Dr. Bill Gibson (biography and disclosure)
What I did before
To be honest, Direct-to-Consumer Genetic Testing (“DTC” Genetic Testing) wasn’t really on my radar until a patient showed up on my doorstep with a result in hand. Broadly speaking, I used to see three main applications for sequencing a large number of genes in a specific individual. The first application was for ancestral origin: looking at DNA sequences all over the genome to get a better idea of a specific person’s ancestry. Because different bits of the genome map track back to different regions of the world map, we can get a rough idea of where someone’s historic (and sometimes prehistoric) relatives used to live. Medical applications of this approach seem pretty limited, so some genomicists call it “recreational genomics.” It doesn’t really require full sequencing of the human genome, but many private sequencing providers offer it. The second application is diagnostic: to find rare mutations that have a large effect on someone’s health. These rare mutations truly cause disease. Examples include connective tissue mutations (heart disease), mutations in muscle proteins (muscular dystrophy), certain cancers (e.g. breast cancer, ovarian cancer and colon cancer), mutations in deafness genes and mutations in brain-, eye- and nerve proteins (neurological disease and/or vision loss). These types of rare diseases are often caused by one or two major mutations in a gene, leading to dominant or recessive inheritance. For some of them, any one among a large group of genes could harbour mutations, so a “one gene at a time” search becomes really long and expensive. In these situations, high-throughput sequencing can be a cost-effective way to get almost all of the genes in one go, letting us quickly find the type of mutation(s) we are looking for. The third application was semi-predictive: to find common genes that have a small effect on someone’s health, but that may have a large effect on health when added (or multiplied) together. The hope behind this type of testing is that these sequences can be assembled into a profile that would have some ability to classify a person into “high-risk” or “low-risk” categories for common diseases, such as heart disease and stroke, diabetes, dementia, asthma and so forth.
What changed my practice
I was referred a patient who had paid privately for high-throughput DNA sequencing, and then had her genome interpreted by several specialists in the field. I was asked to decide whether a particular rare mutation explained a number of unusual features in her medical history. These features were consistent with a common disease, though many of her close relatives also had the same disease. Thus, the presence of a dominant gene in the family was another plausible explanation. Like the result of any test I might have ordered myself, the patient’s data might help me to do my job, but it wouldn’t do my job for me. I therefore also looked at her family history, and it became clear that there was no way to be certain that this specific mutation was causing her health problems. We are now trying to get better evidence to match the gene with the disease by testing other family members for the same mutation found by the private genome sequencing company. We are not sequencing all of their genes, just the one area where the mutation is found. We can do this because we have a research lab, but this is beyond the scope of what many family docs could do in the clinic.
This case made me more skeptical about the current medical usefulness of DTC genetic testing for common diseases, whether it finds rare mutations or common variants. That third application now seems to be much more distant in the future.
What I do now
From the perspective of one’s own ancestry, DTC genetic tests can return data that is interesting, and sometimes surprising, but it is difficult to imagine a good rationale for changing medical management on the basis of ancestry. It must be remembered that the “family history” one learns from ancestry testing is not the same as the traditional “family history” that a family doc would collect as part of routine care.
From a medical perspective, I advise people who are thinking about DTC genetic testing to think of it like the over-the-counter (OTC) medical tests available in many pharmacies. An OTC pregnancy test is designed to answer a focused question: whether or not a woman is pregnant. An OTC cholesterol test answers a similarly focused question. I tell people that DTC genetic tests very rarely return results that are medically actionable, unless originally done to answer a focused question about a rare genetic disease.
That said, most DTC genetic testing is still not well suited to detecting rare disease-causing mutations. Patients who have concerns about a specific genetic disease, or an unusual problem that may be genetic in nature, are best served by a referral to the Provincial Medical Genetics Program or to the BC Cancer Agency’s Hereditary Cancer Program. Though DTC genetic testing may uncover major mutations that cause cancer or other diseases, current methods “flag” a large number of rare DNA changes that aren’t really there (false positive mutations) and fail to flag an unknown number true mutations – thus generating false negative results. This makes it hard for a practitioner to know if the real mutation he or she is looking for will actually be found the list returned by the company. Targeted testing of specific candidate genes through a certified lab (such as the Molecular Genetics Lab at Children’s and Women’s Health Centre of BC) offers the highest sensitivity and specificity in the context of rare genetic diseases.
Although DTC genetic tests can return a “panel” of common DNA sequences that match up some of the time with common diseases, these “panels” don’t reliably tell you who will get sick, when they will get sick, or how sick they will get. A multi-generation family medical history (three generations if known) remains the gold standard to assess risk for common diseases; DTC genetic testing doesn’t yet do better than that.
Being a Careful Consumer – A Checklist:
Practitioners who are asked by a patient about DTC genetic testing may suggest that the patient consider the following points:
1. What do you really want to find out from the test?
Ancestry tests can be fun and interesting, but should not be confused with tests that yield medically actionable information. If the patient is seeking a definitive diagnosis for a condition such as chronic pain, hormonal problems or other poorly-defined conditions, DTC testing is likely to raise more questions than it answers. If the patient is seeking a definitive diagnosis for a rare genetic condition, this is best pursued through their local Medical Genetics service.
2. How long has the company been doing DTC testing, and how many tests have they done?
There are many players in this field, and it is much easier for a company to generate DNA sequence data than it is to interpret it. Patients should avoid new start-up companies that appear to be cashing in on a trend.
3. Is the testing offered in a lab that abides by CAP- and CLIA-regulations?
Patients may not know what the Clinical Laboratory Improvement Amendments or the College of American Pathologists are, but the lab should! Labs offering testing should comply with one or both of these. Labs outside of North America should have their procedures certified by an appropriate independent body.
4. Is the company doing the test itself, or is it acting as a “broker”?
Some companies collect the DNA and send it off elsewhere for sequencing. A careful consumer or GP should know where the testing is being done, and also what is the real “value added” by the company collecting the sample.
5. Who is actually doing the interpretation of the test?
Any test that is intended to be medically actionable should be signed out of the lab by someone with the legal authority to take responsibility for medical reports – like a pathologist, medical microbiologist, molecular geneticist, and so forth. It should not be a series of fields generated by a computer.
References and Further Reading:
Ashley E et al., “Clinical assessment incorporating a personal genome.” The Lancet, Volume 375, Issue 9725, Pages 1525 – 1535, 1 May 2010
This group of authors suggest a way to start with population-prevalence data, throw in common DNA variants from whole genome sequencing and calculate an individual’s risk for developing common disease. Their algorithm generates a percentage risk, but the accuracy of this risk assessment (e.g. whether or not 10% of people who receive a 10% risk for type 2 diabetes from the calculation will actually go on to get type 2 diabetes) is not yet known.
Useful Websites to find out more:
UBC CPD Webinas:
|Monday Jan 28, 2013 7-8 pm PST||Genomics in Primary Care: What’s Ready for Prime Time? Part A with Dr. Linlea Armstrong and Dr. Bob Bluman||Register|
|Monday March 7, 2013 7-8 pm PST||Genomics in Primary Care: What’s Ready for Prime Time? Part B with Dr. Linlea Armstrong and Dr. Bob Bluman||Register|
BC Clinical Genomics Network
How to Make a Referral to the BC Cancer Agency Hereditary Cancer Program
(Site contains downloadable referral form)
How to Make a Referral to the Provincial Medical Genetics Program
(Site contains downloadable referral forms)
How to Order a Specific Genetic Test from BC’s Molecular Genetic Laboratory
(Site contains interactive test menu and downloadable requisition – please note that this lab does not offer direct-to-consumer genetic testing)
Genetics Home Reference
American College of Medical Genetics and Genomics Statement on DTC Genetic Testing