Another year, another Workshop on HIV Drug Resistance. We have become accustomed to seeing large amounts of new information at this meeting, at least some of which has a major and immediate impact in our approach to patient care.

Well, for 2002, we might be a bit disappointed, at least as far as NNRTIs are concerned. Resistance to this class of agent is perhaps the best understood of all. In one interesting protocol, the GUESS study, a group of 12 experts were presented with as series of 45 complex viral genotypes, selected based on the availability of phenotypic resistance test results for all antiretroviral agents and showing a 4-fold increase in IC₅₀ for at least one drug [1]. For the NNRTIs, the experts predicted the phenotype (based on genotypic information alone) with great precision - 66% to 69% for the three NNRTIs. Clearly, new "blood" is needed in this class.

Elegant crystallographic studies have demonstrated that, at least for nevirapine, mutations at codons 100, 106 and 108 give rise to significant perturbations in the necessary interaction of the drug with the aromatic side chains at codons 181 and 188 [2]. This explains why mutations at any one of these sites leads to a significant reduction in susceptibility to nevirapine. These findings indicate the need for new agents, the activity of which is less dependent on the interactions with the aromatic side chains, thereby reducing the number of pathways to resistance. TMC125 may be such a drug [3]. In vitro, the only way to select isolates that were highly resistant to TMC125 was to incubate an isolate that already carried the K103N and Y181C mutations in the presence of at least 200 nM TMC125. The addition of L100I and other polymorphisms (the significance of which remains unclear) conferred high-grade resistance. Clinical trials are eagerly awaited to see if these observations will be associated with a clinical benefit in vivo.

Although it has been known that the G190E mutation confers resistance to NNRTIs, it is rarely seen in vivo. This is likely due to its impact on replicative capacity. Such reduced fitness is also associated with other mutations, such as V106A, P225H, M230L and P236L [4]. This may be why these changes are less frequently seen in vivo. Now, a patient has been described who was treated with an efavirenz-containing combination [5]. A compensatory change occurred as an insertion of an arginine residue between codons 100 and 105 leading to a virus that was quite viable and highly resistant to delavirdine, efavirenz and nevirapine. It would be interesting to see how such an isolate fared against TMC125, and whether the insertion would restore the fitness of other mutants.

Primary NNRTI resistance is alive and well. First, it appears that the best tool to assess such resistance is genotypic (rather than phenotypic) testing. In one interesting study of 25 isolates with NNRTI resistance mutations, only 20 were identified as resistant by phenotypic testing [6]. In a large, North American study of 509 patients identified during the first months of infection between 1995 and the present time, the incidence of NNRTI resistance was quite stable over time, at 6.6% to 7.3%, while primary resistance to NRTIs and PIs is now <1% [7]. This may relate to the fact that genotypic changes at codons 103, 181 and 188 (accounting for the majority of cases of NNRTI resistance) are not associated with decreased replicative capacity, a finding that contrasts with the mutations that account for most of the primary resistance to drugs in other classes [8]. This is supported by the findings of a San Francisco-based study of recently infected individuals [9]. In this cohort, 8% of isolates carried primary NNRTI resistance, 63% of which was accounted for by the K103N mutation [10]. Although some of these isolates were less "fit", most were not and their carriage was not associated with a lower plasma viral load at set point, although CD4 cell counts may have been slightly higher [11].

Primary NNRTI resistance has also been measured in a number of other observational studies. Thankfully, little or no resistance has been identified in the Ivory Coast and Viet Nam, nor in 37 pregnant women studied in South Africa [12,13]. Results of other studies are summarized in the table below, showing significant variability from one country to the other, but with prevalence approaching 10% in some areas [14-20].

Let us finish on a positive note. At the Intentional AIDS Conference, Haubrich et al presented an interesting study of NNRTI hypersusceptibility (NNRTI^HS) and its significance [21]. Patients with significant prior exposure to NRTIs (but naive to NNRTIs) began a new regimen that was to include NNRTIs. NNRTIHS was defined as a fold-change in IC₅₀ of < 0.4 in the patient sample compared with a reference wild-type control sample. Of 177 patients, NNRTI^HS was detected in 24% for efavirenz, 17.5% for delavirdine, and 20% for nevirapine. The phenomenon was associated with decreased NRTI susceptibility, particularly to zidovudine and abacavir. The mean decrease in HIV-1 RNA six months after starting a new NNRTI-containing regimen was greater for the 21 patients with NNRTI^HS compared with the 77 patients without NNRTI^HS (1.2 vs. 0.8 log₁₀ copies/mL, P =0.016). If NNRTIs are not used as part if the initial treatment regimen, they may have particular benefit following virologic breakthrough, especially if broad NRTI resistance is present.

In summary, much of the new information about NNRTI resistance presented over the past weeks reminds us of the need for new drugs in this class. Such agents appear to be in development and ready to enter clinical trials. Enhanced knowledge about the mechanisms of NNRTI resistance and the interesting phenomenon of hypersusceptibility may better inform drug development and move us closer to the goal of sequencing NNRTIs, opening new treatment options for our patients living with HIV/AIDS.