| |
Genetic Screening to Prevent Abacavir Hypersensitivity Reaction: Are We There Yet?
|
| |
| |
Clinical Infectious Diseases July 1, 2006;43:103-105
Elizabeth J. Phillips
Division of Infectious Diseases, Head Clinical Pharmacology, British Columbia Centre for Excellence in HIV/AIDS, St. Paul's Hospital, Vancouver, British Columbia, Canada
(See the study article by Rauch et al. below)
"...,..,A large, randomized, controlled trial is planned that will enroll subjects from sites in Europe and Australia; this trial will provide more-definitive information on the utility and generalizability of HLA-B*5701 screening in diverse populations. At this time, HLA-B*5701 screening is best positioned as a screening test for abacavir-naive populations and should not be used as a rationale for rechallenge in abacavir-exposed individuals...."
Pharmacogenetics is hardly a new field, with early discoveries of genetic differences in drug-metabolizing enzymes dating back to the early 1960s. However, despite the increasing knowledge base regarding variation in drug response and toxicity related to pharmacogenetic factors and the simplification and increased availability of molecular technologies, there have been major barriers to the introduction of genetic testing into routine clinical practice [1]. Factors favoring uptake of a genetic test into clinical practice are shown in table 1. The lack of incorporation of testing into practice has related primarily to (1) the availability of laboratory markers that correctly phenotype drug response or toxicity (e.g., drug levels for anticonvulsants or international normalized ratio for warfarin), (2) the lack of standardized clinical or laboratory guidelines or algorithms for the interpretation of genetic test results, and (3) the paucity of evidence-based medicine showing that the prospective use of pharmacogenetic screening positively impacts patient outcome.
Abacavir hypersensitivity reaction is a potentially life-threatening disease that occurs in 5% of patients initiating therapy with this drug. It sets itself apart from the majority of other drug toxicities encountered in clinical practice-which tend to be multifactorial in etiology, dose related, and pharmacologically predictable-in that its occurrence appears to be largely unrelated to dose, and its clinical presentation is more severe in cases of reexposure. Furthermore, clinical diagnosis of this disease (which most commonly presents 9-11 days after initiation of therapy) has been confounded by an overlap of symptoms and signs with those of other drug hypersensitivities, viral infections, and immune restoration diseases. Although the occurrence of abacavir hypersensitivity reaction has been more common in Caucasian subjects (occurring in 8% of such subjects who initiate abacavir therapy), other clinical factors have not been useful in the prediction of this disease, and the risk of morbidity and even mortality on rechallenge precludes future use of abacavir in an individual who has been labeled with abacavir hypersensitivity syndrome.
A major breakthrough occurred in 2002 with the publication by 2 independent groups of a strong association between a specific HLA (HLA-B*5701) and abacavir hypersensitivity reaction [2, 3]. In this issue of Clinical Infectious Diseases, one of these groups-from Perth, Western Australia-describes their experience with the implementation of prospective screening for HLA-B*5701 in a single-center cohort since January 2002, compared with their experience in the prescreening era from January 1998 through December 2001 [4]. Convincingly, this study has shown that, of 148 patients found to be HLA-B*5701 negative and exposed to abacavir since January 2002, only 6 patients (4%) had abacavir discontinued within 6 weeks; all 6 of these patients had presentations clinically inconsistent with abacavir hypersensitivity reaction (nonspecific symptoms, single-symptom disease, or symptoms that did not respond to abacavir dechallenge). These patients were also found to have negative results of patch testing, which has emerged as a promising diagnostic technique for confirmation of abacavir hypersensitivity reaction [5]. Of note, among 151 patients exposed to abacavir, each of the 3 patients who developed convincing clinical evidence of abacavir hypersensitivity reaction had been HLA-B*5701 positive on screening, with 2 of these 3 patients being inadvertently exposed to abacavir because of a lack of review of HLA results before abacavir prescription and 1 of these 3 patients being exposed on the basis of his own consent. All 3 of these patients had strongly positive abacavir patch test results. Overall, despite the prescription of abacavir for 3 patients known to be HLA-B*5701 positive, this group was able to show a statistically significant decrease in the incidence of "true" abacavir hypersensitivity reaction, from 8% in the era before genetic screening for HLA-B*5701 to 2% after the implementation of genetic screening.
There are several important conclusions and implications of this study that should be highlighted. First, this is one of the first studies in which prospective implementation of a pharmacogenetic screening test that requires a yes or no interpretation has been shown to convincingly reduce the occurrence of a defined toxicity. Perhaps of even more interest, this study has shown not only that implementation of genetic screening was associated with a significant reduction in the incidence of true abacavir hypersensitivity but also that the incidence of overall early abacavir discontinuations decreased from 16.5% in the prescreening era to 6% after screening was implemented [4]. One might interpret this as implying that we have become better at the clinical diagnosis of abacavir hypersensitivity reaction over time; however, other groups not using screening have, in fact, seen an increase in early discontinuation of abacavir therapy since 2001 [6]. An alternate explanation is that genetic screening provides additional information that allows clinical pharmacovigilence to be applied more intelligently to an individual patient population.
Where does this leave us in terms of the application of genetic screening for HLA-B*5701 to our own clinic populations initiating abacavir therapy? First, there are several practical issues with regard to broad-based application of screening to HIV clinic populations. Clearly, the utility, cost-effectiveness, and generalizability of genetic screening to a given clinic population will depend on knowledge of the prevalence of HLA-B*5701 in that population, which has been shown to be closely linked to the incidence of abacavir hypersensitivity reaction in a given population. In clinics serving a predominantly Caucasian population, the overall incidence of abacavir hypersensitivity is 5%-8%, which means that, from a feasibility and cost-effectiveness standpoint, only 14 patients would need to be screened to prevent 1 case of abacavir hypersensitivity reaction. However, this ratio would be much less cost-effective in certain Asian and African black populations, for whom the prevalence of HLA-B*5701 is <1% (figure 1). The wide range of frequencies of abacavir hypersensitivity reaction described in clinical trials (0%-14%) is at least in part attributable to ethnic or racial differences in enrollees in these studies. Regardless, the clinical implications of HLA-B*5701 positivity appear to be the same, and we assume that an HLA-B*5701-positive African black subject is just as likely to experience abacavir hypersensitivity reaction as an HLA-B*5701-positive Caucasian subject. In addition, data from patch testing studies, which to date have shown 100% correlation between positive patch test results and HLA-B*5701 positivity (and which have identified no other genetic markers independently associated with abacavir hypersensitivity reaction), have been reassuring [8, 9].
At the end of the day, we have to be clear about both what we can and what we cannot currently accomplish with HLA-B*5701 screening for abacavir hypersensitivity reaction. The data from Western Australia [4] reassures us that genetic screening can prevent true abacavir hypersensitivity and, intriguingly, may also lower the rate of false-positive diagnosis of abacavir hypersensitivity syndrome. Screening, therefore, shows promise to be cost-effective, both by lowering the morbidity associated with true hypersensitivity reactions and by reducing inappropriate early discontinuation of therapy. Importantly, a point that is often overlooked is that screening should actually significantly increase the safety of abacavir use by promoting proactive avoidance of abacavir use in the HLA-B*5701-positive population, who would be at risk for death as the result of inadvertent rechallenge with abacavir.
Genetic screening for abacavir hypersensitivity reaction also warrants a few words of caution. Screening should promote a more intelligent pharmacovigilence, but it should in no way be used as a substitute for clinical judgment or pharmacovigilence. A large, randomized, controlled trial is planned that will enroll subjects from sites in Europe and Australia; this trial will provide more-definitive information on the utility and generalizability of HLA-B*5701 screening in diverse populations. At this time, HLA-B*5701 screening is best positioned as a screening test for abacavir-naive populations and should not be used as a rationale for rechallenge in abacavir-exposed individuals. Because high-resolution HLA typing is costly, is performed in specialized immunology and transplant laboratories, and has long turnaround times, more-rapid and cheaper validated molecular tests, such as the recently published HLA-B57 sequence-specific amplification and flow cytometric techniques, will be welcomed [10, 11]. Rauch et al. [4], in their description of 2 HLA-B*5701-positive patients who were inadvertently exposed to abacavir despite positive screening results, emphasize an important point: genetic screening will simply not work unless the test results are acted upon! Ensuring the reliable and timely transmission of information from laboratory to clinician will be vital to the safety and success of such genetic screening programs.
Prospective Genetic Screening Decreases the Incidence of Abacavir Hypersensitivity Reactions in the Western Australian HIV Cohort Study
A. Rauch,1,2 D. Nolan,1 A. Martin,1 E. McKinnon,1 C. Almeida,1 and Simon Mallal1
1Centre for Clinical Immunology and Biomedical Statistics, Perth, Australia; and 2Division of Infectious Diseases, University Hospital Berne, Switzerland
Clinical Infectious Diseases July 1, 2006;43:99-102
Abacavir therapy is associated with significant drug hypersensitivity in about 8% of recipients, with retrospective studies indicating a strong genetic association with the HLA-B*5701 allelle. In this prospective study, involving 260 abacavir-naive individuals (7.7% of whom were positive for HLA-B*5701), we confirm the usefulness of genetic risk stratification, with no cases of abacavir hypersensitivity among 148 HLA-B*5701-negative recipients.
Abacavir drug hypersensitivity represents an idiosyncratic, multisystem inflammatory reaction that occurs in 5%-8% of white patients initiating this commonly prescribed HIV drug [1, 2]. Clinical symptoms appear in >90% of cases within 6 weeks of commencing abacavir treatment [1] and include fever, rash, gastrointestinal symptoms, and lethargy or malaise. A history of definite abacavir hypersensitivity precludes any further use of abacavir, because rechallenge can evoke severe and potentially life-threatening reactions.
Clinical risk factors associated with abacavir hypersensitivity have been described, with relative protection associated with African racial origin, male sex, and more-advanced HIV disease stage in a meta-analysis of >8000 subjects in 34 clinical trials [2]. Higher CD8+ T cell count at the time of initiating abacavir treatment has also been associated with abacavir hypersensitivity [3]. However, genetic susceptibility conferred by the presence of a specific HLA allele-HLA-B*5701-appears to represent the dominant risk factor for abacavir hypersensitivity [4-8]. The frequency distribution of this genetic marker in different populations is likely to provide a rational basis for racially defined differences in susceptibility [7, 9], whereas the critical role of HLA-B*5701 in generating and directing CD8+ T cell-dependent, HLA-restricted immune responses suggests a key role for this genetic variant in the pathogenesis of an abacavir-specific immune response [6, 8].
Following the recognition of a strong predictive association between HLA-B*5701 carriage and abacavir hypersensitivity in our study population in 2001, a prospective testing strategy was commenced on the basis that HLA-B*5701 testing could sharply discriminate individual risk of developing abacavir hypersensitivity into low-risk (<1%) and high-risk (>70%) groups [4]. Abacavir prescription was therefore avoided in HLA-B*5701-positive patients in the Western Australian HIV cohort from 2002 onwards. Here, we describe a single-center cohort study involving all abacavir-naive individuals starting or changing antiretroviral therapy (ART) during a study period from January 2002 through July 2005 (a total of 260 subjects). The study included 121 initial ART prescriptions and 214 nucleoside reverse-transcriptase inhibitor (NRTI) treatment changes among 178 abacavir-naive patients (39 of whom where also considered in the analysis of ART-naive patients). As a component of routine clinical care, high-resolution HLA class I and class II typing (HLA-A, HLA-B, HLA-C, HLA-DR, and HLA-DQ) was undertaken for all subjects by direct DNA sequencing, as described elsewhere [4, 6]. Diagnostic classification of abacavir hypersensitivity reactions in this study population was performed using standardized clinical criteria [1, 2, 4, 6]. Therefore, a diagnosis of abacavir hypersensitivity required the presence of at least 2 symptoms of fever, rash, nausea, vomiting, headache, respiratory and gastrointestinal symptoms, lethargy, myalgia, or arthralgia, occurring <6 weeks after initial abacavir exposure. Patients stopping abacavir therapy within 6 weeks after initiation because of symptoms that did not meet the diagnostic criteria for abacavir hypersensitivity were also assessed. Adjunctive epicutaneous patch testing [8] was performed in all suspected cases of abacavir hypersensitivity and in unclear cases for which patients provided consent. All cases in which abacavir treatment was discontinued within 6 weeks after commencing therapy with the drug were examined, both in the prospective study period (January 2002 through July 2005) and in the period before the introduction of pharmacogenetic screening (January 1998 through December 2001).
As shown in table 1, patients starting abacavir therapy did not differ significantly from those starting therapy with alternative antiretroviral drugs with respect to sex, age, ethnicity, and route of HIV infection. Patients not commencing abacavir therapy had more-advanced HIV disease (i.e., significantly higher HIV loads, lower CD4+ cell counts, and more-frequent AIDS-defining events), compared with patients starting abacavir therapy, reflecting the limited use of abacavir for patients with more-advanced HIV disease during the 2002-2005 study period, when abacavir was frequently incorporated into triple-NRTI regimens (i.e., zidovudine-lamivudine-abacavir therapy). Overall patterns of NRTI drug prescription are presented in table 2.
Among 121 treatment-naive patients initiating first-line HIV therapy, 112 (92.6%) were found to be negative for the HLA-B*5701 allele, of whom 42 (37.5%) started abacavir therapy (tables 1 and 2). In 2 of these cases, abacavir was discontinued within 6 weeks after initiation because of symptoms that were not consistent with drug hypersensitivity (diarrhea in 1 patient and headache probably related to concurrent zidovudine treatment in 1 patient). Of the 9 ART-naive, HLA-B*5701-positive patients identified by genetic screening, 1 patient started abacavir treatment before the review of HLA results. This 48-year-old white man subsequently developed a typical abacavir hypersensitivity reaction, characterized by the development of a widespread maculopapular rash, fever, and constitutional symptoms 6 days after commencing abacavir therapy. Treatment was discontinued 2 days later, with resolution of symptoms within 24 h after discontinuation of therapy. A positive epicutaneous patch test result was subsequently recorded with vesicular, erythematous skin changes following application of both 1% and 10% abacavir concentrations in a petrolatum vehicle.
During this study period, 178 ART-experienced but abacavir-naive patients switched NRTI therapy. From this group, 164 (92.1%) were identified as HLA-B*5701-negative patients. Abacavir was prescribed for 106 (64.6%) of these individuals; 90 patients started abacavir therapy at the time of their first treatment change (tables 1 and 2), and 16 patients started abacavir therapy subsequently (at the second or third change in ART). In 4 cases, abacavir was discontinued within 6 weeks. One patient developed nonspecific symptoms of headache and cold sweats that were not responsive to abacavir withdrawal within 1 week after withdrawal of therapy, and 2 patients developed symptoms of nausea and vomiting attributed to nevirapine (including rash in 1 patient) that were associated with negative epicutaneous patch test results. One patient suffered from HIV-associated dermatosis that was present before starting abacavir therapy. Of the 14 ART-experienced, HLA-B*5701-positive patients who were screened, 2 started abacavir therapy. In 1 case, HLA test results were not reviewed before drug prescription, and 1 patient made an informed choice to initiate abacavir therapy because HLA typing demonstrated the absence of HLA-B57-associated ancestral haplotype markers in central and class II major histocompatibility complex regions (i.e., C4A6 and HLA-DR7/HLA-DQ3). This patient also had limited treatment options because of previous treatment complications (e.g., severe lipoatrophy). Both of these white patients experienced abacavir hypersensitivity reactions, confirmed with epicutaneous patch test results that were strongly positive to 1% and 10% abacavir concentrations.
Overall, the incidence of abacavir hypersensitivity in this prospectively screened population was 3 (2.0%) of 151 patients, with no cases diagnosed among the 148 HLA-B*5701-negative patients who received abacavir treatment. Therefore, the incidence of abacavir hypersensitivity decreased significantly (P = .01, by Fisher's exact test), compared with the 8% prevalence of abacavir hypersensitivity before genetic screening (figure 1), a result that was consistent with published data in predominantly white patient cohorts [2, 3, 10]. The proportion of patients stopping abacavir therapy because of symptoms that did not meet diagnostic criteria for drug hypersensitivity during the first 6 weeks of therapy also decreased, from 8.5% before genetic screening to 4.0% after introduction of screening (P = .1, by Fisher's exact test) (figure 1). These results, although reinforcing the positive predictive value of HLA-B*5701 testing for these patients, also demonstrate that failure to review HLA results before abacavir prescription can have significant adverse consequences. Therefore, a system has now been developed whereby HLA-B*5701-positive results are routinely noted in the allergy field of the pharmacy system database, ensuring that abacavir is not dispensed to HLA-B*5701-positive patients without prior explicit knowledge and consent of the treating clinician.
The prevalence of HLA-B*5701 carriage in this population was 20 (7.7%) of 260 patients, all of whom were identified as white, which is consistent with expected results from population-based HLA data. It is also notable that high-resolution HLA typing methods used in this study were able to discriminate the HLA-B*5701 allele from closely related alleles, such as HLA-B*5702 (in 2 cases), HLA-B*5703 (in 1 case), and HLA-B*5801/5802 (in 9 cases). Subjects with these HLA alleles were prescribed abacavir (including 2 subjects with HLA-B*5702, 1 subject with HLA-B*5703, and 3 subjects with HLA-B*5801) without incident, indicating that susceptibility is mapped specifically to the HLA-B*5701 allele and its related major histocompatibility complex haplotype.
We, therefore, conclude that prospective genetic screening with avoidance of abacavir prescription in patients carrying the susceptibility marker HLA-B*5701 has had a dramatic impact on the incidence of abacavir hypersensitivity in the Western Australian HIV Cohort. We acknowledge that assessing the broader clinical implications of this approach to abacavir prescribing in racially diverse HIV-infected populations requires further study, which may be conducted in light of the expected frequency distribution of HLA-B*5701 and its associated major histocompatibility complex haplotype in these population groups [9]. However, given the otherwise favorable safety profile of abacavir once drug hypersensitivity is excluded [11], this pharmacogenomic approach to risk stratification would appear to pave the way for more confident abacavir prescription.
|
|
| |
| |
|
|
|
