Change on the Horizon

While the never-skip-a-dose wisdom still holds, in a few settings researchers and doctors today are closely supervising HIV positive subjects who have halted their antiretroviral therapy. The theory is that alternating periods of HAART treatment with regulated withdrawals of drug therapy may serve as a means of inducing immune system control of HIV.

Specifically, it is hoped that viral load increases seen during periods of HAART cessation will trigger the proliferation of HIV-specific, naive CD4 cells that are present during acute (initial) HIV infection and may become undetectable, particularly in the presence of maximally suppressive HAART. A sufficient CD4 cell response facilitates vigorous cytotoxic T-lymphocyte (CTL) activity, which is crucial for establishing and maintaining viral suppression without drugs. Treatment interruptions have also allowed researchers to examine the effect of shifts in the dominant HIV quasispecies (strain) in persons whose HIV has become resistant to many drugs due to genetic mutations.

Background

This experimental approach to anti-HIV drug therapy derives in large part from the problems of drug adherence and resistance. Although some drug regimens have been simplified and a few drugs, such as ddI (Videx), have been reformulated to reduce their pill burden, most people taking HAART find staying 100% adherent to their combination therapy to be difficult. In most cases, the burdensome number of pills required, with their strict dosing and dietary requirements, and a host of side effects make perfect adherence a major challenge. As a result, many people stop therapy on their own, either for short, inconsistently scheduled intervals often referred to as "drug holidays" or for long periods of time.

Perhaps inevitably, the issue of nonadherence has crossed over into the realm of research, and the informal "drug holidays" taken by many people have been reconfigured into closely monitored experiments of structured, or strategic, treatment interruption (STI). Close monitoring--especially frequent (often weekly or monthly) testing and analysis of virologic and immunologic markers--is a defining characteristic of STI research and an essential safety component of any protocol involving treatment interruptions.

The concept of STI was first described by Franco Lori, MD, of the Research Institute for Genetic and Human Therapy (RIGHT) in Washington, DC, and Pavia, Italy. Dr. Lori treated an individual, now widely known as the "Berlin patient," who began taking HAART (indinavir [Crixivan]/ddI/hydroxyurea [HU]) during acute HIV infection (day 57) with a baseline viral load of approximately 85,000 copies/mL. Although this regimen lowered the 29-year-old man's viral load below 500 copies/mL, he stopped all drugs on his own after two weeks due to epididymitis (a testicular infection).

His virus rebounded above 5,000 copies/mL, but was again suppressed when he resumed antiretroviral therapy a week later. This individual suffered acute hepatitis A infection four months later and halted his anti-HIV drugs. Although he restarted HAART after this second, two-week break, his virus did not rebound during the drug-free period. Less than two months later (day 176), he discontinued therapy altogether. Remarkably, his viral load has remained undetectable without any anti-HIV therapy for more than three years, and vigorous HIV-specific CTL responses appear to be ongoing. (See also "Conference Coverage" in BETA, Summer 1999.)

Dr. Lori first reported on this case at the 5th Conference on Retroviruses and Opportunistic Infections (CROI) in February 1998. While the medical establishment greeted this report with understandable skepticism, Dr. Lori's "discovery" sparked considerable interest among persons taking anti-HIV therapy. The possibility of containing HIV without a steady diet of drugs seemed almost too good to be true, but hardly any data supported the theory.

Reports involving STI, however, began to surface anecdotally and at scientific meetings. Veronica Miller, PhD, of the Klinikum der J.W. Goethe Universität in Frankfurt, Germany, created a stir at meetings in Glasgow (November 1998) and Chicago (February 1999) when she reported that STI led to the reemergence of wild-type (nonmutated) virus in some persons with chronic HIV disease who were not responding to many of the approved anti-HIV drugs.

Encouraged that STI might be more than a fantasy and indeed have several applications, the Foundation for AIDS and Immune Research (FAIR), Project Inform, and Treatment Action Group (TAG) convened an STI workshop July 30-August 1, 1999, in Newton, MA. At this gathering, researchers and advocates discussed the potential uses of STI, various study designs for researching the efficacy of treatment interruptions, as well as the need to establish an observational database and clinical practice guidelines. Since then, interest in STI therapy has mushroomed and an increasing number of studies have been initiated. At least 22 presentations concerning treatment interruptions and STI were given at the 7th CROI in San Francisco in early February 2000, by which time news of this experimental treatment strategy was appearing in the popular press.

Target Treatment Populations

At present, researchers are studying the utility of STI in three different populations: persons with primary, or early, HIV disease whose virus has been fully suppressed with HAART; persons with chronic HIV disease whose virus has been fully suppressed with HAART; and persons with chronic HIV disease who display resistance to several anti-HIV medications and whose only treatment option is so-called mega-HAART (normally consisting of more than five different drugs). Although a few reports have shown STI to be beneficial, it is important to remember that many studies of treatment interruptions to date have not yielded encouraging results.

Primary HIV Infection

Since the concept of STI derived from the "Berlin patient's" experience, it was initially believed that only those persons who had started HAART during acute HIV infection would benefit from STI. If the host immune system is still relatively intact, as during acute infection, it seemed plausible that a cycle of treatment interruptions might provide enough exposure to HIV antigens, which are lost when HAART suppresses the virus below extremely low limits of quantification, to reestablish effective CTL control over viral replication.

A glimpse at what might be achieved using STI in this population was presented as a late breaker at the 37th Annual Meeting of the Infectious Diseases Society of America (IDSA), held November 18-21, 1999, in Philadelphia. Eric Rosenberg, MD, of Brigham and Women's Hospital in Boston and colleagues first observed a group of 25 persons (8% female, 12% African-American, and 12% Latino) who started HAART during primary HIV infection, before their bodies had produced antibodies to the virus.

By analyzing each participant's CD4 cell responses to HIV antigens in vitro, Dr. Rosenberg's team found that initiating antiretroviral therapy during the period of HIV seroconversion resulted in a greatly enhanced CD4 lymphoproliferative response (LPR) against the virus. (LPR refers to lymphocyte growth when HIV antigens are added to them in the laboratory.) A strong LPR is considered desirable, as it is implicated in the ability of long-term nonprogressors (LTNPs) to keep their viral load persistently low and their CD4 cell counts stable. (LTNPs are a small subset of HIV-infected individuals who, without any anti-HIV medication, show no signs of HIV-related symptoms or illness for a period of at least ten years.)

Two persons from the group of 25 then willingly submitted to treatment interruptions to test whether their strong, HAART-induced LPR could adequately suppress the virus without drugs. The first participant, a 32-year-old man with an undetectable HIV RNA viral load, stopped HAART for approximately three weeks. His plasma viral load rebounded to 17,000 copies/mL, while his LPR rate, measured in terms of a stimulation index (SI), surpassed 100. (A high SI rate corresponds to increased CD4 reactivity to HIV; the lowest significant measurement is 10.)

At this point, he went back on HAART and his viral load plummeted to an undetectable level. He again stopped therapy approximately four months later due to acute hepatitis A infection. Two weeks after beginning this second STI, his viral load increased to 37,000 copies/mL but three days later began a steady decline to 400 copies/mL without the aid of medication. This individual's SI reading during the second treatment interruption was greater than 800, a very high level for an HIV positive person. The LPRs in this individual were directed toward surface, regulatory, and core HIV proteins as well as the reverse transcriptase enzyme. This broad-based CD4 cell response was accompanied by ever greater increases in CTL activity as the STI cycles progressed (this participant later underwent a third STI).

The second subject began with a baseline LPR SI level measuring approximately 20. He went on a three-week STI, causing his viral load to increase to 120,000 copies/mL and his SI to rise to 130. After he stopped therapy a second time, his viral load measurements were quite different: over a four-month, drug-free period, his viral load never climbed above 5,000 copies/mL and was at times undetectable. This implies that he achieved a lower viral set-point, or the point at which viral load appears to level off. CTL activity also increased. Significantly, no drug-resistant mutations were detected.

Some favorable trends emerged in this study: viral load increases corresponded with increases in HIV-specific LPRs, which presumably augmented viral suppression during successive STIs. In addition, neither participant experienced the adverse effect of detectable drug resistance; each application of HAART repeatedly and effectively suppressed viral load levels after various durations of hiatus. However, a cohort of two persons is regrettably small; the results yielded from this study therefore might not be applicable to other HIV positive people.

A similar report by Dr. Lori, Julianna Lisziewicz, PhD (Lori's co-director at RIGHT), and colleagues appeared in the January 22, 2000 issue of the Lancet. Three persons with acute HIV infection were given a HAART regimen containing hydroxyurea for three weeks followed by a one-week STI in imitation of the "Berlin patient." Drug therapy was restarted and again halted in a controlled setting up to five different times; HAART was resumed when viral load reached 5,000 copies/mL.

One of the three subjects was not able to control HIV replication using this approach; his HIV RNA levels rebounded far above the 5,000 copies/mL mark each time he stopped therapy (except for the first time, when he elected to resume HAART after a 0.118 log, or 1.3-fold, increase). The second subject stopped treatment only twice. During the second, six-month STI he was able to maintain a viral load of approximately 2,000 copies/mL, nearly 1 log (10-fold) lower than baseline. When his viral load subsequently climbed above 5,000 copies/mL and the same HAART regimen was restarted, this individual's viral load decreased to undetectable levels.

The third subject, a Washington, DC, resident later dubbed the "Washington patient," displayed an especially favorable reaction to STI. The time to viral rebound above 5,000 copies/mL increased in this individual after each successive treatment interruption; his daily viral load increases dropped from a high of 0.264 log (1.8-fold) per day after the second interruption to 0.036 log (1-fold) per day after the fifth. In addition, his viral load remained 2-3 logs lower than baseline values--as low as subject two's viral load--for nearly five months during the final STI, suggesting a lower viral set-point.

Based on the results of this study, it appears that a hydroxyurea-containing regimen (possibly with a protease inhibitor [PI]) may in some cases allow a person to stop drug treatment several times with increasingly longer duration until viral breakthrough. That the reintroduction of antiretroviral therapy in all cases swiftly reduced viremia to below the limit of detection suggests that inducing drug resistance via STI may occur less readily than feared, although the durability of this effect is unknown. (It may be that HAART maintains clinical health despite an increase in the amount of virus in the plasma. See below for information on the likely immunologic advantages of PI-based regimens.)

The three-person Lori-Lisziewicz study in many ways mirrors that of Dr. Rosenberg. Using HIV as a sort of "booster shot" to reinvigorate the immune system is an intriguing hypothesis that seems to be borne out in at least some persons. Nevertheless, much larger cohorts will need to be monitored and trials will have to be randomized to better understand the full, long-term ramifications of STI in people with acute HIV infection. If it does eventually prove beneficial, STI may be used in this population to offer at least some people true holidays from drug therapy.

Chronic HIV Infection with Viral Suppression

Drs. Lori and Lisziewicz also found STI beneficial in some persons with chronic disease whose HIV, and immune systems, have been suppressed with HAART. Before Drs. Lori and Lisziewicz presented their study as a poster at this year's CROI, very little data supported the use of STI in this population. These two researchers believe that, like a vaccination, starting and stopping drug therapy may stimulate the immune system, allowing it to recover despite chronic infection and revert to its fully functioning, normal state to control HIV and postpone disease progression.

The RIGHT researchers' match-controlled experiment, known as the PANDA study, featured two groups. One nine-person group, named the PANDA cohort, had been taking only hydroxyurea and ddI; the term PANDA refers to the rarity of finding a cohort in the developed world consistently following a two-drug anti-HIV regimen. (This group originally had ten members, who first underwent an eight-week STI in an unmatched study done by RIGHT. When the group showed a rise in viremia [plasma virus] at week six followed by a decrease at week eight, the researchers decided to initiate the match-controlled study. One PANDA member moved to Brazil before the second study began.)

The other, non-PANDA group was treated with standard-of-care, PI-based HAART containing at least three drugs. Five of the eight persons in this treatment arm took d4T (Zerit)/3TC (Epivir)/saquinavir (Fortovase)/ritonavir (Norvir); two took d4T/3TC/indinavir (Crixivan); and one person took AZT (Retrovir)/3TC/saquinavir/
ritonavir.

The subjects in both cohorts entered the study with similar CD4 and CD8 cell levels (approximately 550 cells/mm3 and 850 cells/mm3, respectively). Yet even though both groups had been taking anti-HIV drugs for more than two years, the PANDA cohort had an average viral load of 549 copies/mL--low but not undetectable--while the majority in the HAART group (6 of 8) had a viral load below 50 copies/mL. The STI was planned to continue for eight weeks; antiretroviral therapy was to be restarted if a participant's viral load increased beyond 10,000 copies/mL or CD4 cell count dropped to fewer than 200 cells/mm3.

The results were fairly startling. By week 6 of the treatment interruption, most of the HAART subjects (5 of 8) were required to resume drug therapy. None of the PANDA subjects had to abandon the eight-week STI at any time. By all measurements, the PANDA group received the most benefit from stopping drug therapy. The average viral load in the PANDA group increased to only 1,896 copies/mL by week eight, while the average viral load among the HAART subjects increased to 16,863 copies/mL by the sixth week. CD4/CD8 cell ratios, another measure of immune system status, fell below 0.3 among HAART subjects by week 6, whereas they remained fairly stable at about 0.7 in the PANDA group; a measure of approximately 2 is seen in healthy individuals.

CD4 cell counts likewise did not fall significantly in the PANDA group. Among HAART subjects, however, they dropped by approximately 200 cells/mm3. Similarly, CD8 cell levels rose in the HAART group but remained consistent among the PANDAs.

Reintroduction of anti-HIV therapy triggered a prompt reversal of viremia values to their respective pretreatment levels in both groups. CD4 and CD8 cell measurements generally returned to their pretreatment levels in the HAART group, while they were simply maintained in the PANDAs.

The excellent response to STI among the PANDAs suggests that persons with chronic HIV infection may achieve positive results with treatment interruptions. (Had this trial been randomized and the PANDA cohort much larger, the data would have been more persuasive.) In speaking with BETA, Drs. Lori and Lisziewicz explained that the low but detectable levels of HIV RNA in the PANDA group apparently kept the immune system "alert" and better able to control the virus. Developing resistance to ddI was not an issue, since, according to Dr. Lori, the unique mechanism of hydroxyurea compensates for ddI resistance.

When asked to estimate how many STIs may be necessary for optimal viral control by the immune system alone, Dr. Lori responded that the answer is not known. "It is important to remember," he added, "that success can be measured in other ways. For example, if a person [using STI] experiences less toxicity, needs only half his medication, which will cost half as much, etc., that can be considered an advantage."

Chronic HIV Infection without Viral Suppression

Dr. Veronica Miller has been fueling the surge of interest in STI by suggesting that it may help chronically infected persons with inadequate viral suppression.

At the Second International Workshop on Salvage Therapy for HIV Infection, held May 19-21, 1999, in Toronto, Dr. Miller presented interim data on the Frankfurt HIV cohort. This group of 39 antiretroviral-experienced, highly drug-resistant persons had undergone a treatment interruption lasting at least two months between 1996 and 1999 after being on a minimum of two PI-containing regimens for at least a year. Virus from all 39 subjects displayed genotypic resistance to an average of eight drugs at baseline. The STI was initiated not to prime the immune system as in the studies mentioned above, but as a brief respite from drug therapy before starting a regimen of mega-HAART.

As in her previous findings, Dr. Miller reported that a majority (in this case, 26 of 39, or 66%) of subjects experienced a total reversion to wild-type virus as the dominant strain during the STI, as measured by genotypic and phenotypic resistance analyses. Viral loads increased by a median of 0.71 log copies/mL, or 5-fold (i.e., 0.98 log copies/mL [9-fold] in those who shifted to wild type, but only 0.34 log copies/mL [2-fold] in those who did not).

Interestingly, baseline CD4 cell counts were higher among those whose virus had shifted (a median of 180 cells/mm3) compared with those whose virus had not (a median of 50 cells/mm3). Persons with "shifty" viruses (Dr. Miller's term) apparently had less severe HIV disease; it has not been determined whether a CD4 cell level exists below which shifting to wild type will not occur. A potentially troublesome outcome of this experiment was that CD4 cell levels dropped among all subjects by a median of 89 cells/mm3--increasing the risk of contracting opportunistic diseases or infections (OIs)--and were slow to return to baseline levels upon reintroduction of therapy. Steeper declines seen in the wild-type "shifters" may account for this slower rate of CD4 cell replacement.

One benefit of reverting to wild type was made clear when mega-HAART was given to all 39 participants following the STI in Dr. Miller's study. Plasma viral load fell an average of 1.02 log copies/mL (10-fold) among the nonshifters, yet those with wild type experienced an impressive 2.8 log copies/mL (630-fold) drop. In addition, the shifters' response was swift. Seventy-two percent of those who shifted had viral loads below 500 copies/mL after only eight weeks, compared with a mere 10% among the nonshifters.

Several reasons might explain why those who reverted to wild type displayed such results. It may be, for instance, that the drug-resistant virus of those who shifted was somehow less able to reproduce than that of the nonshifters. Some researchers have found this to be true (see the following paragraph); nevertheless, drug-resistant virus can be surprisingly durable, as in the case of strains resistant to non-nucleoside reverse transcriptase inhibitors (NNRTIs).

Alternatively, the shifters may possess an immunologic trait, such as more robust HIV-specific responses, that more readily transposes mutant virus back into wild type without the selective pressure of anti-HIV drugs. In any case, more pronounced fluctuations in viral load and CD4 cell levels are associated with wild-type virus during STI cycles in two small studies. Blunting precipitous rises in viral load and CD4 cell count drops will no doubt be a priority of future STI studies in this population.

In a poster presentation at this year's CROI, Steven Deeks, MD, and colleagues from the University of California, San Francisco (UCSF) reported an even higher percentage of shifters (88%, or 16 of 18) during a 12-week STI. The purpose of this small study was to examine CD4 cell responses during virologic "failure" (inability to suppress the virus). As in the Frankfurt cohort, Dr. Deeks's subjects were all heavily pretreated candidates for salvage therapy, although this study was also randomized to include a group who continued taking a partially suppressive PI-based regimen.

Of those undergoing STI, the time to wild-type reversion, though abrupt in each case, ranged from 2 to 15 weeks based on a follow-up analysis. Because Dr. Deeks's team examined so many variables, a more detailed--and not entirely agreeable--picture of STI emerged. For example, the PI-susceptible wild-type virus still had some NRTI resistance in seven of the subjects. In addition, drug-resistant virus identical to the baseline variety was cultured from peripheral blood mononuclear cells, or PBMCs (i.e., lymphocytes and monocytes) in four of eight persons measured.

Since drug-resistant virus was not eradicated but rather was "outgrown" by wild type, drug-resistant virus could possibly reestablish itself as the dominant strain after restarting therapy. (For more information on residual viral replication in the context of STI, see "HIV Persists despite HAART.") The wild-type virus, however, replicated more efficiently (a median 67.1% replicative fitness) than the drug-resistant form (22.3% replicative fitness). (Wild-type virus in a person who has never taken drug therapy displays 100% replicative fitness.) Whether a presumably more virulent wild-type virus is preferable to a strain with less replicative capacity remains to be determined.

Not surprisingly, reversion to wild type was associated in this study not only with a sharp viral rebound (0.82 log copies/mL [6-fold] median gain) but also with a decline in CD4 cells (a median of 94 cells/mm3) statistically equal to the decline seen in the Frankfurt cohort.

This study illustrates some of the risks and unanswered questions related to strategic treatment interruptions, which unfortunately could not be elucidated with such a small sample population and within the short time parameters of the evaluation. It also would have been worthwhile to view the results--and to compare them with Dr. Miller's data--had drug therapy been restarted in this cohort and another STI initiated.

Analysis

Despite widespread interest in STI and some reports of progress, the theory remains unproven and potentially dangerous, given the risk of developing life-threatening OIs. No definitive statements can be made until the shortcomings of STI research are addressed. For example, all the studies mentioned above suffer from their extremely small size. Many researchers who have done retrospective analyses, like Dr. Miller, are constrained by the limited number of persons who have followed STI in a clinical setting thus far (after all, Dr. Lori introduced the concept of STI a mere two years ago).

These same studies are further hampered by their short follow-up and design flaws. Retrospective analyses by their very nature produce less focused and less useful data than do randomized clinical trials. Trials that are neither randomized nor properly controlled--though sometimes offering tantalizing results--can rarely be of more than anecdotal value.

One late-breaker presentation at this year's CROI demonstrated some of the limitations of current STI research. Lead author Xia Jin, MD, and colleagues from the Aaron Diamond AIDS Research Center in New York introduced the possibility that STI might be used as a component of anti-HIV therapy that includes not only HAART but also a vaccine.

In this study, four persons were given a series of therapeutic vaccinations with ALVAC 1452, a recombinant (genetically altered) canarypox vaccine, and recombinant gp160 after taking a HAART regimen that included AZT, 3TC, and indinavir for approximately 2.5 years (i.e., since their period of acute infection). After the last vaccine injection, all four participants chose to discontinue HAART. Two subjects witnessed a rapid rise in HIV RNA, with replication doubling times of 13 and 23 days--much like the rates seen when HAART is withdrawn in unvaccinated, virally suppressed persons.

The other two experienced a delay in viral rebound; their viral load became detectable 68 and 85 days, respectively, after HAART was stopped. At the time of the presentation, the study authors noted that these two "delayed rebounders" had been off antiretroviral therapy for up to eight months, and their respective viral loads were measured as 6,309 (3.8 log) and 316 (2.5 log) copies/mL. These measurements were lower than their baseline pre-HAART levels, indicating a possible lowering of the viral set-point, much like a handful of other subjects in STI-related studies. In the two persons with delayed viral rebound, Dr. Jin's team also noted increases in their CTL responses to several HIV proteins (the other subjects showed almost no HIV-specific CTL activity).

Yet despite the promising data related to the "delayed rebounders," this study was undermined by a number of drawbacks, including a tiny sample population, short-term follow-up, and the absence of an unvaccinated control group. Since similarly promising data have been gathered among subjects in other STI experiments without a vaccine component, it is questionable what role the vaccine played in this study.

Though certainly not unique to STI studies, inconclusive data are a recurring problem. Significantly, there is no consensus on how long an STI should be; some researchers have been setting arbitrary time limits, while others choose to let an STI continue indefinitely until virologic boundaries have been crossed. It is no wonder that STI studies often seem to contradict one another. For example, despite a few intriguing reports from the 7th CROI mentioned in this article, numerous studies presented at the same conference found no benefit to interrupting HAART treatment. (Abstracts and posters from the 7th CROI can be viewed by visiting www.retroconference.org.)

Limited studies notwithstanding, the promise of STI--including an improved quality of life for persons taking HAART--appears to warrant further
investigation. Optimism based on what little is known, however, should not obscure the significant dangers of stopping antiretroviral therapy. If STI causes the depletion of CD4 cells and delays their return, it may in the end prove more harmful than beneficial.

Similarly, a return to pathogenic (disease-causing) wild-type virus is likely desirable only if the ability to suppress the virus is restored. STI-induced drug resistance, though not detected as often as expected, remains a serious risk. Different drugs exhibit different half-lives; if a combination regimen is suddenly halted, the virus may mutate in the presence of suboptimal drug levels and become less sensitive to some medications--particularly NNRTIs. In another scenario, drug-resistant HIV lying dormant in resting lymphocytes may suddenly be activated and begin replicating if those cells are ever called to respond to an antigen. The issue of dormant HIV should not be overlooked. Some experts believe that reseeding latent reservoirs of replication-competent HIV with virus that proliferates during STI could ultimately cause considerable damage (see "HIV Persists Despite HAART").

Conclusion

It is far too early to say whether or not STI will ever be in widespread clinical use. As with any highly experimental strategy, a certain amount of time must pass before enough data can be amassed to inform treatment decisions. As Dr. Lisziewicz recently pointed out, scientists are still "searching for the right [STI] recipe." Only large, well-controlled, and well-designed studies can move that search forward. One thing is clear: STI is not a panacea, nor is it likely to be an advantageous treatment option for many people. It may take years to firmly establish the appropriate durations and successions of STI as well as the profile of persons who will benefit from treatment interruptions, if they ever prove useful.

Atzori, C. and others. Indinavir and other HIV protease inhibitors decrease Pneumocystis carinii in vitro growth. 7th Conference on Retroviruses and Opportunistic Infections. San Francisco. January 30-February 2, 2000. Abstract 245.

Deeks, S.G. and others. Delayed immunologic deterioration among patients who virologically fail protease inhibitor-based therapy. 7th CROI. Abstract 236.

Deeks, S.G. and others. Virologic and immunologic evaluation of structured treatment interruptions (STI) in patients experiencing long-term virologic failure. 7th CROI. Abstract LB10.

Jin, X. and others. Discontinuation of HAART after a course of therapeutic vaccination with ALVAC 1452 and rgp160 may be associated with delayed viral rebound kinetics. 7th CROI. Abstract LB12.

Kilby, J.M. and others. Significant delay in plasma vRNA rebound during a scheduled treatment interruption in HIV-1 chronically infected patients previously on effective therapy. 7th CROI. Abstract 359.

Lori, F. and Lisziewicz, J. Personal communication. February 1, 2000.

Lori, F. and others. Control of viremia after structured treatment interruptions. 7th CROI. Abstract and poster 352.

Lori, F. and others. Intermittent drug therapy increases the time to HIV rebound in humans and induces the control of SIV after treatment interruption in monkeys. 6th CROI. Chicago. January 31-February 4, 1999. Abstract LB6.

Lori, F. and others. Structured treatment interruptions to control HIV-1 infection. The Lancet 355(9200): 287-288. January 22, 2000.

Miller, V. and others. Association of viral load, CD4 cell count, and treatment with clinical progression in HIV patients with very low CD4 cell counts: the EuroSIDA cohort. 7th CROI. Abstract 454.

Miller, V. and others. Mega-HAART, resistance and drug holidays. 2nd International Workshop on Salvage Therapy for HIV Infection. Toronto. May 19-21, 1999. Abstract 030.

Rosenberg, E.S. and others. Generation and maintenance of HIV-1 specific T helper cell responses in persons treated during acute HIV-1 infection and augmentation of these responses following structured treatment interruption. 37th Annual Meeting of the Infectious Diseases Society of America. Philadelphia. November 18-21, 1999. Abstract LB725.