To date, few data have been reported on emerging drug resistance patterns in Africa. In an oral presentation at the International AIDS Conference in Mexico City on August 5th, delegates heard that drug resistance was widespread among people failing antiretroviral treatment in Malawi, due to lack of viral load testing to identify virologic treatment failure.
Patterns of resistance mutations among this group were also described in data presented by Mina Hosseinipour of the UNC Project, Lilongwe.
People who are allowed to continue on a treatment regimen after virologic failure (i.e., after viral load has rebounded to consistently detectable levels) are highly likely to develop resistance to one or more of their treatment drugs, and possibly cross-resistance to other drugs in the same treatment classes.
However, since many resource-limited settings are unable to provide viral load monitoring, virologic failure is seldom detected. Instead, treatment failure is gauged by clinical progression or immunologic failure (falls in CD4 cell count). As virologic treatment failure almost always occurs first, reliance on clinical or immunologic signs of treatment failure virtually guarantees the development of drug resistance.
Over 130,000 HIV-positive people in Malawi have started antiretroviral treatment. First-line therapy for all patients is stavudine (d4T), lamivudine (3TC) and nevirapine (NVP); patients may be switched from d4T to zidovudine (AZT) or nevirapine to efavirenz for reasons of toxicity. Those who fail first-line treatment are started on a combination of AZT, 3TC, tenofovir, and lopinavir/ritonavir (Kaletra). According to Malawian national guidelines, clinical failure is defined as the development of a new Stage 4 condition, and immunologic failure is defined as a decline in CD4 cell counts to 50% of the pre-treatment CD4 cell count baseline; patients are only switched to second-line therapy when either of these conditions occurs.
Due to inability to afford routine viral load testing, "we are rarely able to catch virologic failure," stated Dr Hosseinipour. "We simply hope that there are no lopinavir resistance patterns in patients failing first-line therapy, so that patients have the best chance of success with second-line treatment."
In this prospective study, the research team evaluated the prevalence and patterns of resistance mutations emerging in patients failing first line antiretroviral therapy in Lilongwe and Blantyre. A total of 101 patients confirmed as meeting the definition of treatment failure (according to the definition above), from December 2005 to July 2007, were evaluated. Virologic failure was confirmed by viral load testing; genotypic analysis was then performed on blood samples with HIV viral loads of at least 1000 copies/ml. Resistance mutations were defined as per IAS-USA guidelines.
In the 101 patients analysed, the mean CD4 count was 121 cells/mm3, mean viral load was 135,984 copies/ml, and mean duration on antiretroviral therapy was 38 months. Ninety-six patients were identified as having virologic failure with viral loads over 1000 copies/ml. HIV RNA in two samples could not be amplified, leaving 94 samples for analysis.
Resistance patterns identified
Five percent of the samples were wild-type, with no mutations identified.
NNRTI mutations (181C, 103N, 106M, 188L, 190A) were seen in 93% of the samples. The 3TC-associated M184V (or M184I) mutation was seen in 81% of the samples. These were most commonly seen together: no samples showed only an M184 mutation, and just 2% had only an NNRTI-associated mutation.
The so-called TAMs (thymidine analogue mutations), associated with resistance to AZT and d4T, were identified in 56% of the samples – one or two TAMs in 32%, three or more in 25% (three or more thymidine analogue mutations are likely to substantially reduce the response to tenofovir in second-line treatment). The most common pattern overall was an accumulation of NNRTI mutations, plus M184, plus one or more TAMs (the most common being 215F/Y).
Tenofovir-associated mutations (K65R or K70E) were seen in 23% of the samples, and tenofovir mutations plus TAMs in 7%. What Dr Hosseinipour referred to as "pan-nucleoside mutation combinations" occurred in 17%: these corresponded to K65R or K70E plus additional multi-nucleoside resistance mutations, most commonly the Q151M complex and/or (more rarely) 69 insertions.
Risk factors for resistance
CD4 cell counts below 100 cells/mm3 significantly increased the risk of K65R or K70E mutations (adjusted odds ratio, 6.1; 95% confidence interval [CI], 1.47 to 25.0) and of pan-nucleoside mutations (odds ratio, 9.8; 95% CI, 1.16 to 82.9). Use of AZT lowered the risk of both types of mutations, with an odds ratio of 0.18 (95% CI, 0.04 to 0.94) for K65R/K70E, and 0.12 (95% CI, 0.014 to 0.978) for pan-nucleoside mutations. (All figures were from a multivariate analysis adjusting for sex, viral load, CD4 count, and AZT use.)
However, both CD4 cell counts 3 and AZT use increased the risk of developing more than three TAMS: the odds ratio for CD4 cells was 5.57 (95% CI, 1.41 to 22.02), and for AZT use, 3.4 (95% CI, 1.07 to 10.97).
Study limitations included not knowing the exact duration of virologic failure, of using phenotypic resistance associations for clade B virus that might not necessarily apply to clade C or other virus, and of being unable to determine any potentially archived resistance beyond that detected in circulating blood.
However, the researchers were able to conclude that, "upon first line ART failure diagnosis which relied on clinical and CD4 monitoring, extensive resistance was present to NRTI/NNRTI agents, including 17% of patients who acquired pan-NRTI resistant virus and an additional 56% who likely had reduced susceptibility to multiple NRTIs." Including AZT in the first-line regimen "appears protective for the emergence of tenofovir and pan-nucleoside resistance mutations but increases the risk for developing multiple (three or more) TAMs."
These findings "have profound implications for the optimal time to switch treatment and the choice of second line therapy for developing countries." Depending on the second-line combination chosen, between 22% and 50% of Malawian patients with first-line failure may have no fully active drugs in their second regimen apart from Kaletra.
In a question and answer session, Dr Hosseinipour reiterated that "viral load monitoring does have to come into the forefront. We have to find some feasible method of measuring viral load in resource-limited settings, if not necessarily the standard assays now in use. Other studies all point to the need to use virologic monitoring to ensure treatment success."
Asked whether specific drugs appear to select for K65R, Hosseinipour stated that "all K65Rs except two were in people who'd been on d4T, and the other two were only on AZT for short periods of time, so it seems it's very clearly d4T that's selecting for K65R."
(The questioner, McGill University's Mark Wainberg, stated that this was in agreement with their own data, and that "we should stop referring to K65R as a tenofovir mutation and acknowledge that it's selected by multiple drugs including NRTIs.")
Reference:
Hosseinipour M et al. Resistance profile of patients failing first line ART in Malawi when using clinical and immunologic monitoring. Seventeenth International AIDS Conference, Mexico City, abstract TUAB0105, 2008.