 |
|
|
| |
Genetic Screening Test for Abacavir Hypersensitivity -- PREDICT-1: Description of a Novel Randomised Prospective Study Design to Determine the Clinical Utility of Prognostic Screening for HLA-B*5701
|
| |
| |
Reported by Jules Levin
8th International Congress on Drug Therapy in HIV Infection
November 12-16, 2006, Glasgow, UK
S Hughes, K Parry-Billings, N Givens, D Wong, E Loeschel, P Parada, D Zamoryakhin, C Goodwin, A Hughes, J Muir, L Curtis and D Thorborn
GlaxoSmithKline, United Kingdom
INTRODUCTION
The abacavir (ABC) hypersensitivity reaction (HSR) is a well characterised adverse event, affecting approximately 5% of individuals exposed to ABC in clinical trials. Retrospective analyses have identified HLA-B*5701 carriage as the dominant risk factor for ABC HSR. More recently, the deliberate avoidance of
ABC in adults carrying the HLA-B*5701 allele resulted in a dramatic reduction in the incidence of ABC HSR in the Western Australian (Rauch 2006) and Brighton, UK (Reeves 2006) cohorts. It is clear that observational cohorts and case controlled studies do provide valuable information on target genes of interest
and when significant associations are established between genes and response to ART this may help guide prescribing decisions.
However in the hierarchy of clinical research designs, the results of randomised
controlled trials are considered the highest level of evidence as randomisation/stratification is the only method for controlling for both known and unknown prognostic factors. In the case of HLA-B*5701 screening, expert opinion suggests that the clinical utility of HLA-B*5701 screening remains to be proven in a randomised clinical trial (BHIVA 2006) and that only well designed prospective clinical trials can answer specific questions relating to HLA-B*5701 screening (Fox 2006).
As the PREDICT-1 study (described here) is the first randomised prospective study to assess the clinical utility of genetic screening in any patient population, then the burden of proof needs to be high. The robust study design presented here should provide the scientific, medical and patient communities with confidence that the results will be a true reflection of the clinical utility of screening for HLA-B*5701 in HIV-1 infected adults. The results of this ongoing study will be presented in 2007.
Study Overview
The PREDICT-1 study (CNA106030) is a phase IV, randomised, multi-centre, double-blind study to evaluate the clinical utility of prospective genetic screening (HLA-B*5701) for susceptibility to ABC HSR (see figure 1).
Co-primary objectives:
- To determine if screening HIV-1 infected adults for HLA-B*5701 prior to treatment with ABC-containing highly active antiretroviral therapy (HAART), results in a significantly lower incidence of clinically-suspected HSR compared with current standard of care, which does not include genetic screening.
- To determine if screening HIV-1 infected adults for HLA-B*5701 prior to treatment with ABC-containing HAART, results in a significantly lower incidence of immunologically-confirmed HSR (clinical diagnosis plus epicutaneous patch testing [EPT]) compared with current standard of care, which does not include genetic screening or EPT.
Randomisation & HLA-B*5701 Screening
ABC-naïve subjects are randomised 1:1 (following the return of routine laboratory screening results) to receive or not receive a prospective genetic screen for HLA-B*5701. Randomisation is stratified by (i) caucasian vs. non-caucasian subjects; (ii) ART-naïve vs. ART-experienced; (iii) intention to introduce NNRTIs by or at Day 1. Subjects randomised to receive local SOC without prospective screening will receive retrospective HLA-B*5701 screening at study end. For subjects assigned to the prospective screening arm of the study, the randomisation process triggers the screening of samples collected during the laboratory
screening visit. To allow for blinding, subjects do not proceed to their planned Day 1 visit until the investigator is notified that the subject should proceed. HLA-B*5701 screening is performed by two methodologies on all subjects: full allelic typing (Perth, Australia) and Sequence-specific oligonucleotide probe technology, SSOP/sequence based typing (Labcorp, US). HLA-B*5701 positive subjects randomised to the test arm will be excluded from progressing to their Day 1 visit and receiving ABC as part of the study.
Epicutaneous Patch Testing (EPT)
EPT is performed on all subjects with clinically suspected HSR no earlier than 6 weeks after diagnosis and on 100 ABC-tolerant controls shortly after their Week 6 visit, according to a modification of the method of Phillips et al. (2002). All EPT materials are provided from a central location as single-use standard kits containing sufficient dermatological patches, swabs and contactant preparations for one test with readings at 24 and 48 hours. Contactants comprise 1% and 10% abacavir (as ground Ziagen commercial tablet formulation) in white petrolatum, plus petrolatum and Ziagen excipient controls. High definition digital photographs are taken before and after removal of the patches (see figure 2), and assessed by an independent clinical evaluation committee blinded to the status of the subject (HSR or ABC-tolerant) and the contents of the individual wells.
Statistical Methods
Sample Size Calculation
Assumed underlying incidence of clinically-suspected HSR in the overall study population is 8%:
- 5% immunologically-confirmed HSRs + 3% 'false positives'
- 91% of HSRs seen in 6 week study = apply 0.91 of all these estimates
Total planned sample size of 1806 randomised subjects provides at least 90% power to detect:
- a 50% reduction from 7.3% to 3.6% in clinically-suspected HSRs
- an 80% reduction from 4.6% to 0.91% in immunologically-confirmed HSRs
Based upon the expected recruitment pattern for the study and the hypothesised high rates of ABC HSR in a Caucasian population, it is calculated that the study also has at least 90% power to test the co-primary endpoint in the Caucasian population.
Analysis Methods
HSR rates will be compared using logistic regression, adjusting for randomisation strata. Stepwise regression will also be performed to investigate a fuller set of covariates of interest. In addition to comparing HSR rates between arms, the sensitivity, specificity, negative and positive predictive values will be calculated.
Recruitment
1994 subjects were screened between 24th April - 29th September 2006 in 262 sites across 19 countries. The following table represents the numbers of subjects screened by country. Each row also highlights the centres with the highest number of screened subjects.
DISCUSSION
The ability to identify patients at risk of a severe adverse event as a result of HAART or those that may gain a specific benefit from a treatment modality is the current goal of pharmacogenetics within the field of HIV medicine. To date, genetic variation has been linked to the metabolism of antiretroviral drugs, drug hypersensitivity and HIV associated lipodystrophy. Most of these investigations have been hypothesis generating although a small number of observational studies have reported an impact on local clinical practices. Results from previous clinical studies on HLA-B*5701 and its association with ABC HSR may be confounded by study populations, design issues or case ascertainment, therefore
the clinical applicability of these previous data to a non-research setting within a broad population is unclear.
This is the first randomised, blinded, prospective study designed to determine the clinical utility of screening for a specific pharmacogenetic marker, HLA-B*5701, in the management of HIV patients. This design could be adapted by others to determine the utility of pharmacogenetic screening and the use of other medicines.
In the absence of definitive data from well designed clinical studies, GSK does not recommend routine use of prognostic screening for HLA-B*5701 outside of a clinical research setting. Pharmacovigilance remains the cornerstone of successful ABC risk management.
CONCLUSIONS
When the burden of proof is high, particularly relating to the safety of HAART, then only the most rigorous testing of a hypothesis will assure the medical, scientific and regulatory communities that the data are credible.
PREDICT-1 is a unique landmark pharmacogenetic study. It is the first trial of its kind in any therapy area, designed to help clinicians to ensure a safer, more rational and individualised use of treatment.
The use of randomisation, stratification, blinding and an independent committee to assess specific endpoints, were considered key design elements.
This is one of the largest prospective randomised HIV clinical studies to be undertaken (∼2,000 screened) and demonstrates that large prospective clinical trials can be recruited rapidly in centres with highly motivated healthcare professionals and well informed study volunteers.
|
| |
|
|
|
|
|
