Given that this meeting was held only three weeks after the 8^th CROI, it was not surprising that there was some overlap between the content of the two meetings. However due to the smaller size of the meeting and the European atmosphere, there was considerably more discussion.

On the topic of nucleoside-associated mutations (NAMs), data were presented that showed that a 10-fold increase in AZT resistance conferred resistance to ddC (1.5 fold), ddI and d4T (2 fold) and 3TC and ABC (2.5 fold). Moreover, the presence of AZT resistance mutations plus 44D, 69D, 75M and 118I were strongly associated with d4T resistance (Abstract 6) [1].

Resistance data were also presented on patients failing AZT/3TC/ABC in study CNA 3005 (Abstract 13) [2]. Interestingly, the vast majority (32/40) of patients' virus harbored 184V but no additional mutations. Moreover, this 184V-only genotype was retained at week 80 in 15 patients who continued the triple-NRTI regimen past week 36 despite. This occurred in the presence of incomplete virologic rebound to pre-treatment levels.

Data on abacavir resistance were also presented by two groups. Schmidt et al. found that AZT mutations by themselves could confer up to 4-fold reduced ABC susceptibility, and that addition of the 184V mutation, but not E44D and/or V118I, had a significant effect (Abstract 14) [3]. Wiese et al. determined that ABC-resistant viruses lacking 184V had the 118I change in all cases (causing low level resistance; Abstract 23) [4].

The findings of a large analysis looking into the prevalence of the G333D/E mutation were presented (Abstract 17) [5]. This mutation has previously been shown to contributing to dual-NRTI resistance in a subset of patients during AZT/3TC combination therapy. In a population of patients failing triple therapy who had previously received AZT/3TC, the prevalence of the 333 change was 12%. In 73% of patients, AZT and/or 3TC mutations were also present. The significance of the 333D/E remains to be determined in prospective studies.

The results of a trial in which tenofovir (TDF) was added to stable regimens in 189 patients were presented (Abstract 40) [6]. A durable decline in viral load of 0.6 log10copies/ML was seen with 300 mg TDF. A mean change of 2-fold resistance to TDF was observed (range 0.4-6) in four patients harboring HIV with the 65R mutation; this change has also been observed in vitro. In addition, the classical thymidine-associated mutations (TAMs) developed. However, it remains to be determined to what extent this was attributable to concomitant thymidine analog therapy.

Confusing data were presented regarding the development of Kaletra resistance in a trial comparing Kaletra plus d4T/3TC with nelfinavir plus d4T/3TC (Abstract 41) [7]. No changes in resistance to Kaletra were detected in the patients failing Kaletra, while 32% of failing patients had resistance to nelfinavir in the nelfinavir arm. Another unexpected finding was that only 42% of the failing patients in the Kaletra arm developed the 184V mutation compared to 82% in the nelfinavir arm. No clear explanation for these differences were given, but it remains to be determined to what extent compliance issues were responsible.

A large number of presentations dealt with the development and performance of genotype interpretation/conversion systems. The prospective Havana study showed that the use of genotyping combined with a decision support algorithm (Retrogram 1.2, Virology Networks) led to improved outcomes in a patient population consisting of 316 ART-experienced patients (Abstract 34) [8]. A retrospective comparison of the value of the Virtual Phenotype (Virco, Cambridge, UK) compared to rules developed by Resistance Collaborative Group back in 1999 showed that the Virtual Phenotype performed better. This was not very surprising given that the rules were quite old and the latest version of the Virtual Phenotype was used. It remains to be determined to what extent the difference in performance was merely a function of difference in the age of the two systems (Abstract 35) [9]. Interestingly comparison of different interpretation/conversion systems showed that systems using algorithms performed better than simple rules-based systems when applied to outcome in a clinically followed cohort (Abstract 37) [10]. A key limitation of these comparative studies is their retrospective design, and prospective studies are urgently needed in order to gain insight into the true value of the different conversion systems. In one study evaluating the value of prospective genotyping, it was shown that HIV RNA levels at failure, specific drug mutation patterns and recycling drug strategies influenced therapy outcome.

Finally, various novel methods for detecting phenotypic or genotypic resistance were discussed in a technology session. Most of the assays presented differed from current ones in that they showed greater sensitivity or enabled the testing of larger numbers of samples. The clinical value of using tests with increased sensitivity to needs to be established, however.

See Dr. Mark Wainberg's report from the 3^rd European Symposium on the Clinical Implications of HIV Drug Resistance