18.3.1 Acute Treatment of Mania

A recent synthesis of trials for mania found antipsychotics remain the most effective treatments, using response rate as an outcome measure, thirty-seven trials demonstrating a response rate of 49.7%, compared to seven trials showing response rate of 49.1% for lithium and eight trials indicating a response rate of 48.4% for anticonvulsants (15). Most trials are brief (around three weeks and response were broadly defined as <50% in ratings on a mania scale). This is broadly similar to a 2011 meta-analysis that measured change in scale, with haloperidol, the archetypal D₂ blocker, showing the greatest effect size (16).

18.3.2 Acute Treatment of Depression

Examination of antipsychotics licensed for acute depression indicates heterogeneity among compounds (17), most of which do not have direct effects on the dopamine system (olanzapine plus fluoxetine, lurasidone, quetiapine). A recent trial of cariprazine showed some effects on depression symptoms, using the Montgomery Asperg Depression Rating Scale (MADRS), with a least-square reduction of 2.5 versus placebo at 1.5 mg and 3 for 3 mg daily (18). It is also worth noting lack of efficacy at 3 mg in a prior trial (though difference of 4 on MADRS for 1.5 mg) (19), and contrasting these differences to those for other licensed antipsychotics, for example, quetiapine, where the mean difference in MADRS was approximately 15 for 300 mg/day (allowing for effects on sedation and weight gain, this difference is striking (20).) The findings with cariprazine can be contrasted with those of aripiprazole (see later), which may be a result of cariprazine having more agonistic properties (21).

18.3.3 Maintenance Treatment

A 2017 systematic review and meta-analysis of SGAs in maintenance treatment included fifteen RCTs, lasting from six months to two years, and one observational study lasting four years (22). This examined monotherapy and adjunctive therapy to lithium, sodium valproate, or lamotrigine. Antipsychotics included olanzapine (four trials), quetiapine (four trials), aripiprazole (three trials), risperidone (three trials), and ziprasidone (one trial). Meta-analyses demonstrated antipsychotic monotherapy superior to placebo, reducing overall relapse risk (olanzapine: RR 0.52 (95% CI 0.38–0.71), two studies; quetiapine: HR 0.37 95% CI 0.31–0.45), two studies; risperidone: RR 0.61 (95% CI 0.47–0.80), two studies). It should be acknowledged that the quality of the studies was inferior. As adjunct to mood stabilizers (lithium/valproate/lamotrigine), given to people who had responded to acute treatment, efficacy was seen for aripiprazole (RR 0.65, 95% CI 0.50–0.85; two studies), olanzapine (RR 0.49 (95% CI 0.27–0.91; one study), quetiapine (RR 0.38, 95% CI 0.32–0.46; two studies), and ziprasidone (RR 0.62, 95% CI 0.40–0.96; one study). One trial with risperidone long-acting injection (LAI) in people with bipolar 1 disorder and four or more episodes in the prior year, was not statistically significant in the meta-analysis for relapse to any mood episode, though did show benefit in a fifty-two-week follow-up compared to placebo as an adjunct to treatment as usual, with a 2.3-fold decreased risk of relapse to any mood episode (22, 23). Adjunctive quetiapine was the only drug that reduced manic (RR 0.39, 95% CI 0.30–0.52; two studies) and depressive (RR 0.38, 95% CI 0.29–0.49; two studies) episodes. All but one study had an enriched design, that is, patients were taking the drug prior to randomization. (A form of selection bias.) Two of the RCTs included people with bipolar 2 disorder. Accounting for side effects, discontinuation rates as adjunct varied from a hazard ratio of 0.66 (ziprasidone) to 0.89 (aripiprazole), with weight gain (defined as an increase of >7%) noted when meta-analyzing all antipsychotics.

Lurasidone was not examined in this review. This has an FDA license as monotherapy and adjunctive treatment to lithium and divalproex for acute treatment of bipolar depression. Following a six-week double-blind placebo-controlled RCT of lurasidone monotherapy or adjunctive treatment with lithium or divalproex, participants were randomized to an extended six-month trial of lurasidone as monotherapy or adjunct. Though not the primary outcome, treatment‐emergent mania occurred in 1.3% in the monotherapy group, and in 3.8% in the adjunctive group. Among extension study baseline responders, 10.2% met post hoc criteria for depression relapse during six months of treatment in the monotherapy group, 10.2% meeting relapse criteria in the adjunctive therapy group. The nature of the trial makes it challenging to compare depression and mania relapse to other treatments, though the low incidence of manic relapse should be noted (23, 24). A recent twenty-six-week double-blind placebo-controlled trial of asenapine maintenance therapy in 253 people with bipolar disorder found a statistically significantly longer time to recurrence of any mood episode (manic or depression), HR 0.16 for manic episode, HR 0.35 for depressive episode, though not for mixed episodes (the study may have been underpowered for this, these being post hoc analyses)(24, 25). RCTs of LAIs should be added to this literature. A randomized placebo-controlled fifty-two-week trial of aripiprazole depot showed a beneficial effect in bipolar 1 illness, with a hazard ratio of 0.45 in recurrence of any mood episode, with a signal predominantly for preventing manic episodes, mirroring the evidence for oral aripiprazole (25, 26). Similar efficacy is also seen for risperidone LAI, versus placebo, from two trials, of eighteen and twenty-four months’ duration (26–28), in relapse prevention, with a combined risk ratio of 0.42 for manic, hypomanic, or mixed symptoms, though not for depression relapse. A review summarizing three trials of LAI versus oral antipsychotics found no difference in relapse rates, though sensitivity analysis showed benefit in people with rapid cycling illness (29).

18.5.1 Presynaptic System

Dopamine synthesis capacity is measured by quantifying the uptake of the enzyme AADC, a precursor to dopamine, though this is acknowledged as not being the rate-limiting step in dopamine production (this being tyrosine hydroxylase)(55).

The endogenous production of dopamine can also be assessed indirectly, using amphetamine challenge.

As recognized for tracers such as raclopride, endogenous dopamine competes with some tracers for binding at D_2/3 receptors. At a simplified level, D₂ receptors exist in low- and high-affinity states for agonists such as dopamine, with D₂ high the functional state in the striatum. Antagonists will bind at both D₂ high and low states, and agonists will compete for D₂ high states with endogenous dopamine, are therefore more vulnerable to competition by endogenous DA, and therefore more sensitive. This led to the development of D₂ agonist radiotracers, as listed earlier. These are more sensitive to the effects of amphetamine challenge on binding potential (BP) of D₂ receptors than raclopride. For example, Narendran et al. demonstrated 1.5 times the effect of amphetamine on BP in striatum relative to raclopride, using [¹¹C] NPA (56), and Shotbolt et al. showed a similarly large change in BP with PHNO (in healthy volunteers) (57).

18.5.2 Extra-Striatal Imaging of the Dopamine System

While F-DOPA PET has a very good signal-to-noise ratio in the striatum (Figure 18.3), it is poorer at quantifying dopamine synthesis capacity in extra-striatal regions, the same test–retest study that showed good intraclass coefficients (ICC) for striatum showing poorer ICC in extra-striatal regions (75). Specific regions of poor reliability (ICC<0.5) were hippocampus, amygdala, and medial frontal gyrus. Ki^cer in the thalamus, posterior cingulate cortex, anterior orbital gyrus, and medial frontal gyrus was equal or less than that of adjacent white matter, which raises doubts regarding the validity of measuring Ki^cer in these regions. Partial volume correction for the white matter for F-DOPA PET in extra-striatal regions has been suggested, based on finding greater Ki^cer in white matter compared to gray matter, despite no evidence to suggest appreciable AADC in white matter (58).

Figure 18.3 A normative map of dopamine synthesis capacity as measured with FDOPA PET.

Given that extra-striatal regions such as the limbic and cortical dopamine systems have lower density populations of D₂ receptors, tracers such as [¹¹C] raclopride and [¹²³I] IBZM have less utility here, due to signal-to-noise ratio, with low-affinity or high nonspecific binding, respectively. Therefore high-affinity tracers such as [¹⁸F] fallypride and [¹¹C] FLB457 were developed, the former used in striatal and extra-striatal regions, the latter used only in extra-striatal regions, due to its ultra-high affinity. These tracers are used in dopamine release paradigms, and problems that limit their use in stimulant paradigms include the fact that stimulant effects on the prefrontal cortex occur outwith the dopamine synapse and therefore the process of displacement is dependent on diffusion, as well as effects of COMT in the cortex. While the stimulant challenge has shown a difference between fallypride and FLB 457 (56), task-based paradigms have failed to show a difference (59). The agonist tracer PHNO has been used in extra-striatal regions such as the substantia nigra, where D₃ receptors predominate (PHNO has a high affinity for these receptors) though it has limited utility in cortical regions (60).