Bipolar Diagnosis: Navigating Between Scylla and Charybdis

May 1, 2007

When a new patient with depression enters your practice, you face a diagnostic dilemma. If you miss bipolar disorder (BD), and prescribe an antidepressant, you can do harm. But if you call a unipolar depression "bipolar," you may also do harm, because lithium, anticonvulsants, and atypical antipsychotics carry significant risk as both short- and long-term treatments. In addition, the label of "BD" currently carries much more stigma than the term "depression."

When a new patient with depression enters your practice, you face a diagnostic dilemma. If you miss bipolar disorder (BD), and prescribe an antidepressant, you can do harm. But if you call a unipolar depression "bipolar," you may also do harm, because lithium, anticonvulsants, and atypical antipsychotics carry significant risk as both short- and long-term treatments. In addition, the label of "BD" currently carries much more stigma than the term "depression."

Many concerns have been raised about the overdiagnosis of BD,1,2 especially in child and adolescent psychiatry.3 In the very young, this diagnosis is particularly difficult: the differential is broad; the stigma is particularly burdensome and potentially long-lasting; and treatments pose their gravest risk, especially given the potential duration of exposure and unknown developmental effects. At the same time, systematic review of existing data support the opposite concern-that of continued underdiagnosis of BD-at least in adults.4,5

Why is our profession simultaneously concerned about overdiagnosis and underdiagnosis? This paradox stems in part from limitations of the DSM system. A categorical system is, by its nature, forced to create divisions that include some patients and exclude others, even if there is no recognizable juncture of separation. It cannot identify patients whose symptoms and signs fall below official thresholds yet who will eventually manifest a bipolar outcome over time and in their response to treatment. By contrast, a diagnostic approach that can recognize varying degreesof bipolarity (usually referred to as a "dimensional system") avoids the problem of potentially arbitrary division. However, it creates other problems, including a lack of precise boundaries with resultant increased interobserver variation (at minimum) and substantial logistical problems (eg, coding). Moreover, a major overhaul in our diagnostic system would create its own chaos, which is a significant barrier against change.

Nonmanic bipolar markers
Nevertheless, mood researchers increasingly emphasize the importance of nonmanic bipolar markers (also known as "external validating features"6 or colloquially as "soft signs"). These illness features, enumerated below, are statistically associated with an eventual bipolar course. In an example of this trend, the International Society for Bipolar Disorders recently convened 25 bipolar specialists to prepare literature reviews for each of several problematic aspects of bipolar diagnosis, requesting specific recommendations for interim changes in DSM-IV. Scheduled for publication in mid-2007, these reviews include examinations of the lower boundary of hypomania in BDII6; presence of nonmanic bipolar markers in patients who present with depression7 and mixed states8; and the diagnostic implications of a bipolar spectrum perspective.9 All of these emphasize the diagnostic importance of nonmanic bipolar markers.

However, this is not a new concept. Pies10 examined bipolar recognition from Aristotle to Kraepelin, demonstrating that the latter's broad view of bipolarity had deep historical roots. Nevertheless, an important transition in diagnostic thought came 5 years ago with a review by Ghaemi and colleagues11 in which they summarized data supporting the use of 11 nonmanic markers in the process of diagnosing bipolar disorder:

  • Repeated (4 or more) episodes of major depression.
  • Early age at onsetof mood disturbance (before age 25).1
  • Family historyof BD, particularly a first-degree relative.
  • Hyperthymicpersonality.
  • Atypical depression (hypersomnia, hyperphagia, leaden anergy, rejection hypersensitivity).
  • Brief episodes of major depression, eg, lasting less than 3 months.
  • Psychosis.
  • Postpartumonset.
  • Hypomaniawhen patient is given an antidepressant.
  • Loss of patient responseto an antidepressant agent.
  • Three or moreantidepressant agents ineffective.

This approach was also adopted by Mitchell and colleagues,7 who proposed using these features to adjust the probability of BD when patients present with depression, independent of the assessment of manic symptoms. These nonmanic bipolar markers have also been incorporated into a diagnostic system that is currently being tested in the Systematic Treatment Enhancement Program for Bipolar Disorder.12 The Bipolarity Index organizes nonmanic bipolar markers into 5 dimensions (easier to remember and document than 11 soft signs):

  • Hypomania or mania.
  • Family history of mood and substance use problems.
  • Patient's age at the onset of mood symptoms.
  • Illness course and other features generally only visible over time.
  • Response to medications (antidepressants and mood stabilizers).

The index thus places 4 additional dimensions of bipolarity alongside the DSM-IV criteria, openly changing the diagnostic question from "Does the patient have bipolar disorder?" to "How much bipolarity does the patient have?"13

Another system for memorizing and routinely assessing nonmanic bipolar markers was recently prepared by Pies14; his whiplashED mnemonic appears in Table 1.


WHIPLASHED 14: nonmanic markers
Worse or wired when taking an antidepressant
Hyperthymic temperament or subthreshold
Hypomania Irritability
Psychomotor retardation or agitation
Loaded family history of mood/substance use disorders
Abrupt onset or termination/brief episodes of depression
Seasonal or postpartum depression
Hyperphagia or Hypersomnia
Early age at onset
Delusions, hallucinations, or other psychotic features

DIGFAST15: manic symptoms
Distractibility (poorly focused, multitasking)
Insomnia (decreased need for sleep)
Grandiosity (inflated self-esteem)
Flight of ideas (racing thoughts)
Activities (increased goal-directed activities)
Speech (pressured)
Thoughtlessness (risk-taking behaviors)

Screening for manic symptoms
As its progenitors emphasize,13 the Bipolarity Index supplements but does not replace DSM-IV. Clinicians must still screen for manic or hypomanic symptoms to assess the first dimension of the index. Examples of systematizing this screening in routine clinical care include the Digfast mnemonic (originated by Falk15), the Mood Disorders Questionnaire (MDQ),16 MDQ-Adolescent,17 the Bipolar Spectrum Diagnostic Scale (BSDS),18 and the Hypomania Checklist (HCL-32).19 These questionnaires are self-report instruments that take just a few moments to complete and are available in various forms on the Internet (Table 2).

While the MDQ may be best suited for a primary care practice (because of its quick and easy administration and interpretation), the BSDS and HCL-32 may be preferable in a psychiatric practice, because they have greater sensitivity for BDII18,20,21 (although potentially at the expense of lower specificity20). Zimmerman and colleagues22 argued that the sensitivity and specificity of the MDQ may be too low to justify its use as a screening instrument.

However, another recent analysis23 suggests that these instruments are sufficiently accurate when used to augment an educated clinician's hunch-which, according to the above data, should be based in part on an assessment of nonmanic bipolar markers. If the clinician's degree of suspicion is high, the predictive value of a positive MDQ or BSDS is between 85% and 90% (with one outlier study at 65%). Conversely, if the clinician's suspicion is low and the test is negative, the negative predictive value is over 90% (no outliers). Thus, these tests are particularly good for confirming a negative impression and acceptably strong for affirming a suspicion of BD, but their results must be evaluated with great caution when they run counter to an informed clinical hunch.23

What should be done when one's hunch and a screen for manic symptoms (eg, MDQ, BSDS, HCL-32, or DIGFAST) conflict? If clinical suspicion is high but the screening is negative, the clinician must consider that the patient lacks insight into his/her manic symptoms and thus, collateral information is needed. If clinical suspicion is low but the test is positive, there is a high risk of a false positive, following the categorical approachof DSM-IV. But is the test result still meaningful in some way? Does such a result indicate some form of subthreshold BD? Current research provides scant data to address this question (a biological marker of some sort could help here). In the meantime, the answer depends on how one conceptualizes bipolarity.

What are we modeling?
The Figure shows several different ways of thinking about the diagnosis of BD. With the categorical DSM approach there is no middle ground; a patient either has unipolar depression or 1 of 3 discrete levels of bipolarity. By contrast, the data on nonmanic bipolar markers suggest that patients can have many different degrees of bipolarity. If so, does bipolar propensity increase exponentially (Model B) or linearly (Model C)? This will probably remain unclear until we have a better understanding of the etiology of BD.

Could a patient even have some degree of bipolarity with no hypomania or mania at all, as indicated by the pair of asterisks in the Figure? This theory was proposed by Ghaemi and colleagues,11 who characterized bipolar spectrum disorder asmajor depression accompanied by multiple nonmanic bipolar markers. Family history of BD in a first-degree relative and antidepressant-induced hypomania are given particular weight. Although this concept was found useful in a small pilot study24 and is a logical extension when thinking in spectrum terms, it is not a validated entity. Do such patients have more adverse reactions to antidepressants than purely unipolar patients? The answer to this important question remains unknown.

At present, clinicians must prove that bipolarity is present by looking for manic symptoms. If none are found, the patient is presumed unipolar. If the search is conducted cursorily, or not at all, the diagnosis defaults to major depression; and if a medication is used, it will almost certainly be an antidepressant. Routine use of tools described herein could improve the accuracy of this rule-out process, especially in primary care.

However, at a conceptual level, bipolar depression might be a better default assumption when a patient presents with depression. With this approach, the burden of proof lies in demonstrating that manic markers, including all 5 dimensions of the Bipolarity Index, are notpresent. This might lower the incidence of antidepressant prescriptions for patients for whom antidepressants may worsen their condition. But what if only a few indicators of bipolarity are present? When should a clinician begin to be concerned about using an antidepressant? For now, physicians' approaches in this situation are likely to be determined primarily by their level of concern about antidepressant risks in patients with BD.25 For an example of a data-based and very concerned view, see the recent review by El-Mallakh and colleagues.26

Several investigators have presented data showing that antidepressant monotherapy in BDII can be associated with good outcomes, at least in the short term and for some patients.27,28 But clearly at some point along the continuum (Figure), antidepressants become, if not contraindicated, at least less than optimal initial therapy. Currently psychiatrists use DSM-IV criteria as a proxy indicator of which patients should receive mood stabilizers. (When the primary target is depression, we would presumably select from those that have demonstrated antidepressant efficacy.) However, this assumes that bipolarity is a step function, as shown for the DSM-IV model in the Figure. If B or C are better models, as appears to be the case, we should think more broadly about which patients can safely be given an antidepressant.

Do continuum models risk overpromotion of mood stabilizers for patients who do not have BD? This is an obvious and concerning logical extrapolation. Historically, we have protected patients from the risks of the medication treatments for BD and the stigma of the label by setting the bar high for making this diagnosis, especially in children. However, another way to protect patients would be to set the bar high for anypharmacological approach, given the data on the efficacy of such alternatives as psychotherapy, exercise, and light therapies. This approach would protect against both the Scylla of underdiagnosis and antidepressant exposure and the Charybdis of overdiagnosis and mood-stabilizer risks-as long as a more aggressive approach, where potentially necessary, is outlined in advance and ready for launch.


References1. Patten SB. Does almost everybody suffer from a bipolar disorder? Can J Psychiatry. 2006;51:6-8.
2. Sobo S. Mood stabilizers and mood swings: in search of a definition. Psychiatr Times. 1999;16(10):36-42. Available at: Accessed April 10, 2007.
3. Harris J. The increased diagnosis of "juvenile bipolar disorder": what are we treating? Psychiatr Serv. 2005; 56:529-531.
4. Ghaemi SN, Sachs GS, Chiou AM, et al. Is bipolar disorder still underdiagnosed? Are antidepressants overutilized? J Affect Disord. 1999;52:135-144.
5. Albanese MJ, Clodfelter RC Jr, Pardo TB, Ghaemi SN. Underdiagnosis of bipolar disorder in men with substance use disorder. J Psychiatr Pract. 2006;12:124-127.
6. Vieta E, Suppes T. Bipolar II disorder. Bipolar Disord. In press.
7. Mitchell PB, Goodwin GM, Johnson GF, Hirschfeld RMA. Diagnostic guidelines for bipolar I depression: a probabilistic approach. Bipolar Disord. In press.
8. McIntyre RS, Benazzi F, Maj M. Mixed states: diagnostic consensus and future research vistas. Bipolar Disord. In press.
9. Phelps JR, Angst J, Katzow J, Sadler J. Validity and utility of the bipolar spectrum model. Bipolar Disord. In press.
10. Pies R. The historical roots of the "bipolar spectrum": did Aristotle anticipate Kraepelin's broad concept of manic-depression? J Affect Disord. 2007 Jan 12; [Epub ahead of print].
11. Ghaemi SN, Ko JY, Goodwin FK. "Cade's disease" and beyond: misdiagnosis, antidepressant use, and a proposed definition for bipolar spectrum disorder. Can J Psychiatry. 2002;47:125-134.
12. Sachs GS. Strategies for improving treatment of bipolar disorder: integration of measurement and management. Acta Psychiatr Scand Suppl. 2004;422:7-17.
13. Saenger E. The Bipolarity Index as a tool for assessment and creating rapport: an expert interview with Gary Sachs, MD. Medscape. Psychiatry & Mental Health. 2005;10. Available at: Accessed January 28, 2007.
14. Pies R. Is it bipolar depression? Whiplashed aids diagnosis. Curr Psychiatry Rep. 2007;6:80-81.
15. Ghaemi SN. Mood Disorders: A Practical Guide. Philadelphia: Lippincott Williams & Wilkins; 2003:13.
16. Hirschfeld RM, Williams JB, Spitzer RL, et al. Development and validation of a screening instrument for bipolar spectrum disorder: the Mood Disorder Questionnaire. Am J Psychiatry. 2000;157:1873-1875.
17. Wagner KD, Hirschfeld RM, Emslie GJ, et al. Validation of the Mood Disorder Questionnaire for bipolar disorders in adolescents. J Clin Psychiatry. 2006;67:827-830.
18. Ghaemi NS, Miller CJ, Berv DA, et al. Sensitivity and specificity of a new bipolar spectrum diagnostic scale. J Affect Disord. 2005;84:273-277.
19. Angst J, Adolfsson R, Benazzi F, et al. The HCL-32: towards a self-assessment tool for hypomanic symptoms in outpatients. J Affect Disord. 2005;88:217-233.
20. Vieta E, Sanchez-Moreno J, Bulbena A, et al. Cross validation with the mood disorder questionnaire (MDQ) of an instrument for the detection of hypomania in Spanish: the 32 item hypomania symptom check list (HCL-32). J Affect Disord. 2006 Dec; [Epub ahead of print].
21. Carta MG, Hardoy MC, Cadeddu M, et al. The accuracy of the Italian version of the Hypomania Checklist (HCL-32) for the screening of bipolar disorders and comparison with the Mood Disorder Questionnaire (MDQ) in a clinical sample. Clin Pract Epidemol Ment Health. 2006;2:2.
22. Zimmerman M, Posternak MA, Chelminski I, Solomon DA. Using questionnaires to screen for psychiatric disorders: a comment on a study of screening for bipolar disorder in the community. J Clin Psychiatry. 2004;65: 605-610.
23. Phelps JR, Ghaemi SN. Improving the diagnosis of bipolar disorder: predictive value of screening tests. J Affect Disord.2006;92:141-148.
24. Ghaemi SN, Hsu DJ, Ko JY, et al. Bipolar spectrum disorder: a pilot study. Psychopathology. 2004;37:222-226.
25. Phelps J. Bipolar disorder: particle or wave? DSM categories or spectrum dimensions? Psychiatr Times. 2006;23(8):76-82.
26. El-Mallakh R, Karippot A, Ghaemi SN. Antidepressants in bipolar depression. In: El-Mallakh R, Ghaemi SN, eds. Bipolar Depression: A Comprehensive Guide. Washington, DC: American Psychiatric Publishing; 2006.
27. Amsterdam JD, Shults J. Fluoxetine monotherapy of bipolar type II and bipolar NOS major depression: a double-blind, placebo-substitution, continuation study. Int Clin Psychopharmacol. 2005;20:257-264.
28. Parker G, Tully L, Olley A, Hadzi-Pavlovic D. SSRIs as mood stabilizers for bipolar II disorder? A proof of concept study. J Affect Disord. 2006;92:205-214.