A young woman faints at a party and her friends take her to hospital. Routine investigations include an ECG, electrocardiography, which shows a relatively prolonged QT interval. This might be part of the normal human variation in QT intervals, or it might indicate she has long QT syndrome. If so, this could explain her faint and mean she is at risk of further faints or, if the heart rhythm disturbance is severe, sudden death.
A heart rhythm panel gene test reveals a variation in the LQT1 gene: a single letter T is replaced by a letter A in a potentially important part of the gene. This particular genetic code variation has not been described before in the several thousand individuals that the lab has tested.
What to make of this finding? Should the young woman now be regularly reviewed by cardiologists, offered medication or even an implantable cardiac defibrillator? Or should the finding be ignored until more evidence is accumulated that it is causally related to her faint? Should her immediate family members be offered a genetic test for this variant together with cardiology assessments?
Such scenarios feature increasingly frequently in everyday medical practice. The last few decades have seen the cost of determining the genetic sequence of an entire human genome drop one million-fold, with a similar magnitude increase in the speed of analysis.
These technical advances have enabled great leaps in our understanding of the mechanisms and aetiology of health and disease and have led to much optimism about the ability of genomics to transform healthcare. However, the gap between discovering genomic contributions to diseases and testing an individual's genome for personal, actionable, information about health is wider than the current discourse on genomics might suggest.
If certain genes are known to be associated with certain diseases, then is analysing these genes in patients who have a family or personal history of such a condition not a logical next step? There are several (overlapping) reasons that such a leap should only be taken with caution:
The degree of normal human genetic variation (variation in our genetic codes not linked to disease) has not yet been fully determined. There isn't a normal genome against which others can be compared. Although much of our genomes are the same, they are also subtly different - there are probably around six million differences per genome. Where a difference clearly disrupts the gene message - and the gene is known to have a particular function - interpretation is sometimes easier, but in the clinical example above there just isn't enough evidence to know for sure that the variant is important.
Recent advances have improved the quality of genome testing. Almost the whole genome can be analysed accurately to predict the presence or absence of a particular genetic variant, but the clinical validity of the variant (how well it is related to the presence or absence or future risk of a disease) requires much more information than the just the sequence itself.
Many common diseases are multifactorial - several different genetic variants interact with several different environmental factors to trigger the disease. We don't yet know how all these interactions work in different people, let alone what all the different factors are. Therefore, identifying any one, or even more, factors in such complex causations doesn't help predict the presence or the absence of the disease in an individual very well. In other words, the clinical utility - whether the test can provide helpful information about treatment, management, or prevention of a disease - is low.
The genetic background on which a variant lies may also affect the risk it confers. This means that the same variant found in a woman of, say, African ancestry might confer a different risk in a man of northern European ancestry. Some populations have been studied in greater detail than others, meaning far less is known about what is normal variation for some people.
Before the advent of new technologies, genetic tests involved screening particular genes based on signs, symptoms or family histories. Tests were expensive, so only those with the strongest family history, or the most characteristic features, would be offered a test. These were also the most likely to have an underlying clear-cut genetic variant that was easy to interpret. As tests get cheaper, this approach is reversed - genomic tests are done when the diagnosis is unclear, or in order to predict what signs or symptoms someone might develop. Interpreting the results from such testing is much harder because there are many shades of grey between disease-causing and completely benign genetic variants.
BioNews recently highlighted a New England Journal of Medicine paper that demonstrated the gap between a sequence and its interpretation beautifully. While the genetic codes were reliably analysed, the clinical interpretation often provided very different take-home messages. In our example above, it is clearly important to know what, if any, risk this genetic variant confers. At what level of risk can our patient be reassured, and at what level should treatments be started? How certain do we need to be?
To answer these questions we need to gather a much better evidence base than currently exists. Large-scale international data-sharing exercises will be necessary to provide some of these answers. The more clinical data that is attached to genomic variation, the more useful it is, but the greater the potential concern that an individual's privacy might thereby be breached, or that the governance of the data will become too difficult or problematic.
Patients such as our hypothetical case above often express surprise and disappointment that concerns about genetic privacy mean that their result is less interpretable than it might have been. Efforts such as the database outlined by Rehm et al. are absolutely crucial, together with an intelligent debate about how, and when, genetic privacy is threatened and how it can best be protected . Without this, our progress will be slow, and we will not be able to maximise the benefits to healthcare that the genomic revolution offers us.