New Research Suggests ‘Biomarkers’ May Predict Risk
Suicide isn’t like a heart attack. People typically don’t reveal early symptoms to their doctor—morbid thoughts, for example, instead of chest pain—and there’s no equivalent of a cholesterol or high blood pressure test to identify those at most risk of killing themselves. “We are dealing with something more complex and less accessible,” says Alexander Niculescu III, a psychiatrist at the Indiana University School of Medicine in Indianapolis. So some researchers are eager to find physical signs, called biomarkers, that can be measured in the bloodstream to signal when a person is at a high likelihood of committing suicide.
Over the past decade, Niculescu and his colleagues have been refining a method for identifying biomarkers that can distinguish psychological states. The technique depends on blood samples taken from individuals in different mental states over time—for example, from people with bipolar disorder as they swing between the disorder’s characteristic high and low moods. The researchers test those samples for differences in the activity, or expression, of genes for of different proteins. After screening the blood samples, the scientists “score” a list of candidate biomarker genes by searching for related results in a large database of studies by other groups using a program that Niculescu compares to the Google page-ranking algorithm. In previous published studies, Niculescu and other groups have used the technique to probe for biomarkers in disorders such as bipolar disorder, psychosis, and alcoholism.
In the new study, the team tested whether the approach could be used to identify people experiencing suicidal “ideation”—thoughts ranging from feelings of worthlessness to specific plans or attempts at suicide. The study required finding a rare group of people who switch dramatically from zero to high levels of suicidal ideation, Niculescu says. Because those with bipolar disorder are at a far higher risk of suicide than the general population—one in three patients attempt it—the team recruited 75 men with that diagnosis. Many were war veterans in their mid-20s to late 60s, receiving care at the Indianapolis VA Medical Center, he says.
Over the course of three interviews spaced 3 to 6 months apart, only nine of the 75 men experienced an extreme shift from zero suicidal thoughts to high suicidal ideation, Niculescu says. (Most had more chronic thoughts of suicide, he notes.) After screening the blood samples from this small group, the researchers identified five top candidates for biomarkers. Of these, a protein made by the gene SAT1, which is involved in cellular damage and stress, was “head and shoulders” above the others at distinguishing the two mental states, Niculescu says. On average, SAT1 expression was significantly elevated when the men’s interviews suggested they were experiencing suicidal ideation.
In work by other groups, the activity of SAT1 has been found to increase in response to a number of stressors such as toxins, infection, and lack of oxygen, Niculescu says. To confirm that SAT1 activity was elevated in patients who had actually committed suicide, the group examined blood samples from nine men who died by hanging, gunshot wounds, or slit wrists. (No drug overdoses were included because they would have likely altered brain chemistry.) All nine men had strikingly elevated levels ofSAT1 expression—higher, even, than those who had the highest rates of suicidal ideation in the previous study, the team reports today in Molecular Psychiatry.
Finally, to determine if the presence of elevated SAT1 and several other biomarkers that their work had flagged could predict future suicidal behavior, the group conducted a long-term study of two at-risk populations—42 men with bipolar disorder and 46 men with schizophrenia. The biomarkers were, to a small, but statistically significant degree, able to distinguish between patients who were later hospitalized for suicidal behavior and those who were not, the team reports. When combined with other simple psychological measures of anxiety and mood, the biomarkers became potentially far more helpful, Niculescu says—in the bipolar group, they were more than 80% predictive of future hospitalization for suicide.
“It’s unlikely these markers are specific” for the complex behavior of suicide, but rather for contributing factors such as abnormal cellular function due to stress, Niculescu says. If SAT1 or the other biomarkers are to be used in the clinic, scientists will need to make sure that they hold up in larger and more diverse populations, such as women and patients with other mental disorders, he says. Even if validated, the biomarkers will likely never be adequate on their own to predict how likely a person is to commit suicide, Niculescu adds. Rather, they need to be combined with other information, such as substance abuse history and stress, much like a FICO score from the credit bureau, he says.
Gustavo Turecki, a psychiatrist at McGill University in Montreal, Canada, was “particularly happy” when he read the study’s results because it fits well with his own lab’s previously published work, which has also found evidence that the production ofSAT1’s protein is elevated in the brains of suicidal individuals after they die. Although he’s hopeful that the findings will eventually illuminate the biological processes that underlie suicidal behavior, Turecki doubts that these biomarkers will ultimately help predict suicide in the clinic. The small sample sizes in Niculescu’s studies and the fact that all the patients were male and had bipolar disorder or schizophrenia means that the study is “not representative of the universe of people with suicidal behavior,” he says. More important, he says, is that the vast majority of people who die by suicide don’t see a doctor first.
“Any one test isn’t going to be a perfect predictor,” agrees psychologist Matthew Nock of Harvard University, an expert in developing behavioral exams to predict suicide. However, looking for biological warning signs of suicide is “an important piece of the puzzle,” he says. “Hopefully we’ll see a lot more in this line of research.