The high rate of replication that is found throughout the course of HIV infection and the variability of HIV, coupled with the relative inaccuracy of the enzyme HIV reverse transcriptase, are the main reasons for the frequent occurrence of copying errors in the transcription of viral genetic information. HIV replicates at the rate of around 10⁸ to 10¹⁰ virus particles per day, probably giving rise daily to about 3 × 10^-3 spontaneous changes (mutations) in its genetic sequence. The ultimate size of a viral population containing a mutation is probably determined by three concurrent factors: the forward mutation frequency, the replicative capability of the mutated virus and the “age” of the viral population containing the mutation i.e. how long ago this population was generated. With the on-going production of genetic variants of HIV there is then a continuous selection for the “fittest” virus population.

Sub-optimal ART regimens that allow replication of HIV to continue in the presence of antiretroviral drugs, encourage the growth of viral populations that are carrying a genetic mutation which protects against these drugs. It is likely that many of these drug resistance mutations already exist before any antiretroviral drug is introduced and are further encouraged to proliferate under the selective pressure exerted by drug treatment.

Antiretroviral therapy can minimise the emergence of drug resistance in two ways:

• by maximising and sustaining the suppression of viral replication
• by using drugs where multiple mutations are required before resistance can occur.

In recent years laboratory testing for antiretroviral drug resistance has become available raising the possibility of using resistance testing to guide therapeutic choices. The testing methods in use, however, are still hampered by technical complexity, poor sensitivity, difficulties in interpretation and high cost. The place of resistance testing in every day clinical practice remains to be clearly defined because while it appears useful in patients experiencing treatment failure, its utility in other situations (treatment naive patients, patients who have failed on multiple regimens or pregnant HIV-positive women) is still under investigation. When performing testing for antiretroviral drug resistance, it is important to ensure that this is done while the patient is still on therapy in order to maximise accuracy.

Cross resistance among the available classes of antiretroviral drugs is common and is an important consideration when assessing the possibility of sequencing (replacing one drug with another) should it become necessary to change a therapeutic regimen (Table 3). Cross-resistance implies that a population of virus resistant to one drug in a class is also resistant to other drugs of the same class. This is particularly liable to occur with the NNRTIs especially if they are used as part of a regimen that produces incomplete suppression of viral replication. The NNRTIs in general present a very low “genetic barrier” to resistance because a single mutation is sufficient to produce resistance.

PIs and NRTIs are more robust in this respect since multiple mutations are required to confer resistance to drugs in these classes.