Key points

 

• Adherence is the key to preventing drug resistance
• In the era of robust DTG- and PI-based regimens, and where TDF and FTC/3TC can be recycled across multiple regimens, the role of resistance testing is more limited than previously.
• Resistance testing may not detect archived mutations to particular drugs if the patient is not receiving these drugs at the time of resistance testing.

As a result of transcription errors and recombination, HIV that is replicating can accumulate mutations that lead to drug resistance. Durable viral suppression by ART is required to limit the chances of drug resistance developing. Intermittent drug adherence, as opposed to a total lack of ART, provides a greater opportunity for resistance to develop, by exposing replicating virus to sub-therapeutic ART drug concentrations. Similarly, drugdrug interactions may result in prolonged exposure to sub-therapeutic drug levels and hence the opportunity for drug resistance to develop.

ARV drug resistance mutations are summarised in Table 9.
 

A baseline resistance test is not generally indicated. We recommend a baseline resistance test to guide firstline regimen choice only where there is a recent history of sexual exposure to a person with known resistance to the InSTI drug class, or where pre-exposure prophylaxis (PrEP) with a cabotegravir-containing regimen has been taken within the past year. A resistance test is no longer indicated for patients with breakthrough infection while taking oral PrEP, as FTC/3TC and TDF/TAF can be recycled in subsequent ART regimens regardless of resistance.

Resistance testing is generally only possible if the VL is > 500 copies/mL. However, in the era of DTG- and PI-based therapy, we generally recommend it only be performed with a 2-3 consecutive VL > 1000 copies/mL, which would satisfy the definition of virological failure.

Resistance to DTG as part of initial ART is extremely rare. DTG resistance in those with a prior history of treatment failure is less so; approximately 1-3% of patients developed resistance in this scenario within 1-3 years in clinical trials. ^{22, 83, 84} In most scenarios, we recommend testing for integrase gene resistance only when a patient with virological failure has been on DTG for > 2 years. Exceptions that may warrant earlier integrase gene testing include when a patient has virological failure following a drug-drug interaction that would have substantially decreased DTG concentrations (e.g. magnesium supplementation at the same time as DTG for a prolonged period), a patient with known previous integrase resistance, or a patient with exposure to DTG monotherapy.

For PI-based regimens, sufficient resistance mutations to cause virological failure typically take at least 2 years to develop; therefore, in most cases, we recommend only performing a resistance test after the patient has been on a PI-based regimen for at least this long. Exceptions include exposure to subtherapeutic PI drug levels as a result of drug-drug interactions (e.g. not doubling the dose of LPV/r when using RIF-based TB treatment), and those with known previous PI resistance.

A resistance test at failure of NNRTI-based therapy is not routinely recommended. The EARNEST and SELECT trials showed that without the use of a resistance test to decide which NRTIs to use in second-line therapy, virological outcomes were good and equivalent to a boosted PI + RAL regimen. ^{85, 86} However, where resources permit, resistance testing may offer some advantages:

• A resistance test may offer reassurance that some first-line NRTIs can be safely recycled (e.g. ABC) if they are shown to be susceptible. Note, however, that TDF and FTC/3TC can be safely and effectively recycled even in the face of highlevel resistance, and thus testing for resistance to these agents is of little or no value. ⁴
• A resistance test may identify drug resistance to drugs (such as second-generation NNRTIs) that may be important to identify should the patient require third-line ART in future.

Most labs offer combined reverse transcriptase (RT) and >protease (PR) resistance testing with the optional addition of integrase (IN) resistance testing on request. For patients on an integrase inhibitor-based regimen, it may be possible to request integrase testing alone without RT/PR testing, depending on the lab, which reduces the cost of resistance testing. In these patients, identifying resistance to TDF or 3TC/ FTC via RT testing is no longer be of relevance since they can be recycled in subsequent regimens without significant loss of efficacy, and hence RT testing is not indicated.

Current commercial tests have been licensed for specimens with a VL of at least 1000 copies/mL. Nevertheless, many in-house assays can detect VLs of 500–1000 copies/mL. In general, most commercial HIV resistance tests detect mutations if they are present in > 10–20% of the HIV subpopulations in the sample. ⁸⁷

Common pitfall: Performing a resistance test in patients with a low or undetectable VL. Commercial assays may not be successful in samples where the VL is < 500–1000 copies/mL.

A key concept in interpreting resistance tests is archived resistance. After reverse transcription from its RNA template, HIV inserts a DNA copy of itself into the host genome. Some of the cells that HIV infects are extremely long-lived, and essentially provide an ‘archive’ of HIV variants over time. Thus, mutations that are known to have been present at one point in time can be assumed to be present for the lifetime of the patient, even if they are not visible on the patient’s latest resistance test.

A second key concept is that of the wild-type virus, which is the naturally-occurring HIV strain free of drug-resistance mutations. In most cases, this form of the virus replicates more efficiently than viral strains that have acquired resistance. Therefore, when drug pressure is removed, the wild-type forms of the virus will predominate, even though the resistant strains have been archived and can become predominant again later if the drug pressure subsequently changes in ways favourable to these strains.

• A prominent exception to this is the signature mutation of EFV and NVP, namely K103N, which imposes no significant fitness cost on the virus. Even after these drugs are stopped, the K103N strains may persist at detectable levels for several years.
• Resistance testing should, therefore, only be performed when the patient is still taking their ART regimen, or up to a maximum of 4 weeks after discontinuation.
• The absence of any identified resistance mutations implies that non-adherence is the cause for a raised VL. This does not exclude the possibility of archived resistance however, which may only become detectable once the patient is back on ART that suppresses the wild-type strain.
• Any significant drug resistance mutations identified by resistance testing can be assumed to be present for the lifetime of the patient, even if subsequent resistance testing fails to show these mutations (e.g. as a result of worsened adherence or an ART switch). Therefore, where a patient has had more than one resistance test, the results of all resistance tests should be combined.
• Conversely, it is only possible to identify mutations reliably for drugs that the patient was currently taking when the resistance testing was performed, and for drugs affected by cross- resistance. ‘Susceptible’ results to drugs for which there is no drug pressure may be unreliable due to archived resistance if the patient has been exposed to that drug or certain drugs of the same class in the past.

Common pitfall: Performing a resistance test in the absence of drug pressure. If a patient has interrupted or discontinued therapy for more than a few weeks, there is little purpose for a resistance test. In this scenario, it is highly likely that replication of wild-type virus will overtake and obscure any resistant strain, rendering mutations undetectable by conventional resistance tests.