The use of abacavir has been limited by the occurrence of abacavir hypersensitivity syndrome, a distinct entity constituting fever, malaise, respiratory symptoms and gastrointestinal upset. The rash is a late manifestation of the syndrome and occurs in 70% of cases.⁵⁹ Evidence suggests that most cases present within the first 2–3 weeks of therapy, at a median time of 8 to 9 days, and certainly all cases will have occurred within 6 weeks of continuous therapy.⁶⁰ The reported incidence of abacavir hypersensitivity syndrome is 2-9%.^8,61,62

Immunological mechanisms for abacavir hypersensitivity

CD8⁺ T-cells recognize abacavir, inducing a well-characterized immune response, in both prior-sensitized patients (exposed T-cells)⁶³ and abacavir naïve individuals (HLA-B*57:01 positive patients).⁶⁴  Confirmation of CD8⁺ T cell involvement following abacavir exposure has been confirmed in biopsies of positive abacavir patch tests taken from affected individuals.⁶⁵ Taken together, these findings demonstrate a class I HLA-restricted CD8⁺ T-cell mechanism for the abacavir hypersensitivity syndrome. In 2012, evidence from three separate groups^66-68 demonstrated that abacavir activates T cells through a new mechanism of drug hypersensitivity called the “altered peptide repertoire model”.  In this model, abacavir binds non-covalently to a portion of the antigen binding cleft of HLA-B*57:01. The initial binding and ensuing immune response does not require drug metabolism. Peptide elution studies suggested that the change in the peptide binding groove that follows causes an alteration in the repertoire of self-peptides that bind. These self-peptides were previously not presented during thymic development and the establishment of T cell tolerance/anergy and, therefore, are able to induce T-cell responses by T cells that were not previously tolerized to ‘self’.

The fact that only 55% of patients with HLA-B*57:01 develop abacavir hypersensitivity syndrome identifies a need for future models to explain the ‘positive predictive value (PPV) gap’ and HLA-B*57:01 positive abacavir tolerance.⁸ Recently, an HLA-B*57:01-transgenic mouse model provided some insight into a potential mechanism to explain why 45% of HLA-B*57:01-positive individuals tolerate abacavir. The altered peptide model suggests that in all HLA-B*57:01-positive abacavir-exposed patients, abacavir would bind noncovalently to the floor of the peptide-binding groove of HLA-B*57:01 and alter the repertoire of endogenous peptides presented to CD8⁺ T cells; however, the recent study by Cardone et al suggests that in the absence of CD4⁺ T-cell depletion, dendritic cells remain in an immature state and there is tolerance to the altered peptide repertoire.^69,70 Factors outside of the MHC such as endoplasmic reticulum aminopeptidase (ERAP1) allotypes that favour peptide trimming may be important drivers of susceptibility in HLA-B*57:01 positive individuals.⁷¹

Pharmacogenomic screening for susceptibility to abacavir hypersensitivity syndrome

Since 2008, international guidelines have recommended the use of HLA-B*57:01 screening before abacavir prescription and abacavir hypersensitivity syndrome as a clinical entity has disappeared in the developed world, where these guidelines are currently followed.  As a result, HLA-B*57:01 is the most widely used genetic screening test in clinical practice today.⁷² The PREDICT-1 study was the first double-blind randomized controlled study to test the utility of a genetic marker in reducing a specific drug toxicity.  It demonstrated a significant reduction in abacavir hypersensitivity reactions in people randomized to real-time HLA-B*57:01 testing when abacavir was withheld from those carrying HLA-B*57:01.⁸ In this study, where patch testing was a co-primary endpoint with clinically diagnosed abacavir hypersensitivity, HLA-B*57:01 testing eliminated patch test positive (immunologically-confirmed) abacavir hypersensitivity syndrome. In the PREDICT-1 study patients naive to abacavir (n = 1956) were divided into a prospectively screened group, in which those carrying the predictive allele were excluded from abacavir therapy, and a control group who were treated with abacavir without prior screening. Following 6 weeks of observation there was a clear reduction in subjects with immunologically confirmed AHR: 0% in the screening group and 2.7% in the control group. HLA B*57:01 screening for susceptibility to abacavir hypersensitivity demonstrated a positive and negative predictive value of 55% and 100% respectively. Screening for HLA-B*57:01 is recommenced prior to commencement of abacavir in any HIV positive individual in the developed world where populations carrying HLA-B*57:01 are represented.