Screening for Bipolar Disorder in Primary Care

This clinical scenario was drawn from my own practice. I’ll tell you what I plan to do, but I’m most interested in crowdsourcing a response from all of you to collectively determine best practice. So please answer the polling question and contribute your thoughts in the comments, whether you agree or disagree.
A 22-year-old man with no significant past medical history presents and reports 4 months of fatigue. He says he has had poor sleep during that time — difficulty falling sleep as well as middle insomnia. He also has been more irritable and has less interest in his normal activities at work and with family. He admits to feeling mildly depressed, and his Patient Health Questionnaire 9 (PHQ-9) score is 21, indicating moderate to severe depression with no suicidal ideation.
He recalls having similar symptoms during his first year of college and was treated with fluoxetine. Because it did not help, he stopped the drug after 4 months. He got some counseling and worked on his diet, sleep, and exercise habits, and his symptoms improved. He has no other chronic illnesses and is not taking any medications.
The patient denies other mental health symptoms but does admit that he can “get really stressed for no reason.” He stopped using alcohol and marijuana several years ago and has no other substance use history. He lives alone and works at a bank. 
You tell the patient that major depressive disorder (MDD) is at the top of your differential diagnosis but you would like to conduct further tests to be certain. 
It’s reasonable to suspect MDD in this patient. His PHQ-9 score demands intervention to improve his symptoms. A basic workup featuring a complete blood count, TSH, and comprehensive metabolic panel is appropriate to assess for confounding conditions. However, it is not necessary to wait until these results are known before starting pharmacologic therapy in seemingly clear-cut cases of mood disorders. Bupropion could be a good agent for this patient, in part because it does not promote sexual dysfunction, but I would not prescribe any antidepressant until I performed a screening test for an important disorder that can readily mimic MDD: bipolar disorder (BD). From among these options therefore, I would have the patient complete the MDQ.
BD affects approximately 2% of the world’s population, although the prevalence of BD in the United States might be twice that figure. Overall, the prevalence of BD is similar among women and men. Bipolar I disorder features mania that is severe enough result in hospitalization, whereas bipolar II disorder is associated with less severe symptoms of mania (hypomania). Both bipolar I and II disorders are more likely to present clinically with depression rather than mania. 
Unfortunately, the average duration between a first episode of BD and a clinical diagnosis is nearly 9 years, and a prior diagnosis of depression tends to increase this lag time. Clinicians need to do better in identifying BD because the response to treatment is usually more robust earlier in the disease process.
Patients who present with symptoms of depression should be asked about previous episodes of increased activation or elevated mood. Some patients may not recall hypomanic episodes, or they may feel some shame or fear, leading to underreporting. In these cases, a friend or family member who has known the patient well over time can be invaluable in identifying BD.
The Mood Disorder Questionnaire (MDQ) is an effective and very clinically efficient tool to screen for BD. It consists of 17 questions, the first 13 of which focus on different symptoms of mania/hypomania. A previous review found that the MDQ had adequate internal consistency and a sensitivity between 73% and 90% in the distinction between BD and unipolar depression. However, the review also noted that the sensitivity of the MDQ may be lower in community vs inpatientsettings.
A positive MDQ test does not confirm the diagnosis of BD. Further is required, and engaging with a mental health specialist should be strongly considered when BD is suspected. Still, even if the patient can access a specialist, that appointment could be months away. Should primary care step into the breach and offer treatment of suspected BD in these cases? If so, what types of therapy offer the best balance of efficacy and safety? I will follow up this case with some new information and a focus on therapy for BD in another article next month.