors (NNRTIs), 13 (9.8%) to nucleoside reverse transcriptase inhibitors (NRTIs), and 7 (5%) to protease inhibitors (PIs). We found genotypic resistance to NRTIs and PIs in 3.8% of samples and to NRTIs and NNRTIs in 3.0%. Mutations associated with resistance to all NRTIs were detected in 6 (4.8%) samples, to all NNRTIs in 13 (9.8%) samples, and to all PIs in 3 (2.4%). Three patients presented resistance to all NRTIs and all PIs, preserving only NNRTIs as a therapeutic option. In addition, 2 other samples showed complete resistance to NRTIs and NNRTIs, leaving only PIs as active drugs.
Among the 26 children, we detected resistance mutations in 7 (26.9%): 5 presented with NNRTI resistance; 3 had genotypic evidence of resistance to NRTI; and 1 showed the mutation I50V in the Pro region, which is related to resistance to amprenavir. Two children showed resistance mutations to NRTIs and NNRTIs. We evaluated 1 mother-child pair: the mother's genotypic profile showed the mutations V75M, M184V, D30N, M36I, L63P, and N88D, but the analysis of the sample obtained from her baby revealed that she transmitted an HIV-1 strain harboring only the M184V mutation.
The phylogenetic analysis of the total pol region shows that of all the samples, 117 (84%) belong to subtype B with a low bootstrap value but highly significant ML (P < 0.001); 12 (8%) were subtype F and 3 (2.5%) were subtype C, both with P < 0.001 for ML. Twenty-one strains were classified as recombinants: 20 (14%) B/F and 1 (0.5%) B/C, with 61% and 89% bootstrap values, respectively (P < 0.001 for ML). Because this is the first time a study has found subtype C virus in Salvador, this may suggest a recent introduction of this subtype in the city. The presence of the 2 B/F recombinant clusters (69% and 100% bootstrap values and P < 0.001) suggests at least 2 distinct recombinant events. The bootscanning analysis of the B/F strains showed 2 different profiles of these 2 recombinant clusters, reinforcing that possibility.
Brazil has a much lower prevalence of HIV primary resistance than other developed countries do.6 This provides a strong argument favoring the Brazilian policy of ARV drugs. Preliminary findings in Bahia have already raised some concerns about the real extent of drug resistance, however. A previous report detected 20% azidothymidine (AZT) resistance in peripheral blood mononuclear cells (PBMCs) of drug-naive subjects using a line probe assay (LiPA).7
The present work shows a worrisome scenario. The prevalence of primary genotypic drug resistance is much higher than previously estimated for Brazilian patients, and for the first time, we see evidence of resistance to more than 1 class of ARV drugs in drug-naive patients. The high prevalence of resistance to NNRTIs in our study is in accordance with what is generally seen in other parts of the world.8,9 This is a likely consequence of the wide use of these drugs in first-line regimens and the tendency of NNRTI mutations to persist because they do not significantly compromise the replicative capacity of the virus.10 Also, the known low genetic barrier of that class of drugs contributes to the easier selection of resistance in a setting of nonstrategic interruption of treatment. The high frequency of resistance to other drug classes was surprising, however.
The high rate of genotypic resistance among drug-naive children vertically infected was also unexpected. The Brazilian guideline still accepts prophylaxis with AZT or AZT plus lamivudine (3TC) for pregnant women who do not fulfill the criteria for ARV treatment. Our results probably reflect the late diagnosis of HIV infection for most mothers, absence of proper prophylaxis against vertical transmission, and poor adherence among those who used ARV drugs but did not avoid mother-to-child transmission.
The analysis of viral subtypes provides additional evidence supporting the previous report on the circulation of non-B subtypes in Brazil. Despite the fact that only one third of our samples could be sequenced for subtype confirmation, the finding of 12 subtype F, 3 subtype C, and 21 recombinant forms confirms the high genetic diversity of HIV-1 in this region. The finding of an almost complete resistance profile in 1 patient harboring a subtype F strain is intriguing. It has been reported that in non-B subtypes, resistance to specific ARV drugs could evolve along different pathways than in B strains.11 This subtype F case showed complete resistance to all but NNRTI drugs, raising the possibility that multidrug-resistant non-B subtypes are being transmitted.
The present report indicates that HIV primary resistance in Brazil may be underestimated and suggests the need for evaluating specific populations. Our findings reinforce the recent decision from the Brazilian Ministry of Health making mandatory the perinatal screening for HIV-1 antibodies among women not tested during pregnancy, along with resistance testing for those women with a viral load greater than 5000 RNA copies/mL. Resistance testing for children of HIV-infected mothers may be required to avoid future therapy failure because of transmitted HIV-1-resistant strains. The high prevalence of HIV-1 B/F recombinant strains in this population is also of note. Given these peculiarities, monitoring the genetic diversity of HIV-1 is probably important for the development of vaccines and treatment strategies.
The authors acknowledge Bayer for kindly providing the tests and equipment for viral sequencing. They also thank Michele Portela for her technical support in the evaluation of the samples.
Célia Pedroso, BSc*
Artur T. L. Queiroz, MD†
Luis C. Alcântara, MD, PhD†
Jan F. Drexler, MD*
Ricardo S. Diaz, MD, PhD‡
Carlos Brites, MD, PhD*
*Infectious Diseases Service Federal University of Bahia Hospital Salvador, Bahia, Brazil
†Advanced Public Health Laboratory Gonçalo Moniz Research Center Oswaldo Cruz Foundation (LASP/CPqGM/FIOCRUZ) Salvador, Bahia, Brazil
‡Federal University of São Paulo São Paulo, Brazil
§CREAIDS (Bahia State AIDS Referral Center)
1. Palella FJ Jr, Deloria-Knoll M, Chmiel JS, et al, for the HIV Outpatient Study Investigators. Survival benefit of initiating antiretroviral therapy in HIV-infected persons in different CD4+ cell strata. Ann Intern Med
2. Idemyor V. Human immunodeficiency viruses and drug therapy: resistance and implications for antiretroviral therapy. AIDS Rev
3. de Mendoza C, Gallego O, Soriano V. Mechanisms of resistance to antiretroviral drugs-clinical implications. Pharmacotherapy
4. Boden D, Hurley A, Zhang L, et al. HIV-1 drug resistance in newly infected individuals. JAMA
5. Bangsberg DR, Perry S, Charlebois ED, et al. Non-adherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS
6. Soares MA, Brindeiro RM, Tanuri A. Primary HIV-1 drug resistance in Brazil. AIDS
. 2004;18(Suppl 3):S9-S13.
7. Brites C, Bahia F, Gilbert M, et al. Evaluation of resistance to reverse transcriptase inhibitors (RTI) in HIV-1-infected patients, before and after 6 months of single or double antiretroviral therapy. Braz J Infect Dis
8. Little SJ, Holte S, Routy JP, et al. Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med
9. Ross LL, Lim ML, Liao Q, et al. Prevalence of antiretroviral drug resistance and resistance mutations in antiretroviral therapy naive HIV-infected individuals from 40 US cities during 2003. Presented at: 44th Interscience Conference on Antimicrobial Agents and Chemotherapy; 2004; Washington, DC.
10. Wrin T, Gamarnik A, Huang W, et al. Drug resistance is associated with impaired PR and RT function and reduced replication capacity: characterization of recombinant viruses derived from 200 HIV-1 infected patients [abstract 2116]. Presented at: 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Canada, 2000.
11. Spira S, Wainberg MA, Loemba H, et al. Impact of clade diversity on HIV-1 virulence, antiretroviral drug sensitivity and drug resistance. J Antimicrob Chemother