 | |
 |
John G. Bartlett, MD
Chief, Division of Infectious Diseases, Professor of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA. |
 | |
| This will be a brief presentation of the
guidelines recently developed by the Department of Health and Human Services
(DHHS). Although it is in draft form at the present time, the data presented
will probably not be modified substantially in the final publication. These
guidelines are the product of a 35-member Panel that included scientists
involved in HIV research, clinicians, representatives of the consumers and
various federal agencies including those that have a major impact on decisions
regarding health policies and allocation of resources (Table 1). The panel
limited the review to drugs that are already licensed in the U.S. for
antiretroviral treatment (Table 2). This was done to avoid interference with the
FDA which has the legal authority and the expertise to evaluate drugs. In terms
of future plans, the 35-member Panel is appointed through the year 1999 and has
now formulated the strategy to update the recommendations based on availability
of newly approved drugs and new data as they evolve. The specific method to do
this is through the creation of a subcommittee of 15 members that have monthly
conference calls and will make recommendations for review and approval by the
full Panel. This is regarded as a particularly important component of Panel
activities because changes in recommendations for treatment of HIV infection
occur with such extraordinary frequency. Great emphasis was placed on HIV RNA
levels since this relatively new laboratory technology has proven to be a
critical component of determining prognosis as well as response to treatment.
With regard to prognosis, the major data is from John Mellors based on the
experience of the Multicenter AIDS Cohort Study (MACS) using stored sera and
clinical evaluations for over 1,600 patients with HIV infection enrolled for
longitudinal observations over a ten-year period. The results of this analysis
are summarized in Table 3 which shows the baseline HIV RNA levels at the time of
entry in 1984 with subsequent observations for the relative hazard of AIDS or
death, median survival and the negative CD4 slope. The results show the power of
HIV RNA levels for determining clinical outcome. It might be noted
parenthetically that the FDA held a consensus conference in June 1997, for
analysis of these levels and clinical correlations from over 5,000 patients who
had participated in 16 clinical trials. Again, there was an excellent
correlation between HIV RNA levels and clinical events and survival that was
consistent regardless of gender, initial CD4 cell strata and risk category. As a
result of this analysis, the FDA now accepts HIV RNA levels as an adequate
parameter to judge the relative merits of drugs for licensing as a substitute
for outcome based on clinical complications and survival. There are three
commercially-available assays to determine HIV RNA levels: RT-PCR (Roche), bDNA
(Chiron), and NASBA (Organon). Currently, the most commonly used assays have
thresholds of about 400-500 copies/mL. The next generation tests will have a
threshold of around 20-50 copies/mL, and the interpretation of tests with this
threshold for making therapeutic decisions is presently not known. Specifically,
the stated goal of therapy is to achieve no detectable virus and to sustain this
activity for prolonged periods. This, however, was recommended with the
assumption that the threshold would be 400-500 copies/mL. In terms of
comparisons between assays, there is a modest difference between RT-PCR and bDNA
assays, so that values with the former are approximately double those of the
latter. There are insufficient data to make these comparisons with the NASBA
assay. There were actually two Panels involved in the development of the DHHS guidelines: One Panel, chaired by Dr. Charles Carpenter, dealt with the principles of HIV management. The second Panel utilized the principles document to formulate specific recommendations for therapy. Major conclusions of the Principles Panel were the following:
For developing the therapy guidelines, four questions were addressed: 1) When to initiate therapy; 2) What regimen to initiate; 3) When to change therapy; and, 4) What regimen to change to. 1) Therapy should be initiated for three patient populations (Table 5). First, patients who have the acute HIV syndrome or those who are seen within six months of seroconversion. Data to support this recommendation are sparse, but the Panel felt that this was the optimal time to treat since it probably represents the best opportunity to make a major impact on the long term outcome; the anecdotal experience to date supports this thesis. The second patient category is patients with symptomatic HIV infection; virtually all experts agree with this indication. The third patient population is more controversial because it deals with arbitrarily selected thresholds for VL and CD4 count in asymptomatic patients. The Panel agreed that treatment should be initiated for asymptomatic patients with a CD4 cell count <500/mm3 or a VL exceeding 10,000 (bDNA) or >20,000 (RT-PCR). In each patient category, but especially asymptomatic patients, the strength of the recommendation is determined by patients acceptance, the probability of adherence, and the prognosis as determined by VL and CD4 cell count. 2) Once the decision has been made to initiate therapy, the second question concerns the initial regimen (Table 6). The preferred regimen is two nucleoside analogues plus a PI. For the nucleosides, the five accepted regimens are summarized in Table 6. Other combinations of nucleosides (NRTIs) are unaceptable due to possible antagonism, lack of supporting clinical data and/or overlapping toxicity. For the PIs, the preferred drugs are ritonavir, indinavir, nelfinavir, and saquinavir (Fortovase formulation). Another combination that is in frequent clinical use that appears highly effective in terms of VL response is ritonavir plus saquinavir combined with one or two nucleosides. The alternatives, considered less effective because of less HIV suppression, is two nucleosides combined with nevirapine and two nucleosides combined with the hard gel form of saquinavir. The third category is two nucleosides; these are not generally recommended because, although clinical benefit has been clearly demonstrated, viral suppression is usually not sustained for extended periods. The fourth category is drug regimens that are considered contraindicated; this includes any drug used as monotherapy or combinations of nucleosides not included in the preferred listing above. Strong emphasis is placed on patient acceptance and ability to adhere to a complex medical regimen that requires multiple pills, complicate dosing instructions and substantial risk of toxicity. The counterbalancing risks and benefits to initiating treatment are summarized in Table 7. It is important to emphasize the consequences of poor adherence since this will lead to incomplete viral suppression, HIV replication and evolution of resistance that precludes future benefit of these drugs. The decision is obviously influenced by prognosis as dictated by VL and CD4 cell count since these are independent determinants of the speed of progression. 3) The third question deals with indications to change therapy (Table 8). If there is toxicity or intolerance, a single substitution is acceptable. If the patient is currently receiving a suboptimal therapy, there will be a difference according to the results achieved. If the patient has no detectable virus, one option is to simply follow the patient carefully with frequent VL assays and make appropriate substitutions when virus becomes detectable. The alternative would be to add a drug that will achieve a regimen in the preferred category. Many experts favor the latter tactic because the long term outcome with a double nucleoside regimen will usually be a failure. The most difficult scenario is the patient who fails to achieve the goal of no detectable virus using a regimen in the preferred category. The probability of failure by this definition can usually be predicted by the failure to decrease VL by at least one log after 1-2 months of treatment. However, the maximum suppression is generally achieved at 4-6 months of treatment. For patients who fail, it is important to determine if this is a true failure of the regimen, or if this reflects non-adherence. Fine tuning of recommendations to change or not change therapy must take into account the extent of viral suppression achieved. This is difficult because many patients have limited options that are attractive due to prior drug exposures with resistance or toxicity, thus limiting future options. If there has been a substantial reduction in VL, but persistent detectable virus at <10,000 copies/mL, some would favor continuation of the same regimen with careful follow-up measures of VL. |
Similarly, a patient who achieves no detectable virus, but has occasional values <5,000 copies/mL may be a candidate for close follow-up with the same regimen if alternative options are limited. It should be noted that the primary definition of failure in the guidelines is based on VL measurements. Additional or alternative definitions are based on the CD4 cell count changes or the development of complications of HIV infection. These alternative considerations are probably important to include in the overall assessment, but VL is considered the paramount indicator since this represents a direct measurement of the effect of treatment on the putative agent. Some patients develop clinical complications simply because immune reconstitution is not possible with virtually any regimen.
4) Decisions regarding alternative regimens are relatively easy in most patients who have toxicity or in patients who have prior substandard treatment. The most difficult decisions are in patients who fail with the optimal or preferred regimens because studies in this population are sparse and the relatively few studies that are available indicate there is no alternative regimen that is likely to achieve the goal of sustained viral suppression to undetectable levels (Table 9). The best results have been achieved with the combination of ritonavir plus saquinavir, or nelfinavir plus nevirapine. These drugs should be combined with nucleoside analogues to comprise the "quadruple regimen". Another rule when changing the regimen due to failure to achieve no detectable virus is to use at least one new nucleoside, and it is preferable to use an entirely new menu of drugs. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Controversial Issues There are a number of controversial issues in these guidelines. One of the most important is the arbitrary thresholds selected for CD4 cell counts and VL. While the goal is no detectable virus, there is not a clear consensus on the level intended as discussed above. The threshold selected by the Panel is 400-500 copies/mL since this was the assay used at the time that the guidelines were written. With the third generation assays that detect HIV-RNA at 20-50 copies/mL, many will prefer this lower threshold as the standard. It seems intuitively clear that the lower VL is a more idealistic goal, but the available evidence suggests that this threshold with the newer tests may not be realistic. For example, recent experiences in HIV care centers suggest that at least 50% of patients treated with triple therapy failed to achieve no detectable virus for a sustained period of time. This difference between the experience in clinical practice and therapeutic trials reflects the difference between patients who participate in trials and patients who are seen in "real life". Factors that correlate with viral failure include non-adherence, high VL and low CD4 cell counts and extensive prior exposures to antiretroviral agents. The result will be multiple changes in regimens for many patients leading to limited therapeutic options, often very early in the course of treatment. Nevertheless, nearly all agree that lower is better. There is also a controversy about the recommendation for the most highly active regimens at the initiation of treatment due to the limitations on future options in the event of failure. The hope is that future drug development will result in the addition of several new drugs that will widen options. Particularly attractive are regimens that avoid exposure to PIs early in the course. This reflects the concern for possible "class resistance" to the PIs meaning that incomplete HIV suppression with one PI will result in resistance to all agents in the class. This concern seems well established with indinavir and ritonavir, and it is possibly true with saquinavir and nelfinavir. There is also controversy regarding methods to deal with regimens that require temporary discontinuation of a single component due to toxicity or non-availability. Some have suggested that the best tactic in this setting is to discontinue all drugs. When this is done, there is recurrence of virus, often within one week, usually with "wild-type HIV" that is susceptible to all agents. The theory is that renewal of triple therapy should be highly effective because wild-type virus is usually susceptible to all agents. One of the major problems is that the time required for evolution of resistance is not well characterized in terms of the number of missed doses or days of treatment that will result in resistance. This is probably quite variable depending on the stage of disease and the agent in question. Pregnancy The prior recommendation was AZT during the second and third trimesters, with intravenous administration during delivery and then oral administration to the newborn for six weeks. The rationale is that this is the only regimen that has established merit for preventing perinatal transmission. The new recommendation of the Panel is to treat the pregnant woman according to the standard guidelines as enumerated above. A limiting factor in these recommendations is that no agents other than AZT, 3TC and nevirapine have established safety during pregnancy. It is consequently important that these drugs be given with informed consent regarding safety for the infant; this particularly applies to the first trimester when no agent has established safety. Occupational Exposure Multiple studies with a total of over 6,000 exposures to needlestick injuries from a source with established HIV infection indicate a risk of 0.3%. The risk is increased with a deep injury, when there is visible blood on the device, when the needle has been placed in a vein or artery of the source and when the source is in late stage disease, presumably reflecting a time of high viral load. These risks are cumulative, so that the total risk is substantially higher when two or more apply. A retrospective case-control study indicated that AZT given prophylactically will reduce the risk of transmission by 79%. Although the risk is reduced with mucocutaneous exposure, there are at least five health care workers who have had HIV transmission with this mechanism. The only body fluid that is associated with HIV transmission in the workplace is blood or bloody body fluids. Based on these observations, the current guidelines are to treat the exposed health care worker with AZT combined with 3TC. The recommendation for 3TC is based on the observation that, in combination with AZT, it is simply a more potent antiviral regimen. There is also a recommendation for the addition of indinavir for patients with a high risk injury or from a source strain judged likely to have resistance to AZT and/or 3TC. Nelfinavir has been added as another third drug option. Dosing regimens with these recommendations are standard and the duration of treatment is one month. The time of initiation of treatment is felt to be critical. It is preferable to initiate prophylaxis within 1-2 hours after exposure, although prophylaxis is offered up to one week after exposure. Sexual or Needle Sharing Exposure This is a controversial area because there are no data. The assumption is that prophylaxis with AZT works for the exposed health care worker, so it should be equally effective with HIV exposure in other settings. Thus, some have recommended that the same recommendations for the exposed health care worker be applied in alternative high risk settings such as unprotected intercourse, oral receptive intercourse or needle sharing. As with the health care worker, prophylaxis should be started as quickly as possible, but it should be offered for up to 72 hours after exposure and it should be continued for one month. Cost-effectiveness of Antiretroviral Drug Treatment Richard Moore and I have done a cost analysis of AZT versus a triple drug regimen to determine cost-effectiveness of the latter. This showed that the cost of triple therapy was U$S 10,000-18,000 per year of life preserved. Subsequent analyses have shown similar results. To judge the merit of this therapy, comparison should be made with other health care strategies that have become accepted standards. By comparison, the cost per year of life preserved for mammography is U$S 30,000; for hemodialysis it is U$S 50,000, and for coronary bypass surgery it is U$S 113,000. Thus, the new therapy according to initial analysis is acceptable when compared to other practices that have become medical standards. Who Should Care for HIV-infected Patients? The DHHS guidelines strongly state that HIV-infected patients should be cared for by an "expert". Furthermore, prior studies showed that care by a physician with HIV care experience resulted in a significant prolongation of survival for HIV-infected patients. The problem is that, despite the consensus that there should be an expert, there is no consensus of the definition of an expert. Some have used the requirement of caring for at least 50 patients with HIV infection; others favor the use of a clinical competency exam. A certifying board exam is under consideration by the American Board of Internal Medicine; a similar competency exam is currently available on the Hopkins Website. Experience on Guidelines We have reviewed the experience of approximately 2,000 patients who use the Moore Clinic at Johns Hopkins Hospital for HIV care. Nearly all are candidates for triple therapy according to the guidelines enumerated above. Nevertheless, only about 60% have received protease inhibitors. The presumed explanation is that the remaining 40% are not ready for this complex medical regimen. Among the 60% who have received triple therapy, about 60% have achieved the goal of no detectable virus using an assay with a threshold of 400 copies/mL. Of those who achieved the goal of no detectable virus, 30% subsequently had detectable virus during follow-up evaluations. The implications of these observations is that about 60% of those who have been treated with triple therapy have been failures according to the definition used in DHHS guidelines. In total, only about 25% have achieved the goal of no detectable virus for a sustained period. Despite what appears to be a disappointing experience, the overall impact of these drugs has been incredible. Comparison of 1997 with 1995 shows the rate of hospitalizations has decreased 76%, the frequency of opportunistic infections (Cytomegalovirus disease, disseminated M. Avium infection, P. Carinii pneumonia, toxoplasmosis and cryptosporidiosis) has decreased 50-100%, and the mortality rate is substantially reduced. A similar experience has been noted at other HIV/AIDS care centers as well. The implication is that the drugs have had a dramatic effect on outcome. Furthermore, analysis of costs for 1997 versus 1995 shows a substantial reduction in total costs for patients with CD4 counts <50/mL; the major reduction is in hospital costs. For patients with CD4 counts of 50-200/mL, the costs are comparable due to cost shifting with the decrease in hospital cost balanced by the increase in the pharmacy bill. |
