For some time after the discovery of HIV, the technology for measuring virus was relatively insensitive and it appeared that the degree of T-cell destruction was much greater than could be accounted for by the level of HIV replication. In the past several years, sensitive assays to measure HIV RNA in plasma have been developed. Studies of viral dynamics utilizing these assays have revealed that HIV-infected patients often have extremely high rates of viral production of 10⁹ to 10¹⁰ or more virions per day. There is now an appreciation that in the absence of effective therapy for HIV, nearly all HIV-infected individuals have active HIV replication that leads to immune system damage and progression to AIDS.

Also, a number of studies have shown that the plasma HIV RNA levels are a good measure of the level of HIV replication and are good predictors of subsequent CD4 decline and clinical deterioration. In particular, high viral loads are associated with a substantially higher rate of future disease progression (and CD4), and regimens that effectively suppress the viral loads are associated with better outcomes. CD4 counts provide a measure of the present status of the immune system and the current risk of developing opportunistic complications. These two parameters can thus be used together to assess the disease status of patients and as a guide for treatment decisions.

It is now generally accepted that a goal of anti-HIV therapy should ideally be to completely inhibit HIV replication in a patient to the extent possible. This concept was recently articulated by a panel convened by the National Institutes of Health to define the principles of therapy for HIV infection. Suppression of viral replication will stop the process of HIV-induced immune destruction and permit some reconstitution of the immune system to occur. As is described later, one of the most exciting recent advances in AIDS therapy was the finding that combinations of a PI with two NRTIs could suppress HIV replication in many patients to the point where HIV RNA could not be detected in the plasma for many months, even with highly sensitive assays.

Because of the ability of proviral HIV to integrate into host cells in a latent form, an absence of plasma viral RNA does not indicate that patients have been cured or rendered free of HIV, or even that all HIV replication in the body has been stopped. It has been hypothesized that with sufficiently prolonged effective therapy all the HIV-infected cells might eventually die. However, it has recently been shown that long-lived memory T cells can act as a reservoir for HIV, indicating that anti-HIV therapy would have to be maintained for years for these cells to die out. This goal may become even more elusive as we learn more about other potential reservoirs (such as the central nervous system).

At the same time, there is now some interest in strategies to activate such cells to produce HIV so they can be eliminated, either through a viral-induced cytopathic effect or through immunologic destruction. Also, if sufficient immune reconstitution can be attained, it may be possible to substantially control HIV replication for long periods of time, possibly with long-term or intermittent anti-HIV therapy. Thus, in the future it may be conceivable to attain long-term survival of HIV-infected patients even without complete eradication of their disease.

HIV has an extremely high mutation rate (approximately one mutation for every one to three genomes copied), and resistant strains can emerge to any of the available drugs. Resistance to reverse transcriptase inhibitors is associated with amino acid substitutions in the reverse transcriptase, whereas resistance to PIs is associated with amino acid substitutions in HIV protease. For some of the available drugs (e.g., lamivudine or nevirapine), a single mutation in the pol gene encoding reverse transcriptase can confer high-level (100-fold to over 1000-fold) resistance.

By contrast, resistance to certain other drugs is of a lesser extent or requires the development of several mutations. There is evidence that because of the high replication rate of HIV and the high mutation rate, patients generally harbor viruses resistant to antiviral drugs even before antiretroviral therapy is initiated. Under the selective pressure of anti-HIV therapy, resistant strains can then rapidly predominate. In patients given monotherapy with lamivudine or nevirapine, for example, clinically significant outgrowth of resistant strains can occur within 4 weeks of the initiation of therapy. In fact, the emergence of HIV resistance is perhaps the most challenging obstacle to the development of prolonged effective anti-HIV therapy, and it is of utmost importance to choose and manage antiretroviral therapy in individual patients in a way to forestall the development of resistance.

It has recently been shown that the emergence of resistant strains can be substantially delayed or even prevented by highly potent combination regimens that suppress HIV replication to undetectable levels. This was initially shown in a study in which patients received a PI (indinavir), two NRTIs (zidovudine plus didanosine), or the three drugs together. During a 24-week period, patients on the triple-drug combination had significantly fewer mutations conferring resistance either to the PI or the NRTI than the respective arms in which patients received just the PI or the NRTI. Thus, regimens that effectively suppress HIV replication (at least below the limits of detection of plasma HIV RNA) can forestall or prevent the emergence of resistant strains. This provides a strong rationale for achieving and maintaining complete HIV suppression whenever possible. Indeed, drugs such as nevirapine whose activity in other settings may be limited by the rapid emergence of resistance may be of greater value when used in regimens that achieve complete HIV suppression.