Recommendations for pharmacogenetic testing in clinical practice guidelines in the US

Abstract

Purpose

Pharmacogenetic testing can identify patients who may benefit from personalized drug treatment. However, clinical uptake of pharmacogenetic testing has been limited. Clinical practice guidelines recommend biomarker tests that the guideline authors deem to have demonstrated clinical utility, meaning that testing improves treatment outcomes. The objective of this narrative review is to describe the current status of pharmacogenetic testing recommendations within clinical practice guidelines in the US.

Summary

Guidelines were reviewed for pharmacogenetic testing recommendations for 21 gene-drug pairs that have well-established drug response associations and all of which are categorized as clinically actionable by the Clinical Pharmacogenetics Implementation Consortium. The degree of consistency within and between organizations in pharmacogenetic testing recommendations was assessed. Relatively few clinical practice guidelines that provide a pharmacogenetic testing recommendation were identified. Testing recommendations for HLA-B*57:01 before initiation of abacavir and G6PD before initiation of rasburicase, both of which are included in drug labeling, were mostly consistent across guidelines. Gene-drug pairs with at least one clinical practice guideline recommending testing or stating that testing could be considered included CYP2C19-clopidogrel, CYP2D6-codeine, CYP2D6-tramadol, CYP2B6-efavirenz, TPMT-thiopurines, and NUDT15-thiopurines. Testing recommendations for the same gene-drug pair were often inconsistent between organizations and sometimes inconsistent between different guidelines from the same organization.

Conclusion

A standardized approach to evaluating the evidence of clinical utility for pharmacogenetic testing may increase the inclusion and consistency of pharmacogenetic testing recommendations in clinical practice guidelines, which could benefit patients and society by increasing clinical use of pharmacogenetic testing.

Clinical use of biomarker tests, including genetic tests, requires evidence of clinical utility.^1,2 The clinical utility of a genetic test refers to the likelihood that the information from the test, by prompting an intervention, will result in an improved clinical outcome.^3-6 Some groups have extended the definition of clinical utility to include other outcomes such as economic and societal benefits.^5,7 Pharmacogenetic (PGx) testing targets a subset of genetic biomarkers—gene variants that influence drug response. Among stakeholders in the PGx arena, there has been ongoing debate regarding what defines sufficient evidence of clinical utility to recommend PGx testing.^8-10

The lack of consensus on the level and type of evidence of clinical utility needed to recommend PGx testing has led to confusion as to which PGx tests should be used in clinical practice. There is variability in testing recommendations by clinical society and the Food and Drug Administration (FDA) and in coverage by payors.¹¹ This inconsistency is partly due to differences in the types of recommendations and evidence required for each. The decision to order a PGx test is different from the decision about how to use available PGx results for prescribing decisions. The Clinical Pharmacogenetics Implementation Consortium (CPIC) and Dutch Pharmacogenetics Working Group (DPWG) have provided guidelines with consensus recommendations for using PGx information in treatment decisions.^12,13 Recently, DPWG added testing recommendations to their guidelines based on a relatively simple scoring rubric that considers the clinical importance of the outcome, level of evidence for this outcome, number needed to genotype, and inclusion of PGx information in the drug label.¹⁴ While CPIC and DPWG guidelines are valuable resources, most practicing clinicians are unaware of these guidelines and rely on clinical practice guidelines (CPGs) published within their clinical specialty to guide which tests they order and how they use the results for drug therapy decision-making.^15,16 For the purpose of this study, CPGs are clinical practice guidelines from medical societies and do not include CPIC or DPWG guidelines.

The Standardizing Laboratory Practices in Pharmacogenomics (STRIPE) Collaborative Community was established to accelerate the adoption of PGx in clinical practice in the US. STRIPE comprises various stakeholders across the PGx testing ecosystem, including clinicians, researchers, clinical laboratories, health systems, manufacturers, academic centers, professional societies, patient advocates, payors, and FDA.^17,18 The STRIPE Study Designs Task Force (SDTF) was created to provide recommendations on study design methods to optimize the clinical utility of PGx testing through 3 charges: (1) collating PGx testing recommendations from clinical, payor, and regulatory organizations in the US; (2) developing consensus on the type and strength of evidence required to demonstrate clinical utility; and (3) making recommendations for study designs to demonstrate clinical utility. This manuscript describes the findings of one aspect of the first SDTF charge, to collate PGx testing recommendations within CPGs from clinical organizations in the US, to inform consensus development around an evidence threshold for clinical utility in future work. A separate manuscript describing the payor coverage for PGx testing in the US has been submitted for publication elsewhere (Patel et al, manuscript under review).

Methods

A clinical subcommittee of the STRIPE Collaborative Community SDTF was created that includes PGx expert clinicians in diverse therapeutic areas, each of whom is included as a coauthor, among other PGx experts. The subcommittee identified gene-drug pairs with well-established associations with drug response by reviewing gene-drug pairs with CPIC guidelines¹² or listed in the FDA table of PGx associations¹⁹ and reviewed gene-drug pairs discussed in CPGs. CPGs from US-based clinical organizations were reviewed for information regarding PGx testing for 21 gene–drug (or drug class) pairs chosen that were categorized into 5 therapeutic areas (Table 1). All these gene-drug pairs have evidence of clinical actionability according to CPIC.

Table 1.

List of Pharmacogenetic Associations Included in the Study

Area	Associations
Cardiology	CYP2C19-clopidogrel CYP2C9-warfarin SLCO1B1-statins
Pain and general medicine	CYP2D6-codeine, CYP2D6-tramadol CYP2C9-NSAIDs HLA-B*58:01-allopurinol CYP2C19–proton pump inhibitors RYR1–volatile anesthetics
Infectious disease	HLA-B*57:01-abacavir CYP2B6-efavirenz NAT2-isoniazid CYP2C19-voriconazole
Psychiatry and neurology	CYP2C19-antidepressants, CYP2D6-antidepresseantsa CYP2D6-antipsychoticsb CYP2D6-atomoxetine HLA-B*15:02-carbamazepine, HLA-B*15:02-oxcarbazepine, HLA-B*15:02-phenytoin; HLA-A*31:01-carbamazepine, HLA-A*31:01-oxcarbazepine, HLA-A*31:01-phenytoin
Oncology	TPMT-thiopurines, NUDT15-thiopurinesc DPYD-fluoropyrimidinesd UGT1A1-irinotecan CYP2D6-tamoxifen G6PD-rasburicase

Abbreviation: NSAIDs, nonsteroidal anti-inflammatory drugs.

^aIncluding tricyclic and serotonin reuptake inhibitors.

^bIncluding risperidone, aripiprazole, and brexpiprazole.

^cIncluding 6-mercaptopurine and 6-thioguanine.

^dIncluding 5-fluorouracil and capecitabine.

A systematic search for all CPGs was not attempted due to the inherent challenges with identifying all CPGs and defining which CPGs should be included in a review (eg, whether a CPG with a single passing mention of a drug of interest should be included). Instead, the most prominent guidelines from US-based clinical practice organizations were identified by the subject matter experts within each therapeutic area and, in some cases, supplemented by searching other tertiary information sources, including Lexicomp, guideline databases, relevant governmental websites (eg, for the Centers for Disease Control and Prevention [CDC]), and internet searches. CPGs were reviewed to identify guidelines that discussed clinical use of the relevant drug and mentioned the gene-drug association. These guidelines were then reviewed to collect any information regarding the PGx gene-drug pair (eg, “associated with pharmacokinetics” or “associated with toxicity”), year, PGx testing recommendation (eg, “testing recommended,” “testing not recommended,” or “testing may be considered”), and any stated evidence to support the recommendation. If a guideline gave an explicit testing recommendation, this information was used in the review. If the guideline mentioned the PGx association but did not contain an explicit testing recommendation, the guideline was included with the information on association, but no inference was made regarding a testing recommendation. These data were used to identify areas of inconsistency, in which guidelines from the same or different organizations had different recommendations regarding whether to order PGx testing. Other information, such as recommendations for how to treat patients with available genetic information, was not the focus of this analysis and was not considered.

Results

Cardiology guidelines.

Cardiology CPGs were reviewed for information regarding PGx testing for clopidogrel, warfarin, and statins. The American Heart Association (AHA) copublishes, along with other clinical guidelines, evidence-based guidelines for clinical indications in which clopidogrel is used. In 2010, after FDA added a boxed warning about CYP2C19 to the clopidogrel drug label, AHA and the American College of Cardiology (ACC) determined that the evidence was insufficient to recommend genetic testing.²⁰ AHA has made differing recommendations within various guidelines between 2011 and 2014. The 2012 focused update on management of patients with unstable angina/non–ST-elevation myocardial infarction recommends considering testing,²¹ whereas the 2011 percutaneous coronary intervention²² and 2014 management of patients with non–ST-elevation acute coronary syndromes²³ guidelines recommend against testing. In 2016, ACC/AHA published a guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease,²⁴ which updated 6 separate guidelines published between 2011 and 2014. The 2016 guideline states that routine CYP2C19 testing is not recommended. An international expert consensus panel reviewed the evidence in 2019 and determined that testing is not recommended in patients with stable coronary artery disease or acute coronary syndrome, with the exception that testing in patients with stable coronary artery disease to escalate treatment in carriers of loss-of-function alleles may be considered in specific clinical scenarios, although the relevant scenarios were not specified.²⁵ The most recent AHA guideline (2022 guideline for coronary artery revascularization) does not mention CYP2C19 genetic testing.²⁵

The 2021 update to the CHEST guideline for antithrombotic therapy for venous thromboembolic disease does not mention CYP2C9 or VKORC1 testing for warfarin dosing.²⁶ Evidence reviews by the American College of Chest Physicians²⁷ and American College of Medical Genetics²⁸ in 2008 determined that there was insufficient evidence of clinical benefit to recommend genetic testing to guide warfarin dosing. Similarly, the most recent 2018 ACC/AHA CPGs for use of statins do not mention SLCO1B1 testing.²⁹

CPG recommendations for cardiology are summarized in Table 2.

Table 2.

Cardiology Clinical Practice Guidelines With Pharmacogenetics Information

Gene-drug pair	Guideline	Testing recommendation (recommendation strength)
CYP2C19-clopidogrel	ACCF/AHA: Clopidogrel clinical alert: approaches to the FDA “boxed warning” (2010)20	Insufficient evidence to recommend
	ACCF/AHA/SCAI: Guideline for percutaneous coronary intervention (2011)22	Not recommended (class III)
	ACCF/AHA: Focused update incorporated into the ACCF/AHA 2007 guidelines for the management of patients with unstable angina/non–ST-elevation myocardial infarction (2012)21	May be considered (class IIB)
	Guidelines for management of patients with non–ST-elevation acute coronary syndromes (2014)23	Not recommended
	ACC/AHA: Guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease (2016)24	Not recommended
	Updated expert consensus statement on platelet function and genetic testing for guiding P2Y12 receptor inhibitor treatment in percutaneous coronary intervention (2019)30	Testing to escalate treatment may be considered Testing to determine which drug would remain when dual antiplatelet therapy de-escalation is being considered is not recommended
CYP2C9-warfarin, VKORC1-warfarin	ACCP: Pharmacology and management of vitamin K antagonists (2008)27	Insufficient evidence to recommend (class 2C)
	ACMG: Pharmacogenetic testing of CYP2C9 and VKORC1 alleles for warfarin (2008)28	Insufficient evidence to recommend

Abbreviations: ACC, American College of Cardiology; ACCF, American College of Cardiology Foundation; ACCP, American College of Chest Physicians; ACMG, American College of Medical Genetics; AHA, American Heart Association; FDA, Food and Drug Administration; SCAI, Society for Cardiovascular Angiography and Interventions.

^aBold indicates partial or full recommendations for testing. Underline indicates partial or full recommendations against testing.

Pain and general medicine guidelines.

CPGs that discuss the use of codeine, tramadol, nonsteroidal anti-inflammatory drugs (NSAIDs), allopurinol, proton pump inhibitors (PPIs), and volatile anesthetics were reviewed for PGx testing recommendations. CPGs from the Veterans Affairs/Department of Defense (VA/DoD) and American Academy of Pain Medicine (AAPM) both describe the association of CYP2D6 genetics with codeine activation and efficacy; VA/DoD concluded that evidence for testing was lacking.³¹ Guidelines from the National Comprehensive Cancer Network (NCCN) for adult cancer pain state that PGx testing may be considered and reference CPIC guidelines for dosing.³² There was inconsistency across AAPM guidelines, as some guidelines mention the PGx association^33,34 but others do not.³⁵ Similarly, PGx is not mentioned in guidelines that discuss codeine and/or tramadol from CDC,³⁶ the American Academy of Neurology,³⁷ or the American College of Rheumatology (ACR).³⁸ Within CPGs for NSAID use, the Society of Critical Care Medicine guidelines for the prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the intensive care unit briefly mention that PGx data were reviewed but do not discuss any PGx associations or comment on testing.³⁹ No other PGx references were found in any other CPGs that discuss use of NSAIDs, such as the CPGs on arthritis from ACR.^38,40 ACR does recommend testing for HLA-B*58:01 before treatment with allopurinol for patients of Southeast Asian descent and for African Americans, but not for patients of other ancestral backgrounds.⁴¹

Several CPGs from the American College of Gastroenterology (ACG) address the use of PPIs in various clinical contexts. CPGs for the management of gastroesophageal reflux disease mention the association of CYP2C19 PGx with PPI response but state that testing has no role in practice.⁴² CPGs for the treatment of Helicobacter pylori infection also mention the associations but do not discuss testing,⁴³ whereas older guidelines that discuss PPIs do not mention PGx at all.^44-46 Finally, CPGs from the Occupational Safety and Health Administration⁴⁷ and the American Society of Anesthesiologists (ASA)^48,49 do not mention RYR1 testing for patients receiving volatile anesthetics.

CPG recommendations for pain and general medicine are summarized in Table 3.

Table 3.

Pain and General Medicine Clinical Practice Guidelines With Pharmacogenetics Information

Gene-drug pair	Guideline	Testing recommendation (recommendation strength)a
CYP2D6-codeine, CYP2D6-tramadol	VA/DoD: Use of opioids in the management of chronic pain (2022)31	NR: associated with response
	AAPM: Using clinical laboratory tests to monitor drug therapy in pain management patients (2018)33	NR: associated with response
	AAPM: Rational urine drug monitoring in patients receiving opioids for chronic pain: consensus recommendations (2017)34	NR: associated with pharmacokinetics
	NCCN: Adult cancer pain (2023)32	May be considered
CYP2C9-NSAIDs	ACCM/SCCM: Prevention and management of pain, agitation/sedation, delirium, immobility, and sleep disruption in adult patients in the ICU (2018)39	NR: pharmacogenetics data exist
HLA-B*58:01-allopurinol	ACR: Guideline for the management of gout (2012)41	Testing recommended in certain ancestral groups (A)
CYP2C19-PPIs	ACG: Diagnosis and management of gastroesophageal reflux disease (2022)42	Testing not recommended
	ACG: Treatment of Helicobacter pylori infection (2017)43	NR: associated with response

Abbreviations: AAPM, American Academy of Pain Medicine; ACCM/SCCM, American College of Critical Care Medicine/Society of Critical Care Medicine; ACG, American College of Gastroenterology; ACR, American College of Rheumatology; ICU, intensive care unit; NCCN, National Comprehensive Cancer Network; NR, no recommendation regarding pharmacogenetic testing; NSAIDs, nonsteroidal anti-inflammatory drugs; PPIs, proton pump inhibitors; VA/DoD, Veterans Affairs/Department of Defense.

^aBold indicates partial or full recommendations for testing. Underline indicates partial or full recommendations against testing.

Infectious disease guidelines.

CPGs for the treatment of human immunodeficiency virus (HIV) and other infectious diseases were reviewed for PGx testing recommendations for abacavir, efavirenz, isoniazid, and voriconazole. Infectious disease CPGs were available from various organizations, including CDC, the Department of Health and Human Services (DHHS), the Infectious Diseases Society of America (IDSA), and the International Antiretroviral Society–USA (IAS-USA). FDA drug labeling states that all patients should be screened for HLA-B*57:01 before initiating abacavir therapy.⁵⁰ Consistent with this requirement, there was consensus across CPGs for HIV that PGx testing before initiation of abacavir is recommended, including in guidelines from DHHS,^51,52 IDSA,⁵³ and IAS-USA.⁵⁴

Use of efavirenz is discussed in various CPGs from CDC, DHHS, IDSA, and other organizations, including the American Thoracic Society (ATS). DHHS guidelines for the use of antiretroviral agents in pediatrics recommend CYP2B6 genotyping before considering efavirenz in children younger than 3 years old.⁵² Other CPGs from DHHS⁵⁵ and ATS/CDC/IDSA⁵⁶ mention the association of CYP2B6 genotype with efavirenz pharmacokinetics but do not comment on testing. The same ATS/CDC/IDSA guidelines state that some patients are NAT2 rapid acetylators and that this could affect systemic concentrations of isoniazid;⁵⁶ however, no other CPGs for isoniazid comment on NAT2. Voriconazole use is discussed in many infectious disease CPGs from the above organizations and others, including the American Society for Transplantation and Cellular Therapy and the American Society of Transplantation. None of these CPGs comment on CYP2C19 testing explicitly, although several IDSA and CDC guidelines mention that CYP2C19 genetics affects voriconazole pharmacokinetics.^57-61

CPG recommendations for infectious diseases are summarized in Table 4.

Table 4.

Infectious Disease Clinical Practice Guidelines With Pharmacogenetics Information

Gene-drug pair	Guideline	Testing recommendation (recommendation strength)a
HLA-B*57:01-abacavir	DHHS: Use of antiretroviral agents in adults and adolescents with HIV (2021)51	Testing recommended (AI)
	DHHS: Use of antiretroviral agents in pediatric HIV infection (2022)52	Testing recommended
	IDSA: Primary care guidance for persons with HIV (2020)53	Testing recommended
	IAS-USA: Antiretroviral drugs for treatment and prevention of HIV infection in adults: 2020 recommendations of the International Antiviral Society–USA Panel (2020)54	Testing recommended (A1a)
CYP2B6-efavirenz	DHHS: Use of antiretroviral agents in pediatric HIV infection (2022)52	Testing recommended in patients <3 years old
	DHHS: Prevention and treatment of opportunistic infections in adults and adolescents with HIV (2022)55	NR: associated with pharmacokinetics
	ATS/CDC/IDSA: Treatment of drug-susceptible tuberculosis (2016)56	NR: associated with pharmacokinetics
NAT2-isoniazid	ATS/CDC/IDSA: Treatment of drug-susceptible tuberculosis (2016)56	NR: associated with pharmacokinetics
CYP2C19-voriconazole	IDSA: Blastomycosis (2008)57	NR: associated with pharmacokinetics
	IDSA: Aspergillosis (2016)58	NR: associated with pharmacokinetics
	IDSA: Candidiasis (2016)59	NR: associated with pharmacokinetics
	CDC: Prevention and treatment of opportunistic infections in children with and exposed to HIV: Candida infections (2022)60	NR: associated with pharmacokinetics
	CDC: Prevention and treatment of opportunistic infections among HIV-exposed and HIV-infected children: fungal infections—aspergillosis (2009)61	NR: associated with pharmacokinetics

Abbreviations: ATS, American Thoracic Society; CDC, Centers for Disease Control and Prevention; DHHS, Department of Health and Human Services; HIV, human immunodeficiency virus; IAS-USA, International Antiretroviral Society–USA; IDSA, Infectious Diseases Society of America; NR, no recommendation regarding pharmacogenetic testing.

^aBold indicates partial or full recommendations for testing.

Psychiatry and neurology guidelines.

Psychiatry CPGs from the American Psychiatric Association (APA),^62,63 VA/DoD,⁶⁴ Anxiety and Depression Association of America (ADAA),⁶⁵ American Academy of Child and Adolescent Psychiatry (AACAP),⁶⁶ American Academy of Pediatrics (AAP), and American Epilepsy Society (AES) were reviewed for PGx recommendations for antidepressants, antipsychotics, atomoxetine, and carbamazepine, oxcarbazepine, or phenytoin.⁶⁷ Similar to the other disease states reviewed, most psychiatry CPGs did not comment on PGx testing. Two organizations mention PGx testing for antidepressants; AACAP discourages use of testing for antidepressant therapy, citing limitations to the published evidence,⁶⁶ and VA/DoD states that there is insufficient evidence to recommend for or against testing.⁶⁴ For antipsychotic therapy, APA acknowledges the potential for PGx testing and includes, verbatim, recommendations derived from the FDA table of pharmacogenomic biomarkers in drug labeling¹⁹ within their guidelines.⁶³ PGx testing to guide attention-deficit/hyperactivity disorder treatment was absent in all CPGs reviewed except for one from AAP, which does not recommend testing due to insufficient evidence of clinical utility.⁶⁷ None of the CPGs commented on HLA-A or HLA-B testing for treatment with carbamazepine, oxcarbazepine, or phenytoin.

CPG recommendations for psychiatry and neurology are summarized in Table 5.

Table 5.

Psychiatry and Neurology Clinical Practice Guidelines With Pharmacogenetics Information^a

Gene-drug pair	Guideline	Testing recommendation (recommendation strength)b
CYP2C19-antidepressants, CYP2D6-antidepressants	VA/DoD: Management of major depressive disorder (2022)64	Insufficient evidence to recommend
	AACAP: Clinical use of pharmacogenetic tests in prescribing psychotropic medications for children and adolescents (2020)66	Not recommended
CYP2D6-antipsychotics	APA: Practice guideline for the treatment of patients with schizophrenia (2020)63	Defer to FDA
	AACAP: Clinical use of pharmacogenetic tests in prescribing psychotropic medications for children and adolescents (2020)66	Not recommended
CYP2D6-atomoxetine, CYP2D6-stimulants	AAP: Clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity disorder in children and adolescents (2019)67	Not recommended

Abbreviations: AACAP, American Academy of Child and Adolescent Psychiatry; AAP, American Academy of Pediatrics; APA, American Psychiatric Association; FDA, Food and Drug Administration; VA/DoD, Veterans Affairs/Department of Defense.

^aGuidelines from the American Epilepsy Society do not mention HLA testing, so these are not included in the table.

^bUnderline indicates partial or full recommendations against testing.

Oncology guidelines.

Oncology CPGs from NCCN, the American Society of Clinical Oncology (ASCO), and Children’s Oncology Group (COG) were reviewed for PGx testing recommendations for thiopurines, fluoropyrimidines, irinotecan, tamoxifen, rasburicase, and allopurinol. There was inconsistency between organizations regarding testing TPMT and/or NUDT15 before initiating thiopurine treatment. NCCN guidelines for treatment of pediatric⁶⁸ and adult⁶⁹ acute lymphoblastic leukemia (ALL) recommend considering testing for TPMT and NUDT15 before initiating thiopurines, whereas ALL guidelines from ASCO⁷⁰ and COG⁷¹ do not comment on PGx testing.

There is also inconsistency within and between organizations for DPYD testing before initiating fluoropyrimidine treatment. NCCN colon cancer guidelines⁷² include a recommendation against DPYD testing, whereas NCCN guidelines that discuss fluoropyrimidine treatment for rectal⁷³ and breast⁷⁴ cancer do not mention DPYD. The 2006 ASCO colorectal cancer biomarker testing guidelines⁷⁵ state that there was insufficient evidence to recommend testing, whereas the 2017 update⁷⁶ does not mention DPYD. The 2022 ASCO metastatic breast cancer biomarker guidelines mention that DPYD polymorphisms increase capecitabine toxicity but do not provide a testing recommendation.⁷⁷

NCCN guidelines for colon⁷² and rectal⁷³ cancer comment on the availability of UGT1A1 testing for irinotecan but not on its use in practice. Similar to DPYD, ASCO guidelines for biomarkers in metastatic colorectal cancer do not comment on UGT1A1.⁷⁶ There was general consistency regarding testing CYP2D6 before initiating tamoxifen treatment: the NCCN⁷⁴ guidelines and 2016 ASCO guidelines recommend against testing;⁷⁸ however, CYP2D6 is not mentioned in the 2022 ASCO update.⁷⁹

Although the FDA label includes a boxed warning for hemolysis risk in patients with G6PD deficiency, there is inconsistency in NCCN guidelines regarding G6PD testing before initiating rasburicase. NCCN CPGs for treatment of B cell⁸⁰ and T cell⁸¹ lymphomas and chronic lymphocytic leukemia/small lymphocytic lymphoma⁸² stat that G6PD testing is required. Alternatively, CPGs for pediatric aggressive mature B cell lymphomas⁸³ state that G6PD testing should be considered in male patients. Finally, it was mentioned previously that ACR CPGs recommend HLA-B*58:01 testing before initiating allopurinol treatment; NCCN CPGs that discuss use of allopurinol^80-83 do not comment on HLA-B testing.

CPG recommendations for oncology are summarized in Table 6.

Table 6.

Oncology Clinical Practice Guidelines With Pharmacogenetics Information

Gene-drug pair	Guideline	Testing recommendation (recommendation strength)a
TPMT-thiopurines, NUDT15-thiopurines	NCCN: Pediatric acute lymphoblastic leukemia (2022)68	Consider testing (2A)
	NCCN: Adult acute lymphoblastic leukemia (2022)69	Consider testing
DPYD-fluoropyrimidines	NCCN: Colon cancer (2022)72	Not recommended
	ASCO: Metastatic breast cancer biomarkers (2022)77	NR: associated with toxicity
UGT1A1-irinotecan	NCCN: Colon cancer (2022)72	Not recommended
	NCCN: Rectal cancer (2022)73	Not recommended
CYP2D6-tamoxifen	NCCN: Breast cancer (2022)74	Not recommended
G6PD-rasburicase	NCCN: B cell lymphoma (2022)80	Testing required
	NCCN: T cell lymphoma (2022)81	Testing required
	NCCN: Chronic lymphocytic leukemia/small lymphocytic lymphoma (2023)82	Testing required
	NCCN: Pediatric aggressive mature B cell lymphomas (2022)83	Consider testing

Abbreviations: ASCO, American Society of Clinical Oncology; NCCN, National Comprehensive Cancer Network; NR, no recommendation regarding pharmacogenetic testing.

^aBold indicates partial or full recommendations for testing. Underline indicates partial or full recommendations against testing.

Discussion

Clinical uptake of PGx has been limited, in part due to the limited inclusion and consistency of PGx testing recommendations within CPGs.^8-10 The first charge of the STRIPE SDTF was to collate PGx testing recommendations within CPGs from various clinical organizations in the US, to inform future work of this task force that will understand evidence thresholds used to define clinical utility. This review found relatively few CPGs that included explicit recommendations for PGx testing. Additionally, PGx testing recommendations for the same gene-drug pair were sometimes inconsistent between organizations and between different CPGs from an individual organization (eg, for G6PD), suggesting that there may be a lack of consensus on the evidence required to recommend PGx testing.

The major exceptions to the lack of recommendations on PGx testing within CPGs were the consistent recommendations to test for HLA-B*57:01 before initiating treatment with abacavir (Table 4) and several NCCN guidelines requiring G6PD testing before initiating rasburicase (Table 6). Inclusion of these recommendations may be due to testing requirements on the FDA labels for abacavir⁵⁰ and rasburicase,⁸⁴ although this was not the case for HLA-B*15:02 for carbamazepine and oxcarbazepine, for which testing requirements are also included in FDA drug labeling. The two other PGx testing recommendations found within CPGs were specific to individual patient subgroups. The recommendation to test for HLA-B*58:01 before initiating allopurinol (Table 3) was specific to certain ancestry groups in which this allele is more common, and the recommendation to test for CYP2B6 before initiating efavirenz (Table 4) was specific to children less than 3 years of age. Testing recommendations in specific subgroups, particularly based on race/ancestry, have been criticized for being potentially inequitable and logistically challenging.⁸⁵

Several CPGs commented that testing may be considered, including testing for TPMT and/or NUDT15 before initiating thiopurine chemotherapy (Table 6) and CYP2C19 before initiating clopidogrel (Table 2). The changing positions within CPGs from AHA and other cardiology organizations may reflect the ongoing debate regarding CYP2C19 testing for clopidogrel treatment,⁸ which has evolved as more randomized controlled trials have reported evidence of clinical benefit,^8,86-88 but also may reflect more nuanced differences between the guidelines in terms of the clinical contexts in which their recommendations apply. Testing was not recommended for other potentially clinically actionable gene-drug pairs, including CYP2C19 and CYP2D6 with antidepressants (Table 5) and DPYD with fluoropyrimidine chemotherapy (Table 6), despite prospective trials demonstrating clinical benefit from testing for both examples.^89-92

One of the objectives of this review of CPGs was to determine the level of consistency between and within clinical organizations. We identified examples in which a CPG from one organization recommended against PGx testing while a similar CPG from another organization did not make a testing recommendation, such as the recommendations regarding DPYD testing in patients with colon cancer in NCCN⁷² and ASCO⁷⁶ guidelines, respectively. Similarly, there were instances where various CPGs from the same organization had differing recommendations, such as the changing AHA recommendations for CYP2C19 testing for clopidogrel (Table 2) and the variation between NCCN guidelines for G6PD and DPYD testing (Table 6). However, these differences between guidelines may be due to differences in the strength of evidence or clinical contexts. Aside from the possible example of clopidogrel, which depends on whether newer guidelines were concurrent with or replaced older guidelines, there were no other examples in which one CPG recommended or required testing and another CPG from the same organization explicitly recommended against testing.

Consistent guidance on PGx testing by CPGs would provide clinicians clarity on best practices, increase adoption of PGx testing, improve patient outcomes, and ensure healthcare dollars are spent optimally.⁷ This review demonstrates not only inconsistency between CPGs but also limited inclusion of PGx testing recommendations and a lack of information regarding the process for review and evaluation of PGx evidence. Ideally, CPGs that discuss PGx testing should provide an explicit testing recommendation (eg, CYP2C19 testing for voriconazole) (Table 4). Inconsistency between guidelines is particularly concerning because it causes confusion among clinicians and disparities in care between patients. CPG committees should include PGx expertise and describe what PGx evidence was reviewed to inform PGx testing recommendations. A consensus is needed regarding the level of evidence needed to justify PGx testing recommendations. This ongoing work within the STRIPE SDTF will be a major step toward reaching the goal of consistent PGx testing recommendations within CPGs and improved medical care.

Several strengths and limitations of this review should be noted. The SDTF convened a subcommittee that included experts from diverse therapeutic areas to generate a list of potentially clinically useful gene-drug pairs and to assist with identifying the major CPGs for each therapeutic area. However, this was not a systematic review of all gene-drug pairs or all CPGs. CPIC guidelines exist for many other gene-drug pairs for which genotype-guided treatment may improve clinical outcomes, such as with CYP2D6-ondansetron⁹³ and CYP3A5-tacrolimus.⁹⁴ However, our clinical experts selected only gene-drug pairs that, in their opinion, would have sufficient clinical benefit to justify recommendations to order pretreatment PGx testing. The committee reviewed hundreds of CPGs, many of which mention these 21 commonly used drugs. However, very few CPGs mention, describe, or recommend PGx testing. This could be for appropriate reasons, such as the drug being mentioned only in passing in a CPG focused on an entirely different topic, or less appropriate reasons, such as the absence of PGx information in a CPG directly related to the use of the drug in a relevant patient cohort. On the basis of the objective of describing PGx recommendations in CPGs, the SDTF committee decided to focus this review on only CPGs that mention PGx, rather than attempting to quantify the prevalence of PGx information in all CPGs. Additionally, due to lag times in updating CPGs, some CPGs may not reflect the most recent literature. Finally, some inconsistencies between CPGs may be appropriate, as they reflect differences in the strengths of evidence for testing of the same gene-drug pair in different clinical contexts. For example, a 2019 consensus statement determined that CYP2C19 testing may be considered to escalate clopidogrel treatment but is not recommended to determine which drug should remain when de-escalation of dual antiplatelet therapy is being considered.³⁰ Similar context-specific recommendations may be appropriate for other drugs that are used in different patient populations, such as for the use of fluoropyrimidine chemotherapy in patients with colorectal vs breast cancer.

Conclusion

In conclusion, there are limited examples of PGx testing requirements or recommendations in CPGs, which likely contributes to the lack of uptake of PGx testing in clinical practice. Further work, ongoing within the STRIPE SDTF and in other organizations broadly, is needed to clearly define the level and type of evidence of clinical utility needed for PGx test recommendations in CPGs and to provide guidance on appropriate study designs to generate the evidence to demonstrate clinical utility. A standardized approach to evaluating the evidence of clinical utility of PGx testing may increase the inclusion and consistency of PGx testing recommendations in CPGs. Consistent recommendations for PGx testing in CPGs would likely increase the clinical use of PGx testing, which could improve patient outcomes and benefit society by preventing unnecessary, unsafe, and ineffective treatment.

Key Points.

• Relatively few clinical practice guidelines from clinical organizations in the US provide pharmacogenetic testing recommendations.

• Pharmacogenetic testing recommendations for the same gene-drug pair are often inconsistent between organizations and are sometimes inconsistent between different guidelines from the same organization.

• A standardized approach to evaluating the evidence of clinical utility for pharmacogenetic testing may increase the inclusion and consistency of pharmacogenetic testing recommendations in clinical practice guidelines in the US.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Disclosures

Dr. Hertz is an informal, unpaid scientific advisor to Advocates for Universal DPD/DPYD Testing (AUDT), a nonprofit advocacy organization. Dr. Bousman is founder and CEO of Sequence2Script, Inc. Dr. Patel serves as a paid consultant for VieCure and Clarified Precision Medicine. Dr. Orlando is a founder of MeTree&You. Dr. Voora has received consulting fees from GTI Labs and Sanford Health. The other authors have declared no potential conflicts of interest.