THEY are called 鈥渋ncidental鈥� findings, but for the people concerned they can be anything but. Every once in a while during research on human subjects, a scientist uncovers unlooked-for information that has serious implications for a volunteer鈥檚 health.
Incidental findings are already causing dilemmas in projects that use brain imaging, during which tumours and other abnormalities sometimes come to light (see 鈥淲e need to talk about your scan鈥�). But they are poised to create even more far-reaching problems in the field of human genetics, where genome-wide studies are accumulating vast amounts of data that may have a bearing on health. Just how much of this information should be fed back to the volunteers who provided DNA samples?
In contrast with companies that scan people鈥檚 genomes and advise them on predispositions to disease (see 鈥淕enetics get personal鈥�), the norm in human genetics research has been to tell volunteers that no results will be returned. There can be exceptions, however. Researchers have long wrestled with what to do if they find that a man is not the biological father of his partner鈥檚 child. And they may disclose an incidental finding with life-threatening implications, such as when a chromosomal deletion confers a risk for a dangerous heart arrhythmia. In a case like this, seeing a cardiac specialist could save the volunteer鈥檚 life.
Advertisement
鈥淭he norm has been to tell volunteers that no results will be returned 鈥� but there can be exceptions鈥�
In the era of genomics, such ad hoc decision-making will not be good enough. Researchers looking for common genetic variants, called single nucleotide polymorphisms (SNPs), now routinely use 鈥淪NP chips鈥� that scan an individual鈥檚 DNA for a million of them at a time. Typically, research geneticists are trying to find which SNPs are associated with a disease, by asking which ones tend to appear in people with the condition but not in healthy controls. In the process, they also gather information about other SNPs that may have clinical significance.
To take one example, responses to the blood-thinning drug warfarin are affected by common SNPs in two genes, and this information can help prevent someone being given a dangerous dose of the drug. So should researchers routinely check their data for these SNPs, and send the results back to all their volunteers?
The problem with going down this road is that it would turn researchers into clinical geneticists. The researchers concerned often have no medical training, and their labs are not usually certified to provide clinical genetic tests. Also, it is difficult to see where any such extensions to their role would end. While relatively few SNPs have so far been associated with disease, the number will grow as more studies of genome-wide associations are done.
Some research geneticists are now starting to send some information back to their volunteers. One , based at the Coriell Institute for Medical Research in Camden, New Jersey, is recruiting 10,000 volunteers who will have their genomes scanned using SNP chips. The results will be used to try to uncover any links between SNPs and the risk of common conditions such as heart disease. In addition, volunteers who want to know will be given findings of medical relevance from their genetic scans if an oversight board thinks that harm can be alleviated in some way. 鈥淭he aim is to allow participants to benefit from the improved health that genome-informed medical practice will bring,鈥� the project鈥檚 website explains.
That sounds great, but most researchers lack the resources to adopt a similar approach. Do we really want research geneticists to be obliged to give their subjects鈥� genomes a complete health check, and to provide genetic counselling? Applied across the board, such a burden could bring research grinding to a halt.
So it is telling that the first set of guidelines on how to manage incidental findings comes from an expert panel funded by the US . Its recommendations provide a useful starting point for discussion. Sensibly, the NHGRI says more research on the consequences of different approaches for communicating incidental findings will be needed before a policy can be decided. For the sake of scientific progress, let鈥檚 hope those in charge of human genetics research draw a clear distinction between clinical medicine and the research that is needed to underpin future medical advances.