Secondary Logo

Journal Logo


Identification of accessory mutations associated with high-level resistance in HIV-1 reverse transcriptase

Cane, Patricia Aa; Green, Hannahb; Fearnhill, Estherb; Dunn, Davidbon behalf of the UK collaborative group on HIV Drug Resistance

Author Information
doi: 10.1097/QAD.0b013e3280129964



The introduction of HAART has greatly improved the prognosis of HIV-infected patients. However, resistance often develops to the drugs used as a result of the acquisition of mutations within the genes encoding the target proteins. The determination of the mutations developed by HIV-1 that are associated with drug resistance has been by analysis of virus passed in vitro in the presence of drug and also by phenotypic and genotypic analysis of large numbers of clinical samples from patients failing therapy. The significance of mutations in conferring reduced drug susceptibility has been confirmed by site-directed mutagenesis in molecular clones. The lists of mutations considered to be of clinical importance are regularly summarized by the International Aids Society – USA panel (IAS) [1].

It has been reported that additional mutations other than those designated by the IAS are associated with treatment with antiretroviral drugs. For example, Gonzales et al. [2] analysed a total of 1210 reverse transcriptase (RT) sequences from patients known to be on antiretroviral therapy. They showed that the number of RT changes relative to consensus increased from a median of four in untreated individuals to 14 in heavily treated patients. Whereas most of the increase in mutations could be ascribed to recognized resistance changes, they also reported nine previously undescribed mutations (at codons 20, 39, 43, 203, 208, 218, 221, 223 and 228) that were associated with nucleoside analogue reverse transcriptase inhibitor (NRTI) therapy, and suggested that these were accessory mutations as they occurred almost exclusively with other ‘standard’ resistance mutations. Rhee et al. [3] also described the development of specific drug resistance-associated mutations and their correlation with NRTI alone or NRTI and non-nucleoside reverse transcriptase inhibitor (NNRTI) treatment. Both these reports analysed RT codons between 1 and 240 from subtype B HIV-infected patients. In addition, Berkhout et al. [4] have identified alternative amino acid substitutions at known resistance-associated positions.

This study examines the association of novel mutations in codons up to position 400 of RT with the accumulation of standard primary thymidine analogue mutations (TAMs) or NNRTI resistance-associated mutations within a large UK database [5]. This approach was chosen to supplement and extend the reports of Gonzales et al. [2], Rhee et al. [3] and Svicher et al. [6], who demonstrated novel mutations associated with treatment, in order to focus upon those mutations associated with high-level resistance.


The UK HIV Drug Resistance Database pools the results of resistance tests performed as part of routine clinical care in the United Kingdom. All resistance tests in this analysis were based on nucleotide sequencing of the pol gene. Data were received from seven laboratories, which used a number of different in-house and commercial systems [5]. As a result of the variety of sequencing methods used, not all samples provided data on all codons.

Subtype B HIV-infected patients more than 16 years of age, who had at least one genotypic resistance test after starting antiretroviral therapy were identified from the database. For each test, the number of TAMs and NNRTI resistance mutations were determined (TAMs: M41L, D67any, K70R, L210W, T215any, K219any, 65–69 insertions and deletions; NNRTI: L100I, K103any (not R), V106M/A, V108I, Y181any, Y188any, G190any, P225H, M230L, P236L). The test with the greatest number of TAMs or NNRTI resistance mutations was selected for patients tested more than once (36%), creating two datasets for analysis containing one test per patient.

For the purpose of this paper, the most commonly occurring amino acid at a codon will be referred to as wild type, and all other amino acids as mutations. One-way analysis of variance was used to test the association between the amino acid distribution and the accumulation of TAMs/NNRTI resistance mutations at each codon, with P values less than 0.01 considered statistically significant [7]. Accessory mutations were identified at codons where there was (i) a statistically significant association, and (ii) the difference in the proportion of the wild-type amino acid between samples showing no TAMs/NNRTI resistance mutations and samples showing four or more TAMs or three or more NNRTI resistance mutations was greater than 5%. Positions already established as drug resistance positions by IAS [1] were not included in this analysis.

The strength of the association between accessory mutations and specific known primary resistance mutations was expressed through risk ratios: the probability of an accessory mutation in samples in which the primary resistance mutation was present, compared with samples in which it was absent. Interaction tests, using log-binomial regression, were conducted to see if the risk ratios differed across the specific set of TAMs/NNRTI resistance mutations, with P values less than 0.01 considered statistically significant.


Treatment data were available from 2460 patients (77%). Of these, 56% had been exposed to both stavudine and zidovudine, 16% to stavudine only, and 27% to zidovudine only. With respect to NNTRI exposure, 12% of patients had received both nevirapine and efavirenz, 29% nevirapine only, and 19% efavirenz only.

Accessory mutations associated with thymidine analogue mutations

Resistance tests from 3162 patients were available for the analyses. Of these samples, 47, 9, 9, 13 and 22% had none, one, two, three or four or more TAMs, respectively. A total of 24 codons were identified in which accessory mutations were associated with the accumulation of TAMs using the criteria described in the methods (Table 1).

Table 1
Table 1:
Amino acid distribution at 24 codons where accessory mutations were associated with the accumulation of thymidine analogue mutations.

The top 10 positions that showed the greatest difference in the proportion of the wild-type amino acid between zero and four or more TAMs were codons 43 (24.1% decrease in the proportion of the wild-type amino acid between none and four or more TAMs), 359 (21.0%), 228 (19.4%), 371 (18.1%), 208 (17.9%), 39 (15.3%), 203 (15.2%), 356 (14.1%), 122 (13.8%), and 218 (13.8%). Of these 43, 208, 218, and 228, were non-polymorphic in naive patients (i.e > 99% homogeneity), whereas 39, 122, 203, 356, 359, and 371 were polymorphic. Interestingly, at codons 83 and 381, the proportion of samples with the wild-type amino acid increased with increasing numbers of TAMs (80.8% with no TAMs to 91.6% with four or more TAMs, and 93.6–99.3%, respectively).

Overall, 7% of the 3162 samples had no accessory mutations, 42% had one to two, 33% three to four, and 18% had more than five. The number of accessory mutations increased with the number of TAMs (Fig. 1a). For example, of samples with no TAMs, 10, 53, 31 and 6% had zero, one to two, three to four and five or more accessory mutations, respectively, whereas in samples with four or more TAMs the respective percentages were 2, 19, 36 and 43%. In addition, changes in the proportion of the wild-type amino acid at the top 10 accessory TAMs positions with increasing numbers of TAMs are illustrated in Fig. 2a. As expected, at non-polymorphic codons (43, 208, 218, 228) the proportion of samples with the wild-type amino acid was over 97% for samples with no TAMs, decreasing with the increasing number of TAMs. However, the pattern was not the same for all positions; for codons 208 and 218 the proportion of the wild-type amino acid remained high for samples with one or two TAMs, and only decreased in samples with three or more TAMs, whereas there was a steady decrease with increasing numbers of TAMs for codons 43 and 228. At polymorphic codons (39, 122, 203, 356, 359, 371), the proportion tended to be lower for samples with no TAMs (67–96%) but generally still decreased with the increasing number of TAMs.

Fig. 1
Fig. 1:
Number of accessory mutations per sample in relation to the number of (a) thymidine analogue mutations and (b) non-nucleoside reverse transcriptase inhibitor mutations. NNRTI, Non-nucleoside reverse transcriptase inhibitor; TAMs, thymidine analogue mutation.JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU1/v/2017-07-25T100134Z/r/image-tiffFive or more;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU2/v/2017-07-25T100134Z/r/image-jpegthree to four;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU3/v/2017-07-25T100134Z/r/image-tiffone to two;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU4/v/2017-07-25T100134Z/r/image-tiffzero.
Fig. 2
Fig. 2:
Frequency of the wild-type amino acid for the top 10 accessory mutations with (a) thymidine analogue mutations and (b) non-nucleoside reverse transcriptase inhibitor mutations. NNRTI, Non-nucleoside reverse transcriptase inhibitor; TAMs, thymidine analogue mutation. (a)JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU5/v/2017-07-25T100134Z/r/image-tiff39;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU6/v/2017-07-25T100134Z/r/image-tiff43;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU7/v/2017-07-25T100134Z/r/image-tiff122;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU8/v/2017-07-25T100134Z/r/image-tiff203;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU9/v/2017-07-25T100134Z/r/image-tiff208;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU10/v/2017-07-25T100134Z/r/image-tiff218;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU11/v/2017-07-25T100134Z/r/image-tiff228;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU12/v/2017-07-25T100134Z/r/image-tiff356;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU13/v/2017-07-25T100134Z/r/image-tiff359;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU14/v/2017-07-25T100134Z/r/image-tiff371. (b)JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU15/v/2017-07-25T100134Z/r/image-tiff20;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU16/v/2017-07-25T100134Z/r/image-tiff43;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU17/v/2017-07-25T100134Z/r/image-tiff101;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU18/v/2017-07-25T100134Z/r/image-tiff122;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU19/v/2017-07-25T100134Z/r/image-tiff179;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU20/v/2017-07-25T100134Z/r/image-tiff203;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU21/v/2017-07-25T100134Z/r/image-tiff221;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU22/v/2017-07-25T100134Z/r/image-tiff228;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU23/v/2017-07-25T100134Z/r/image-tiff359;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU24/v/2017-07-25T100134Z/r/image-tiff371.

Accessory mutations associated with non-nucleoside reverse transcriptase inhibitor resistance

Of the 3162 samples available, 52, 24, 17 and 7% had zero, one, two, and three or more NNRTI mutations, respectively. A total of 25 codons were identified in which accessory mutations were associated with the accumulation of NNRTI primary mutations (Table 2). The top 10 positions showing the greatest difference in the proportion of the wild-type amino acid between zero and three or more NNRTI were at codons 101 (32.9% decrease in the proportion of the wild-type amino acid between zero and three or more NNRTI mutations), 221 (24.4%), 122 (16.0%), 359 (14.4%), 228 (13.4%), 179 (12.8%), 371 (12.5%), 203 (11.9%), 43 (10.6%), and 20 (9.4%). Of these, positions 43, 101, 221 and 228 were non-polymorphic in naive patients, whereas positions 20, 122, 179, 203, 359, and 371 were polymorphic. As seen in the TAMs analysis, the number of accessory NNRTI mutations increased with the number of NNRTI primary mutations (Fig. 1b and Fig. 2b).

Table 2
Table 2:
Amino acid distribution at 25 codons where accessory mutations were associated with the accumulation of non-nucleoside reverse transcriptase inhibitor mutations.

Several positions, including 43, 122, 203, 228, 359 and 371, were identified in which accessory mutations were associated with both the accumulation of TAMs and NNRTI mutations.

Association of accessory mutations with specific primary resistance mutations

The data presented above show that the development of accessory mutations is associated with the number of TAMs or NNRTI primary resistance mutations. The data were therefore further analysed to determine whether particular primary resistance mutations were associated with certain accessory mutations. Accessory mutations in which the risk ratio was significantly different across the TAMs are shown in Fig. 3a. For example, an accessory mutation at codon 43 was 15.9 times more likely to occur with 41L than the wild type at position 41, but only 3.6 times more likely to occur with 67N than with the wild type at position 67.

Fig. 3
Fig. 3:
Relative risk of association of accessory mutations with particular (a) thymidine analogue mutations and (b) non-nucleoside reverse transcriptase inhibitor mutations. (a)JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU25/v/2017-07-25T100134Z/r/image-tiff41L;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU26/v/2017-07-25T100134Z/r/image-tiff67N;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU27/v/2017-07-25T100134Z/r/image-jpeg70R;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU28/v/2017-07-25T100134Z/r/image-tiff210W;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU29/v/2017-07-25T100134Z/r/image-tiff215F;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU30/v/2017-07-25T100134Z/r/image-tiff215Y;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU31/v/2017-07-25T100134Z/r/image-tiff219E;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU32/v/2017-07-25T100134Z/r/image-tiff219Q. (b)JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU33/v/2017-07-25T100134Z/r/image-tiff100I;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU34/v/2017-07-25T100134Z/r/image-tiff103N;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU35/v/2017-07-25T100134Z/r/image-jpeg106M;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU36/v/2017-07-25T100134Z/r/image-tiff106A;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU37/v/2017-07-25T100134Z/r/image-tiff108I;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU38/v/2017-07-25T100134Z/r/image-tiff181I;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU39/v/2017-07-25T100134Z/r/image-tiff181C;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU40/v/2017-07-25T100134Z/r/image-tiff188L;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU41/v/2017-07-25T100134Z/r/image-tiff190S;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU42/v/2017-07-25T100134Z/r/image-tiff190A;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU43/v/2017-07-25T100134Z/r/image-tiff225H;JOURNAL/aids/04.02/00002030-200702190-00007/math_7MMU44/v/2017-07-25T100134Z/r/image-tiff230L.

The most striking associations were changes at codons 43, 208, and 223 (and to a lesser extent at codons 39 and 98) with the previously described TAM1 cluster of 41L, 210W and 215Y, and changes at codon 218 with the TAM2 cluster of 67N, 70R, and 219E/Q (and to a lesser extent at codon 60; Fig. 3a). Changes at positions 122, 196 and 228 were associated with TAMs from both pathways.

Differences relative to the consensus at codons 83 and 381 (which showed a negative trend with the increasing numbers of TAMs) were negatively associated with all TAMs (data not shown), but in particular 219Q [risk ratio (RR) 0.13, 95% confidence interval (CI) 0.06–0.29] and 215Y (RR 0.12, 95% CI 0.03–0.5), respectively.

Figure 3b shows the associations of the accessory mutations with particular NNRTI mutations. Codon 43 mutations were most strongly associated with 181I, as were changes at codon 179 (as well as with 188L and 106M). Substitutions at codon 221 were strongly linked with mutations at codon 181. Changes at positions 218 and 228 were associated with 100I as well as 181I and 181C, respectively. Changes at position 101 were most strongly associated with 190S/A, 359 with 106M and 203 with all NNRTI primary resistance mutations.


This report confirms and extends the descriptions of additional RTI resistance-associated mutations by Gonzales et al. [2], Rhee et al. [3], and Svicher et al. [6]. The method of analysis described here was based on analysing sequences with respect to major recognized resistance mutations rather than on therapy histories, other than the requirement that the patients had received treatment. However, it is reasonable to assume that patients with high numbers of primary resistance mutations had either received multiple therapies or alternatively had continued with failing therapy despite a rising viral load. This approach was taken in order to facilitate the sensitive detection of accessory mutations associated with actual high-level genotypic resistance, rather than just with treatment. We also quantified the appearance of accessory mutations with increasing number of TAMs and NNRTI mutations together with specific associations between mutations.

The data confirm the accumulation of the increasing numbers and complexity of accessory mutations as TAMs and NNRTI mutations develop. Previous studies have reported data on codons 1–240; this paper extends the dataset to include the region encompassing codons 240–400, and further accessory mutations in this additional region that includes part of the connection domain of the molecule have now been identified. Many codon positions, but not all, are associated with both TAMs and NNRTI resistance mutations, as would be expected as patients are very likely to have been treated with both NRTI and NNRTI.

The secondary mutations identified in this study in association with TAMs include those previously described as occurring with NRTI treatment (20, 39, 43, 203, 208, 218 and 228 [2,3]) together with many previously unrecognized mutations, namely at positions 35, 60, 83, 101, 122, 284, 322, 356, 359, 360, 371 and 381 (Table 1). The reason for the apparent increased sensitivity of the analysis method used here relates to the determination of resistance being based on the accumulation of primary mutations rather than treatment histories. This results in much higher prevalence levels of the secondary mutations. For example, changes at position 43 were seen in 7.5% of NRTI-treated patients as reported by Rhee et al. [3], whereas in this study such changes were observed in approximately 1% of sequences with no TAMs rising to 24% of sequences in samples with four or more TAMs. This method of analysis thus accentuates those mutations that accumulate with increasing resistance, so allowing the identification of additional resistance-associated mutations.

The significance of the changes at polymorphic positions 83 and 381 to become closer to the wild type with increasing resistance caused by TAMs is unclear. It may be that the alternative amino acids at these positions are not structurally compatible with the primary resistance mutations. It was observed that K to R (wild-type) substitutions at codon 83 were particularly frequent with mutations at codon 219, so it would be interesting to explore whether such natural polymorphisms influence the pathways taken to resistance.

Various reports have indicated further positions associated with resistance. For example, Sturmer et al. [8] found that H208Y together with R211K and L214F were associated with high-level zidovudine resistance. Whereas this analysis confirmed the importance of H208Y in vivo, there was no indication that changes at positions 211 and 214 were more prevalent with increasing levels of resistance.

Of the positions identified in this study as being associated with high-level NNRTI resistance, all except codon 43 are recorded on the Stanford web site as being associated with NNRTI usage [9]. Codons 101 and 179 (together with codon 98, which showed only a 7.6% change in prevalence between zero and three or more NNRTI mutations) have previously been described as associated with NNRTI resistance [10,11]. Position 221 has been mentioned as being associated with NNRTI treatment [3], but the high prevalence of this mutation with primary NNRTI mutations suggests that it may have an important role to play in either the development of NNRTI resistance or compensatory activity.

It has been shown that high-level resistance to efavirenz is associated with the acquisition of multiple RT mutations, and that the resistance conferred by the most frequently occurring mutation, namely K103N, can be greatly enhanced by additional mutations [10,11]. Mutations associated with increasing numbers of NNRTI resistance mutations were also identified in this study. Position 101 is not a designated NNRTI resistance mutation in the current IAS list [1], but mutations at this position did occur at a low frequency in the absence of other NNRTI mutations, so it seems likely that it is a primary resistance-conferring mutation and this would account for its high prevalence as the other NNRTI mutations accumulate. It has previously been reported that K101E was observed in 13.8% of efavirenz-treated patients either as a single NNRTI mutation (4.8%) or linked with other NNTRI mutations [12]. In addition, in-vitro resistance to efavirenz conferred by K103N can be greatly increased by mutations at codon 101 [10]. A recent report also showed that K101P/H/Q mutations were associated with reduced phenotypic susceptibility to NNRTI [13], with K101P conferring the greatest reduction in susceptibility. However, that paper also mentioned 14 other non-IAS positions associated with reduced susceptibility to NNRTI, but of these only three were concordant with those identified in Table 2, namely positions 179, 218 and 221. In this report, K101P does not feature as a common additional mutation, and the predominant mutations identified here at position 179 was V179I, not V179D as observed by Parkin et al. [13].

The most striking non-polymorphic secondary mutation associated with accumulating NNRTI mutations other than position 101 in this analysis was at position 221. Changes in amino acids relative to the consensus sequence were observed at this position in 24% of samples with three or more primary NNRTI mutations, with half of the sequences showing H221Y and the other half showing a mixture of Y and H at this position. The presence of a high level of mixtures at this position could be interpreted as continuing evolution of a resistant virus population either to show higher levels of resistance or greater fitness. In addition, V179D has been reported as an NNRTI mutation. This mutation did not figure significantly in this dataset. However, V179I did appear to be significantly associated with the development of high-level NNRTI resistance, although this particular change is more commonly seen as a polymorphism in non-B subtypes [14].

Clearly, there was likely to be some overlap between these series of analyses as a result of treatment often containing both NRTI and NNRTI drugs. Nevertheless, some changes are restricted to one series only, indicating that they are specifically associated with either NRTI or NNRTI treatment such as substitutions at codon 221 with NNRTI mutations. There are known interactions between NNRTI mutations and TAMs, for example, the NNRTI mutations at positions 100 and 181 are known to reduce T215Y-mediated zidovudine resistance, whereas multiple TAMs can reduce the level of efavirenz resistance associated with K103N [9].

The new mutations described here could either increase resistance to anti-HIV drugs or act to restore replication fitness. Those increasing resistance might well be further detrimental to fitness, so the effect of the acquisition of a particular mutation could be neutral, reduce or increase fitness. In addition, when the new mutation sites are polymorphic, it cannot be excluded that certain genetic backgrounds predispose towards the selection of particular pathways for the development of resistance rather than contributing to resistance or fitness per se. This will only be resolved when extensive longitudinal datasets become available.

It is unlikely that the accessory mutations described here have a significant influence on the degree of susceptibility to particular drugs compared with the primary resistance mutations, so their relevance to algorithms for predicting the drug resistance phenotype will probably be limited. However, analysis of secondary mutations could potentially play a future role in the prediction of the replicative capacity of the virus. In order for this to become a reality, the influence of each of the newly identified resistance-associated mutations on viral fitness in the context of highly resistant virus will need to be explored.


The authors would like to thank all the clinicians, virologists, data managers, and research nurses in participating centres who have assisted with the provision of data. A full list is available at They also thank Caroline Sabin and Deenan Pillay for critical reading of the manuscript.

Sponsorship: The UK HIV Drug Resistance Database receives funding from the Department of Health.

For members of the UK Collaborative Group on HIV Drug Resistance see Appendix.


1. Johnson VA, Brun-Vezinet F, Clotet B, Conway B, Kuritzkes DR, Pillay D, et al. Update of the drug resistance mutations in HIV-1: fall 2005. Top HIV Med 2005; 13:125–131.
2. Gonzales MJ, Wu TD, Taylor J, Belitskaya I, Kantor R, Israelski D, et al. Extended spectrum of HIV-1 reverse transcriptase mutations in patients receiving multiple nucleoside analog inhibitors. AIDS 2003; 17:791–799.
3. Rhee S-Y, Fessel WJ, Zolopa AR, Hurley L, Liu T, Taylor J, et al. HIV-1 protease and reverse-transcriptase mutations: correlations with antiviral therapy in subtype B isolates and implications for drug-resistance surveillance. J Infect Dis 2005; 192:456–465.
4. Berkhout B, Back NKT, de Ronde A, Jurriaans S, Bakker M, Parkin NT, et al. Identification of alternative amino acid substitutions in drug-resistant variants of the HIV-1 reverse transcriptase. AIDS 2006; 20:1515–1520.
5. UK Group on Transmitted HIV Drug Resistance. Time trends in primary resistance to HIV drugs in the United Kingdom: multicentre observational study. BMJ 2005; 331:1368–1370.
6. Svicher V, Sing T, Santoro MM, Forbici F, Rodriguez-Barrios F, Bertoli A, et al. Involvement of novel human immunodeficiency virus type 1 reverse transcriptase mutations in the regulation of resistance to nucleoside inhibitors. J Virol 2006; 80:7186–7198.
7. Rothman KJ. No adjustments are needed in multiple comparisons. Epidemiology 1990; 1:43–46.
8. Sturmer M, Staszewski S, Doerr H-W, Larder B, Bloor S, Hertogs K. Correlation of phenotypic zidovudine resistance with mutational patterns in the reverse transcriptase of human immunodeficiency virus type 1: interpretation of established mutations and characterization of new polymorphisms at codons 208, 211 and 214. Antimicrob Agents Chemother 2003; 47:54–61.
9. Rhee SY, Gonzales MJ, Kantor R, Betts BJ, Ravela J, Shafer RW. Human immunodeficiency virus reverse transcriptase and protease sequence database. Nucl Acids Res 2003; 31:298–303.
10. Young SD, Britcher SF, Tran LO, Payne LS, Lumma WC, Lyle TA, et al. L-743, 726 (DMP-266): a novel, highly potent nonnucleoside inhibitor of the human immunodeficiency virus type 1 reverse transcriptase. Antimicrob Agents Chemother 1995; 39:2602–2605.
11. Winslow DL, Garber S, Reid C, Scarnati H, Baker D, Rayner MM, et al. Selection conditions affect the evolution of specific mutations in the reverse transcriptase gene associated with resistance to DMP 266. AIDS 1996; 10:1205–1209.
12. Bacheler LT, Anton ED, Kudish P, Baker D, Bunville J, Krakowski K, et al. Human immunodeficiency virus type 1 mutations selected in patients failing efavirenz combination therapy. Antimicrob Agents Chemother 2000; 44:2475–2484.
13. Parkin NT, Gupta S, Chappey C, Petropoulos CJ. The K101P and K103R/V179D mutations in human immunodeficiency virus type 1 reverse transcriptase confer resistance to nonnucleoside reverse transcriptase inhibitors. Antimicrob Agents Chemother 2006; 50:351–354.


UK Collaborative Group on HIV Drug Resistance Steering Committee:

Sheila Burns, City Hospital, Edinburgh; Sheila Cameron, Gartnavel General Hospital, Glasgow; Patricia Cane, Health Protection Agency, Porton Down; Ian Chrystie, St Thomas' Hospital, London; Duncan Churchill, Brighton and Sussex University Hospitals NHS Trust; Valerie Delpech, Deenan Pillay, Health Protection Agency, London; David Dunn, Esther Fearnhill, Hannah Green, Kholoud Porter, MRC Clinical Trials Unit (coordinating centre), London; Philippa Easterbrook, Mark Zuckerman, King's College Hospital, London; Anna Maria Geretti, Royal Free NHS Trust, London; Rob Gifford, Paul Kellam, Deenan Pillay, Andrew Phillips, Caroline Sabin, Royal Free and University College Medical School, London; David Goldberg, Health Protection Scotland, Glasgow; Mark Gompels, Southmead Hospital, Bristol; Tony Hale, Health Protection Agency, Leeds; Steve Kaye, St Marys Hospital, London; Andrew Leigh-Brown, University of Edinburgh; Chloe Orkin, St Bartholemews Hospital, London; Anton Pozniak, Chelsea and Westminster Hospital, London; Gerry Robb, Department of Health, London; Erasmus Smit, Health Protection Agency, Birmingham Heartlands Hospital; Peter Tilston, Manchester Royal Infirmary; Ian Williams, Mortimer Market Centre, London.

14. Rhee S-Y, Kantor R, Katzenstein DA, Camacho R, Morris L, Sirivichayakul S, et al. HIV-1 pol mutation frequency by subtype and treatment experience: extension of the HIVseq program to seven non-B subtypes. AIDS 2006; 20:643–651.

Accessory mutations; HIV drug resistance; reverse transcriptase inhibitors; viral fitness

© 2007 Lippincott Williams & Wilkins, Inc.