Recent Medicare rebates for key pharmacogenomic tests: advancing equitable access to prevent serious drug reactions

The Royal College of Pathologists of Australasia (RCPA) has been instrumental in promoting the effective and safe use of medicines through evidence-based pharmacogenomic testing in the Asia Pacific region. In 2024, the RCPA's multidisciplinary pharmacogenomics advisory group published an initial list of consensus-based guidance for prescribers, and the group performs regular updates based on literature review and consensus, adding three drugs in 2025 (irinotecan, lamotrigine and venlafaxine).1 The guidance categorises drug-gene pairs as ‘Recommended’, ‘Consider’ or ‘No Current Consensus’ for testing.2 The Recommended category includes tests that identify people at risk of severe hypersensitivity reactions, exposure-related adverse drug reactions or therapeutic failure (Table 1). The Consider category includes drug-gene pairs with established evidence for therapeutic failure or adverse events, while No Current Consensus refers to drug-gene pairs with evidence for functional associations but where testing is not yet widely recommended due to limited evidence (Table 2). A recently published Australian Prescriber primer on using pharmacogenomic tests in clinical practice provides practical guidance on which drugs to test for, when to order testing, and how to interpret results, focussing on the list of RCPA Recommended tests.3 In parallel to guidance efforts, the RCPA has championed and led applications to the Medical Services Advisory Committee (MSAC) for Medicare rebate consideration of key Recommended pharmacogenomic tests. Through these efforts, rebated testing for dihydropyrimidine dehydrogenase deficiency (DPYD gene) to predict fluoropyrimidine-induced toxicity commenced 1 November 2025.4 Additionally, rebated human leukocyte antigen (HLA) testing to identify people at increased risk of serious carbamazepine/oxcarbazepine-induced hypersensitivity will be introduced in 2026.5 These recent examples expand the currently rebated list of pharmacogenomic tests beyond TPMT for thiopurine treatment and HLA for abacavir treatment,6, 7 representing a significant advancement in equitable access to pre-treatment screening for life-threatening adverse drug reactions that occur in a vulnerable minority of individuals taking these medications. This article details the recent expansion of Medicare-rebated pharmacogenomic tests in Australia, particularly DPYD genotyping to prevent fluoropyrimidine toxicity and HLA testing to avoid severe hypersensitivity reactions to carbamazepine and oxcarbazepine, and presents some considerations for clinical practice. We also discuss the evidence underpinning MSAC decisions and consider other RCPA Recommended tests that currently lack Medicare funding. Fluoropyrimidines (5-FU and capecitabine) are widely used cancer chemotherapy drugs. People with deficient dihydropyrimidine dehydrogenase (DPD) enzyme activity risk severe and potentially life-threatening toxicities, notably mucositis, diarrhoea, cytopenias and, less commonly, central neurotoxicities and acute cardiac toxicities.8 Four well-characterised genetic variants in the DPYD gene cause reduced enzyme activity in 3% to 9% of European populations (Box 1).9, 10 One of the variants (c.1905+1G>A) has been reported in South Asian, East Asian and Middle Eastern people experiencing fluoropyrimidine-related toxicity, although allele frequencies are low at a population level (3%–0%).11 None of these variants is common in Tiwi First Nations Australians,12 while variant frequencies in Māori and Pacific Islander populations, to our knowledge, have not been characterised.13 Complete DPD deficiency (0.01%–0.5% of Caucasian populations) can be fatal with exposure to these chemotherapy drugs.9, 10 The four loss-of-function alleles individually have a high specificity for predicting severe toxicity risk (~80%). Recent Australian feasibility studies have demonstrated that DPYD genotyping can be performed with a turnaround time of a median of 5 business days for two regional local health districts in New South Wales (range 1–30 days), which generally aligns with therapeutic timeframes and was considered acceptable by key stakeholders.14 The Cancer Institute NSW eviQ Cancer Treatments Online portal provides genotype-based dosing recommendations based on international consensus guidelines (detailed in Box 1).9, 15, 16 DPYD variants with genotyping tests available in Australia and New Zealand (common name in brackets): Testing should be discussed with all patients before starting fluoropyrimidine therapy. eviQ dose recommendations for variant carriers‡: Heterozygous carriers (single copy of any one variant): Homozygous or compound heterozygous carriers (two copies of the same variant or two different variants): For people tolerant after cycle 1 or 2, consider dose increase in 12.5% increments of the standard dose (detailed on eviQ). All patients should be closely monitored for toxicity regardless of genotype, as absence of these four variants does not eliminate risk of severe toxicity. The MSAC determined that DPYD testing has superior clinical effectiveness and non-inferior safety compared with no testing for those with loss-of-function variants.17 Evidence limitations for key outcome studies were acknowledged; the seven included studies were found to have serious to critical risk of bias, small sample sizes and lack of Australian-specific data.18-24 In economic evaluations, the incremental cost-effectiveness ratio (ICER) to avoid severe fluoropyrimidine-related toxicity was $67 910 per patient, resulting in a dominant economic strategy. There were uncertainties in the evidence for estimates of key model drivers, including toxicity-related hospital length of stay; however, the dominant position was maintained across most drivers in sensitivity analyses. A large national rollout study (GeneScreen 5-FU, n = 5000 patients), supported by the Medical Research Future Fund, is currently underway and will evaluate cost-effectiveness and clinical impact within the Australian healthcare system.25 An important technical consideration for DPYD testing is the direct detection of the causal variant c.1129-5923C>G (HapB3) rather than its proxy, c.1236G>A. While earlier studies predominantly used c.1236G>A as a surrogate, recent evidence has demonstrated incomplete linkage disequilibrium between these variants, meaning c.1236G>A does not reliably identify all individuals carrying the c.1129-5923C>G variant.26 This limitation was acknowledged by MSAC, which emphasised the importance of directly testing for the causal variant. The Medicare item lists c.1129-5923C>G; however, many guidelines still refer to c.1236G>A. DPYD genotyping of the four alleles has limited sensitivity, and approximately 25% to 30% of people who test negative may still experience toxicity, with these estimates predominantly representing Caucasian populations.27 Several phenotypic assays have been proposed to characterise pretreatment DPD activity. Direct measurement of activity in peripheral blood mononuclear cells is impractical for clinical applications, while endogenous plasma uracil concentrations or uracil/dihydrouracil ratio (U/UH2) have a long history but are controversial. Despite inclusion in European Medicines Agency recommendations and some national guidelines,28, 29 measurement of plasma uracil for DPD phenotyping continues to be debated due to inconsistent evidence for its predictive value, preanalytical stability concerns and lack of prospective validation of dose adjustment.30 A urinary thymine challenge test was investigated as a potential phenotyping method in a New Zealand clinical trial, but despite initial promise of improved sensitivity over plasma uracil and U/UH2,31, 32 it did not discriminate severe gastrointestinal toxicity in a clinical trial of patients receiving fluoropyrimidine-based chemotherapy.33 Efforts are underway to address some of the pre-analytical challenges with plasma uracil for phenotyping, including dried blood spotting of samples prior to analysis, and the application of DPD inhibitors. Therapeutic drug monitoring (TDM) enables individualised dose optimisation for all patients, regardless of their genetic or phenotypic status. Pre-treatment DPYD genotyping remains important, as severe fluoropyrimidine-related toxicity following initial dosing occurs rapidly in DPD-deficient patients. Additionally, genetic-based dose adjustment often results in underdosing.34 This is recognised in genotype-based dosing guidelines, e.g., eviQ recommends incremental dose increases of approximately 12.5% of the standard dose for patients with DPYD variants who tolerate the reduced dose after cycle 116; however, this may be insufficient for a significant proportion of patients. TDM is recommended for prolonged 5-FU infusions and is possible for oral capecitabine, but target ranges for the latter are not established.35, 36 For 46-h 5-FU infusions, the therapeutic target range is an area under the concentration time curve of 20 to 30 mg·h/L; this is supported by extensive evidence from both observational and interventional studies.35 While a 5-FU assay is not yet available in routine clinical practice in Australia or New Zealand, recent clinical trials have demonstrated both the feasibility and clinical value of drug monitoring. The multicentre PREDICT-5FU trial showed that TDM-guided 5-FU dosing significantly increased the proportion of patients achieving target exposure from 36% to 58% (n = 38).37 With standard body surface area dosing, 61% of patients were underexposed, compared with 39% underexposed with TDM guidance; 3% were overexposed in each treatment condition. This demonstrates that conventional dosing results in subtherapeutic exposure in the majority of patients (~60%), aligning with previous findings.35 An economic evaluation showed that pharmacokinetically guided 5-FU dosing is cost-saving in the Australian healthcare system, providing better health outcomes (1.751 vs 1.291 quality-adjusted life-years (QALYs)) at lower total cost ($32 564 vs $36 379) compared with standard body surface area-based dosing, with an ICER value per QALY consistently below the willingness-to-pay threshold across sensitivity analyses ($50 000).38 HLA genotyping to prevent severe drug hypersensitivity reactions addresses a critical safety concern, particularly for medications often initiated in primary care, including allopurinol and carbamazepine. Specific HLA alleles are causally linked to life-threatening reactions including Stevens-Johnson syndrome/toxic epidermal necrolysis (SJS/TEN) and drug reaction with eosinophilia and systemic symptoms (DRESS). The mechanisms involve formation of HLA drug T-cell complexes through direct drug binding to HLA epitopes, drug modification of presented peptides or drug-altered peptide repertoires, creating targets for T-cell activation.39 Carbamazepine is a widely used medication for focal epilepsy, trigeminal neuralgia and bipolar affective disorder, while oxcarbazepine is a treatment option for partial or generalised seizures. Although severe hypersensitivity reactions are rare, they can be devastating, with mortality rates up to 30% for severe manifestations.40 HLA-B*15:02 shows a particularly strong association with carbamazepine-SJS/TEN: 72% to 100% of cases in Asian populations carry this allele.41 On the other hand, HLA-A*31:01 shows particularly strong association with DRESS syndrome, present in 37%–67% of cases across diverse populations.41 Not all patients who test positive for these alleles will develop severe hypersensitivity reactions, resulting in low positive predictive values. Similarly, a negative test does not definitively exclude risk, and reactions are possible in non-carriers. Testing is nevertheless recommended due to the severity of reactions and availability of alternative treatments that are equally effective and safe (Box 2).41, 42 HLA alleles associated with severe hypersensitivity reactions: Testing should ideally be conducted before initiating carbamazepine or oxcarbazepine therapy for any indication (epilepsy, trigeminal neuralgia, bipolar disorder). Testing may be conducted during treatment initiation in cases of urgent clinical need (e.g., severe trigeminal neuralgia requiring immediate pain management). For positive test results, consider alternative therapy based on indication. Cross-reactivity may occur: avoid structurally similar agents if possible (phenytoin, lamotrigine, phenobarbital, primidone). Reactions are possible with a negative test. At signs of a fever or rash, patients should be advised to cease the drug and seek medical advice. The clinical relevance of HLA genotyping is particularly significant in Australia's ethnically diverse population, with approximately 17.4% of Australians reporting wholly or partial Asian ancestry.43 Risk allele prevalence varies, e.g., HLA-A*31:01 is prevalent in up to 10% of Korean and North Chinese people and up to 15% of Japanese people, while HLA-B*15:02 prevalence can reach 20% in Malay, Filipino and Vietnamese people.44 In contrast, HLA-A*31:01 is present in approximately 5% of European people and other Asian populations, while HLA-B*15:02 is relatively rare in Europeans and Japanese people (