If you can't switch, better to stay on failing treatment than stop it, studies show

Patients on failing drug regimens maintain stable CD4 counts with detectable viral load levels up to 10,000 copies/ml, as long as they remain on HIV treatment, a study presented at the Ninth International Congress on Drug Therapy in HIV in Glasgow showed this week.

In contrast, patients who stop treatment begin to suffer CD4 cell declines at viral load levels as low as 500 copies/ml. It is therefore better to stay on a failing therapy than to stop, the investigators concluded.

Meanwhile, another study found that it is better to stay on therapy in the opposite scenario, where patients achieve an undetectable viral load but never achieve a good CD4 response. Such patients had a much lower risk of AIDs or death than patents who came off therapy, the study found.

Stay on treatment if you have a detectable viral load...

The first study, co-ordinated by the Royal Free and University College Medical School (UCMS), reviewed European patients in seven countries, but will likely have greatest applicability to resource-poor settings where patients are more likely to be maintained on treatment while having a detectable viral load, presenter Amanda Mocroft said.

“We don’t actually know the rate of CD4 count change or at what level of viral load CD4 counts start to decline in developing countries,” she commented.

To address this gap in knowledge, the survey looked at the relationship between viral load and rate of CD4 count change in 7231 European patients from 1997 to 2004.

The study looked at CD4 ‘slopes’. In order to calculate a CD4 slope for a patient, there had to be at least three CD4 measurements taken over a time period when a patient’s viral load and therapy regimen were both stable, with the viral load varying by no more than 0.5 logs (no more than threefold). The researchers could then use the three CD4 tests to calculate a slope, expressed as the average number of CD4 cells gained (or lost, in the case of a negative slope) per year.

‘Stable’ antiretroviral therapy (ART), in this case, meant no change in any regimen component over the duration of the CD4 slope, even in cases of failure.

The majority of patients included in the survey were white men. Over a quarter had had AIDS diagnoses, with an average pre-therapy minimum CD4 count of 168 and a baseline viral load of 4.6 logs (40,000). The 7231 patients contributed about 59,000 CD4 slopes between them, an average of six per patient.

The study confirmed that being on ART was better than not taking it, regardless of viral load. Patients on stable ART with viral loads under 500 copies/ml had an average gain of about 40 CD4 cells a year.

If the viral load was between 500 and 10,000 copies/ml, the CD4 count neither increased nor declined as long as patients stayed on ART. But at viral loads over 10,000 the CD4 slope became negative and patients who stayed on failing ART lost an average 42 cells a year.

Patients who had never started therapy, however, lost 40 cells a year at any level of detectable viral load (defined for this study as over 500 copies/ml). When viral load rose above 10,000 copies/ml, CD4 cell loss accelerated to around 70 cells a year.

Worst off were people on treatment interruptions, and this study added yet more weight to the evidence that once you start on ART, it is a bad idea to stop. Patients on treatment interruptions lost nearly twice as many cells at viral loads between 500 and 10,000 as patients who’d never started ART (about 70 cells a year) and lost about 110 cells a year at viral loads above 10,000 copies/ml – nearly three times as many as patients maintained on so-called ‘failing’ ART.

The types of drug used in the ART regimen also had an effect on the CD4 slope. Patients on protease inhibitors (PIs - boosted and unboosted) had better CD4 increases than patients on other regimens, about 50 versus 25 cells a year at viral loads below 500 copies/ml. At viral loads above 10,000 copies/ml everyone experienced CD4 declines but patients taking PIs still had slower declines. Mocroft therefore recommended that second-line regimens should all contain a boosted PI.

Stay on treatment if you have a low CD4 count...

Another cohort study also found that it was much better to stay on therapy than come off it in the situation where, despite achieving viral undetectability, patients do not make a full CD4 count recovery. Wendy Bannister, also of University College Medical School, told the conference that patients with CD4 counts under 200 who came off ART were ten times more likely to die than patients who remained on ART – even if those on ART had a detectable viral load.

Bannister was reporting on findings from the EUROSIDA Cohort, a group of patients from seven European countries. To be included in the study, patients had to have a CD4 count under 200 within six months of a viral load test. The patients were then divided into three groups according to whether they were on ART with a viral load under 500 (Group A), on ART with a viral load over 500 (Group B), or not on ART (Group C).

Clearly, patients could be in any of the three groups at different times, and Bannister gave an example of a fictitious patient who’d originally been on failing dual therapy and had stopped it (group C), subsequently started triple combination therapy but still had a detectable viral load (group B), then finally became undetectable (group A).

The three groups were similar, though the ones off therapy were less likely to be male and more likely to be injecting drug users, a possible confounding factor. Forty-five per cent of those off ART and 40% of those on it but with a detectable viral load had CD4 counts under 100 compared with 20% of those with an undetectable viral load. The average viral load in patients off ART was 72,000 copies/ml and in those on ART but detectable, 25,500 copies/ml.

The combined rate of AIDS or death was 5% a year in group A, 12% in group B and no less than 60% in group C. However when it came to the risk of death alone, patients on ART (groups A and B) had an equal risk of dying (about 3% a year) regardless of viral load, whereas in group C patients the risk of death was about 30%.

When this was further resolved into deaths from AIDS conditions and non-AIDS conditions, being on ART and having undetectable viral load protected patients from AIDS-related conditions, but viral undetectability did not protect them from non-AIDS conditions; indeed those in group B had the lowest risk of death from non-AIDS-related conditions. The risk of AIDS deaths was 0.4%, 1% and 11% in groups A,B and C respectively while of non-AIDS deaths it was 2%, 1% and 11%.

The tenfold higher risk of death and disease in people off ART can be partly explained by terminally ill people being taken off ART or having died before starting it, Bannister commented; but the difference between AIDS and non-AIDS deaths is intriguing and warrants further investigation.