Psychotropic Medication of Acute Episodes of Mood Disorders: Current Prescription Attitude in Two Psychiatric Wards in Cagliari, Italy

4.1. Old Antimanic Agents (Lithium and First-Generation Antipsychotics)

The first modern medication in mood disorders was introduced almost 70 years ago but still represents the gold standard. In fact, the Australian psychiatrist John Cade described the antimanic properties of lithium in 1949 [11]. Five years later, Mogens Schou and his colleagues implemented the first placebo-controlled double-blind trial in psychopharmacology, confirming the antimanic properties of lithium [12]. Subsequent controlled studies in the early 1970s provided evidence that lithium is also able to prevent recurrences of bipolar disorder [13]. Thereafter, there has been a great evolution in the design of clinical trials across the various phases of bipolar disorder. The efficacy of treatments for acute mania has been established, replicated and accepted widely [14]. The armamentarium has extended, as witnessed by the great variety of compounds prescribed to the patients from this study. As controlled trials often conclude for non-inferiority of a new active compound versus a comparison active medication, the therapeutic choice is left to the preferences of the individual clinicians.

With regard to acute mania, many drugs have been approved in different countries since the 1970s. For example, the US Food and Drug Administration (FDA) approved lithium in 1970 and the neuroleptic chlorpromazine in 1973. Other medications have long been used in acute mania but not necessarily approved by regulatory authorities. Paradoxically, haloperidol, which is not FDA-approved has been used as a comparison treatment in many modern trials of acute mania, as reviewed by Cipriani and colleagues [15]. Haloperidol is not promoted any more by the manufacturer due to its long stay in the market, nevertheless, it was one of the most prescribed individual medication to the patients from this study.

4.2. Clozapine

Another example of the discrepancies between the real world and randomized clinical trials is clozapine, which has been used as monotherapy or as add-on therapy in particular in the case of treatment-resistant mania, but whose effects were addressed by uncontrolled trials alone [16, 17].

Clozapine was prescribed to a small proportion of patients from this study (approximately 5%), supposedly those with treatment-resistant episodes. One reason for clozapine's limited use is its side-effect profile (including constipation, hypersalivation, weight gain), and its association with the serious adverse effect of agranulocytosis [18]. The latter is fairly uncommon and can be prevented in long-term maintenance by monthly blood-counts, mandatory in most countries, including Italy. However, it must be underscored that clozapine may have antisuicide properties, as shown at least in patients diagnosed with schizophrenia [19, 20]. According to a Swedish population-based cohort study, clozapine shows a positive benefit-risk ratio when comparing the benefit of reducing suicide and suicide attempts to the risk of agranulocytosis [21].

4.3. Anticonvulsants

Carbamazepine was prescribed to less than 10% of patients discharged from SPDC-1 and SPDC-2. Carbamazepine has been used in the treatment of various phases of bipolar disorder since the early 1980s [22] but was eventually approved by FDA in 2004 for acute mania alone. In this case, the development of an extended-release capsule formulation prompted the two randomized placebo-controlled trials required for approval [23, 24]. In Europe, the use of carbamazepine for mood disorders was launched in the 1980s but is now in decline for several reasons, including its relatively slow time to efficacy and its side-effect profile (dizziness, ataxia, diplopia, fatigue, and nausea). A slow titration is usually recommended in the treatment of acute episodes (much slower compared to valproate). The declining use of carbamazepine could also be attributed to a reduced interest by the manufacturer.

There have been some studies comparing carbamazepine to lithium in the maintenance phase of mood disorder, finding it somewhat less effective [25, 26]. In our experience, adding carbamazepine to lithium can restore prophylactic effectiveness in patients with bipolar disorder who proved refractory to lithium monotherapy [27]. In Italy, there is an example of the influence of “non-scientific” reasons for the changing attitude of prescribing anticonvulsants in mood disorders. Despite valproate’s antimanic properties had been reported since the 1960-1970s, Italian psychiatrists had to prescribe valproate “off-label” for many years or use the analog valpromide that had been approved earlier as “coadjuvant in depressive states and manic excitement, and in psychomotor agitation”. Eventually, after the introduction of a prolonged-release valproate formulation in Italy, the manufacturer applied for and obtained approval for treatment and prevention of mania associated with bipolar disorder in the 2000s. In the US, valproate semisodium had been approved in 1995 for acute mania and a dramatic change in prescription patterns followed at the expense of lithium [28-30].

The relatively small number of patients (3%) treated with lamotrigine at discharge from SPDC-1 and SPDC-2 can be attributed to the probable excess of patients with predominant manic than depressive episodes. Moreover, lamotrigine did not demonstrate efficacy in four out of five placebo-controlled clinical studies in the acute treatment of bipolar depression [31], as opposed to its prophylactic effect on depressive recurrences [32]. However, there has been a debate about this so-called “disconnect between initial judgments of lamotrigine versus its real-world effectiveness in managing bipolar disorder” [33, 34].

Focusing on the challenging problem of bipolar depression, Ketter [35] stated that “some FDA-approved treatments are helpful during acute and maintenance phases of therapy, but there is a significant unmet need for effective bipolar depression treatments with favorable side-effect profiles. Newer agents offer the promise of improvements in tolerability, but additional research is needed to actualize this promise into better treatments for patients struggling with bipolar depression” [35].

Other anticonvulsants, such as gabapentin, topiramate, oxcarbazepine, and pregabalin, were also prescribed to some patients included in this survey, even if their efficacy has not been established by controlled studies [36-38].

4.4. Second-Generation Antipsychotics

Between 2000 and 2004, there has been an “explosion” of controlled trials that resulted in FDA approval for acute mania of the second-generation antipsychotics olanzapine, risperidone, quetiapine, ziprasidone and aripiprazole [2]. Thereafter, other agents (asenapine, lurasidone, paliperidone, cariprazine, loxapine inhalation powder) have been studied and approved by FDA for either manic, mixed, or depressive phases of bipolar disorder or schizoaffective disorder. However, approval and current availability of individual agents vary across other countries.

4.5. Real World Versus Randomized Clinical Trials

As the acute response to treatment was beyond the scope of this survey, we focused on medication prescribed at discharge from the hospital. In fact, a hospital stay is one of the best opportunities to start (or reinforce) prophylaxis of mood disorders. Prophylaxis should be taken into account because mood disorders are often recurrent, affect the quality of life, and are associated with long-term increased mortality particularly due to suicide [5].

Unfortunately, controlled studies do not generally test long-term effectiveness because of the elevated costs of such trials. There are even discrepancies in the definitions of phases of mood disorders (e.g. relapses versus recurrences) and their therapy (e.g. continuation, maintenance, prophylaxis, etc.) [2]. It must also be underlined that the term “mood stabilizer” is not used by regulatory agencies.

Cipriani and colleagues [39] in their review of study designs of long-term studies for bipolar disorder underscored the problem of “enrichment design”: all randomized clinical trials selected patients who responded to acute treatment to increase the treatment effect observed in the long-term phase. By contrast, in the prescribing information sheets for all second-generation antipsychotics, the reported indication was 'maintenance treatment of bipolar disorder'. They concluded that “extrapolation of results from enrichment studies to the more general population of patients should be carried out cautiously because average treatment benefits are likely to be less in unselected patients” [39].

The results from our “real world” survey indicate that, even if the medication has been prescribed in compliance with the indications of regulatory agencies for acute episodes of mood disorders, the prescription attitude has differed significantly among hospital clinicians serving at the two psychiatric wards studied. Individual specialists make their own decision about the patients under their care.

Some choices might have been influenced by the current trend to limit hospital stays as much as possible (indeed, the mean stay averaged 8-10 days). This could be one of the reasons why benzodiazepines and more sedative first-generation or second-generation antipsychotics (e.g. promazine or quetiapine) were prescribed to most patients to control agitation, irrespective of diagnosis. Supposedly, sedative drugs were maintained at the time of discharge because the task of tapering or withdrawing them was left to the outpatient specialists once the agitation improved. With regard to benzodiazepines, it must be underscored that, in the 1980-1990s, prior to the advent of Divalproex and second-generation antipsychotics, adjunctive treatment with lorazepam or clonazepam was studied for acute mania in the US. For example, one double-blind study [40] concluded that lorazepam may offer an efficacious and safe alternative to haloperidol as an adjunctive treatment to lithium in the clinical management of the early phase of manic agitation in a subgroup of bipolar patients. However, as reviewed in a meta-analysis published in 2004, trial designs with benzodiazepines were considered heterogeneous and patient number was limited. The conclusion was that clonazepam is efficient and safe in the treatment of acute mania, but the results remain inconclusive for lorazepam [41].

Valproate may be prescribed more often than lithium thanks to its more rapid titration to achieve a quick therapeutic response [42]. In fact, for both humanitarian and economic reasons, interest in rapidly reducing symptoms of mania with pharmacologic loading strategies has increased [42].

Current biological treatment of acute agitation or aggression in the inpatient setting has recently been reviewed [42]. Among the pharmacologic agents studied in randomized controlled trials, second-generation antipsychotics have the best evidence to support efficacy [43].

4.6. Combined Medication

The extensive prescription of combined regimens is not unexpected. In fact, it has been shown that combined antipsychotic/mood stabilizer have efficacy advantages over antipsychotic or mood stabilizer monotherapy at least in acute mania, and should be considered as first-line therapy [44, 45]. However, the lower tolerability of combinations should also be taken into account.

In a systematic review, Fornaro and colleagues [46] reported that regardless of type or current mood episode polarity of bipolar disorder, prevalence rates up to 85% and 36% were found using the most permissive (two or more medications at once) and the most conservative (four or more) operative definitions for polypharmacy, respectively. Moreover, lithium prescription rates ranged from 13% to 33% in patients receiving polypharmacy according to conservative and permissive definitions, possibly suggesting a reduced need for augmentation of combination strategies for those cases with a favorable lifetime lithium response and/or long-lasting treatment as well as less likelihood of lithium response over the time most severe cases possibly exposed to a more complex polypharmacy overall.

4.7. Current Positions

With regard to individual agents, Bauer and Gitlin [47] commented that although first-generation antipsychotics are also effective, unquestionably, the second-generation antipsychotics now dominate the pharmacotherapy of acute mania. This reflects both their clear efficacy as shown in studies and meta-analysis, their ease of administration, relatively more benign side effect profiles (certainly compared to first-generation antipsychotics), and their rapidity of response compared to lithium.

However, the ideal mood stabilizer has to be effective against both mania and depression, without any risk of increasing the incidence of opposite episodes. Unfortunately, some antipsychotics can increase the risk of depression and antidepressants can trigger mania if used as monotherapy [48]. This problem is sometimes overlooked in the context of acute episodes, because the aim may be the short-term response.

4.8. Differences Between SPDC-1 and SPDC-2

The different lithium prescribing attitude on discharge deserves a particular discussion. In some patients, prior nonresponse or side effects may have influenced the confirmation or the discontinuation of lithium therapy: However, this cannot be ruled out from our data. It should be noted that abandoning lithium prophylaxis may be associated with an increase of suicide in high-risk patients, even in the event of an incomplete mood stabilization [49, 50].

No similar evidence is available regarding currently prescribed alternatives to lithium, including carbamazepine [51], valproate [30], and gabapentin [52]. Moreover, even a systematic review of data from randomized trials, that necessarily focused on shorter-term treatment periods, confirmed that lithium is effective in the prevention of suicide, deliberate self-harm, and death from all causes in patients with mood disorders, compared to placebo and other active treatments [53, 54].

Coryell reviewed early (1973-1976) and recent (2000-2003) lithium versus placebo trials, as well as those assigned to alternative mood stabilizers versus placebo and concluded that differing success rates in early and more recent maintenance trials strongly suggest a cohort effect in which the lithium responders have been relatively unavailable for being recruited into recent maintenance trials [6].

In an evaluation of real-world long-term outcome, lithium has recently been found as the most effective mood-stabilizer in a study of the comparative effectiveness of pharmacologic treatments in the prevention of psychiatric re-hospitalization in a Finnish nationwide cohort of 18,018 patients with bipolar disorder followed up for a mean of 7.2 years [55].

Similar conclusions were reported by a recent systematic review of non-randomized controlled observational studies [56]. Maintenance lithium monotherapy was associated with improved outcome compared with another mood stabilizer in monotherapy, including valproate, lamotrigine, olanzapine, quetiapine, unspecified anticonvulsants, carbamazepine/lamotrigine, unspecified atypical antipsychotics and unspecified antipsychotics. Lithium in combination with other mood stabilizers was found to be superior to lithium monotherapy in some analyses, but not in others [56].

The significant differences observed between SPDC-1 and SPDC-2 warrant some interpretation. Both wards are located in the same hospital and derive from the division of one single SPDC occurred in 2008. There was a tendency to keep the original staff of the old SPDC in SPDC-1, while the SPDC-2 personnel was new. Indeed, the mean age of the psychiatrists serving in 2016 differed between SPDC-1 and SPDC-2 (59.2 vs 50.8; p = 0 .017). We can speculate that the different professional experience has influenced both the diagnostic and the prescription attitude: perhaps SPDC-1 clinicians have been prone to diagnose patients with classical nosographic entities compared to their younger SPDC-2 colleagues who have often concluded for unspecified or for schizoaffective disorders. Similarly, the older SPDC-1 clinicians have preferred first-generation antipsychotics and the younger SPDC-2 clinicians have preferred second-generation antipsychotics. We can hypothesize that experienced clinicians keep their long-lasting habits and are less prone to novelty.

4.9. Tolerability

Another aspect that could be overlooked in the choice of therapy for acute episodes is the tolerability of psychotropic drugs. Cipriani and colleagues [15] underscored that evidence comparing drugs is poor, but clinicians and patients should consider different side effect profiles as an important issue to inform their choice. For example, their Cochrane review comparing haloperidol with other active treatments concluded that haloperidol is associated with less weight gain than olanzapine, but with a higher incidence of tremor and other movement disorders [15].

Efficacy advantages of co-treatment with antipsychotics and mood-stabilizers versus monotherapy should be balanced with its greater adverse event burden, including weight gain-related, extrapyramidal, tremor, sedation/somnolence, vomiting, and glucose/lipid-related adverse events [57]. With regard to the maintenance phase, lithium appears the least likely to cause substantial weight gain among mood stabilizers studied in randomized trials [6].

Aripiprazole can induce gastrointestinal disturbances (nausea and constipation) and movement disorders (in particular akathisia). Comparative trials with medicines other than haloperidol and lithium are few, so the precise position of aripiprazole in therapy remains unclear [58].

Aripiprazole appears to have a superior tolerability profile when used as maintenance treatment. Side effects like a headache, insomnia, and extrapyramidal side effects, such as tremor and akathisia may be treatment limiting in some cases [59]. Aripiprazole was also associated with higher levels of high-density lipoprotein, but no difference in extrapyramidal symptoms in the maintenance phase versus placebo or in comparison with other medication such as haloperidol or lithium [60].

Besides tolerability of the acute and the immediately following phase, that have been addressed by controlled trials, the choice of clinicians should consider potential adverse events associated with longer-term treatment in the real world. For example, the aforementioned recent Finnish nationwide cohort study examined hospitalizations owing to physical illness (besides psychiatric re-hospitalizations) for 7.2 years [55]. In the comparison between the three most frequently used medication, lithium was associated with the lowest incidence of subsequent hospitalization owing to physical illness, followed by valproate and quetiapine [55].

It must be underscored that the 7.2-year follow-up could not reveal any adverse event associated with longer-term exposure, including the decline in renal function observed in patients treated with lithium for decades [61, 62].

Acute treatment with valproate has commonly been associated with nausea, vomiting, diarrhea, sedation, and tremor. Long-term treatment can also induce alopecia and polycystic ovarian disease. Hepatic and pancreatic toxicity have been reported, as well as an unusual encephalopathy associated with hyperammonemia. Dealberto reviewed fourteen cases of encephalopathy in the psychiatric setting published between 1995 and 2005: Ammonia concentrations varied across the reported cases from normal to ten-fold the upper limit of the normal range [63]. Raja and Azzoni found a correlation between valproate concentration and ammonemia in patients hospitalized in a psychiatric ward [64]. We confirmed that valproate can increase ammonemia [65]. Median concentrations were significantly higher in a group of 42 outpatients (20 µmol/L) treated with valproate for a median of 48 months compared with the group of 44 outpatients treated with other mood stabilizers (12.5 µmol/L). However, clinically relevant concentrations were rare [65].