Important Distinctions Between Bipolar I and Bipolar II Depression
Terence A. Ketter, MD
The Department of Psychiatry and Behavioral Sciences and the Bipolar Disorders Clinic, Stanford School of Medicine, Stanford, California
Patients with bipolar disorder can spend up to half of their time with depressive symptoms ().1,2 By definition, patients with bipolar II disorder compared with those with bipolar I disorder experience less severe mood elevation episodes (ie, hypomanias rather than manias), which thus do not entail psychosis, psychiatric hospitalization, or severe consequences (such as financial or legal problems and divorce). However, bipolar II disorder ought not to be considered “bipolar light,” as patients with bipolar II disorder compared with those with bipolar I disorder spend more time depressed and have comparable functional impairment and risk of suicide. Also, bipolar II disorder compared with bipolar I disorder is associated with more rapid cycling and comorbid anxiety disorders.3
AV 1. Depressive Symptoms in Patients with Bipolar I and Bipolar II Disorder (00:19)
Although several agents are available for the treatment of acute manic and mixed episodes, treatment options for bipolar depression are far more limited. Indeed, the US FDA has only approved 2 treatments for acute bipolar depression:
- quetiapine (a SGA) for bipolar I depression and bipolar II depression, and
- olanzapine plus fluoxetine (a SGA plus an antidepressant) for bipolar I depression (but not for bipolar II depression)
Other treatment options for acute bipolar depression include the mood stabilizers lamotrigine and lithium, which are FDA approved for bipolar I disorder maintenance (but not for acute bipolar depression), and antidepressants, which (aside from fluoxetine combined with olanzapine) lack FDA approval for any phase of bipolar disorder. In addition, recent studies4,5 support bipolar I depression efficacy for the SGA lurasidone (both as monotherapy and added to lithium or valproate) and for the low-affinity dopamine transport inhibitor armodafinil (added to bipolar I maintenance treatments other than quetiapine), and so may ultimately receive FDA approval for bipolar I depression.
To determine which medication is best for individual patients, clinicians need to devise treatment plans that integrate patient preferences regarding potential benefits (therapeutic effects) and harms (side effects) of interventions.
To measure the therapeutic benefit of a medication, clinicians can use number needed to treat (NNT), which indicates the number of patients who must be treated with 1 intervention compared with another in order to be able to expect 1 additional good outcome (eg, 1 additional response, which is 1 additional patient with an at least 50% decrease in symptoms). The smaller the NNT, the larger the therapeutic advantage.
NNT is commonly used as a metric of how much better a medication is compared with placebo (ie, a metric of therapeutic effect size). In such applications, NNT can be calculated by dividing 100% by how much greater the medication response rate is compared with placebo.
Example:drug response = 50%, placebo response = 25%
drug response advantage is 25% (50% – 25%)
so, NNT = 100% ÷ 25% = 4
that is, 4 patients need to be treated to expect 1 additional patient to respond compared with placebo
NNT is rounded up to the next highest whole number to keep results from being overly optimistic. In general, FDA-approved treatments for bipolar disorder have single-digit NNTs, indicating that they yield response rates that are at least 10% greater than that seen with placebo. Negative NNTs are problematic, as they indicate treatments are less likely than placebo to yield response.
Prescribing any drug means balancing potential benefits (therapeutic effects) and potential harms (side effects). Number needed to harm (NNH) is commonly used as a metric of how much worse a medication is compared with placebo (ie, a metric of side effect size). NNH indicates the number of patients who must be treated with 1 intervention compared with another in order to be able to expect 1 additional bad outcome (eg, 1 additional patient with a side effect). NNH can be calculated by dividing 100% by how much greater the medication side effect rate is compared with placebo.
Example:drug sedation rate = 40%, placebo sedation rate = 20%
drug sedation disadvantage is 20% (40% – 20%)
so, NNH = 100% ÷ 20% = 5
that is, 5 patients need to be treated to expect 1 additional patient to experience sedation compared with placebo
Like NNT, NNH is rounded up to the next higher whole number. Higher NNHs represent better outcomes, with double digits (preferably high double digits) commonly being considered adequate, as they indicate treatments are less than 10% worse than placebo with respect to side effect rates. Single digit NNHs are problematic, as they indicate treatments yield side effects rates that are at least 10% greater than with placebo. Negative NNHs are particularly desirable, as they indicate treatments are less likely than placebo to yield side effects.
AV 2. Bipolar Depression Medications: Tolerability Vs Efficacy (0:42)
For psychotropic medications, NNHs tend to range from low (greater side effects risk, ie, less desirable) with SGAs to intermediate with mood stabilizers to high (lower side effects risk, ie, more desirable) with antidepressants. Clinicians whose patients are particularly concerned about avoiding side effects may consider starting treatment with more tolerable agents. A potential problem with such a “tolerability first” approach is that more tolerable (less side effect likelihood) treatments may also have less efficacy (less therapeutic effect likelihood), which can be particularly challenging when encountering clinically urgent mood symptoms (). Conversely, an “efficacy first” approach, which may be considered when encountering clinically urgent mood symptoms, may entail poorer tolerability, which can be particularly challenging when encountering patients who are particularly concerned about avoiding side effects. Thus, when selecting treatments, providers and patients need to carefully collaborate to balance the needs for efficacy and tolerability.
Applying Efficacy and Tolerability Knowledge to Patient Treatment
Clinicians and patients both want to select treatments that make benefit more likely than harm, or that entail a lower NNT than NNH. However, the variable natures of the need for efficacy and the types of adverse event influence how clinicians and patients view benefit:risk assessments. For example, the prospective risks of sustained sedation and weight gain (some of the most common side effects) commonly cause patients more concern than risks of mild transient side effects (such as mild transient nausea or akathisia).
AV 3. NNT and NNH in Acute Bipolar Depression Studies (0:31)
For patients experiencing severe acute depression (especially if functioning is significantly impaired and/or there are clinical safety concerns), the need for efficacy mitigates side effect concerns. Providers may find integrating NNT and NNH benefit:risk data for available medications with patient preferences can help optimize clinical decision making ().6
Second-generation antipsychotics. Quetiapine is the only medication FDA approved for the treatment of acute depression in both bipolar I and II disorders. A review6 of acute bipolar depression studies of quetiapine versus placebo revealed an NNT of 6 for response, and NNHs of 6 for sedation and 19 for clinically significant (at least 7%) weight gain. Thus, the likelihood of sedation with quetiapine is similar to that of response, as the NNH and NNT for these outcomes are both 6. This means that quetiapine, although effective, is as likely to cause harm as benefit.
The olanzapine plus fluoxetine combination is FDA approved for the treatment of acute bipolar I (but not bipolar II) depression. However, for some patients, the potential therapeutic advantage of this combination may be offset by potential tolerability problems. In a review6 of acute bipolar depression studies, olanzapine plus fluoxetine yielded an NNT of 4 for response and an NNH of 6 for weight gain and 12 for sedation. Thus, the likelihood of weight gain with the olanzapine plus fluoxetine combination is comparable to that of response, as the NNH and NNT for these outcomes are 6 and 4, respectively. This means that the olanzapine plus fluoxetine combination, although effective, is comparably likely to cause harm and benefit.
Hence, both of the FDA-approved treatments for acute bipolar depression, although effective, are comparably likely to cause harm (sedation with quetiapine and clinically significant weight gain with the olanzapine plus fluoxetine combination) and benefit. Nevertheless, clinicians may consider the approved therapies for patients with more severe acute depressive episodes, as the compelling need for efficacy may mitigate potential tolerability risks.
Olanzapine monotherapy, unlike quetiapine monotherapy and the olanzapine plus fluoxetine combination, is not FDA approved for acute bipolar depression. In a review6 of acute bipolar depression studies, olanzapine monotherapy had an NNT of 12 for response and NNHs of 6 for clinically significant weight gain and 7 for sedation.6 Thus, the likelihood of weight gain with olanzapine monotherapy is approximately twice that of response, as the NNH and NNT for these outcomes are 6 and 12, respectively. This means that the olanzapine monotherapy is approximately twice as likely to cause harm as benefit.
In controlled acute bipolar I depression trials,7–9 aripiprazole monotherapy, ziprasidone monotherapy, and ziprasidone adjunctive therapy (added to lithium or valproate), although relatively well tolerated (with less sedation than with quetiapine and less weight gain than with the olanzapine plus fluoxetine combination), yielded therapeutic effects statistically similar to those seen with placebo, and thus none of these interventions received FDA approval for the treatment of acute bipolar I depression. Hence, efficacy disadvantages appeared to mitigate the potential tolerability advantages of these interventions.
Lurasidone is FDA approved to treat schizophrenia (but, to date, not for bipolar disorder). However, recent controlled trials4,5 have suggested lurasidone efficacy in acute bipolar I depression. A monotherapy study4 showed significantly higher response rates (P<0.001) with lurasidone (20–60 mg/d or 80–120 mg/d) compared with placebo, producing a dose-pooled NNT for response compared with placebo of 5. For the most reported adverse events, lurasidone had (dose-pooled) NNHs of 15 for akathisia and 17 for nausea. In addition, NNHs were 25 for sedation and -493 for clinically significant weight gain (NNH was negative as there was numerically less clinically significant weight gain with lurasidone than with placebo).
Similarly, an adjunctive (added to lithium or valproate) study5 of 20 to 120 mg/d lurasidone yielded an NNT of 7 for response and NNHs of 16 for nausea, 19 for sedation, 30 for akathisia, and -51 for clinically significant weight gain (again, NNH was negative as there was numerically less clinically significant weight gain with lurasidone than with placebo). These data suggest that lurasidone may provide an incremental benefit:risk improvement over the FDA-approved bipolar depression treatments. Specifically, like the FDA-approved treatments, lurasidone had a single-digit NNT, but unlike such treatments, lurasidone had double-digit or negative (rather than single-digit) NNHs, indicating benefit was substantially more likely than harm. Further research is needed to examine the efficacy of lurasidone treatment in both bipolar I and II depression.
Mood stabilizers. Lamotrigine has FDA approval for bipolar I maintenance treatment (with particular ability to delay depressive episodes), but not for acute bipolar depression, due to its limited efficacy in the latter. A pooled analysis6 yielded a low double-digit NNT for response compared with placebo of 12. However, in situations involving substantial side effect concerns and lacking immediate clinical urgency, this efficacy limitation may be mitigated by tolerability strengths, as NNHs were -34 for clinically significant weight gain (NNH was negative as there was numerically less clinically significant weight gain with lamotrigine than with placebo) and 42 for sedation. These results indicate that lamotrigine could be considered for patients who are particularly concerned about side effects, such as clinically significant weight gain or sedation, and are not suffering from urgent (ie, severe) acute depressive symptoms. Although lamotrigine is not FDA approved for bipolar II disorder, in a study10 of patients with rapid cycling (ie, 4 or more episodes per year), mood stability for 6 months with lamotrigine compared with placebo was more likely in patients with bipolar II disorder (NNT = 4) than in those with bipolar I disorder (NNT = 13).
Lithium, the classic mood stabilizer, has FDA approval for bipolar I maintenance treatment (with particular ability to prevent manic episodes), but not for acute bipolar depression. In a review6 of acute bipolar depression studies, lithium had an NNT for response compared with placebo of 15, which is not in the ideal single-digit range, but its tolerability was better than FDA-approved treatments, with an NNH of -112 for clinically significant weight gain (NNH was negative as there was numerically less clinically significant weight gain with lithium than with placebo) and 20 for sedation.
Carbamazepine and valproate have FDA approval for the treatment of acute mania but not for other phases of bipolar disorder, as multicenter, randomized, double-blind, placebo-controlled studies are lacking. Thus, neither of these interventions received FDA approval for the treatment of acute bipolar depression. Hence, inadequate evidence of efficacy appears to mitigate any potential tolerability advantages of these interventions over the FDA-approved acute bipolar depression treatments.
Stimulant-like agents. Armodafinil is a wakefulness promoting, low-affinity dopamine transporter inhibitor being investigated as adjunctive treatment to SGAs and/or mood stabilizers in patients with acute bipolar I depression. In a double-blind study,11 the response rate with armodafinil 150 mg/d was 46% compared with 34% for the placebo group, producing an NNT for response compared with placebo of 9. The NNH was -37 for clinically significant weight gain (NNH was negative as there was numerically less clinically significant weight gain with armodafinil than with placebo) while headache, diarrhea, and nausea rates were comparable between groups. This finding in patients with acute bipolar I depression needs to be replicated, and data are also needed for armodafinil in bipolar II depression.
Antidepressants. Antidepressants remain controversial for treating bipolar depression. Although they generally have good somatic tolerability profiles, their efficacy in bipolar depression is questionable and concerns have been raised regarding the risk of switching to mood elevation. The APA practice guidelines12 do not recommend antidepressant monotherapy, but severely ill patients may benefit from adjunctive (added to antimanic agents) antidepressant therapy. A meta-analysis13 of acute bipolar depression studies (most patients had bipolar I disorder) showed an NNT for response compared with placebo of 36, with antidepressants not being significantly more effective than placebo. However, switch rates were low, with an NNH of 209. Thus, inadequate efficacy limits the usefulness of antidepressants for acute bipolar depression, despite their generally good somatic tolerability.
Another study14 examined treatment-emergent switch from depression to mania with antidepressants versus placebo. Those with bipolar II depression experienced more manic switches with TCAs (11.2%) than with SSRIs (3.7%) or placebo (4.2%). Manic switch rates with these antidepressants in patients with unipolar major depressive disorder were substantially lower at 0.5%, 0.7%, and 0.2%, respectively. The APA guidelines12 warn that TCAs have a greater risk for causing a manic switch and do not recommend their use. Overall, clinicians need to be aware that antidepressants may not be as effective for bipolar depression as other agents are, but their somatic safety and tolerability profiles may make them appealing to patients particularly concerned about side effects.
When selecting treatments for patients with bipolar depression, clinicians need to consider both efficacy and tolerability of interventions—NNT and NNH are useful tools for such assessments. The FDA-approved bipolar depression treatments (quetiapine monotherapy and the olanzapine plus fluoxetine combination) have generally adequate efficacy but commonly have inadequate tolerability, as they entail comparable risks of harms (side effects) and benefits (therapeutic effects). Lurasidone monotherapy and adjunctive therapy could potentially represent advances compared with the FDA-approved bipolar depression treatments by yielding enhanced tolerability with similar efficacy. Antidepressants and the mood stabilizers lamotrigine and lithium generally have better tolerability than the FDA-approved bipolar depression treatments, but this potential advantage is mitigated by efficacy limitations. However, preliminary data suggest that adjunctive armodafinil may ultimately prove to represent an advance compared with these non-approved bipolar depression treatments by yielding enhanced efficacy with comparable tolerability. However, more data are needed to confirm these preliminary observations with armodafinil. In addition, for the above-mentioned bipolar depression treatments (with the sole exception of quetiapine), more research is needed to establish efficacy and tolerability in bipolar II depression. Patients with bipolar depression have varying efficacy needs (according to the degree of urgency for therapeutic effects) as well as diverse concerns regarding side effects. Clinicians who understand the efficacy and tolerability of treatments may find that using NNTs and NNHs can facilitate in helping patients select interventions, accounting for the at-times conflicting needs to maximize benefits (therapeutic effects) and minimize harms (side effects).
- Understand how to calculate the NNT and NNH of medications based on study results—lower NNTs reflect better efficacy and higher NNHs reflect better tolerability
- Evaluate the efficacy and tolerability of treatments for patients with bipolar depression
- Incorporate patient preferences into treatment planning that strives to maximize therapeutic effects and minimize side effects
aripiprazole (Abilify), armodafinil (Nuvigil and others), asenapine (Saphris), bupropion (Wellbutrin, Aplenzin, and others), carbamazepine (Carbatrol, Equetro, and others), citalopram (Celexa and others), clozapine (Clozaril, FazaClo, and others), desvenlafaxine (Pristiq), escitalopram (Lexapro and others), fluoxetine (Prozac and others), lamotrigine (Lamictal and others), lithium (Lithobid and others), lurasidone (Latuda), mirtazapine (Remeron and others), olanzapine (Zyprexa and others), olanzapine/fluoxetine combination (Symbyax and others), paroxetine (Paxil, Pexeva, and others), quetiapine (Seroquel and others), risperidone (Risperdal and others), sertraline (Zoloft and others), venlafaxine (Effexor and others), ziprasidone (Geodon and others)
APA = American Psychiatric Association, NNH = number needed to harm, NNT = number needed to treat, TCA = tricyclic antidepressant, SGA = second-generation antipsychotic, SSRI = selective serotonin reuptake inhibitor, US FDA = United States Food and Drug Administration
Take the online posttest.
- Judd LL, Akiskal HS, Schettler PJ, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry. 2003;60(3):261–269. PubMed
- Judd LL, Akiskal HS, Schettler PJ, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry. 2002;59(6):530–537. PubMed
- Goodwin FK, Jamison KR. Conceptualizing manic-depressive illness: the bipolar-unipolar distinction and the development of the manic-depressive spectrum. In: Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression. 2nd ed. New York, NY: Oxford University Press; 2007:18–19.
- Loebel A, Cucchiaro J, Silva R, et al. Lurasidone monotherapy for the treatment of bipolar I depression: results of a 6-week, double-blind, placebo-controlled study [abstract NR3-72]. In: New Research Abstract Book of the 165th Annual Meeting of the American Psychiatric Association; May 5–9, 2012; Philadelphia, PA.
- Loebel A, Cucchiaro J, Silva R, et al. Lurasidone adjunctive to lithium or valproate for the treatment of bipolar I depression: results of a 6-week, double-blind, placebo-controlled study [abstract PS-120]. In: 2012 NEI Global Psychopharmacology Congress Abstracts. CNS Spectr. 2012;17(4):235–236. Full Text
- Srivastava S, Ketter TA. Clinical relevance of treatments for acute bipolar disorder: balancing therapeutic and adverse effects. Clin Ther. 2011;33(12):B40–B48. PubMed
- Thase ME, Jonas A, Khan A , et al. Aripiprazole monotherapy in non-psychotic bipolar I depression: results of 2 randomized, placebo-controlled studies. J Clin Psychopharmacol. 2008;28(1):13–20. PubMed
- Lombardo I, Sachs G, Kolluri S, et al. Two 6-week, randomized, double-blind, placebo-controlled studies of ziprasidone in outpatients with bipolar I depression: did baseline characteristics impact trial outcome? J Clin Psychopharmacol. 2012;32(4):470–478. PubMed
- Sachs GS, Ice KS, Chappell PB, et al. Efficacy and safety of adjunctive oral ziprasidone for acute treatment of depression in patients with bipolar I disorder: a randomized, double-blind, placebo-controlled trial. J Clin Psychiatry. 2011;72(10):1413–1422. Full Text
- Calabrese JR, Suppes T, Bowden CL, et al, for the Lamictal 614 Study Group. A double-blind, placebo-controlled, prophylaxis study of lamotrigine in rapid-cycling bipolar disorder. J Clin Psychiatry. 2000;61(11):841–850. Full Text
- Calabrese JR, Frye MA, Yang R, et al. A double-blind, placebo-controlled, multicenter trial of adjunctive armodafinil in adults with major depression associated with bipolar I disorder [abstract W29]. In: American College of Neuropsychopharmacology (ACNP) 51st Annual Meeting Abstracts. Neuropsychopharmacology. 2012;38:S314–S446. Abstract
- American Psychiatric Association. Treating bipolar disorder: a quick reference guide. http://psychiatryonline.org/content.aspx?bookid=28§ionid=1663219. Published April 2002. Accessed March 21, 2013.
- Sidor MM, Macqueen GM. Antidepressants for the acute treatment of bipolar depression: a systematic review and meta-analysis. J Clin Psychiatry. 2011;72(2):156–167. Full Text
- Peet M. Induction of mania with selective serotonin re-uptake inhibitors and tricyclic antidepressants. Br J Psychiatry. 1994;164(4):549–550. PubMed
- American Psychiatric Association. Practice guideline for the treatment of patients with bipolar disorder. http://psychiatryonline.org/content.aspx?bookid=28§ionid=1669577. Published April 2002. Accessed March 21, 2013.