EACS: One in five UK patients remains on failing therapy for over a year

An analysis of a large cohort of UK patients found that one in five patients remains on a failing regimen for more than a year, and one in ten for more than two years, the eleventh European AIDS Conference (EACS) was told last week. The average gap between failing therapy and change of drug regimen was 4.8 months.

It also found that only half of those who had to change therapy due to viral load rebound did so in accordance with the British HIV Association (BHIVA) guidelines.

Neither of these situations had improved in recent years, since the guidelines were published. In fact doctors have been less likely to stick to guidelines in recent years. The UK CHIC cohort is one of the largest patient groups under study in Europe. It now numbers 25,274 patients – about 57% of all those seen for HIV treatment and care in the UK – who attend ten of the largest HIV treatment centres.

The study presented at the conference reported on patients starting HIV therapy between the beginning of 1998 and the end of 2004, and specifically the group of patients who experienced virological failure to this first HIV drug regimen, where ‘failure’ was defined as two or more viral loads over 400 copies/ml within the space of six months. It looked at whether and how soon they changed to a second regimen, what that regimen consisted of, and whether some re-suppressed their HIV without a change in treatment. It discounted patients who changed within the first six months of therapy, in order to eliminate those who had to change due to short-term toxicities.

In total 7,931 patients in the cohort started therapy between 1998 and 2005, of whom 6,480 (82%) achieved an undetectable viral load. Of these 694 (10.7%) experienced a viral load rebound as defined by the study, with the average time to rebound being 1.4 years after the start of therapy. There was no relationship between the gender, race or sexuality of patients and the likelihood of rebound. The average viral load at rebound was 4,848 copies/ml.

Seventy per cent of patients who failed eventually changed their drug therapy, but this doesn’t mean 30% stayed on a failing regimen. Nearly a quarter of patients (22%) re-suppressed their HIV within six months of the two detectable viral load readings, and this eventually rose to 27% after two years.

However this still meant that 36% remained on a failing regimen with a detectable viral load after six months, 21% after a year, and 10% after two years. There was no relationship between the year of failure and the likelihood of a change in therapy; patients failing in 2004 were no more likely to change therapy than those who failed in 1998.

The average time from virological failure to change of therapy was 7.2 months, but once those who spontaneously re-suppressed viral load were taken out of the picture, this shortened to 4.8 months. Not unexpectedly, those with lower CD4 counts and higher viral loads at rebound were more likely to change therapies.

Fewer than half of those who changed therapy had it changed in accordance with current BHIVA guidelines, which state:

“The physician should construct a new HIV treatment which includes at least two (or preferably three) active agents guided by HIV resistance testing and by the patient’s previous antiretroviral (ARV) history. The use of an agent from a new drug class is likely to be more effective.”

Only half (242 out of 496) of the patients whose new regimen was known had it changed in accordance with these guidelines. Forty-six patients received one or zero new drugs (‘zero’ in this context meaning that an old drug was ‘recycled’); 73 received no new class of drugs; and 135 received both less than two new drugs and no new class. Thirty-sex per cent of patients who had to change therapy received only ‘recycled’ drugs.

The likelihood that changes would be in line with guidelines actually decreased as time went on. Compared with patients between 1998 and 2000, patients failing in 2001-2002 were 36% less likely to have their therapy changed in line with guidelines and in 2003-2005 only half as likely. Patients on an NNRTI-based regimen were 22% less likely to have it changed than those on other regimens, as were patients on four or more drugs.

What is going on? Presenter Caroline Sabin of the Royal Free and University College Medical School explained that UK-CHIC, like all cohorts, collects limited information, so the figures were difficult to interpret. Physicians might be adopting a ‘wait and see’ approach; patients might be reluctant to change; some of the patients might have viral loads which hovered just above detectability; resistance tests may have been unclear or found no resistance.

“However,” Sabin added, added, “these were people on first-line HAART regimens so they shouldn’t have to be waiting for new drugs to come along to construct a second-line regimen.”

One audience member commented that he “often heard from UK colleagues that they have budget problems,” but Sabin said this should not be an issue in second-line regimens.

Another questioner pointed out that it was appropriate to change just one drug if the cause of virological failure was known – such as a specific resistance mutation. However this still does not explain why patients continue to be maintained on failing regimens, nor why ‘guideline adherence’ has got worse in recent years.