Treatment training manual

Appendix 8: UK Resistance Guidelines (BHIVA) summary

Section 14 in the UK monitoring guidelines (BHIVA 2011) includes the current recommendations for resistance testing. The main text and table is reproduced below.

Please see the full guidelines for the context and for references.

Ref: Asboe D et al. British HIV Association guidelines for the routine investigation and monitoring of adult HIV-1-infected individuals 2011. HIV Medicine (2012), 13, 1–44.
Download PDF file.

14. Resistance testing

14.1 Initial HIV-1 diagnosis

The prevalence of antiretroviral drug resistance among treatment-naïve patients in the UK is around 8% [1].

Although previous estimates may have been confounded by selection bias, prevalence rates have been declining over recent years [2]; however, rates are now showing a possible slight increase.

While the highest rates of resistance are seen in patients born in the UK [3], rates are increasing in countries currently expanding access to ART [4–6] and may soon start to rise among immigrant populations as a result [7]. In some cases, the presence of resistance in an apparently treatment-naïve patient may in fact reflect previous undisclosed therapy.

There is increasing evidence to indicate that transmitted resistance negatively impacts on treatment responses, particularly in the context of nonnucleoside reverse transcriptase inhibitor (NNRTI)-based regimens [8–17].

Most transmitted drug resistance affects reverse transcriptase and protease inhibitors (PIs), although transmitted integrase inhibitor resistance has started to emerge.

Although transmitted resistance often remains detect- able in plasma for several years, gradual reversion to low-frequency and archived mutants occurs over time [18–24]. Reversion may occur through intermediates (or ‘revertants’, e.g. T215D/N/S from T215Y/F).

Genotypic tests should therefore be used in treatment-naïve individuals as they allow the detection of such mutations that do not confer phenotypic resistance but may signal the presence of more substantial resistance. Detection of such revertants should be interpreted as an indication that fully resistant mutants are present as either low-frequency quasispecies or archived resistance.

Both genotypic and phenotypic resistance assays provide results based on the majority population of circulating viruses at the time of sampling. The level of detection of mutant viruses is around 20–30% of the population in genotypic assays and probably less in phenotypic assays. Low-frequency mutants can impact negatively on responses to therapy in the context of NNRTI-based regimens (reviewed in [12, 15–17, 25, 26]). Assays with increased sensitivity for detection of resistance mutations are under development but can be considered primarily as research tools in most circumstances at the current time [16].

14.1.1 Recommendations (Table 14.1)

  • Testing for resistance is recommended in all newly diagnosed patients. This includes patients with acute seroconversion, established infection or infection of unknown duration, regardless of demographic characteristics, ethnicity or risk group (Ia).
  • Baseline resistance testing should include the polymerase and protease genes. Testing for susceptibility to integrase and entry inhibitors is not recommended routinely in naïve patients at present, although this area is kept under active review (IIb).
  • The most appropriate sample is the one closest to the time of diagnosis (Ia) and this should preferably be tested at the time of initial presentation (IV).

Table 14.1 Recommendations on when to perform resistance testing

14.2 ART-naïve

The possibility exists that the resistance profile obtained at diagnosis may change in patients who acquire a new infection. The true risk of HIV-1 superinfection remains to be determined but may be significant in persons who continue to be exposed to new sources of the virus [27], especially in early stages of the initial infection [28]. Triggers to repeat resistance testing prior to starting ART may include a sudden increase in viral load, a sudden drop in the CD4 T-cell count, and a recurrence of symptoms of acute HIV infection [29,30]. It should be noted, however, that most patients with sudden changes in viral load and CD4 T-cell counts do not have evidence of superinfection [29,30]. In a London cohort study of 47 homosexual men who showed an increase in viral load of greater than 0.5 log10 copies/mL during routine monitoring, two (4%) showed evidence of superinfection and a change in the initial drug susceptibility profile as determined by repeat sequencing of the reverse transcriptase and protease genes [30].

14.2.1 Recommendations (Table 14.1)

  • For patients who have not undergone resistance testing at the time of diagnosis, testing is recommended before starting therapy (Ia). Whenever possible, a plasma sample collected as close as possible to the time of diagnosis should be retrieved for retrospective testing (Ia). When a stored sample is not available a current sample should be tested (IV).
  • Following resistance testing at the time of diagnosis, repeat testing is not routinely recommended prior to starting therapy, although it should be considered in selected persons who may have experienced reinfection (IIb).

14.3 Post treatment initiation

In patients without evidence of drug resistance by routine methods, a suboptimal virological response to first-line therapy (a viral load reduction of less than 1 log10 copies/mL by 4 weeks) may signal the emergence of drug-resistant variants that were initially present at low frequency and therefore undetectable by routine testing.

14.3.1 Recommendations (Table 14.1)

  • In patients without evidence of drug resistance at diag- nosis by routine genotypic methods, a suboptimal viro- logical response to first-line therapy (a viral load reduction of less than 1 log10 copies/mL by 4 weeks) should prompt resistance testing at that time (IV).

14.4 ART-experienced

The prevalence of drug resistance has declined among treatment-experienced patients in the UK as a result of improved management of ART and treatment failure. Currently, approximately half of treated patients undergoing testing show evidence of resistance and around 3% have evidence of triple-class resistance affecting the nucleoside reverse transcriptase inhibitors (NRTIs), NNRTIs and PIs [1]. There are no national data on the prevalence of resistance to integrase and entry inhibitors, but integrase inhibitor resistance in particular is expected to grow with expanded use of the class. Patients who experience virological failure while on chemokine receptor 5 (CCR5) antagonists may show a change to chemokine receptor 4 (CXCR4)-using virus upon repeat tropism testing, or maintain the R5 tropism. In approximately one-third of R5 failures, the virus exhibits phenotypic resistance to the antagonist. Although certain mutations in the glycoprotein 120 (gp120) V3-loop appear to play a key role, the genotypic predictors of the resistance profile have not been clearly elucidated. Therefore, genotypic resistance testing is not routinely recommended for patients failing CCR5 inhibitor treatment at this time [31–34].

While it is recommended that confirmation of virological rebound is obtained in patients with previously undetect- able viral load prior to performing a resistance test, it should be noted that mutations conferring or increasing resistance may accumulate if a patient is left on a failing regimen [35]. Resistance testing of viral load ‘blips’ (defined as a single viral load measurement greater than 50 copies/mL preceded and followed by values less than 50 copies/mL) is unlikely to yield significant information [36], whereas testing of confirmed low-level viraemia is highly informative [37–39]. Whereas a viral load cut-off of 1000 copies/mL has been traditionally recommended for resistance testing, specialized testing can achieve high success rates at lower levels of viraemia [37–39].

Resistant mutants selected during therapy are rapidly outgrown by wild-type virus once therapy is discontinued [40]. To be informative, resistance testing should therefore be performed on samples taken while the patient is still on therapy. Resistance testing undertaken when a patient has discontinued therapy for more than 2 weeks should be interpreted with caution as the extent of underlying resistance is likely to be underestimated. Despite the apparent disappearance of resistance, however, resistant mutants persist at low frequency in the plasma quasispe- cies and as archived resistance in latently infected cells [41], and can re-emerge rapidly if selective pressure is reintroduced. Therefore, resistance should be considered long-lasting. Interpretation of resistance should take into account the results of all tests performed during the patient’s treatment history (‘cumulative genotype’) [42]. Patients who simultaneously interrupt all drugs in an NNRTI-based regimen are likely to experience a prolonged period of NNRTI monotherapy with a resulting risk of resistance that may or may not be detectable by routine methods, but may have an effect on treatment responses once NNRTI-based therapy is resumed [43–45]. Pending further data, the potential emergence and impact of NNRTI resistance should be taken into consideration in these patients [46].

The interpretation of resistance test results is complex. Although informative interpretation systems have been developed for both genotypic and phenotypic results, none is entirely accurate, and all are subject to change as new data become available. Interpretation is especially difficult with new drugs and this problem affects both genotypic and phenotypic resistance assays. Expert advice should be sought with complex or unusual resistance profiles. Sufficient information on treatment history should be provided to optimize interpretation of resistance test results in the laboratory.

14.4.1 Recommendations (Table 14.1)

  • Viraemia should be confirmed before performing a resistance test in treated patients (IV). However, further assessment should be undertaken promptly because of the risk of accumulation of mutations, particularly in patients taking regimens with a low genetic barrier (IIb).
  • Resistance testing is recommended in all treated patients experiencing confirmed viraemia and changes in therapy should be guided by the results of resistance testing in these patients (Ia).
  • For patients showing viraemia while receiving integrase inhibitors or enfuvirtide (T20), resistance testing should be undertaken promptly in laboratories offering the tests (IIb).
  • For patients experiencing viraemia while receiving CCR5 antagonists, repeat tropism testing should be performed (Ia). If the virus is confirmed as R5, the presence of resistance to CCR5 antagonists should be suspected (Ia), although testing for this is not routinely available at present.
  • The level of viraemia at which resistance testing can be performed reliably is just above 50 copies/mL in many specialised laboratories. Resistance testing where viral load levels are less than 1000 copies/mL can provide useful information and clinicians are encouraged to discuss and agree the required viral load cut-off for testing with their service providers (IV). Laboratories should review the optimal methodology for resistance testing at low viral load levels (III).
  • Resistance testing should preferably be performed on samples taken while the patient is still on therapy (IIb).
  • Resistance testing by routine methods is not recommended after unstructured interruption of NNRTIs because of suboptimal sensitivity in this context (IIa), although selection of NNRTI resistance should be considered possible (IIb).
  • Resistance test results should be interpreted in the context of the patient’s entire treatment history and the results of all tests performed in a patient should be taken into account to guide optimal treatment selection (IIb).

Last updated: 1 September 2014.