It’s an increasingly common scene, playing out in clinics all over the country: A patient comes in with a worried look and a fat printout from 23andMe. She wonders what all those scary red boxes mean, and whether cancer, Alzheimer’s, or some other bad disease is lurking in her genes.

The doctor, with little training in genetics and even less time for guiding patients through the nuances of risk assessment, tries to look unfazed.

Some doctors react with dismissal. Others give polite reassurance that these new genomic tests are still experimental, and that “genes are not destiny.” Still others may praise the patient for being proactive and drop a few vague lines about “methylation pathways” before diverting attention to the safer ground of routine blood work.

Even early adopters who welcome the personalized medicine revolution are scrambling to make clinical sense of the genomic universe now open to anyone with a computer and a few hundred bucks.

From a purely technological viewpoint, genome testing is definitely ready for prime time. Clinicians, by and large, are not.

At the recent annual meeting of the Institute for Functional Medicine, devoted exclusively to the “omics revolution,” more than half of the attendees indicated by show of hands that they’ve had patients come in expecting interpretations of gene data obtained from direct-to-consumer (DTC) tests. Only a handful of practitioners feel they have the skills to fulfill those expectations. Still fewer have the time.

Base Pairs & Big Data

There’s no doubt genomics will play an important role in healthcare. The federal government’s Precision Medicine Initiative—a massive research and policy effort aimed at personalizing care through genome and microbiome testing—will be a big driver.

So will the cohort of Silicon Valley entrepreneurs investing millions in businesses at the crossroad of genetics and computer science. Code is code, after all, and tech moguls are betting heavily on companies that can turn base pairs into Big Data.

Dozens of companies now offer gene testing. Some, like Pathway Genomics and Myriad Genetics, require a physician’s order. Many others, like GenePlanet and 23andMe, do not.

23andMe, which leads the pack in DTC genomics, earns more from the data it amasses than from the test kits it sells. The company has aggregated full genomes from over 800,000 people worldwide, and parlayed that into several multi-million dollar deals with pharma and biotech firms eager to mine the data mountain.

This is despite a 2013 FDA action that stopped 23andMe from marketing its once-popular disease risk assessment service (visit 23 Skidoo: FDA’s Genome Clampdown Gets Mixed Reviews). Under the injunction, the company can still sell genome tests but without any health risk analysis.

That hasn’t changed the game too much. It’s easy enough for patients or their practitioners to upload raw 23andMe data into Genetic Genie, Promethease, or any of the other online interpretation tools now available—some free, some charging small fees.

And then what?

Innovations in genomics, and their commercialization, have advanced largely without clinicians. As a result, people are swimming in genetic information that nobody knows quite what to do with.

Which SNPs Matter?

There are 80 million known variants in the human genome, according to the National Human Genome Research Institute. Which ones really matter?

That’s the question vexing anybody interested in how genomics fits in day-to-day practice, and it was a central theme at IFM’s conference.

“Genomics is the ultimate look at our individuality. The promise is healthcare that is more predictive, preventive, and personalized. But are the tests reliable & relevant?” questioned Helen Messier, MD, a physician and geneticist, who serves as medical director of Healix Health in Vancouver, BC.

Speaking at IFM, she said a clinically meaningful test must meet three criteria:

• Analytical Validity: Does the test accurately and reproducibly detect genetic variants on a methodological level?
• Clinical Validity: Does it reliably show variants that have predictive value?
• Clinical Utility: Does it provide actionable information for early disease detection, prevention, or treatment?

The tests and interpretations now available vary widely across these three parameters. Add to that the fact that genome structure—all 3 billion base pairs of it—is a lot simpler than gene function.

“We have a map, but do we really know what the terrain is?” asked Dr. Messier.

Why Patients Get Gene Tests

DTC gene tests provide a lot of curious factoids about ancestry and physical traits, as well as important information on responsiveness to common drugs, the capacity to clear drugs and toxins, and the risk of a small number of gene-linked diseases.

According to J. Scott Roberts, PhD, of the University of Michigan’s Department of Health Behavior, consumers seek gene tests for a variety of reasons.

Based on a survey of 1,848 people who got gene screens from Navigenics, he said 74% sought ancestry data. A nearly equal number (72%) wanted disease risk assessments. Fifty-two percent sought drug response info, and only 31% were concerned about carrier status for frank genetic disorders.

The top diseases of concern among test-seekers were: Heart disease (68%); Breast cancer (67%); Alzheimer’s (66%); Melanoma (59%); Prostate cancer (59%); Diabetes (56%), and Colon cancer (52%).

The vast majority (85%) was confident in the accuracy of the tests, and 84% felt they got what they paid for. Two-thirds (65%) said the data made them feel “more in control” of their health.

Dr. Roberts, who published a review looking at consumer behavior following gene tests found that, “Neither the health benefits envisioned by proponents (increases in positive health behavior) nor the worst fears expressed by critics (catastrophic psychological distress, undue burdens on the healthcare system) have materialized” (Roberts JS, Ostergren J. Curr Genet Med Rep. 2013: 1 (3): 180-200.)

By and large, he says, people do not act on gene test results.

In part, that may be because predicting the odds of complex chronic diseases is not so easy, and recommendations deduced from DTC gene tests are often conflicting

Conflicting Signals

Using herself as an example, Dr. Messier said that based on 3 SNPs, her 23andMe report (obtained before the FDA injunction) said she should take curcumin. But based on two other SNPs, the same report tells her to avoid the herb.

Recommendations may also vary considerably between gene testing companies. With regard to omega-3 intake, one company recommended she increase intake of omega-3 and omega-6. A second company, looking at the same SNPs reported that she should avoid omega‐3s and avoid over-consumption of omega‐3 rich fish.

“Racial heterogeneity is an important factor,” she said. “Always consider the ethnic background of our patient.”

For example, Dr. Messier’s 23andMe report shows that she has a 3-fold higher risk of rheumatoid arthritis (RA), though the absolute risk is still just 13%.

How did the company determine that risk? Dr. Messier explained that they looked at variations in 9 different genes for which studies have shown links to RA. “I looked at these genes myself, and for 3 of them I actually had lower risk of RA, and for 5 of them my risk was higher. Some were completely irrelevant to people of Northern European ancestry,” she explained.

For instance, the PADI4-94 polymorphism is associated with RA in several studies of East Asian populations, but showed no such association in a large UK population. In another example, she explained that single nucleotide polymorphisms (SNPs) in the CYP1A2 gene alter the activity of an enzyme that detoxifies caffeine. Yet there are marked differences between Swedes and Koreans with regard to actual caffeine metabolism, even if they show the same SNP.

GWAS’up?

Though it presents an air of exactitude, genomics is an imperfect science. The risk estimates that undergird widely used tests are based largely on genome-wide association studies (GWAS). Researchers scan full genomes and look for common gene variations that associate with specific traits. They then define predictivity in terms of odds ratios.

The truth is, most SNPs confer fairly small risk of specific diseases, and on their own have limited predictive value. GWAS show associations but not necessarily causality. The situation is further complicated by the fact that gene expression is extremely malleable.

“It is genomic origami. In every cell you have the same piece of paper. How you fold that paper determines if you get a paper airplane or a duck,” said Jeffrey Bland, PhD. “What shapes the DNA to create health? The environment, the epigenetic modulators.

Speaking at the IFM conference, Dr. Bland acknowledged a bit of collective disillusionment after the euphoria of the Human Genome Project. It was a towering intellectual and technological achievement, but it did not radically transform healthcare—not yet, anyway.

Part of the problem, said Dr. Bland is that GWAS studies give statistically significant findings for populations, but you don’t see the same penetrance on the individual level. “SNPs are not the whole story.”

Dr. Bland said he remains optimistic about genomics as a key to improving health care.

The real revolution, he believes, is in the movement away from a disease risk model that views heredity as a collection of potential disease states laying in wait. “It’s not just about reducing risk; it is about discovering the genes that lead to resilience and repair,”

Genes in Perspective

Nutritional genomics begins to do that, provided genetic information is carefully cross-referenced against metabolic tests and biomarkers, says Bridget Briggs, MD, a family physician who practices functional medicine in Temecula, CA.

In an interview with Holistic Primary Care, Dr. Briggs says she gets gene tests on all her patients. Typically, she has them get 23andMe tests, and then uses various online tools to analyze the gene data. But she never looks at the gene data in isolation.

“It gives an incredible amount of information about susceptibilities, but then I need to look at metabolic studies, stool analysis, and intracellular nutrient levels to see if someone’s actually expressing those susceptibilities.”

For example, if someone shows a SNP suggesting a methylcobalamin deficiency, but the intracellular methylcobalamin is high, the SNP is probably not being expressed and from a clinical perspective it is not really problematic.

“You can have a polymorphism for celiac disease, but if there are no inflammatory symptoms, and no inflammatory blood markers, you need not freak the patient out by telling them never to eat gluten again,” she says.

What’s important is not what SNPs show up in the test, but which ones are being expressed. “To understand that you need to look at metabolic markers and intracellular nutrient levels.”

Genetic “Bypasses”

If there’s one set of genes that have become the undisputed stars of genome science, it’s the ones coding for enzymes involved in methylation.

An essential biochemical process, methylation plays a key role in neurotransmission, amino acid metabolism, hormone and drug detoxification, vitamin metabolism, cell membrane repair, DNA synthesis and many other things.

For many clinicians, methylation is the entry point into genomics. For one, there’s a strong body of science underscoring the clinical relevance of SNPs that affect methylation-related enzymes like methylene tetrahydrofolate reductase (MTHFR), methionine synthase (MTR), and catechol O-methyltransferase (COMT).

Secondly, SNPs that negatively affect methylation are common in the general population, and associated with increased risk of CVD, Alzheimer’s, depression, and other worrisome conditions.

Thirdly, gene-based methylation defects can be counterbalanced by targeted nutritional and lifestyle interventions.

This information can enable practitioners to help a patient do a “genetic bypass,” says Dr. Briggs. By this she means that one can supplement with specific forms of nutrients downstream of the defective enzyme.

For example if someone is homozygous for a SNP called C677T, the MTHFR enzyme functions at a 70% lower capacity than in someone who does not have the SNP. This means that the conversion of folate to its active 5-MTHF form is greatly reduced, as is the individual’s ability to produce SAMe, Both are essential for neurotransmitter synthesis.

By supplementing with 5-MTHF, one bypasses the defective MTHFR enzyme. This is just one of many examples of how genomic information can be used to guide nutritional supplementation.

Help on the Way

A number of companies are providing tools to help clinicians get a handle on genomics by filtering out the noise and focusing on SNPs with the most clinical relevance. These include:

Integrated Genetic Solutions: A software system that cross-tabulates genetic information with a large number of biomarkers, features from patient history, and conventional blood work, enabling physicians to customize nutritional, pharmaceutical, and exercise regimens.

Pure Genomics: Developed and promoted by Pure Encapsulations, this analytic toolset is centered on the 8 most clinically relevant SNPs associated with methylation. Clinicians load patients’ 23andMe data into the system to obtain guided nutritional recommendations.

GeneSight: A lab focused on assessment of responsiveness to psychotropic, analgesic, and attention-related medications. GeneSight also provides info on MTHFR function, and can be used to predict potential for adverse effects from various common drugs.

Dr. Briggs, who will be a featured speaker at HPC’s upcoming Heal Thy Practice conference in October, is the medical director of EpigeneticsRx. The San Diego based education company is launching a 14-part webinar series to help more physicians understand how genetics influence metabolism, and how gene testing fits into day-to-day practice. EpigeneticsRx is drawing its guidance not only from the medical literature, but also from a network of practicing clinicians utilizing genome testing.

There are many reasonable concerns about gene testing, especially when sold directly to consumers: overhyped benefits, lack of methodological validation and standardization, easily misunderstood information, discordant reports, and risk of genetic discrimination by employers or insurers.

The latter issue is supposed to be prevented by the federal Genetic Information Nondiscrimination Act of 2008 (GINA), which prohibits companies from discriminating based on family history or genetic data. However, many people do not realize that GINA does not apply to long-term care insurance, disability, or life insurance companies.

Genome enthusiasts say none of these issues should dissuade clinicians from engaging with genetic testing and the patients who want it.

The reality is that when a patient comes in with a 23andMe report, she is telling you she’s concerned, proactive, and wishing to take some responsibility for her health. Those are great patients with whom to be working.

And, says Dr. Briggs, insurers are beginning to see it that way, too.

She said she’s been able to get insurance coverage for genomic tests, as well as stool analysis, intracellular nutrient measurement, and many other functional medicine tests for virtually all of her HMO, PPO, Medicare and Medicaid patients.

It takes a thorough understanding of the coding system, and learning what tests to order from which labs, but she contends that gene-based personalized medicine is entirely feasible within an insurance-based practice.

As clinicians begin to catch up with the technological (and commercial) developments, genomics could very well be standard-of-care within a few years.

END