Tenofovir resistance may develop in more than half of patients failing treatment in sub-Saharan Africa

More than half of people who experienced failure of a tenofovir-based antiretroviral regimen in sub-Saharan Africa had resistance to tenofovir, a meta-analysis of drug resistance studies published in Lancet Infectious Diseases has shown.

The study found that the prevalence of tenofovir resistance after first-line failure ranged from 20% in Western Europe and North America to 56-60% in sub-Saharan Africa.

The authors say that their findings suggest that somewhere between 7.5 and 17.5% of people who start treatment in sub-Saharan Africa, with a regimen composed of tenofovir, efavirenz and either lamivudine or emtricitabine, will develop tenofovir resistance after one year, based on current rates of treatment failure. This projection assumes a treatment failure rate of between 15 and 35%, depending on how it is measured, and is in line with recent World Health Organization estimates.

This study does not reflect the prevalence of HIV drug resistance in all people on treatment, nor the prevalence of drug resistance in people who have not yet started treatment, but only the prevalence of drug resistance in people who started specific drug regimens and subsequently experienced virological failure of that treatment regimen.

The TenoRes Study

The TenoRes collaborative study was designed to assess the prevalence of tenofovir resistance in people experiencing failure of first-line treatment. Tenofovir is a recommended component of first-line antiretroviral regimens in World Health Organization (WHO) HIV treatment guidelines and in European and United States treatment guidelines.

The study took data from 44 treatment cohorts in 26 countries in North America (3 studies), Europe (13 studies), Latin America (4 studies), Asia (3 studies) and sub-Saharan Africa (24 studies, sub-divided into 4 studies in the East African region, 10 studies in the Southern African region and 7 studies in West and Central Africa) where drug resistance data were available after first-line treatment failure. Participant data were eligible for inclusion in the analysis where virological failure had occurred more than four months after beginning first-line antiretroviral treatment consisting of tenofovir plus either lamivudine or emtricitabine, and a non-nucleoside reverse transcriptase inhibitor (either efavirenz or nevirapine).

The researchers carried out a series of analyses designed to identify regional, demographic and disease-specific differences in the risk of developing resistance to tenofovir and to other drugs in the failing regimen in the 44 cohort studies, comprising 1926 people.

They calculated country-level and regional risks of developing tenofovir resistance in order to take into account potential confounding factors such as differences in the delivery of care between treatment facilities or countries.

The meta-analysis found a wide variation in pre-treatment CD4 cell count between regions, from 44 cells/mm³ in Latin America and 89-104 cells/mm³ in the African regions, to 199 cells/mm³ in Western Europe. The variation in baseline viral load was most pronounced within sub-Saharan Africa, ranging from 4.8 log₁₀ (63,000) copies in southern Africa to 5.58 log₁₀ (316,000) copies/ml in Eastern Africa. Pre-treatment viral load was significantly higher in Eastern and Western/Central Africa and in Latin America compared to other regions.

Tenofovir resistance more frequent at CD4 counts below 100

The analysis showed that regardless of region, tenofovir resistance was associated with a lower pre-treatment CD4 cell count. People who started treatment with a CD4 cell count below 100 cells/mm³ were 50% more likely to develop resistance after treatment failure, but this difference did not seem to be driven by viral load. There was no significant difference in the risk of developing tenofovir resistance between people with baseline viral load above or below 100,000 copies/ml, nor in the risk of developing tenofovir resistance according to viral load at treatment failure.

However, the difference in the risk of tenofovir resistance according to baseline CD4 cell count was largely driven by two regions, Southern Africa and North America. Whereas there was no significant difference in the risk of tenofovir resistance in people with pre-treatment CD4 cell counts below 100 cells/mm³ in most regions, the risk of developing tenofovir resistance was 27% higher in Southern Africa (odds ratio 1.27, 95% CI 1.09-1.48) and 34% higher in North America (OR 1.34, 95% CI 1.06-1.69) among people with CD4 cell counts below 100 compared to those starting treatment with CD4 cell counts above 100.

Although researchers suggest that the regional difference in tenofovir resistance is due to less frequent viral load monitoring, they also found a significant difference in the development of tenofovir resistance according to viral subtype and the other elements of the regimen, factors which also vary from one region to another.

In the case of viral subtype, the researchers found that people with subtype C virus (which predominates in Southern Africa) were 2.44 times more likely to develop tenofovir resistance after treatment failure compared to people with other non-B subtypes (OR 2.44, 95% CI 1.66-3.59). To reduce the effect of confounding factors the researchers restricted this analysis to patients receiving care in Western Europe and to non-B subtypes found in migrant populations in order to reduce any bias caused by the health system or by socio-economic factors.

Tenofovir resistance more frequent when taken with lamivudine or nevirapine

Considering the other elements of the antiretroviral regimen at the time of treatment failure, the researchers made two findings that may explain regional variations in tenofovir resistance. They found that people taking nevirapine were almost 50% more likely to develop resistance to tenofovir than people taking efavirenz (OR 1.46, 95% CI 1.28-1.67), and almost 50% more likely to develop resistance to tenofovir if they received lamivudine rather than emtricitabine (OR 1.48, 95% CI 1.20-1.82).

These findings suggest that, in studies where the median year of initiation of antiretroviral therapy was between 2008 and 2011, what is being detected is as much an effect of the use of out-of-date treatment regimens as it is an effect of late treatment initiation – both of which have been strongly warned against in updated WHO treatment guidelines since 2013. Guidelines now identify efavirenz as the preferred agent over nevirapine, and recommend treatment initiation before the CD4 cell count falls below 350 cells/mm³ – and preferably soon after diagnosis, regardless of CD4 cell count.

In the case of lamivudine WHO guidelines have been less firm, with an update in 2014 noting that studies suggest clinical equivalence. Since then a Dutch cohort study of 4740 people has shown a doubling of the risk of virological failure in people taking lamivudine compared to emtricitabine, and the risk was more pronounced in people taking nevirapine than in those taking efavirenz. 

The TenoRes study also found that of those who developed tenofovir resistance, 83% developed the M184V mutation conferring resistance to lamivudine and emtricitabine, and 78% developed resistance to nevirapine and efavirenz.

The findings indicate the need for further options for first-line treatment, say the authors, and also for second-line treatment, where WHO guidelines recommend the recycling of tenofovir. Improving access to viral load testing, increasing the frequency of viral load tests and development of cost-effective resistance screening could also contribute to reducing resistance.

The researchers also raised concerns about the possibility of onward transmission of tenofovir-resistant virus, and its impact on first-line treatment for the newly infected and tenofovir-containing pre-exposure prophylaxis (PrEP) for prevention of HIV infection.

"If resistant strains of HIV were significantly less effective at spreading in people, we would expect to see lower levels of the virus in patients with the resistant strain," said Dr Ravindra Gupta of University College London. "However, we found that virus levels were no lower in individuals with the resistant strain and were high enough to be fully infectious. We certainly cannot dismiss the possibility that resistant strains can spread between people and should not be complacent. We are now conducting further studies to get a more detailed picture of how tenofovir resistant viruses develop and spread."