Skip directly to content

Stopping Therapy


tagline 1999

Frankfurt HIV Clinic Cohort Study of Total Treatment Interruptions Followed by Mega-HAART

Not So Unforgiving After All?
Viral Load Post TTI to Mega-HAART
in WT-Shifters & Non-Shifters
N26/39 (66.7%)13/39 (33.3%)
change in RNA (Wk8, mean)- 2.9- 0.78
RNA > 500 (Wk24)19/24 (79%)1/9 (11%)
TTI = total treatment interruption; WT = wild-type virus


In Toronto, Dr. Veronica Miller presented an expanded dataset derived from results she had shown in Glasgow (November 1998) and Chicago (February 1999) on the Frankfurt HIV clinic cohort. The initial study impetus was not a drug holiday per se, but rather the introduction of so-called mega-HAART in heavily pre-treated, highly drug-resistant HIV patients after a variety of clinical courses, some of which had included a total treatment interruption (TTI) due to the lack of available viable treatment regimens. In Glasgow and Chicago she showed that three out of five patients (60%) who chose to undertake a TTI before initiating mega-HAART experienced a reversion of viral genotype to wild-type. At the Second International Workshop on Salvage Therapy for HIV Infection, Veronica Miller expanded the dataset to demonstrate that the same result occurred in 26/39 (66.67%) such patients. Moreover, among those whose virus shifted to wild-type, the introduction of mega-HAART caused a significant and profound drop in viral load and was not always associated, at least in the short term, with the re-emergence of genotypic resistance or drug failure.

Traditional retrovirologic dogma had held that the virus is usually, if not always, "genetically unforgiving", that the emergence of drug-resistance is a one-way street, and that even if wild-type (WT) virus emerged during a drug holiday, resistance was sure to re-emerge rapidly once the selective pressure of any given drug or a cross-resistant drug analogue were reimposed. Miller's data suggested that, at least in the short term, in a surprising two-thirds of patients who'd experienced a TTI, this was not the case. Her presentation, undoubtedly the most provocative and intriguing one at the Salvage Therapy Conference, generated a considerable stir.

Veronica presented data on 85 highly pretreated patients who were given mega-HAART. Of these, 81% had previously received six or more antiretrovirals. The baseline viral load, at the time they initiated mega-HAART (often after a TTI) was 5.21 log10 and the baseline CD4 count 108/mm3. The mega-HAART included six (64%), seven (30%) or eight (6%) drugs. Median follow-up was twelve months. Of the 85 people enrolled at Frankfurt, ten had experienced a total treatment interruption (TTI) of at least two months prior to initiating mega-HAART. Miller had pre- and post-TTI viral isolates from nine of these ten patients, and sent them off to Virco for resistance testing. Of these nine, six (66.67%) had experienced a shift to wild-type (WT) virus measured both genotypically and phenotypically. Their isolates had previously demonstrated 100-fold resistance to AZT and 3TC, 1,000-fold resistance to the NNRTIs and high-level resistance to the protease inhibitors as well.

Because of these intriguing results, and to increase the relatively small dataset, Miller set out to do a drug holiday analysis on fifty patients from the Frankfurt cohort. She looked retrospectively at all patients treated since January 1996 who had received at least two protease inhibitor-containing regimens for over one year and who had discontinued treatment for at least two months. Of fifty such patients, pre- and post-TTI resistance samples were available for 39 patients. By phenotypic and genotypic analysis, 26/39 (66.67%) of these TTI-to-mega-HAART patients experienced a complete reversion to wild-type (WT) virus during the TTI. Their virus had previously been resistant to a median of eight drugs (range 2-11). There was no shift in 13/39 (33.3%), who had received the same median number of previous antiretrovirals.

Among the 26 WT-shifters, the change in CD4 levels was not dependent on baseline viral load. Both groups lost about 125 CD4 cells. Among non WT-shifters, viral load increased more dramatically in those beginning with a low viral load. (Viral load appeared to rise to an equal physiologically possible level in both WT-shifters and non-shifters.)

Among the non-shifters, the CD4 loss was small if baseline viral load was low (-24 cells/mm3) and large if baseline viral load was high (-88 cells) -- the expected result. However, according to Veronica, "what I found interesting was that the CD4 cells were equally high for patients with high and low viral load if they had 'shifty 'viruses, whereas in patients with non-shifty viruses, the CD4 counts were low if the VL was high and vice versa, as you would expect. In the patients with the shifting virus populations, the high CD4 counts despite high virus load could reflect the 'discordance' that people like Stephen Deeks and the Swiss cohort have talked about. It also ties in with our data (also presented at Glasgow, and currently being reviewed by Annals of Internal Medicine) showing that clinical progression in patients receiving PIs who have high viral loads is significantly lower than patients not receiving PIs with equally high viral loads. In other words, we saw an independent treatment effect for PIs."

However, the responses to mega-HAART after TTI differed significantly between the WT-shifters and the non-shifters, as might be expected. The shifters experienced a -2.8 log10 drop in plasma viral load, whereas the non-shifters dropped just -1.02 log. After eight weeks of mega-HAART -- a very early time point -- 72% of shifters were beneath the limit of quantitation (BLQ), <500 HIV RNA copies/ml, versus just 10% of the non-shifters. Thus, the response to rescue therapy/mega-HAART was much better in the WT-shifters. Miller pointed out that because of these dramatic differences, both during the TTI and after the initiation of mega-HAART, viral load and CD4 counts need to be very closely monitored in this population.

Might there be immunologic differences between the shifters and the non-shifters? For example, did shifters have stronger specific anti-HIV responses, thus driving re-emergent viral isolates back to more "fit" wild-type virus, while non-shifters had such blunted specific anti-HIV responses that there was no such selective pressure. Veronica is interested in working with researchers who could use the new single cell interferon gamma release assay (such as Picker/Koup in Dallas or Roederer at Stanford) to quantify HIV-specific responses. San Diego's Diane Havlir assured her that these assays are possible when samples are shipped overnight on ice.

Another hypothesis would examine whether the resistant virus of the shifters was less fit than that of the non-shifters (some resistant viruses, such as those with NNRTI resistance, can be very durable, as shown by the MDR patient who transmitted an MDR, NNRTI resistant virus even though he'd been off nevirapine for over two years, suggesting no reduced fitness). John Mellors, clearly flabbergasted, asked how durable the virologic responses were in the shifters. Veronica answered that they looked good in the short term, but that follow up is not that far out. Some patients' virus remains suppressed after 11 months or more. "If it shifted rapidly, it wasn't very happy [fit] to begin with," she commented. "We're going to do fitness studies on the pre- and post-virus isolates." Towards the end of an animated Q&A period, ACTG co-chair Connie Benson -- echoing a similar caveat by Doug Richman in Chicago -- sternly warned that, however intriguing these results might be, they required careful further study in the context of controlled clinical trials with extensive laboratory analysis, and that "it would be wrong to encourage patients to do this on their own."