Comparison of a rule-based algorithm with a phenotype-based algorithm for the interpretation of HIV genotypes in guiding salvage regimens in HIV-infected patients by a randomized clinical trial: the mutations and salvage study.

BACKGROUND: There is still considerable uncertainty as to the best algorithm for interpreting human immunodeficiency virus (HIV) genotyping results. METHODS: A total of 318 subjects with HIV RNA levels of >1000 copies/mL were enrolled in 41 centers throughout Italy from 2001 through 2003, stratified on the basis of their drug history, randomized (1:1) to 2 arms to have their treatments modified on the basis of the results of HIV genotyping (as interpreted by virtual phenotype analysis or with use of a rule-based interpretation system), and followed up for 48 weeks. At least 1 nucleoside reverse-transcriptase inhibitor and 1 protease inhibitor had to be included in any new regimen; nonnucleoside reverse-transcriptase inhibitor-naive patients were also prescribed a nonnucleoside reverse-transcriptase inhibitor. Only drugs licensed in Italy were allowed. The primary end point was a decrease in HIV RNA level to <400 copies/mL by week 12 according to on-treatment analysis. RESULTS: The mean (+/- standard deviation) values at baseline were as follows: HIV RNA level, 4.1+/-0.74 log(10) copies/mL; CD4(+) T lymphocyte count, 410+/-262 cells/microL; reverse-transcriptase mutations, 4.8+/-2.9; and protease mutations, 2.8+/-2.5. There were 133 patients (41.8%) who were nonnucleoside reverse-transcriptase inhibitor naive and protease inhibitor experienced, 63 patients (19.8%) who were nonnucleoside reverse-transcriptase inhibitor experienced and protease inhibitor naive, and 122 patients (38.4%) who were 3-class experienced. A total of 192 patients completed 12 weeks of the treatment regimen assigned at baseline; at 12 weeks, 66.3% of patients in the virtual phenotype arm and 71.3% of patients in the rule-based interpretation arm had HIV RNA levels of <400 copies/mL (P = .46). No statistically significant difference between arms was observed by intention-to-treat analysis. CONCLUSION: Both the virtual phenotype and rule-based interpretation methods of HIV genotyping can guide the selection of effective antiretroviral drugs for a salvage regimen.

A Grid-based HIV expert system.

OBJECTIVES: This paper addresses Grid-based integration and access of distributed data from infectious disease patient databases, literature on in-vitro and in-vivo pharmaceutical data, mutation databases, clinical trials, simulations and medical expert knowledge. METHODS: Multivariate analyses combined with rule-based fuzzy logic are applied to the integrated data to provide ranking of patient-specific drugs. In addition, cellular automata-based simulations are used to predict the drug behaviour over time. Access to and integration of data is done through existing Internet servers and emerging Grid-based frameworks like Globus. Data presentation is done by standalone PC based software, Web-access and PDA roaming WAP access. The experiments were carried out on the DAS2, a Dutch Grid testbed. RESULTS: The output of the problem-solving environment (PSE) consists of a prediction of the drug sensitivity of the virus, generated by comparing the viral genotype to a relational database which contains a large number of phenotype-genotype pairs. CONCLUSIONS: Artificial Intelligence and Grid technology are effectively used to abstract knowledge from the data and provide the physicians with adaptive interactive advice on treatment applied to drug resistant HIV. An important aspect of our research is to use a variety of statistical and numerical methods to identify relationships between HIV genetic sequences and antiviral resistance to investigate consistency of results.

Tenofovir resistance and resensitization.

Human immunodeficiency viruses in 321 samples from tenofovir-naïve patients were retrospectively evaluated for resistance to this nucleotide analogue. All virus strains with insertions between amino acids 67 and 70 of the reverse transcriptase (n = 6) were highly resistant. Virus strains with the Q151M mutation were divided into susceptible (n = 12) and highly resistant (n = 8) viruses. This difference was due to the absence or presence of the K65R mutation, which was confirmed by site-directed mutagenesis. Viral clones with various combinations of the mutations M41L, K70R, L210W, and T215F or T215Y were analyzed for cross-resistance induced by thymidine analogue mutations (TAMs). The levels of increased resistance induced by single, double, and triple mutations at the indicated positions could be ranked as follows: for mutants with single mutations, mutations at positions 41 > 215 > 70; for mutants with double mutations, mutations at positions 41 and 215 > 70 and 215 = 210 and 215 > 41 and 70; for mutants with triple mutations, mutations at positions 41, 210, and 215 > 41, 70, and 215. Viral clones with M184V or M184I exhibited slightly increased susceptibilities to tenofovir (0.7-fold). Almost all clones with TAM-induced resistance were resensitized when M184V was present (P < 0.001). Among the viruses in the clinical samples, the rate of tenofovir resistance significantly increased with the number of TAMs both in the samples with 184M and in those with 184V (P = 0.005 and P = 0.003, respectively). A resensitizing effect of M184V was confirmed for all samples exhibiting at least one TAM (P = 0.03). However, accumulation of at least two TAMs resulted in more than 2.0-fold reduced susceptibility to tenofovir, irrespective of the presence of M184V. Decision tree building, a classical machine learning technique, was used to generate models for the interpretation of mutations with respect to tenofovir resistance. The application of previously proposed cutoffs for a reduced response to therapy and treatment failure demonstrated the central roles of positions 215 and 65 for 1.5- and 4.0-fold reduced susceptibilities, respectively. Thus, clinically relevant resistance may be conferred by the accumulation of TAMs, and the resensitizing effect of M184V should be considered only minor.

Comparison between rules-based human immunodeficiency virus type 1 genotype interpretations and real or virtual phenotype: concordance analysis and correlation with clinical outcome in heavily treated patients.

We compared 2 rules-based genotype interpretation systems and real or virtual phenotype through a retrospective analysis of a prospective trial. Genotypes were determined with VircoGEN II (VIRCO) and were interpreted with either RetroGram 1.4 or TRUGENE HIV-1 (guidelines 3.0) or original virtual phenotype (Virtual Phenotype; VIRCO), as available in the year 2000. Among 188 human immunodeficiency virus (HIV) type 1 isolates, overall concordance (kappa agreement) was observed for the 2 rules-based systems, whereas striking discordances were noted between them and real and virtual phenotype interpretations for stavudine, didanosine, zalcitabine, abacavir, and amprenavir (kappa<0.4). Clinical evaluation of a subset of 173 patients showed that both rules-based sensitivity scores were independently associated with HIV RNA loads <400 copies/mL at week 16 of during-treatment analysis (TRUGENE: odds ratio [OR], 2.90; 95% confidence interval [CI], 1.52-5.52; P=.001; RetroGram: OR, 2.34; 95% CI, 1.21-4.55; P=.012), whereas, in contrast to real or virtual phenotype, interpretations according to biological cut-offs were not (OR, 1.91; 95% CI, 0.77-4.76; P=.162).

Selection of zidovudine resistance mutations and escape of human immunodeficiency virus type 1 from antiretroviral pressure in stavudine-treated pediatric patients.

The relationship between clinical changes in stavudine activity and stavudine resistance was investigated in 16 human immunodeficiency virus (HIV)-infected children who received stavudine monotherapy for 18 months. Seven patients responded well to stavudine therapy, 3 experienced transient reductions in virus load, and all others had no detectable virologic response. In both the responders and nonresponders, no changes in stavudine susceptibility or specific baseline/emergent mutations in reverse transcriptase were observed. Only posttherapy HIV isolates from transient responders had elevated IC(50) values for stavudine. In 2 of the 3 transient responders, substitutions at codons 41, 210, and 215 were selected. The significance of these mutations was confirmed in viral competition experiments, site-directed mutagenesis, and in vitro selection. Selection of mutations previously associated with zidovudine resistance can be an important mechanism through which HIV may escape stavudine. The effect of these mutations on phenotypic stavudine susceptibility is relatively small but apparently large enough to be clinically significant.