There are no definitive data from randomized controlled trials with clinical end points to inform when to start highly active antiretroviral therapy (HAART) in asymptomatic HIV-infected patients.1 Shifts in clinical practice toward both early and delayed initiation of HAART have been guided by observational cohort data and expert opinion1; the current trend is toward earlier initiation.1,2 This current trend is driven by studies demonstrating lower morbidity and mortality and fewer adverse events associated with the initiation of sustained HAART at higher CD4 cell counts3-7 and by the advent of more potent, more convenient, and better-tolerated antiretroviral (ARV) agents.
Even in the current treatment environment, there remain drawbacks to initiating HAART earlier. The financial cost of treatment is a valid consideration in both developed and developing countries. However, even if cost were not a factor and the clinical benefits of earlier HAART were validated in a randomized clinical trial, some clinicians remain concerned that earlier treatment at higher CD4 cell counts may increase the risk of developing treatment-limiting ARV resistance and possibly exhaust treatment options that might have been preserved had treatment been delayed. Data regarding this potential problem are lacking.
To assess the risk of developing HIV drug resistance on an initial HAART regimen based on the CD4 cell count at which HAART was initiated, we analyzed data collected from a large prospective cohort of HIV-infected patients receiving care in 9 US HIV clinics participating in the HIV Outpatient Study (HOPS). We describe herein the frequency of HIV drug resistance among patients who experienced virologic failure after successful virologic suppression, according to CD4 cell count at which they initiated HAART.
The HOPS is an ongoing, prospective, observational cohort study of HIV-infected patients receiving care in 9 participating HIV clinics (1 public, 4 university, and 4 private) in 8 US cities (Chicago, IL; Denver, CO; Long Island, NY; Oakland/San Leandro, CA; Philadelphia, PA; Tampa, FL; and Washington, DC) since March 1993. Patient data, including sociodemographic characteristics, symptoms and signs, diagnoses, treatments, and laboratory values, are abstracted from medical charts and entered into an electronic database (Clinical Practice Analyst; Cerner Corporation, Vienna, VA) by trained staff. These data are reviewed for quality and analyzed centrally. To date, the HOPS has collected information on more than 8400 patients in approximately 300,000 clinical encounters. The HOPS protocol is reviewed annually and approved by the Centers for Disease Control and Prevention and each local site's institutional review board since its inception. Informed consent is obtained from all patients. The study protocol conforms to the guidelines of the US Department of Health and Human Services for the protection of human subjects in research. The present analysis is based on the HOPS database through September 30, 2006.
We included in the analyses all HOPS patients who met the following criteria: entered HOPS either ARV naive or on their first HAART regimen in 1999 or later, started ARV therapy with a HAART regimen, had at least 2 subsequent HOPS visits, and had a complete ARV drug history. We defined initial virologic suppression as a single quantitative HIV RNA (viral load) of <1000 copies per milliliter after starting HAART and virologic failure as a single viral load of >1000 copies per milliliter after initial virologic suppression. Among patients who failed their initial suppressive HAART regimen and who had an HIV genotype resistance test performed at or after virologic failure, we assessed the frequency of major resistance mutations based on the 2006 guidelines of the International AIDS Society-USA8 at the time of this test, stratified by the CD4 cell count at which HAART had been initiated (0-199, 200-349, and ≥350 cells/mm3).
Descriptive and bivariate analyses were conducted with SAS version 9.1 (SAS institute, Cary, NC). We used the Yates corrected χ2 test or the Fisher exact test for statistical comparisons of categorical variables and the Wilcoxon rank sum test for continuous variables. Two-sided P values for trend were calculated with the Cochran-Armitage exact test.
Of 8224 patients who participated in the HOPS through September 30, 2006, 760 patients met inclusion criteria for analysis. Six hundred and eighty-three of these patients (90%) achieved initial virologic suppression on first HAART, of whom 243 subsequently met criteria for virologic failure during observation time; 78 of these patients (32%) had a genotype resistance test done at or after virologic failure (Fig. 1). The percentages of patients with virologic suppression who experienced virologic failure were similar (P = 0.41) within each of these 3 CD4 cell count strata at which HAART was initiated (Fig. 1). Patients who started HAART at lower CD4 cell counts had a shorter time interval between suppression and failure (P = 0.026). In addition, the proportion of patients with virologic failure who had a subsequent HIV genotype resistance testing done at or after this failure was the greatest among persons in the lowest CD4 cell count stratum (P = 0.043). Compared with those with virologic failure without a genotype test, those with a genotype test were more likely to be black and have an AIDS diagnosis; they were less likely to have private insurance (data not shown). Median Log HIV RNA proximal to the time of the genotype test were 4.12, 3.98, and 3.92 for the 0-199, 200-349, and ≥350 cells per cubic millimeter strata, respectively (P = 0.34).
Baseline characteristics of the 78 patients with virologic failure and a genotype test according to their CD4 cell count at HAART initiation are shown in Table 1. At HAART initiation, 46 of the 78 patients had a CD4 cell count between 0-199 cells per cubic millimeter, 14 had a CD4 cell count between 200-349 cells per cubic millimeter, and 18 had a CD4 cell count >350 cells per cubic millimeter. Patients in the lowest CD4 cell count stratum were more likely to have high-risk heterosexual behavior as an HIV risk factor and have a higher HIV RNA at initiation of HAART.
The frequency of all HIV resistance mutations at virologic failure was 50% among patients who started HAART with CD4 counts of 0-199 cells per cubic millimeter (23 of 46) or 200-349 cells per cubic millimeter (7 of 14) and 22% (4 of 18) among patients who started HAART with CD4 cell counts >350 cells per cubic millimeter (P = 0.062) (Fig. 2). The frequency of nucleoside reverse transcriptase inhibitor (NRTI)-associated genotype mutations was 48% (22 of 46), 31% (4 of 13), and 11% (2 of 18) among patients who started NRTI-containing HAART within each of the same CD4 cell count strata, respectively (P = 0.005). We observed a similar difference across the same CD4 cell count strata for nonnucleoside reverse transcriptase inhibitor (NNRTI)-associated mutations among patients who started NNRTI-containing HAART (P = 0.040) and for protease inhibitor (PI)-associated mutations among persons who started PI-containing HAART (P = 0.063) (Fig. 2), a third of whom were on a ritonavir-boosted PI with a similar prevalence of resistance to the overall PI group (data not shown).
Among HOPS participants who achieved virologic suppression on their initial HAART regimen, and subsequently had virologic failure and underwent genotype resistance testing, we observed a higher frequency at virologic failure of ARV resistance mutations among patients who started HAART at a lower CD4 cell counts compared with patients who started at higher CD4 cell counts. We observed this trend for the detection of any resistance and for detection of resistance for each of the 3 major ARV drug classes; the trend was statistically significant for NRTI and NNRTI resistance. Furthermore, although the fraction of patients who were observed with virologic failure was not statistically different across the 3 CD4 cell count strata, the median time to failure was longer for patients who started HAART at higher CD4 cell counts (Fig. 1). Thus, patients who started HAART at a higher CD4 cell count experienced the greatest durability of their initial HAART regimen and had lower frequency of resistance mutations at virologic failure despite the fact that their virus received the longest periods of exposure to the selective pressures of ARVs that can induce resistance mutations. Although we had only a modest number of patients with requisite data for this analysis, our preliminary findings should be reassuring to clinicians and alleviate the concerns for accelerating acquisition of treatment-limiting resistance by starting HAART at CD4 cell counts ≥350 cells per cubic millimeter. Our findings, although focusing only on resistance after failing a first HAART regimen, are consistent with those of Phillips et al9 who found that the rate of extensive triple class failure was lower among patients who started HAART at higher CD4 cell counts (for each 100 cell higher CD4 cell count at treatment initiation, hazard ratio = 0.68, P < 0.001) in the UK Collaborative HIV Cohort.
The basis for the observed differences in the frequency of resistance mutations may be biological, behavioral, or both. Although we included in our analyses only the patients who achieved initial virologic suppression to reduce the potential for confounding by level of adherence to HAART (ie, nonadherent patients may be more likely to start HAART at lower CD4 counts and also be more likely to develop resistance and fail virologically), we did not have systematic data on adherence levels to examine and control for this potential confounding. Additionally, our data were collected in the course of routine clinical care for HIV infection; genotype resistance testing was not performed at scheduled intervals and was ordered at the discretion of the treating physician. Among persons who did fail, genotype resistance testing was only performed for a minority of patients (78 of 243 or 32% overall). Furthermore, the persons who started HAART in the lowest CD4 count stratum and failed virologically were most likely to get a genotype test (Fig. 1); it is not clear whether this disparity in testing rates could bias our findings toward overestimating or underestimating the differences in frequency of resistance mutations at first HAART failure by the CD4 cell count at which HAART was started. There was also the largest time interval between virologic failure and the genotype test in the lowest stratum; however, the viral load at the time of genotype testing was similar across all 3 strata.
Finally, although it is reasonable to assume that the patients who did not fail virologically had no clinically significant (ie, associated with reduced virologic response) resistance mutations (an assumption which would have been difficult to verify, even if we had appropriate blood specimens, because of the technical limitations of performing genotype resistance tests among patients with low viral loads), we do not have data for these analyses. Also, we do not know what proportion of patients may have harbored preexisting drug-resistant HIV before starting ARV therapy; however, published data10 suggest that the number of ARV-naive persons with detectable resistance mutations on genotype testing remains low (5%-15%) and thus would likely have been infrequent in our cohort. Due to scant testing in the HOPS to account for resistance mutations before start of HAART at the time of this analysis, and due to variable frequency of genotype resistance testing among the patients we followed, we did not evaluate the incidence of resistance mutations but rather presented frequency of resistance mutations at first HAART failure.
Ideally, the prevalence and incidence of clinically relevant resistance mutations should be evaluated in randomized controlled clinical trials with long-term follow-up. Many clinical trials are limited by small numbers of participants and select patients without preexisting comorbidities who were most likely to adhere to their HAART regimens, thus limiting the generalizability of their findings. Follow-up in many clinical trials is also on the order of weeks rather than years; the impact of acquired drug resistance on clinical outcomes may not become evident until the second or third HAART regimen years hence; thus large multisite observational HIV cohorts may remain the best option for assessment of the clinical consequences of resistance mutations.
Randomized controlled clinical trials to evaluate the effectiveness and adverse events associated with starting HAART at different CD4 cell counts are currently being considered. We urge investigators planning these studies to include assessments of the incidence of and risk factors for developing detectable and clinically relevant HIV resistance mutations. Data from our relatively small observational study suggest no gross evidence of harm in terms of increased risk of developing resistance mutations on first HAART regimen among patients initiating therapy at CD4 cell counts ≥350 cells per cubic millimeter.
1. Panel on Antiretroviral Guidelines for Adult and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents
. Department of Health and Human Services; January 29, 2008. Available at: http://aidsinfo.nih.gov
. Accessed May 21, 2009.
2. Panel on Antiretroviral Guidelines for Adult and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents
. Department of Health and Human Services; October 16, 2006. Available at: http://aidsinfo.nih.gov
. Accessed May 21, 2009.
3. Kitahata MM, Gange SJ, Abraham AG, et al. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med
4. When To Start Consortium. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet
5. Palella FJ Jr, Baker RK, Moorman AC, et al. Mortality in the highly active antiretroviral therapy era: changing causes of death and disease in the HIV outpatient study. J Acquir Immune Defic Syndr
6. May M, Sterne JA, Sabin C, et al. Prognosis of HIV-1-infected patients up to 5 years after initiation of HAART: collaborative analysis of prospective studies. AIDS
7. Lichtenstein KA, Armon C, Buchacz K, et al. Initiation of antiretroviral therapy at CD4 cell counts >/=350 cells/mm3
does not increase incidence or risk of peripheral neuropathy, anemia, or renal insufficiency. J Acquir Immune Defic Syndr
8. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug resistance mutations in HIV-1: Fall 2006. Top HIV Med
9. Phillips A, Leen C, Wilson A, et al. Risk of extensive virological failure to the three original antiretroviral drug classes over long-term follow-up from the start of therapy in patients with HIV infection: an observational cohort study. Lancet
10. Wheeler W, Mahle K, Bodnar U, et al. Antiretroviral drug-resistance mutations and subtypes in drug-naive persons newly diagnosed with HIV-1 infection, US, March 2003 to October 2006. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 648.
APPENDIX I: THE HOPS INVESTIGATORS
The HOPS Investigators include the following investigators and sites: J.T.B. K.B., Tony Tong, and Anne Moorman, Division of HIV/AIDS Prevention, National Center for HIV, STD, and TB Prevention (NCHSTP), Centers for Disease Control and Prevention, Atlanta, GA; Kathleen C. Wood, Rose K. Baker, James T. Richardson, Darlene Hankerson, and C.A., Cerner Corporation, Vienna, VA; Frank J. Palella, Joan S. Chmiel, Aditya Chawla, and Onyinye Enyia, Feinberg School of Medicine, Northwestern University, Chicago, IL; Kenneth A. Lichtenstein and Cheryl Stewart, National Jewish Medical and Research Center Denver, CO; John Hammer, Benjamin Young, Kenneth S. Greenberg, Barbara Widick, and Joslyn D. Axinn, Rose Medical Center, Denver, CO; Bienvenido G. Yangco and Kalliope Halkias, Infectious Disease Research Institute, Tampa, FL; Douglas J. Ward and Charles A. Fiorentino, Dupont Circle Physicians Group, Washington, DC; Jack Fuhrer, Linda Ording-Bauer, Rita Kelly, and Jane Esteves, State University of New York, Stony Brook, NY; Ellen M. Tedaldi, Ramona A. Christian and Faye Ruley, Atiya Nimmons, Temple University School of Medicine, Philadelphia, PA; and Richard M. Novak, J.U. and Andrea Wendrow, University of Illinois at Chicago, Chicago, IL.