Guidance on Submitting De-Identified Variant-Level Information to ClinVar When Direct Patient Consent Was Not Sought

Download Guidance for Clinical Laboratory Submission of Variants to ClinVar

The following document describes dataset attributes that enable submission of de-identified variant-level information to ClinVar when direct patient consent was not sought. This approach has been reviewed by NHGRI. The NIH Genomic Data Sharing Policy is not applicable to ClinVar submissions, because it involves variant-level information and not large-scale human genomic data. De-identified variant-level information obtained by laboratories during the course of fee-for-service clinical testing may continue to be submitted to ClinVar WITHOUT the explicit consent of the patients in whom these variants were observed. Information submitted may include the following:

  • Laboratory name
  • Complete variant definition
  • Pathogenicity assertion
  • Disease and inheritance pattern upon which the assertion was based, if applicable
  • Evidence upon which the assertion was based, including, but not limited to:
    • Summary of available literature
    • Summary of genetic, computational and functional analyses published or performed internally
    • Summary of the laboratory’s case-level experience with the variant, i.e. number of individuals in whom the variant has been identified, summary and counts of the phenotypic and demographic features of those individuals, etc. For example: “This laboratory has observed this variant in 10 individuals: 4 males, 6 females. All cases had clinical diagnoses of hypertrophic cardiomyopathy, 8 had conduction system disease, and 2 were reported to have onset in infancy. 9 of the individuals were Caucasian, and 1 was Asian.
  • Information on additional variants observed in individual patients as it relates to the pathogenicity assertion of the variant being described in ClinVar. For example, a laboratory may want to note that the reason Variant A was considered “Likely benign” is because the patient was also found to have Variant B in cis, a variant with strong evidence for pathogenicity for a dominant disorder. Similarly, reporting the in trans variant for a recessive condition or both components of an unbalanced translocation are acceptable.

It should be noted that some variants (including but not limited to structural variants) may only have been observed in one individual. However, this does not preclude sharing of these variants with their associated phenotype in ClinVar given that identification of these cases would require detailed knowledge of the case. For example: “This laboratory has observed this variant in a single African-American male. This individual was reported to have developmental delay and dysmorphic features.”

Importantly, highly sensitive information should be excluded from this type of sharing when less than 5 cases have been identified.

More specific individual-level information, such as the distinct phenotypes of each individual observed in a particular laboratory’s experience with a variant, may be accepted and displayed by ClinVar, provided the proper level of consent has been obtained. It is up to the individual laboratory to discuss with their specific institutional review board (IRB) to determine what level of consent is necessary for the submission of information beyond what is explicitly listed here.

Information that WOULD NOT be submitted to ClinVar includes the following:

  • Patient name, DOB, medical record number, or any other personal identifying information
  • Large genomic datasets (e.g., a VCF file from a whole genome or exome sequencing study, all variants identified in a large gene panel, a cytogenomic microarray raw data file, etc.)

Last updated: July 27, 2015