Low level of drug-resistant strains common in untreated patients, but may not affect treatment

A French study of newly-infected, drug-naïve HIV patients published in the journal AIDS has found that low level populations of HIV with drug resistance mutations which were undetectable by conventional resistance tests were present in nearly one in three people. About one in seven patients had HIV drug resistance detectable by conventional tests, and taken together 42.4% of patients had some detectable resistance.

However the study also found that while patients with large amounts of drug-resistant HIV had a poorer response to treatment, with slower viral load declines, the response to treatment was not significantly worse in patients with small amounts of drug-resistant virus.

It also found that newly-infected patients with a majority population of drug resistance mutations tended to have lower viral loads and higher CD4 counts at baseline, and preserved that CD4 advantage over the first six months of treatment. So, although resistance may have an impact on viral load response, it has little impact on clinical outcome, at least in the short-term.

The Aquitaine Cohort is a study of patients treated at five hospitals in south-west France; established in 1987, it is one of the longest-established cohort studies in Europe.

For this particular study patients were selected who:

• Were diagnosed with HIV between 1996 and 2005
• Were recently infected
• Were subsequently treated with antiretroviral drugs
• Had at least one CD4 count and viral load done before starting treatment and within 18 months of the estimated date of infection.
• Had a stored blood sample that could be tested for resistance

Samples that showed no evidence of HIV drug resistance on a conventional genotyping test were then re-tested using ultrasensitive resistance tests tuned to detect small amounts of virus with one or more of three specific resistance mutations: the K103N non-nucleoside (NNRTI) mutation, the M184V 3TC/FTC NRTI mutation, and the L90M protease inhibitor mutation.

One hundred and seventy-two patients were included in the study. About two-thirds were gay men and one-third heterosexual, with only small numbers of other risk groups. Four out of five were men, and one in six had a subtype of HIV other than type B, the one common in the developed world. Their average age was 32, their median viral load at the time of the taking of the blood sample was 52,000 copies/ml and their CD4 count quite high at 426 cells/mm³.

Conventional genotypic testing found that 23 patients (13.4%) had HIV drug resistance, presumably transmitted from others. Resistance to nucleoside (NRTI) drugs was detected in 11.6% of patients, to NNRTIs in 6.4% and to protease inhibitors in 4.1%.

Seventy-three patients’ samples were then re-tested for the three minority resistance mutations. No example of the L90M protease inhibitor mutation was found. However 21% of those tested for the K103N mutation were found to have it and 25% of those tested for M184V. Virus with K103N comprised 2.5% to 7.1% of the total viral population in patients who had it, and M184V 1.3% to 23.5% of the population.

Sixteen out of the total group of 172 patients (9.3%) were found to have had resistance to at least one of the drugs in their initial regimen.

The 23 patients with majority resistant HIV had significantly lower viral loads at the time their sample was given (6500 versus 48,000 copies/ml) and higher CD4 counts (560 versus 438 cells).

After treatment was started all patients had similar viral load declines in the first month but in months one to six patients with majority drug-resistant virus had viral load declines that were about ten times slower than patients without them – about a 30% decline per month compared with a threefold decline in those without drug resistance.

However the two groups had similar CD4 rises in the six months - about 120 cells in the resistant patients and 95 in the non-resistant patients – so, at least up to the sixth month after the start of therapy, immune recovery was similar.

These differences were not seen in patients with minority resistant virus. They started with a similar viral load to patients with wild-type virus and although their CD4 count was somewhat higher at baseline (534 versus 451 cells), this difference was not statistically significant (p=0.14). CD4 count rises in the six months after treatment was started were essentially the same and although the viral load decline was slightly smaller in the minority-resistant group (about a 2.5-fold decline per month compared with a 3.5-fold decline) this was also not significant (p=0.23).

The authors of the paper point out that some studies of patients with minority resistant virus have found worse clinical outcomes in them. They suggest that their study might not have had the statistical power to pick out a difference and that one of the mutations they measured, M184V, both produces HIV that is less fit and reverses the effect of AZT and tenofovir mutations that might also have been present as minority variants.

However in general the paper shows that in the first six months on therapy at least, the presence of resistance mutations does not cause treatment failure, especially in patients with small amounts of resistant virus.