Highly active antiretroviral therapy (HAART) significantly reduces morbidity and mortality in HIV disease. The likelihood of HIV disease progression among patients undergoing antiretroviral treatment (ART) can be predicted by plasma HIV RNA viral load (VL) and CD4 cell count responses 3 to 9 months after initiating HAART, independent of pretreatment values.1 Although immunologic and virologic responses predict HIV disease progression, the relation among variable responses is still not well understood.2-9
Approximately 20% to 40% of HIV-infected patients initiating HAART exhibit discordant responses (immunologic response in the absence of a virologic response or the opposite).4,5,7,10-13 Virologic-only responses have been associated with baseline CD4 cell counts, whereas immunologic-only responses have been associated with VL at baseline,14 and virologic-only and immunologic-only responses have been associated with infection of multiresistant viral strains.15 Studies examining discordant immunologic and virologic responses have shown significant associations with AIDS-defining events or mortality;4,7,8 however, different definitions of immunologic or virologic response have been used, and the results have been inconsistent.
To understand the clinical consequence of discordant immunologic and virologic responses to HAART better, we examined the associations of these responses with the risk of an opportunistic infection (OI) or death and the contribution of patient characteristics. We further evaluated patient and treatment characteristics associated with differential immunologic and virologic response combinations.
The University of Alabama at Birmingham (UAB) Outpatient HIV Clinic (1917 Clinic) has been providing primary care to HIV-positive patients since 1988. The clinic began prospectively to collect information on all patients in a longitudinal, observational database in January 1992. During the study period, trained personnel extracted clinical and treatment data from medical records daily, which were entered into the database. Laboratory data were downloaded from the hospital laboratory information system directly into the database. Results of laboratory tests performed outside of UAB are entered into the database manually. The UAB database protocol was approved by the UAB Institutional Review Board for the conduct of prospective and retrospective studies.
The inclusion criteria for this study were (1) HIV-infected patients attending the 1917 clinic; (2) HAART initiation after January 1, 1995; (3) lack of ART use before treatment initiation at the clinic; (4) availability of baseline CD4 cell count and VL measurements at regimen initiation (closest value before HAART initiation); and (5) availability of at least 1 CD4 cell count and 1 plasma HIV-1 RNA measurement (VL) between 3 and 9 months after starting HAART. Patients were followed until August 31, 2004. HAART was defined as any combination of ≥3 antiretroviral (ARV) drugs (nonnucleoside reverse transcriptase inhibitor [NNRTI]-based, protease inhibitor [PI]-based, or triple-nucleoside reverse transcriptase inhibitor [NRTI]-only regimens).
Patients were categorized into 4 groups on the basis of immunologic (CD4+ or CD4−) and virologic (VL+ or VL−) responses after 3 to 9 months of HAART initiation (intent-to-treat analysis): concordant positive responders (CD4+/VL+), concordant nonresponders (CD4−/VL−), and discordant responders (CD4+/VL− or CD4−/VL+). Immunologic response was defined as an increase of ≥50 cells/μL in the CD4 count at 3 to 9 months after baseline. Virologic response was defined as achievement of an undetectable plasma HIV RNA level, with varying thresholds for VL because of different ultrasensitive assays being used over time. When multiple test results were available at 3 to 9 months of follow-up, the maximum CD4 cell count and the minimum VL measurement were used. Occurrence of an OI or death was ascertained until last contact with the patient after treatment began. Patients who did not experience death or an OI were censored at their last clinic visit date within the study period.
SAS software, version 9.1.3 (SAS Institute, Cary, NC) was used for statistical analyses. Patient characteristics were analyzed by χ2 statistics. The Kaplan-Meier method was used to compare time to development of an OI or death between therapeutic response groups starting at 6 months after therapy initiation. Cox proportional hazards model analyses were estimated, bivariate and multivariate, to quantify the association of clinical progression, starting at 6 months after therapy initiation, with short-term immunologic and virologic response to therapy. Logistic regression by stepwise model selection was used to analyze the associations between patient, clinical, and HAART regimen characteristics and 6-month therapeutic response. Independent variables included age, race (nonwhite vs. white), gender, mode of HIV acquisition (ie, injection drug use [IDU], men who have sex with men [MSM]), baseline CD4 cell count and HIV VL, and type of HAART regimen (PI-based, NNRTI-based, or NRTI-only). Furthermore, hepatitis B and C virus infection (determined by serology [ie, hepatitis B surface antigen (sAg) and hepatitis C antibody (Ab), respectively]) were evaluated as covariates.
Among 404 study patients, 50.7% were nonwhite (with 10 [2.5%] of them being non-African American) and 76.2% were male. The mean age (±SD) was 37.8 (±9.5) years, and 42 patients (10.4%) were older than 50 years of age. Regarding mode of HIV acquisition, half (50.5%) of study patients were classified as MSM and 7.4% reported a history of IDU. At baseline, the average (±SD) CD4 cell count and VL were 213 cells/μL (±260) and 5.4 (±5.8) log10 HIV RNA copies/mL, respectively. Baseline CD4 counts were <50 cells/μL in 31% of patients, 50 to 199 cells/μL in 25%, 200 to 350 cells/μL in 22%, and >350 cells/μL in 22%. Six months after initiating HAART, 289 (71.5%) patients were still on their initial regimen; the remainder had changed to another regimen. More than half (58%) of the patients were initially treated with a PI-based HAART regimen, with 28 (6.9%) patients having received boosted PI-based regimens. CD4 cell counts used for categorizing responses to HAART were obtained at a median of 194 days after HAART initiation, and VL measures were obtained at a median of 144 days after HAART initiation. Of the 404 ARV-naive patients, 70.5% (n = 285) had concordant positive immunologic and virologic responses (CD4+/VL+), 15.8% (n = 64) had an immunologic response only (CD4+/VL−), 8.7% (n = 35) had a virologic response only (CD4−/VL+), and 5.0% (n = 20) had a concordant nonresponse (CD4−/VL−) 6 months after HAART initiation. During a median follow-up time of 38 months, 35 (9%) experienced an OI and 25 (6%) died (Table 1).
As a result of missing CD4 cell counts and VL measurements at baseline and/or 6 months of follow-up, 75 individuals were excluded from the analysis because they could not be categorized within one of the response groups. Excluded individuals were more likely to be nonwhite (P = 0.003), were less likely to be male homosexual (P = 0.012), were different in type of therapy received at 6 months (P < 0.0001), and were more likely to experience an OI or death after 6 months of therapy (P = 0.0001). Kaplan-Meier analyses on all 479 individuals stratified by race (white vs. nonwhite) did not demonstrate any racial differences in the development of an OI or death (data not shown), however.
Kaplan-Meier curves (Fig. 1) summarizing time to an OI or death indicated that clinical progression differed significantly according to immunologic and virologic responses 3 to 9 months after initiating HAART (P < 0.001). The prognosis was best for the CD4+/VL+ group, worst for the CD4−/VL− group, and intermediate for the 2 discordant response groups. No statistical difference in prognosis was found between the immunologic response-only (CD4+/VL−) group and the virologic response-only (CD4−/VL+) group (P = 0.424). Statistical differences were found between CD4+/VL+ responses and the immunologic-only response (P = 0.0006) but not the virologic-only response (P = 0.190). No statistical difference was found between CD4−/VL− responses and the immunologic-only response (P = 0.061), but statistical differences were seen between CD4−/VL− responses and the virologic-only response (P = 0.026). In an adjusted multivariate Cox proportional hazards analysis (Table 2), the likelihood of an OI or death was lowest in the CD4+/VL+ group and highest in the CD4−/VL− group. Compared with the CD4+/VL+ group, the discordant response groups had an adjusted relative hazard (RH) of 2.28 (95% confidence interval [CI]: 1.3 to 4.0) and the CD4−/VL− group had an adjusted RH of 4.83 (95% CI: 2.1 to 11.1) for an OI or death. Regarding patient characteristics, being male was associated with the development of an OI or death univariately (adjusted RH = 2.20, 95% CI: 1.05 to 4.6). Baseline CD4 cell counts showed a significant inverse association with development of an OI or death (adjusted RH per 50 cells/μL = 0.93, 95% CI: 0.88 to 0.99; P = 0.022). We found no significant difference in the development of an OI or death among individuals on different initial regimen types (ie, NNRTI-based vs. PI-based and NNRTI-based vs. NRTI-only). Furthermore, the development of an OI or death was not associated with race, age, or mode of HIV acquisition when tested univariately.
Multivariate logistic regression models were used to evaluate patient, clinical, and HAART regimen characteristics associated with variable 3- to 9-month immunologic and virologic responses. Hepatitis B and C coinfection was excluded from the final multivariate models, because coinfection with either virus was not statistically associated with 6-month immunologic and virologic responses in bivariate analyses. Classification in the immunologic response-only group (CD4+/VL−) was significantly associated with nonwhite race (odds ratio [OR] = 2.83, 95% CI: 1.5 to 5.5) in comparison to the concordant positive responders (CD4+/VL+ group). Classification in the virologic response-only group (CD4−/VL+ group) was not significantly associated with any patient, treatment, or clinical characteristics in comparison to the concordant positive responders (CD4+/VL+ group). No association was observed between type of discordant response and gender, MSM, age, and type of initial HAART regimen (ie, PI-based vs. NRTI-only and NNRTI-based vs. NRTI-only). Classification in the concordant unfavorable response group (CD4−/VL−) was significantly associated with nonwhite race (OR = 6.50, 95% CI: 1.7 to 25.7) and NRTI-only regimens (PI-based OR = 0.07, 95% CI: 0.02 to 0.30; NNRTI-based OR = 0.02, 95% CI: 0.004 to 0.13).
Our study of ART-naive patients indicated that different immunologic and virologic responses 3 to 9 months after initiating HAART were associated with long-term clinical progression, irrespective of the type of treatment used (eg, NNRTI-based vs. PI-based). In contrast to previous studies, our definition of successful virologic response after treatment initiation (<50 copies/mL) was in accordance with current HIV treatment guidelines. Interpretation of previous studies has been limited because (1) subjects were treatment-experienced, (2) the thresholds of the virologic response were more liberal, (3) the definitions of immunologic response varied,4,7,8 and (4) only mortality (not OI) was considered as a long-term clinical outcome.
Moore et al7 followed 1527 treatment-naive individuals and, as in our study, found an independent association of each type of discordant response to HAART with mortality and no difference among types of discordant responses. OIs as a risk of disease progression were not included, however, and virologic response was defined as <350 copies/mL. Other studies found significant differences between the 2 types of discordant responses in subsequent outcomes.4,5 A study by Piketty et al8 showed accelerated clinical progression after both types of discordant responses, but progression was more rapid in the CD4+/VL− group than in the CD4−/VL+ group. Grabar et al4 found that CD4−/VL+ responders progressed more rapidly, however. Taken together, these studies suggest that the prognostic difference, if any, between the 2 discordant response groups is probably dependent on patient characteristics. Further study of treatment-naive patients using standardized definitions of immunologic and virologic responses and clinical progression are needed to resolve this issue.
In our study, as in others, a lower baseline CD4 cell count had significant associations with risk of an OI or death.7,16 Interestingly, male gender was also univariately associated with the risk of OI development or death. Similar to previous studies, the current study did not detect an association between the type of HAART regimen (PI-based or NNRTI-based) and the risk of an OI or death.4,5,7,8
Overall, most patients (70.5%) in our study experienced a favorable concordant response (CD4+/VL+) at 3 to 9 months after initiating HAART, whereas nearly a quarter of patients (24.5%) experienced discordant responses (15.8% CD4+/VL−, 8.7% CD4−/VL+) and a small proportion (5.0%) experienced a concordant unfavorable response (CD4−/VL−). In the evaluation of patient, ART, and clinical characteristics associated with each type of treatment response, only nonwhite race was independently related to an immunologic-only (CD4+/VL−) response relative to those with a concordant positive response (CD4+/VL+). Furthermore, nonwhite race was also associated with a concordant unfavorable response (CD4−/VL−) relative to a concordant favorable response (CD4+/VL+). Because >95% of our nonwhite population identified as African American, these findings are consistent with results from the AIDS Clinical Trial Group 5095 study, in which non-Hispanic black patients had a significantly shorter time to virologic failure.17 Inferior long-term clinical outcomes are well described in racial minorities relative to white patients with HIV infection18,19 and may, in part, be explained by a greater likelihood of short-term virologic nonresponse, as observed in this study. Further research is needed to understand better the reasons why short-term virologic failure is more common in racial minorities, including evaluation of differential adherence, toxicity, and plasma concentrations of ARV drugs relative to HIV-infected white patients.
Not surprisingly, a concordant unfavorable 3 to 9-month treatment response (CD4−/VL−) was more common in patients treated with NRTI-only regimens compared with those on PI- or NNRTI-containing regimens. These findings are consistent with clinical trial data20 and are reflected in more recent HIV ART guidelines.1
Our findings must be interpreted with respect to the limitations of this study. Because of the modest sample size, we may have had insufficient statistical power to differentiate the effects on prognosis between the 2 discordant groups. Furthermore, all cohort studies of ARV response are compromised to some degree by loss of subjects attributable to toxicity and drug intolerance as well as loss to study follow-up. The current study included only those treatment-naive patients who remained on HAART 6 months after initiating therapy, including those who had changed regimens. Individuals stopping ART altogether and those lost to clinic follow-up during the first 6 months after initiating therapy were not included in this analysis and could have biased our analyses.
In conclusion, among treatment-naive patients initiating HAART, discordant responses 3 to 9 months after initiation of therapy predict long-term clinical outcomes relative to concordant favorable responses, namely, shorter time to an OI or death. Therefore, short-term increases in CD4 cell count and achievement of an undetectable plasma HIV RNA VL within 3 to 9 months of starting therapy provide important prognostic information for treatment-naive patients initiating HAART. Future studies with larger samples are needed to clarify the differential long-term clinical outcomes among patients with discordant responses (CD4+/VL− and CD4−/VL+) further, if significant differences indeed exist.
The authors thank all the members of the University of Alabama at Birmingham HIV Outpatient Clinic Database Group for their input and assistance with the submission of this manuscript.
1. Hammer SM, Saag MS, Schechter M, et al. Treatment for adult HIV
infection: 2006 recommendations of the International AIDS Society-USA Panel. JAMA
2. Mellors JM, Munoz A, Giorgi JV, et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV
-1 Infection. Ann Intern Med
3. O'Brien WA, Hartigan PM, Daar ES, et al. Changes in plasma HIV
RNA levels and CD4+ lymphocyte counts predict both response to antiretroviral therapy and therapeutic failure. Ann Intern Med
4. Grabar S, Le Moing V, Goujard C, et al. Clinical outcome of patients with HIV
-1 infection according to immunologic and virologic response after 6 months of highly active antiretroviral therapy
. Ann Intern Med
5. Piketty C, Castiel P, Belec L, et al. Discrepant responses to triple combination antiretroviral therapy in advanced HIV
6. Spritzler J, Mildvan D, Russo A, et al. Can immune markers predict subsequent discordance between immunologic and virologic responses to antiretroviral therapy? Clin Infect Dis
7. Moore DM, Hogg RS, Yip B, et al. Discordant immunologic and virologic responses to highly active antiretroviral therapy
are associated with increased mortality and poor adherence to therapy. J Acquir Immune Defic Syndr
8. Piketty C, Weiss L, Thomas F, et al. Long-term clinical outcome of human immunodeficiency virus-infected patients with discordant immunologic and virologic responses to a protease inhibitor-containing regimen. J Infect Dis
9. Hughes MD, Daniels MJ, Fischl MA, et al. CD4 cell count as a surrogate endpoint in HIV
clinical trials: a meta-analysis of studies of the AIDS Clinical Trials Group. AIDS
10. Marimoutou C, Chene G, Mercie P, et al. Prognostic factors of combined viral load and CD4+
cell count responses under triple antiretroviral therapy, Aquitaine Cohort, 1996-1998. J Acquir Immune Defic Syndr
11. Mehta SH, Lucas G, Astemborski J, et al. Early immunologic and virologic responses to highly active antiretroviral therapy
and subsequent disease progression among HIV
-infected injection drug users. AIDS Care
12. Nicastri E, Chiesi A, Angeletti C, et al. Clinical outcome after 4 years follow-up of HIV
-seropositive subjects with incomplete virologic or immunologic response to HAART. J Med Virol
13. Tuboi SH, Brinkhof MW, Egger M, et al. Discordant responses
to potent antiretroviral treatment in previously naive HIV
-1-infected adults initiating treatment in resource-constrained countries: the Antiretroviral Therapy in Low-Income Countries (ART-LINC) collaboration. J Acquir Immune Defic Syndr
14. Renaud M, Katlama C, Mallet A, et al. Determinants of paradoxical CD4 cell reconstitution after protease inhibitor-containing antiretroviral regimen. AIDS
15. Sufka SA, Ferrari G, Gryszowka VE, et al. Prolonged CD4+ cell/virus load discordance during treatment with protease inhibitor-based highly active antiretroviral therapy
: immune response and viral control. J Infect Dis
16. Egger M, May M, Chene G, et al. Prognosis of HIV
-1-infected patients starting highly active antiretroviral therapy
: a collaborative analysis of prospective studies. Lancet
17. Gulick RM, Ribaudo HJ, Shikuma CM, et al. Three- vs four-drug antiretroviral regimens for the initial treatment of HIV
-1 infection: a randomized controlled trial. JAMA
18. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report
, 2004. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2005.
19. Whong MD, Shapiro MF, Boscardin WJ, et al. Contribution of major diseases to disparities in mortality. N Engl J Med
20. Gulick RM, Ribaudo HJ, Shikuma CM, et al. Triple-nucleoside regimens versus efavirenz-containing regimens for the initial treatment of HIV
-1 infection. N Engl J Med
Keywords:© 2008 Lippincott Williams & Wilkins, Inc.
clinical outcomes; discordant responses; highly active antiretroviral therapy; HIV