CONCISE COMMUNICATION

Transcription

1 1688 CONCISE COMMUNICATION Vertical Transmission of Multidrug-Resistant Human Immunodeficiency Virus Type 1 (HIV-1) and Continued Evolution of Drug Resistance in an HIV-1 Infected Infant Victoria A. Johnson, 1 Christos J. Petropoulos, 2 Charles R. Woods, 3 J. Darren Hazelwood, 1 Neil T. Parkin, 2 Carol D. Hamilton, 4 and Susan A. Fiscus 5 1 University of Alabama at Birmingham School of Medicine and Birmingham Veterans Affairs Medical Center; 2 ViroLogic, South San Francisco, California; 3 Wake Forest University, Winston-Salem, 4 Duke University Medical Center, Durham, and 5 University of North Carolina at Chapel Hill To confirm the vertical transmission of multidrug-resistant (MDR) human immunodeficiency virus type 1 (HIV-1) and to assess its impact on further evolution of drug-resistant virus in an infant, proviral DNA amplified from infected peripheral blood mononuclear cell cultures was sequenced to identify reverse transcriptase (RT) and protease (PR) mutations. The infant had proviral DNA with evidence of RT mutations (M41L, L74V, and T215Y) and 3 PR substitutions (K20R, M36I, and V82A). After delivery, the mother s proviral DNA had the same substitutions. Phylogenetic analyses of these HIV-1 RT and PR sequences indicated epidemiological linkage. Plasma drug susceptibility was determined by using a recombinant virus assay. Plasma HIV-1 obtained after the infant s birth demonstrated reduced susceptibility to zidovudine and ritonavir. Thus, vertical transmission of MDR HIV-1 was demonstrated in the setting of detectable maternal plasma viremia. Further accumulation of broad MDR in the infant s virus to 3 antiretroviral classes occurred, despite postnatal therapy. Horizontal and vertical transmission of drug-resistant human immunodeficiency virus type 1 (HIV-1) during primary HIV-1 infection has been documented. Hecht et al. [1] first described the horizontal transmission of an HIV-1 variant resistant to multiple reverse transcriptase inhibitors (RTIs) and protease inhibitors (PIs). In contrast, vertical transmission of multidrugresistant (MDR) HIV-1 has not been described. During an on- Received 31 August 2000; revised 6 February 2001; electronically published 23 April Presented in part: 6th National Conference on Retroviruses and Opportunistic Infections, Chicago, 31 January 4 February 1999 (abstract 266). This project was approved by institutional review boards for the rights of human subjects at the University of Alabama at Birmingham School of Medicine, University of North Carolina at Chapel Hill, and Wake Forest University. Financial support: National Institutes of Health (grants AI-40876, AI , and AI to V.A.J.; AI and HL to C.D.H.); Core Laboratory Research Facilities, University of Alabama at Birmingham (UAB) School of Medicine, UAB Center for AIDS Research (CFAR), and Birmingham Veterans Affairs Medical Center to V.A.J.; Pediatric AIDS Clinical Trials Group (contract 97PVCL06) and University of North Carolina CFAR (P30HD37260) to S.A.F; Metrolina AIDS Project, Ryan White Title IV (#6 H12 HA00061) to C.R.W.; Centers for Disease Control and Prevention (CDC; grant ) and Department of Veterans Affairs via CDC Memorandum of Understanding to C.D.H. Reprints or correspondence: Dr. Victoria A. Johnson, Division of Infectious Diseases, University of Alabama at Birmingham School of Medicine, THT 229, rd Ave. S., Birmingham, AL (vjohnson@ uab.edu). The Journal of Infectious Diseases 2001;183: by the Infectious Diseases Society of America. All rights reserved /2001/ $02.00 going population-based retrospective investigation of perinatal HIV-1 transmission in North Carolina, we identified an HIV- 1 infected infant who had evidence of vertical MDR HIV-1 transmission. We also documented continued viral evolution over time during sequential therapies of the infant. Subjects and Methods Mother. The mother was diagnosed with HIV-1 infection in July Her initial CD4 T lymphocyte count was 281 cells/ mm 3, and percentages remained 28%. She initiated zidovudine (Zdv) at a dose of 500 mg/day in August This was increased to 600 mg/day in June From , she received sequential monotherapy or combined therapies, including Zdv, lamivudine (3TC), stavudine (), didanosine (ddi), indinavir (Idv), and ritonavir (Rtv) (table 1). She had variable medication adherence. Plasma HIV-1 RNA levels were ,380 copies/ml. Although she had no AIDS-defining illnesses, she failed to gain weight and miscarried in June She had a subsequent full-term pregnancy from June 1997 to January 1998; therapies included Zdv/ Rtv/ or Zdv/Rtv/ddI. In November 1997, she stopped ddi because of nausea and restarted on her own, combined with Zdv/ Rtv. Her plasma HIV-1 RNA level was 102,500 copies/ml. Intrapartum intravenous Zdv therapy was added during the infant s delivery in January 1998, in accordance with the Pediatric AIDS Clinical Trials Group (ACTG) protocol 076 regimen [4]. Infant. The infant was born by spontaneous vaginal delivery at 38.4 weeks of gestation; the mother s membranes had ruptured 12 h earlier. The infant received the complete ACTG 076 regimen [4] and never was breast-fed postpartum. The infant tested positive

2 JID 2001;183 (1 June) Vertical Transmission of MDR HIV Table 1. Subject, month/year Mother and infant human immunodeficiency virus (HIV) disease status and antiretroviral history. CD4 T lymphocyte count, cells/mm 3 (%) Plasma HIV-1 RNA level, copies/ml Prior antiretroviral therapy and key events Available peripheral blood specimens Mother 7/ (24) HIV diagnosis; started Zdv monotherapy 2/ (27) Zdv 1/ (24) Zdv 4/94 Plasma a 2/ (22) Zdv 5/95 Plasma a 8/ (20) Patient requested continued monotherapy 11/95 Zdv; 3TC added 2/ (16) Zdv and 3TC 4/ (21) Zdv; patient reports stopping 3TC because of nausea sometime in past several months 6/ (22) 62,320 Zdv; added Idv; 8/96 patient self-decreased dose to twice daily because of nausea; then learned she was pregnant; then miscarried 10/ (20) Zdv; Rtv and added 12/ (21) 527 Zdv, Rtv, and 2/97 18,017 Zdv, Rtv, and 6/ (21) 61,307 Zdv; patient learned she was pregnant and self-decreased Rtv and 9/ (17) 364,380 Zdv; patient convinced to restart Rtv, and ddi added (instead of ) 11/ (14) 102,500 Zdv; patient self-decreased Rtv to 500 mg every 12 h, stopped ddi because of nausea, and self-added back 4/ (14) 134,036 Delivered of child 1/98 with addition of intrapartum intravenous Zdv; had selfdiscontinued Rtv; switched to new regimen of, 3TC, and postpartum 6/ (16) 4889, 3TC, and 7/98 43,996, 3TC, and Plasma a ; PBMC b 8/ (18) 25,303, 3TC, and 10/ , 3TC, and Plasma a ; PBMC b Infant 1/98 Month of birth 2/98 Zdv monotherapy since birth (1/98) including Plasma a ; PBMC b intrapartum intravenous Zdv regimen 3/ (32) 1,751,000 Zdv discontinued; started new regimen of Plasma a ; PBMC b 3TC,, and Nfv 4/ ,391 3TC,, and Nfv 5/98 378,135 3TC,, and Nfv; added Plasma a ; PBMC b 7/ ,550 3TC,, Nfv, and Plasma a ; PBMC b 9/ ,112 3TC,, Nfv, and 10/ ,169 3TC,, Nfv, and Plasma a ; PBMC b NOTE. 3TC, lamivudine; ddi, didanosine;, stavudine; Idv, indinavir; Nfv, nelfinavir;, nevirapine; PBMC, peripheral blood mononuclear cells; Rtv, ritonavir; Zdv, zidovudine;, no data available. a Plasma virus phenotype by Phenosense HIV assay [2]. b PBMC virus genotype by Affymetrix Genechip nucleic acid sequencing [3]. for HIV-1 infection by peripheral blood mononuclear cell (PBMC) coculture and HIV-1 DNA in February 1998, while receiving Zdv monotherapy. Results were confirmed in March Virus load was HIV RNA copies/ml, and CD4 T lymphocyte count was 3090 cells/mm 3 (32% total lymphocytes). In March 1998, Zdv monotherapy was switched to 3TC//nelfinavir (Nfv). Nevirapine () was added in May Despite aggressive therapy, the plasma HIV-1 RNA levels never fell below 76,169 copies/ml. Patient specimens. Sequential peripheral blood specimens were obtained from the mother and the infant. Before delivery, 2 plasma samples were available from the mother (April 1994 and May 1995). After delivery, anticoagulated peripheral blood samples were collected from the mother and the infant and were processed for plasma and virus recovery. Samples obtained nearest to the infant s birth in January 1998 were a July 1998 specimen from the mother and a February 1998 specimen from the infant. Virus load. PBMC HIV-1 DNA was determined by polymerase chain reaction assay, and plasma HIV-1 RNA levels were determined by Amplicor HIV-1 Monitor test (both Roche Diagnostic Systems).

3 1690 Johnson et al. JID 2001;183 (1 June) HIV-1 drug-resistant mutations. The entire protease (PR) coding region and the first 242 codons of reverse transcriptase (RT) were sequenced by GeneChip (Affymetrix) nucleic acid sequence technology by using proviral DNA amplified from PBMC cocultures. Sequences were aligned with reference sequences from the HIV Sequence Database by use of CLUSTALX Multiple Sequence Alignment Program (version 1.8; available at The final sequence alignment was manually adjusted for optimal alignment, and gaps were stripped to ensure that an equal number of bases was compared. Pairwise evolutionary distances were estimated by using Kimura s 2-parameter method, to correct for superimposed substitutions. We constructed phylogenetic trees by the neighbor-joining method. Reliability of topologies was estimated by performing bootstrap analyses with 1000 replicates. HIV-1 drug susceptibility testing. HIV-1 drug susceptibility was determined by the PhenoSense HIV-1 assay (ViroLogic) by using uncultured patient plasma, as described elsewhere [2]. Patient viruses with IC 50 values 2.5-fold above the reference virus IC 50 values were considered to have reduced drug susceptibility for all drugs, except those with clinically determined cutoff values: 1.7- fold for ddi and and 4.5-fold for abacavir. Results Genotypic analyses of HIV-1 RT- and PR-coding regions from mother and infant PBMC-derived viruses. Table 2 depicts HIV-1 RT mutations in PBMC-derived isolates from the mother and infant. At delivery in January 1998, the mother was switched from Zdv//Rtv to 3TC//. Two samples, dated July and October 1998, were available from the mother. The July 1998 virus (GenBank accession no. AF333243) contained the M41L, T215Y, L74V, and V106A mutations, which are associated with reduced susceptibility to Zdv and, ddi, and nonnucleoside RTIs (NNRTIs), respectively. The October 1998 virus (GenBank accession no. AF333244) acquired the L210W, M184V, and K103N mutations, which are associated with reduced susceptibility to Zdv and, 3TC, and the NNRTI class, respectively. Four samples, dated February, May, July, and October 1998, were available from the infant. The February 1998 virus (GenBank accession no. AF333239) contained the M41L, L74V, and T215Y mutations that also were identified in the maternal sample. The infant received postnatal Zdv treatment; therefore, the possibility that M41L and T215Y mutations emerged after birth cannot be eliminated. However, this seems unlikely, given the short duration of postnatal Zdv exposure at the February 1998 time point. Because the infant did not receive ddi therapy before or after birth, the presence of the L74V mutation strongly suggests mother-to-infant transmission of ddi-resistant virus. Therapy was switched to 3TC//Nfv in March The May 1998 virus (GenBank accession no. AF333240) acquired the 3TC resistance mutation 184V, whereas the M41L, L74V, and T215Y RT mutations persisted. Therapy was intensified with. The July 1998 virus Table 2. Mutations in peripheral blood mononuclear cell derived human immunodeficiency virus 1 reverse transcriptase (RT) and protease (PR) coding regions selected by RT inhibitors and PR inhibitors over time (month/year). Coding region, wild type Mother Infant 7/98 10/98 2/98 5/98 7/98 10/98 RT M41 L L L L L L L74 V a V V V V V K103 a N a N V106 a A a A A Y181 a C a C/G/Y M184 a V a V V V L210 W T215 a Y a Y Y Y Y Y PR L10 I I K20 R R R R R R M36 I M/I I I I M/I V82 a A a A A A A A NOTE., Wild type. a Primary drug-selected mutations [5]. (GenBank accession no. AF333241) acquired the NNRTI mutation Y181C, whereas the M41L, L74V, M184V, and T215Y mutations persisted. By October 1998 (GenBank accession no. AF333242), the NNRTI mutations K103N, V106A, and a mixture of Y181C/G/Y were identified, and the nucleoside RTI (NRTI) mutations M41L, L74V, M184V, and T215Y persisted. Table 2 depicts PR substitutions in the HIV-1 isolates derived from PBMC samples from the mother and infant. The mother s July 1998 virus obtained during /3TC/ therapy contained the V82A mutation, despite discontinuation of PR inhibitor (PI) therapy at delivery in January The V82A mutation is strongly associated with Rtv and Idv resistance and possible saquinavir and Nfv cross-resistance. Other substitutions (i.e., L10I, K20R, and M36I) were evident and persisted in October The infant s February 1998 virus also contained the V82A mutation detected in the mother s July 1998 virus sample, even though the infant never received PI therapy. We interpret this as evidence of mother-to-infant transmission of PI-resistant virus. Other substitutions at K20R and M36I also seen in the maternal isolates were identified in the infant s virus, although a third, L10I, was not. Thus, 3 amino acid substitutions associated with resistance to PIs were identified in both the mother and the child at the time points closest to delivery. These substitutions persisted in the infant s virus at all subsequent time points during ongoing Nfv therapy. Phylogenetic relationships of HIV-1 RT and PR coding sequences from mother and infant PBMC-derived viruses. HIV- 1 RT and PR sequences derived from mother and infant isolates were very closely related (i.e., 198% sequence identity). Moreover, phylogenetic analyses revealed a tight cluster of mother and infant viral sequences that was supported by 100% of bootstrap values (figure 1). These data confirm the authenticity of the virus isolates under study.

4 JID 2001;183 (1 June) Vertical Transmission of MDR HIV The infant s February 1998 virus exhibited reduced phenotypic susceptibility to Zdv (31-fold) and Rtv (6-fold). This virus lacked the L74V mutation and was susceptible to ddi, although the PBMC-derived virus at this time point contained the L74V mutation. This likely reflects the presence of an archival form of ddi-resistant virus in PBMC that was no longer circulating in plasma. Discontinuation of maternal ddi therapy 2 months before delivery is consistent with this interpretation. Despite the infant s switch from Zdv monotherapy to /3TC/Nfv in March 1998, plasma HIV-1 RNA levels decreased to only 317,391 copies/ml in April The May 1998 specimen had reduced Zdv susceptibility (4-fold), reduced ddi susceptibility (2-fold), reduced susceptibility (1.8-fold), high-level 3TC resistance (1180-fold), abacavir cross-resistance (5-fold), and persistently reduced Rtv susceptibility (5-fold). The increased phenotypic Zdv susceptibility in the presence of Zdv and 3TC resistance conferring RT mutations (table 2) likely reflected the resensitization of Zdv susceptibility by the M184V mutation. In May 1998, was added to the infant s 3TC//Nfv therapy. By July 1998, there was reduced susceptibility (161-fold), delavirdine (Dlv) cross-resistance (14-fold), and persistently reduced ddi, 3TC, and Rtv susceptibility. By October 1998, broad NNRTI resistance was selected to (1400-fold), Dlv (47-fold), and efavirenz (3-fold). Thus, in!1 year, the infant s virus exhibited significantly reduced susceptibility to NRTIs, NNRTIs, and PIs. Figure 1. Phylogenetic relationships of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase and protease coding sequences obtained from mother and infant virus isolates at different time points. Sequences were aligned with subtype B reference sequences by using CLUSTALX Multiple Sequence Alignment Program (version 1.8; available at Phylogenetic tree was constructed using the neighbor-joining method and was rooted by using C.IN.94IN11246 as an out-group. Bootstrap values at each node represent percentage of 1000 bootstrap replicates that support the branching order. Only bootstrap values 70% are shown. Scale bar represents 0.01 nucleotide substitution per site. Maternal-infant specimens collected in July (mother, 07/98) and February 1998 (infant, 02/98) were closest to the infant s birth in January Additional specimens were obtained from the infant in May, July, and October 1998 (infant, 05/98, 07/98, and 10/98) and from the mother in October 1998 (mother, 10/98). Subtype B and C reference sequences from the Los Alamos HIV- 1 Sequence Database are provided as controls: B.US.JRCSF, B.US.JRFL, B.US.MNCG, B.US.NY5CG, B.US.DH123, and C.IN.94IN Drug susceptibility testing of plasma-derived HIV-1 isolates from mother and infant. Table 3 depicts phenotypic testing results from plasma samples obtained from both mother and infant. By delivery in January 1998, the mother had received continuous Zdv therapy for 17 years. Phenotypic testing of 2 early maternal plasma samples collected 45 months (April 1994) and 32 months (May 1995) before delivery showed large reductions in Zdv susceptibility (43-fold and 49-fold, respectively). Discussion This study documents vertical MDR HIV-1 transmission. The infant s virus had multiple RT and PI mutations in the first specimen obtained after birth at a time the child was receiving only postpartum Zdv therapy. PBMC-derived viruses obtained from the mother 6 and 9 months after delivery contained the same RT and PR mutations detected in the infant s first specimen. HIV-1 obtained from the infant after birth exhibited reduced susceptibility to both Zdv and Rtv. Phylogenetic analyses also indicated epidemiological linkage. Taken together, these data support vertical MDR HIV-1 transmission. Because the baby was not breast-fed postpartum, this possible route of perinatal HIV-1 acquisition was highly unlikely. One contributing factor for vertical MDR HIV-1 transmission was the mother s advanced HIV-1 disease stage during pregnancy. Elevated plasma viremia levels are associated with increased risk of vertical HIV-1 transmission [6, 7]. In our study, maternal plasma HIV-1 RNA levels were 1500 copies/ml at least 6 times after 1990 and were 61, ,380 copies/ml in the last 7 months of pregnancy. In contrast to our current practice of treating with triple-drug regimens, this mother began treatment over a decade ago. Consequently, she received sequential 1- and 2- drug regimens that are now known to be suboptimal. Rupture of the mother s placental membranes occurred at least 12 h before delivery, which also increased HIV-1 transmission risk [8]. An-

5 1692 Johnson et al. JID 2001;183 (1 June) Table 3. Mother-to-infant transmission of plasma-derived human immunodeficiency virus with reduced susceptibility to zidovudine (Zdv) and ritonavir (Rtv) by sampling time (month/year). Drug Mother Fold resistance vs. control (NL4-3) a Infant 4/94 5/95 7/98 10/98 2/98 3/98 5/98 7/98 10/98 NRTI Zdv ddc ddi TC Abc NNRTI Dlv Efv PI Rtv Idv Nfv Sqv Therapy at time of sampling Zdv Zdv 3TC 3TC Zdv Zdv 3TC Nfv 3TC Nfv 3TC Nfv NOTE. Values considered to have reduced drug susceptibility are in bold. 3TC, lamivudine; Abc, abacavir; ddi, didanosine; Dlv, delavirdine;, stavudine; Efv, efavirenz; Idv, indinavir; Nfv, nelfinavir; NRTI, nucleoside reverse-transcriptase inhibitor; NNRTI, non-nrti;, nevirapine; PI, protease inhibitor; Sqv, saquinavir. a Virus IC 50 values 2.5-fold greater than reference virus IC 50 values are considered to have reduced drug susceptibility for all drugs, except for those with clinically determined cutoff values: 1.7-fold for ddi and and 4.5-fold for Abc [2]. other contributing factor was medication intolerance and variable adherence to her complex multidrug regimens. Combined Zdv/, which the patient chose to self-administer, may result in pharmacologic antagonism. There also was evidence of maternal Zdv-resistant HIV-1 for at least 3 years before delivery. The mother received prolonged Zdv therapy from 1990 to During her 1998 pregnancy, she was treated with Zdv [9 11]. Zdv administration is recommended during the intrapartum period for HIV-infected pregnant women, regardless of the antepartum regimen or evidence of maternal Zdv resistance [11]. Zdv likely will remain a part of the recommended regimen for perinatal HIV-1 transmission prophylaxis until studies demonstrate comparable safety and activity of other regimens. In our study, Zdv likely was an ineffective barrier to vertical HIV-1 transmission. This supports the potential clinical utility of drug resistance testing in HIV-1 infected pregnant women, at least in women with extensive prior antiretroviral therapy and poorly controlled plasma viremia. This may allow for the identification of active agents and may best prevent vertical HIV-1 transmission. This infant s MDR HIV-1 population continued to accumulate RT mutations during ongoing plasma viremia, which resulted in greater reductions in drug susceptibility. Although the child remains clinically well to date, the long-term consequences of drug resistance to all 3 HIV-1 therapeutic classes are unknown. Future treatment options are also unclear and likely depend on new agent availability. This child s outcome may well differ from reports in previous clinical trials and observational studies [12 14] where more limited drug resistance (i.e., Zdv alone) [14] was evident in infants isolates at the time of vertical HIV-1 transmission. We identified this vertical MDR HIV-1 transmission episode during a review of perinatal HIV-1 transmission in North Carolina [15]. Few instances of vertical drug-resistant HIV-1 transmission were detected, and only this case of vertical MDR HIV-1 transmission was identified, perhaps reflecting minimal prior Zdv or antiretroviral maternal exposure. Although, in the case described, HIV-1 isolates were not obtained from the mother and infant at the exact time of delivery, the viruses were well characterized both genotypically and phenotypically at multiple time points, including determination of the mother s viral resistance status several years before delivery. Phylogenetic analyses also demonstrated the close relationships of HIV-1 RT and PR coding sequences at all time points, which indicated epidemiological linkage. Our retrospective analyses were limited by specimen availability. We did genotypic analyses by using proviral HIV-1 DNA from cultured infected PBMC and phenotypic analyses with uncultured plasma HIV-1 RNA. Despite this, there was only one discordant finding between the PBMC- and plasmaderived viruses at the RT L74 position, as mentioned above. Of note, the viral genotypic and phenotypic approaches used in this study are likely to assess majority or predominant virus populations. Therefore, clinically significant minority populations may have been missed, and the degree of MDR viruses

6 JID 2001;183 (1 June) Vertical Transmission of MDR HIV may have been underestimated. As increasing numbers of HIV- 1 infected women throughout the United States accumulate longer duration of antiretroviral exposure and also elect to become pregnant, it is likely that more cases of vertical MDR HIV-1 transmission will be identified. We have documented vertical MDR HIV-1 transmission in an infant. Future studies should assess prevalence and incidence of MDR HIV-1 in antiretroviral-experienced pregnant women and their impact on vertical HIV-1 transmission. These data encourage investigation of alternative agents and non Zdvcontaining combined regimens among antiretroviral-experienced pregnant women. These studies will require long-term safety monitoring of antiretroviral-experienced mothers and infants for potential combined toxicities. Our study also supports the role of viral drug resistance testing in antiretroviral-experienced pregnant women. Clinical application of such data will be particularly relevant when Zdv-resistant HIV-1 isolates are detectable before and during pregnancy and when maternal plasma viremia is not well controlled. Acknowledgments We thank Beatrice H. Hahn, Feng Gao, and Ruth A. Rhodes (University of Alabama at Birmingham [UAB] School of Medicine) and Colombe Chappey (ViroLogic, South San Francisco, California) for phylogenetic sequence analysis assistance; Jennifer L. Koel, Ruth A. Rhodes, and Clinton D. Nail (UAB School of Medicine), Ada Cachafeiro and Melissa Kerkau (University of North Carolina, Chapel Hill), and Kay Limoli, Wei Huang, Gabrielle Heilek-Snyder, and Jeannette M. Whitcomb (ViroLogic) for technical assistance; Ross McKinney and Megan Valentine (Duke University, Durham, North Carolina) and Dara Garner-Edwards (North Carolina Baptist Hospital, Winston- Salem) for patient care and medical chart review; and Jennifer Bell (UAB School of Medicine) for manuscript preparation. References 1. Hecht FM, Grant RM, Petropoulos CJ, et al. Sexual transmission of an HIV- 1 variant resistant to multiple reverse-transcriptase and protease inhibitors. N Engl J Med 1998;339: Petropoulos CJ, Parkin NT, Limoli KL, et al. A novel phenotypic drug susceptibility assay for human immunodeficiency virus type 1. Antimicrob Agents Chemother 2000;44: Kozal MJ, Shah N, Shen N, et al. Extensive polymorphisms observed in HIV-1 clade B protease gene using high-density oligonucleotide arrays. Nat Med 1996;2: Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group. N Engl J Med 1994;331: Hirsch MS, Brun-Vezinet F, D Aquila RT, et al. Antiretroviral drug resistance testing in adult HIV-1 infection: recommendations of an International AIDS Society USA panel. JAMA 2000;283: Garcia PM, Kalish LA, Pitt J, et al. Maternal levels of plasma human immunodeficiency virus type 1 RNA and the risk of perinatal transmission. Women and Infants Transmission Study Group. N Engl J Med 1999;341: Mofenson LM, Lambert JS, Stiehm ER, et al. Risk factors for perinatal transmission of human immunodeficiency virus type 1 in women treated with zidovudine. Pediatric AIDS Clinical Trials Group Study 185 Team. N Engl J Med 1999;341: Van Dyke RB, Korber BT, Popek E, et al. The Ariel Project: a prospective cohort study of maternal-child transmission of human immunodeficiency virus type 1 in the era of maternal antiretroviral therapy. J Infect Dis 1999;179: Carpenter CC, Cooper DA, Fischl MA, et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society USA Panel. JAMA 2000;283: Public Health Service Task Force recommendations for the use of antiretroviral drugs in pregnant women infected with HIV-1 for maternal health and for reducing perinatal HIV-1 transmission in the United States. Centers for Disease Control and Prevention. MMWR Morb Mortal Wkly Rep 1998;47(RR-2): Panel on Clinical Practices for Treatment of HIV Infection convened by the US Department of Health and Human Services and the Henry J. Kaiser Family Foundation. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Available at trtgdlns.html. Accessed 5 February Fiscus SA, Adimora AA, Schoenbach VJ, et al. Perinatal HIV infection and the effect of zidovudine therapy on transmission in rural and urban counties. JAMA 1996;275: Cooper ER, Nugent RP, Diaz C, et al. After AIDS Clinical Trial 076: the changing pattern of zidovudine use during pregnancy, and the subsequent reduction in the vertical transmission of human immunodeficiency virus in a cohort of infected women and their infants. Women and Infants Transmission Study Group. J Infect Dis 1996;174: Colgrove RC, Pitt J, Chung PH, Welles SL, Japour AJ. Selective vertical transmission of HIV-1 antiretroviral resistance mutations. AIDS 1998;12: Fiscus SA, Adimora AA, Schoenbach VJ, et al. Trends in human immunodeficiency virus (HIV) counseling, testing, and antiretroviral treatment of HIV-infected women and perinatal transmission in North Carolina. J Infect Dis 1999;180: