Precision medicine offers the promise of an accurate assessment of individual risk for serious conditions like hypertrophic cardiomyopathy (HCM).

But the utility of genetic tests may be limited by the lack of diversity of people included in the underlying genetic databases used to assess risk, according to "a cautionary tale of broad relevance to genetic diagnosis" published in the New England Journal of Medicine. (A link to the study is expected to be active after 5:00 pm EDT Wednesday.)

Harvard researchers first identified five genetic variants that had been previously linked to HCM. But the studies that established these links were small and did not include a substantial number of blacks in the control groups. The variants would not have been linked to HCM, they wrote, if the database had contained more genetic information from black Americans, because these variants occurred much more frequently in the black American population (ranging from 2.9% to 27%) than in the white population (ranging from 0.02% to 2.9%). The higher occurrence of the variants in the black population is an indication that the variants were benign and not associated with HCM.



The researchers then examined records from a single laboratory and found seven individuals -- five black Americans and two of unspecified ancestry -- whose "benign variants" had been "misclassified as pathogenic in genetic reports." The consequences of the misdiagnoses almost certainly had consequences for many people beyond the seven individuals, because these "misclassifications invalidate risk assessments undertaken in relatives, requiring a chain of amended reports and management plans."

The findings, the authors wrote, "show how health disparities may arise from genomic misdiagnosis." These disparities, though, are not related to access to care or other often discussed issues but "to the historical dearth of control populations that include persons of diverse racial and ethnic backgrounds."

The authors suggested the problem uncovered by the study is unlikely to be limited to HCM. "We believe that what we're seeing in the case of hypertrophic cardiomyopathy may be the tip of the iceberg of a larger problem that transcends a single genetic disease," said the first author of the study, Arjun Manrai, PhD, of Harvard, in a press release.

"Our study powerfully illustrates the importance of racial and ethnic diversity in research," said senior investigator Issac (Zak) Kohane, MD, PhD, also of Harvard. "Racial and ethnic inclusiveness improves the validity and accuracy of clinical trials and, in doing so, can better guide clinical decision-making and choice of optimal therapy. This is the essence of precision medicine."

A frequent critic of precision medicine, Michael Joyner, MD, of the Mayo Clinic, commented: "I think even precision medicine skeptics (like me) see the potential value of DNA variant testing in rare diseases like hypertrophic cardiomyopathy. However, the big story in the last 5 to 10 years has been the emergence of many potential causal variants (versus the hypothesized few) and the genotype-positive, phenotype-negative patient. The other big story is that many or even most 'normal' people have DNA variants that have been associated with disease. So, many variants with variable penetrance."

This new study, said Joyner, points to an additional limitation of precision medicine that has not been generally appreciated. The study "shows the problem is even more challenging when genetic information about DNA risk variants is applied to patients from different ethnic and racial groups."

"Addressing it will require more complete data from highly diverse patient groups, better data sharing, and integration of very granular decision support tools into electronic health records. Given the challenges and controversies about data sharing and worries about the underperformance of electronic health records, the chances of a genomic Tower of Babel emerging for many patients is a real possibility. That such a Tower of Babel might make the problems of health disparities worse and not better is another concern."

Joyner concluded, "All of this makes the once-hoped-for and unambiguous 'biological orthopedic surgery' paradigm stemming from genotype = phenotype = treatment unlikely for most patients and most conditions."

Sekar Kathiresan, MD, of the Broad Institute, said the paper shows that "interpretation of DNA sequence variants for clinical meaning evolves as our knowledge of either genotype and/or phenotype improves."

He pointed out that the problems reflected in the new paper are now being addressed by ongoing research. "Our knowledge of genotype has improved dramatically in past few years as the number and ethnic diversity of people have improved. The paper highlights a problem from a few years ago when the number/ethnic diversity of sequenced individuals was limited."

Kathiresan also pointed out that "to help with the problem identified in the NEJM paper, we at Broad have built a very diverse, large frequency database for mutations involving 60,000 humans."

That ExAC (Exome Aggregation Consortium) database was also mentioned by the authors of the NEJM paper as a resource for new studies to be "well-powered for ruling out pathogenicity even for rarer variation."