Use of a multi-gene pharmacogenetic panel reduces adverse drug effects

Abstract

Swen et al.¹ examine the utility of multi-gene pharmacogenetic testing in a large multi-national cohort. They show fewer adverse drug reactions among patients receiving testing and prescribing recommendations based on genotype results compared with those receiving usual care.

Swen et al.¹ examine the utility of multi-gene pharmacogenetic testing in a large multi-national cohort. They show fewer adverse drug reactions among patients receiving testing and prescribing recommendations based on genotype results compared with those receiving usual care.

Main text

Variation in genes encoding for drug-metabolizing enzymes, drug transporters, and drug targets is well recognized to contribute to inter-patient differences in drug response.² A number of medical centers have implemented pharmacogenetic testing into practice to help guide drug therapy.³ To facilitate implementation, the Clinical Pharmacogenetics Implementation Consortium (CPIC) and Dutch Pharmacogenomics Working Group (DPWG) provide genotype-based prescribing recommendations, which address approximately 20 genes and over 100 drugs to date, including very commonly used drugs, e.g., antidepressants, opioids, antiplatelets/anticoagulants, proton pump inhibitors, and statins.^4,5,6

Data suggest over 90% of individuals have at least onegenetic variant likely to influence drug response (i.e., actionable pharmacogenetic variant),⁷ and over 60% of patients in primary care clinics are prescribed at least one drug with pharmacogenetic guidance.⁸ These data suggest the potential population impact of pharmacogenetic testing is high, and it would follow that genotyping patients using a multi-gene panel that captures variants with implications for responses to multiple medications across the patient's lifetime could potentially improve drug-related outcomes. While data support the feasibility of pharmacogenetic testing of a single drug-gene combination for improved clinical outcomes,⁹ less is known about the feasibility and clinical utility of multi-gene testing to guide prescribing of multiple medications.

Swen et al.¹ describe the effect of pharmacogenetic testing, using a 12-gene panel, on preventing adverse drug reactions in patients enrolled across seven countries and a wide range of settings. The Preemptive Pharmacogenomic Testing for Preventing Adverse Drug Reactions (PREPARE) study, published in The Lancet, was a cluster-design crossover study in which participating countries were randomized to genotype-guided prescribing or standard clinical care with crossover after 19 months.

The study enrolled adult patients newly starting a drug addressed in DPWG guidelines, referred to as the index drug. Patients enrolled from countries randomized to genotype-guided prescribing had their pharmacogenetic results relevant to the index drug returned within seven days. Once available, remaining results were delivered to the provider through a decision support solution and to the patient via a card with results embedded in a QR code. While DPWG recommendations were provided, the ultimate prescribing decision was left to the provider. Patients in the control arm underwent pharmacogenetic testing at the end of their study participation. The study’s primary endpoint was the occurrence of an adverse reaction to the index drug, assessed via questionnaire 12 weeks after drug initiation.

A total of 6,944 patients (97.7% of European, Mediterranean, or Middle Eastern ancestry) were enrolled from 55 hospitals, clinics, community health centers, or pharmacies; 93.5% had at least one actionable variant, 75.4% had >1 actionable variant, and 25.2% had an actionable variant relevant to their index drug. Among patients with an actionable variant, 21% in the genotype-guided group and 28% in the control group had an adverse reaction considered drug-related and clinically relevant over the 12-week follow-up period, equating to 30% lower risk (odds ratio [OR] 0.70, 95% confidence interval 0.54–0.91, p = 0.0075) in the genotype-guided group. In the population overall, 21% in the genotype group and 29% in the control group had an adverse drug reaction (p < 0.0001). During the up-to-18-month follow-up period, 14% of patients were prescribed a second drug with DPWG recommendations.

The investigators should be congratulated for completing this large, complex, multi-national trial. Likely owing to their effective educational strategy and mechanism for returning pharmacogenetic results, there was a high recommendation acceptance rate by providers, demonstrating the feasibility of the investigator's approach to pharmacogenetic implementation. The study also demonstrated a significant reduction in adverse drug reactions in the genotype group, which was observed in both the subset of participants with an actionable genotype and the study population overall. Given nearly identical adverse event reductions in the genotype-guided arm in those with an actionable variant and the overall population, one must assume there was a significant reduction in adverse events in those lacking an actionable genotype. Such a finding is unexpected and raises the question about whether the pharmacogenetic testing directly led to reduction of adverse drug events or rather to a more careful approach to drug therapy overall, thus reducing adverse drug events. Future studies are needed to evaluate whether pharmacogenetic testing is associated with the equivalent of a placebo effect or the findings of Swen and colleagues in their non-actionable group have a different explanation, which they propose in their paper. Also unknown is whether study results extend to individuals of non-European ancestry, though data are clear that the associations between pharmacogenes and drug responses are common across populations. Nonetheless, this is the first large multi-national pharmacogenetic implementation study to show clinical benefit with broad implementation of panel-based pharmacogenetic testing. It provides a framework and argument for the feasibility and clinical utility of widespread clinical implementation of pharmacogenetics.

The investigators referred to their testing strategy as preemptive. However, the major eligibility criterion was that patients were starting a drug addressed in DPWG guidelines, and testing in this regard is considered reactive versus preemptive. Preemptive testing (i.e., testing prior to medication decisions) offers the advantage of having results available at the point of prescribing. Because rapid return of results is not needed, samples can be batched for genotyping, reducing technical time and cost. However, it may be months or years before results are used for prescribing decisions. Indeed, only 14% of PREPARE study patients were prescribed a second drug addressed by pharmacogenetic guidelines over the up-to-18-month follow-up. The approach taken by the PREPARE investigators, whereby testing is ordered to guide prescribing of a specific medication but includes genotypes useful for guiding future prescribing of multiple medications, may be the most clinically feasible and defensible approach to advancing the clinical utility of pharmacogenetic testing at present, particularly in the absence of clear data on outcomes and questionable reimbursement with preemptive testing. Findings from the PREPARE study provide some confidence that this approach will improve drug-related outcomes, and the authors should be congratulated for this important study that significantly advances the field.