The 7th Conference on Retroviruses and Opportunistic Infections (CROI) continued to live up to its billing as the preeminent HIV meeting of the year. While no startling new information was presented, some reports about transmission risks have significant public health implications, and there were further encouraging reports on the new class of anti-HIV drugs, the "entry" inhibitors. Unfortunately, no significant advances were reported in the management of treatment-related metabolic side effects. This report will highlight the most important presentations, in our estimation.
More information can be obtained at the excellent 7th CROI Web site (www.retroconference.org), including searching and viewing by abstract number. (We have included abstract numbers in our reviews so you can identify the ones you want more details about.) In addition, STEP reported extensively from the CROI (www.thebody.com/step/retro2000/coverage.html).
HIV Transmission Risks and Public Health Implications
Four presentations at the 7th CROI had potentially significant public health implications:
- Risk of HIV transmission based on viral load (HIV RNA).
- Risk of HIV transmission by oral sex.
- Effectiveness of post-exposure prophylaxis (PEP) following possible sexual or needle exposure to HIV.
- Possibility of becoming re-infected with a more aggressive strain of HIV.
Viral Load and Risk of Heterosexual Transmission
Thomas Quinn from Johns Hopkins University presented a study conducted in Uganda on the effect of viral load on the risk of heterosexual transmission of HIV when one partner in a couple is HIV-infected and the other is not ("discordant couples"). A study of 15,127 people in Uganda followed 415 HIV-discordant couples for 30 months. The HIV-negative partner became infected in 90 of 415 (21.7%) discordant couples. (The transmission rate from males to females was basically the same as females to males.) The incidence of transmission was highest among 15- to 19-year-olds and among uncircumcised males. The higher the viral load, the greater the risk of transmission, with no transmissions occurring when the viral load was less than 1,500. It should be cautioned that this does not mean that it is not possible to transmit HIV when the HIV RNA load in the blood is very low, or non-detectable, and this was a study of heterosexual couples only. The significant implication for public health is that lowering viral load may result in fewer new HIV infections (Abstract 193).
Primary HIV Infections Associated with Oral Transmission
Researchers from San Francisco studied the risk of HIV transmission by oral sex in a group of people recently infected with HIV. The study followed 122 persons from June 1996 through June 1999. In 20 cases (16.4%), the route of HIV infection appeared, on initial evaluation, to be oral sex. After follow-up investigation, eight cases (6.6%) were classified as likely oral sex transmission. However, even this estimate may be low, due to exclusion of people where the method of transmission was uncertain. The men believed oral sex represented no or minimal risk. (The researchers did not discuss whether or not ejaculation occurred during oral sex, nor attempt to evaluate the risk of pre-ejaculatory oral sex.)
The conclusion was that despite a lower transmission risk than anal sex, oral sex may be an important mode of HIV transmission due to its frequency, and that standardized investigation of HIV transmission via oral sex is needed to understand the epidemiology and risk of this behavior. Dr. Anthony Fauci, the director of the NIAID, commented that he always thought it was possible to transmit HIV by oral sex (Abstract 473).
Final Results from the San Francisco Post-exposure Prevention (PEP) Project
This feasibility study conducted in San Francisco evaluated the effectiveness of post-exposure prophylaxis (PEP) following possible sexual or needle exposure to HIV. The study covered PEP in persons reporting sexual, injection drug use, or other nonoccupational exposures to HIV within the past 72 hours. All were offered four weeks of antiretroviral therapy, received risk reduction and medication adherence counseling, and were followed at 1, 2, 4, 26, and 52 weeks post-exposure. Over a 16-month period, 401 participants were enrolled. Condoms were not used in 64% of sexual exposures; condom failure occurred in 36%. Some 41% of participants knew their exposure source was HIV-infected. All but five persons elected to use antiviral therapy; AZT/3TC combination twice daily was most common (88%). A four-week course of therapy was completed by 80% of the people. In follow-up, no HIV seroconversion attributable to the initial exposure has occurred. The researchers concluded that it is feasible to provide PEP within 72 hours to persons reporting sexual or other nonoccupational exposures to HIV and such persons for up to one year (Abstract 196). (There is no way to know if any of these people would have become HIV-infected had they not received PEP, as the risk of HIV transmission from receptive anal sex is about 1% to 2% per exposure.)
HIV-1 Reinfection: A Documented Case?
Dr. Jonathan Angel presented documentation regarding a suspected infection of a super strain of HIV-1 that was resistant to all current highly active antiretroviral therapy (HAART) regimens. The infection led to a rapid advance of HIV disease in the infected person. Dr. Angel described a situation where Patient A with non-progressive HIV infection was possibly re-infected by Patient B with an HIV infection containing several drug-resistant mutations in their virus. If confirmed, this would be the first documented report of what appears to be infection of an HIV-1 drug-resistant virus in a person already infected with HIV-1. This report poses serious questions as to the safety of unprotected sex between people who are both HIV+ and on HAART. While many people who are HIV+ are making the choice to have unprotected sex with others who are also HIV+, such decisions could be putting them at great risk of acquiring a drug-resistant virus. More studies are needed to conclusively verify that reinfection does occur in people (Abstract LB2).
Lipodystrophy: More Questions than Answers
The metabolic changes experienced by people with HIV on HAART received much attention. While the risk factors for developing these complications, and their reversibility, continue to remain unknown, the 7th CROI did present some data. Risk is increased by:
- Age -- older people are more at risk.
- Having symptomatic HIV disease.
- Effective viral load suppression (the better the drugs work to suppress the virus, the more likely it is they will cause lipodystrophy).
- Increased duration on protease inhibitors.
Lipodystrophy also appears to be associated with elevated cholesterol and triglyceride levels, and possibly an increase in the cortisol-DHEA ratio. What is not known is whether these changes are specifically related to drug toxicity of HAART, the natural course of HIV, or a combination of things.
The effect of switching drug combinations to reduce or reverse the body composition changes is being explored. Much of the study data is conflicting, making the decision to change therapies because of lipodystrophy difficult for both patient and provider. Of note was a study of 36 people, which evaluated the reversibility of lipoatrophy (loss of fat from the face, arms, and legs) when stopping the drug d4T. After nine months, four patients reported their body shape returned to normal, seven reported "major improvement," and another 21 reported "partial improvement." While this study is too small to draw the conclusion that stopping d4T will partially or fully reverse lipodystrophy, it does point to the need for larger long-term studies of stopping d4T (Abstract 52).
Another study of switching drug combinations involved substitution of nevirapine (Viramune) for a protease inhibitor (PI). This study followed 40 people who were switched from two nucleoside reverse transcriptase inhibitors (NRTIs) plus one PI to two NRTIs plus Viramune. At 24 weeks, no improvements were noted in the fat redistribution syndrome, but some improvements were noted in triglyceride and cholesterol levels. Clearly, larger long-term studies are needed to evaluate the efficacy of this change, and inclusion of the specific NRTIs in the pre- and post-switch regimens need to be evaluated (Abstract 45).
A group of abstracts emphasized the critical relationship between adherence (how many pills you miss) and decreases in viral load, as well as the need to directly measure adherence, rather than relying on self-reporting. Dr. Margaret Fischl presented an abstract comparing people on several AIDS Clinical Trial Group (ACTG) studies who were in jail, where they took all of their prescribed pills under directly observed therapy (DOT), to people who were on the same ACTG studies, but not in jail. At 48 weeks, 80% of the people in jail had HIV RNA loads below 400, compared to only 50% of the people who were not in jail and responsible for their own medications. Curiously, the incidence of side effects was higher in the people not in jail. While it would seem logical to expect that the more of their pills a person took, the more side effects they might have, Dr. Fischl commented that people who have better control of HIV replication might have fewer side effects, another possible, and unexpected benefit, of improved adherence (Abstract 71).
Another study compared patient self-reports of missed doses to data kept by pill bottle caps with a computer chip, which records how often the bottle is really opened. While people reported taking 81% of their prescribed doses, the pill cap data found they actually only took 56% of the pills. Researchers found a strong correlation between viral load suppression and adherence. The investigators are also planning HIV resistance tests to determine if resistance developed more often in people who were less adherent, accounting for the differences in suppression of HIV RNA observed (Abstract 69).
The Latest on Structured Treatment Interruption
The 12 abstracts reporting on various approaches and results from structured treatment interruptions (STIs) drew the most people of all the afternoon posters (Abstracts 349-360). It should be noted that these reports are generally of a small number of people in each study, include both people treated for primary HIV infection and chronic infection, and are not randomized trials with adequate comparison groups.
There are two types of STIs or drug holidays:
- Intermittently stopping all HAART drugs, when you have complete HIV suppression. (The theory is that the short, controlled, bursts in HIV replication act as an auto-vaccine.)
- Stopping all HAART drugs before starting a new regimen, when the current HAART regimen is failing. (The theory is that the drug-sensitive virus returns, decreasing the population of drug-resistant virus, thus making the new regimen more effective.)
Franco Lori presented data on a small group of chronically infected people treated with ddI and hydroxyurea (HU), the "PANDA" cohort (Abstract 352). When compared to matched people on HAART, when therapy was stopped for eight weeks, the 10 people treated with only ddI and HU had much smaller increases in HIV RNA. All people were restarted on their antiretroviral therapy after 8 weeks. It is unclear whether better immunologic control or some specific effect of the ddI and HU regimen caused the smaller increases in HIV RNA. This was not a randomized study so it must be interpreted cautiously.
Most of the posters reported rapid increases in HIV RNA as soon as HAART was stopped, usually to pre-HAART levels. One striking exception was a poster presented by Kendall Smith about a study that used low-dose Interleukin-2 (IL-2) after HAART was stopped (Abstract 355). All of the people received at least three months of IL-2 while on HAART, followed by continued low-dose IL-2 (2 million units daily) after HAART was stopped. This dose of IL-2 had very few side effects. After a rise in HIV RNA during the first two weeks of stopping HAART, the HIV RNA load dropped, and only small transient decreases were observed in CD4 cell counts. Continuous low-dose IL-2 might stimulate immunologic control of HIV, but a larger randomized trial is needed to verify these intriguing results.
The Swiss Intermittent Treatment Trial slide presentation was an early look at an ongoing study using STIs to induce HIV-specific immune responses. All people in this trial had HIV RNA loads below 50 and CD4 counts above 300. The trial was designed to stop HAART for two weeks, then restart HAART for eight weeks, and repeat this cycle four times, in 120 people. Preliminary data was presented on the first 99 people after the first cycle only. No real conclusions can be drawn yet, but one reassuring aspect is that, to date, most patients in the study appear to be able to regain viral suppression to undetectable levels using the same drug combinations they were on before the STI. We can infer that drug-resistant strains of the virus are not being created during the drug holiday. However, some people rebounded to over 1 million copies of HIV RNA in two weeks, and the researchers did not tell us what happened to their CD4 counts during the two-week STI. Others researchers have observed a very significant decrease in CD4 counts, even during short periods of treatment interruption. Also, while some improvements were noted in some measures of immune response to HIV, no improvements were noted in others (Abstract 458).
Another report examined STI in people with good virologic suppression (HIV RNA load under 20) on their current treatment regimen. Ten HIV-infected people who had been treated with d4T, 3TC, and ritonavir or indinavir for 52 weeks, and had HIV RNA loads below 20 for longer than 32 weeks, underwent three four-week cycles of STI, separated by six-month periods back on their drugs. A rebound in plasma viral load was detected in all cases. At the second stop, in four of the nine patients, HIV RNA initially rebounded to similar levels prior to HAART, but then spontaneously dropped somewhat thereafter. These four patients developed laboratory evidence of strong HIV-specific immune responses. After the first, second, and third, stop, no mutations associated with resistance to NRTIs or PIs were detected. After the third stop, a spontaneous drop in HIV RNA and improvements in HIV-specific immune responses were detected in three of seven evaluable patients. Also, the HIV RNA set-point, or the level at which the HIV RNA stabilized, was lower in four of seven people after the third stop, compared to their pre-HAART viral loads. However, the CD4 cell counts decreased after the third stop to the level they were prior to starting antiretroviral drugs (Abstract LB 11).
These preliminary studies, based a very small number of people, show that STI may induce effective immune responses against HIV-1 antigens in some people, which may result in a decrease in plasma HIV RNA loads. Whether these potential "benefits" of STI will offset the risk of a significant decrease in CD4 cell counts and result in a net clinical benefit remains to be determined by larger, ongoing studies, with longer follow-up.
Steven Deeks from San Francisco reported on a study of STIs among 18 patients experiencing virologic failure on PI-based therapy. Eligible people had virologic failure on a PI for more than 12 months and an HIV RNA load above 2,500. Studies included measurement of viral replicative fitness, which compares the strength of the drug-resistant virus to non-drug-resistant ("wild type") virus. Interrupting therapy for 12 weeks was associated with an average decrease of 94 CD4 cells and an increase in HIV RNA load of almost 10 times (0.82 log). However, virus in the blood in 16 of 18 people reverted to non-PI-resistant virus. NRTI resistance persisted, but often at much reduced levels. Replicative fitness (compared to a wild-type virus) increased from about 22% to 67%. However, PI-resistant virus was cultured from blood cells 12 to 36 weeks after therapy discontinuation in four of eight patients who had tests showing loss of PI-resistance in the blood. This demonstrates that the drug-resistant virus was still around in at least half of the people in this small trial (Abstract LB10).
Dr. Deeks stated that this study suggests antiretroviral therapy may be associated with continued immunologic and virologic benefit, despite high-level drug resistance and significant HIV RNA loads. This benefit may reflect, in part, the maintenance of a "less fit" or less aggressive virus. Therefore, it may be better to stay on a "failing" regimen than to be on no drugs at all. Thus, the safety and efficacy of drug holidays as a therapeutic strategy remains to be determined, and may result in immunologic deterioration.
Dr. Calvin Cohen presented a preliminary analysis of a study (VIRA3001) that asked providers to choose a therapy for people who had developed resistance to their first, PI-containing, therapy. Study participants were randomized into two groups, with the first group's providers given the results of a phenotypic resistance test, and the second group's providers not given such information. The providers who used the phenotypic test placed people on at least three active drugs 77% of the time, compared to the second group, who were so placed 68% of the time. The group who received the phenotype resistance test had a median viral load drop of 1.27 log, compared to a median viral load drop of 0.75 log for the second group. This study suggests that providers might benefit from the additional information provided by this test (Abstract 237).
Does Drug-resistant HIV Reseed the Latent Resting Pool?
A very interesting study isolated HIV DNA, and also cultured HIV from the latent pool of HIV-infected lymphocytes, from lymph nodes of people who had good virologic suppression on HAART (Abstract 238). Researchers looked for evidence of drug resistance in the latently infected cells. This is of potential clinical significance, since it is thought that these cells are the possible source of HIV production when HAART is stopped. They found significant evidence of drug resistance in the HIV and DNA from people who had treatment prior to their current regimen, and had only very small elevations in HIV RNA, or "blips," while on their current regimen therapy. It is generally thought that this pool of infected cells is established very early in HIV infection, and would not be seeded by drug-resistant virus later in infection, so these findings are of some concern. Whether or not this will be clinically significant remains to be seen.
Where Does the Virus Come Back From when HAART is Stopped?
Dr. Chun from the National Institutes of Health (NIH) presented a study comparing the virus that is present in lymph node reservoirs to that which reappears in the blood when HAART is stopped. Based on genetic studies, in seven of nine patients the virus that rebounded in the blood appeared to be genetically different from the residual virus in the lymph nodes. This raises the possibility that there are other reservoirs in the body from which viral rebound can occur, such as the brain or genital tract. This report is one of the first that demonstrates that this may be a clinically significant concern (Abstract 239).
Interleukin-2 May Not Add Much
Dr. Stellbrink presented the results of a randomized trial of Interleukin-2 (IL-2) studying its effect on the latent HIV-infected reservoirs, and rates of HIV rebound when antiretroviral therapy is stopped. While the group receiving IL-2 had a slightly greater increase in CD4 cell counts, there was no significant difference in decreases in HIV RNA load. When HAART was stopped, after eight weeks, five of five people who did not receive IL-2, and five of seven people who did, had prompt rebound of their viral loads to levels similar to their pre-HAART loads. These results raise serious questions about the benefit of adding IL-2 to HAART (Abstract 240).
A New Class of Drugs: Entry Inhibitors
An excellent review of the process of HIV entry into the cell, which results in HIV infection, was presented by Peter Kim (Session 30); in addition, a symposium reported on the current state of trials of entry inhibitors (Session 47). There are three critical events that need to occur for HIV to enter a cell and infect it:
- First, the HIV particle attaches to the cell, coupling the gp120 protein of the virus with the CD4 cell surface receptor.
- Second, the gp120 HIV protein binds to one of two cell surface receptors, CCR5 or CXCR4 (called "chemokine receptors").
- Third, a change in the shape of another HIV protein, gp41, results in fusion of the HIV membrane with the cell membrane and injection of the HIV genetic material into the cell.
Two new drugs, T-20 and T-1249, being developed by Trimeris appear to inhibit the third step, fusion, by blocking the necessary shape change of gp41. T-20 is furthest in development and has shown very impressive decreases in HIV RNA load in people who have been treated with many antiretroviral drugs. Both of these compounds are small proteins, and must be administered by injection under the skin (subcutaneously). T-20 is currently being given twice a day, but T-1249 may need to be administered only once a day. Both drugs so far appear to have very few side effects other than some mild inflammation at the injection site. After up to 32 weeks of T-20 administration, with or without HAART, impressive decreases in HIV RNA loads are being observed in people with multiple drug resistance. Trimeris is now developing manufacturing capability to produce larger quantities of T-20 to begin larger trials in the middle of this year.
Current FDA-Approved Drugs
Combivir (ZDV/3TC) Plus Nelfinavir or Nevirapine in HIV-Infected, Naive Patients
This randomized multicenter study (the COMBINE study) compared a regimen of zidovudine (ZDV) and Epivir (3TC) plus either nevirapine (Viramune, an NNRTI) or nelfinavir (Viracept, a PI) at 1,250mg twice a day in HIV-infected, naive patients. After 6 months, the group taking the NNRTI nevirapine had better virologic suppression -- 74% had HIV RNA loads below 200, compared to 60% for the people taking the PI. Considering only the patients who did not drop out of the study, the nevirapine group had 91% with HIV RNA loads below 200, versus 82% for the nelfinavir group. Like a prior study that compared another NNRTI, efavirenz (Sustiva), to a PI, indinavir (Crixivan), this study confirms the potential equivalency, or even superiority, of an NNRTI compared to a PI in people without prior therapy. There is some concern that in people with high viral loads (above 100,000) an NNRTI might not provide as good viral suppression, but the authors did not report the results based on baseline HIV RNA (Abstract 510).
Ritonavir Intensification in Indinavir Recipients
This study of people with detectable levels of HIV on Indinavir (IDV) regimens asked whether adding ritonavir (RTV) would enhance antiretroviral potency and viral suppression, even in patients with reduced IDV susceptibility. Patients on therapy of IDV three times a day plus two NRTIs, with viral loads between 50 and 50,000, switched to combination regimen of RTV/IDV 400mg/400mg by dose titration of RTV (200mg for two days, then 300mg for three days). Of the 17 people who stayed on the study for the full 16 weeks, 59% had HIV RNA loads below 400, and 53% had HIV RNA loads below 50. IDV minimal blood levels increased 272%, peak levels decreased, and total daily levels remained similar. The study confirms that IDV blood levels were substantially improved in patients receiving RTV/IDV, 400mg/400mg twice a day, versus IDV at 800mg three times a day and that adding RTV permitted twice daily dosing of IDV without fasting. Short-term virologic and immunologic improvements were seen in a majority of patients, likely due to the improvement of IDV blood levels. However, adding RTV increased cholesterol and triglycerides in seven of 27 people in this study (Abstract 534).
People who had failed one prior PI, and never received an NNRTI, were randomized to a blinded ABT-378/ ritonavir dose (400mg/100mg compared to 400mg/200mg twice a day) plus nevirapine (Viramune) and two NRTIs (at least one new). After 48 weeks, 49 of 58 people (84%) on treatment had viral loads below 400 (70% based on intent-to-treat). Of significant interest is that seven of the 11 patients who had more than a four-fold reduced susceptibility to ABT-378 at baseline had a viral load below 400 at week 48. The blood levels achieved with ABT-378 are very high and may even suppress HIV in people with some resistance to ABT-378. In general, ABT-378 was relatively well tolerated, except for some moderate nausea and diarrhea (Abstract 532).
ABT-378/Ritonavir in Antiretroviral-naive Patients
Trip Gulick from New York reported results of a dosing trial of ABT-378, which contains a small amount of ritonavir and results in very high levels of ABT-378 in the blood, providing a potentially very high barrier to drug resistance. Two groups of antiretroviral-naive patients received d4T/3TC with a randomly assigned, blinded ABT-378/ritonavir dose taken twice a day. Group I received 200mg/100mg, and Group II received 400mg/100mg. At week 48, 83% to 86% of patients on treatment had HIV RNA loads below 50. After 72 weeks, 25 of 27 patients (93%) in Group I, and 45 of 45 patients (100%) in Group II had HIV RNA loads below 400. All doses have been well tolerated to date, with only one discontinuation through 72 weeks. As other studies report, even low doses of ritonavir cause elevations in cholesterol and triglycerides. Certainly, ABT-378 is a potent PI, and whether it will be used as first line or second line therapy remains to be seen (Abstract 515).
FTC Plus ddI Plus Efavirenz in Treatment-naive Patients
A study from France (ANRS 091) reported on a 24-week trial of ddI plus efavirenz in a regimen that included a new NRTI: FTC. This was a 24-week pilot study in 40 antiretroviral-naive adults. All three drugs were taken once a day at bedtime, with doses of FTC at 200mg, ddl at 400mg if body weight exceeded 60kg (about 132 pounds) or 250mg if body weight was below 60kg, and efavirenz (Sustiva) at 600mg. At week 12, 39 of 40 people had a viral load below 400, and 20 of 40 had a viral load less than 20. The average increase in CD4 count at week 12 was 120 cells. At week 24, 24 of 28 patients achieved an HIV RNA load of less than 20. This preliminary report showed that FTC plus ddI plus efavirenz is a potentially potent, and well-tolerated, once-a-day HAART regimen (Abstract 518).