INTRODUCTION
Drug resistance testing has been used for many years to support the treatment of infectious diseases. In virology, resistance testing has been constrained, until recently, by a lack of appropriate technology, and difficulty in routinely propagating relevant viruses. Gene sequencing was established as a research tool over 30 years ago and now academic and commercial groups have developed and modified research technologies to form ‘kits’ and/or systems for evaluation for clinical use.
The challenge is to develop a clinical ‘service’ that will detect or predict changes in susceptibility associated with antiretroviral drug resistance, using a system that is rapid, reliable, reproducible and quality assured.
Assuming that technologies available withstand this scrutiny we are left to determine how results of resistance testing are related to virological and clinical outcomes, and their relevance for improving patient care.
Intuitively it would seem obvious that an understanding of HIV-1 drug resistance profiles should help in the management of their therapy. However, within the setting of a chronic viral infection where drug responses are equally influenced by adherence, pharmacological tolerance and drug resistance it may be difficult to demonstrate this unequivocally.
Antiretroviral drug resistance has been described retrospectively, in vitro and in vivo, to be associated with poor virological and clinical outcomes (Table 1a) and short-term prospective, randomized controlled trials have, in some studies, demonstrated the short term benefits of resistance testing over standard of care (Table 1b). It remains to be clarified in which patient groups resistance testing is of long term benefit, or indeed if resistance testing without measures of adherence and/or drug pharmacokinetics, in the long term, is likely to be useful at all.
TABLE 1
- Retrospective studies
Study |
Resistance test |
Significant outcomes |
1. 1994: CONCORDE SUBSTUDY |
Genotyping |
ZDV MTs predicted virological and clinical outcome |
2. 1995: ACTG 116B/117 |
Phenotyping |
ZDV resistance associated AIDS/death |
3. 1995: ACTG 116B/117 |
Genotyping |
ZDV MTs associated with AIDS/death |
4. 1998:FRANKFURT COHORT |
Phenotyping & Genotyping |
Resistance ZDV/3TC associated with failure |
5. 1999: UNIVERSITY CLINIC |
Genotyping |
PI MTs associated with RIT-SAQ failure |
6. 1999: HOSPITAL CLINIC |
Phenotyping |
Number of PI/NRTI sensitive associated with positive response |
7. 1999:SWISS HIV COHORT |
Genotyping |
PI/NRTI MTs predict virological response |
8. 2000: FRANKFURT COHORT |
Phenotyping |
PI/NRTI sensitivity associated with virological response to Mega-HAART |
9. 2000: HOSPITAL CLINIC? |
Genotyping |
MT at baseline to a new NRTI predicts virological failure of new regimen |
- F Brun-Vezinet,S Kaye, F.Ferchal, C Loveday,C, Buffet-Janvresse, R Tedder, J Darbyshire, JP Aboulker AZT Resistance:A Case Controlled Study from Concorde. Third Int. Workshop on HIV Drug Resistance.Kauai, Hawaii. 2-5 August 1994. Abstr.49
- D’Aquila RT,Johnson VA,Welles SL et al.ZDV resistance and HIV-1 disease progression during antiretroviral therapy.Ann Int Med.1995, 122: 401-408
- Japour AJ, Welles S, D’Aquila RT et al. Prevalence and clinical significance of ZDV resistance mutations in HIV isolates from patients after long term ZDV therapy. J Infect.Dis.1995. 171:1172-1179
- Miller V, Phillips, Rottman C et al. Dual resistance to ZDV and 3TC in patients treated with ZDV/3TC combination therapy: association with therapy failure. J Inf ect Dis. 1998, 177: 1521-1532
- Zolopa AR, Shafer R, Warford A et al. HIV-1 genotypic resistance patterns predict response to saquinavir/ritonavir therapy in patients in whom previous protease inhibitor therapy failed. Ann Int Med. 1999, 131: 813-821
- Deeks SG, Hellman NS, Grant RM et al. Novel 4 drug salvage treatment regimens after failure of HIV-1 protease inhibitor-containing regimen: antiviral activity and correlation of baseline phenotype drug susceptibility to virological outcome. J Infect Dis, 1999, 179: 1375-1381
- Lorenzi P, Opravil M, Hirschel B et al. Impact of drug resistance mutations on virological response to salvage therapy. AIDS, 1999, 13: F17-21
- Miller V, Cozzi-Lepri A, Hertogs K et al. HIV drug susceptibility and drug response to Mega-HAART regimen in patients from the Frankfurt cohort. Antiviral Therapy 2000, 5:56-71
- Van Vaerenbergh K, Van Leuthem K, Van Wijngaerden E et al. Baseline HIV-1 genotypic resistance to a newly added nucleoside analogue is predictive for virological failure of the new therapy. AIDS Res Hum Retroviruses 2000. 16: 529-537
- Prospective Studies Resistance Test
Study |
Resistance test |
Significant outcomes |
1. 1999: VIRADAPT (RCT) |
Genotyping |
Better virological response after 6 months (p |
2. 1999: VIRADAPT (open Study) |
Genotyping |
Benefits of genotyping maintained for 1 year |
3. 1999: GART (RCT) |
Genotyping |
Better virological response after 8 weeks (p |
4. 2000: VIRA 3001 (RCT) |
Phenotyping |
Better virological response after 16 weeks (p |
5. 2000: HAVANA (RCT) |
Genotyping |
Better virological response after 6months(p |
6. 2001: NARVAL (RCT) |
Phenotyping |
No difference for SOC & Genotyping, phenotyping but better virological response after 6 months (p |
MTs = mutations; PI = protease inhibitors; NRTI = nucleoside reverse transcriptase inhibitors; SOC = standard of care; RCT = randomized controlled trial
- Durant J, Clevenbergh F, Halfon F et al. Drug resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet. 1999, 353: 2195-2199
- Clevenbergh F, Durant J, Halfon F et al. Persisting long-term benefit of antiretroviral genotypic guided treatment for HIV infe4cted patients failing HAART. Antiviral Ther. 1999 (Suppl 1) 42: Abstr 60
- Baxter JD, Mayers DL, Wentworth DN et al. A randomised study of antiretroviral management based on plasma genotypic resistance testing in patients failing therapy.CPCRA 046 study team for the Terry Beirn Community Programs for Clinical Research on AIDS. AIDS 2000. 14: F83-93
- Cohen C, Kessler H, Hunt S et al. Phenotypic resistance testing significantly improves response to therapy: final analysis of a randomized controlled trial (VIRA 3001). Antiviral Ther. 2000. 5 (Suppl 3): 67
- Havana – CROI, 2001
- Meynard JL, Wray M, Mourand-Joubert L et al. Impact of treatment guided by phenotypic and genotypic resistance tests on the response to antiretroviral therapy: a randomized controlled trial (NARVAL,ANRS 088). Antiviral Ther. 2000, (Suppl 3): 67-68
THE TESTS
Resistance tests available for determination of of HIV-1 drug susceptibility (phenotyping) and detection of mutations associated with drug resistance (genotyping) are amongst the most complex ever addressed for use in routine clinical care. In terms of reviewing performance and quality these issues divide naturally into:
- Clinical sample and request. This is an important determinant of the ability to generate a useful result (Table 2)
- The technical process. A range of different assays are available, summarised in Table 3
- Validation of the technical process and printout. Assay quality control means the assimilation of extra samples at all stages of the test process, in parallel with clinical samples, in order to identify changes in sensitivity in the assay, and the generation of potentially false results. These extra controls increase the overall cost of the test. External quality assurance schemes test the quality of the overall process, from receipt of sample in the laboratory, to production of the final report. Such schemes are being developed for HIV drug resistance.
All these quality aspects of the test are assessed as part of laboratory accreditation procedures.
- Generation of clinical report. Genotypic assays produce large volumes of detailed gene sequence information, from which must be identified those changes thought to be important in determining drug susceptibility. This comprises two steps. Firstly, the identification of any mutations which are different from the presumed "drug sensitive" virus, and secondly, providing an interpretation of these changes, with regard to specific drugs.. Software programmes are available to undertake these analyses, however they need regular updating, in the light of new knowledge on drug resistance, and differences between them still exist. A selection of programmes available for providing an interpretation of drug resistance associated mutations is given in Table 4. Table provides guidance on the association of specific mutations with resistance to specific drugs.
- Clinical interpretation and advice. The factors involved in optimizing the clinical utility of resistance test results are given in Table 6.
Table 2
Practicalities in requesting resistance test
1. Request test on sample taken before starting first-line therapy, before stopping therapy or before changing therapy
2. In drug-experienced patients, do not request test when off therapy >2 weeks. In such circumstances, consider testing stored samples from the time of drug failure.
3. Request test on sample with viral load >1,000c/ml. EDTA.
4. Attempt to construct at least one specific question before requesting resistance tests, since this will assist in interpretation of results.
5. Discuss with the laboratory providing the test the most appropriate format for reporting.
6. Since testing of stored samples is likely to become increasingly requested in drug-experienced patients, close liaison with the laboratory undertaking viral load testing is required to ensure that relevant samples are stored appropriately.
Table 3
(a)
AVAILABLE TESTS
Genotype
Techniques |
Advantages |
Disadvantages |
Examples and Availability |
|||||
Identification of specific changes in relevant areas of the HIV genome associated with drug resistance. This is undertaken by gene sequencing or hybridisation |
Rapid (turnaround time 2-3 weeks). Detects minority population of virus (down to ~10-20%), so leads to early identification of resistance. Moderately expensive (£200-£300). Key mutations clearly associated with drug failure |
Hybridisation based assays do not detect all possible mutations. Complex pattern of mutations difficult to interpret. |
Sequencing Hybridisation |
Applied Biosystems (Viroseq) Visible Genetics (Trugene) VIRCO Virologics In-house Innogenetics (Innolipa) Affymetrix |
Clinical Virology Laboratory Clinical Virology Laboratory Commercial & Clinical Virology Laboratory ???
Commercial Clinical Virology Laboratory Clinical Virology Laboratory Clinical Virology Laboratory |
Table 3
(b)
Phenotype
Techniques |
Advantages |
Disadvantages |
Examples and Availability |
||||
Classically undertaken by infection of PBMC, but in the main surpassed by producing a recombinant virus incorporating relevant genes from patient sample, which can be tested for drug susceptibility |
Quantitative values for fold resistance against all drugs generated Teminology to clinicians |
Phenotypic resistance may only become apparent after appearance of specific mutations Expensive (£400-£500) Time consuming (>4 weeks) Relevance of assay cut-offs, and biological cut-offs to clinical practice unclear |
VIRCO (Antivirogram) Virologics (Phenosense) Viralliance ( Phenoscript) |
Commercial Laboratory Commercial Laboratory Commercial Laboratory |
Table 4
Some sources of information assisting in interpretation of resistance data
Guidelines by the International AIDS Society [137] |
French guidelines (Error! Reference source not found.) |
Guidelines by the US Dept of Health and Human Services (Error! Reference source not found.) |
British Guidelines (Error! Reference source not found.) |
Schinazi website table (Error! Reference source not found.] |
Los Alamos website table (Error! Reference source not found.) |
Stanford website database (Error! Reference source not found.] |
Published specific genotypic algorithms [64,65,120] |
Stand alone software based on specific genotyping algorithms (such as the Retrogramñ software, Error! Reference source not found.) |
Commercial interpretation systems (such as the VirtualPhenotype, Error! Reference source not found.) |
Association of specific mutations with drug resistance
Table 5
Most common HIV-1 resistance mutations against currently available drugs
Name (Abbreviation) |
Time to resistance (1) |
Most common resistance mutations in the target protein (2) |
Nucleoside analogue reverse transcriptase inhibitors (NRTIs) |
||
Zidovudine (AZT) |
Intermediate |
M41L, D67N, K70R, L210W, T215Y/F, K219Q (combined mutations needed for high level resistance) |
Didanosine (ddI) |
Long |
K65R, L74V, M184V |
Zalcitabine (ddC) |
Long |
K65R, T69D, V75T, M184V, (L74V) |
Stavudine (d4T) |
Long (insufficient data) |
V75T M41L, D67N, L210W, T215Y/F, K219Q |
Lamivudine (3TC) |
Fast |
M184I/V, (K65R) |
Abacavir |
Intermediate (not enough in vivo data) |
K65R, L74V, Y115F, M184V M41L, D67N, L210W, T215Y/F, K219Q (combined ZAMS? mutations needed for high level resistance) |
Most common NRTI cross resistance mutation patterns (3) |
(A62V, S68G, V75I, F77L, F116Y, Q151M, combined mutations) (T69S-XX, usually T69S-SS, in combination with other NRTI mutations) |
|
Non-nucleoside analogue reverse transcriptase inhibitors (NNRTIs) |
||
Nevirapine |
Fast |
A98G, L100I, K103N, V106A, V108I, Y181C/I, Y188C, G190A |
Delavirdine |
Fast |
K103N/T, Y181C, P225H,P236L |
Efavirenz |
Fast |
L100I, K101E/Q, K103N, V106A, V108I, G179N/E/D,Y188C/H/L, G190E/A/S, P225H (combined mutations needed for high level resistance) |
Most common NNRTI cross resistance mutations (3) |
K103N, Y181C |
|
Protease inhibitors (PIs) |
||
Saquinavir |
Intermediate |
L10I, G48V, I54L/M/V, L63P, G73S, V82A, L90M (combined mutations needed for high level resistance) |
Ritonavir |
Intermediate |
L10I, K20R, V32I, L33F, M36I, M46I/L, I54L/M/V, L63P, A71V, V82A/F/S/T, I84V, L90M (combined mutations needed for high level resistance) |
Indinavir |
Intermediate |
L10I/R/V, K20R/M, L24I, V32I, M46I/L, I54L/M/V, L63P, A71T/V, G73S, V82A/F/T, I84V, L90M (combined mutations needed for high level resistance) |
Nelfinavir |
Intermediate |
D30N, M36I, M46I, L63P, A71V, V77I, I84V, N88D, L90M (combined mutations needed for high level resistance) |
Amprenavir |
Intermediate? (not enough in vivo data) |
L10F/I/V, V32I, L33F, M46I/L, I47V, I50V, I54V/L/M, V82I, I84V (combined mutations needed for high level resistance) |
Lopinavir |
(not enough data) |
Any 1 of M46I/L, V82A/F/L, I84V, L90M + any 7 of L10I/R/V, K20M/R, L24I/V, M46I/L, F53L, I54L/M/V, L63P, A71T/V, V82A/F/V, I84V, L90M |
Notes
1. Key mutations, clearly associated with reduced drug susceptibility, are in bold
2. Our understanding of the impact of new mutations and interaction of mutations is ever changing. Therefore, the table above should be viewed as general guidance only, rather than a definitive list of all mutations.
3. The "time to resistance" column provides relative information for the particular drugs, rather that specific time periods. The actual time taken for resistance to develop to any one particular drug is determined by a multitude of factors such as co-exisiting therapies, adherence and pharmacological factors.
(1) Time to resistance is very variable and highly dependent on the efficacy of the combination therapy. Therapy choice should not primarily be based on a longer period to resistance since this is often associated with a less powerful drug. Fast: resistance generally develops within weeks of initiation of monotherapy, intermediate: resistance generally develops within a few months of initiation of monotherapy, long: resistance has either been poorly described in the literature or has developed only years after the initiation of monotherapy.
(2) Specific drug related mutations indicated are the ones observed under monotherapy. Mutations shown in bold are generally considered either primary or key mutations, their clinical use in patient follow-up is currently still under investigation. Mutations between brackets are cross-resistance mutations selected during therapy with another drug.
(3) Combination therapy can select for the mutations mentioned as cross-resistance mutations.
(4) There is increasing evidence for at least some degree of cross-resistance between AZT and d4T, as well as other nucleoside analogues.
(5) The phenotypic implications of these mutations have not been added, since assay cut-offs and their biological and clinical implications are still unclear. Clinically relevant cut-offs are only available for lopinavir, for which 10-fold reduced susceptibility is associated with an increased risk of non-response.
Aminoacid codes are: A (alanine), C (cysteine), D (aspartic acid), E (glutamic acid), F (phenylalanine), G (glycine), H (hystidine), I (isoleucine), K (lysine), L (leucine), M (methionine), P (proline), Q (glutamine), R (arginine), S (serine), T (threonine), V (valine), W (tryptophane), Y (tyrosine).
The data in this table are adapted from reference 1. [Vandamme A-M, Houyez F, Banheghi D et al. Laboratory guidelines for the practical use of HIV drug resistance tests in patient follow up. Antiviral Therapy 2001 (in press)].
Table 6
Aspects to consider when interpreting resistance results
1. For genotypic assays, the "weight" of specific mutations for causing resistance is dependent on the mutation and drug in question. In addition, mutations can interact to enhance or suppress resistance to specific drugs.
2. For phenotypic assays, the precise fold resistance required to confer a lack of drug response in vivo remains unclear for most drugs.
3. Adherence is very likely to impact on the resistance result; however, the precise manner in which differing degrees of adherence have an effect remains unclear.
4. Consider drug history. Previous drug failure may have been accompanied by mutations not detectable at a later time point. It should be assumed that these mutations remain archived and can rapidly re-emerge within the majority of virus population.
5. Discussion of results in the context of clinical picture with a virologist/clinician with expertise and experience in this area is strongly recommended. This will enhance the body of knowledge and experience within each clinical centre.
6. Geographical origin of virus. The interpretation of resistance mutations may be different between subtype B and other virus subtypes. If the virus is known to be of a non-B subtype, or this is suspected through the likely geographical origin of infection, then expert advice on interpretation should be sought.
TABLE 7 RECOMMENDATIONS
Patient Group |
Recommendations |
Comment |
Primary HIV Infection (PHI) |
Recommend |
Within the UK, a national surveillance programme provides real-time resistance testing free of charge, to monitor transmission of drug resistant viruses (2). If therapy indicated, use results to guide therapy, or modify therapy if result obtained after initiation. |
Drug naïve, chronic infection |
Consider |
Resistance result may not reflect presence of transmitted resistant virus due to ongoing virus evolution, so testing may underestimate presence of resistance. Also consider testing baseline specimen if poor initial response to therapy |
Drug experienced- 2nd line or salvage |
Recommend |
Interpret in light of clinical and therapeutic history |
Pregnancy |
Recommend |
Do not delay therapy where late maternal presentation, and modify therapy when result obtained |
Paediatrics |
Recommend |
Indicated prior to therapy if mother had detectable viral load while receiving therapy, and also at times of therapy failure |
Post exposure prophylaxis |
Recommend |
Undertake if sample available from source patient, but use result to modify therapy rather than delay onset of PEP. |
Notes:
1) Ensure that laboratory undertaking viral load assays store aliquots of plasma samples appropriately, in order that retrospective resistance testing can be undertaken from relevant time points e.g prior to therapy, at times of previous drug failure
2) UK Collaborative Group on Monitoring Transmission of HIV Drug Resistance, is a collaboration between the PHLS and UK HIV Seroconverter Register. Plasma samples should be sent to PHLS Antiviral Susceptibility Reference Unit, Birmingham Public Health Laboratory, Birmingham Heartlands Hospital, Birmingham B9 5SS