The Risks and Benefits of Expanding the DSM Diagnostic Criteria for Bipolar Disorder

Controversy exists regarding the diagnostic boundary for bipolar disorder, as I discussed in my recent Commentary in The Journal of Clinical Psychiatry. Critics of the DSM-IV criteria have suggested that the thresholds should be lowered. In support, they cite studies in both clinical and general population samples demonstrating that individuals with subthreshold levels of bipolar pathology differed from depressed subjects without subthreshold levels of bipolar symptoms in comorbidity, personality, family history, and longitudinal course. The DSM-5 Mood Disorders Work Group is considering expanding the boundary of bipolar disorder by reducing the duration required to define a hypomanic episode. However, lowering the minimum number of symptoms required for a diagnosis of mania or hypomania is apparently not under consideration.

Expanding the diagnostic criteria for bipolar disorder carries with it potential risks and benefits. Wherever the diagnostic threshold is set, diagnostic error will still exist. Despite appropriate inquiry, patients might not recall or report prior hypomanic/manic episodes, thus resulting in false negative diagnoses. Or false negative diagnoses could result from the failure to make inquiry or the failure to make appropriate inquiry. False positive diagnoses are also a problem. It’s sometimes difficult to determine if prior hypomanic/manic episodes occurred independently of substance use, thereby resulting in false positive diagnoses. Transient episodes of affective instability and emotional lability associated with borderline personality disorder might be confused with hypomanic episodes, also resulting in false positive diagnoses.

So, the question is not whether diagnostic error exists, but rather which type of error predominates. And then, how much will shifting the diagnostic threshold impact the relative number of each of these types of diagnostic errors? Also important to consider are the clinical consequences of each type of error, and which error is more difficult to correct after it has been made.

With the existing DSM-IV diagnostic criteria, overdiagnosis (ie, false positives) is already a problem. If the diagnostic threshold is lowered, how many more patients will be misdiagnosed with bipolar disorder because brief periods of symptoms that are characteristic of cluster B personality pathology (ie, affective instability, behavioral impulsivity, or irritability and anger) will be incorrectly considered to indicate hypomania? The frequency of overdiagnosis due to lowering the diagnostic threshold must be contrasted against the frequency of underdiagnosing “true” bipolar disorder because the observed or recently occurring hypomanic syndrome did not last long enough to qualify as a DSM-IV hypomanic episode. While underdiagnosis due to insufficient duration can occur, it’s likely that patients who manifest recurrent hypomanic episodes of presumably insufficient duration during the course of treatment ultimately will be diagnosed with bipolar disorder and treated accordingly.

Advocates of lowering the diagnostic threshold emphasize the costs associated with missed diagnoses. However, a more balanced approach toward the question of where to set the diagnostic threshold recognizes that both false positive and false negative diagnoses are associated with adverse consequences. Unrecognized bipolar disorder is associated with the underprescription of mood stabilizers, an increased risk of rapid cycling, and increased costs of care. Overdiagnosed bipolar disorder is associated with overtreatment with unneeded medications and consequent exposure to potential side effects and medical risk as well as the potential failure to offer more appropriate treatments.

In trying to decide where to set the threshold for diagnosing bipolar disorder and minimizing diagnostic errors of all types, another consideration is whether one type of diagnostic error is likely to be more long-lasting and more difficult to correct than another. It’s more difficult to take away the diagnosis once it has been established than to add the diagnosis once a hypomanic episode occurs. Once a depressed patient is diagnosed with bipolar disorder, the re-occurrence of a hypomanic or manic episode is not necessary to retain the diagnosis. In fact, the lack of recurrence could be viewed as treatment success. The patient with a false positive diagnosis of bipolar disorder who is doing well on an antidepressant and a mood stabilizer is unlikely to have the mood stabilizer discontinued or the diagnosis corrected. On the other hand, a patient with a false negative diagnosis is more likely to have it changed from major depressive disorder to bipolar disorder upon the emergence of a hypomanic or manic episode. Thus, a false negative diagnosis of nonbipolar depression is easier to correct than a false positive diagnosis of bipolar disorder.

Strong evidence supporting the expansion of bipolar disorder’s diagnostic boundary would come from prospective follow-up studies demonstrating that individuals with subthreshold bipolarity are at high risk for developing bipolar disorder (as currently defined). If the majority of individuals with subthreshold symptoms of bipolar disorder develop manic or hypomanic episodes during prospective follow-up, this would indicate that the higher threshold results in more false negative than false positive diagnoses. As described in my recent Commentary, however, the results of 4 prospective 3- to 17-year follow-up studies indicated that, while subthreshold bipolarity was a risk factor for the future emergence of bipolar disorder, the vast majority of individuals did not develop bipolar disorder.

Supporters of an expanded concept of bipolar disorder indicate that the accurate identification of subthreshold forms of bipolar disorder is clinically important because of the treatment implications. Yet, the literature reviews advocating the expansion of the diagnostic boundary have not identified a single controlled study of the efficacy of mood stabilizers in the treatment of subthreshold bipolar disorder. The risk of medically significant side effects, on the other hand, is well established.

With recognition of the inherent imperfect reliability and validity of diagnoses based on the retrospective application of any set of symptom criteria, I conclude that the risk from the potential unforeseen consequences of lowering the diagnostic threshold is too great to change the diagnostic boundary of bipolar disorder in DSM-5 because of the following factors:

• the opportunity for clinicians to change a depression diagnosis to a bipolar diagnosis upon the emergence of a hypomanic or manic episode during the course of treatment
• the low likelihood that a false positive diagnosis of bipolar disorder will ever be changed, because the absence of future hypomanic/manic episodes will be viewed as treatment success
• the low likelihood that individuals with subthreshold bipolarity will experience threshold episodes during prospective follow-up
• the lack of a single controlled study demonstrating the efficacy of mood stabilizers in the treatment of subthreshold bipolarity
• the possible medically significant side effects associated with mood stabilizers

Financial disclosure:Dr Zimmerman had no relevant personal financial relationships to report.​