Skip to content

The Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) is the annual meeting of the American Society for Microbiology. In 2004, it happened in Washington DC from 30 Oct to 2 Nov. Some new drugs were presented along with data on new strategy initiatives. Here, a quick roundup.

Reverset (RVT) was presented by Rob Murphy (Northwestern, Chicago) (abstr H-1130). RVT is a D enantiomer (‘mirror image’) of an earlier drug, d4FC, and has an intracellular half-life of some 17 hours. Murphy only explained the QD treatment-experienced group, 8 people at 200mg. There were 4 other groups of 8 (at varying dosages, supposedly to be presented at some future date?), including a placebo group. At baseline, 5/8 had the 184, 4/8 had 3 TAMs (RVT is a cytosine analog like 3TC and FTC). No one had the K65R and 2/8 had no mutations. On top of a ‘failing’ regimen, there was an average 0.8log drop at 10 days. It was not clarified that of the 4/8 who got <400, what baseline mutations they had. TDF might do a little less well. There was less of a viral load drop with those who had the 184. Headache was seen in 31% of the participants, a cold in 28%, neutropenia in 19%, hypertriglyceridemia in 13% (at 10 days).

Graeme Moyle presented the TIMS study (abstr H-1131), of TDF+Trizivir (n=57) vs EFV + CBV (n=56). At 48 weeks in this naïve trial, there was a high discontinuation rate, of 30%. In the quadruple arm, CD4s rose 165 (more than doubled!), cholesterol remained stable, while the triglycerides fell 0.3. The problem with the quadruple was the abacavir hypersensitivity, at 10.5%. In the triple arm, CD4s rose 120 (not statistically different from the quad arm), triglycerides rose slightly at 0.2 more, total cholesterol rose 25%. The ITT analysis of those below detection was not statistically dissimilar between the two arms, a single class quad vs a double class triple.

Victor Soriano presented what he called a paradoxical CD4 response to TDF+ddI combinations (abstr H-1132). His patients started at the high dose ddI (400 mg), then lowered it when the data became available, to 250 mg. The CD4 loss was greater at the higher dose, but even at the lower dose, the CD4 loss was clear. This despite complete viral suppression. Is it because ddI and TDF are both adenosines which cause an additive/synergistic cytostatic effect (ie, competing, thus limiting ATP production) or is it a mitochondrial issue (which also might limit replication of CD4s)?

Richard Nettles of Johns Hopkins did an analysis of blips (abstr H-1134). Are they important? Do they have meaning? Do certain populations have more of them? Are there clinical correlates? His analysis of some 10 people showed 9 with at least 2 blips, and that at least at 5% of the time, people are experiencing blips, of some 60 hours duration. Over time, 90% of the population will intermittently experience (regardless of demographic, clinical, or HAART variables) low-level detectable viremia in a setting of sustained suppression. They are most likely due to assay variation and are not clinically significant elevations of viral RNA (ie, no new resistance, so not ‘real’, based on frequent genotypic analyses).

Graeme Moyle presented another paper (H-1135) on the epidemiology and predictive factors for chemokine receptor use. He saw that in his patient population, R5 prevalence <100 CD4s was 60%, >100 CD4s, R5 prevalence was in 85% of people, with viral load, <5000 copies, 88% of people had R5 prevalence, while >100,000, 72% had R5. X4 was mainly seen in those with a viral RNA of >11,550, CD4s <155, and Natural Killers below 48.

Viral Load CD4 CD4%
R5 35,800 307 18.7
X4 66.228 237 15.4


CCR5 usage declines with absolute CD4 cell count, although the majority of patients with low CD4 count continue to have CCR5-using virus. Co-receptor usage is not influenced by viral clade or exposure to currently approved antiretrovirals. No change in co-receptor use was noted with treatment experience (vs Demarest, below)—it’s still the early days.

J Demarest (GSK) presented on co-receptor tropism (abstr H-1136). 7% of 442 samples were unphenotypable. He did not look at CD4 nadir or change in CD4, but in experienced patients, prior PI use was 4.16 times as likely to see a shift (p=.0085), age was a slight cofactor (1.05 times, p=.039), as was being non-white (experienced only) with a 5.13 times rate of change in tropism, (p=.0015). 82% of all samples tested were R5-tropic (67% exp, 88% naive), 16% R5X4 (28% exp, 12% naive), and <2% X4 (5% exp, 0% naive). Decreasing CD4/CD8 ratio was significantly associated with X4 tropism in naive subjects (p<0.0001).

Chris Petropoulis (of Virologic) talked about EI escape (abstr H-1137), the mechanism of which can be due to probably 4 things: competitiveness, non-competitiveness (allosterism)—which diminishes binding, a receptor switch, and receptor escape. Increases in the IC50 describe resistance. Poster 192.

Charles Hicks’ late breaker (abstr H-1137a) on Resist 1 (Tipranavir/r) was a 24 wk analysis. Resist 1 was a phase III, randomized, controlled, open-label, multicenter trial of TPV/r vs CPI/r in 620 people with more than one primary PI mutation, although less than 3 at primary positions. Expert advice on constructing OBR was taken advantage of 30% of the time (to build the Comparitor PI arm). Change in viral load > 1 log and CD4 from baseline were primary endpoints.

(Tipranavir, a ‘salvage’ drug, could be launched by the middle of 2005. The sponsor applied for U.S. and E.U. approval to sell and market tipranavir/r in October 2004. One day before a decision was to be made on priority review, they withdrew the application for ‘data formatting’ reasons, and re-submitted the next day (22 Dec 04), effectively re-setting the clock.)

The trial was 90% male, 20% black. Lopinavir was used by 60% of the CPI participants, T-20 by 36%. 8% of the patients had HCV, HBV in 11%. Baseline data – 123 CD4s, viral load was 4.8 logs, there were a median of 15 protease mutations, 12 prior ARVs, 4 prior PIs.

N TPV/r 313 CPI/r 317
Discon @ 8w 15 33
Discon @ 24w 48 139
Viral Failure 13 109
Adv Events 25 9
AEs 3/4 22.8% 18.1%
AE Discon 9.3% 5.2%
ALT Discon 6 1
Resp >1 log 41% 22%
RNA Change -0.88 -0.28
<400 34.7% 16.5%
w/T20 47%
<50 25.1% 10.3%
w/T20 22%
CD4s +36 +6
Lab ab’s 21.7 trigs 12.5
4.2 chol 0


More analysis from TORO: working the T-20 data. The median time to failure in the T-20 group (from the TORO-1 and 2 studies) was eight weeks. The investigators note, “our results reinforce the concept that regimens that do not maximise the chances of suppression negatively impact future chances of success.” These findings are likely to raise questions about whether it is appropriate to add less than 2 new drugs to people in salvage situations, and that salvage, once it is defined, will need to be based on at least 2 active agents. Both the Kaletra-as-new-therapy arm and the total number of new agents significantly added to chances of success beyond 8 weeks with T-20.

Steve Piscitelli talked about GSK’s entry inhibitor, 873140 (abstr H-1137b). Of the 40 subjects, 10% were female, 50% non-white. Exposure to 873140 is 30% better with a moderately fat meal. It “occupies” 97% of the R5 receptors at all doses, from 200mg QD to 600mg BID, (significance not clear, in-house test – not validated). It has prolonged binding – the viral load nadir was reached 2 days after stopping, something seen with other EI’s, although the time to debinding was dose-related (data not shown). Lots of side effects (from gr1 through gr 3) were reported. 600 mg BID is the MTD. So, although the occupancy doesn’t vary with dose, the response does (from -0.46 200 mg QD to -1.66 600 BID). QD vs BID will be explored.

G Van’t Klooster gave us the latest on tx-tx interactions with TMC 125 (abstr A-1827).

Affected Drug Interacting Drug AUC Cmax
TMC125 RTV -46% (27-59) -32% (15-45)
TMC125 EFV -41% (33-48) -18% (8-28)
TMC125 NVP -55% (NA) -36% (NA)
TMC125 IDV +51% (20-90) +51% (16-97)
IDV TMC125 -46% (38-54) -28 (11-42)
SQV TMC125 -52% (20-71) -46% (14-66)


Brian Gazzard presented the 24 week data from Truvada + EFV vs Combivir + EFV (abstr H-1137c). Truvada was presented as the clear winner, which is probably why Gilead threw away any significance of presenting or publishing the 48 week data (they didn’t care—they won at 24w, let’s celebrate now!). It was predicated on an 85% power to detect a 13% margin of non-inferiority using TLOVR.

N 255 254
Women 14% 13%
White 56% 61%
CD4 233 241
<200 41% 42%
Discon’s 11% 21% p=.003
Anemia 0% 5%
AEs 3% 9% p=.008
<400 TLOVR 87% 78% p=.01
<50 73% 65% p=.038
CD4 +129 +111 p=.021 AAUCMB


Adverse events leading to study regimen discontinuation (most common: anemia, nausea, fatigue, vomiting) were fewer for TDF+FTC arm (3%) vs CBV arm (9%). One interesting note is that Gazzard was asked why he presented so much “preliminary” data. This is a 48 wk trial, and presenting so much of the 24 wk data may influence those on therapy. Gazzard responded with 2 things, one that wk 48 has been reached by just about everyone (yes, it takes 6 months to write a presentation) and that there are some capitalistic forces even beyond the principal investigator’s control…

Chris Petropoulous (H-1722c) defined viral fitness for us—in vivo, pre-defined drug and immunologic pressures. Replicative capacity (RC)—in vitro, when there is no drug pressure, there is no immune selection. Virulence is the ability to induce disease. Resistant viruses may replicate less well. Virological failure is not necessarily clinical failure. Drug pressure trumps replicative capacity. More resistance-associated mutations equals a less fit virus. Keep resistant virus non-wild type and it is more viral. An NIH study has 3 questions—Does RC predict the pace of progression? Can RC distinguish those people who might be ready for treatment interruptions? Can RC mark those who should stay on failing regimens?

John Mellors (U Pittsburgh) explained that there is always a frequency, normally low, of drug resistant variants in the viral population. Minor variants make up some 1011 variants / day. Visible (measurable, quantifiable) alleles must be at least in 25% of the RNA. ACTG 398 is a study looking at a switch to a second NNRTI. 56% of the participants were naïve, and there were less failures than expected (in a class that is defined as cross-resistant). Experience is a predictor, separate from resistance (failure). In the Gilead 902 and 907 studies, with a switch to TDF, the L74V predicted the K65R (which may mean that the 65 is already there as a quasi-non measurable species). When the 74 goes down, the 65 comes up. Prior regimens that select for the 74 would lead to the 65. In A5086—single genome sequencing saw really low 3 class mutants that predicted 3-class failure after treatment interruption. The PHPT 2 trial, Jourdain, NEJM, 2004, suggests that there is partial activity of recycled drugs.

Abstract H-580, Dr Cal Cohen of the Clinical Research Initiative of New England, Boston. (10/31/04) Alcorn K

Experienced patients: Kaletra + Tenofovir lowers both

Antiretroviral-experienced HIV+ patients taking lopinavir/r (Kaletra) at the same time as the nucleotide analogue tenofovir (Viread) may require increased doses of Kaletra to achieve adequate drug exposure, according to a poster presented at ICAAC (Abstract A-445). The present study found significant reductions in blood drug concentrations of lopinavir. Seven of the 18 participants were female. The minimum concentration (Cmin) of lopinavir was reduced by 34% when tenofovir was given (p = 0.04). Similarly, ritonavir’s Cmin was reduced by 44% (p = 0.014). In contrast, the maximum concentrations (Cmax) of both drugs were not affected by tenofovir. The authors suggest that the dose of Kaletra may need to be raised in antiretroviral-experienced patients taking tenofovir. They suggested that therapeutic drug monitoring may help to choose the appropriate dose adaptation for individual patients and highlighted a protocol for assessing whether dose adjustment of lopinavir might be needed:

Lopinavir level greater than 4µg/ml: plasma Cmin predictive of treatment success, no dose  adjustment necessary.

Lopinavir level between 2.5 and 4µg/ml: measure intracellular lopinavir concentration; if not possible, and virological response is poor, increase Kaletra dose to 533/133mg.

Lopinavir level below 2.5µg/ml: increase Kaletra dose to 533/133mg. (For more, please see (10/31/04) Gadd C)

Hypersensitivity—High Incidence During Primary Infection

The potentially life-threatening hypersensitivity reaction to abacavir (Ziagen) is often presaged by transient falls in CD4 and CD8 cell counts, according to research presented at ICAAC (Abstract H-171). In another study, investigators found that 18% of individuals with primary HIV infection who took abacavir developed the hypersensitivity reaction (Abstract H-168). Up to 10% of HIV+ individuals who take abacavir will develop a hypersensitivity reaction to the drug within the first few months of treatment. For symptoms, please go to

Five cases of abacavir hypersensitivity were detected during an AIDS Clinical Trials Group (ACTG) study in which patients received Trizivir (AZT/3TC/abacavir), adding efavirenz after seven days. All five patients who developed a hypersensitivity reaction were white, four were male, and the mean age was a little over 36 years. Several conclusions are drawn by the investigators:

There is a high rate of hypersensitivity reaction to abacavir when the drug is used to treat primary HIV infection, and its use should be avoided at this time.

HLA-B*5701 is associated with a risk of abacavir hypersensitivity reaction, and maybe more importantly, the lack of this gene may be a ‘go ahead’ to use abacavir without hypersensitivity worries (although the test to do this is not readily available).

Levels of CD8 cell count and viral load may be independent risk factors for hypersensitivity, although this may be correlated with the HLA-B*5701 polymorphism.

Hypersensitivity may be related to later primary infection. (For more, please see (10/30/04) Carter M)

Treatment Breaks Riskier if NNRTIs Involved

Following on Steve Taylor’s seminal look at NNRTI half-life at the 2004 CROI, individuals who interrupt highly active antiretroviral therapy (HAART) that contains a non-nucleoside reverse transcriptase inhibitor (NNRTI), or resume HAART with an NNRTI, are more likely to have resistance than patients who were taking a protease inhibitor before taking a treatment break, according to Abstract H-576.

“Discontinuation of NNRTI-HAART is risky”, say the investigators. They recommend that if NNRTI based-treatment is going to be interrupted, NRTIs should be continued for two weeks after the NNRTI is stopped. They also warn of the presence of secondary NRTI mutations, not the 103, but precursors to it, which may be harder to see or objectively qualify. These minolrity populations, quasi-species if you will, may be laying the land for the eventual rise of the K103N. (For more, please go to (10/30/04) Carter M)

HIV Infection May Increase COPD Risk

Smokers who have HIV infection are more likely to develop chronic obstructive lung disease than smokers who are not infected with HIV, according to results reported at CHEST 2004, the 70th annual meeting of the American College of Chest Physicians. Smokers with 40 pack-years of smoking “had a 5.5-fold increased risk for developing COPD, while 40 pack-year smokers who were not HIV positive had an odds ratio of 3.” K Crothers of Yale noted that COPD risk increases with age and with the growth of HAART therapy “HIV patients are now living long enough to develop and die of comorbid diseases such as COPD.”

Clinicians treating HIV patients “should consider pulmonary function testing for any patients that develop respiratory symptoms.” After adjusting for COPD risk factors such as age, race and pack-years smoking, “the HIV-infected veterans were 59% more likely to have COPD.” (For more, please see Reuters Health Information Services (10/25/04) Conroy M)

Some Resistant HIV Strains Show Greater Transmission

Certain drug-resistant strains of HIV may be more transmissible than others, Spanish researchers report in the October 15 issue of Clinical Infectious Diseases (39:8, p. 1231, 2004). “Drug resistance mutations were recognized in 17% (15 of 89) of recent HIV-1 seroconverters,” stated V Soriano. After “comparison of resistant genotypes in this population and the resistance profile in a large group of potential transmitters,” he said, “we found that HIV [variants] with some specific drug-resistant genotypes… were more efficiently transmitted than viruses with other drug resistance mutations.” Soriano, of Hospital Carlos III, Madrid, identified strains that were resistant to NRTIs (14.6%), NNRTIs (3.4%) and PIs (3.4%). Strains with 41L, 215Y/F and 181C in the RT gene, and 46L in the protease gene were transmitted to a greater degree than were viruses with other drug resistance mutations—184V and 103N in the RT gene and 82A/S/T and 90M in the protease gene. “Hypothetically,” Soriano concluded, “a larger compromise in the replication capacity of viruses carrying some resistance mutations might explain those findings.” (For more, see Reuters Health Information Services (10/26/04))

The majority of individuals whose lowest-ever CD4 cell count never went below 250 appear able to interrupt treatment safely, for at least one year, according to a mostly prospective United States study published in the November 1 issue of the Journal of Acquired Immune Deficiency Syndromes (37:3, p.1351, 2004). Researchers undertook an examination of the consequences of interrupting highly active antiretroviral therapy (HAART) for any reason. One hundred and seven patients in Dallas were followed for a median of 252 days. Almost 90% of participants had a lowest-ever pre-HAART CD4 cell count greater than 250. The median lowest-ever CD4 cell count was 377 cells. At the time of interrupting therapy, only two individuals had a CD4 cell count below 350 and 61% had an undetectable viral load.

CD4 cell count decline was rapid in the first 60 days (a median of 65 cells / month) but leveled off to a relatively constant 8 cells a month throughout the rest of the follow-up period. Similarly, the increase in HIV viral load was rapid in the first 60 days (2.54 log) but then remained constant without a significant increase. After 13 months of follow-up, the median CD4 cell count remained above the pre-HAART CD4 cell count, and the median viral load was “slightly less” than pre-HAART levels (22,400 vs. 8320 copies/ml).

Two individuals suffered from the effects of increased viral loads, one with ‘retroviral rebound syndrome’ and the second with fevers and malaise, without the complete syndrome. 38% restarted HAART during the follow-up period. The median CD4 cell count was 388 after a median interruption of 6.4 months.

Factors which predicted the combined endpoint were a lower pre-HAART CD4 cell count, the total number of prior drugs, and being older.

Stopping therapy in patients with a pre-HAART CD4 cell count greater than 250 cells is “safe and is not associated with any adverse outcomes.”
(For more, please go to (10/26/04) Bernard E)


Back To Top