2.1 The Evidence Framework: NNT, NNH, and Treatment Guidelines

Modern bipolar pharmacotherapy relies on three standardized metrics to compare agents across studies. These metrics appear throughout this module, so a brief orientation is worthwhile.

The Number Needed to Treat (NNT) answers a simple question: how many patients must receive this medication for one additional patient to achieve a meaningful benefit (typically a 50% or greater reduction on a standardized depression scale, or full clinical remission) compared to placebo? A lower NNT indicates stronger efficacy. An NNT of 5 means that for every five patients treated, one achieves a benefit they would not have experienced on placebo.

The Number Needed to Harm (NNH) asks the same question for side effects: how many patients must be exposed to the medication for one additional patient to experience a specific adverse event (such as significant weight gain or severe somnolence) compared to placebo? A higher NNH indicates better tolerability. An NNH of 3 for sedation means one in every three patients will experience clinically significant drowsiness attributable to the drug.

The ratio of NNH to NNT yields the Likelihood to be Helped or Harmed (LHH). Values well above 1 indicate that the medication is much more likely to help than to cause a specific side effect (Citrome, 2014, Journal of Clinical Psychiatry).

These metrics directly inform the hierarchical recommendations of the major international treatment guidelines. The 2023 update of the CANMAT/ISBD guidelines (Keramatian et al., 2023, Focus) provides the most comprehensive current consensus, stratifying treatments into first-line, second-line, and third-line tiers based on efficacy, tolerability, and safety evidence. For acute bipolar I depression, the CANMAT/ISBD guidelines identify quetiapine, lurasidone, lithium, lamotrigine, and cariprazine as first-line monotherapy options, with lumateperone added as a newer option supported by Level 2 evidence. Lurasidone and lamotrigine are also endorsed as first-line adjunctive treatments when added to lithium or valproate. For bipolar II depression, quetiapine is the only first-line monotherapy; lumateperone is classified as second-line for BD-II given the limited clinical experience in that population. These recommendations are broadly aligned with updates from NICE and the APA (Keramatian et al., 2023, Focus).

2.2 Quetiapine: Strong Efficacy, Significant Tolerability Cost

Quetiapine is one of the most extensively studied monotherapies for acute bipolar depression. Its efficacy was established in the landmark BOLDER (BipOLar DEpRession) I and II trials, large, 8-week, randomized, double-blind, placebo-controlled studies that intentionally enrolled patients with both BD-I and BD-II depression in a 2:1 ratio reflecting real-world demographics. Patients were randomized to quetiapine 300 mg/day, 600 mg/day, or placebo (Calabrese et al., 2005, American Journal of Psychiatry; Thase et al., 2006, Journal of Clinical Psychopharmacology).

Both dosage arms demonstrated rapid, statistically significant improvements on the MADRS (Montgomery-Asberg Depression Rating Scale, a clinician-administered tool that scores depressive severity from 0 to 60, with higher scores indicating worse depression) and HAM-D (Hamilton Rating Scale for Depression, a similar clinician-rated scale), with separation from placebo evident as early as week one. Pooled data from all four BOLDER and EMBOLDEN trials showed HAM-D improvements of 14.00 points in the 300 mg group and 13.57 points in the 600 mg group, compared to 10.88 points with placebo (p < 0.001) (Young et al., 2013, International Journal of Bipolar Disorders).

To put these results in perspective, researchers use effect sizes (a standardized measure of the magnitude of difference between treatment and placebo, where values around 0.2 are considered small, 0.5 medium, and 0.8 or above large). The BOLDER studies reported strong effect sizes: in the BD-I subgroup, Cohen's d was 1.09 for the 600 mg dose and 0.91 for 300 mg, both in the large range. In the BD-II subgroup, effect sizes were 0.64 and 0.56, respectively, in the moderate range (Thase, 2008, Neuropsychiatric Disease and Treatment). These BD-I subgroup effect sizes derive from the original individual trials and likely overestimate the effect in broader populations; a pooled BD-I post-hoc analysis of the BOLDER I and II trials (N = 694) yielded more conservative effect sizes of 0.78 and 0.80 for the 300 mg and 600 mg doses, respectively (Weisler et al., 2008, Journal of Clinical Psychiatry). Network meta-analysis estimates are smaller still (MADRS change of -4.80 vs. placebo in the Kadakia 2021 NMA), reflecting the expected attenuation when pooling across more heterogeneous populations.

Quetiapine remains one of the very few agents with proven efficacy specifically in BD-II depression. A pooled post-hoc analysis of BD-II patients from the BOLDER I and II trials confirmed its antidepressant effect in this population (Weisler et al., 2008, Journal of Clinical Psychiatry). A larger pooled analysis of 776 BD-II patients from all four BOLDER and EMBOLDEN trials subsequently confirmed these findings (Young et al., 2013, International Journal of Bipolar Disorders).

The aggregate NNT for clinical response is approximately 6, and the NNT for full remission is approximately 6 (Citrome, 2014, Journal of Clinical Psychiatry). Rates of treatment-emergent mania or hypomania were low and numerically comparable to placebo (3.2% for the 600 mg dose vs. 3.9% for placebo) (Calabrese et al., 2005, American Journal of Psychiatry).

The tolerability burden. Why does quetiapine cause so much drowsiness and weight gain when other antipsychotics used for bipolar depression do not? The answer lies in which receptors it binds to beyond the ones responsible for its antidepressant effect.

Quetiapine's antidepressant action is driven in part by its active metabolite norquetiapine, a breakdown product that the body produces from quetiapine, which has its own pharmacological effects. Norquetiapine inhibits norepinephrine reuptake (blocks the recycling of norepinephrine back into nerve cells, keeping more of it available in the synapse), providing an SNRI-like effect that contributes to the antidepressant response.

The side-effect burden comes from a different set of receptors. The parent compound has potent affinity for histaminergic H1 receptors (the receptors that mediate drowsiness and appetite, the same ones that antihistamines like Benadryl block) and muscarinic M1 receptors (involved in cognition and wakefulness). These receptor interactions are what make quetiapine significantly more sedating and weight-promoting than agents like lurasidone or cariprazine, which do not bind to H1 or M1 receptors.

The NNH for sedation and somnolence is 3 to 4 (Citrome, 2014, Journal of Clinical Psychiatry). For every 3 to 4 patients treated, one will experience clinically significant somnolence attributable to the medication. The NNH for clinically relevant weight gain (7% or more increase from baseline) is 16, with an average absolute weight gain of 1.17 kg over 6 to 8 week trial periods compared to placebo (Citrome, 2014, Journal of Clinical Psychiatry).

Note on industry funding: The BOLDER and EMBOLDEN trial programs were sponsored by AstraZeneca, the manufacturer of quetiapine. This is the standard funding model for pivotal registration trials, but clinicians should be aware of the sponsorship when interpreting effect sizes and tolerability presentations.

2.3 Lurasidone: Comparable Efficacy with Metabolic Neutrality

Lurasidone was approved by the FDA in 2013 for bipolar I depression, both as monotherapy and adjunctive to lithium or valproate. It is important to note that the PREVAIL trials enrolled only BD-I patients; lurasidone's use in BD-II depression is off-label, though clinical use in this population is common. It represents a significant advance in the risk-benefit profile of bipolar depression pharmacotherapy.

The clinical bottom line for lurasidone's receptor profile is that it does not cause the sedation and weight gain that limit quetiapine, because it does not bind to the receptors responsible for those effects. The pharmacology behind this: lurasidone has potent antagonism (blocking action) at dopamine D2, serotonin 5-HT2A, and 5-HT7 receptors, and partial agonism (a dimmer-switch effect, partially activating rather than fully blocking) at 5-HT1A. The 5-HT7 receptor blockade is particularly relevant clinically because it is associated with improved cognition and circadian rhythm regulation in preclinical models. What sets lurasidone apart from quetiapine is its negligible affinity for histaminergic H1 and muscarinic M1 receptors (the receptors responsible for sedation and weight gain), which means it is structurally protected against the tolerability problems that limit quetiapine (Citrome, 2011, International Journal of Clinical Practice).

The clinical evidence comes from the PREVAIL trial program. PREVAIL 2, a Phase III double-blind, 6-week trial of 505 adults with BD-I depression, tested lurasidone monotherapy at two flexible dose ranges (20--60 mg/day and 80--120 mg/day) against placebo. Both dose ranges produced statistically significant, near-identical MADRS improvements: 15.4 points from a baseline of 30, compared to 10.7 points with placebo. Both dose ranges achieved a similar effect size (Cohen's d = 0.51), suggesting a relatively flat dose-response curve for efficacy (Loebel et al., 2014, American Journal of Psychiatry). Combined data from PREVAIL 1 and 3 confirmed significant superiority as adjunctive therapy to lithium or valproate (Citrome, 2013, Expert Opinion on Pharmacotherapy).

The NNT for clinical response is 5, directly comparable to quetiapine and OFC (Citrome, 2014, Journal of Clinical Psychiatry).

Metabolic and tolerability profile. The NNH for clinically relevant weight gain with lurasidone monotherapy is 29 for the 20--60 mg dose range. At the higher dose range (80--120 mg), weight gain was statistically indistinguishable from placebo, effectively rendering lurasidone metabolically neutral at either dose. It induces minimal changes in lipid profiles and glycemic parameters. The NNH for severe somnolence is 130, in stark contrast to quetiapine's NNH of 3 (Citrome, 2014, Journal of Clinical Psychiatry).

The limiting side effects. Lurasidone's primary tolerability challenges are akathisia and nausea. In real-world retrospective cohort studies comparing 1-year discontinuation rates, akathisia was the most commonly reported adverse event (17.7% of patients) and the leading cause of discontinuation (Kishi et al., 2024, TH Open). Lurasidone must be taken with food (at least 350 calories) to ensure adequate absorption.

Note on industry funding: The PREVAIL trial program was sponsored by Sunovion Pharmaceuticals, the manufacturer of lurasidone.

**In plain language for families: **Quetiapine is one of the most effective medications for bipolar depression, and it works for both bipolar I and bipolar II. The tradeoff is that it commonly causes significant drowsiness and some weight gain. For some patients, these side effects are manageable and worth it for the antidepressant benefit. For others, the sedation is too heavy to function well at work or school, and your doctor may choose a different medication. This is a conversation about which tradeoffs matter most to you.

Important Note: Akathisia vs. Agitation, A Critical Distinction

Akathisia (intense inner restlessness with an urge to move) is frequently misdiagnosed as worsening anxiety, emerging mixed features, or psychomotor agitation from the depression itself. This misidentification can lead clinicians to increase the dose of the offending antipsychotic, which worsens the akathisia and may increase suicide risk. When a patient on lurasidone or cariprazine reports new-onset restlessness, agitation, or an inability to sit still, screen specifically for akathisia using the Barnes Akathisia Rating Scale (BARS) before adjusting the dose upward. Management includes dose reduction, dose splitting, or adjunctive propranolol or short-term benzodiazepines.

PsychHQ

First-Line Pharmacotherapy: Quetiapine & Lurasidone