Suicide diagnostic kit or media hype?

Wouldn’t it be beneficial to be able to identify individuals at risk of suicide with a blood test? This idea has been widely covered in the media following the publication of a genetic study in The American Journal of Psychiatry at the end of July. The study reported the identification of genetic markers on a particular gene (called SKA2) that appeared to be associated with suicidal behaviour.

By testing whether or not an individual carries this marker, it was proposed that it would be possible to predict their risk of suicide with high confidence. This idea was directly suggested in a press release quoting Dr Kaminsky, Assistant Professor at Johns Hopkins University and lead author of the study. The published study is, however, much more cautious and concluded that the findings are only suggestive and need to be replicated.

A closer look at the study reveals several weaknesses. The sample size of the cohort in which the gene was identified was relatively small, and it is not possible to exclude that it might be a false positive result (that is, finding an association that doesn’t actually exist; that is why independent replication  is so important in all areas of research). The authors used additional cohorts to replicate the findings, but these were also quite small leaving the validity of this association open to question. But even if the association was genuine, the risk of dying by suicide for the individuals carrying this genetic factor was very low and not informative for being used in any diagnostic sense.

Furthermore, there was a total lack of biological characterisation of the mechanism by which the SKA2 gene might be associated with a greater risk of suicide. Suicide can be unquestionably triggered by environmental factors. We know people attempt or die by suicide for many different reasons, sometimes related to their mental health, or as a consequence of other diseases or stressful events, making it difficult to summarise or generalise what leads to suicide. While the environmental component is certainly strong, we cannot exclude that some individuals might also be at higher or lower risk of suicide because of their genetic background. We need to better understand what those genetic risk factors might be, but it is likely to be a combination of many different factors, each contributing a very small effect.

Cecile Janssens, Professor of Translational Epidemiology at Emory University, brings balance into the story with this piece carefully describing the limitations of the study and the media hype around it. Professor Janssens explains how risk for complex and multifactorial traits cannot be predicted by testing a single gene. This  is different to monogenic disorders, which are caused by mutations occurring in single genes. For example, mutations in the CFTR and HBB genes lead to cystic fibrosis or sickle-cell anemia, respectively. In these cases where there is a direct one-to-one correspondence between possessing a gene and the outcome, it makes sense to talk about diagnostic tests. Unfortunately, Professor Janssens’s well-written piece has not received much attention compared to the more sensational media coverage of the current study.

It is not uncommon to see media reports of genes being identified for any sort of complex disease or trait, from autism to happiness. In the study described here, the over-interpretation of the findings seems to have originated, at least partly, from the institutional press release. As a rule, such reports should immediately raise a red flag with the reader; two red flags if there is also the suggestion of diagnostic tests based on these findings.

Guintivano, J., et al.(2014). Identification and Replication of a Combined Epigenetic and Genetic Biomarker Predicting Suicide and Suicidal Behaviors. The American Journal of Psychiatry. DOI: 10.1176/appi.ajp.2014.14010008