Early treatment of asymptomatic human immunodeficiency virus type 1 (HIV-1) infection remains controversial. In the AIDS Clinical Trials Group 019 study, zidovudine was shown in 1990 to slow the clinical progression to AIDS in infected but asymptomatic subjects. (1) However, a follow-up of those subjects found no evidence of longer survival with the use of zidovudine. (2) Furthermore, the Concorde study found that there was not only no survival benefit from early treatment with zidovudine, but also no effect on the overall progression of disease. (3) Now, in this issue of the Journal, Volberding et al. report the further results of the AIDS Clinical Trials Group study, which show that immediate zidovudine therapy, as compared with deferred treatment, in asymptomatic persons with CD4 lymphocyte counts of 500 or more cells per cubic millimeter does not prolong the disease-free period or confer a survival benefit. (4) At the same time, Kinloch-de Loes et al. show that the use of zidovudine earlier, during primary HIV-1 infection and six months thereafter, results in a detectable improvement in the clinical course as well as an increase in the CD4 cell count. (5) The seemingly contradictory nature of these new findings, although attributable in part to differences in study design and study subjects, can be explained in the light of recent observations on the pathogenesis of HIV-1.
Three sets of recent scientific findings and therapeutic developments converge to favor an aggressive interventional strategy early in the course of HIV-1 infection. First, over the past several years, it has been shown that newly infected persons generally harbor a relatively homogeneous population of HIV-1, in contrast to the diverse swarm of viruses found in persons with chronic infection. (6,7,8) This monoclonality or oligoclonality of the virus around the time of seroconversion is believed to result from a "bottleneck" effect exerted by as yet undefined selective forces during the transmission of HIV-1. A relatively homogeneous viral population is less likely to contain preexisting variants with drug resistance and is thus more likely to respond to therapy. These theoretical considerations support the early treatment of HIV-1, as is the case with any treatable infectious disease.
Second, the recent demonstration of the highly dynamic nature of HIV-1 replication in vivo has important implications for early therapeutic intervention. (9,10,11) In infected persons this replication is continuous, with very rapid kinetics. Half the virus in plasma is cleared and replenished every two days or less, and overall rates of virion production average nearly 1 billion particles per day. Over a typical course of HIV-1 infection lasting about 10 years, thousands of replication cycles occur, resulting in a total production of close to 10 trillion virions. Given this relentless, high-level viral replication in conjunction with the known mutation rate of HIV-1 reverse transcriptase of approximately 1 in 1000 to 1 in 10,000 per base, (12) it is expected that viral variants will be generated rapidly. (Therapeutically, the generation of viral diversity in HIV-1 infection may be likened to the development of metastasis in cancer.) Simple calculations suggest that after a few years of infection, every viable mutation at every position in the 10-kb genome will occur; certain combinations of double mutations may also emerge. Therefore, monotherapy as we know it is doomed to fail, especially in the case of antiviral agents to which HIV-1 can become resistant with a substitution of a single base. In the long run, effective treatment must instead force the virus to mutate simultaneously at multiple positions in one viral genome. This is best achieved by using a combination of multiple, potent antiretroviral agents.
From our understanding of viral dynamics (9,10) and levels of plasma viremia during different stages of HIV-1 infection, (13,14,15,16,17) it can be estimated that the number of virions produced during the typical one-month duration of primary infection approximates the number produced in several subsequent years of asymptomatic infection. Consequently, the effect of zidovudine therapy during the seroconversion period is expected to be greater than the effect of similar treatment for a year during the asymptomatic period. This concept provides a potential explanation for the apparent discrepancy between the results of Volberding et al. (4) and those of Kinloch-de Loes et al. (5) Moreover, a recent study demonstrates that the plateau concentration of plasma viral RNA after primary HIV-1 infection is predictive of the long-term clinical outcome. (18) Treatment at the time of seroconversion may lower the initial viral plateau ("set point") and thereby improve the subsequent clinical course.
Although these theoretical considerations argue strongly for early aggressive treatment, we have been limited to date by the lack of sufficiently potent drugs. When used individually, the currently available nucleoside inhibitors of reverse transcriptase -- zidovudine, didanosine, zalcitabine, and stavudine -- are relatively weak antiretroviral agents that at best lower the viral load by 0.7 log. (14,19) It could be argued that such imperfect therapy does not adequately test the concept of early treatment. As a crude analogy, although no one disputes the merit of early diagnosis and treatment of breast cancer, it is doubtful that a therapy that decreased the tumor burden by only 80 percent would be clinically beneficial, regardless of the timing of its administration. It is therefore most encouraging that substantial progress is being made in developing more potent antiretroviral agents and regimens. Novel inhibitors of HIV-1 protease, such as ritonavir (formerly known as ABT-538) (20) and MK-639, (21) lower plasma viremia by about 2.0 log. (9,10) Certain non-nucleoside blockers of reverse transcriptase, such as nevirapine, have an inhibitory effect of 1.0 to 1.5 log, (9) whereas the regimen of zidovudine plus lamivudine has shown an activity of about 1.7 log in vivo. (19) These advances suggest that we may now have a number of better weapons with which to fight HIV-1. Collectively, these drugs provide the third rationale for early intervention.
Given the measurable benefit of zidovudine treatment in primary infection observed by Kinloch-de Loes et al., (5) imagine what might be achieved if HIV-1 replication were reduced by 3.0 log or more. When safe and effective treatment regimens are defined, the time will be right to test the concept of ablative treatment in HIV-1 infection. To stack the deck in favor of success, we should exert maximal antiviral pressure (using the optimal regimen) on the virus when it is most homogeneous -- during the initial phase of infection. We should bear in mind the lessons learned from the treatment of tuberculosis and childhood leukemia, in which monotherapy resulted in transient responses that were quickly followed by relapses. It was aggressive combination chemotherapy at the outset that led to cures. Optimistically, we can hope that such an approach will become possible in patients infected with HIV-1.
David D. Ho, M.D.
Aaron Diamond AIDS Research Center
New York University School of Medicine
New York, NY 10016