How reliable is an undetectable viral load test?

How often do people become potentially infectious in between viral load tests, if their previous viral load was undetectable? This is a factor that obviously needs to be considered in models of the efficacy of treatment as prevention models. If viral loads were essentially no guide to a person’s infectiousness a month after a test, then using quantities like community viral load to predict HIV incidence would be meaningless.

One study has found that once a person has achieved an undetectable viral load, at least in a European setting, the chances that their next viral load will be over 1000 (a viral load below which infection is regarded as very uncommon) is only about 2%, and only 1% if they started HIV therapy with a fully suppressive HAART combination and don’t have a history of suboptimal regimens.¹

Investigators from the Swiss HIV Cohort looked at the likelihood of successive viral load tests being undetectable. “By analogy”, they write, “this likelihood also applies to sexual contact some time after a measured undetectable viral load, a situation where real-time measurements are not practicable.”

The study population involved 6168 patients taking antiretroviral therapy who had at least two successive viral load measurements below 50 copies/ml between May 2003 and the end of 2007. The median interval between viral load measurements was three months. About one third had originally started HIV treatment in the pre-HAART era, and had taken suboptimal dual or single therapy.

Viral load rebounded to above 50 copies/ml on at least one occasion in 43% of patients, and above 1000 copies/ml in 7% of patients.

Eighty per cent of patients reported that they had not missed a single dose in the previous four weeks, which was this study’s definition of 100% adherence. In all patients, if none or one dose was missed there was a 14% to 15% chance that the next viral load would be detectable, though not necessarily at a high level. If more than one dose of HIV treatment was missed, the probability of it remaining undetectable between tests was 70%. However, in individuals who started HIV therapy with a triple-drug combination, there was only a 5% chance that the next viral load would be over 50 copies/ml if the previous one was undetectable.

Overall, there was a 98% probability that viral load would not rebound from undetectable to the potentially infectious level of 1000 copies/ml or above in successive tests. If an individual had started HIV treatment with a triple combination of drugs this increased to 99%.

Most increases in viral load were low and transient. In 66% of patients it rebounded to below 200 copies/ml, and in two-thirds of patients whose viral load increased to below 1000 copies/ml it was once again undetectable in the following test. The same was true for 30% of individuals whose viral load increased to over 1000 copies/ml.

The investigators examined rebounds in viral load to 1000 copies/ml or more in greater detail. In 78% of cases there was an explanation for this - for example, poor adherence. However, “for the remaining 22%, no plausible explanation could be found. The data therefore leave open the possibility that unexplained rises in viral load above 1,000 copies/ml, although rare, may occur.”