JAIDS Journal of Acquired Immune Deficiency Syndromes:
Epidemiology and Social Science
The Impact of Different Definitions on the Estimated Rate of Transmitted HIV Drug Resistance in the United Kingdom
Green, Hannah MSc*; Tilston, Peter MSc†; Fearnhill, Esther*; Pillay, Deenan FRCPath‡§; Dunn, David T PhD*; on behalf of the UK Collaborative Group on HIV Drug Resistance1
From the *HIV Group, Medical Research Council Clinical Trials Unit, London, United Kingdom; †Department of Clinical Virology, Manchester Royal Infirmary, Manchester, United Kingdom; ‡Centre of Virology, Department of Infection, Royal Free and University College Medical School, London, United Kingdom; and §Centre for Infections, Health Protection Agency, London, United Kingdom.
Received for publication February 25, 2008; accepted July 2, 2008.
The UK HIV Drug Resistance Database is partly funded by the Department of Health. Additional support is provided by Boehringer Ingelheim, Bristol-Myers Squibb, Gilead, Tibotec (a division of Janssen-Cilag Ltd), and Roche.
Presented as poster presentation at 13th Annual Conference of the British HIV Association, April 25-28, 2007, Edinburgh, United Kingdom (abstract P118) and as poster presentation at 2nd International Workshop on HIV Transmission, August 26-28, 2007, Washington, DC (abstract 33).
The views expressed in the publication are those of the authors and not necessarily those of the Department of Health.
1Details on UK Collaborative Group on HIV Drug Resistance Steering Committee are listed in Appendix.
Correspondence to: Hannah Green, Msc, Medical Research Council Clinical Trials Unit, 222 Euston Road, London NW1 2DA (e-mail: email@example.com).
Background: The use of different lists of resistance mutations has resulted in estimates of transmitted HIV drug resistance (THDR) that are often not comparable.
Methods: We estimated the rate of THDR based on the 3 definitions: ≥1 major mutation(s) listed on the International AIDS Society (IAS)-USA drug resistance mutation (DRM) 2006 list; ≥1 surveillance drug resistance mutation(s) (SDRM) on the published list by Shafer et al; and low-level/intermediate/high-level resistance to ≥1 drug(s) according to the Stanford HIVdb interpretation algorithm. Analyses were based on genotypic resistance tests conducted during 1997-2005 on antiretroviral therapy-naive patients and reported to the UK HIV Drug Resistance Database. The effect on THDR rates of revisions to the IAS-DRM list was also examined.
Results: Overall, 10.0%, 9.2%, and 10.4% of the 8272 samples available for analysis were classified as having THDR by the IAS-DRM, SDRM, and Stanford definitions, respectively; however, there was discordance for 244 (3%) samples. Changes in the version of the IAS-DRM list over time resulted in 4%-7% differences in the estimated rate of THDR, which increased from 4%-5% during 1997-2000 to 5%-7% during 2001-2005.
Conclusions: The choice of genotypic definition had a minor influence on the estimated rate of THDR. The SDRM list is recommended for epidemiological estimates of THDR as it has been designed with such studies in mind.
Transmitted HIV drug resistance (THDR) has the potential to limit a patient's future treatment options due to decreased effectiveness of first-line and subsequent antiretroviral regimens. At a population level, the surveillance of THDR provides useful information on spread of drug-resistant HIV, which may influence public health policy. Surveillance programs exist in many countries, and numerous national and regional estimates of the THDR have been predicted. However, it is argued that differences in the genotypic definition of THDR in studies estimating THDR have resulted in estimates that are often not comparable.1,2
To try and overcome this problem, a surveillance drug resistance mutation (SDRM) list has been developed specifically for epidemiological estimates of THDR.3 The list was constructed based on 4 principles that the mutations should be commonly recognized as causing or contributing to resistance; the mutations should be nonpolymorphic in untreated patients (<1%); the mutations should be applicable to all subtypes; and the list should be simple, unambiguous, and parsimonious. Publicly available data on the associations between genotype and antiretroviral treatment were used to choose the drug resistance mutations (DRMs), and the proposed list was applied to 5 published studies of primary HIV infection (see Shafer et al3 for full details).
Here, we assess the extent to which the estimated rate of THDR in the United Kingdom is altered by using the SDRM list, compared with the International AIDS Society (IAS)-USA DRM list and the Stanford HIVdb genotypic resistance interpretation algorithm, which have previously been used in studies in the United Kingdom.4-6 Finally, we assess the effect on THDR rates of revisions to the IAS-DRM list, reflecting accumulating knowledge on associations between existing drugs and specific mutations and the licensing of new drugs with distinctive resistance profiles.
Findings are based on genotypic test results reported to the UK HIV Drug Resistance Database, which aims to collect all tests conducted as part of routine clinical care nationwide.6 Virology laboratories provide data on an annual basis; this analysis includes resistance tests reported up to the end of 2005.
Analysis was restricted to the first test on patients older than 16 years who were antiretroviral therapy-naïve at the time of sampling. Patients' antiretroviral therapy status was classified from information recorded on the resistance test request form and via linkage with the UK Collaborative HIV Cohort Study.7
Estimates of THDR were compared based on the following 3 definitions.
1. IAS-USA guidelines: THDR was defined as the detection of 1 or more major DRMs in the 2006 IAS-USA guidelines8 and any mutation at T215 in reverse transcriptase (RT).
2. SDRM list: THDR was defined as the detection of 1 or more SDRMs as proposed by Shafer et al.3
3. Stanford HIVdb algorithm: The algorithm is based on a matrix of scores for each drug-mutation combination; these are summed across all mutations in the sample, and individual drug susceptibility is classified as “sensitive” (total score <9), "possible low-level" (10-14), "low-level" (15-29), "intermediate" (30-59), or "high-level" resistance (≥60). The 3 highest levels (low-level, intermediate, and high-level resistance) were collapsed to represent any reduced susceptibility (resistance), comparable to the presence of a single DRM or SDRM; THDR was defined as resistance to 1 or more drugs, according to version 4.3.0 of the Stanford HIVdb algorithm.9 Mutations included in 3 definitions are shown in Tables 1-3.
Sensitivity and specificity of the IAS-DRM list and Stanford algorithm were defined, taking the SDRM list as the gold standard, that is, sensitivity was defined as the proportion of samples correctly classified as having THDR and specificity as the proportion of samples correctly classified as “not” having THDR, compared with the SDRM list. The kappa coefficient was used as a measure of agreement between the definitions.
The IAS-DRM list is updated in the autumn of each year, whereas the Stanford algorithm is updated more regularly. The impact of changes in the IAS-DRM list on estimates of THDR was assessed (by defining THDR as the detection of 1 or more major DRMs) in each of the annual IAS-USA guidelines (plus any mutation at T215 in RT) from 2000 to 2006.10-15,8
Discordance Between THDR Definitions
Overall, 830 (10.0%), 764 (9.2%), and 864 (10.4%) of the 8272 samples available for analysis were classified as having THDR by the IAS-DRM, SDRM, and Stanford algorithm definitions, respectively. In pairwise comparisons, there was concordance in classifying samples as having THDR or not having THDR for 8106 (98.8%), 8090 (97.8%), and 8132 (98.3%) samples, comparing IAS-DRM and SDRM, IAS and Stanford algorithm, and SDRM and Stanford algorithm definitions, respectively (kappa 0.88, 0.88, and 0.90, respectively). Although the rates of THDR were similar for the 3 definitions, there was discordance for 244 (26%) of the 952 samples classified as having THDR by at least 1 of the 3 definitions (Table 4). Sixty-eight (7% of 952) samples were classified as having THDR by the IAS-DRM list but not by the SDRM list or the Stanford algorithm, largely driven by mutations A62V (n = 17) and V108I (n = 33) in RT and L33F (n = 12) in protease. These recognized DRMs were excluded from the SDRM list as they have been reported as polymorphic (prevalence of >1% in untreated persons) in 1 or more subtypes; A62V reported in 17% of persons with subtype A sequences; V108I in 0.6%-1.1% of persons with subtypes A, CRF02, D, and G; and L33F in 1.0% of subtype A and 1.2% of CRF01_AE sequences.3 Fourteen (1%) samples were classified as having THDR by the SDRM list only; all due to protease mutations, F53L (n = 11) and G73C/S (n = 3). F53L is nonpolymorphic (<0.5%) across all subtypes3 but is listed as a minor DRM in the IAS list and classified as “sensitive” to all protease inhibitors (PI) by the Stanford algorithm. Seventy-two (7%) samples were classified as having THDR by the Stanford algorithm only, largely driven by resistance to 1 or more nonnucleoside reverse transcriptase inhibitors (NNRTIs) (n = 61), in particular low-level resistance caused by the V179E mutation (n = 28), which was excluded from the SDRM list as it has been reported in 4.3% of untreated persons with subtype G sequences.3
Forty-eight (5%) samples were classified as THDR by the IAS-DRM list and the Stanford algorithm but not by the SDRM list. Protease mutation M46L (n = 26) is a major PI-DRM on the IAS list and classified as being associated with low-level resistance to nelfinavir and atazanavir by the Stanford algorithm but was excluded from the SDRM list due to a reported prevalence of 1.5% in subtype G sequences3. Thirty-six (4%) samples were classified as THDR by the SDRM list and the Stanford algorithm but not by the IAS-DRM list [including RT mutations: T69D (n = 8) and K101E (n = 15)].
The Stanford algorithm had higher sensitivity (97.4%) than the IAS-DRM list (93.5%) overall; for nucleoside reverse transcriptase inhibitor (NRTI), 98.8% and 96.6% and for NNRTI THDR, (100% and 93.0%, respectively. However, the classification of PI THDR had the lowest sensitivity for both definitions (91.0% Stanford algorithm and 89.0% IAS-USA list). Specificity was high for both definitions and across all drug classes (between 99.1% and 99.8%) due to the predominance of wild-type sequences with no resistance mutations in this analysis.
The 3 lists gave similar estimates of THDR over time (Fig. 1); for example, in 2005, THDR to any drug class was 9.0%/8.5%/9.6% (IAS-DRM list/SDRM list/Stanford algorithm, respectively), NRTI THDR 5.2%/4.9%/5.0%, NNRTI THDR 4.3%/4.2%/5.2%, and PI THDR 1.8%/1.6%/1.8%, respectively. Regardless of which definition was used, the rate of THDR declined after a peak in 2001-2002 and seems to have now stabilized with no significant change between 2004 and 2005.
Changes in the IAS-DRM List
The first IAS-DRM list was published in 2000 and did not include several recognized DRMs, which have appeared in all lists between 2001 and 2006 (Tables 1-3). Since 2000, other changes in the RT-DRM lists have included the removal of E44D and V118I in 2005 and M230L in 2006 due to lack of evidence that these mutations cause clinically significant resistance on their own. T69D was removed in 2005 when zalcitabine became commercially unavailable, and K70E and V106M were added in 2006 and 2003, respectively, due to new studies showing their association with tenofovir and nevirapine/efavirenz resistance, respectively. The greatest number of changes to the IAS-DRM list over time has been in protease due to the licensing of new drugs such as atazanavir, tipranavir, and more recently darunavir and accumulating knowledge about lopinavir resistance (Tables 1-3).
The impact of these changes in the IAS-DRM list on the estimated prevalence of THDR is illustrated in Figure 2A. Although trends over time were similar, the estimated prevalence of THDR varied markedly according to which annual IAS-DRM list was used, with differences in THDR increasing from 4%-5% during 1997-2000 to 5%-7% during 2001-2005. Estimates of THDR to any class were highest using the 2003 list, 3%-6% higher than the 2006 list. This was largely driven by the inclusion of the L33V protease mutation in 2003, the only difference between the 2003 and 2004 lists (Fig. 2B). The prevalence of this mutation varied over time between 1% and 3% (Fig. 3). Similarly, the higher prevalence of PI THDR in 2004 compared with 2006 was due to the inclusion of L33I, which increased in prevalence from 0.3% in 1997 to more than 1% from 2001 onwards (I84A/C, the only other mutation that was different between the 2004 and 2006 lists and was not found in any patients in this analysis). Estimates of PI THDR using lists developed in recent years were higher than pre-2003 due to the inclusion of mutations associated with reduced susceptibility to tipranavir including L33F and V82L (Fig. 3). In RT, the exclusion of V118I, and to a lesser extent E44D and T69D, in the 2005 and 2006 lists lowered the prevalence of NRTI THDR to between 5% and 8%, compared with between 8% and 13% using 2002-2004 versions (Fig. 2C and Fig. 3). There has been very little change in the mutations thought to be associated with resistance to NNRTIs, and therefore, little difference in the prevalence of NNRTI THDR was seen between the lists (data not shown).
The presence of resistance-associated mutations in virus from drug-naive individuals implies transmission of resistance from treated patients and is therefore a marker of sexual risk within a population. Surveillance of THDR provides important information on the epidemiology of HIV transmission and the impact of sexual health initiatives in HIV-infected individuals. The cost-effectiveness of screening treatment-naive patients is also highly sensitive to baseline prevalence of resistance.16 A list of DRMs designed to track the transmission of resistant virus for epidemiological purposes is necessary to support these public health policies, especially as the wider use of antiretroviral drugs in resource-limited countries may lead to a rapid spread of resistance in those countries and in industrialized countries where an increasing proportion of new diagnoses are from migrants.
Most studies of THDR have typically estimated rates between 4% and 25%17; latest figures from the United Kingdom indicate a rate of 9% (2005), having dropped from a peak of 13.6% in 2002.4 Stable or decreasing rates over time between 8% and 10% have also been reported from other European countries,18,19 whereas increases have been reported from parts of the United States.20 Estimates of THDR are dependent on a number of interrelated variables including the study population, the study period and geographical location, the time between infection and testing, and the accuracy of antiretroviral treatment data. Variation in these parameters across studies has made direct comparisons difficult. However, as resistance is most frequently measured using genotypic methods, 1 parameter that could potentially be standardized is the genotypic definition of THDR.
The majority of reports to date have defined resistance either in relation to the annually updated IAS-DRM list8 or using the Stanford genotypic resistance interpretation algorithm.9 The use of the IAS-DRM list for epidemiological purposes has a number of limitations; first, it is primarily designed as a tool for clinicians in the identification of key DRM and may therefore contain genetic determinants associated with suboptimal response to specific therapies but that are not epidemiologically relevant to the study of THDR. Second, it also excludes the “sentinel” (or revertant) mutations at RT 215, which, although not conferring resistance directly, may act as markers for the earlier presence of major DRMs at this position. Finally, the list is not always applied consistently across studies with additional substitutions at RT codons 69, 190, and 215, which are not included in the IAS-DRM list, also being considered as THDR markers. Conversely, substitutions at RT 44 and 118 have frequently been omitted, even before their removal from the IAS-DRM list in 2005.15,17,21
We compared the estimated rate of THDR in the United Kingdom using the SDRM list to define THDR with the conventional definitions using the IAS-DRM list or the Stanford interpretation algorithm. The 3 definitions did not lead to differences in overall THDR rates of more than 2%, with estimates of THDR slightly lower using the SDRM list, in agreement with other studies, mainly due to the exclusion of mutations reported as polymorphic in more than 1 subtype. Hurt et al22 identified overall THDR rates in 126 acutely infected patients of 16.7%/15.9%/12.7% by IAS-DRM 2006 list/SDRM list/Stanford algorithm, respectively, using a conservative Stanford definition of intermediate and high-level resistance only. NNRTI mutations were the most prevalent at 7.9% by SDRM and 9.5% by IAS-DRM 2006. Similarly, Liu et al23 in a larger group of primary infections identified overall rates of 15%/16%/14% by IAS-DRM 2005/SDRM/Stanford (intermediate and high-level resistance), respectively. Overall rates of THDR in our study were lower at 10.0%/9.2%/10.4% (IAS-DRM 2006/SDRM/Stanford), probably due to the inclusion of chronic and acute infections.
Overall THDR rates using IAS-DRM list and the Stanford algorithm were similar at 10.0% and 10.4%, respectively. Previous comparisons of IAS and Stanford have generally shown a high level of concordance, although this has been dependent on the choice of threshold for the Stanford algorithm. If a more conservative definition of intermediate/high-level resistance was adopted, THDR rates were reported as 5%-6% lower than those derived from the IAS list. The most commonly identified mutations responsible for this difference were substitutions at RT 118 and 215 and 33 in protease. However, some of these reports used the 2004 IAS-DRM list, which included V118I and L33I, which have been excluded from subsequent lists. If, as in this study, lower levels of (Stanford defined) resistance are also included, overall rates differ by <1%.18,23,24,25
The effect of annual updating can lead to marked alterations in apparent THDR rates; 4%-7% differences in the estimated rate of THDR were seen in this study depending on which version of the IAS-DRM list was used. Although this should not affect the pattern of temporal trends in which the entire dataset is consistently analyzed, it may compromise the validity of comparisons between cross-sectional studies from different years. Changes to the composition of the IAS list over time showed that the highest levels of resistance were predicted by the 2003 and 2004 lists, most prominently for the PIs. The major contribution to this effect was the inclusion of L33I/V (2003) or L33I alone (2004) in protease (Fig. 3), both of which were subsequently removed as major mutations due to their lack of association with PI therapy.26 NRTI resistance was also higher in pre-2005 lists, largely driven by the inclusion of V118I. This mutation does not seem to compromise NRTI-based treatment,27 is polymorphic in a number of subtypes,3 and has therefore been removed from more recent versions. Similarly, the definition of resistance using the Stanford algorithm, which is also regularly updated, has varied between studies6,18,23 influencing estimates of THDR rates and subtly affecting prescribing practices in those centers that rely on this interpretation. As new drugs and new drug classes are introduced, updates to the SDRM list will also be required. In addition, as more sequence data from non-B subtypes become available, it is likely that some mutations will be reclassified with regard to their polymorphic status, whereas other subtype-specific DRMs may be identified. A subtype-specific SDRM list would more accurately reflect the transmission of resistance, however may be impractical and difficult to implement as the number of intersubtype recombinants increases.3
Care should be taken when estimating THDR in relation to DRMs associated with recently released antiretrovirals, as very few patients will have been exposed to the new drugs, and therefore, it is very unlikely that viruses with resistance associated to the new drugs would have yet been transmitted to others. However, for this concern to have an impact on the estimates of THDR, the mutation would have to occur as a polymorphism in the absence of therapy. Such mutations, by definition, would be excluded from the SDRM list.
In summary, we have measured THDR rates in a large treatment-naive population using 3 definitions of transmitted resistance. The choice of definition had a minor influence on the estimated rate of THDR over an 8-year period. However, THDR rates were sensitive to the version of the IAS-DRM list used, and direct comparisons of studies using different versions of DRM lists or interpretation algorithms should be undertaken with caution. By standardizing the genotypic definition of THDR, differences seen in estimates of THDR across studies will not be due to the mutation list or algorithm used. The SDRM list is recommended for epidemiological estimates of THDR as it has been designed to include transmitted resistance-associated mutations regardless of their effect on drug susceptibility.
1. Pillay D. Current patterns in the epidemiology of primary HIV drug resistance in North America and Europe. Antivir Ther
2. Geretti AM. Epidemiology of antiretroviral drug resistance in drug-naive persons. Curr Opin Infect Dis
3. Shafer RW, Rhee S, Pillay D, et al. HIV-1 protease and reverse transcriptase mutations for drug resistance surveillance. AIDS
4. HIV Drug Resistance in the United Kingdom: Data to the End of 2005. Health Protection Report
. Health Protection Agency: London, UK. Vol 1. No. 31; August 3, 2007.
5. UK Collaborative Group on HIV Drug Resistance, UK Collaborative HIV Cohort Study, and UK Register of Seroconverters. Evidence of a decline in transmitted HIV-1 drug resistance in the UK. AIDS
6. UK Group on HIV Transmitted Drug Resistance. Time trends in primary resistance to HIV drugs in the United Kingdom: multicentre observational study. BMJ
7. Sabin CA, Hill T, Lampe F, et al. Treatment exhaustion of highly active antiretroviral therapy (HAART) among individuals infected with HIV in the United Kingdom: multicentre cohort study. BMJ
8. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1: Fall 2006. Top HIV Med
10. 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
11. D'Aquila RT, Schapiro JM, Brun-Vezinet F, et al. Update of drug resistance mutations in HIV-1. Top HIV Med
12. D'Aquila RT, Schapiro JM, Brun-Vezinet F, et al. Drug resistance mutations in HIV-1. Top HIV Med
13. Johnson VA, Brun-Vezinet F, Clotet B, et al. Drug resistance mutations in HIV-1. Top HIV Med
14. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1:2004. Top HIV Med
15. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1:Fall 2005. Top HIV Med
16. Sax PE, Islam R, Walensky RP, et al. Should resistance testing be performed for treatment-naive HIV-infected patients? A cost-effectiveness analysis. Clin Infect Dis
17. Booth CL, Geretti AM. Prevalence and determinants of transmitted antiretroviral drug resistance in HIV-1 infection. J Antimicrob Chemother
18. Masquelier B, Bhaskaran K, Pillay D, et al. Prevalence of transmitted HIV-1 drug resistance and the role of resistance algorithms. J Acquir Immune Defic Syndr
19. Wensing AMJ, van de Vijver DA, Andarano G, et al. Prevalence of drug-resistant HIV-1 variants in untreated individuals in Europe: implications for clinical management. J Infect Dis
20. Shet A, Berry L, Mohri H, et al. Tracking the prevalence of transmitted antiretroviral drug-resistant HIV-1. A decade of experience. J Acquir Immune Defic Syndr
21. Weinstock HS, Zaidi I, Heneine W, et al. The epidemiology of antiretroviral drug resistance among drug-naive HIV-1-infected persons in 10 US cities. J Infect Dis
22. Hurt C, McCoy S, Kerkau M, et al. Frequency of transmitted drug resistance among acute and recently HIV-infected patients in North Carolina is similar to reports from more urbanized areas. Antivir Ther
23. Liu L, May S, Richman D, et al. Comparison of resistance algorithms to determine prevalence of transmitted drug resistance. Antivir Ther
24. Ross LL, Williams V, Wine B, et al. Changes in the regional prevalence of HIV-1 drug resistance-associated mutations and in the prevalence on non-clade B subtypes in antiretroviral therapy-naive HIV-infected patients in the United States from 2000-2006. Antivir Ther
25. Dunn DT, Green H, Matthias R, et al. Prevalence of reduced drug susceptibility in treatment-naive patients in the UK. Antivir Ther
26. Kozal MJ, Huppler Hullsiek K, Leduc R, et al. Prevalence and impact of HIV-1 protease codon 33 mutations and polymorphisms in treatment-naive and treatment-experienced patients. Antivir Ther
27. Milhailidis C, Dunn D, Pillay D, et al. Effect of isolated V118I mutation in reverse transcriptase on response to first-line antiretroviral therapy. AIDS
UK Collaborative Group on HIV Drug Resistance Steering Committee
Jane Anderson, Homerton University Hospital, London; David Asboe and Anton Pozniak, Chelsea and Westminster Hospital, London; Sheila Burns, Royal Infirmary of Edinburgh; Sheila Cameron, Gartnavel General Hospital, Glasgow; Patricia Cane, Health Protection Agency, Porton Down; Ian Chrystie, Guy's and St Thomas' NHS Foundation Trust, London; Duncan Churchill, Brighton and Sussex University Hospitals NHS Trust; Duncan Clark, St Bartholomew's and The London NHS Trust; Valerie Delpech and Deenan Pillay, Health Protection Agency, Centre for Infections, London; Linda Lazarus, Expert Advisory Group on AIDS Secretariat, Health Protection Agency, London; David Dunn, Esther Fearnhill, Hannah Green, and Kholoud Porter, Medical Research Council Clinical Trials Unit, London; Philippa Easterbrook and Mark Zuckerman, King's College Hospital, London; Anna Maria Geretti, Royal Free NHS Trust, London; Paul Kellam, Deenan Pillay, Andrew Phillips, and Caroline Sabin, Royal Free and University College Medical School, London; David Goldberg, Health Protection Scotland, Glasgow; Mark Gompels, Southmead Hospital, Bristol; Antony Hale, Leeds Teaching Hospitals NHS Trust; Steve Kaye, St Mary's Hospital, London; Svilen Konov, Community Advisory Board; Linda Lazarus, Department of Health; Andrew Leigh-Brown, University of Edinburgh; Nicola Mackie, St Mary's Hospital, London; Chloe Orkin, St Bartholomew's Hospital, London; Erasmus Smit, Health Protection Agency, Birmingham Heartlands Hospital; Peter Tilston, Manchester Royal Infirmary; Ian Williams, Mortimer Market Centre, London; and Hongyi Zhang, Addenbrooke's Hospital, Cambridge.
Addenbooke's Hospital, Cambridge (Hongyi Zhang); Department of Virology, St Bartholomew's and The London NHS Trust (Duncan Clark, Ines Ushiro-Lumb, Tony Oliver, and David Bibby); Belfast Health and Social Care Trust (Suzanne Mitchell); HPA Birmingham Public Health Laboratory (Erasmus Smit); Chelsea and Westminster Hospital, London (Adrian Wildfire); Dulwich Hospital, London (Melvyn Smith); Royal Infirmary of Edinburgh (Jill Shepherd); West of Scotland Specialist Virology Laboratory, Gartnavel, Glasgow (Alasdair MacLean); Guy's and St Thomas' NHS Foundation Trust, London (Ian Chrystie); Leeds Teaching Hospitals NHS Trust (Diane Bennett); Specialist Virology Centre, Liverpool (Mark Hopkins) and Manchester (Peter Tilston); Department of Virology at Royal Free Hospital, London (Clare Booth and Ana Garcia-Diaz); St Mary's Hospital, London (Steve Kaye); and University College London Hospitals (Stuart Kirk).
Medical Research Council Clinical Trials Unit, London (David Dunn, Esther Fearnhill, Hannah Green, Kholoud Porter, and Kate Coughlin). Cited Here...
This article has been cited 1 time(s).
transmitted resistance; HIV; mutations; genotypic; definitions
© 2008 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.