Guides › HLA pharmacogenomics

HLA-B*57:01 and abacavir: what the CPIC guideline says

Updated 2026-10-30 · allele examples checked against IPD-IMGT/HLA 3.66.0 on 2026-10-30

The Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for abacavir says that carriers of HLA-B*57:01 have a "significantly increased risk of abacavir hypersensitivity", and its recommendation for carriers is "Abacavir is not recommended", classified as strong. The association is CPIC evidence level A, with the testing label "Testing Required".

This guide summarises what the guideline says and how its terms map onto HLA nomenclature. It is not prescribing advice. The guideline itself says it is intended "only to assist clinicians in decision making", and that "it remains the responsibility of the health-care provider to determine the best course of treatment for the patient". Read the guideline page before acting on anything here.

The guideline and its update

The guideline was published by Martin and colleagues in Clinical Pharmacology & Therapeutics in 2012. Its stated purpose is to "allow the interpretation of clinical HLA-B genotype tests so that the results can be used to guide the use of abacavir for the treatment of HIV". Choice of antiretroviral therapy in general is outside its scope.

A 2014 update reviewed the literature published since and concluded that "none of the evidence would change the therapeutic recommendations in the original guideline", so the 2012 recommendations remain current. The update added implementation material, such as workflow diagrams for storing a result in an electronic health record. The maintained version is the ClinPGx guideline page.

The recommendation table

The 2012 paper's Table 2 has two rows:

The body of the guideline adds that screening "should be performed in all abacavir-naive individuals before initiation of abacavir-containing therapy", and qualifies the carrier recommendation: in abacavir-naive carriers, abacavir "should be considered only under exceptional circumstances when the potential benefit ... outweighs the risk". Those conditions belong with the table whenever it is quoted.

How results are reported

The guideline defines two results and nothing in between. Given the codominant expression of HLA-B, results are "'positive' (HLA-B*57:01 being present in one or both copies of the HLA-B gene) or 'negative' (no copies of HLA-B*57:01 are present), with no intermediate phenotype."

So one copy and two copies are the same result for this guideline. A typing that reports B*57:01 on either chromosome is positive.

The guideline also says HLA-B status "has no effect on abacavir pharmacodynamics or pharmacokinetics; it only influences the likelihood that an HSR will occur". This is not a dosing guideline.

What the test can and cannot tell you

The 2012 guideline is explicit about the limits of the test, and these statements are as much a part of it as the table:

• Positive predictive value. "The positive predictive value of HLA-B*57:01 genotyping is ~50%", meaning a significant number of carriers would have tolerated the drug.
• Negative predictive value. The guideline gives a negative predictive value of ">99%" for immunologically confirmed hypersensitivity.
• A negative result does not remove all risk. "Patients testing negative for HLA-B*57:01 also have a 3% risk of developing a clinically diagnosed HSR", so counselling and monitoring remain standard practice.
• Testing does not replace monitoring. "Genotyping is not a replacement for appropriate patient education and clinical monitoring."
• It predicts one reaction only. HLA-B*57:01 "is not predictive of any other adverse reactions", nor of whether abacavir will work.

For background, the abstract states that without genetic prescreening hypersensitivity affects about 6% of patients; the body text gives "~5–8%" in the first six weeks. If you quote a figure, quote it with its source and wording.

Getting the allele name right

The guideline is about one specific allele name, and lookups fail in predictable ways.

It is a two-field name. B*57:01 names a protein. In IPD-IMGT/HLA release 3.66.0 it covers 92 named alleles, all encoding that protein; a typing resolved to any of them carries B*57:01 as its two-field name. The two-field guide explains why.

A first-field result is not enough. B*57 includes other alleles. In release 3.66.0 B*57:03 covers 20 named alleles; the guideline mentions it only as "the closely related allele HLA-B*57:03" in a passage about viral load, not as a risk allele, and in the 2026-10-26 snapshot of the open hla-pgx-open table B*57:03 has no associations. A result of B*57 cannot say whether B*57:01 is present.

Targeted tests are what the guideline prefers. It says: "It is preferable to perform only specific tests for HLA-B*57:01 because more extensive HLA genotyping does not add clinically useful information with regard to abacavir treatment." It notes that several HLA-B genotyping methods are commercially available, and does not prescribe one.

Old spellings must be normalised. Reports written before April 2010 give the same name as B*5701. It means B*57:01, and should be read as such with a note that the spelling is deprecated, not rejected as unknown. The conversion guide covers older names.

Looking it up with HLA-Verify

HLA-Verify carries the open hla-pgx-open table, which joins CPIC guideline rows and PharmGKB clinical annotations to HLA allele names checked against the pinned release. POST /v1/pgx/screen normalises each reported name first and then joins on the two-field form: sent B*5701, it returns the current name B*57:01, the flag deprecated_name and the abacavir CPIC level A associations, with the table snapshot and release named in the response.

The pharmacogenomic endpoints are decision support only and not a medical device. They report what the guideline says; they do not decide anything about a patient. For the other alleles in CPIC HLA guidelines, see the summary of HLA alleles in drug guidelines.

Frequently asked questions

What does the CPIC guideline recommend for HLA-B*57:01 carriers?

Its table says "Abacavir is not recommended" for carriers, classified as strong, and the text adds that use should be considered only in exceptional circumstances. The decision belongs to the clinician using the guideline.

Is one copy of HLA-B*57:01 enough for a positive result?

Yes. The guideline defines positive as B*57:01 present in one or both copies of HLA-B, with no intermediate result.

Does a negative HLA-B*57:01 test rule out abacavir hypersensitivity?

No. The guideline states that patients who test negative still have a 3% risk of a clinically diagnosed hypersensitivity reaction and that monitoring remains necessary.

Is HLA-B*57:03 covered by the abacavir guideline?

The guideline's recommendations are for HLA-B*57:01. It mentions B*57:03 only in a passage about viral load, and B*57:03 had no associations in the 2026-10-26 hla-pgx-open snapshot.