Recent understanding of the pathogenesis of human immunodeficiency virus (HIV) infection has important implications for clinical and laboratory monitoring of HIV disease progression. As the disease is characterized by a virus-mediated immune suppression, viral load and CD4 cells are meaningful clinical markers to monitor during the course of HIV infection. In addition, these and other surrogate markers may be useful in evaluating agents to treat HIV disease (1,2). Monitoring therapy outcomes with meaningful clinical markers may prove especially useful in asymptomatic individuals in whom the clinical course of the disease may be as long as 10 years or more, making large clinical end point studies technically difficult.
The absolute CD4 count has been the primary surrogate marker in HIV clinical trials, as this marker has demonstrated a correlation with the clinical stage of HIV infection (3). Although of great clinical utility, absolute CD4 counts may predict only 40% of the clinical benefit of zidovudine (AZT), as estimated in some studies (4). In addition, spurious increases in absolute CD4 counts have been reported in HIV-infected individuals who are co-infected with human T-lymphotropic virus type 1 (HTLV-1) (5). Few studies have examined the percentage of CD4 cells as a potential surrogate marker (6).
Early clinical events, such as oral hairy leukoplakia and oral candidiasis, are common clinical manifestations of HIV infection and, in some studies, have been shown to be predictive of clinical outcome (7). As clinicians must evaluate the effect of potential therapies on the early, asymptomatic phase of the disease, we have examined the relationship between two potential markers of disease progression: CD4 percentage and early clinical events.
We examined the relationship between the rate of change in the percentage of CD4 cells over time and early clinical events in 103 asymptomatic, HIV-1-seropositive patients who were randomized to receive either 100 μg of HIV-1 immunogen, an inactivated, gp120-depleted HIV-1 virus in incomplete Freund's adjuvant (IFA), or IFA alone at 0, 3, and 6 months. The results of this trial revealed a beneficial effect of the HIV-1 immunogen on surrogate markers of peripheral blood mononuclear cells (PBMC) proviral DNA and CD4 percentage, which are described elsewhere (8).
The data from the present study arose from a parent study designed as a randomized, double-blind, adjuvant-controlled trial for 1 year, with an unblinded no-treatment follow-up phase. The primary endpoint of the trial was viral burden as measured by polymerase chain reaction (PCR) DNA or PBMCs during the first year of the trial. Subjects were randomized at nine medical centers and were followed periodically. As this study focused primarily on surrogate markers; clinical progression was added prior to unblinding as a secondary parameter, as defined by the 1993 revised classification system for HIV infection and expanded AIDS surveillance case definitions for adolescents and adults by the Centers for Disease Control (CDC) (9). The use of AZT was permitted, at the discretion of each investigator following Food and Drug Administration (FDA) approval for expanded indications in March 1990. Informed consent was obtained from each subject, and approval from an Institutional Review Board was granted at each center prior to the start of the study.
One hundred and three asymptomatic, HIV-1-seropositive subjects were enrolled at nine investigational sites (U.S. Naval Medical Center-San Diego, Vermont Cancer Center, Arizona Cancer Center, Northwestern University Hospital, North Shore University Hospital, Long Island Jewish Medical Center, Graduate Hospital, USC County Hospital, and San Francisco General Hospital). Inclusion criteria were a positive HIV-1 enzyme-linked immunosorbent assay (ELISA), a positive HIV-1 Western blot, two positive viral cultures, and two positive PCRs for HIV-1 DNA. CD4 cell counts were required to be >550/mm3 and to comprise >20% of the total lymphocytes in the peripheral blood. Subjects were required to have a delayed type hypersensitivity response to two or more of the seven antigens in the Merieux cell-mediated immunity (CMI) multitest panel and to demonstrate an antibody response to recall (tetanus toxoid) and neo (inactivated rabies vaccine) antigens.
Measurements of the percentage of CD4 cells were performed serially in subjects for a mean of 538 days (range 176-694) in the cohort. Three subjects in the IFA control group and four subjects in the HIV-1 immunogen group were lost to clinical follow-up during the second year when inoculations were no longer given. Subjects lost to follow-up were similar to the original cohort in baseline demographic characteristics. The baseline (mean ± S.E.) percentage of CD4 cells in the cohort was 31.5 (± 0.74). The baseline (mean ± S.E.) absolute CD4 count in the cohort was 656.1 (± 17.3).
Measurements of the percentage of CD4 cells were performed at each study site by a qualified flow cytometry laboratory every 8 weeks during the first year and every 24 weeks during the second year. Absolute CD4 counts were calculated from both lymphocyte counts performed at a central laboratory and the percentage of CD4 cells.
Clinical status was evaluated monthly during the first year and every 24 weeks during the second year. Complete clinical evaluations were performed at baseline, and at weeks 24 and 52 during the first year, and every 24 weeks during the second year. A first clinical event on study was defined as progression from baseline to a confirmed new CDC category (to category B or C, or to AIDS-related death). Clinical events were documented on case report forms during the study and verified with investigators retrospectively.
The rate of change (slope) was calculated by least squares regression during the first year or until a clinical event, which-ever occurred first, and is described elsewhere (10). One subject was excluded from the slope analysis because of having only one time point. When examining the relationship between CD4 cells and early clinical events, data were combined for both the immunogen and IFA control subjects. The mean percentage of CD4 cells was determined for all subjects during the study, as well as prior to and after an early clinical event (11). Using the two sample Wilcoxon Rank Sum test, the median change in the slope of CD4 percentage of absolute CD4 count was compared between those subjects who had an early clinical event and those who did not (12). A Cox regression analysis examined the relationship between the slope values and the risk of having a clinical event using the slope values as an ordered categorical variable (12). The four categories were defined by the <25th (< -4), 25th (-4 to -0.5), 50th (-0,5 to 2.5), and >75th (>2.5) quartiles, Confidence intervals (CI) were determined for the Wilcoxon Rank Sum and Cox Regression tests. All p values presented are two-tailed.
The percentage of CD4 cells in the peripheral blood was evaluated serially over time in both groups. The mean decline over time (slopes) in the percentage of CD4 cells and absolute CD4 counts for the entire cohort was 1.47 per year and 50.5 cells per year, respectively.
There was a total of 19 clinical events, of which 16 were CDC category B events. Twelve subjects had oral hairy leukoplakia, three had oral candidiasis, one had Kaposi's sarcoma, one had chronic herpes, one had peripheral neuropathy, and one had HIV-related dementia; 83 remained asymptomatic. Subjects were stratified by whether or not they had an early clinical event. The mean decline in the CD4 percentage was 5.03 ± 2.04 (SE) per year in those subjects who had an early clinical event compared with 0.66 ± 0.57 per year in the subjects who remained asymptomatic. There was a greater median decline over tine in the percentage of CD4 cells in subjects who had a first clinical event (4.05) than in those who did not have a clinical event (O) (p = 0.054, CI = -7.79; +0.02). The mean decline in absolute CD4 cells was 90.3 cells per year for those subjects experiencing an event, compared with 41.3 cells per year for those without an event. The median decline in absolute CD4 cells for those who had an event was 72.0 cells per year compared with 39.7 cells per year for those who had no events (p = 0.40). In the subjects who became symptomatic, the greatest decline in CD4 percent occurred within 10 weeks prior to a clinical event (11% mean decrease from baseline). The mean CD4 percentage in subjects who had an early clinical event and in those who remained asymptomatic is graphed in Fig. 1.
To determine if changes in the CD4 percentage over time were related to the time to a first clinical event, subjects stratified by slope categories were plotted on a survival curve. The relationship between the time to a clinical event and slope categories is displayed graphically in Fig. 2. Subjects with the most negative CD4 percentage slope (< -4) progressed to an early clinical event more rapidly than the other groups. There was approximately a fivefold risk of having a clinical event in the group with the most negative CD4 percent slope (< -4) compared with the next stratified group (-4 to -0.5) (p = 0.045, CI = 0.04; 0.97).
The present report describes a relationship between the decline in the percentage of CD4 cells and early clinical events during follow-up of a cohort of asymptomatic, HIV-1-seropositive individuals who were randomized to receive three inoculations of a gp120-depleted, inactivated HIV-1 immunogen or IFA. Subjects who had an early clinical event were more likely to have a greater decline in CD4 percentage than those who did not have a clinical event. Subjects who had the greatest decline in CD4 percentage progressed more rapidly to a clinical event and had a fivefold risk of having an early clinical event. A similar trend was demonstrated for absolute CD4 cell counts but was not found to be statistically significant.
The use of absolute CD4 counts as a marker in clinical trials is made difficult by the inherent large intra-assay variability. Absolute CD4 counts comprise both the absolute lymphocyte count and the percentage of CD4 cells. Greater variability has been seen in measurements of absolute CD4 counts than in the percentage of CD4 cells alone (13). In addition, some studies have suggested that measurements of absolute lymphocyte counts of seropositive subjects may be imprecise on some hematology instruments when assayed at a time more than 48 h after venipuncture, a practical issue to consider when using this marker in large clinical trials (14).
We have examined the relationship between a laboratory marker of CD4 percentage and early clinical events. A previous study had reported a greater decline in the percentage of CD4 cells in subjects who had an AIDS-defining symptom than those who had no events (10) and, interestingly, subjects who developed AIDS had a median CD4 percentage slope of -8, whereas those who did not develop AIDS had slopes of -2. A slope of < -7 increased the relative hazard of developing AIDS 35 times. A similar analysis in our healthier cohort, reported here, revealed a median change of -4.05 in those who had an early clinical event compared with 0 in subjects who did not have an event. In our study, subjects with a slope of < -4 had a fivefold risk of having an early clinical event and progressed more rapidly to a clinical event.
Early clinical events may be meaningful markers of disease progression, particularly in asymptomatic, HIV-1-infected subjects, who are now known to harbor a reservoir of HIV provirus in lymph node tissue and who may not develop full-blown AIDS for up to 10 years (15). Both oral candidiasis and oral hairy leukoplakia are early clinical markers that have been shown in some studies to be predictive of eventual AIDS and death (16). In one study, subjects with oral hairy leukoplakia showed an 83% chance of developing AIDS within 31 months (17). Although some clinical trials of early HIV infection, such as the Concorde and other studies (18), have used early clinical events as endpoints, further validation of these markers is warranted and may expedite evaluation of therapies for HIV infection.
This paper has described a relationship between CD4 percentage and early clinical events during a clinical trial in which subjects received the HIV-1 immunogen or IFA alone, a finding that is consistent with a recent report describing CD4 percentage as independently predictive of AIDS (19). Examining markers that are predictive of disease progression should identify new surrogate markers. Future studies are warranted to examine the extent to which changes in a marker can predict the clinical impact of new therapies.
Acknowledgment: The authors would like to thank the study coordinators at the sites, as well as the clinical research teams at Rhône-Poulenc Rorer, Collegeville, Pennsylvania, and The Immune Response Corporation, Carlsbad, California. Statistical consultation was provided by Ted Smith of Bio-Pharm, Blue Bell, Pennsylvania. We also would like to thank Stephen Lagakos, Steven Richieri, Anne Daigle, Richard Trauger, Francois Ferre, Fred Jensen, Steve Basta, Jon Allen, and Susie Stacy. The authors are also indebted to the late Dr. Jonas Salk for critical review of this manuscript and to the subjects who have participated in this trial.
1. Kessler DA. Issues in approving drugs for AIDS treatment. Regulatory Affairs
2. Moss RB, Ferre F, Trauger R, Jensen F, Daigle A, Richieri SP, Carlo DJ. Inactivated HIV-1 immunogen: impact on markers of disease progression. J Acquir Immune Defic Syndr
3. Stein DS, Korvick JA, Vermund SH. CD4+
lymphocyte cell enumeration for prediction of clinical course of human immunodeficiency virus infection: a review. J Infect Dis
4. Choi S, Lagakos SW, Schooley RT, Volberding PA. CD4+
lymphocytes are an incomplete surrogate marker for clinical progression in persons with aymptomttic HIV infection taking zidovudine. Ann Intern Med
5. Schechter M, Harrison LE, Halsey NA, Trade G, Santino M, Moulton LH, Quinn TC. Co-infection with human T-cell lymphotropic virus type I and HIV in Brazil. JAMA
6. Taylor JMG, Fahey JL, Detels R, Giorgi JV. CD4 percentage, CD4 number and CD4:CD8 ratio in HIV infection: which to choose and how to use. J Acquir Immune Defic Syndr
7. Greenspan D, Schiodt M, Greenspan JS, Pindborg J. AIDS and the Mouth
1992; Chicago, Illinois: Mosby YearBook.
8. Trauger RJ, Ferre F, Daigle AE, et al. Effect of immunization with an inactivated gp120-depleted HIV-1 immunogen in IFA on HIV-1 immunity, viral DNA, and CD4 percent. J Infect Dis
9. Centers for Disease Control. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR
December 1992; 41.
10. Burcham J, Marmor M, Dubin N, et al. CD4% is the best predictor of development of AIDS in a cohort of HIV-infected homosexual men. AIDS
11. Margolick JB, Munoz A, Donnenberg AD, et al. Failure of T-cell homeostasis preceding AIDS in HIV-1 infection. Nature Medicine
12. Statistical Analysis System,
Version 6, SAS Institute Inc., Cary, NC.
13. Gelman R, Eudey L. Biostatistical considerations for quality assessment of immunologic measurements used in clinical and longitudinal studies. Clin Immunol Immunopathol
14. Koepke JA, Smith-Jones M. Lymphocyte counting in HIV-positive individuals. Sysmex Journal International
15. Embretson J, Zupanic M, Ribas Jl, et al. Massive covert infection of helpter T cells: lymphocytes and macrophages by HIV during the incubation period of AIDS. Nature
16. Lifson AR, Hilton JF, Westernhouse JL, et al. Time from HIV seroconversion to oral candidiasis or hairy leukoplakia among heterosexual and bisexual men enrolled in three prospective cohorts. AIDS
17. Greenspan D, Greenspan JS, Hearst HG, et al. Relation of oral hairy leukoplakia to infection with the human immunodeficiency virus and the risk of developing AIDS. J Infect Dis
18. Cooper DA, Gatell JM, Kroon S, et al. Zidovudine in persons with asymptomatic HIV infection and CD4+
cell counts greater than 400 per cubic millimeter. N Engl J Med
19. Henrard DR, Phillips JF, Muenz LR, et al. Natural history of HIV-1 cell-free viremia. JAMA