Viral rebound during antiretroviral therapy (ART) is just as likely to be detected by a failure to pick up medication from a hospital pharmacy as by a decline in the CD4 cell count, a large study from southern African clinics shows. The results, published today in PLoS Medicine, are likely to add to the debate about what methods are most suitable for spotting treatment failure in the absence of viral load testing.
Regular viral load testing can detect rebound of viral replication. If virus levels rise above the limits of detection of the test, treatment may be switched to prevent the development of resistance, or patients may receive intensive adherence counselling if they have reported recent problems in taking all their medication. Improved adherence may be enough to re-suppress viral load below 50 copies/ml, without the need for a switch to a more expensive second-line regimen.
Where viral load testing is not available, the World Health Organization currently recommends using a decline in the CD4 cell count as a trigger for switching treatment. But the CD4 cell count may take a long time to decline to a level that will cause concern, and throughout this period viral replication will continue unchecked, with the result that the patient develops greater levels of resistance to the drugs he or she is taking. This resistance may reduce the effectiveness of second-line treatment, especially where second-line drug options are restricted on the grounds of cost.
Researchers at the University of Pennsylvania School of Medicine and Johns Hopkins Bloomberg School of Public Health in the United States and Aid for AIDS, a disease management programme in South Africa, set out to determine whether monitoring adherence levels might provide an accurate early warning of virologic treatment failure. Since viral load rebound – or failure to suppress viral load on treatment – is usually due to poor adherence, they reasoned that keeping an eye on patient adherence might provide clinicians with sufficient information to know when to intervene in the absence of laboratory monitoring.
The study used the crudest measure of adherence available – the failure to pick up medication from a hospital pharmacy at a scheduled monthly appointment. Records were available through a computerised system for 1,982 patients who had started NNRTI-based ART and who had CD4 count and viral load data available at baseline and after 6 or 12 months of treatment.
The study looked for two forms of treatment failure:
- Lack of virologic response at 6 months or 12 months (viral load above 1,000 copies/ml)
- Viral rebound above 1,000 copies/ml after previous viral load below 400 copies/ml.
Adherence, calculated by the number of pharmacy refill claims, was expressed as a percentage of the total months since initiating treatment in which medicines were collected.
Lack of virologic response was detected in 25% of patients with viral load measurements available at 6 months, and in 26% of patients with measurements available at 12 months. Viral rebound was detected in 14% of patients who had initially achieved undetectable viral load (n=1101), after an average follow up of 1.75 years.
Adherence levels predicted virologic failure at 6 and 12 months more accurately than CD4 cell count changes, and this relationship persisted when the threshold of virologic failure was raised to 10,000 copies/ml.
There was no significant difference between adherence and CD4 count in their accuracy at predicting virologic breakthrough.
When different levels of adherence or CD4 count decline were compared for sensitivity – the percentage of virologic failures missed) and specificity (the percentage who changed regimens despite viral suppression), adherence greater than 90% was substantially more sensitive than any CD4 cell measurement. However, switching on the basis of CD4 cell count decline would have resulted in far fewer unnecessary switches to second-line treatment.
Detecting breakthrough viremia: performance of various measures
| Adherence | % virologic failures missed | % regimen changes despite viral suppression |
| Adherence | 83% | 3% |
| 75% | 4% | |
| 66% | 8% | |
| 58% | 16% | |
| 36% | 36% | |
| 28% | 54% | |
| CD4 | % virologic failures missed | % regimen changes despite viral suppression |
| CD4 | 85% | 3% |
| CD4 decrease | 85% | 2% |
| 77% | 4% | |
| 68% | 9% | |
| 56% | 15% |
The authors suggest that “adherence monitoring could be used to pursue focused virologic or genotypic testing in settings where these assays are available to some but not all”, and also note that adding adherence measurement to CD4 counting could allow viral load testing to be postponed in cases where there appears to be a very low risk of virologic failure already being present.
“Clinics able to perform viral load assessments in all patients routinely could use adherence monitoring to guide decision-making on timing of these tests. For example, patients with perfect adherence and a CD4 count increase of more than 100 cells at 6 mo[nths] could have their viral load assessed at 12 mo[nths], or patients with adherence
The authors note that their measure of adherence – pharmacy refills – may overestimate actual adherence because patients may not take all the drugs they collect, and they say that other methods of adherence measurement should also be assessed. However pharmacy claims in many settings will be the easiest data to evaluate, particularly where patient records are not computerised or staff time to carry out more involved questioning about adherence is limited.
Provision of adherence data to patients at the point of care in order to motivate good adherence, say the authors, could be either simple or complex.
"In Botswana for example, patients present to providers with a paper pharmacy card on which dates of ART dispensation (and pill counts) are noted. Providers conceptually can calculate adherence directly using these cards. A more complex approach for clinics with computers would be to link pharmacy and patient care records electronically, so that a program would automatically supply the pharmacy refill adherence. The ability of adherence to identify patients at high risk of virologic failure…should be considered a reason for clinics to organise these data in a way that can be used in simple algorithmic approaches to patient care.”
Further reading
Models of adherence support from around the world were reviewed last September in HIV & AIDS Treatment and Practice, NAM's newsletter on HIV treatment in resource-limited settings.
Debates about what indicators to use in order to switch treatment were also discussed in HIV & AIDS Treatment in Practice last year.
Bisson GP et al. Pharmacy refill adherence compared with CD4 count changes for monitoring HIV-infected adults on antiretroviral therapy. PLoS Medicine 5 (5): e109. doi:10.1371/journal.pmed.0050109