Pharmacogenomic Implications of Genes on the American College of Medical Genetics Secondary Findings List
Health Services and Implementation
-
Primary Categories:
- Health Care Inequities and health disparities
-
Secondary Categories:
- Health Care Inequities and health disparities
Introduction:
Historically, the clinical genomics (CGx) and pharmacogenomics (PGx) domains have remained largely distinct. CGx assesses genetic variants associated with disease risk and pathology, while PGx evaluates genes associated with pharmacokinetics, pharmacodynamics, and side effect risks. However, in our institutions, we have identified multiple cases that demonstrate the intersection between CGx and PGx. In one case, a diabetic individual taking liraglutide tested positive for a pathogenic variant in RET, consistent with a diagnosis of multiple endocrine neoplasia type 2 (OMIM:171400). This diagnosis is a contraindication for GLP-1 agonists. In another, a child with a pathogenic variant in SDHD was prescribed atomoxetine, which had the potential to interfere with the metanephrine levels used to screen for pheochromocytoma.
Commonly utilized knowledgebases (PharmGKB and GeneReviews) provided limited pharmacogenomic information on pathogenic variants. To investigate the breadth of this gap, we cross-referenced the genes listed on the ACMG Secondary Findings (ACMG-SF) list with the knowledgebases and drug labeling to identify those with overlapping CGx and PGx implications.
Methods:
For each gene on the ACMG-SF list, we searched two common clinical sources of genomic information: GeneReviews (“Agents/Circumstances to Avoid” section) and the PharmGKB (“Drug Label Annotation” and “Variant Annotation” sections). Targeted therapies for specific germline/somatic variants were excluded. Interactions were rated on a 1-4 scale by pharmacogenomics-trained pharmacists and a genetic counselor: Level 1 indicates FDA or guideline-driven recommendations, Level 2 indicates a reasonable likelihood of clinically significant interactions based on published literature, Level 3 indicates emerging evidence for an interaction but no recommendations, and Level 4 indicates no/weak interactions. We defined Level 1 or 2 as the threshold for clinical actionability. We then performed a follow-up systematic review through the FDA Label drug product label database to determine if any drug’s label “Contraindications” and “Warnings and Precautions” section referenced a Level 1 or 2 disease/risk phenotype.
Results:
Of the 81 genes reviewed, 30 (37%) had Level 1 or Level 2 interactions. PharmGKB identified 3/4 (75%) Level 1 interactions and 5/26 (19%) Level 2 interactions. GeneReviews identified 3/4 (75%) Level 1 interactions and 19/26 (73%) Level 2 interactions. Neither source identified 7/30 (23%) Level 1 or Level 2 interactions, which includes the ones seen in the clinical cases. The FDA Label Database had contraindications or warnings labeling specific to 4/4 (100%) Level 1 Interactions and 7/26 (27%) Level 2 Interactions. Examples of both interactions include risks inferred by the disease pathology (e.g. stimulants and catecholaminergic polymorphic ventricular tachycardia) and less obvious associations (e.g. Marfan syndrome (OMIM:154700) and fluoroquinolones and Loeys-Dietz syndrome (OMIM:613795, 614816, 610168) and over-the-counter decongestants).
Conclusion:
Our results emphasize a need for greater research on medication implications of hereditary disease genes. We found that over 1/3 of genes on the ACMG-SF list carried some risk for medications otherwise unrelated to the condition. There is the concern for patient harm if these interactions are not known, assessed, and communicated appropriately via education and electronic clinical decision support.
There is no singular source that can be utilized to identify these significant interactions. Information is sparse and scattered throughout primary literature and drug labeling, making risk identification/triage challenging. ClinPGx, a PGx database, has recently begun working with ClinGen to integrate their databases and we propose this project as a natural fit for these two workgroups.
As clinical practice continues to move towards broad preemptive genetic testing, practitioners must be aware of both CGx and PGx implications of genetic findings. More health systems are employing clinical pharmacists to manage PGx implementation programs and we suggest that their scope should expand to support CGx. This multidisciplinary approach promises to provide the highest quality of care to patients with genetic conditions.
Historically, the clinical genomics (CGx) and pharmacogenomics (PGx) domains have remained largely distinct. CGx assesses genetic variants associated with disease risk and pathology, while PGx evaluates genes associated with pharmacokinetics, pharmacodynamics, and side effect risks. However, in our institutions, we have identified multiple cases that demonstrate the intersection between CGx and PGx. In one case, a diabetic individual taking liraglutide tested positive for a pathogenic variant in RET, consistent with a diagnosis of multiple endocrine neoplasia type 2 (OMIM:171400). This diagnosis is a contraindication for GLP-1 agonists. In another, a child with a pathogenic variant in SDHD was prescribed atomoxetine, which had the potential to interfere with the metanephrine levels used to screen for pheochromocytoma.
Commonly utilized knowledgebases (PharmGKB and GeneReviews) provided limited pharmacogenomic information on pathogenic variants. To investigate the breadth of this gap, we cross-referenced the genes listed on the ACMG Secondary Findings (ACMG-SF) list with the knowledgebases and drug labeling to identify those with overlapping CGx and PGx implications.
Methods:
For each gene on the ACMG-SF list, we searched two common clinical sources of genomic information: GeneReviews (“Agents/Circumstances to Avoid” section) and the PharmGKB (“Drug Label Annotation” and “Variant Annotation” sections). Targeted therapies for specific germline/somatic variants were excluded. Interactions were rated on a 1-4 scale by pharmacogenomics-trained pharmacists and a genetic counselor: Level 1 indicates FDA or guideline-driven recommendations, Level 2 indicates a reasonable likelihood of clinically significant interactions based on published literature, Level 3 indicates emerging evidence for an interaction but no recommendations, and Level 4 indicates no/weak interactions. We defined Level 1 or 2 as the threshold for clinical actionability. We then performed a follow-up systematic review through the FDA Label drug product label database to determine if any drug’s label “Contraindications” and “Warnings and Precautions” section referenced a Level 1 or 2 disease/risk phenotype.
Results:
Of the 81 genes reviewed, 30 (37%) had Level 1 or Level 2 interactions. PharmGKB identified 3/4 (75%) Level 1 interactions and 5/26 (19%) Level 2 interactions. GeneReviews identified 3/4 (75%) Level 1 interactions and 19/26 (73%) Level 2 interactions. Neither source identified 7/30 (23%) Level 1 or Level 2 interactions, which includes the ones seen in the clinical cases. The FDA Label Database had contraindications or warnings labeling specific to 4/4 (100%) Level 1 Interactions and 7/26 (27%) Level 2 Interactions. Examples of both interactions include risks inferred by the disease pathology (e.g. stimulants and catecholaminergic polymorphic ventricular tachycardia) and less obvious associations (e.g. Marfan syndrome (OMIM:154700) and fluoroquinolones and Loeys-Dietz syndrome (OMIM:613795, 614816, 610168) and over-the-counter decongestants).
Conclusion:
Our results emphasize a need for greater research on medication implications of hereditary disease genes. We found that over 1/3 of genes on the ACMG-SF list carried some risk for medications otherwise unrelated to the condition. There is the concern for patient harm if these interactions are not known, assessed, and communicated appropriately via education and electronic clinical decision support.
There is no singular source that can be utilized to identify these significant interactions. Information is sparse and scattered throughout primary literature and drug labeling, making risk identification/triage challenging. ClinPGx, a PGx database, has recently begun working with ClinGen to integrate their databases and we propose this project as a natural fit for these two workgroups.
As clinical practice continues to move towards broad preemptive genetic testing, practitioners must be aware of both CGx and PGx implications of genetic findings. More health systems are employing clinical pharmacists to manage PGx implementation programs and we suggest that their scope should expand to support CGx. This multidisciplinary approach promises to provide the highest quality of care to patients with genetic conditions.
)
){kind=logo}
){kind=logo}
){kind=logo}
){kind=logo}
){kind=logo}
){kind=logo}
)
){kind=logo}
)
)
)
)
)
)
){kind=logo}
){kind=logo}
){kind=logo}