Bipolar disorder (BD) can be a difficult condition to diagnose because its signature symptoms–episodes of abnormal, often persistent, highs and lows–are related to one another in different ways in different people. We often think of highs and lows as mutually exclusive opposites. Yet in BD they are not opposites but are sometimes “mixed” in varying degrees of intensity.

One can be depressed, for instance, and yet for brief intervals– say, a couple of days–display certain features of mania, or a less severe form of mania called hypomania (for example, elation, increased energy, decreased need for sleep, rapid speech, irritability, a tendency toward risky behavior). It’s also possible to experience milder or “subthreshold” symptoms that aren’t classified as either manic or depressive. In some patients, depression may be the dominant mood; in others, there will be distinct periods of mania and depression of varying duration, and in others very rapid changes in mood. A fairly new term, bipolar spectrum disorder (BPSD), covers the full range–on the one hand, full-blown BD featuring depression plus at least one period of mania or hypomania, but also subthreshold depressive and/or manic mood symptoms. BPSD is an umbrella term that emphasizes that the manifestations of BD exist in a continuum.

Identifying patterns–in moods, behaviors, brain activity, gene activation, even the body’s metabolism–can distinguish different sub-groups of patients, and is a major objective of research being conducted by many of the Foundation’s grantees. Describing these patterns and determining their prevalence in a growing range of illnesses from psychosis and schizophrenia to depression and suicidality–is now leading to the development of the first tools to predict risk, as well as the course a disorder will take in specific individuals, a major achievement that is decades in the making.

At the University of Pittsburgh’s School of Medicine and Western Psychiatric Clinic, Boris Birmaher, M.D., Endowed Chair in Early Bipolar Disorder and 2013 recipient of the Colvin Prize for Mood Disorders Research, has for the past 17 years led a highly impactful study that exemplifies how the analysis of a single, large patient cohort over an extended period of time can generate the kind of knowledge needed to improve patient care.

Dr. Birmaher heads The Pittsburgh Bipolar Offspring Study, or BIOS, which is looking at the mental health of children born to a parent with a diagnosis of bipolar disorder. By the early 2000s, when BIOS got underway, it was already clear that there was no more powerful factor affecting a child’s risk of developing BD. By age 21, about 3.4 percent of the general population will be diagnosed with BD, a rate that Dr. Birmaher’s group and many others assumed was far higher in children with at least one parent with the diagnosis. But how much higher? No one knew for sure.

There were lots of other unknowns. Was there a way to predict which high-risk children would “convert” to the illness, and if so, which form of it and at what point in their development? Just as important, was there a biological or behavioral pattern –a “signature”–for high-risk children who probably would not develop BD? What was the risk that children of affected parents would develop other psychiatric or behavioral issues?

Joining Dr. Birmaher in this work from its inception have been 2001 Distinguished Investigator and 2006 Ruane Prize for Child and Adolescent Psychiatric Research recipient David A. Brent, M.D. at the University of Pittsburgh and David Axelson, M.D., currently the Director of Child Psychiatry at the Nationswide Children’s Hospital in Columbus, Ohio.

As the study has progressed, they have been joined by 2014 Independent Investigator and 2007 Young Investigator Benjamin I. Goldstein, M.D., Tina Goldstein, Ph.D., Danella Hafeman, M.D., Ph.D., and 2008 Young Investigator Dara Sakolsky, M.D., Ph.D., who are also at the University of Pittsburgh.

First BIOS Results

In 2009, the BIOS study generated its first headlines. Six years after contacting over 1,600 people living within 200 miles of Pittsburgh, they assembled an initial study cohort of 388 children of 233 parents with BD, plus 251 children of 143 demographically matched control parents.

Before BIOS, various experts estimated that children of BD parents aged six to 17 would have anywhere from two to seven times the risk of developing BD symptoms as compared with children of parents without BD. BIOS showed the risk to be 14 times higher.

It also revealed a two- to three-fold greater incidence in these high-risk children of developing any mood or anxiety disorder. Families in which both parents had BD generally had more offspring with BD spectrum disorders than families with one affected parent. And a very important finding from the study revealed that in children of affected parents who developed BD, episodes began during childhood, usually before age 12, most often manifesting with sub-threshold manic symptoms, and to a lesser degree, depression. Fully 85 percent of the children who developed BD had comorbid conditions–usually anxiety disorders, disruptive behavior and/or ADHD–that typically preceded the onset of BD.

The study made clear that children of parents with bipolar illness were indeed at very high risk of developing the disorder themselves. But there was a ray of light in the first analysis of data from the study. “Because nearly half the children of parents with BD have not yet manifested any diagnosable psychiatric illness, there is a great need and opportunity for primary prevention in this high-risk population,” Dr. Birmaher and colleagues concluded.

Two years later, in February 2010, the BIOS team announced more newsworthy results. While the first results had analyzed children of school age, this time the focus was on children of preschool age. In a group of 121 preschoolers, aged two to five, of 83 parents with bipolar disorder, the risk of developing ADHD was calculated to be eight times that of a matched control sample consisting of 102 children of 65 parents. Children of parents with BD also had six times the risk of having two or more other psychiatric disorders.

Again, there was a ray of hope generated by these worrying results. At the time of the report, only three of the 121 preschool children of bipolar-diagnosed parents had developed mild depressions, and none had developed BD. The remainder, particularly those with ADHD, were much more likely than children of control parents to have subclinical manic and depressive symptoms. “We believe there is a window of opportunity for prevention in the high-risk group of kids,” Dr. Birmaher said at the time of the study’s release.

Another report from the BIOS team appeared in 2016. In the pages of the American Journal of Psychiatry, Drs. Hafeman and Birmaher and the BIOS team now were able to measure the risk that children of bipolar parents would show warning signs, sometimes called a “prodrome” period by doctors. Children of BD parents with symptoms of depression, anxiety, unstable mood, and subclinical manic symptoms were at high risk to develop BD. The risk of developing BD increased to almost 50 percent in children with these symptoms whose parents had developed BD before age 21.

A Calculator to Measure Risk

This result highlighted a familiar problem. A major depressive episode is known to be a warning sign of risk for conversion to bipolar disorder. But only a minority of depressed young people will ever experience mania or hypomania and therefore receive a BD diagnosis and treatments specific to BD, as opposed to depression. Among other things, antidepressants may not help a young person whose depression is just the prelude to mania and BD. Those with the diagnosis are usually treated with mood stabilizers including lithium, anti-seizure and antipsychotic medications.

The BIOS team’s 2016 paper that identified prodromal symptoms before the onset of mania drew from results across the study cohort as a whole, but did not identify the individual risk for specific children. To address this issue, a paper published by the BIOS team in August 2017 in JAMA Psychiatry brought hopeful news.

Based on a study cohort that now numbered 412 children of parents with BD–of whom 54 had themselves developed BD during the follow-up period of the study–the team was now able to construct a risk calculator. Based on established criteria for assessing risk for BD–mood and anxiety symptoms, general psychosocial functioning, and age of one’s parent when she or he began to suffer from a mood disorder–the risk calculator was tested in the BIOS study population, where the researchers had observed some of the high-risk young people initially enrolled actually develop the illness over the course of the study.

Estimating the preliminary or prodrome period for BD at anywhere from two to 10 years, depending on the individual, the team noted that all of the early warning signs were not in themselves specific to BD. One could be anxious or habitually defy authority or be irritable or have sleep disturbances or be depressed–and not go on to develop mania and BD.

Yet such symptoms as factored into the risk calculator tested by the team succeeded with a 70 percent accuracy of “predicting” which of the high-risk young people in the BIOS study did go on to receive a BD diagnosis within five years of their “check-in” assessment. The accuracy was by no means perfect, but it was almost exactly equal to that used in risk assessments for heart disease and colorectal cancer that are widely adopted in medicine.

Dr. Birmaher and colleagues caution that the risk calculator is not yet ready for clinical use because it needs to be tested in sample populations not involved in the BIOS study. Yet the tool does give a sense, finally, of what doctors should look for in trying to assess whether a specific young patient is at high risk of developing BD within the next five years.

Preventive interventions can be undertaken in those whose risk is found to be high. The tool is equally valuable for researchers, who now can pay particularly close attention to those thought to be likely to develop the illness but who have not yet done so. These are ideal candidates for state-of-the-art brain imaging and other monitoring tools, which have a good chance of discovering telltale biomarkers that will make predicting who will get sick ever more accurate in the years to come.

— Written By Peter Tarr, Ph.D.

Click here to read the Brain & Behavior Magazine's September 2018 issue