Despite its potency, highly active antiretroviral therapy (HAART) does not control HIV-1 replication in some patients. Viral resistance to antiretroviral drugs represents the principal cause for this failure1,2 and often forces the clinician to withdraw current therapy and prescribe alternative medications. Because of the wide cross-resistance within each antiretroviral class, however, the sequential use of drugs is frequently problematic, and the most effective combination cannot be predicted solely on the basis of the treatment history. Testing viral susceptibility is now within clinicians' reach and has been encouraged by international guidelines.3-7 In fact, several trials, both retrospective1,8-11 and perspective,12-18 have demonstrated that genotypic12-15 or phenotypic16 assays for the evaluation of HIV-1 drug resistance are better predictors of virologic outcome than the standard of care. In general, genotypic testing is faster, more accessible, and less costly than phenotypic testing; however, the interpretation of mutation patterns is complex and requires expert advice.17,18 As an alternative approach for resolving mutation patterns, Virco (Virco NV, Mechelen, Belgium) has developed the genotype/virtual phenotype (vPt), which uses a large genotypic-phenotypic database based on sequence data to deduct the viral phenotype. Clinical trials are ongoing to validate the correlation between vPt and real phenotype (rPt) prospectively with regard to the virologic and immunologic outcome.19 By intuition, phenotypic assays, which provide direct quantitative measures of drug susceptibility, would be a better instrument than genotyping, but whether phenotype is actually more advantageous than genotype is still debatable, and clinicians are perplexed about the relative utility of each type of assay20 and undecided on the choice when requiring resistance testing.
The Narval study, which evaluated genotyping and phenotyping side by side, found no significant difference in the virologic response to treatment modifications based on genotyping, phenotyping, or standard of care.21 In a subgroup of patients with limited protease inhibitor (PI) experience, however, a significant benefit was observed in the genotyping arm. Whether or not the results of this trial can be regarded as definite, there is undoubtedly a need to evaluate the validity of genotyping or phenotyping in various clinical settings thoroughly, with particular attention to the extent of previous exposure to antiretroviral drugs.
In the context of the MaSTeR (Management Standardizzato di Terapia AntiRetrovirale) Italian Study Group, 2 randomized multicenter studies were designed with the aim of establishing the relative strengths and weaknesses of genotypic and phenotypic HIV-1 resistance assays to select a new antiretroviral regimen. Between May and July 2000, the Genotipo-Fenotipo di Resistenza (GenPheRex) study22 enrolled a "multifailure" patient population (with exposure to >6 drugs and at least 2 years of exposure to antiretrovirals), but no significant difference was demonstrated between the 2 approaches. As a logical completion of the GenPheRex study, the present article reports the final results of the 6-month follow-up of the phenotype/genotype (PhenGen) study in which the clinical usefulness of vPt versus rPt was evaluated for choosing a new regimen after HAART failure in a population of anti-HIV-positive subjects with limited antiretroviral experience.
PATIENTS AND METHODS
The patients participating in the PhenGen study were adult (age >18 years) HIV-positive subjects with a plasma viral load (pVL) between 2000 and 200,000 copies/mL during therapy and a CD4 cell count >200 cells/μL. Patients with ≥1 previous failure (defined as a plasma HIV RNA load >2000 copies/mL in at least 2 consecutive determinations), including at least 1 HAART regimen (nonnucleoside reverse transcriptase inhibitor [NNRTI]- or PI-based regimen) were enrolled if their total antiretroviral experience did not exceed 6 antiretroviral drugs, each of which was administered for at least 12 consecutive weeks. Patients had to be on treatment at the time of entry into the study. The exclusion criteria included pregnancy, active drug abuse, and any active opportunistic infection (patients on primary or secondary antiopportunistic prophylaxis were admitted). All participants provided written informed consent. The protocol was approved by the Ethical Committee of the coordinating center (Clinic of Infectious Diseases, University of Foggia, Foggia, Italy).
Study Design and End Points
The PhenGen study was a multicenter study that recruited patients from 30 Italian infectious disease centers participating in the national MaSTeR cohort study and was designed as a prospective, randomized, open-label study comparing genotype with vPt and rPt for prediction of virologic response to a new regimen administered to patients failing previous antiretroviral therapy. The new treatment choice was also supported by the expert advice of 3 physicians with proven experience (from centers with at least 500 patients on therapy) regarding antiretroviral therapy and interpretation of resistance tests, who had access to the patients' clinical data (including sex, age, CD4 cell count, viral load, history of treatment, and causes of therapy change or interruptions). Following the recommendations of the coordinating center, the expert panel was instructed to attempt to adjust the new regimen to include previously used drugs (that still resulted in active results at testing) rather than new medications so as to reserve the latter for future options. The expert suggestion was communicated to physicians before therapy choice by telefax or e-mail and could be discussed directly with experts if requested by physicians.
The primary end points of the study were to determine (1) the proportion of patients with an undetectable pVL (HIV RNA <400 copies/mL) at months 3 and 6, (2) the mean change in plasma HIV RNA level at months 3 and 6, and (3) the mean change in CD4 cell count throughout the follow-up period. Secondary end points were to establish (1) the median duration of time to achieve a pVL level of <400 copies/mL, (2) the proportion of patients developing an AIDS-defining clinical event throughout the study period, (3) the occurrence of adverse events and/or toxicities limiting treatment adherence, and (4) the rate of physician compliance with the expert advice.
Study Procedures and Laboratory Assessments
At week −4, eligible patients were randomized at a 1:1 ratio into 2 treatment arms chosen according to the results of either a genotype with vPt (vPt arm) or a genotype with rPt (rPt arm). Randomization was carried out by the coordinating center using a matrix of casual numbers to generate the random allocation sequence, which was communicated to the participating center. Plasma samples collected at the time of entering the PhenGen study (week −4) were used for assessment of genotypic or phenotypic drug resistance. Samples were analyzed by Virco using the Virco genotyping system (VircoGen) with subsequent determination of the vPt (VirtualPhenotype, Virco) and the Antivirogram phenotype (Antivirogram, Virco).
Phenotypic data were expressed as "fold resistance" and calculated by dividing the 50% inhibitory concentration (IC50) of each medication for the recombinant virus from the patient by the IC50 for the wild-type reference strain. Viruses with a fold resistance value greater than or less than the biologic cutoff were interpreted as sensitive or resistant to each single drug, respectively. According to the manufacturers' instructions, the cutoff values were as follows: zidovudine (ZDV), 4.0; lamivudine (3TC), 4.5; didanosine (ddI), 3.5; zalcitabine (ddC), 3.5; stavudine (d4T), 3.0; abacavir (ABC), 3.0; nevirapine (NVP), 8.0; delavirdine (DLV), 10.0; efavirenz (EFV), 6.0; indinavir (IDV), 3.0; ritonavir (RTV), 3.5; nelfinavir (NFV), 4.0; saquinavir (SQV), 2.5; amprenavir (APV), 2.5; and lopinavir (LPV), 2.5. Tenofovir DF was also tested with a cutoff value equal to 3.0 as of November 2001. For the vPt, the cutoff values were the same as those for the rPt, with some exceptions: ddI, 2.0; ddC, 2.0; and d4T, 1.8. By January 2002, the cutoff value for d4T was lowered to 1.75. A drug was considered active or inactive if either the rPt or vPt was interpreted as sensitive or resistant, respectively.
Results were mailed to the participating centers and the coordinating center within 2 weeks for the vPt and within 4 weeks for the rPt. The new antiretroviral regimen was determined by the clinician in each center, who used information derived from the genotype or phenotype in the corresponding arm; centers also received expert advice recommending treatment within 1 week from reception of resistance test results. Patients initiated the new regimen within 4 weeks from the availability of test results. The following drugs were available at the beginning of the study: ZDV, ddI, d4T, 3TC, ABC, NVP, EFV, RTV, IDV, NFV, SQV, and APV. During the study, LPV/ritonavir (LPV/r) became commercially available, and tenofovir DF was also available in some centers in the expanded access program.
Plasma HIV-1 RNA, CD4 cell count, adverse events, and any therapy variation or interruption were recorded monthly for a period of 6 months. CD4 cell counts were performed locally using standard flow cytometry. Plasma HIV-1 RNA viral load was measured by the standard method used in each participating center (Amplicor reverse transcriptase polymerase chain reaction [RT-PCR] [Roche], NASBA [Organon-Tecknika], or b-DNA [Chiron]). The highest accepted threshold value for HIV RNA was 400 copies.
The local physician also provided a judgment regarding the patient's adherence based on the reported doses missed in the weeks before each study visit (scarce: >50% of missed doses, good: <5%-10% of missed doses, and very good: no missed doses).
All data were collected at each participating center and sent to the coordinating center.
At the end of the study, complete protease and RT sequences encompassing codons 1 through 400 were available for all patients in the vPt and rPt arms. Genotypic data were reported as sequence changes with respect to a consensus B. Mutations at all positions in the protease and RT domains were recorded; mutations were classified as resistance-associated mutations according to the International AIDS Society (IAS)-USA Drug Resistance Mutations Group.23 Nucleoside reverse transcriptase inhibitor (NRTI)-associated mutations (NAMs) that are associated with cross-resistance to NRTIs were those listed by the IAS23 and included the M41L, E44D, D67N, K70R, V118I, L210W, T215YF, and K219Q mutations. Universal protease-associated mutations (UPAMs) comprised the L33F, V82AFTS, I84V, and L90M mutations.24 For the purpose of analysis, mixtures of mutant and wild-type viruses were considered as mutants, as were viruses with multiple amino acid substitutions, including those indicating resistance. For substitutions at positions not indicated by the IAS, only those mutations detected in >20% of patients were considered.
When the sample size in each group is 150, a 2-group, 2-sided t-test with an α-error = 0.05, has a 73% power to reject the null hypothesis that the difference in mean decrease in pVL between the 2 groups is ≥0.3 log copies/mL. This favors the alternative hypothesis that the means of the 2 groups are equivalent, assuming that the expected difference in means is 0.0 and the common standard deviation (SD) is 1.0.
Descriptive statistics were produced for the demographic, clinical, and viroimmunologic characteristics of all cases. The mean and SD are presented for normally distributed variables, and the median and interquartile range (IQR) are presented for nonnormally distributed variables. Minimum-maximum ranges are also presented. The Student t test (rank sum test or median test for skewed distributions) was used to compare quantitative variables, and the Pearson χ2 test (Fisher exact test where appropriate) was used for categoric variables.
The primary analysis was performed according to an intent-to-treat (ITT) approach, where the denominator was composed of all randomized patients for whom amplification was successful and who initiated the prescribed treatment. The principal outcome indicator was the pVL mean decrease (log copies/mL) at 6 months from starting therapy. Undetectability (pVL<400 copies/mL) at 3 and 6 months as well as mean CD4 cell count increase was also compared between the 2 groups.
For the on-treatment (OT) analysis, patients who switched, discontinued, or suspended the initial therapy (for any reason) were excluded. In addition, the response predictors (log copies/mL and CD4 cell count over time) were investigated by means of an autoregression of order 1 regression models (generalized estimating equations [GEEs]). Univariate and multivariate models were fitted using a forward-fitting strategy. Differences between models were tested by means of the likelihood ratio test (LRT). The variables considered were age; sex; baseline CD4 cell count; baseline pVL; resistance test used; number of drugs experienced; number of treatment lines; number of NRTIs, NNRTIs, and PIs experienced; viral load never <400 copies/mL; total number and individual resistance-associated mutations; number of sensitive drugs in the new regimen; adherence to therapy (defined as reported doses missed); and adherence to the expert advice.
Between February 2001 and March 2002, 303 of the 328 patients consecutively evaluated were randomized (152 in the vPt arm and 151 in the rPt arm). The remaining 25 patients were excluded because they did not meet the inclusion criteria. The flow of participants through each stage of the study is summarized in Figure 1. A total of 111 and 108 patients received a new treatment in the vPt and rPt arms, respectively. Either no genetic material could be amplified or susceptibility testing did not meet quality assurance standards for 26 (8.6%) patients in both arms. The new treatment was not initiated in an additional 37 (12.2%) patients, including 6 subjects who did not turn out for the baseline visit and 31 patients who refused therapy or dropped out of the study between week −4 and baseline. Finally, extensive drug resistance was determined in the remaining 21 (6.9%) subjects; these patients with a mean CD4 cell count >500 cells/μL underwent a supervised interruption of therapy before starting a new regimen. Overall, these 84 patients (41 in the vPt arm and 43 in the rPt arm) were excluded from the primary analysis.
The baseline characteristics of patients enrolled in the study are listed in Table 1. Patients were well matched for age, sex, Centers for Disease Control and Prevention (CDC) stage, mean number of previous regimens and drugs, median CD4 cell count (342 cells/μL in the vPt arm vs. 351 cells/μL in the rPt arm), and pVL (4.22 vs. 4.15 log10 copies/mL, respectively). Similarly, prior treatment history did not differ between the 2 groups: all subjects had previously been exposed to a median of 5 antiretroviral drugs and 3 regimens, and 52% of patients in the vPt arm and 46% in the rPt arm were naive to NNRTIs. The proportions of patients with first-line HAART failure, second-line HAART failure, and third-line HAART failure were 27%, 53%, and 20% in the vPt arm compared with 25%, 54%, and 21% in the rPt arm. After excluding the 84 patients who did not start therapy, the 2 groups remained comparable. Overall, 101 of 111 patients in the vPt arm and 104 of 108 patients in the rPt arm completed follow-up. Twenty-seven (26.7%) of 101 patients in the vPt arm and 29 (27.9%) of 104 patients in rPt arm changed therapy or interrupted treatment at least once. The reasons for premature changing or interrupting therapy were as follows: refusal to continue therapy (43%), adverse drug reactions (28%), lack of efficacy (13%), development of diseases that required hospitalization (7%), and death (8%). The principal characteristics (including sex, age, clinical status, number of prior regimens and antiretroviral medications, and mode of acquisition of HIV infection) of patients changing therapy did not differ from those of patients who continued the prescribed regimen.
The proportion of isolates that were resistant to antiretroviral drugs in each arm at baseline is shown in Figure 2A and is compared with the proportion of subjects who were receiving each single medication at the time of testing (see Fig. 2B). No significant differences were found between the vPT and rPT arms for the proportion of isolates resistant to each drug. Among NRTIs, as expected, a low prevalence of ddI- and d4t-resistant strains was found in both arms: in fact, although >50% and up to 27% of failed patients in each arm were treated with d4T and ddI, respectively, resistant isolates did not exceed 12%. Conversely, AZT-resistant isolates in the vPt arm and ABC-, NNRTI-, and PI-resistant isolates in both arms surpassed the percentage of patients receiving the specific drug-based regimens. Even though all patients were naive to tenofovir DF, a total of 6.2% isolates (5.9% in the vPt arm and 6.6% in the rPt arm) with ascertained susceptibility to the drug demonstrated low to moderate resistance to tenofovir. Similarly, in the overall population, only 1 and 4 patients were receiving APV and LPV/r, respectively; in the vPt arm, 18.8% and 14.8% of the tested isolates were resistant to APV and LPV, respectively, whereas in the rPt arm, a reduced susceptibility to these PIs was detected in 14.1% and 12.9% of isolates, respectively.
Mutational Resistance Profiles
The mean number of total and resistance-associated mutations (as listed by the IAS) detected within the protease and the RT domains is reported in Table 2. Figure 3 shows the prevalence of protease (3A) and RT (3B) mutations in isolates from the vPt arm and the rPt arm. Overall, a similar mutation distribution between the 2 arms was noted.
No patient had the simultaneous presence of all 4 UPAMs: 49% of isolates showed at least 1 UPAM in the vPt arm and 38% in the rPt arm, and 17% of isolates had more than 2 UPAMs in each arm. At least 3 NAMs were present in 50% of patients tested in the vPt arm and in 51% of patients tested in the rPt arm, whereas more than 6 NAMs were observed in 6% and 5% of isolates in the 2 arms, respectively (see Table 2). Multi-NRTI resistance because of the presence of the Q151M complex was observed in 6 cases (3 in each arm), whereas no isolates showing the 69 insertion were found (see Fig. 3).
Prescribed Treatment and Outcomes
Resistance testing strongly influenced the choice of antiretroviral agents for the new regimen, because the drugs most frequently resulting in sensitivity (d4T and ddI among the NRTIs and IDV and LPV/r among the PIs) were also the most frequently prescribed (Fig. 4).
Of the patients not previously exposed to NNRTIs, these drugs were administered to 70% of patients in the vPt arm and 61% of patients in the rPt arm. The preferred regimen prescribed at baseline was a combination of 2 NRTIs and 1 PI (alone and with a booster dose of RTV) (41%), followed by the association of 2 NRTIs with 1 NNRTI (36%).
A total of 83 (75%) of 111 patients in the vPt arm and 87 (80%) of 108 patients in the rPt arm received a new regimen containing 3 active drugs (defined as any medication for which no evidence of genotypic or phenotypic resistance was found, independent of previous use). The proportion of patients receiving at least 1 recycled drug was 78.4% in the vPt arm versus 84.4% in the rPt arm (P = 0.7). In addition, when the patients were divided according to reception/rejection of expert advice by the clinicians, the mean number of recycled drugs did not differ between the 2 groups (1.2 ± 0.8 vs. 1.3 ± 0.8 recycled drugs). As shown in Figure 4, the most frequently recycled drugs were the NRTIs: d4T (>40% of cases) and ddI (>25% of cases), followed by AZT and 3TC in both arms; among the PIs, IDV in the vPt arm (6%) and SQV in the rPt arm (3%) were readministered.
The new therapy corresponded to the expert advice in 61% of cases in the vPt arm and 56% of cases in the rPt arm. The reasons for not using the drugs suggested by the expert panel were either the patient's concern about toxicity or problems concerning patient adherence as estimated by the physician.
Treatment compliance was considered very good by the physician in 25% of patients, good (5%-10% of missed doses) in 54%, and scarce in the remaining 25%, with no significant difference between the 2 arms of the study.
When using an ITT approach with missing values considered as failures, the proportion of patients with a pVL of <400 copies/mL was 62.5% in vPt arm and 61.4% in rPt arm (P = 0.9) at month 3 and 54.8% and 52.6%, respectively, at month 6 (P = 0.9), without any statistical difference noted between the 2 groups. No statistical differences were found in the mean log of HIV RNA pVL decrease (−1.46 vs. −1.56, respectively, P = 0.6 at month 3; and −1.35 and −1.37, respectively, P = 0.8 at month 6; Fig. 5A) or in the mean absolute CD4+ cell count variation at month 3 (+46 vs. +56 cells in the vPt and rPt arms, respectively, P = 0.6) and at month 6 (+55 vs. +46 cells in the vPt and rPt arms, respectively, P = 0.7; see Fig. 5B). The OT analysis provided similar results with respect to the proportion of patients with a pVL of <400 copies/mL (37.5% vs. 41.8%, P = 0.6 at month 3; and 54.5% vs. 51.6%, respectively, at month 6, P = 0.7), the mean log of HIV RNA pVL decrease (−1.70 vs. −1.69, respectively, P = 0.9 at month 3; and −1.39 vs. −1.59, respectively, P = 0.3 at month 6), and the mean absolute CD4+ cell count change at month 3 (+46 vs. +57 cells in the vPt and rPt arms, respectively, P = 0.6) and at month 6 (+45 vs. +63 cells in the vPt and rPt arms, respectively, P = 0.4).
The results of univariate and multivariate analysis of factors that might influence the virologic outcome are shown in Table 3. Univariate analysis revealed that the CDC clinical stage; pVL at baseline; number of total amino acid changes; RT and protease mutations; number of NAMS; and resumption of 3TC, ABC, NVP, EFV, and IDV (even when sensitive to resistance testing) were all predictive of virologic failure (HIV RNA level >400 copies/mL). At multivariate analysis, variables independently associated with failure of the new regimen were pVL at baseline (odds ratio [OR] = 1.81, 95% confidence interval [CI]: 1.10-3.01; P = 0.021), number of NAMS (OR = 1.21, 95% CI: 1.08-1.35; P < 0.001), number of protease mutations (OR = 1.15, 95% CI: 1.08-1.22; P < 0.001), and resumption of IDV (OR = 4.63, 95% CI: 1.29-16.64; P = 0.019). At univariate and multivariate analysis, the patient's adherence to the prescribed regimen (OR = 0.23, 95% CI: 0.12-0.45; P < 0.001), number of active drugs included in the new combination therapy (OR = 0.55, 95% CI: 0.38-0.78; P = 0.001), and adherence to the expert panel suggestions (OR = 0.37, 95% CI: 0.23-0.59; P < 0.001) predicted virologic response.
At week 24, 49 adverse events were reported (11 considered serious) affecting 12% of the patients in the vPt arm and 14% in the rPt arm. Only 2 AIDS-defining events were reported (both in the vPt arm). Five deaths occurred: 4 in the vPt arm as a result of malignancies in 3 cases (lung cancer, bone marrow aplasia, and testicular lymphoma) and an automobile accident in 1 case and 1 in the rPt arm as a result of hepatocellular carcinoma.
The aim of this study was to verify which assay is more beneficial-the rPt or the genotype plus vPt-when used for choosing a new HAART regimen in failed patients. No standard-of-care arm was included in the PhenGen study, because when this trial was designed, previous studies12-16 had already ascertained that resistance testing significantly improved the virologic response of salvage therapy compared with the standard of care. The results of our trial demonstrate that 62% of patients in the vPt arm versus 61% in the rPt arm had an undetectable pVL at month 3 and 55% versus 53% at month 6, respectively, thereby indicating that both approaches permit a good response to therapy in more than half of our patients with little previous drug exposure. For the most part, however, therapeutic decisions guided by genotypic data seemed to be slightly more beneficial virologically and immunologically.
The results of the PhenGen study are not unique and are similar to those of the GenPheRex and VIHRES25 studies, in which no differences were found between the use of vPt or rPt in a heavily pretreated population, as well as those of the RealVirfen study,26,27 where the slight superiority of the vPt over the rPt did not reach statistical significance. Overall, these observations are supported by the results of an even smaller study28 comparing antiretroviral resistance susceptibility testing of patient HIV-1 strains using genotypic and phenotypic methods. In this study, the concordance between the 2 approaches was 81% for NRTIs, 91% for NNRTIs, and 90% for PIs; therefore, phenotypic and genotypic susceptibility seemed to provide similar results. This conclusion is not unexpected; resistance is currently assessed by DNA sequencing or by measuring the drug concentration required to inhibit viral replication using recombinant technology, which, in turn, utilizes the same plasma-derived PCR product as in the genotype assay.
The crucial point still remains the interpretation of vPt and rPt results, which are dependent on criteria that are frequently updated and improved. As a result, the explanation of mutation patterns is complex, and the clinical relevance of susceptibility fold changes for each single medication in the phenotypic assay can occasionally vary when used in the context of combination regimens. In our study, genotyping and phenotyping were performed by Virco, and genotyping results were interpreted with the vPt, which uses a large database containing paired real phenotypic and genotypic profiles for retrieving the IC50 values of the virus with the sequence most similar to that of the patient. Therefore, the reciprocal results from the vPt and rPt are somewhat foreseen. Previous studies22,29 have already suggested an optimal correlation between the vPt and rPt, and Perez-Elias et al30 found that the vPt performed as well as or better than the rPt for guiding the selection of antiretroviral therapy in patients failing more than 1 regimen, thereby encouraging clinicians to use the vPt rather than the less-accessible rPt. When considering the use of genotyping for larger populations than those currently studied, however, cost-free algorithms that permit computer-assisted interpretation of mutation patterns would be the optimal solution. Actually, discordances between available genotypic drug resistance interpretation algorithms have already been reported31,32; furthermore, whether these expert systems work as well as the vPt has not yet been clarified. In a substudy33 of the GenPheRex trial, in which sequence data were retrospectively analyzed by means of Retrogram 1.4 or Trugene HIV-1 3.0, the interpretation of mutation patterns diverged from the rPt and vPt, particularly for susceptibility to d4T, ddI, ddC, ABC, and APV. After a sensitivity score for drugs in the salvage regimen was assessed by each method and correlated with the treatment response, however, computer-assisted interpretations provided a closer association with the virologic outcome compared with the vPt and rPt.
In our study, the use of expert recommendations significantly improved response to therapy. Fifty-nine percent of patients followed expert advice, and at multivariate analysis, the acceptance of these suggestions was independently associated with a better outcome. Similar virologic benefit results were found with the use of expert advice in the Havana trial14 independent of contemporary use of a genotype test. Overall, these results have been further confirmed by a retrospective study aiming to assess the impact of expert advice in routine clinical practice. In this latter study, patients who received HAART regimens based on expert advice and genotypic resistance testing had significantly better virologic outcomes than patients who did not receive expert advice-recommended antiretroviral regimens.34 In a stratified analysis, however, both of these studies14,34 showed that virologic response was improved by experts' suggestions only in a subset of patients for whom the second HAART regimen was failing, whereas no benefit was found in patients with more than 3 HAART failures. In the PhenGen study, the patients' drug experience was limited, and at multivariate analysis, the advantage conferred by expert advice was independent of the number of previous HAART regimen failures and the number of active drugs in the new combination. Therefore, in agreement with the above-mentioned studies,14,34 we conclude that in highly retreated patients for whom no additional active drugs are available, a good virologic response is somewhat impeded, regardless of whether expert suggestion is followed. Conversely, we believe that expert opinion is helpful in less-experienced patients, even after a first HAART failure when a correct choice of therapy would preserve future options. A distinct advantage was always noted for those patients who followed the expert advice, in spite of the simplicity of the individual patient's report for both assays in the PhenGen study, even if the access to experts by all clinicians, including those who did not adhere to these suggestions, might have represented a bias reducing the added value of the expert advice. In a recent report, Salama et al35 attempted to ascertain whether clinicians have adequate knowledge of relevant genotypic mutations and indicated that this information is markedly limited, even among clinicians caring for large numbers of patients, thus suggesting the need to furnish more update for HIV specialists and provide the opportunity of close contact with a recognized expert in this field.
In the present study, the expert panel was requested to adjust the new regimen with the aim of saving some therapeutic options for the future and, when possible, to recycle previously used drugs to which the patient virus was still sensitive. Overall, in 80% of patients, the new regimen included at least 1 past medication. ddI and d4T were the most frequently recycled drugs because of the lowest rates of resistance in both arms. The biologic cutoff values that were adopted for testing susceptibility to d4T and ddI are, at present, unacceptable,36 however, and the clinical cutoff values have yet to be established. Moreover, the isolates from one half of patients in the PhenGen study had at least 3 NAMs. Although the actual impact of NAMs on d4T resistance is still controversial,37 and we presume it will eventually be lower than that to AZT, d4T activity is undoubtedly diminished by NAMs. In fact, 3 or more of these mutations are associated with a relevant reduction in d4T sensitivity,38,39 and approximately 10% to 15% of individuals failing therapy with ddI also develop AZT resistance mutations instead of the classic L74V mutation.40 Finally, the presence of AZT resistance mutations, particularly at position 215, leads to a diminished response to subsequent therapy, including ddI or d4T.41 In spite of these questionable aspects, the recycling of these 2 drugs in our study was not associated with a worse outcome. The most likely explanation for this apparent incongruity is that when including these drugs in a regimen, the eventual suboptimal performance of d4T and ddI is compensated for by the activity or synergistic effect of the other medications. Conversely, at univariate analysis, the recycling of 3TC, ABC, NVP, and EFV was associated with virologic failure. The rapid re-emergence of a preexisting population of mutant viruses resistant to drugs with a low genetic barrier is not surprising; the real unexpected result of our study was that this effect vanished at multivariate analysis. Similarly, a few explanations can be proposed for the negative impact (at univariate and multivariate analysis) of recycling IDV in a small number of our patients. The emergence of a predominantly IDV-resistant virus population may be delayed for months, even in the presence of ongoing IDV therapy42,43; this means that IDV resistance could have been underestimated. Moreover, Zhang et al44 suggested a common pathway for IDV mutations, with M46L/I and V82A first to appear, followed by I54V or A71V/T. Overall, in our population, up to 20% of patients had M46L/I, 16% had V82A, 15% had I54V, and 38% had A71V/T at baseline, which probably determined a rapid appearance of resistance to IDV when it was readministered. This did not occur for other PIs, perhaps only because they were recycled less.
Actually, no difference regarding the number of recycled drugs in the new regimen was observed between patients who followed the expert advice and those who did not. This means that although resistance tests identify drugs for which viral strains retain susceptibility, expert intervention produces the most appropriate combination regimen in which these drugs may be beneficially included.
Finally, because our trial did not focus on assessment of patient adherence, its evaluation in this study was only based on the clinician's judgment; nevertheless, it correlated well with the virologic outcome.
The present study was designed to evaluate patients with limited drug resistance and is the natural completion of another Italian trial, the GenPheRex study,22 in which a multifailure patient population previously treated with more than 6 drugs and at least 2 years of exposure to antiretrovirals was enrolled between May and July 2000. As expected, because of the extensive antiretroviral experience of these patients, an undetectable pVL after a 6-month period was detected in only 24% and 20% of patients in the genotype and phenotype arms, respectively. The GenPheRex study and the PhenGen study had the same design, and resistance tests were performed in the same centralized laboratory; as to the question of multidrug failure or less drug experience, when considered together, the GenPheRex and PhenGen studies indicate a similar superior efficacy for rPt and vPt assays in patients with treatment options.
The authors thank all the participants in the PhenGen study. They also thank Paulene Butts for her helpful review of the manuscript.
1. DeGruttola V, Dix L, D'Aquila R, et al. The relation between baseline HIV drug resistance and response to antiretroviral therapy: re-analysis of retrospective and prospective studies using a standardized data analysis plan. Antivir Ther.
2. Lorenzi P, Opravil M, Hirschel B, et al. Impact of drug resistance mutations to be on virologic response to salvage therapy. AIDS.
3. The EuroGuidelines Group for HIV Resistance
. Clinical and laboratory guidelines for the use of HIV-1 drug resistance testing as part of treatment management: recommendation for the European setting. AIDS.
4. US Department of Health and Human Services. Guidelines for the use of antiretroviral agents in adults and adolescents; February 4, 2002. Available at: http://hivatis.org/trtgdlns.html
5. Hirsch MS, Brun-Vezinet F, Clotet B, et al. Antiretroviral drug resistance testing in adults infected with human immunodeficiency virus type 1:2003 recommendations of an International AIDS Society-USA Panel. Clin Infect Dis.
6. Pozniak A, Gazzard B, Anderson J, et al. British HIV Association (BHIVA) guidelines for the treatment of HIV infected adults with antiretroviral therapy. HIV Med.
7. Vandamme AM, Houyez F, Banhegyi D, et al. Laboratory guidelines for the practical use of HIV drug resistance tests in patients follow-up. Antivir Ther.
8. Zolopa AR, Schafer RW, Warford A, et al. HIV-1 genotypic resistance patterns predict response to saquinavir-ritonavir therapy in patients in whom previous protease inhibitor therapy had failed. Ann Intern Med.
9. Harrigan PR, Montaner JS, Hoggs RS, et al. Baseline resistance profile predicts response to ritonavir-saquinavir therapy in a community setting. AIDS.
10. Deeks SG, Hellman NS, Grant RM, et al. Novel four-drug salvage treatment regimens after failure of a human immunodeficiency virus type 1 protease inhibitor-containing regimen. J Infect Dis.
11. Call SA, Saag MS, Westfall AO, et al. Phenotypic drug susceptibility testing predicts long-term virologic suppression better than treatment history in patients with human immunodeficiency virus infection. J Infect Dis.
12. Baxter JD, Mayers DL, Wentworth DN, et al. A randomised study of antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy. AIDS.
13. Durant J, Clevenbergh P, Halfon P, et al. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet.
14. Tural C, Ruiz L, Holtzer C, et al. Clinical utility of HIV-1 genotyping and expert advice: the Havana trial. AIDS.
15. Cingolani A, Antinori A, Rizzo MG, et al. Usefulness of monitoring HIV drug resistance and adherence in individuals failing highly active antiretroviral therapy: a randomised study (ARGENTA). AIDS.
16. Cohen CJ, Hunt S, Sension M, et al. A randomized trial assessing the impact of phenotypic resistance testing on antiretroviral therapy (VIRA3001 Study Team). AIDS.
17. Haubrich R, Demeter L. International perspectives on antiretroviral resistance. Clinical utility of resistance testing: retrospective and prospective data supporting use and current recommendations. J Acquir Immune Defic Syndr.
18. Shafer RW. Genotyping testing for HIV-1drug resistance. Clin Microbiol Rev.
19. Larder BA, Kemp SD, Hertogs K. Quantitative prediction of HIV-1 phenotypic drug resistance from genotypes: the virtual phenotype
) [abstract 63]. Antivir Ther.
20. Parkin N, Chappey C, Maroldo L, et al. Phenotypic and genotypic HIV-1 drug resistance assays provide complementary information. J Acquir Immune Defic Syndr.
21. Meynard JL, Vray M, Morand-Jubert L, et al. Phenotypic or genotypic resistance testing for choosing antiretroviral therapy after treatment failure: a randomised trial. AIDS.
22. Mazzotta F, Lo Caputo S, Torti C, et al. Real versus virtual phenotype
to guide treatment in heavily pretreated patients: 48-week follow-up of the Genotipo-Fenotipo di Resistenza (GenPheRex) trial. J Acquir Immune Defic Syndr.
23. D'Aquila RT, Schapiro JM, Brun-Vèzinet F, et al. Drug resistance mutations in HIV-1. International AIDS Society-USA. Top HIV Med.
24. Gathe J, Kohlbrenner VM, Pierone G, et al. Tipranavir/ritonavir demonstrates potent efficacy in multiple protease inhibitor experienced patients: BI 1182.52 [abstract 179]. Presented at the 10th Conference on Retroviruses and Opportunistic Infections, Boston, February 10-14, 2003.
25. Blanco JL, Valdecillos G, Arroyo JR, et al. A prospective randomized study on the usefulness of genotypic resistance tests versus real phenotypic resistance tests in heavily pretreated patients with virological failure (VIHRES Study) [abstract uPeB4624]. In: Program and Abstracts of the 14th International AIDS Conference (Barcelona). Barcelona: Prous Science, S.A.; 2002:421.
26. Perez-Elias MJ, Garcia I, Munoz V, et al. A randomized, prospective study of phenotype
(P) versus virtual phenotype
(virtualP) testing for patients failing antiretroviral therapy (ARVT): final analysis [abstract H-1079]. In: Programs and Abstracts of the 42nd Interscience Conference on Anti-microbial Agents and Chemotherapy (San Diego). Washington, DC: American Society for Microbiology; 2002:268.
27. Perez-Elias MJ, Garcia-Arata I, Moreno S, et al. Agreement degree between simultaneous real phenotype
and virtual phenotype
in patients from the Realvirfen study [abstract 808]. Antivir Ther.
28. Dunne AL, Mitchell FM, Coberly SK, et al. Comparison of genotyping and phenotyping methods for determining susceptibility of HIV-1 to antiretroviral drugs. AIDS.
29. Ferrer E, Podzamczer D, Arnedo M, et al. Genotype
at baseline and at failure in human immunodeficiency virus-infected anti-retroviral-naive patients in a randomized trial comparing zidovudine and lamivudine plus nelfinavir or nevirapine. J Infect Dis.
30. Perez-Elias M, García-Arata I, Munoz V, et al. A randomized, prospective study of phenotype
versus virtual phenotype
testing for patients failing antiretroviral therapy [abstract 586-T]. In: Programs and Abstracts of the 9th Conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retrovirology and Human Health; 2002:269.
31. Schmidt B, Walter H, Schwingel E, et al. Comparison of different interpretation systems for genotypic HIV-1 drug resistance data. Antivir Ther.
32. Kijak GH, Rubio AE, Pampuro SE, et al. Discrepant results in the interpretation of HIV-1 drug-resistance genotypic data among widely used algorithms. HIV Med.
33. Torti C, Quiros-Roldan E, Keulen W, et al. Comparison between rule-based human immunodeficiency virus type 1 genotype
interpretations and real or virtual phenotype
concordance analysis and correlation with clinical outcome in heavily treated patients. J Infect Dis.
34. Badri SM, Adeyemi OM, Max BE, et al. How does expert advice impact genotypic resistance testing in clinical practice? Clin Infect Dis.
35. Salama C, Policar M, Cervera C. Knowledge of genotypic resistance mutations among providers of care to patients with HIV. Clin Infect Dis.
36. Shulman NS, Hughes MD, Winters MA, et al. Subtle decreases in stavudine phenotypic susceptibility predict poor virologic response to stavudine monotherapy in zidovudine-experienced patients. J Acquir Immune Defic Syndr.
37. Lafeuillade A, Tardy JC. Stavudine in the face of cross-resistance between HIV-1 nucleoside reverse transcriptase inhibitors: a review. AIDS Rev.
38. Demeter L, Nawaz T, Morse G, et al. Development of zidovudine resistance mutations in patients receiving prolonged didanosine monotherapy. J Infect Dis.
39. Miller V, Ait-Khaled M, Stone C, et al. HIV-1 RT genotype
and susceptibility to RT inhibitors during abacavir monotherapy and combination therapy. AIDS.
40. Winters M, Shafer R, Jellinger R, et al. HIV-1 RT genotype
and drug susceptibility changes in infected individuals receiving dideoxyinosine monotherapy for 1 to 2 years. Antimicrob Agents Chemother.
41. Shulman NS, Machekano RA, Shafer RW, et al. Genotypic correlates of a virologic response to stavudine after zidovudine monotherapy. J Acquir Immune Defic Syndr.
42. Harrigan PR, Larder BA. Extent of cross-resistance between agents used to treat human immunodeficiency virus type 1 infection in clinically derived isolates. Antimicrob Agents Chemother.
43. Havlir DV, Hellmann NS, Petropoulos CJ, et al. Drug susceptibility in HIV infection after viral rebound in patients receiving indinavir-containing regimens. JAMA.
44. Zhang Y-M, Imamichi H, Imamichi T, et al. Drug resistance during indinavir therapy is caused by mutations in the protease gene and in its gag substrate cleavage sites. J Virol.
Participants in the PhenGen study are as follows: G. Pastore, N. Ladisa, and P. Maggi (Bari); G. Carosi, F. Castelli, C. Torti, and L. Tomasoni (Brescia); A. Mandas, F. Pigliaru, and S. Manca (Cagliari); G. Angioni, S. Angioni, and G. Abeltino (Cagliari); P. Bellissima and S. Bonfante (Caltagirone); F. Ghinelli and L. Sighinolfi (Ferrara); F. Mazzotta, S. Lo Caputo, and P. Pierotti (Firenze); A. Gallo and L. Sarracino (Frosinone); M. Toti, T. Carli, and E. Donati (Grosseto); A. M. Orani and P. Perini (Lecco); A. Scasso and M. De Gennaro (Lucca); G. Todaro and G. Magaraci (Messina); M. Moroni, A. d'Arminio Monforte, and T. Bini (Milano); A. Cargnel and C. Atzori (Milano); A. Lazzarin and N. Gianotti (Milano); A. Chirianni and M. Gargiulo (Napoli); N. Abrescia and M. D'Abbraccio (Napoli); C. Izzo and T. Pizzella (Napoli); G. Marani Toro and G. Parruti (Pescara); D. Dionisio and A. Vivarelli (Pistoia); A. Antinori and G. Liuzzi (Roma); N. Pasquale and V. Tozzi (Roma); F. Resta and G. Buccoliero (Taranto); G. Di Perri, A. Sinicco, S. Bonora, and S. Audagnotto (Torino); F. Branz and P. Delle Foglie (Trento); P. Grossi and C. Basilico (Varese); and A. Poggio and V. Mondino (Verbania).