Guidelines for the Use of Antiretroviral Agents in HIV-Infected Adults and Adolescents

Use of Testing for Plasma HIV RNA Levels and CD4+ T-Cell Count in Guiding Decisions for Therapy

Decisions regarding initiation of or changes in antiretroviral therapy should be guided by monitoring the laboratory variables of plasma HIV RNA (viral load) and CD4+ T-cell count, as well as the clinical condition of the patient. As discussed in Principle 2, results of the two laboratory tests give the physician important information about the virologic and immunologic status of the patient and the risk for progression to AIDS. It should be noted that HIV viral load testing has been approved by the Food and Drug Administration (FDA) only for the reverse transcriptase polymerase chain reaction (RT-PCR) assay (Roche) and only for the determination of disease prognosis. However, data presented at an FDA Advisory Committee for the Division of Antiviral Drug Products (14-15 July 1997, Silver Spring, MD) provide further evidence for the utility of viral RNA testing in monitoring therapeutic responses. Multiple analyses in more than 5000 patients who participated in approximately 18 trials with viral load monitoring showed a statistically significant dose-response type association between decreases in plasma viremia and improved clinical outcome based on standard end points of new AIDS-defining diagnoses and survival. This relationship was observed over a range of patient baseline characteristics, including pretreatment plasma RNA level, CD4+ T-cell count, and prior drug experience. Thus, it is the consensus of the Panel that viral load testing is the essential variable in decisions to initiate or change antiretroviral therapies. Measurement of plasma HIV RNA levels (viral load), done by using quantitative methods, should be performed at the time of diagnosis and every 3 to 4 months thereafter in the untreated patient (AIII) (Table 2). CD4+ T-cell counts should be measured at the time of diagnosis and, generally, every 3 to 6 months thereafter (AIII). These intervals between tests are merely recommendations, and flexibility should be exercised according to the circumstances of the individual case. Plasma HIV RNA levels should also be measured immediately before and 4 to 8 weeks after initiation of antiretroviral therapy (AIII). This second measurement allows the clinician to evaluate the initial effectiveness of therapy because in most patients, adherence to a regimen of potent antiretroviral agents should result in a large decrease ([approximate]0.5 to 0.75 log₁₀) in viral load by 4 to 8 weeks. The viral load should continue to decline over the following weeks and in most individuals is below detectable levels (currently defined as <500 RNA copies/mL) by 12 to 16 weeks. The speed of viral load decline and the movement toward undetectable levels are affected by the baseline CD4+ T-cell count, the initial viral load, the potency of the regimen, adherence, prior exposure to antiretroviral agents, and the presence of any opportunistic infections. These individual differences must be considered when the effect of therapy is monitored. However, the absence of a virologic response of the magnitude discussed above should prompt the physician to reassess patient adherence, rule out malabsorption, consider repeated RNA testing to document lack of response, and/or consider a change in drug regimen.

Once the patient is receiving therapy, HIV RNA testing should be repeated every 3 to 4 months to evaluate the continuing effectiveness of therapy (AII). With optimal therapy, viral levels in plasma at 6 months should be undetectable, that is, below 500 copies/mL [2]. If HIV RNA remains detectable in plasma after 6 months of therapy, the plasma HIV RNA test should be repeated to confirm the result and a change in therapy should be considered, according to the guidelines in “Considerations for changing a failing regimen” (BIII). More sensitive viral load assays are in development that can quantify HIV RNA down to approximately 50 copies/mL. Preliminary data from clinical trials strongly suggest that lowering plasma HIV RNA levels to less than 50 copies/mL is associated with more complete and durable viral suppression compared with reducing HIV RNA to levels between 50 and 500 copies/mL. However, the long-term clinical significance of these findings is currently unclear.

When decisions are made about the initiation of therapy, the CD4+ T-lymphocyte count and plasma HIV RNA measurement should ideally be performed on two occasions to ensure accuracy and consistency of measurement (BIII). However, in patients who present with advanced HIV disease, antiretroviral therapy should generally be initiated after the first viral load measurement is obtained in order to prevent a potentially deleterious delay in treatment. It is recognized that the requirement for two measurements of viral load may place a significant financial burden on patients or payers. Nonetheless, the Panel feels that two measurements of viral load will provide the clinician with the best information for subsequent follow-up of the patient. Consistent with Principle 2, plasma HIV RNA levels should not be measured during or within 4 weeks after successful treatment of any intercurrent infection, resolution of symptomatic illness, or immunization. Because commercially available tests differ, confirmatory plasma HIV RNA levels should be measured by the same laboratory using the same technique to ensure consistent results.

A minimally significant change in plasma viremia is considered to be a threefold or 0.5-log₁₀ increase or decrease. A significant decrease in CD4+ T-lymphocyte count is a decrease of more than 30% from baseline for absolute cell numbers and a decrease of more than 3% from baseline in percentages of cells [3, 4]. Discordance between trends in CD4+ T-cell numbers and plasma HIV RNA levels can occur and was found in 20% of patients in one cohort studied [5]. Such discordance can complicate decisions regarding antiretroviral therapy and may be due to many factors that affect plasma HIV RNA testing (see Principle 2). In general, viral load and trends in viral load are felt to be more informative for guiding decisions regarding antiretroviral therapy than are CD4+ T-cell counts; exceptions to this rule do occur, however. For further discussion, refer to “Considerations for changing a failing regimen”; in many such cases, expert consultation should be considered.

Established Infection

Patients with established HIV infection are discussed in two arbitrarily defined clinical categories: 1) asymptomatic infection or 2) symptomatic disease (wasting, thrush, or unexplained fever for ≥ 2 weeks), including AIDS, defined according to the 1993 CDC classification system [6]. All patients in the second category should be offered antiretroviral therapy. Considerations for initiating antiretroviral therapy in the first category of patients are complex and are discussed separately below. Before therapy is initiated in any patient, however, the following evaluation should be performed:

1. Complete history and physical examination (AII),

2. Complete blood count and chemistry profile (AII),

3. CD4+ T-lymphocyte count (AI), and

4. Plasma HIV RNA measurement (AI).

Additional evaluation should include routine tests pertinent to the prevention of opportunistic infections, if not already performed (VDRL test, tuberculin skin test, toxoplasma IgG serology, and gynecologic examination with Papanicolaou smear), and other tests as clinically indicated (e.g., chest radiography, hepatitis C virus [HCV] serologic testing, and ophthalmologic examination) (AII). Hepatitis B virus (HBV) serologic testing is indicated in a patient who is a candidate for the hepatitis B vaccine or has abnormal results on liver function tests (AII), and cytomegalovirus (CMV) serologic testing may be useful in certain individuals, as discussed in the “USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with the human immunodeficiency virus” [1] (BIII).

Considerations for Initiating Therapy in the Patient with Asymptomatic HIV Infection

It has been demonstrated that antiretroviral therapy provides clinical benefit in HIV-infected individuals with advanced HIV disease and immunosuppression [7-11]. Although there is theoretical benefit to treatment for patients with CD4+ T-cell counts greater than 500 cells/mm³ (see Principle 3), no long-term clinical benefit of treatment has yet been demonstrated. A major dilemma confronting patients and practitioners is that the antiretroviral regimens currently available that have the greatest potency in terms of viral suppression and CD4+ T-cell preservation are medically complex, are associated with many specific side effects and drug interactions, and pose a substantial challenge with respect to adherence. Thus, decisions regarding treatment of asymptomatic, chronically infected individuals must balance several competing factors that influence risk and benefit.

Table 3 summarizes some of the factors that the physician and the asymptomatic patient must consider in deciding when to initiate therapy (see also Principle 3). Factors that would lead one to initiate early therapy include the real or potential goal of maximally suppressing viral replication; preserving immune function; prolonging health and life; decreasing the risk for drug resistance due to early suppression of viral replication with potent therapy; and decreasing drug toxicity by treating the healthier patient. Factors weighing against early treatment in the asymptomatic stable patient include the potential adverse effects of the drugs on quality of life, including the inconvenience of most of the maximally suppressive regimens currently available; the potential risk for developing drug resistance despite early initiation of therapy; the potential for limiting future treatment options due to cycling of the patient through the available drugs during early disease; the potential risk for transmission of virus resistant to protease inhibitors [PIs] and other agents; the unknown durability of effect of the currently available therapies; and the unknown long-term toxicity of some drugs. Thus, the decision to begin therapy in the asymptomatic patient is complex and must be made in the setting of careful patient counseling and education. The factors that must be considered in this decision are 1) the willingness of the individual to begin therapy; 2) the degree of existing immunodeficiency as determined by the CD4+ T-cell count; 3) the risk for disease progression as determined by the level of plasma HIV RNA (Table 4, Figure 1; see also Principles document); 4) the potential benefits and risks of initiating therapy in asymptomatic individuals, as discussed above; and 5) the likelihood, after counseling and education, of adherence to the prescribed treatment regimen. In this regard, no individual patient should automatically be excluded from consideration for antiretroviral therapy simply because he or she exhibits a behavior or other characteristic judged by some to lend itself to noncompliance. Rather, the likelihood of patient adherence to a complex drug regimen should be discussed and determined by the individual patient and physician before therapy is initiated. To achieve the level of adherence necessary for effective therapy, providers are encouraged to use strategies for assessing and assisting adherence that have been developed in the context of long-term treatment for other serious diseases; in this regard, intensive patient education regarding the critical need for adherence should be provided, specific goals of therapy should be established and mutually agreed upon, and a long-term treatment plan should be developed with the patient. Intensive follow-up should take place to assess adherence to treatment and to continue patient counseling for the prevention of sexual and drug injection-related transmission.

Initiating Therapy in the Patient with Asymptomatic HIV Infection

Once the patient and physician have decided to initiate antiretroviral therapy, treatment should be aggressive, with the goal of suppressing plasma viral load to undetectable levels. Table 5 and Table 6 summarize the recommendations on when to initiate therapy and what regimens to use. In general, any patient with fewer than 500 CD4+ T cells/mm³ or more than 10 000 (branched DNA [bDNA]) or 20 000 (RT-PCR) HIV RNA copies/mL should be offered therapy (AII). However, the strength of the recommendation for therapy should be based on the readiness of the patient for treatment as well as a consideration of the prognosis for disease-free survival as determined by viral load, CD4+ T-cell count (Table 4, Figure 1), and the slope of the CD4+ T-cell count decline. Note that the values for bDNA shown in the Figure 1 and Table 4 (first line or column) are the uncorrected HIV RNA values obtained from the Multicenter AIDS Cohort Study (MACS). It had previously been thought that these values, obtained on stored heparinized plasma specimens, should be multiplied by a factor of 2 to adjust for an anticipated twofold loss of RNA ascribed to the effects of heparin and delayed processing on the stability of RNA. However, more recent analysis suggests that the reduction ascribed to these factors is 0.2 log or less, so that no significant correction factor is necessary (Mellors J. Personal communication, October 1997). Values for RT-PCR are also shown in Table 4 and the (Figure 1); comparison of the results obtained from the RT-PCR and bDNA assay done by using the manufacturer's controls consistently indicate that the HIV-1 RNA values obtained by RT-PCR are approximately two times higher than those obtained by the bDNA assay [12]. Thus, the MACS values must be multiplied by approximately 2 to be consistent with current RT-PCR values. A third test for HIV RNA, the Nucleic-Acid Sequence Based Amplification (NASBA), is currently used in some clinical settings. However, formulas for converting values obtained from either bDNA or RT-PCR assays to NASBA-equivalent values cannot be derived from the limited data available at this time. This information will be added to the guidelines when it becomes available.

In current practice, there are two general approaches to initiating therapy in the asymptomatic patient: a therapeutically more aggressive approach that would treat most patients early in the course of HIV infection because of the recognition that HIV disease is virtually always progressive, and a therapeutically more cautious approach in which therapy may be delayed because the balance of the risk for clinically significant progression and other factors discussed above are felt to weigh in favor of observation and delayed therapy. The aggressive approach is heavily based on the Principles of Therapy, particularly the Principle that one should begin treatment before significant immunosuppression develops and one should treat to achieve undetectable viremia; thus, all patients with fewer than 500 CD4+ T cells/mm³ would be started on therapy, as would patients with higher CD4+ T-cell counts who have plasma viral loads greater than 10 000 (bDNA) or 20 000 (RT-PCR) HIV RNA copies/mL (Table 5). The more conservative approach to the initiation of therapy in the asymptomatic individual would delay treatment of the patient with fewer than 500 CD4+ T cells/mm³ and low levels of viremia who have a low risk for rapid disease progression, according to the data in Table 4; careful observation and monitoring would continue. Patients with CD4+ T-cell counts greater than 500/mm³ would also be observed, except those at substantial risk for rapid disease progression because of a high viral load. For example, the patient with 60 000 (RT-PCR) or 30 000 (bDNA) HIV RNA copies/mL, regardless of CD4+ T-cell count, has a high probability of progressing to an AIDS-defining complication of HIV disease within 3 years (32.6% if the CD4+ T-cell count is greater than 500/mm³) and should clearly be encouraged to initiate antiretroviral therapy. On the other hand, a patient with 18 000 HIV RNA copies/mL (RT-PCR) and a CD4+ T-cell count of 410 cells/mm³ has a 5.9% chance of progressing to an AIDS-defining complication of HIV infection in 3 years (Table 4). The therapeutically aggressive physician would recommend treatment for this patient to suppress the ongoing viral replication that is readily detectable; the therapeutically more conservative physician would discuss the possibility of initiation of therapy but would recognize that a delay in therapy due to the balance of considerations discussed above is also reasonable. In either case, the patient should make the final decision regarding acceptance of therapy following discussion with the health care provider of specific issues relevant to his or her own clinical situation.

When initiating therapy in the patient naive to antiretroviral therapy, one should begin with a regimen that is expected to reduce viral replication to undetectable levels (AIII). On the basis of the weight of experience, the preferred regimen to accomplish this is two nucleoside reverse transcriptase inhibitors (NRTIs) and one potent PI (Table 6). Alternative regimens have been used; these include ritonavir and saquinavir (with one or two NRTIs) or nevirapine as a substitute for the PI. Ritonavir and saquinavir (hard-gel capsule) dual PI therapy (without an NRTI) appears to be potent in suppressing viremia below detectable levels and has convenient twice-daily dosing; however, the safety of this combination has not been fully established according to FDA guidelines. In addition, this regimen has not been directly compared to the proven regimens of two NRTIs and a PI, and the Panel thus recommends that at least one additional NRTI be used when the physician elects to use two PIs as initial therapy. Substituting nevirapine for the PI or using two NRTIs alone does not achieve the goal of suppressing viremia to below detectable levels as consistently as does combination treatment with two NRTIs and a PI; it should be used only if more potent treatment is not possible. It should be noted, however, that some experts feel that there are currently insufficient data to choose between a three-drug regimen containing a PI and one containing nevirapine in the drug-naive patient; further studies are pending. Likewise, other regimens using two PIs or a PI and a non-nucleoside reverse transcriptase inhibitor (NNRTI) as initial therapy are currently in clinical trials with data pending. Although lamivudine (3TC) is a potent NRTI when used in combination with another NRTI, resistance to 3TC develops rapidly when suppression of virus replication is not complete [21, 22]. Therefore, the optimal use for this agent is as part of a combination of three or more drugs that has a high chance of completely suppressing virus replication. Other agents in which a single genetic mutation can confer drug resistance, such as the NNRTIs nevirapine and delavirdine, should also be used in this manner. Use of antiretroviral agents as monotherapy is contraindicated (DI), except when there are no other options or when the agents are used during pregnancy to reduce perinatal transmission, as noted below. When antiretroviral therapy is initiated, all drugs should be started simultaneously at full dose with the following three exceptions: Dose-escalation regimens are recommended for ritonavir; nevirapine; and, in some cases, ritonavir plus saquinavir.

Detailed information comparing the different NRTIs, the NNRTIs, the PIs, and drug interactions between the PIs and other agents can be found in Table 7, Table 8, Table 9, Table 10, Table 11 and Table 12. Particular attention should be paid to Table 9, Table 10, Table 11 and Table 12 regarding drug interactions between the PIs and other agents because these are extensive and often require dose modification or substitution of various drugs. Toxicity assessment is an ongoing process; assessment at least twice during the first month of therapy and every 3 months thereafter is a reasonable management approach.

Initiating Therapy in Advanced HIV Disease

All patients receiving a diagnosis of advanced HIV disease, which is defined as any condition meeting the 1993 CDC definition of AIDS [6], should be treated with antiretroviral agents regardless of plasma viral levels (AI). All patients with symptomatic HIV infection without AIDS, defined as the presence of thrush or unexplained fever, should also be treated.

Special Considerations in the Patient with Advanced-Stage Disease

Some patients present with opportunistic infections, wasting, dementia, or malignancy and are first diagnosed with HIV infection at this advanced stage of disease. All patients with advanced HIV disease should be treated with antiretroviral therapy. When the patient is acutely ill with an opportunistic infection or another complication of HIV infection, the clinician should consider clinical issues, such as drug toxicity, ability to adhere to treatment regimens, drug interactions, and laboratory abnormalities, when determining the timing of initiation of antiretroviral therapy. Once therapy is initiated, a maximally suppressive regimen, such as two NRTIs and a PI, should be used, as indicated in Table 6. Advanced-stage patients being maintained on an antiretroviral regimen should not have the therapy discontinued during an acute opportunistic infection or malignancy unless there are concerns regarding drug toxicity, intolerance, or drug interactions.

Patients who have progressed to AIDS are often treated with complicated combinations of drugs, and the potential for multiple drug interactions must be appreciated by clinician and patient. The choice of which antiretroviral agents to use must be made with consideration given to potential drug interactions and overlapping drug toxicities, as outlined in Table 7, Table 8, Table 9, Table 10, Table 11 and Table 12. For instance, the use of rifampin to treat active tuberculosis is problematic in a patient receiving a PI, which adversely affects the metabolism of rifampin but is frequently needed to effectively suppress viral replication in these advanced patients. Conversely, rifampin lowers the blood levels of PIs, which may result in suboptimal antiretroviral therapy. Although rifampin is contraindicated or not recommended for use with all of the PIs, one might consider using rifabutin at a reduced dose, as indicated in Table 8, Table 9, Table 10 and Table 11; this topic is discussed in greater detail elsewhere [26]. Other factors complicating advanced disease are wasting and anorexia, which may prevent patients from adhering to the dietary requirements for efficient absorption of certain PIs. Bone marrow suppression associated with zidovudine and the neuropathic effects of zalcitabine, stavudine, and didanosine may combine with the direct effects of HIV to render the drugs intolerable. Hepatotoxicity associated with certain PIs may limit the use of these drugs, especially in patients with underlying liver dysfunction. The absorption and half-life of certain drugs may be altered by antiretroviral agents, particularly the PIs and NNRTIs, whose metabolism involves the hepatic cytochrome p450 (CYP450) enzymatic pathway. Some of these PIs and NNRTIs (ritonavir, indinavir, saquinavir, nelfinavir, and delavirdine) inhibit the CYP450 pathway; others (nevirapine) induce CYP450 metabolism. The CYP450 inhibitors have the potential to increase blood levels of drugs metabolized by this pathway. At times, adding a CYP450 inhibitor can improve the pharmacokinetic profile of selected agents (such as adding ritonavir therapy to the hard-gel capsule formulation of saquinavir) as well as contributing an additive antiviral effect; however, these interactions can also result in life-threatening drug toxicity, as indicated in Table 10, Table 11 and Table 12. Thus, health care providers should inform their patients of the need to discuss any new drugs, including over-the-counter agents and alternative medications, that they may consider taking, and careful attention should be given to the relative risk versus benefits of specific combinations of agents.

Initiation of potent antiretroviral therapy is often associated with some degree of recovery of immune function. In this setting, patients with advanced HIV disease and subclinical opportunistic infections, such as Mycobacterium avium complex or CMV infection, may develop a new immunologic response to the pathogen; thus, new symptoms may develop in association with the heightened immunologic and/or inflammatory response. This should not be interpreted as a failure of antiretroviral therapy, and these newly presenting opportunistic infections should be treated appropriately while the patient is maintained on the antiretroviral regimen. Viral load measurement is helpful in clarifying this association.

Interruption of Antiretroviral Therapy

There are multiple reasons for temporary discontinuation of antiretroviral therapy, including intolerable side effects, drug interactions, first trimester of pregnancy (when the patient so elects), and unavailability of drug. There are no studies and no reliable estimate of the number of days, weeks, or months that constitute a clinically important interruption of one or more components of a therapeutic regimen that would increase the likelihood of drug resistance. If there is a need to discontinue therapy with any antiretroviral medication for an extended time, clinicians and patients should be advised of the theoretical advantage of stopping all antiretroviral agents simultaneously, rather than continuing one or two agents, to minimize the emergence of resistant viral strains (see Principle 4).

Considerations for Changing a Failing Regimen

As with the initiation of antiretroviral therapy, the decision to change regimens should be approached with careful consideration of several complex factors. These factors include recent clinical history and physical examination; plasma HIV RNA levels measured on two separate occasions; absolute CD4+ T-lymphocyte counts and changes in these counts; remaining treatment options in terms of potency, potential resistance patterns from prior anti-retroviral therapies, and potential for compliance and tolerance; assessment of adherence to medications; and preparation of the patient for the implications of the new regimen, which include side effects, drug interactions, dietary requirements, and possible need to alter concomitant medications (see Principle 7). A regimen may fail for many reasons, including initial viral resistance to one or more agents, altered absorption or metabolism of the drug, multidrug pharmacokinetics that adversely affect therapeutic drug levels, and poor patient adherence to a regimen due to either poor compliance or inadequate patient education about the therapeutic agents. In this regard, it is important to carefully assess patient compliance before changing antiretroviral therapy; health care workers involved in the care of the patient, such as the case manager or social worker, may be of assistance in this evaluation. Clinicians should be aware of the prevalence of mental health disorders and psychoactive substance use disorders in certain HIV-infected persons; inadequate mental health treatment services may jeopardize the ability of such individuals to adhere to their medical treatment. Proper identification of and intervention in these mental health disorders can greatly enhance adherence to medical HIV treatment.

It is important to distinguish between the need to change therapy because of drug failure and because of drug toxicity. In the latter case, it is appropriate to substitute one or more alternative drugs of the same potency and from the same class of agents as the agent suspected to be causing the toxicity. In the case of drug failure where more than one drug had been used, a detailed history of current and past antiretroviral medications, as well as other HIV-related medications, should be obtained. Optimally and when possible, the regimen should be changed entirely to drugs that have not been taken previously. With triple combinations of drugs, at least two and preferably three new drugs must be used; this is based on the current understanding of strategies to prevent drug resistance (see Principles 4 and 5). Assays to determine genotypic resistance are commercially available; however, these have not undergone field testing to demonstrate clinical utility and are not FDA-approved. The Panel does not recommend these assays for routine use at the present time.

Three different populations of patients should be considered with regard to a change in therapy: 1) individuals who are receiving incompletely suppressive antiretroviral therapy, such as single- or double-nucleoside therapy, with detectable or undetectable plasma viral load [discussed further below]; 2) individuals who have been on potent combination therapy, including a PI, and whose viremia was initially suppressed to undetectable levels but has again become detectable; and 3) individuals who have been on potent combination therapy, including a PI, and whose viremia was never suppressed to below detectable limits. Although these groups of individuals should have treatment regimens changed in order to maximize the chances of durable, maximal viral RNA suppression, the first group may have more treatment options because they are PI-naive.

Criteria for Changing Therapy

The goal of antiretroviral therapy-to improve the length and quality of the patient's life-is probably best accomplished by maximal suppression of viral replication to below detectable levels (currently defined as <500 copies/mL) sufficiently early to preserve immune function. However, this is not always achievable with a given therapeutic regimen, and frequently regimens must be modified. In general, the plasma HIV RNA level is the most important variable with which to evaluate response to therapy, and increases in levels of viremia that are significant, confirmed, and not attributable to intercurrent infection or vaccination indicate failure of the drug regimen regardless of changes in CD4+ T-cell counts. Clinical complications and sequential changes in CD4+ T-cell count may complement the viral load test in evaluating a response to treatment. Specific criteria that should prompt consideration for changing therapy include the following:

1. Less than a 0.5- to 0.75-log reduction in plasma HIV RNA by 4 weeks following initiation of therapy or less than a 1-log reduction by 8 weeks (CIII).

2. Failure to suppress plasma HIV RNA to undetectable levels within 4 to 6 months of initiating therapy (BIII). In this regard, the degree of initial decrease in plasma HIV RNA and the overall trend in decreasing viremia should be considered. For instance, a patient with 10 (⁶) viral copies/mL before therapy who stabilizes after 6 months of therapy at an HIV RNA level that is detectable but less than 10 000 copies/mL may not warrant an immediate change in therapy.

3. Repeated detection of virus in plasma after initial suppression to undetectable levels, suggesting the development of resistance (BIII). However, the degree of plasma HIV RNA increase should be considered; the physician may consider further, short-term observation in a patient whose plasma HIV RNA increases from undetectable to low-level detectability (e.g., 500 to 5000 copies/mL) at 4 months. In this situation, the patient should be followed very closely. It should be noted, however, that most patients who fall into this category subsequently show progressive increases in plasma viremia that will probably require a change in antiretroviral regimen.

4. Any reproducible significant increase (defined as threefold or greater) from the nadir of plasma HIV RNA levels not attributable to intercurrent infection, vaccination, or test methods except as noted above (BIII).

5. Undetectable viremia in the patient receiving double-nucleoside therapy (BIII). Patients currently receiving two NRTIs who have achieved the goal of no detectable virus have the option of continuing this regimen or may have modification to conform to regimens in the preferred category (Table 6). Experience indicates that most of these patients receiving double-nucleoside therapy will eventually have virologic failure with a frequency that is substantially greater than that of patients treated with the preferred regimens.

6. Persistently declining CD4+ T-cell counts, as measured on at least two separate occasions (see Principle 2 for significant decline) (CIII).

7. Clinical deterioration (DIII). In this regard, a new AIDS-defining diagnosis that was acquired after treatment was initiated suggests clinical deterioration but may or may not suggest failure of antiretroviral therapy. If the antiretroviral effect of therapy was poor (e.g., <10-fold reduction in viral RNA), then a judgment of therapeutic failure could be made. However, if the antiretroviral effect was good but the patient was already severely immunocompromised, the appearance of a new opportunistic disease may not necessarily reflect a failure of antiretroviral therapy but rather a persistence of severe immunocompromise that did not improve despite adequate suppression of virus replication. Similarly, an accelerated decline in CD4+ T-cell counts suggests progressive immune deficiency if there are sufficient measurements to assure quality control of CD4+ T-cell measurements.

A final consideration in the decision to change therapy is the recognition of the still limited choice of available agents and the knowledge that a decision to change may reduce future treatment options for the patient (see Principle 7). This may influence the physician to be somewhat more conservative when deciding to change therapy. Consideration of alternative options should include potency of the substituted regimen and probability of tolerance of or adherence to the alternative regimen. Clinical trials have shown that partial suppression of virus is superior to no suppression of virus. On the other hand, some physicians and patients may prefer to suspend treatment in order to preserve future options or because a sustained antiviral effect cannot be achieved. Referral to or consultation with an experienced HIV clinician is appropriate when one is considering a change in therapy. When possible, patients who require a change in an antiretroviral regimen but do not have treatment options among currently approved drugs should be referred for consideration for inclusion in an appropriate clinical trial.

Therapeutic Options When Changing Antiretroviral Therapy

Recommendations for changes in treatment differ according to the indication for the change. If the desired virologic objectives have been achieved in patients who have intolerance or toxicity, there should be substitution for the offending drug, preferably using an agent in the same class with a different toxicity or tolerance profile. If virologic objectives have been achieved but the patient is receiving a regimen not in the preferred category (such as two NRTIs or monotherapy), there is the option to continue treatment with careful monitoring of viral load or to add drugs to the current regimen to comply with preferred treatment regimens. As discussed above, most authorities feel that treatment with regimens not in the preferred category is associated with eventual failure and recommend the latter tactic. Currently, very few clinical data support specific strategies for changing therapy in patients in whom the preferred regimens that include PIs have failed; however, a number of theoretical considerations should guide decisions. Because of the relatively rapid mutability of HIV, viral strains with resistance to one or more agents often emerge during therapy, particularly when viral replication has not been maximally suppressed. Of major concern is recent evidence of broad cross-resistance among the PIs. Evidence indicates that viral strains that become resistant to one PI will have reduced susceptibility to most or all other PIs. Thus, the likelihood of success of a subsequently administered regimen of a PI plus two NRTIs, even if all drugs are different from the initial regimen, may be limited, and many experts would include two new PIs in the subsequent regimen.

Table 13 summarizes some of the most important guidelines to follow when changing a patient's antiretroviral therapy. Table 14 outlines some of the treatment options available when a decision has been made to change the antiretroviral regimen. As noted in the footnote to Table 14, insufficient data exist to suggest that any of these alternative regimens will be effective, and careful monitoring and consultation with an expert in the care of such HIV-infected patients is desirable. As stated above, a change in regimen because of treatment failure should ideally involve complete replacement of the regimen with different drugs to which the patient is naive. This typically would include the use of two new NRTIs and one new PI or NNRTI, two PIs with one or two new NRTIs, or a PI combined with an NNRTI. Dose modifications may be required to account for drug interactions when using combinations of PIs or a PI and NNRTI (Table 12). In some individuals, these options are not possible because of prior antiretroviral use, toxicity, or intolerance. In the clinically stable patient with detectable viremia for whom an optimal change in therapy is not possible, it may be prudent to delay changing therapy in anticipation of the availability of newer and more potent agents. It is recommended that the decision to change therapy and design a new regimen should be made with assistance from a clinician experienced in the treatment of HIV-infected patients through consultation or referral.

Acute HIV Infection

It has been estimated that at least 50% and as many as 90% of patients acutely infected with HIV will experience at least some symptoms of the acute retroviral syndrome (Table 15) and can thus be identified as candidates for early therapy [27-30]. However, acute HIV infection is often not recognized in the primary care setting because of the similarity of the symptom complex with that of the “flu” or other common illnesses. In addition, acute primary infection may occur without symptoms. Physicians should maintain a high level of suspicion for HIV infection in all patients presenting with a compatible clinical syndrome (Table 15) and should obtain appropriate laboratory confirmation (see below). Information regarding treatment of acute HIV infection from clinical trials is very limited. There is evidence for a short-term effect of therapy on viral load and CD4+ T-cell counts [31], but as yet no outcome data demonstrate a clinical benefit of antiretroviral treatment during primary HIV infection. Clinical trials completed to date have also been limited by small sample sizes; by short duration of follow-up; and, often, by the use of treatment regimens that have suboptimal antiviral activity by current standards. Nevertheless, these studies generally support antiretroviral treatment of acute HIV infection. Ongoing clinical trials are addressing the question of the long-term clinical benefit of more potent treatment regimens.

The theoretical rationale for early intervention, as provided in Principle 10, is fourfold: 1) to suppress the initial burst of viral replication and decrease the magnitude of virus dissemination throughout the body; 2) to decrease the severity of acute disease; 3) to potentially alter the initial viral “set point,” which may ultimately affect the rate of disease progression; and 4) to possibly reduce the rate of viral mutation due to the suppression of viral replication.

The physician and the patient should be fully aware that therapy for primary HIV infection is based on theoretical considerations, and the potential benefits (described above) should be weighed against the potential risks (see below). Most authorities endorse treatment of acute HIV infection based on the theoretical rationale, limited but supportive clinical trial data, and the experience of HIV clinicians.

The risks of therapy for acute HIV infection include adverse effects on quality of life resulting from drug toxicities and dosing constraints, the potential (if therapy fails to effectively suppress viral replication) for the development of drug resistance that may limit future treatment options, and the potential need to continue therapy indefinitely. These considerations are similar to those for initiating therapy in the asymptomatic patient and were discussed in greater detail in the section “Considerations in Initiating Therapy in the Asymptomatic HIV-Infected Patient.”

Whom To Treat during Acute HIV Infection

Many experts recommend antiretroviral therapy for all patients who demonstrate laboratory evidence of acute HIV infection (AII). Such evidence includes detectable HIV RNA in plasma on sensitive PCR or bDNA assays together with a negative or indeterminate result on an HIV antibody test. Measurement of plasma HIV RNA is the preferred method of diagnosis, but a test for p24 antigen may be useful when RNA testing is not readily available. It should be noted, however, that a negative result on a p24 antigen test does not rule out acute infection. When suspicion for acute infection is high, as in a patient with a report of recent risk behavior in association with symptoms and signs listed in Table 15, a test for HIV RNA should be performed (BII). (Patients diagnosed with HIV infection by HIV RNA testing should have confirmatory testing [Table 2].) As noted earlier, individuals may or may not have symptoms of the acute retroviral syndrome. Viremia occurs acutely after infection before the detection of a specific immune response; an indeterminate antibody test result may occur when an individual is in the process of seroconversion.

Apart from patients with acute primary HIV infection, many experts would also consider therapy for patients in whom seroconversion has been documented to have occurred within the previous 6 months (CIII). Although the initial burst of viremia in infected adults has usually resolved by 2 months, treatment during the 2- to 6-month period after infection is based on the likelihood that virus replication in lymphoid tissue is still not maximally contained by the immune system during this time. Decisions regarding therapy for patients who test positive for antibody and who believe the infection is recent but for whom the time of infection cannot be documented should be made by using the “Asymptomatic Chronic Infection” algorithm mentioned previously (CIII). Except in the setting of postexposure prophylaxis with antiretroviral agents [32], no patient should be treated for HIV infection until the infection is documented. In this regard, all patients presenting without a formal medical record of a positive HIV test result, such as those who have tested positive on available home testing kits, should undergo enzyme-linked immunosorbent assay and an established confirmatory test, such as the Western blot (AI), to document HIV infection.

Treatment Regimen for Primary HIV Infection

Once the physician and patient have decided to use antiretroviral therapy for primary HIV infection, treatment should be implemented with the goal of suppressing plasma HIV RNA levels to below detectable levels (AIII). The weight of current experience suggests that the therapeutic regimen for acute HIV infection should include a combination of two NRTIs and one potent PI (AII). Although most experience to date with PIs in the setting of acute HIV infection has been with ritonavir, indinavir, or nelfinavir [2, 33-35], there are insufficient data with which to make firm conclusions regarding specific drug recommendations. Potential combinations of available agents are much the same as those used in established infection (Table 6). It is recognized that these aggressive regimens may be associated with several disadvantages, including drug toxicity, large pill burden, cost of drugs, and the possibility of developing drug resistance that may limit future options; the latter is likely if virus replication is not adequately suppressed or if the patient has been infected with a viral strain that is already resistant to one or more agents. The patient should be carefully counseled regarding these potential limitations, and individual decisions should be made only after the risks and sequelae of therapy are weighed against the theoretical benefit of treatment (see above).

Because 1) the ultimate goal of therapy is suppression of viral replication to below the level of detection, 2) the benefits of therapy are based primarily on theoretical considerations, and 3) long-term clinical outcome benefit has not been documented, any regimen that is not expected to maximally suppress viral replication is not considered appropriate for treating the acutely HIV-infected individual (EIII). Additional clinical studies are needed to further delineate the role of antiretroviral therapy in the primary infection period.

Patient Follow-up

Testing for plasma HIV RNA levels and CD4+ T-cell count and toxicity monitoring should be performed as described above in “Use of Testing for Plasma HIV RNA Levels and CD4+ T-Cell Count in Guiding Decisions for Therapy”-that is, on initiation of therapy, after 4 weeks, and every 3 to 4 months thereafter (AII). Some experts feel that testing for plasma HIV RNA levels at 4 weeks is not helpful in evaluating the effect of therapy for acute infection because viral loads may be decreasing from peak levels even in the absence of therapy.

Duration of Therapy for Primary HIV Infection

Once therapy is initiated, many experts would continue to treat the patient with antiretroviral agents indefinitely because viremia has been documented to reappear or increase after discontinuation of therapy (CII). However, some experts would treat for 1 year and then re-evaluate the patient with CD4+ T-cell determinations and quantitative HIV RNA measurements. The optimal duration and composition of therapy are unknown, and ongoing clinical trials are expected to provide data relevant to these issues. The difficulties inherent in determining the optimal duration and composition of therapy initiated for acute infection should be considered when the patient is first counseled regarding therapy.

Considerations for Antiretroviral Therapy in the HIV-Infected Adolescent

Adolescents with HIV infection who were infected sexually or via injection drug use during adolescence appear to follow a clinical course that is more similar to that of adults with HIV disease than to that of children. In contrast, adolescents who were infected perinatally or via blood products as young children have a clinical course that may differ from that of other adolescents and adults who survive over the long term. Currently, most HIV-infected adolescents were infected sexually during the adolescent period and are in a relatively early stage of infection, making them ideal candidates for early intervention.

Puberty is a time of somatic growth and hormonally mediated changes, with females developing more body fat and males more muscle mass. Although these physiologic changes could theoretically affect drug pharmacology, particularly in the case of drugs with a narrow therapeutic index that are used in combination with protein-bound medicines or hepatic enzyme inducers or inhibitors, no clinically significant impact of puberty has been noted to date with the use of NRTIs. Clinical experience with PIs and NNRTIs has been limited. Thus, it is currently recommended that the dosage of medications used to treat HIV and opportunistic infections in adolescents should be based on Tanner staging of puberty and not on specific age. Adolescents in early puberty (Tanner stages I through II) should be dosed under pediatric guidelines, whereas those in late puberty (Tanner stage V) should be dosed according to adult guidelines. Youth who are in the midst of their growth spurt (Tanner stage III females and Tanner stage IV males) should be closely monitored for medication efficacy and toxicity when adult or pediatric dosing guidelines are chosen.

Considerations for Antiretroviral Therapy in the HIV-Infected Pregnant Woman

Guidelines for optimal antiretroviral therapy and for initiation of therapy in pregnant HIV-infected women should be the same as those delineated for nonpregnant adults (see Principle 8). Thus, the woman's clinical, virologic, and immunologic status should be of primary importance in guiding treatment decisions. However, it must be realized that the potential impact of such therapy on the fetus and infant is unknown. As discussed further below, the decision to use any antiretroviral drug during pregnancy should be made by the woman after discussion with her health care provider regarding the known and unknown benefits and risks to her and her fetus. Long-term follow-up is recommended for all infants born to women who have received antiretroviral drugs during pregnancy.

Women who are in the first trimester of pregnancy and who are not receiving antiretroviral therapy may wish to consider delaying initiation of therapy until after 10 to 12 weeks' gestation because this is the period of organogenesis when the embryo is most susceptible to the potential teratogenic effects of drugs; the risks of antiretroviral therapy to the fetus during that period are unknown. However, this decision should be carefully considered and discussed by the health care provider and the patient and should include an assessment of the woman's health status and the potential benefits and risks of delaying initiation of therapy for several weeks. If clinical, virologic, or immunologic variables are such that therapy would be recommended for nonpregnant women, many of the Panel members would recommend initiating therapy regardless of gestational age. Nausea and vomiting in early pregnancy affecting the ability to adequately take and absorb oral medications may be a factor in the decision regarding treatment during the first trimester.

Some women already receiving antiretroviral therapy may recognize their pregnancy early enough in gestation that concern for potential teratogenicity may lead them to consider temporarily stopping antiretroviral therapy until after the first trimester. Data are insufficient to support or refute the teratogenic risk of antiretroviral drugs when administered during the first 10 to 12 weeks of gestation. However, a rebound in viral levels would be anticipated during the period of discontinuation, and this rebound could theoretically be associated with increased risk for early HIV transmission in utero or could potentiate disease progression in the woman [36]. Although the effects of all antiretroviral drugs on the developing fetus during the first trimester are uncertain, most experts recommend continuation of a maximally suppressive regimen even during the first trimester. If antiretroviral therapy is discontinued during the first trimester for any reason, therapy with all agents should be stopped simultaneously to avoid development of resistance. Once the drugs are reinstituted, they should be introduced simultaneously for the same reason.

The choice of which antiretroviral agents to use in pregnant women is subject to unique considerations (see Principle 8). Currently, minimal data are available on the pharmacokinetics and safety of antiretroviral agents during pregnancy for drugs other than zidovudine. In the absence of data, drug choice will need to be individualized on the basis of discussion with the patient and available data from preclinical and clinical testing of the individual drugs. The FDA pregnancy classification for all currently approved antiretroviral agents and selected other information relevant to the use of antiretroviral drugs in pregnancy is shown in Table 16. It is important to recognize that the predictive value of in vitro and animal screening tests for adverse effects in humans is unknown. Many drugs commonly used to treat HIV infection or its consequences may have positive findings on one or more of these screening tests. For example, acyclovir is positive on some in vitro assays for chromosomal breakage and carcinogenicity and is associated with some fetal abnormalities in rats; however, data on human experience from the Acyclovir in Pregnancy Registry indicate no increased risk for birth defects to date in infants with exposure to acyclovir in utero [38].

Of the currently approved NRTI antiretroviral agents, the pharmacokinetics of only zidovudine and lamivudine have been evaluated in infected pregnant women to date [39, 40]. Both appear to be well tolerated at the usual adult doses and cross the placenta, achieving concentrations in cord blood similar to those observed in maternal blood at delivery. All the nucleosides except didanosine have preclinical animal studies that indicate potential fetal risk and have been classified as FDA pregnancy category C (defined in Table 16); didanosine has been classified as category B. In primate studies, all the NRTIs appear to cross the placenta, but didanosine and zalcitabine appear to have significantly less placental transfer (fetal to maternal drug ratios of 0.3 to 0.5) than do zidovudine, stavudine, and lamivudine (fetal-to-maternal drug ratios > 0.7) [41].

Of the NNRTIs, only nevirapine administered once at the onset of labor has been evaluated in pregnant women. The drug was well tolerated after a single dose and crossed the placenta and achieved neonatal blood concentrations equivalent to those in the mother. The elimination of nevirapine administered during labor in the pregnant women in this study was prolonged (mean half-life after a single dose, 66 hours) compared to that in nonpregnant women (mean half-life after a single dose, 45 hours). Data on multiple dosing during pregnancy are not yet available. Delavirdine has not been studied in phase I pharmacokinetic and safety trials in pregnant women. In premarketing clinical studies, the outcomes of seven unplanned pregnancies were reported. Three of these pregnancies were ectopic, and three resulted in healthy live births. One infant was born prematurely with a small ventricular septal defect to a patient who received approximately 6 weeks of treatment with delavirdine and zidovudine early in the pregnancy.

Although studies of combination therapy with PIs in pregnant infected women are in progress, no data are currently available regarding drug dosage, safety, and tolerance in pregnancy. In mice, indinavir has significant placental passage, but in rabbits, little placental passage was observed. Ritonavir has been shown to have some placental passage in rats. There are some special theoretical concerns regarding the use of indinavir late in pregnancy. Indinavir is associated with side effects (hyperbilirubinemia and renal stones) that could be problematic for the newborn if transplacental passage occurs and the drug is administered shortly before delivery. This is because the immaturity of the metabolic enzyme system of the neonatal liver would probably be associated with prolonged drug half-life, leading to extended drug exposure in the newborn and potential exacerbation of physiologic neonatal hyperbilirubinemia. In addition, because of immature neonatal renal function and the inability of the neonate to voluntarily ensure adequate hydration, high drug concentrations or delayed elimination in the neonate could result in a higher risk for drug crystallization and renal stone development than is observed in adults. These concerns are theoretical, and such effects have not been reported; because the half-life of indinavir in adults is short, these concerns may be relevant only if the drug is administered near the time of labor.

Gestational diabetes is a pregnancy-related complication that can develop in some women; administration of any of the four currently available PIs has been associated with new-onset diabetes mellitus, hyperglycemia, or exacerbation of existing diabetes mellitus in HIV-infected patients [42]. Pregnancy is itself a risk factor for hyperglycemia, and it is not known whether the use of PIs will exacerbate this risk. Health care providers caring for infected pregnant women who are receiving PI therapy should be aware of this possibility. They should closely monitor glucose levels in their patients and instruct their patients in recognizing the early symptoms of hyperglycemia.

To date, the only drug that has been shown to reduce the risk for perinatal HIV transmission is zidovudine administered according to the following regimen: oral administration antenatally after 14 weeks of gestation and continued throughout pregnancy, intravenous administration during the intrapartum period, and administration to the newborn for the first 6 weeks of life [43]. This chemoprophylactic regimen was shown to reduce the risk for perinatal transmission by 66% in a randomized, double-blind clinical trial, pediatric ACTG (AIDS Clinical Trials Group) 076 [44]. Insufficient data are available at present to justify the substitution of any antiretroviral agent other than zidovudine for the purpose of reducing perinatal HIV transmission; further research will address this question. For now, if combination antiretroviral drugs are administered to the pregnant woman for treatment of her HIV infection, zidovudine should be included as a component of the antenatal therapeutic regimen whenever possible and the intrapartum and neonatal zidovudine components of the chemoprophylactic regimen should be administered for the purpose of reducing the risk for perinatal transmission. If a woman does not receive zidovudine as a component of her antenatal antiretroviral regimen (e.g., because of history of non-life-threatening zidovudine-related severe toxicity or personal choice), intrapartum and newborn zidovudine should continue to be recommended; when use of zidovudine is contraindicated in the woman, the intrapartum component may be deleted but the newborn component is still recommended. Zidovudine and stavudine should not be administered together because of potential pharmacologic antagonism. When stavudine is a preferred nucleoside for treatment of a pregnant woman, it is recommended that antenatal zidovudine not be added to the regimen; however, intrapartum and neonatal zidovudine should still be given.

The time-limited use of zidovudine alone during pregnancy for chemoprophylaxis of perinatal transmission is controversial. The potential benefits of standard combination antiretroviral regimens for treatment of HIV infection should be discussed with and offered to all pregnant HIV-infected women. Some women may wish to restrict exposure of their fetus to antiretroviral drugs during pregnancy but may still wish to reduce the risk for transmitting HIV to their infant. For women in whom initiation of antiretroviral therapy for treatment of their HIV infection would be considered optional (e.g., CD4+ count > 500 cells/mm³ and plasma HIV RNA level < 10 000 to 20 000 copies/mL), time-limited use of zidovudine during the second and third trimesters of pregnancy is less likely to induce resistance because of the limited viral replication existing in the patient and the time-limited exposure to the antiretroviral drug. For example, the development of resistance was unusual among the healthy population of women who participated in pediatric ACTG 076 [45]. The use of zidovudine chemoprophylaxis alone during pregnancy might be an appropriate option for these women. However, for women with more advanced disease or higher levels of HIV RNA, concerns about resistance are greater. These women should be counseled that a combination antiretroviral regimen that includes zidovudine for reducing transmission risk would be more optimal for their own health than use of zidovudine chemoprophylaxis alone.

Monitoring and use of HIV-1 RNA levels for therapeutic decision making during pregnancy should be performed as recommended for nonpregnant women. Transmission of HIV from mother to infant can take place at all levels of maternal HIV-1 RNA. In untreated women, higher HIV-1 RNA levels correlate with increased risk for transmission. However, in zidovudine-treated women, this relationship is markedly attenuated [44]. Zidovudine is effective in reducing transmission regardless of maternal HIV RNA level. Therefore, the use of the full zidovudine chemoprophylaxis regimen, including intravenous zidovudine during delivery and the administration of zidovudine to the infant for the first 6 weeks of life, alone or in combination with other antiretroviral agents, should be discussed with and offered to all infected pregnant women regardless of their HIV-1 RNA levels. Health care providers who are treating HIV-infected pregnant women are strongly encouraged to report cases of prenatal exposure to antiretroviral drugs (either administered alone or in combinations) to the Antiretroviral Pregnancy Registry. This registry collects observational, nonexperimental data regarding antiretroviral exposure during pregnancy for the purpose of assessing potential teratogenicity. Registry data will be used to supplement animal toxicology studies and assist clinicians in weighing the potential risks and benefits of treatment for individual patients. The registry is a collaborative project with an advisory committee of obstetric and pediatric practitioners, staff from CDC and NIH, and staff from pharmaceutical manufacturers. The registry allows the anonymity of patients, and birth outcome follow-up is obtained by registry staff from the reporting physician. Referrals should be directed to Antiretroviral Pregnancy Registry, PO Box 13398, Research Triangle Park, NC 27709-3398; telephone 919-483-9437 or 800-258-4263; fax 800-800-1052.

Conclusion

The Panel has attempted to use advances in our understanding of the pathogenesis of HIV in the infected person to translate scientific principles and data obtained from clinical experience into recommendations that can be used by the clinician and patient to make therapeutic decisions. The recommendations are offered in the context of an ongoing dialogue between the patient and the clinician after having defined specific therapeutic goals with an acknowledgment of uncertainties. It is necessary for the patient to be entered into a continuum of medical care and services, including social, psychosocial, and nutritional services, with the availability of expert referral and consultation. In order to achieve the maximal flexibility in tailoring therapy to each patient over the duration of his or her infection, it is imperative that drug formularies allow for all FDA-approved NRTIs, NNRTIs, and PIs as treatment options. The Panel strongly urges industry and the public and private sectors to conduct further studies to allow refinement of these guidelines. Specifically, studies are needed to optimize recommendations for first-line therapy; to define second-line therapy; and to more clearly delineate the reasons for treatment failure. The Panel remains committed to revising their recommendations as new data become available.

Information included in these guidelines may not represent FDA approval or approved labeling for the particular products or indications in question. Specifically, the terms “safe” and “effective” may not be synonymous with the FDA-defined legal standards for product approval.