| |
Primary HIV Drug Resistance and Efficacy of First-Line Antiretroviral Therapy Guided by Resistance Testing
|
| |
| |
JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 41(5) 15 April 2006 pp 573-581
[Clinical Science]
".....We found a prevalence of primary HIV drug resistance of 11.2% in treatment-naive patients before initiation of first-line HAART. Most mutations were seen for NRTI. Only a small proportion of subjects had virus resistant against NNRTI, PI, or multiclass resistance. The most important finding of the study was comparable efficacy of first-line HAART in the group with resistance and the group with wild-type HIV.... NRTI-associated resistance was detected in 8.6%. Revertant variants were found in 4.1%. NNRTI-associated mutations were detected in 3.7%. PI-associated primary mutations and multiclass resistance each were found in 1.5%...... In summary, our study demonstrates similar virologic and immunologic efficacy of first-line HAART guided by genotypic resistance testing in patients harboring primary resistance as compared with patients with wild-type virus....."
authors: Oette, Mark MD*; Kaiser, Rolf PhD ; Daumer, Martin PhD ; Petch, Ruth MD*; Fatkenheuer, Gerd MD ; Carls, Horst MD ; Rockstroh, Jurgen Kurt MD; Schmaloer, Dirk MD ; Stechel, Jurgen MD**; Feldt, Torsten MD*; Pfister, Herbert PhD ; Haussinger, Dieter MD*; the RESINA Study Team
From the *Clinic for Gastroenterology, Hepatology, and Infectious Diseases, University Clinic Dusseldorf, Dusseldorf; Institute of Virology; Clinic I for Internal Medicine, University of Koln, Koln; Private Practice, Dusseldorf; Clinic for Internal Medicine, University of Bonn, Bonn; Westfalendamm Clinic, Dortmund; **Private Practice, Koln, Germany.
Abstract
Background: Primary HIV drug resistance has been associated with poor treatment outcome of first-line highly active antiretroviral therapy (HAART) in several trials. The aim of the study was to assess the efficacy of first-line HAART guided by resistance testing.
Methods: In a prospective multicenter study in the state of Nordrhein-Westfalen, Germany, chronically HIV-infected patients underwent genotypic resistance testing and were monitored for 48 weeks after initiation of HAART.
The project on "primary drug resistance in treatment-naive HIV-infected patients" (RESINA) is a prospective multicenter study in the state of Nordrhein-Westfalen, Germany. With about 18 million inhabitants, it is the largest state of the country and accounts for approximately 21% of HIV cases in Germany.32 Genotypic resistance testing was performed in HAART-naive HIV-1-infected patients before first application of HAART.
Results:
Primary drug resistance was found in 30 of 269 patients'entering the study between January 2001 and December 2003 [11.2%; 95% confidence interval, 7.4-14.9].
In intent-to-'treat analysis, the proportion of patients with viral load below 50 copies/mL after 24 and 48 weeks was 70.0% and 66.7%, respectively, in patients with resistance and 74.1% and 73.6%, respectively, in patients without (P = 0.66 and 0.51).
In on-treatment analysis, the proportions were 80.8% and 83.3%, respectively, in patients with resistance and 81.9% and 85.0%, respectively, in patients without (P'= 0.79 and 0.77).
These results were also valid considering a detection limit of 400 copies/mL.
Conclusions:
The prevalence of primary drug resistance was 11.2% in chronically HIV-infected patients.
HAART guided by resistance testing had similar efficacy in patients with primary drug resistance as compared with patients with wild-type virus. Based on these facts, resistance-adapted first-line HAART is suggested as routine practice.
Background
Despite documented efficacy of highly active antiretroviral therapy (HAART) of HIV infection, viral replication cannot be suppressed sufficiently in a substantial proportion of patients.1 One major reason for this is the emergence of drug-resistant variants. These tend to be common in treated HIV-infected patients as demonstrated in the HIV Health Cost and Utilization Study in the US.2 Phenotypic resistance was found in about half of the analyzed population and in almost 80% of persons with detectable viral replication while treated with HAART. Resistance against antiretroviral drugs in previously untreated HIV-positive patients, defined as primary drug resistance, is of growing relevance. It is characterized by reduced drug susceptibility in phenotypic testing or the presence of resistance-associated mutations in genotypic testing in subjects who were never exposed to antiretroviral compounds. The reason for this phenomenon is infection with resistant virus strains,3 as described for sexual,4 vertical,5 and parenteral6 path of transmission.
Primary drug resistance has been demonstrated for seroconverters in a number of countries in North America and Europe.7,8 Although many virus strains develop reversion to wild type in the absence of selective drug pressure, especially in pretreated patients,9 it has been shown that resistance may persist for years even without treatment in drug-naive and drug-exposed patients.10-12 Accordingly, mutated virus could be isolated in chronically infected HIV patients several years after seroconversion in different countries with a prevalence between 3% and 29%.7,8,13-19 Some studies found no change or a decline in prevalence of primary drug resistance,20,21 whereas others showed a continuous increase.15,22-25 The substantial variation of local epidemiology throughout Europe gives reason for continuous efforts in standardized surveillance in defined geographical areas.25,26
Antiretroviral treatment may have reduced potency in patients having acquired resistant virus. Accordingly, several studies showed reduced efficacy of HAART in patients with primary resistance compared with patients infected with wild-type HIV.7,8,24,27-30 In these investigations, 5 of which were prospective,24,27-30 the decision for a certain combination therapy was not based on genotyping. As a result, viral load decline, time to reach suppression of viral replication, or percentage of patients with viral load below the limit of detection was inferior in the group with primary resistance. Resistance testing was able to predict treatment success, depending on method, in a retrospective analysis.13 Because of these facts, revised Food and Drug Administration guidelines for trials on antiretroviral therapy in treatment-naive patients require separate analyses of the treated population according to the presence of baseline genotypic resistance.31 Currently, no prospective study has investigated the value of resistance testing for tailoring first-line HAART. Thus, the aim of this study was to evaluate the efficacy of first-line HAART guided by genotypic resistance testing by comparing treatment outcomes of patients with drug-sensitive HIV to outcomes of patients with primary resistance.
RESULTS
Between January 2001 and December 2003, patients were prospectively enrolled into the RESINA study. The flow chart of the investigation is demonstrated in Figure 1. Informed consent was obtained from 372 individuals; of these, 346 met inclusion criteria and entered the study. Three individuals withdrew consent before resistance testing. In 269 patients, antiretroviral therapy was initiated; the characteristics of these patients are shown in Table 1. Participants were predominantly middle-aged, male, white Germans with homosexual transmission harboring HIV subtype B. Mean CD4 cell count and viral load as well as clinical staging indicated moderate to advanced stage of the disease. In the observational period, 8 patients died and 12 patients were lost during follow-up.
In 30 of 269 patients, primary HIV drug resistance was identified [11.2%; 95% confidence interval (CI), 7.4-14.9]. NRTI-associated resistance was detected in 8.6% (95% CI, 5.2-11.9). Revertant variants were found in 4.1% (95% CI, 1.7-6.5). NNRTI-associated mutations were detected in 3.7% (95% CI, 1.5-6.0). PI-associated primary mutations and multiclass resistance each were found in 1.5% (95% CI, 0.04-2.9). To test for differences between the subgroups with and without resistance, the distribution of baseline characteristics was compared. The result is demonstrated in Table 1. Univariate analysis identified the following parameters as being significantly associated to the diagnosis of primary resistance: shorter duration of HIV diagnosis, HIV subtype B vs other subtypes, and white ethnicity vs other ethnic origins.
Antiretroviral therapy was prescribed with a substantial variety of combinations. A potent HAART consisting of at least 2 classes of antiretroviral compounds was administered in 220 of 239 patients without resistance (92%) and 27 of 30 patients with resistance (90%). Table 2 displays the characteristics and course of treatment for patients harboring primary resistant virus. Applied regimens complied with the treatment recommendation in all but 3 patients (patients 12, 177, and 291). In 2 of these individuals, the combination was changed according to the result of primary resistance testing after 1 and 3 months (patients 177 and 291). At the end of follow-up, these'individuals had a viral load below the detection limit. Virologic failure occurred in 4 patients: those'with the mutations T215S, V118I, 2-class resistance (T215D, L90M), or 3-class resistance (T215C/S, K103N, V82A, L90M). Five patients were lost to follow-up, and 1'patient died during the'observation period. In all other patients, HAART succeeded in suppressing detectable viral replication.
Virologic failure, defined as not reaching a detection limit of 50 copies/mL at 48 weeks, was recorded in 4 of the patients with resistance (13.3%) and in 31 of the patients without (13.0%). Development of viral load after 24 and 48 weeks is shown in Table 3 and Figure 2 (for the detection limit of 50 copies/mL). Results of virologic outcome were not significantly different between the groups with and without resistance. This finding was valid in intent-to-treat and on-treatment analysis at 24 and 48 weeks considering both 50 and 400 copies/mL as limit of detection.
To test for confounding variables influencing treatment outcome, the groups with and without virologically effective therapy were evaluated in a multivariate analysis. After adjustment for the parameters age, sex, ethnic origin, disease stage of AIDS, CD4 cell count, and viral load, detection of primary drug resistance was not associated with treatment outcome. This could be shown for on-treatment analysis considering a detection limit of 50 copies/mL after 24 weeks [P = 0.97; odds ratio (OR) 0.98; 95% CI, 0.3-2.8] and 48 weeks (P = 0.25; OR 1.75; 95% CI, 0.3-2.9). Moreover, a similar result was found considering a detection limit of 400 copies/mL for the time points of 24 weeks (P = 0.88; OR 1.29; 95% CI, 0.1-3.5) and 48 weeks (P = 0.80; OR 1.70; 95% CI, 0.1-2.2).
The values of CD4 cell counts were as follows: mean ± SD at baseline were 185 ± 157 cells/mL in the subgroup with primary resistance and 175 ± 160 cells/mL in the subgroup without (P = 0.33). After 24 weeks, CD4 cell count was 376 ± 166 cells/mL in the group with resistance and 310 ± 193 cells/mL in the group without (P = 0.14). After 48 weeks, CD4 cell count was 450 ± 238 cells/mL in the group with resistance and 361 ± 199 cells/mL in the group without (P = 0.13).
DISCUSSION
The RESINA project is an ongoing prospective multicenter study on epidemiology of primary HIV drug resistance and treatment outcome in Nordrhein-Westfalen, Germany. Between 2001 and 2003, we studied 269 treatment-naive, chronically HIV-infected patients starting HAART who were followed for 48 weeks in an open cohort in 33 centers. We found a prevalence of primary HIV drug resistance of 11.2% in treatment-naive patients before initiation of first-line HAART. Most mutations were seen for NRTI. Only a small proportion of subjects had virus resistant against NNRTI, PI, or multiclass resistance. The most important finding of the study was comparable efficacy of first-line HAART in the group with resistance and the group with wild-type HIV.
The epidemiologic data are consistent with results from different European countries. For example, the prevalence of'resistance against NRTI in chronically infected patients is 8.6% in this study, 8.0% in France,21 7.8% in Italy,16 and 7.5% in several European countries.17 Differences in epidemiology may be because of varying distribution of transmission route, subtype, or other patients characteristics as well as study design. The latter can be demonstrated, for example, with the inclusion of the V118I mutation into mutation scores. Although classified as resistance-associated mutation,33 it was shown not to be associated with reduced treatment success in a retrospective study.36 We regarded this mutation as resistance of significance in our data set because of the classification of the International AIDS Society panel. Furthermore, the outcome of patients defined as having resistance would have been artificially improved by its omission from the list, so a conservative decision was made to include it.
Subgroups of participants that were overrepresented in the population carrying resistant strains were patients with shorter duration of HIV diagnosis, patients with subtype B, and whites. These seem to be at risk for acquiring primary drug resistance. Up to now, only the Combined Analysis of Resistance Transmission Over Time of Chronically and Acute Infected HIV Patients in Europe study reported on risk groups, with a prevalence of baseline resistance of 11.3% in subjects carrying HIV subtype B compared with 3.3% with non-B subtypes (P < 0.001),17 matching our results. Altogether, the presented epidemiologic data underscore the importance of further surveillance of primary drug resistance with better standardization of methods.
Reports about inferior treatment outcome in patients harboring primary drug resistance have been presented in a growing number of studies.7,8,24,27-30 The first major investigation on this phenomenon was a retrospective study that showed an increased time to viral suppression and a reduced time to virologic rebound in the group with resistance.8 Further studies found a higher rate of virologic failure in the participants with baseline resistance.27-30 In one study, resistant virus before treatment was seen in 34% to 38% of patients with viral failure, and multivariate analysis identified baseline resistance as predictive for viral rebound.27,37 Another investigation found a suppression of viral replication in 27% and 63%, respectively, of patients with resistance and 45% and 82%, respectively, in patients with wild-type virus with P values of 0.05 and 0.02, depending on the applied detection limit.29 Because of varying study settings, the discussed findings cannot be easily translated into our data set. However, they point to a similar direction, raising serious concerns that first-line HAART administered independent of results of resistance testing may lead to a less effective virologic outcome. As discussed earlier, primary drug resistance has to be analyzed separately in trials on HAART in treatment-naive patients according to Food and Drug Administration regulations. Consequently, first-line HAART should be guided by resistance testing. We present data on the outcome of this treatment strategy, which is the first prospective trial on efficacy on resistance-adapted first-line HAART that suggests similar virologic outcome in both arms.
The treatment course of 269 patients was followed for 48 weeks. As shown in Table 2, there was a substantial variety of resistance mutations and applied regimens in the individuals with primary drug resistance. HAART was highly effective in the presented study, as shown in Tables 2 and 3 and Figure 2. In general, the proportion of patients with a viral load below the detection limit was 70% to 80% in intent-to-treat analysis and around 82% to 93% in on-treatment analysis, depending on the detection limit of viral load. More importantly, we were unable to find a significant difference of outcome between the group with resistance and the group with wild-type virus. This fact is valid for both time points (24 and 48 weeks), for intent-to-treat as'well as on-treatment analysis, and for both the'detection'limits of 50 and 400 copies/mL. The result was confirmed by multivariate analysis, which did not determine resistance as a factor of significant influence on virologic success after adjustment for age, sex, ethnic origin, stage of AIDS, CD4 cell count, and viral load. Virologic failure was seen in a comparable proportion of patients. Taken together, virologic efficacy of first-line HAART was similar in patients with wild-type virus and those with primary resistance.
The patients experienced substantial immunologic benefit from HAART. CD4 cell count rose from about 180 to more than 360 cells/mL after 48 weeks. The difference at baseline was small between the groups with and without resistance, but there was a more pronounced increase in the group with resistance compared with the group without that did not reach statistical significance. The different immunologic response may be because of the different distribution of baseline characteristics in the 2'groups, as shown in Table 1. Moreover, the knowledge that resistant virus was found may have resulted in a higher adherence level in the group with baseline resistance.
In summary, our study demonstrates similar virologic and immunologic efficacy of first-line HAART guided by genotypic resistance testing in patients harboring primary resistance as compared with patients with wild-type virus. This finding is in contrast with several trials reporting inferior efficacy of HAART in patients with primary'resistance who received a treatment combination independent of results of genotyping. Our findings contribute to the'growing body of literature on the value of genotypic resistance testing in HAART-naive patients before application of first-line'therapy.
Most HIV specialists of the largest state in Germany worked together for this prospective study. By covering a defined geographical region, selection bias is likely to be low in our investigation. The number of participants allows for an estimate of primary HIV drug resistance in chronically infected HIV-positive patients with a narrow confidence interval, which is an important parameter that has been requested by several scientists.25,26 One limitation of the trial is the open design of treatment evaluation, which may result in a bias of efficacy data because the patients were treated in clinical routine within a prospective cohort. As a consequence, a randomized controlled trial may help determine the true value of resistance-adapted first-line HAART. However, the documented negative impact of primary resistance on treatment outcome in retrospective studies raises ethical concerns about this approach. Moreover, the approximate prevalence of primary resistance of about 11% would require an enormous study population to prove equivalence of treatment strategies. Another limitation of this study is the heterogeneity of the population investigated. As shown in Table 2, the spectrum of mutations and administered HAART regimens was diverse; thus, it is difficult to compare with other trials. A more defined study population with specific genotypic results and treatment combinations would result in a reproducible data set. However, because of the prevalence of 11%, this limitation is difficult to deal with because a very large study population would again be required to produce a subgroup with similar characteristics. Another bias is the lack of data on therapeutic drug monitoring and adherence levels, which have been included into the protocol later. Finally, the'limitations of the method of genotypic resistance testing have'to be taken into account. Because only the predominant quasi-species can be evaluated with the applied method, minor variants may not have been detected. This finding was described recently38 and leads to the concern that the determined prevalence may be an underestimation. Despite these limitations, the prospective application of genotyping for combining HAART compounds is a new approach to improve first-line therapy. In doing this, our study tries to enhance the knowledge about the clinical relevance of primary resistance and may help define rational application of genotypic resistance testing. Further investigations on epidemiology and impact on therapy for primary HIV drug resistance are needed.
Primary HIV drug resistance is a problem of significant public health impact. Based on our observations, 3'conclusions should be drawn. First, the determined prevalence of primary resistance in this study region and in data from other countries makes continuing surveillance of this phenomenon necessary. Second, because of good efficacy of first-line HAART guided by resistance testing especially in the population with primary resistance, this method should also be considered in clinical routine. More data on this topic have to be collected. Third, to prevent transmission of resistant virus, all efforts should be made to enhance effectiveness of HAART regimens.
METHODS
The project on "primary drug resistance in treatment-naive HIV-infected patients" (RESINA) is a prospective multicenter study in the state of Nordrhein-Westfalen, Germany. With about 18 million inhabitants, it is the largest state of the country and accounts for approximately 21% of HIV cases in Germany.32 Genotypic resistance testing was performed in HAART-naive HIV-1-infected patients before first application of HAART. The study received approval by local and collaborating institutional review boards. Across the state, 42 centers specialized in the treatment of HIV, and patients with AIDS were identified and initially asked for participation. Thirty-four centers consisting of 5 hospitals, 5 outpatient units of university clinics, 5 outpatient units of regional hospitals, and 23 private practices collaborated in the project. A substantial part of the HIV-infected subgroup as defined by the study protocol has been reached by covering most treatment centers of the state. Thus, the study population may be regarded as representative for the state of Nordrhein-Westfalen, Germany.
Inclusion criteria were documented HIV-1 infection, eligibility for application of HAART, and informed consent. Exclusion criteria were previous intake of antiretroviral drugs (as determined by self-report) and unwillingness to participate. The following baseline parameters were documented: age, sex, duration of HIV diagnosis, Centers for Disease Control and Prevention (CDC) stage of disease, CD4 cell count, viral load, HIV subtype (subdivision of group M of HIV-1), ethnic origin, and route of HIV transmission.
Genotypic resistance testing was performed at the Institute of Virology, University of Koln, Germany, as previously described.18 Interpretation and classification of clinical significance of novel variants were performed according to current guidelines.33 Resistance-associated mutations were subdivided by their biological relevance: mutations resulting in resistance against nucleoside reverse transcriptase inhibitors (NRTI), nonnucleoside reverse transcriptase inhibitors (NNRTI), protease inhibitors (PI), and multiclass resistance in cases with involvement of at least 2 of these groups. Furthermore, mutations indicating revertant variants of resistant strains were regarded as significant. These isolates (T215C/E/L/S) have been associated with reduced virological efficacy of HAART, although phenotypically sensitive.28,34
Results of resistance testing were made available to assist the treatment decisions of the participating physicians. Interpretation of resistance and recommendations for therapy were given by virologists who are experts in the field with the help of the bioinformatics interpretation system geno2pheno.35 For each patient, an individualized report was issued, which stated whether resistant variants were detected and which drugs should be avoided in case of resistance. The treating physician made the decision about the drug combination. Follow-up was evaluated in an open cohort in clinical routine prospectively. No centralized laboratory measurements were performed. At baseline and time points of 24 and 48 weeks after initiation of HAART, CD4 cell count and viral load were recorded. The status of being below the limit of detection for viral load was calculated for 50 and 400 copies/mL because several study centers used 400 copies/mL as routine procedure.
The statistical analysis was performed with SPSS, release 12.0. Univariate comparisons were applied using the Wilcoxon rank sum test or two-sided Fisher exact test where appropriate. Multivariate analyses were done with the help of logistic regression models. P values less than 0.05 were considered significant; no adjustment for multiple testing was applied.
|
|
| |
| |
|
|
|
