The Body: The Complete HIV/AIDS Resource
Follow Us Follow Us on Facebook Follow Us on Twitter Download Our App
Professionals >> Visit The Body PROThe Body en Espanol
  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

Treatment Issues' Second Survey of Physicians' Treatment Practice

December 1997/January 1998

A note from Since this article was written, the HIV pandemic has changed, as has our understanding of HIV/AIDS and its treatment. As a result, parts of this article may be outdated. Please keep this in mind, and be sure to visit other parts of our site for more recent information!

When formulating this year's survey, we recognized that open-ended questions would limit our ability to quantify and analyze the responses; however, we once again desired a broader discussion than an objective, "check-box" format would allow. In many ways, this resembles a mini-interview with those who responded. We also recognized that questions of this nature, in an environment of complex treatment decisions, would be time-consuming and difficult to answer. As best said by one of our participants, Dr. Mark Katz, "Factors that go into making treatment decisions are so varied that it is truly difficult to format the complex thinking into words."

We sent out 13 more surveys this year (92 vs. 79 last year) hoping to receive at least the same number of responses (36) as last year. We received responses from 37 physicians (40%), 27 of whom participated last year. We would like to express our extreme gratitude to those who participated, a group of extremely busy physicians with vast experience in treating people living with HIV.

We thank the following people for responding:

    Bernard Bihari, New York, NY
    Jeff Birnbaum, Kings County Hospital, Brooklyn, NY
    James Braun, New York, NY
    Carol Brosgart, East Bay AIDS Center, Berkeley, CA
    Lisa Capaldini, San Francisco, CA
    Charles Carpenter, Brown University/Miriam Hospital, Providence, RI
    Cal Cohen, CRI/New England, Brookline, MA
    David Cooper, St. Vincent's Hospital, Sydney, Australia
    Douglas Dieterich, Liberty Medical, New York, NY
    Stephen Follansbee, San Francisco, CA
    Brian Gazzard, Chelsea & Westminster Hospital, London, England
    Howard Grossman, New York, NY
    Roy Gulick, New York University/Bellevue, NY, NY
    David Hardy, Los Angeles, CA
    Keith Henry, Regions Hospital, Saint Paul, MN
    Martin Hirsch, Mass. General Hospital, Boston, MA
    Hans Jäger, Munich, Germany
    Victoria Johnson, University of Alabama, Birmingham, AL
    Jon Kaiser, Davies Medical Center, San Francisco, CA
    Christine Katlama, Hôpital Pitié Salpétrière, Paris, France
    Mark Katz, Kaiser Permanente of So. Cal., CA
    Michael Lederman, Univ. Hospitals of Cleveland, OH
    George Matula, Permanente Medical Group, SF, CA
    Kenneth Mayer, Project Achieve, Pawtucket, RI
    Julio Montaner, St. Paul's Hospital, Vancouver, Canada
    Michael Para, Ohio St. Univ. Hospitals, Columbus, OH
    Frank Rhame, Abbott Northwestern Hospital, Minneapolis, MN
    Michael Saag, University of Alabama, Birmingham, AL
    Paula Schuman, Wayne State University, Detroit, MI
    Kent Sepkowitz, Memorial Sloan Kettering Cancer Center, New York, NY
    Joseph Sonnabend, New York, NY
    Leanna Standish, Bastyr University, Seattle, WA
    Barbara Starrett, New York, NY
    Rona Vail, Gouverneur Hospital, New York, NY
    Stephano Vella, Instituto Superiore Di Sanitá, Italy
    Paul Volberding, University of California, SF, CA
    Todd Yancey, New York, NY

1. How do you decide when to start antiretroviral therapy for your HIV-positive patients? Please be specific -- e.g., symptoms, threshold CD4 counts, threshold viral loads.

The prospect of initiating potentially difficult and lifelong therapies elicited a complex array of responses that may not be fully elucidated in the following summary. Patient-specific factors were frequently mentioned.

Brosgart: HIV therapy must be assumed to be lifelong. Potent combination therapy is not easy for patients. It is a major life-changing commitment. The best approach is to form a partnership with your patient, educate them as to HIV pathogenesis, treatment guidelines, options (prescription and experimental) and work together to find the right regimen at the right time. Even with this approach, adverse drug experiences will play a major role in continued compliance.

Carpenter: For those who do not want to initiate aggressive triple therapy including a protease inhibitor, I generally recommend that no therapy at all, with monitoring of plasma viral loads quarterly, is generally preferable to beginning dual nucleoside therapy. In such instances, I try to work out with each patient a virtually agreed-upon measurement (e.g., when the plasma viral load exceeds 40,000) which will eventually trigger initiation of triple therapy.

Cohen: The primary issue is a person's interest/enthusiasm to take medications. If the patient is ambivalent, I'd rather wait a bit and have them try taking vitamins to see how it feels to take pills regularly.

Gazzard: The first and most important issue would seem to be that the patient and physician are both willing and keen to take therapy for the long term. I think it is unlikely that eradication therapy would prove to be possible for the vast majority of patients.

Gulick: I discuss therapy with all HIV-infected patients. The best time to start therapy is when the patient is ready.

Lederman: I am not eager to begin therapy unless the patient is committed for the long haul and prepared to make adherence to the regimen a major priority.

Mayer: Because my tendency is to offer antiretroviral therapy with three active agents, it is more important for people with low viral loads and higher CD4 counts to understand the implications of chronic therapy. I do not feel the issue is that the patient has to feel that he or she has to go on it immediately when the CD4 count is high and the viral load is modest. It is more important that once they begin antiretroviral therapy that they are willing to commit to trying to be as adherent as possible to the regimen over the long haul.

Para: I don't decide, the patient does. I may recommend it but will go along with almost all patient wishes if they are informed and understand the risk/benefits.

Vail: I believe that ability to adhere to treatment is the strongest decision point for beginning treatment. I would rather have patients hold off on therapy than to take it and not be committed to it.

Vella: It is often a patient decision because of the obvious problems with adherence.

There is agreement that anyone with symptomatic HIV should be treated and there is the universal use of viral load and/or CD4 counts as tools to determine when to recommend therapy in asymptomatic HIV-positive individuals. Most responses to this question resembled the recommendations of the International AIDS Society-USA panel (IAS-USA) that are a bit more aggressive than the guidelines recently developed by the Panel on Clinical Practice for Treatment of HIV Infection convened by the United States Department of Health and Human Services (HHS). The HHS guidelines recommend initiation of therapy in asymptomatic individuals with CD4 counts below 500 or HIV RNA greater than 20,000 as measured by RT-PCR. The IAS-USA guidelines recommend initiation of therapy in those with viral loads of 5,000 to 10,000.

Thirty-one of our 37 survey participants pointed to a specific CD4 count or viral load level that prompts them to recommend therapy. A CD4 cell count below 500 (similar to the HHS guidelines) is the most common value below which therapy is recommended (by 20 survey participants). Five physicians recommend therapy when CD4 counts fall below 300. The most commonly mentioned viral load levels above which therapy is recommended are 10,000 (eight physicians) and 5,000 (seven physicians) which fall within the IAS-USA guidelines. Treating anyone with a viral load above detection was mentioned by six physicians, equal to the number of physicians who would treat when viral load exceeded 20,000 (as per the HHS guidelines).

Statements that reflect the "average" response:

Bihari: I generally urge HIV-positive patients with less than 300 CD4 cells and/or more than 15,000 virions by HIV RNA-PCR to begin antiretroviral therapy, symptomatic or not.

Birnbaum: I recommend therapy if the viral load is greater than 10,000 at any CD4 count but I feel less pressured about starting someone on medication if the CD4 count is above 500.

Grossman: I definitely start therapy for any patient with a CD4 count below 500. Over 500 CD4 cells, I strongly recommend therapy for patients with viral loads above 10,000. I encourage therapy for patients with PCR above 5,000.

Hirsch: Having participated in the IAS-USA and USPHS guidelines process, I am satisfied that they provide useful information on instituting therapy. Symptoms, CD4 counts below 500 and HIV plasma RNA above 5,000 to 10,000 are generally good indications to begin therapy.

Jäger: I recommend therapy if the CD4 count is below 300 and/or there's a significant decrease over several months; if the viral load is above 10,000 copies in at least two consecutive measurements; if there's an AIDS-defining illness or symptoms.

Kaiser: I ask three questions. Are there any HIV-related symptoms? Are the CD4 counts trending downward during the past several tests? Is the viral load above 20,000? If the answer to two of three is affirmative then I usually will recommend antiretroviral therapy.

Katlama: Treat if the viral load is above 5,000 and the CD4 count is below 500 or if the viral load is above 10,000 if the CD4 count is above 500.

Katz: Treatment is especially emphasized for any patient who is symptomatic or who has a CD4 count under 500 but I would say that at least 80% to 90% of patients who start do so because of viral load exceeding a sentinel threshold. We recommend treatment for bDNA values exceeding 5,000 to 10,000 [editor's note: equivalent to about 10,000 to 20,000 using the PCR assay]. I personally am inclined to offer antiretroviral therapy to a patient with any detectable viral load (500 copies per ml) with the possible and often-stated exception of those with known duration of infection exceeding several months to years with CD4 counts over 400 to 500 and if the viral load is under 5,000. I say to myself and to the patient, "Is her/his immune system doing something protective which may well be as effective as our antiretroviral therapy and with less risk of morbidity/side effects and at a lower cost as well?"

Matula: My prejudice is to start with CD4 counts below or near 500 but would follow patient preference to treat or not from CD4 400 to 450 or above if stable. The use of viral load depends on CD4 counts or duration or stability of disease. Long-term non-progressors I prefer to leave untreated (CD4 above 500 for ten or more years) regardless of viral load unless changing rapidly. I may accept 15,000 to 20,000 if the CD4 count is high and the patient does not want treatment. Above 20,000, I encourage therapy.

Saag: When viral load is above 5,000 or CD4 below 500, I recommend therapy. When the viral load is below 5,000 and CD4 above 500, I suggest and consider therapy.

Sepkowitz: I recommend treatment for viral loads above 5,000.

Standish: I am disposed to wait if the CD4 count is greater than 300 and the viral load is steady and low (less than 5,000).

Vail: If the viral load is less than 5,000 to 10,000 and stable, I advise patients that watchful waiting is a reasonable alternative. I also advise therapy generally at CD4 counts below 500 regardless of viral load with the exception of patients who have had a stable CD4 over time that is above 350 with a low viral load.

Vella: I recommend therapy if the CD4 counts are below 500 or RNA above 10,000 to 20,000.

Volberding: I pretty much follow the IAS-USA 1997 guidelines. I consider treatment for all with detectable HIV RNA and urge treatment if the RNA is above 5,000 or the CD4 is below 500.

Some focus on trends.

Bihari: I will also urge patients with higher CD4s and or lower viral loads to begin antiretroviral therapy if they have shown a significant drop in CD4 percent or rise in PCR in recent months.

Capaldini: I recommend combination therapy for patients with constitutional symptoms, rapidly rising viral loads or rapidly falling CD4 counts.

Para: There is no magic number for CD4 or viral load, trends are most important. Stable viral loads of 20,000 to 40,000 with stable CD4 counts above 400 are often followed in asymptomatic patients. If RNA is greater than 40,000, I would usually treat but would still like a few repeat assays. If the CD4 count is below 300, I would usually treat but would like assays repeated a few times.

Some physicians are a bit more aggressive.

Braun: We offer antiretroviral therapy to anybody with a detectable viral load, after discussing the pros and cons of drug therapy, probable rate of progression and carefully evaluating patient motivation.

Dieterich: I recommend treatment at any viral load. Symptoms and CD4 counts do not matter.

Hardy: I use two major criteria for starting antiretroviral therapy: detectable viral load (above 400 copies) and a patient's willingness to begin.

Schuman: I recommend treatment if viral loads are above detectable levels and/or CD4 counts are declining.

Yancy: Assuming that the patient is willing, I recommend treatment if the CD4 is below 500, regardless of PCR or if the PCR is measurable, regardless of CD4 count.

A small but vocal minority questions the consensus view:

Gazzard: While I am well aware of data that at any CD4 count those with a viral load over 5,000 or 10,000 copies are likely to progress, I am unaware that any strategic study has shown that this method of treatment is particularly successful. We in the UK tend to rely on drop in CD4 counts as a measure that treatment is required to prevent irreversible damage to the immune system. We obviously use viral load prior to therapy as an index of how likely it is the CD4 count will start to drop significantly. In my own practice, I would strongly encourage those with a CD4 count of around 400 to initiate therapy.

Henry: Individualize recommendations. The sicker the patient, the higher the HIV level and the lower the CD4 count, the more aggressive I would push for a four-drug regimen. The healthier the patient and the better the "numbers," the more I would suggest patience. I am very patient-focused with no absolute values.

Rhame: The arguments in favor of early therapy are powerful but they are based largely on our understanding of the pathophysiology of HIV disease and late treatment rather than direct clinical trials. While it is heresy, I am unconvinced that the viral load is relevant to the decision about when to start therapy. The question is, "At what point in the progression of HIV disease does one begin to lose the maximum benefit of therapy?" Yes, the high viral load patient will come to that point sooner but I don't see how that bears on the main issue. I suspect that the right way to think about starting therapy is by considering the expected time to regimen failure/resistance (they're probably the same) and the expected life span of the patient. If a regimen has a median duration of HIV undetectability of three years and a patient has a life expectancy of one year, he or she should definitely start it. But if life expectancy is ten years, starting a three-year regimen seems analogous to AZT monotherapy, only not quite as futile. These possibilities make me very anxious about the degree of enthusiasm that the current consensus bodies have for treating high helper-count persons. While it may be right, in the absence of a controlled trial I do not feel comfortable advocating early therapy so forcefully. I am certainly willing to prescribe it, I might take it myself, but I do not promote it aggressively. When a patient appears dedicated to starting a regimen and adhering to it rigorously, I will prescribe it for any patients, weakly advocate it for persons with CD counts above 300, moderately advocate between 200 and 300 and vigorously advocate below 200.

Sonnabend: I do not know when to start antiretroviral therapy in asymptomatic individuals in the absence of any evidence to guide me. I suggest that asymptomatic individuals be observed regularly to try and determine their individual rate of progression and start antiretrovirals if and when there are indications of progression (symptoms, falling CD4 counts, increasing viral loads).

2. What percentage of your HIV-positive patients are treated with monotherapy, dual-combination therapy or triple-combination therapy? How do you decide on the number of drugs to prescribe?

The vast majority of physicians responding to our survey prescribe three-drug combinations to their patients. Thirty-three of our 37 survey participants estimated the percentage of their patients on mono-, dual- or triple-combination therapies. On average, 1% of patients are on monotherapy, 17% are on dual-combination therapy and 82% are on triple- (or more) combination therapy. Reasons most often cited for using less than three drugs included: patients stable on dual therapy initiated before the advent of triple therapy; patient preference; intolerance to protease inhibitors; low baseline viral loads.

Of Those Treated,
Percent on:
Dual Therapy17%
Triple Therapy82%
Protease Inhibitors71%

A few physicians elaborate on the dynamics behind the number of drugs used.

Birnbaum: Some patients are on dual therapy with two nucleosides because they stop the prescribed protease inhibitor and want to continue on at least some medication.

Braun: We are converting all patients to quadruple treatment if the patient, drug tolerance and drug availability permit. Patients with longstanding viral suppression (less than 50 copies) on AZT/3TC/Crixivan or other Crixivan-based three-drug combinations are usually continued on a three-drug regimen.

Cohen: I have approximately 5% of patients on two reverse transcriptase inhibitors and approximately 15% on dual protease combinations. In general, I prescribe three drugs for those with viral loads below 100,000 and four drugs for those with viral loads above 100,000.

Follansbee: I have a few "holdovers" on zidovudine/lamivudine or stavudine/lamivudine who are tolerating therapy, with persistently responding CD4 cell counts and undetectable (less than 500) HIV-1 RNA who are hesitant to change. Otherwise, I only recommend didanosine/stavudine as dual therapy because of intrinsic difficulty with resistance developing in those patients with relatively high CD4 counts (above 400) and modestly elevated HIV-1 RNA levels (1,000 to 20,000) who are hesitant to try HAART. Seventy percent of my patients are on triple or quadruple therapy.

Gazzard: Relatively few patients are started on dual therapy (perhaps 20%). It is clear that dual therapy for those with viral loads of over 20,000 is unlikely to produce viral loads that go below detectable limits. The issue is of primary importance in those drug combinations that induce resistance. This may be less the case with combinations like ddI/d4T where the mechanism of failure may not be related to generation of resistant mutations and therefore going below detectable is a less important issue. The issue of whether to start with two or three drugs is driven largely by the viral load and to some extent by the CD4 count when the patient is first seen.

Grossman: Triple combination is the minimum. Patients who are progressed or very experienced are on as many as seven drugs.

Katlama: Forty percent are on dual combination and 60% are on triple combination based on baseline viral load.

Kaiser: If the CD4 count is below 200 or viral load is above 100,000, always recommend triple therapy. If the CD4 count is above 300 and viral load is below 100,000, start with dual therapy and add a third drug if necessary.

Matula:Those with high CD4 counts, asymptomatic and modest viral loads may try two drugs and monitor the effect over the short term. Three drugs are recommended for very advanced patients (CD4 below 300 or viral loads above 50,000). Four drugs are recommended only in very drug-experienced patients who are failing.

Mayer: A few individuals who have been stable on dualtherapy have elected to remain on it even while appreciating the likelihood that they will have to be switched to three agents, including two new active agents, at some point in the future.

Montaner: Patients who are on dual-nucleoside therapy and are stable with very low viral loads and very stable CD4 counts may choose to continue on that regimen under close supervision. If that is the case, we would offer them a change to a new triple-drug regimen as soon as there is any evidence of change in these parameters. We strongly discourage initiation of dual-nucleoside therapy.

Saag: One percent are on monotherapy which includes a few patients who are not willing to change therapy. Approximately 15% are on dual-nucleoside therapy, mostly ddI/d4T but some residual AZT/3TC or d4T/3TC patients with viral loads below 2,000. Seventy percent are on triple combos, mostly two nucleosides and a protease inhibitor with approximately 15% on two nucleosides and a non-nucleoside, usually as initial therapy. Ten percent are on four drugs if their viral load is above 100,000 and 4% are on five or more drugs, usually protease inhibitor-failure patients.

Vail: My patients who are on dual-drug therapy are folks who started treatment when two drugs were a reasonable option, have stable viral loads less than 10,000, stable CD4 counts and have opted to hold off on switching to a three-drug combination until their viral load increases or CD4 counts drop.

Abbreviations Used In The Following Sections
Short NameGeneric NameBrand NameDrug Class
HUhydroxyureaHydreacell-enzyme inhibitor

NRTInucleoside (analog) reverse transcriptase inhibitor
NNRTInon-nucleoside reverse transcriptase inhibitor
PIprotease inhibitor

3. What are your preferred drug choices for initial therapy (single-agent, dual-agent and triple-agent)?

With eleven antiretroviral drugs approved in the United States and three or four more expected in 1998, the number of potential combinations seems countless. The following table attempts to simply illustrate the number of times a combination was mentioned as potential first-line therapy. A total of 43 combinations were mentioned by our 37 survey participants. Factors considered in choosing first-line drugs include the ease of administration, effects of food intake, drug interactions, toxicity and patient preference. Although dual-combination therapy is used less frequently than triple therapy (as seen in question 2, above), ddI/d4T is the most commonly used dual combo. Triple-combination therapies that include one protease inhibitor are, for the most part, limited to indinavir or nelfinavir and are the most frequently mentioned first-line therapies. Ritonavir and saquinavir, with rare exception, are mentioned only in double protease combination regimens; however, it is important to note that this survey was completed before the commercial availability of saquinavir soft gel capsules (see "Fortovase"). When triple combinations include a non-nucleoside reverse transcriptase inhibitor, frequently in those with lower baseline viral loads, nevirapine is the drug of choice. Delavirdine was only mentioned as potential, albeit rarely used, first-line therapy by two physicians. And finally, only one physician mentioned the nucleoside reverse transcriptase inhibitor ddC as a potential member of a dual-combination regimen.

Number of Times Mentioned as Potential First-Line Antiretroviral Therapy
ddI/d4T +/- HU11
PI (unspecified)/2 NRTIs (unspecified)6
NVP/2 NRTIs (unspecified)5
RTV/SQV/2 NRTIs (unspecified)2
RTV/SQV/1 NRTI (unspecified)2
NNRTI (unspecified)/2 NRTIs (unspecified)2
NFV/2 NRTIs (unspecified)1
NFV/SQV/2 NRTIs (unspecified)1

Examples of thought processes behind first-line treatment choices.

Birnbaum: I prefer 2 NRTIs plus a protease inhibitor or an NNRTI. The initial protease inhibitor choice is nelfinavir 750 mg every eight hours. I am interested in twice-daily dosing but haven't yet prescribed it. I seem to be using a lot of nevirapine because of its twice-daily dosing, lower number of pills and better tolerability. This appears to improve adherence. It also helps a treatment-naïve person get used to the idea of taking multiple medications. The decision to use 2 NRTIs plus 1 NNRTI is also usually based upon a relatively high CD4 count and low viral load.

Cooper: When using triple therapy, we recommend 2 NRTIs (usually AZT/3TC) plus a potent protease inhibitor (usually indinavir) in symptomatics. In asymptomatics, we recommend two NRTIs (usually ddI/d4T) and an NNRTI (usually nevirapine).

Gulick: Currently preferred drug combinations are AZT (or d4T)/ddI/NVP; AZT (or d4T)/3TC/IDV; AZT (or d4T)/3TC/NFV; RTV/ SQV/1 or 2 NRTIs. Choosing between these regimens requires individualization of therapy (based on stage of disease, concomitant illness and medications, adherence and tolerance issues, etc.).

Hardy: In patients with CD4 counts above 500 and viral loads below 10,000, AZT/3TC or d4T/3TC may be an option as long as viral loads remain below detection. For patients with CD4 counts above 400 and viral loads between 10,000 and 50,000, I recommend NVP/d4T/3TC; NVP/AZT/3TC or NVP/ddI/d4T. In those with viral loads above 50,000 or CD4 counts below 400, I recommend AZT/3TC/IDV (or NFV); d4T/3TC/IDV (or NFV); or ddI/d4T/IDV (or NFV).

Katlama: If the viral load is below 40,000, I recommend a dual combination without a protease inhibitor. If the viral load is not below detection in six to eight weeks, I add a third drug. If the baseline viral load is above 100,000, start with a triple combination including a protease inhibitor.

Katz: In choosing the NRTIs of a three-drug combination my first choice is AZT/3TC due to the ease of administration of the Combivir preparation. For patients who won't take AZT, I recommend d4T/3TC. In both cases, however, a rapid and full suppression to undetectable viral load is necessary in order to avoid the codon 184 mutation. Thus, I repeat the viral load in two weeks and if undetectable levels have not been achieved, I rethink the formula. Likewise, for patients who are starting with viral loads above 100,000, I am especially concerned about this possibility and might even propose a regimen that does not contain 3TC. In such cases, I use ddI/d4T. I've found no increased rate of neuropathy (over the 5% to 10% you see with either alone) but the necessity of the empty stomach for proper ddI administration makes this more difficult for even the most compliant patients to adhere (the recent reports of comparable levels achieved with 300 mg once daily might make me feel better about supporting its use up front). For a third drug, if the viral load is under 15,000, I discuss the possibility of nevirapine. If the viral load is higher I suggest nelfinavir as the first-line protease inhibitor due to the relative lack of a food requirement. I rate having to take a medication on an empty stomach as a greater impediment than that of taking a regimen twice or three times daily.

Lederman: I use either 2 NRTIs with a protease inhibitor or RTV/SQV with 1 or 2 NRTIs. For those with high CD4 counts or low viral loads, I have used 1 NNRTI with 2 NRTIs.

Saag: Dual-agent choice is ddI/d4T. If the viral load is below 100,000 my choices for triple therapy are NFV (or IDV)/d4T (or AZT)/3TC. For those with viral loads above 100,000, the choices are NFV/IDV/d4T (or AZT)/3TC, RTV/ SQV/d4T/3TC or NFV/SQV/2 NRTIs.

Sonnabend: Unless the individual is sick with low CD4 counts and really high viral load, I prefer to start with nevirapine and 2 NRTIs; that is, in minimally symptomatic individuals or those just starting to decline. In sicker people, I would use a protease inhibitor (except Invirase) and 2 NRTIs. Which protease inhibitor depends on the patient.

A few physicians choose first-line therapy by paying special attention to preserving future treatment options.

Cohen: For someone with a viral load less than 20,000, we'd consider d4T/ddI/HU as a way to give a regimen with a greater than 50% success rate at reducing viral loads below 400 for months as well as preserving maximal future options.

Gazzard: In those in whom compliance may be difficult, dual therapy is often preferred and those with low viral loads (below 10,000) might be eligible for such therapy, particularly if they were willing to take d4T/ddI. With regard to triple-therapy regimens, I am impressed by the data that non-nucleoside reverse transcriptase inhibitors are able to produce viral loads below detection in those with modest viral loads initially (perhaps up to 100,000 copies). This certainly reserves protease inhibitors for later treatment and is often preferred by the patients themselves.

Grossman: I recommend AZT/3TC/NFV or d4T/3TC/NFV. Some patients start with NVP rather than NFV, but I am mostly holding off on this to avoid cross resistance with DMP 266 (efavirenz, "Sustiva").

A couple of rarely mentioned dual combinations as first-line therapy:

Bihari: IDV/NVP is my first choice for patients I feel can comply with the rigors of the indinavir schedule. For those who cannot because of schedules or temperament, I start with NFV/NVP. I never start with a triple combination unless it is to add 3TC to one of these pairs because the patient has hepatitis B with abnormal liver function tests.

Standish: I recommend starting with dual-combination therapy, usually IDV/3TC or NVP/3TC.

4. What are your second-line treatment strategies? How do you decide when to change therapy? Do you use the ultrasensitive PCR assay and does this influence your treatment decision? Do you use any drug-resistance assays when considering a new treatment strategy? Which do you use and how are these assays helpful?

The choice of second-line treatment strategies is dependent on first-line strategies. For those on dual-nucleoside therapy, the consensus is to change the nucleosides and add a third drug, either a protease inhibitor or a non-nucleoside. For those already on triple combinations comprised of two nucleosides and a non-nucleoside, the consensus is to change the nucleosides and add a protease inhibitor. For those on triple combinations that include a protease inhibitor, the following broad strategies were commonly mentioned: change at least two agents, including the protease inhibitor; change all three drugs if possible; use double protease inhibitors with nucleoside or non-nucleoside reverse transcriptase inhibitors; switch the nucleosides and the protease inhibitors and add a non-nucleoside. One physician first tries increasing the dose of the currently prescribed protease inhibitor while another admits that he is totally unsure of second-line therapy strategies.

The following are specific second-line combinations strategies most often cited:

    RTV/SQV/2 RTIs (nucleoside or non-nucleoside);

Follansbee: The choice of a follow-up regimen depends on the initial regimen and usually incorporates as many "novel" agents with nonoverlapping toxicity to either previous or concurrent agents on "as convenient" a regimen as possible, often trying to utilize pharmacologic interactions to lower doses or frequency of administration.

Montaner: The second-line strategy will depend on what was the first treatment used for the patient. In addition, recent data regarding dual protease inhibitors has generated a fair bit of enthusiasm so that this option is being increasingly offered to selected patients. Again, the safety profile of the medications and the constraints associated with each regimen, as well as patient preferences will play a significant role in trying to decide what will be an appropriate second-line regimen for a given individual.

Starrett: A second-line combination is any combination that is possible and can be up to five drugs, e.g., AZT/3TC/RTV/SQV/DLV.

There was a range of responses to the question on when to switch therapy. Other than drug toxicity, the decision is usually based on increasing viral load but there's a difference of opinion on what level or viral load change warrants the switch. [Note: "detectable" viral load means levels at or above 400 using standard PCR assays unless otherwise noted.]

Number of Times Mentioned as Reason to Change Antiretroviral Therapy
Viral load becoming detectable6
Viral load above 10,0003
Inadequate initial response3
Viral load above 1,0003
Viral load above 5,0002
Viral load increase of 0.5 log (3 fold)2
Rising viral load with constitutional symptoms2
Use the IAS-USA guidelines2
Increased viral load with decreasing CD4 count1
Combination of viral load, CD4 and symptoms1
Rapidly rising viral load1
Viral load above 2001
Viral load above 5,000 for dual nucleoside therapy1
Positive viral load slope between 20 and 2001

Gazzard: Changing therapy is decided on a combination of clinical symptoms, return of viral load towards normal and a dropping CD4 count. The most difficult group is those with a viral load which returns to baseline but a CD4 count which remains elevated.

Matula: When to change therapy is often empiric. It's easy if the viral load increases and the CD4 declines but many patients have a high and steady CD4 count yet the viral load creeps up. All else being equal, I will just observe the patient carefully. I have a real concern of using up the entire armamentarium in patients who will remain clinically well for many, many years, just to address the plasma RNA.

Rhame: I think the notion that a regimen is failing or not failing is simplistic. Partial failure occurs. Furthermore, the target viral load is partly contingent on the pretreatment value. It is much harder to get a viral load of 750,000 to below 400 than it is when an individual starts out at 20,000.

Saag: Failure is defined as any detectable viral load for those taking protease inhibitor combinations or greater than 5,000 in those on dual nucleosides. I accept detectable viral loads below 2,000 in very experienced patients on new regimens with baseline viral loads exceeding 100,000.

Volberding: This is a hard decision in patients with incomplete RNA suppression but with good CD4 counts and/or clinical recovery.

Do You Use?
Ultrasensitive PCR: 10 21
Genotyping assay for resistance: 1022
Phenotyping assay for resistance: 329

Eleven physicians (of 32 responding to the question) use the ultrasensitive PCR assay, but even fewer base treatment decisions on it.

Braun: If the viral load is above 200 on two consecutive draws, I switch therapy. If it is below 200, I follow.

Brosgart: It is unclear whether less than 20 is really different than less than 400 and whether this translates into a survival advantage, delayed disease progression or less resistance.

Cohen: The ultrasensitive is useful between months six and twelve to ensure that the slope of the viral load is negative (i.e., dropping steadily over time). If the viral load is not continuing to drop, I'll consider adding to or intensifying the regimen.

Gazzard: We do have the ultrasensitive PCR available although we are unclear about how best to use it. I think it is important, as a research tool, and it looks as though for most clinical circumstances, falling below detection on a conventional assay predicts an eventual fall below detectable limits on the ultrasensitive assay. There remains, in my view, insufficient data to decide when an ultrasensitive assay should be below detectable and what to do with those with a viral load between 20 and 200.

Rhame: I push for as low a viral load as possible, even using the ultrasensitive assays. For patients with a viral load of 350 on "only" three drugs, I will at least gently push toward the addition of an NNRTI or the SQV/RTV combination.

Sonnabend: Who knows if keeping the viral load at 20 or at 400 makes the slightest difference?

Of the 32 physicians who responded to our resistance assay question, 10 use genotyping (tests that check HIV gene mutations) and only three have ever used direct lab culture tests for reduced susceptibility to specific drugs, or phenotyping. (And for these three, phenotyping was either new or very rarely used). The reliability of genotyping was questioned by those who don't use it and some of those who do. Some physicians do not have access to resistance assays due to third-party payer or managed care limitations.

Bihari: I have stopped using drug resistance assays since I have learned how little correlation they have with clinical indicators of resistance for the protease inhibitors.

Birnbaum: My facility is Medicaid-dependent and I have never had the luxury of phenotype or genotype assays but would welcome the availability of the tests. At present, the resistance game must be played as an educated best guess.

Braun: We just switched from Specialty Labs to Labcorp because the viral load threshold is lower (1,000 vs. 5,000).

Brosgart: I don't find the current generation of genotype assays helpful, nor have they been validated. I look forward to the new genotypic and phenotypic technologies (Virco, Virologic).

Follansbee: I have avoided drug resistance assays since they are nonstandardized and do not reflect compensatory mutations in their analysis. They do not measure previously prevalent viral subspecies that may be temporarily present in plasma in very low numbers because of relatively low "fitness" in the absence of antiretroviral pressure. In other words, a good medication history is worth the price of these tests at this point.

Gulick: I have only used genotyping in limited circumstances (highly experienced patient on a particular regimen to test for substitutions suggesting antiretroviral resistance to the currently taken drugs). In my limited use, I have not found resistance testing particularly helpful (so far).

Henry: This is very complicated. I use genotyping and I use therapeutic drug monitoring.

Lederman: I rarely use genotyping because only a few patients can afford it and I am not persuaded that it has helped much.

Matula: I do not use genotyping clinically. It only measures what is present as a majority now and misses minority genotypes. It is difficult to assess when multiple mutations may be required for resistance. It cannot predict the interaction of multiple agents. It cannot predict resistance with some drugs (i.e., d4T).

Rhame: I think genotyping is most useful in the situation of partial failure. This is a context in which you do not necessarily want to change all three drugs. There are not, after all, an unlimited number of new drug regimens. Being able to change only one or two drugs of a partially failing regimen seems rational when no genotypic resistance to the retained drugs is found. It must be recognized that these assays identify only the predominant virus so genotyping may have to be rechecked on the new regimen to find previously hidden, now emergent, resistance to the newly started drugs in the regimen.

Yancey: I use genotyping in all patients prior to the choice of therapy.

5. With more potent antiretroviral therapy now available, have you seen a change in the incidence of opportunistic infections (and which OIs)? Have you changed your OI prophylaxis strategies (primary or secondary)?

As reported in the literature, opportunistic infections, hospitalizations and deaths have been dramatically reduced by the advent of more potent antiretroviral therapy combinations. Not surprisingly, this finding was confirmed by all 37 physicians responding to our survey. However, a few noted that the incidence of both oral and vaginal candidiasis has not declined and that the incidence of lymphoma may actually be increasing.

Brosgart: There's been a dramatic drop in opportunistic infections for people on treatment including less PCP, CMV and MAC. However, there's still plenty of oral and esophageal candida. Our hospital census is down from 15 to 20 patients daily to five to eight patients daily. Home care is down from 50 to 75 patients to less than ten patients at any given time.

Capaldini: I've had only two patients have an OI in 12 months (I used to diagnose a new OI in my practice twice daily).

Carpenter: There has been a 70% reduction in admission of my patients to the hospital during the current calendar year.

Cooper: We've seen less CMV, MAC, cryptosporidiosis and microsporidiosis. PCP, toxoplasmosis and cryptococcus are about the same. We are seeing more KS, lymphoma and AIDS dementia.

Grossman: Marked decrease in OIs but I've seen CMV in patients with increased CD4 counts.

Hardy: OIs seem to be vanishing in patients responding well to potent combination antiretroviral therapy, especially CMV, MAC, fungal disease (with the exception of oral and vaginal candidiasis) and parasitic enteric disease.

Jäger: We estimate that the incidence of PCP has been reduced to about 25% of the 1995 rate and occurs predominately in patients not on protease inhibitors.

Sonnabend: There is obviously a very welcome change in that OIs are occurring less frequently. I'm not sure about lymphoma.

Seventeen of our 37 survey participants sometimes discontinue primary prophylaxis when there is both a dramatic antiviral response and significant, durable CD4 increases. This is more common for CMV and MAC prophylaxis where regimens tend to be cumbersome or toxic. However, the majority (20 physicians) believe that the data are insufficient and continue primary prophylaxis based on the CD4 nadir before initiation of HAART unless the patient is in a clinical trial for the expressed purpose of testing this hypothesis. From our sample, it seems that physicians from outside the U.S. are a bit more aggressive in stopping primary prophylaxis. There is a bit more caution in stopping secondary prophylaxis. CMV maintenance infusion therapy is sometimes reduced from daily ganciclovir or foscarnet to less frequent cidofovir.

Braun: If the CD4 count is persistently greater than 150% of the prophylaxis threshold for six months, we consider stopping prophylaxis: above 300 for PCP; above 75 for MAC; above 150 for toxo.

Brosgart: I still encourage prophylaxis on the basis of CD4 nadir but a number of patients have switched from daily foscarnet or ganciclovir to biweekly cidofovir maintenance.

Cooper: We stop primary MAC prophylaxis at CD4 counts above 75 and secondary CMV prophylaxis at CD4 counts between 50 and 100. "Passive" stopping of PCP prophylaxis does not seem to get patients into trouble.

Follansbee: I have suggested patients on CMV prophylaxis (oral ganciclovir) stop if their CD4 counts remain persistently above 75. I continue PCP prophylaxis, particularly secondary prophylaxis. I have withdrawn MAC prophylaxis if there has been a question of tolerance of the macrolide. I do not use rifabutin and would have no problem discontinuing it if I assumed the care of someone on it for MAC prophylaxis.

Gazzard: I have ceased to use primary prophylaxis in all circumstances although we continue secondary prophylaxis for most individuals with CMV disease and PCP.

Grossman: I stop MAC prophylaxis if the CD4 counts are above 200 for three months. I continue PCP prophylaxis if the CD4 count was ever below 200 for some time.

Hardy: I have generally not stopped OI prophylaxis in my patients when their CD4 cells have increased over the various thresholds used to start prophylaxis. I have decreased the frequency of CMV retinitis maintenance therapy to cidofovir every three to four weeks.

Jäger: MAC prophylaxis is negligible when CD4 counts remain over 100 for two to three months. If the CD4 count is above 200 and/or greater than 20% for two to three months, PCP prophylaxis can be discontinued.

Matula: I'm reluctant to stop PCP primary prophylaxis once the patient is qualified to start. However, I do not give MAC or CMV prophylaxis if the CD4 count has improved and stayed up for more than three months. I have not changed secondary prophylaxis but have just observed ganciclovir implants and not replaced them if the CD4 count is increased.

Montaner: We have been rather liberal in discontinuing primary prophylaxis in patients who have a healthy rebound in CD4 counts as a result of effective antiretroviral therapy. On the other hand, we have been less enthusiastic regarding discontinuation of secondary prophylaxis. This has been done on a case by case basis and only if there were clinical reasons why this should be considered.

Para: I am gradually cutting back primary prophylaxis depending on the health of the patient and extent of prior immune injury. I would like to see some data!

Rhame: Notwithstanding the consensus recommendation, I have stopped primary anti-MAC prophylaxis in patients with profound CD4 responses. I've even stopped several secondary MAC prophylaxis without any relapse. However, this is done only on a very selective basis where the CD4 cell response seems to be extremely good.

Standish: I have not changed my PCP prophylaxis recommendations, but I think I am starting to feel differently about MAI prophylaxis since it's not very effective and pretty toxic.

6. Have you seen an impact on wasting from highly active antiretroviral therapy combinations? Have you seen any changes in body composition and/or shape (positive or negative) with the use of protease inhibitors? If there are problems, what therapies have you tried and how successful have they been?

There was nearly universal belief that highly active antiretroviral therapy has had a positive impact on wasting, sometimes rather dramatic. There were, however, a few caveats.

Brosgart: Some patients (can't quantify) have increased CD4 counts, decreased viral load but lose weight, both fat and muscle, despite testosterone replacement. They have protease inhibitor taste perversion, nausea and anorexia.

Gazzard: We have certainly seen major problems with weight loss and lipid redistribution in patients on all classes of protease inhibitors.

Hardy: I have had to stop or decrease the dose of anabolic steroids and testosterone replacement treatments due to over-gaining of weight and, interestingly, loss of body fat (sunken cheeks, loss of buttocks).

Jäger: A number of patients continue to lose weight regardless of adequate antiretroviral therapy and some of these people profit from human growth hormone. In both men and women, we have seen positive increases in body weight and consequently an improvement in the patients' feeling about their bodies and health. Another option is the use of steroids.

Vail: I have seen cases of dramatic weight gain on potent antiretrovirals. Less frequently, I have seen viral loads under control and wasting continues. In these cases I work with our nutritionist on a combination of dietary advice, supplements, appetite stimulants where appropriate, and androgenic agents (testosterone, oxandrolone, growth hormone). I have had moderate success with this approach.

Yancy: While I have fewer people with wasting, those that have wasting continue to do so despite treatment.

About half of the survey participants have patients who have experienced alterations in fat distribution and most believe that it is caused by the protease inhibitors. A few believe that it is most commonly seen with indinavir but most say it is a drug class effect. Manifestations of this fat redistribution include increased abdominal fat, increased neck size, decreased facial fat, some muscle wasting of extremities or buttocks, buffalo hump and dorsal pad fat. Since the effect resembles Cushing's syndrome (elevated adrenal production of the hormone cortisol), it is sometimes described as cushinoid or cushings-like (see the "Protease Inhibitor Side Effects" article in this issue).

Capaldini: We have seen "Crix Belly," midline fat distribution associated with leg and arm muscle wasting.

Follansbee: There has been a body habitus shift to more lean body mass, loss of subcutaneous fat, with a more "gaunt" and "aged" look in some. I cannot tell if this is related to a certain protease inhibitor over the others, although in an anecdotal way, it seems more related to indinavir. I have seen a few cases of marked mesenteric and retroperitoneal fat hypertrophy (so called "Crix" belly), including one on ritonavir and one on SQV/NFV.

Gulick: I have seen several patients with "buffalo hump." The first time I saw it, I recommended a biopsy, which showed adipose ("fatty") tissue. I have not tried specific therapy.

Jäger: We see an increase in body fat around the waist and a decrease in muscle tissue on the arms and legs. The changes in composition resemble those seen by Cushing's syndrome, however, usually without the swelling in the face.

Montaner: In recent months we have recognized a small but increasing number of patients who have shown substantial wasting and redistribution of body fat. In our experience, this has been seen exclusively in the context of protease inhibitor-containing regimens. Despite active investigation from a nutritional and endocrinological standpoint, we have not yet been able to characterize the underlying mechanism for this phenomenon.

Starrett: I am starting to see wasting in protease inhibitor-treated patients. Although their body weight is stable, there is more fat and less lean body mass. This includes mesenteric fatty depositions and buffalo hump.

Volberding: "Crix belly" is simply the same old protein-wasting, fat-sparing disorder of HIV itself.

When this "cushings-like" fat redistribution syndrome occurs, only two physicians mentioned specific interventional strategies.

Deiterich: Hypogonadism persists and is the cause of "Crix belly." It responds to testosterone or decadurabolin.

Saag: We encourage aerobic exercise and resistance weight training for the increased abdominal girth with some success.

7. About what percentage of your patients are currently taking protease inhibitors? What adverse events and laboratory value abnormalities are you seeing with the use of protease inhibitors? What interventions have you tried and how well have they succeeded?

Thirty-one of our 37 survey participants estimated the percentage of their patients currently taking protease inhibitors, which averaged 71%.

Adverse events and abnormal laboratory values were pretty much as expected. The most often mentioned adverse events associated with indinavir therapy were kidney stones, increased bilirubin and abdominal bloating. One physician described a patient with increased liver function tests and jaundice that reversed when indinavir was switched to nelfinavir. The most common adverse event noted with nelfinavir therapy was diarrhea. For ritonavir, gastrointestinal adverse events included nausea, abdominal pain and diarrhea; circumoral and peripheral paresthesias (tingling or numbness around the mouth or more generalized) were occasionally mentioned.

The most common adverse events for the entire class of drugs included diarrhea, nausea, abdominal pain and vomiting. One physician noted hair loss in women. The laboratory abnormalities mentioned by approximately half of survey participants were hyperglycemia and lipid elevations, particularly increased triglycerides and cholesterol (see "Protease Inhibitor Side Effects" in this issue). Increased liver function tests were mentioned by about one-third of physicians. Two physicians cited rare "liver failure" and one suggested that two deaths from liver failure may have been protease inhibitor-related. Increased blood levels of the enzymes amylase, CPK and creatinine were each referred to by only one physician.

Very few interventions for these abnormalities and adverse events were noted. Most patients continue taking the drug and rarely require dose reduction or a change in antiretroviral therapy. Diarrhea is partially controlled with loperamide (Imodium), diphenoxylate/atropine (Lomotil), paregoric, opioids and psyllium. One physician treats abdominal discomfort and reflux with metroclopramide (Reglan) and sulcrafate (Carafate). Kidney stones are prevented or treated with hydration. Hyperglycemia (high blood sugar) sometimes requires oral hypoglycemic agents or insulin. The dyslipidemias (abnormal blood lipid levels) are not treated by most. One physician first tries nutritional intervention including fish oil and garlic before prescribing pharmacologic intervention. Several mentioned gemfibrizol (Lopid) for high triglycerides and two physicians prescribe atorvastatin (Lipitor) for high cholesterol levels. Leanna Standish uses a kidney tonic tea, herbs and antioxidants (see below).

Bihari: I have had to switch several patients from indinavir to nelfinavir because of kidney stones or fatigue. Persistent diarrhea with indinavir or nelfinavir improves with Imodium or Lomotil but is the most unpleasant persistent side effect.

Cohen: I have seen those abnormalities commonly known (e.g., dyslipidemias, elevated blood glucose). For lipids, we employ such approaches as fish oil, garlic supplements and other nutritional attempts and, if not successful, we go to pharmacologic interventions with at least some success.

Hardy: Adverse events and lab abnormalities are drug specific although cholesterol increases with all three. Indinavir: Gastrointestinal distress (bloating, nausea but rarely diarrhea); kidney stones have occurred in 5% to 10% of my patients; 80% to 90% have increased total and direct bilirubin; 40% to 50% have mild increases in other liver enzymes. Ritonavir: The overall worst as far as side effects which decrease with continued dosing but never go away completely. Increased triglycerides, nausea, diarrhea, fatigue; peri-oral and peripheral nerve superficial paresthesias; malaise or "ritonavir moments" which occur about two to three hours after dosing. Nelfinavir: Mild, moderate and severe diarrhea in 50% to 60% of patients; less commonly rash and gastrointestinal distress.

Johnson: Lots of diarrhea; stones with indinavir; peri-oral numbness with ritonavir especially in older women. I write Imodium prescriptions with nelfinavir.

Lederman: I have had two deaths associated with liver failure. One patient was on indinavir. I am not certain about the other, but, as I recall, he also was taking indinavir. I cannot be certain that the deaths were due to the drugs as both patients were also receiving nucleoside RTIs. One patient had chronic hepatitis C infection but at post (autopsy) had fatty necrosis. The other patient was negative for hepatitis A, B, and C but had hepatitis at post. In the patient with the fatty liver, the drug was continued until shortly before death. In the other patient, the drug was discontinued a week or two before death but liver deterioration persisted.

Mayer: A couple of patients have developed hyperglycemia and mild liver function test abnormalities. We tend to measure laboratory abnormalities on a monthly basis. One patient did develop glycosuria (sugar in the urine) and had to be started on insulin because of unmanageable diabetes. The patient had been overweight and had a family history of diabetes so it is not clear how much the initiation of protease inhibitor therapy exacerbated an underlying predisposition.

Para: Kidney stones occur in approximately 10% to 20%, not 3%. Acidifying the urine may be helpful. This can be accomplished by eating more meat.

Standish: Kidney problems are common. I'm also seeing high triglycerides and blood sugar. I am trying a botanical formula, "kidney tonic tea" (call Bastyr Clinic for the formula if you want it) for the kidney damage from indinavir. I use herbs like silymarin and lipoic acid for lipid problems. I am also trying carnitine for the high triglycerides but I can't tell if it's helping.

Vail: The main problems I have seen are increased glucose and increased lipids (triglycerides and cholesterol). I have tried diet control but have had to resort to medication or a change in antiretroviral therapy to control it.

Vella: We see a lot of nephrolithiasis (from stones to minor renal problems) in about 10% to 15% of patients. Drinking helps but does not solve the problem with indinavir. We also see the usual gastrointestinal problems with ritonavir, which cause 30% of patients to interrupt the drug during the first one or two months. We have seen much less common but serious side effects such as liver failure.

8. About how often do your patients fail to respond to all currently available protease inhibitors? All currently available nucleoside RTIs? All currently available non-nucleoside RTIs? All antiretrovirals? Why do you think these patients are not responding? Do you have "salvage" drug regimens?

The complexities of "drug failure" were reflected in the range of responses to this question, a question, probably too difficult to answer towards the end of this admittedly cumbersome survey. Therefore, very few survey participants addressed failure of individual drug classes but spoke about treatment failure, in general. That said, the responses to this question were frightening for a number of reasons. On the most basic level, these physicians have differing opinions on, or are unsure of, defining drug failure. Secondly, estimates of overall treatment failure ranged from 0% to 100%! It was unclear what factors were the basis for this broad range. Most commonly, if drug failure is defined as a rising viral load, it was estimated to occur in approximately 20% to 40% of treated patients. These drug failure rates reflect the commercial availability of the first clinically relevant protease inhibitors beginning in March/April 1996, less than two years ago. Cross-resistance and the absence of enthusiasm about treatment alternatives for those who have experienced multiple drug failure underscore the need to develop new antiretrovirals with novel targets.

Many speculate about what are the causes of treatment failure. The usual suspects mentioned include previous antiretroviral drug experience, difficulty with adherence, difficulty with dietary restrictions, adverse events and poor absorption. A few physicians expressed the need for assays to assess drug levels. High baseline viral loads, correlated with drug failure in studies, were cited by only one physician in this survey. The definition of treatment failure is vague.

Capaldini: If defined by a persistently detectable viral load, about 35% of patients experience combination therapy "failure." However, only 10% to 15% of patients actually show clinical evidence of disease progression on HAART.

Cohen: First, we have to define "fail to respond" as I have some patients with persistent suppression despite ongoing detectable viral levels as high as 40,000 for a year or more when compared to a pretreatment baseline viral load of 400,000.

Cooper: Fifty percent experience drug failure if based on viral load but most have "stable" CD4 counts and feel fine. We need a better definition of drug failure.

Jäger: Our definition of treatment failure is an increase in viral load over 10,000 copies in at least two measurements.

Mayer: In many cases, the patients may be partially responding. They have detectable viral loads that are not completely suppressed but maintain a stable plateau between 10,000 and 100,000.

Starrett: Approximately 20% of my patients experience some increase in viral load but their CD4 counts are still increased so we are watching. Having assays to assess drug levels might help.

Those who report drug failure on the extreme ends of the spectrum.

Bihari: Very few experience drug failure. Of 130 patients I have on indinavir plus nevirapine and 25 on nelfinavir plus nevirapine, only five in the first group never achieved nondetectable PCRs. None of the five had indinavir side effects or nondetectability on as much as 1,600 mg. every eight hours. Three of the five now have nondetectable PCRs on nelfinavir plus nevirapine. The other two eventually developed clinical evidence of some resistance to indinavir and nevirapine. Although these two patients are clinically stable with high CD4 counts and PCRs of 15,000 to 30,000, they are waiting for new drugs in development. Only one patient out of 150 broke through with a PCR greater than 400 after having several undetectable PCRs in 13 months on these drugs.

Hirsch: If patients are treated relatively early in the course of their infections with adequate regimens to which they adhere, they rarely experience drug failure over the course of at least two years. For those who experience failure of all regimens, I know of no magic strategy.

Johnson: Seems like 100% eventually experience drug failure: on one drug at six months; on two drugs at 6 to 18 months; on three drugs at 12 to 18 months. Patients are not getting enough drug into all cells to suppress the virus.

Standish: I'm hearing about drug failure but not seeing it yet.

Some more typical experiences.

Birnbaum: So many of my patients are treatment-naïve that I do not see a high drug failure rate overall. Drug failure tends to occur in the long-term survivors who lived through sequential monotherapy.

Braun: Initial treatments success rates are approximately 90%. Second combination success rates are approximately 60%. Third or subsequent combination success rates are approximately 25%.

Brosgart: About 30% of heavily pretreated patients experience drug failure that is probably due to preexisting mutations from prior nucleoside monotherapy. About 10% to 20% of drug-naïve patients experience drug failure either because of preexisting mutations (or infection with drug-resistant virus), but of greater importance is nonadherence.

Follansbee: About 50% ultimately experience drug failure, due to incomplete viral suppression and resistance and, occasionally, drug intolerance or adverse events. I think it is "easiest" to experience failure from all the NNRTIs. Occasionally, in addition to the above, patients do not respond because of malabsorption. It is a shame that there seems to be no pressure to develop inexpensive, easily performed serum or plasma assays for drug levels. It would greatly enhance patient care, at least theoretically.

Katz: It appears that there is more protease cross-resistance than we first imagined. Virus with the 84 and 90 mutations may be resistant to all four approved protease inhibitors. I have often wondered, with the exception of ddI and d4T for which there are specific dosages for varying body weights, why antivirals are dosed the same for a 100-pound and 220-pound person. Are some heavier patients being consistently underdosed and likewise some small patients receiving excessive amounts? Drug-level monitoring would provide such much-needed advice and support here.

Lederman: Drug failure occurs in 20% of patients and is mostly due to poor adherence, unanticipated interactions with other drugs or variability in pharmacokinetics.

Dr. Michael Saag: Estimated Drug Failure Rates
at one year
Protease Inhibitors 20%40%
All NRTIs 20% 40%
All NNRTIs 60% 80%
All antiretroviral less than 2%15 - 20%

A few cited potential "salvage" regimens.

Braun: We use NFL/DLV/ddI/3TC.

Carpenter: The drug regimen that I have used most often has been RTV/SQV/2 RTIs (nucleoside or non-nucleoside).

Follansbee: I have several patients on a multiple-drug regimen with the "best tolerated" and "most convenient" therapy to hopefully push the virus into a relatively "unfit" state and slow disease progression, awaiting better and more potent therapy. Occasionally these regimens employ SQV/NFV, in combination with ddI or d4T and an NNRTI, particularly delavirdine, in hopes it will drive up the levels of the protease inhibitor.

Gulick: In patients with highly resistant virus, I often "go back" to d4T and/or ddI (since genotypic resistance is poorly defined), with or without hydroxyurea. I also may try double protease inhibitor with or without NNRTIs. Further, expanded access programs with abacavir (1592), efavirenz (Sustiva, DMP 266) and adefovir (Preveon, bis-Pom- PMEA) should be available soon (and possibly combined together!).

Hardy: Pan-antiretroviral therapy failure occurs in 10% to 15% of my patients. For these patients I am using abacavir (1592), efavirenz (Sustiva, DMP 266) and a protease inhibitor (the best tolerated one). I'll add adefovir (Preveon, bis-Pom-PMEA) when it becomes available in December 1997 (hopefully).

Jäger: We are currently devising a six-drug treatment regimen for our patients who have experienced drug failure with all other treatment strategies.

Katz: I have limited experience with the use of hydroxyurea but have eyed its development as an option in HIV treatment for at least four years. At doses of 500 mg twice daily, it is generally safe and well tolerated and may provide a drop in viral load. The usual concomitant drop in CD4 counts needs further study, but I sense that the advantage of the suppressed viral replication is worth it for most patients. Fortunately, for patients with fewer remaining options, efavirenz (Sustiva, DMP 266) and abacavir (1592) may be available through expanded access, soon to be joined by adefovir (Preveon, bis-Pom-PMEA).

Mayer: Current regimens include compassionate usage of adefovir (Preveon, bis-Pom-PMEA), efavirenz (Sustiva, DMP 266) and abacavir (1592).

Saag: I try a four- or five- drug regimen of two to three protease inhibitors plus nucleosides and/or non-nucleosides. I try for at least two new drugs in this combination and try to avoid combinations of drugs previously used together, if possible.

Vail: I don't feel very hopeful for folks who are experiencing multi-drug failure now because of the problems with cross-resistance that the newest agents seem to share with existing drugs.

9. Please comment on the proposed U.S. government's treatment guidelines. Other comments on treatment strategies are also welcome (attach additional pages as necessary).

The majority of survey participants are generally satisfied with the guidelines ("OK," "not bad"). Some noted that they should be used as a general guide while making sure that therapy be individualized. This is done best by physicians with significant experience in treating people with HIV. Others believe that while it's nice to have broad guidelines, they tend to lag behind clinical practice and will be obsolete when new drugs are approved in 1998. Two physicians noted that the guidelines are probably more important for third-party and government payers than for practicing physicians and may provide a minimum level of reimbursement. On that theme, someone questioned how it is possible for state ADAP programs to limit the availability of these drugs now that there are federal guidelines.

Five physicians note that the guidelines understate the role of non-nucleoside reverse transcriptase inhibitors in combination therapy, particularly for those with lower viral loads. Another, who believes that the use of protease inhibitors in combination with non-nucleoside reverse transcriptase inhibitors is the reason for the very low breakthrough rate among his patients, complains that the guidelines virtually ignore this potential combination. He also believes that the guidelines should explore the need for therapeutic drug-level monitoring due to the large interpatient kinetic variability of the protease inhibitors (300%). A few physicians stated that the guidelines are too conservative.

Schuman: I think the official recommendations are quite conservative. More sensitive measurement of viral load is needed.

Starrett: Not aggressive enough but they are okay for the consensus as long as we understand there is room for more aggressive therapy.

Yancey: The currently proposed guidelines are not quite aggressive enough from the perspective of viral load ranges and when to start therapy.

Nine physicians believe that the guidelines are too aggressive in treating asymptomatic patients ("overtreatment," "premature"). This warning was more consistently expressed by survey participants from outside the U.S.

Cooper: The recommendations are too aggressive. While biologically plausible, toxicity and compliance issues on a population basis may limit their benefit.

Jäger: The proposed U.S. government's treatment guidelines seem quite aggressive and somewhat distant from the real world. The recommendation to change therapy as soon as any confirmed increase in viral load seems clearly too aggressive. This would quickly burn out all treatment possibilities. Our strategy, which has thus far proven to be a valuable alternative to playing out all our trump cards early, has been to save the protease inhibitors for later in the game.

Kaiser: I do not believe that every patient requires triple combination therapy. When combined with aggressive alternative/natural therapy regimens, dual antiretroviral therapy is highly effective at stabilizing mild-to-moderate HIV disease for very long periods of time, allowing one to save entire classes of antiretrovirals for later when they might be necessary.

Para: They are too aggressive with initiating therapy! Although based on Mellor's data, there's a problem with the analysis. A patient with a viral load of 5,000 to 10,000 may have a higher chance of getting sick in ten years but a low chance of getting sick in three years.

Vail: I believe that the guidelines are a bit too aggressive with asymptomatic, high CD4 count and low viral burden patients and may be pushing patients and providers to begin complex therapies before the patient is ready (and before it may be necessary).

Two physicians were most vocal in their concerns.

Brosgart: Unfortunately, much of the guidelines are based on theory, new paradigms of thought and extrapolation from existing data. While I hope that short-term clinical and virologic benefit translates into long-term benefit, this still remains to be seen. It is getting harder to do clinical endpoint trials. We still do not know how to correctly use these agents (when to start; when to switch; what to start with; which regimens are best to reserve; how do you sequence?). We need clinical trials to address long-term strategy, not just short-term efficacy. The media hype has not helped. AIDS is not over. Therapy is better but there is no cure. The alternative media will play a significant role in encouraging patients to participate in trials to help answer these strategic questions so that our next set of guidelines will really be knowledge-based!

Sonnabend: The recommendations regarding treatment in more advanced disease (where evidence derived from controlled clinical trials is available), will be of great benefit to patients in this category. However, it is far from clear that this will be the case for asymptomatic patients, even some of those with CD4 counts below 500. The potential risks are far from trivial. For an individual facing more than five, or even possibly three years free of disease, instigation of combination antiretroviral therapy with agents whose long-term toxicity is unknown, may in fact have the net effect of shortening that individual's life. The issue of quality of life is also of concern, as is the likelihood of a failure in compliance over a long period, with the attendant risk of the development of resistance, with the possible consequence that effective therapies may be unavailable at later stages of the disease. Faced with such difficulty in recommending when to initiate therapy in asymptomatic individuals, I believe that the panel might have devoted more consideration to the rate of disease progression in individual patients as a factor that should influence the decision as to whether or not to start antiretroviral therapy. Rates of disease progression vary widely, and it might require a 6 to 12 month period of observation to assess this rate in an individual patient. The uncertainty as to when to start therapy with these agents in asymptomatic patients is perhaps the most important issue that needs to be addressed at this time. It was, therefore, most surprising that the panel did not call for controlled clinical trials to resolve this important question. There is the unfortunate implication that in the area of AIDS medicine, convening a panel to make recommendations on such areas of clinical uncertainty has now replaced clinical studies as a means of guiding treatment decisions. It is therefore to be hoped that the Department of Health and Human Services will help to prepare the ground for such trials by educating patients and providers that there is indeed considerable uncertainty regarding when treatment should begin in asymptomatic individuals, and that the clearest answer to this question can only be obtained through appropriate clinical studies.

And finally, one physician said it all in one sentence.

Hardy: We have a lot to learn.

A note from Since this article was written, the HIV pandemic has changed, as has our understanding of HIV/AIDS and its treatment. As a result, parts of this article may be outdated. Please keep this in mind, and be sure to visit other parts of our site for more recent information!

  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

This article was provided by Gay Men's Health Crisis. It is a part of the publication GMHC Treatment Issues. Visit GMHC's website to find out more about their activities, publications and services.
See Also
More Research on HIV/AIDS Treatment Strategies