Pramipexole Augmentation for Treatment-Resistant Unipolar and Bipolar Depression in the Real World: A Systematic Review and Meta-Analysis

Background: Pramipexole is a dopamine full agonist approved for the treatment of Parkinson’s disease and restless legs syndrome. Its high affinity for the D3 receptor and neuroprotective, antioxidant, and anti-inflammatory activity provides a rationale for the treatment of depression. In this paper, we review studies on the effectiveness and safety of antidepressant pramipexole augmentation in treatment-resistant depression. Methods: This comprehensive systematic review and meta-analysis of observational studies on pramipexole–antidepressant augmentation included patients with resistant unipolar and bipolar depression. The primary outcome measure was the treatment response, measured at the study endpoint. Results: We identified 8 studies including 281 patients overall, 57% women and 39.5% with bipolar disorder and 60.5% with major depressive disorder. The mean follow-up duration was 27.3 weeks (range 8–69). The pooled estimate of treatment response was 62.5%, without significant differences between unipolar and bipolar depression. Safety was good, with nausea and somnolence the most frequent side effects. Conclusions: The findings of this systematic review, needing further confirmation, show that off-label use of pramipexole as augmentation of antidepressant treatment could be a useful and safe strategy for unipolar and bipolar treatment-resistant depression.


Introduction
Major depressive disorder (MDD) and bipolar disorder (BD) are the most frequent psychiatric disorders, affecting about 20% of the general population in their lifetime [1]. They are associated with social withdrawal, functional and vocational impairment, medical morbidity, and premature death, including elevated suicide risk [2][3][4][5][6]. Major depressive episodes (MDEs) are the most common presentation of BD [7,8].
The antidepressant (AD) treatment of MDE comprises selective serotonin reuptake inhibitors (SSRIs), serotonin and noradrenaline reuptake inhibitors (SNRIs), monoamine oxidase inhibitors (MOAIs), tricyclic antidepressants (TCAs), and "other" antidepressants (mirtazapine, vortioxetine, amisulpride, trazodone, etc.). ADs are employed in monotherapy for unipolar depression [3,9,10] and combined with mood stabilizers or second-generation antipsychotics for bipolar depression without mixed features, rapid-cycling course, and treatment-emergent manic/hypomanic switch [3,6,11,12]. Still, a substantial subgroup of patients fails to respond to ADs. The results from the STAR*D study indicate that one in three patients with unipolar depression did not achieve symptomatic remission after several antidepressant trials [13,14], and in the STEP-BD study, patients with bipolar depression did not benefit from antidepressant treatment combined with mood stabilizers [15]. activity of pramipexole, showing that its subacute administration (12)(13)(14)(15) days at a peak daily dose of 1.0 mg) in healthy volunteers modifies neural responses to emotional information similarly to that of traditional antidepressants [56].
This drug is bound to plasma protein to a very low (<20%) extent, does not inhibit CYP isoenzymes 1A2, 2C9, 2C19, 2E1, 3A4, or 2D6, is not appreciably metabolized by CYP isoenzymes, and is secreted by the renal tubules (90%). The information leaflet (https: //www.accessdata.fda.gov/drugsatfda_docs/label/2018/020667s036lbl.pdf (accessed on 31 December 2022) does not report any potential interaction of pramipexole with lithium, valproate, and carbamazepine and points out a possible additive sedative effect with antidepressants, antipsychotics, and benzodiazepines. Furthermore, it does not report a risk of metabolic changes or QTc interval prolongation.
In clinical practice, some studies have evaluated the efficacy and safety of pramipexole as an antidepressant in patients with Parkinson's disease and patients with mood disorders.
Regarding the patients with Parkinson's disease, a meta-analysis of 18 randomized controlled trials showed that the improvement of depressive symptoms in the pramipexole treatment group was significantly higher than in the control group without differences in the side effects rate [57].
Regarding the patients with mood disorders, a meta-analysis including 13 studies published up to December 2018 (5 randomized clinical trials and 8 observational studies) and 504 patients (362 with unipolar depression and 142 with bipolar depression) [58] estimated that pramipexole treatment had a 52.2% response rate in the short term and a 62.1% response rate in the long term and a 36.1% remission rate in the short term and a 39.6% remission rate in the long term. In randomized clinical trials (RCTs), the response rate to pramipexole was superior to that of a placebo (mostly in the bipolar depression subgroup) and similar to that of SSRIs. Acceptability and tolerability were good, with nausea being the most frequent side effect.
Although encouraging about the antidepressant properties of pramipexole, this evidence does not allow us to evaluate the effectiveness of pramipexole as AD augmentation in patients with TRD in the real world because of the heterogeneity of the studies included in the review. In fact, 2 of the 13 studies did not include patients with TRD and 5 of the 13 studies evaluated the antidepressant property of pramipexole as monotherapy and not as augmentation of traditional AD.

Aims of the Study
This study's aim is to conduct a systematic review and meta-analysis of observational studies on the effectiveness and safety of pramipexole AD augmentation in patients with unipolar and bipolar TRD.

Literature Search
The PRISMA method [59] was followed in the literature search. Specifically, PubMed/MEDLINE, Cochrane Central Register of Controlled Trials (CEN-TRAL), and Embase databases were searched to identify peer-reviewed articles on the effectiveness and safety of pramipexole for major depressive episodes in unipolar and bipolar depression. The search string used the following terms: mood disorders, depression, affective symptoms, affective disorder, mood disorder, bipolar, mania, manic, hypomania, pramipexole, and dopamine agonists. We searched for ongoing and unpublished studies via Internet searches on ClinicalTrials.gov (www.clinicaltrials.gov (accessed on 31 December 2022)) and on the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (apps.who.int/trialsearch/ (accessed on 31 December 2022).
No beginning date and no language restrictions were applied, and the last publication date was 31 December 2022.

Inclusion and Exclusion Criteria
Retrospective or prospective observational studies and case series using pramipexole as an AD augmentation strategy for treatment-resistant unipolar and bipolar MDE were included.
Randomized clinical trials, observational studies using pramipexole as monotherapy, and observational studies including patients without TRD were excluded.

Population
Male and female patients aged ≥18 years with a primary diagnosis of MDD or BD were included. Studies including patients with other comorbid psychiatric disorders, suicidal thoughts, or serious medical illnesses were not excluded.

Outcome Measures
The primary effectiveness outcome measure was the treatment response, defined as a ≥50% reduction from the baseline of the Montgomery-Asberg Depression Rating Scale (MADRS) [60], Hamilton Depression Rating Scale (HDRS) [61], or Clinical Global Impression (severity) [62].
The secondary effectiveness outcome measure was remission, defined as reaching a sub-threshold score on the depression scale used in the specific study (for instance, a score of HDRS-17 ≤ 7) at the endpoint.
If neither of these three scales were used, we considered other clinician-rated or self-report scales.

Safety Measures
The safety measures were: (1) tolerability, i.e., the number of patients who discontinued the study due to side effects and the number and types of side effects (2) acceptability, i.e., the number of patients who discontinued treatment for any reason; (3) (hypo)mania onset; and (4) suicide attempt.

Data Extraction
Two researchers (S.B. and R.d.F.) read each article and evaluated the completeness of the data extraction independently. A structured data retrieval form was designed to ensure consistency of appraisal for each study. The data included study characteristics (such as lead author, publication year, and journal), participant characteristics (age range, setting, and diagnosis), intervention details (dose range and mean dose of study drugs), and outcomes of interest. The data were extracted from the manuscript and the tables, while PlotDigitizer 2.6. [63] was used to extract the data from the figures. Information on primary and secondary outcomes was extracted by the same two researchers, and disagreements were resolved in a consensus meeting with a third researcher (AT).

Statistical Analysis
A random-effects meta-analysis was performed to obtain the overall pool estimates with a 95% confidence interval (95% CI) of primary, secondary, and safety outcomes; the pool estimates were also estimated and compared between bipolar and unipolar depression.
Between-study heterogeneity was tested by Cochran's Q test and measured with the I2 statistic, which represents the percentage of variance in the estimated effects due to heterogeneity rather than chance and ranges from 0 to 100. An I 2 statistic > 50% was considered indicative of significant heterogeneity.
Forest plots were used to graphically depict the estimates with 95% CI for individual studies and pooled results. All statistical analyses were performed with R version 4.2.0 using the meta and metafor packages.

Selected Studies
The literature search generated 881 records and another 309 records were identified from trial registers. After duplicate removal through EndNote, 1163 titles and abstracts were initially assessed; 1136 articles were excluded from the title and abstract, and 27 articles were retrieved in full text. Nineteen studies were excluded for the following reasons, three were RCTs including patients without TRD [64][65][66], three were RCTs [67][68][69], two included patients without TRD [70,71], three were case reports [72][73][74], two were letters to the editor [54,75], two were trials conducted on individuals with bipolar disorder in the euthymic phase [76,77], three were reviews [31,58,78], and one was a clinical trial on individuals with bipolar disorder treated with quetiapine extended-release and pramipexole, but the results were not disclosed [79] (see flow chart, Figure 1). heterogeneity rather than chance and ranges from 0 to 100. An I 2 statistic > 50% was con sidered indicative of significant heterogeneity.
Forest plots were used to graphically depict the estimates with 95% CI for individua studies and pooled results. All statistical analyses were performed with R version 4.2. using the meta and metafor packages.

Selected Studies
The literature search generated 881 records and another 309 records were identifie from trial registers. After duplicate removal through EndNote, 1163 titles and abstract were initially assessed; 1136 articles were excluded from the title and abstract, and 27 ar ticles were retrieved in full text. Nineteen studies were excluded for the following reasons three were RCTs including patients without TRD [64][65][66], three were RCTs [67][68][69], tw included patients without TRD [70][71], three were case reports [72][73][74], two were letters t the editor [54,75], two were trials conducted on individuals with bipolar disorder in th euthymic phase [76,77], three were reviews [31,58,78], and one was a clinical trial on ind viduals with bipolar disorder treated with quetiapine extended-release and pramipexole but the results were not disclosed [79] (see flow chart, Figure 1). Finally, 8 studies (281 participants) were included in the present systematic review [49,[80][81][82][83][84][85][86]. One study considered only unipolar patients [81], one study considered onl bipolar patients [83], while one study did not differentiate the results with respect to d agnosis [80].

Study Characteristics
The main study characteristics are listed in Table 1.  Three studies were prospective observational studies, two were open-label trials, two were chart reviews, and one was a case series. Six studies used the DSM-IV or the subsequent editions for diagnosis [49,[80][81][82]84,86], one used DSM-III-R [83], and one did not report which diagnostic criteria were used [85]. One hundred and seventy patients (60.5%) had a diagnosis of MDD and one hundred and eleven (39.5%) had a diagnosis of BD.
The mean study sample size was 35 (range 10-116), 57% were women, the mean age was 49.6 years, and the mean age at onset was 33.6 years (76 did not report this information).
Although the pooled proportion of remission in unipolar depression was higher than that in bipolar depression (60.8% vs. 39.4%), no significant difference was found between the two groups (p = 0.200).

Safety Measure
a. Drop-outs due to any adverse event The estimated pooled proportion of dropouts due to adverse events was 11.4% (95% CI 5.0-17.8%), based on seven studies reporting this information (271 participants) (Supplementary Table S1).

b. Drop-outs due to any reason
Overall, when pooling the seven studies with data on drop-outs due to any reason (271 participants), the proportion of dropouts was 29.7% (95% CI 16.2-43.3%). For the bipolar group, the pooled proportion of dropouts was 12.8%, while in the unipolar group, it was 20.4%; however, the difference was not significant (p = 0.424) (Supplementary Table  S2).

a.
Drop-outs due to any adverse event The estimated pooled proportion of dropouts due to adverse events was 11.4% (95% CI 5.0-17.8%), based on seven studies reporting this information (271 participants) (Supplementary Table S1).

b.
Drop-outs due to any reason Overall, when pooling the seven studies with data on drop-outs due to any reason (271 participants), the proportion of dropouts was 29.7% (95% CI 16.2-43.3%). For the bipolar group, the pooled proportion of dropouts was 12.8%, while in the unipolar group, it was 20.4%; however, the difference was not significant (p = 0.424) (Supplementary Table S2). (Hypo)mania onset (Hypo)mania onset was uncommon in both groups (no difference between the groups, p = 0.098); indeed, the global pooled proportion of (hypo)mania was 1.1% (95% CI 0-2.0%) (Supplementary Table S3).

d. Suicide attempt
No patients attempted suicide.

Discussion
To the best of our knowledge, this is the first systematic review of observational studies to evaluate the effectiveness and safety of pramipexole as AD augmentation on treatment-resistant unipolar and bipolar depression. We focused on these studies for multiple reasons. Observational studies include patients with psychiatric and physical comorbid disorders and with suicidal thoughts, often excluded from RCT, providing evidence on response/remission rates and adverse events associated with treatment in the real world [87]. Furthermore, effective dosage and titration of pramipexole as AD augmentation are not yet clearly defined and the flexible prescription in observational studies, based on the clinical judgement of the prescriber according to the tolerability and therapeutic efficacy, provide valuable information on the use of the drug that maximizes tolerability and effectiveness [85,86]. Notably, the present study includes 7 of the 13 studies included in a previous review [58] and a study published later [86]. Five studies of previous revision were excluded because of RCTs and one because it was conducted on patients with non-treatment-resistant depression.
This systematic review and meta-analysis suggest that the augmentation of traditional antidepressants with pramipexole could be an effective strategy for TRD.
The pooled estimates of overall response (62.5%) and remission (48.1%) with pramipexole augmentation are close to the higher boundary of the range of response and remission rates reported in the literature for aripiprazole augmentation (18.5% to 60% and 7.4% to 54%, respectively) [30,88], the strategy for TRD with the strongest evidence of efficacy [31,89,90] and extensively used in clinical practice [33].
Furthermore, response and remission rates for pramipexole are quite higher than those reported for the treatment of unipolar non-resistant depression with ADs (52.9% and 32.6%, respectively) [91].
Regarding the pre-planned subgroup analyses, the response rate was quite similar for individuals with unipolar and bipolar depression (66.0% and 56.7%, respectively), while the remission rate was higher for unipolar than for bipolar individuals (60.8% and 39.4%, respectively), although the difference failed to reach the statistical significance because of the small sample size. This result did not confirm the finding of a previous review [90] showing a significantly higher response rate to pramipexole than to a placebo in bipolar but not in unipolar treatment-resistant and non-resistant depression (RCT studies). Considering the limited information on this topic, evidence supporting the use of pramipexole augmentation for bipolar TRD as well as for unipolar TRD could be useful for clinicians engaged in the routine clinical management of bipolar patients with TRD. This result should be interpreted with caution because ADs use in patients with bipolar depression is controversial and the definition of bipolar TRD is not clearly supported by the data.
The topic needs further investigation in larger samples.
Notably, one study included in the present review [86] compared response and remission rates in patients with (n = 32) and without (n = 84) chronic cerebrovascular comorbid disease showing no statistical differences between the two groups. This evidence, needing further confirmation, could be clinically relevant because patients with depression and chronic cerebrovascular diseases are often poor respondents to ADs [92,93], and data on the effectiveness of lithium or aripiprazole AD augmentation for patients with TRD comorbid with chronic cerebrovascular diseases are very limited [94,95].
Our findings indicate that pramipexole augmentation is a safe strategy for TRD with a drop-out rate due to side effects of 11.4% and due to any reason of 29.7%. Furthermore, during the trials, no patients attempted suicide and only a very low number of patients developed (hypo)mania.
Although the profile of side effects was similar to that usually reported for patients with Parkinson's disease (https://www.drugs.com/sfx/pramipexole-side-effects.html (accessed on 31 December 2022), the frequency of these effects was lower. In patients who did not drop out, side effects often resolved spontaneously or after reducing the pramipexole dose [85,86]. Although the information leaflet reports a potential interaction of pramipexole with antidepressants, antipsychotics, and benzodiazepines (sedative effect), no study included in the present review reports drop-out or side effects due to this interaction.
Transient visual hallucination and confusion occurred respectively in 2 (6.2%) and 1 (3.1%) of the 32 old patients with chronic cerebrovascular comorbid disease included in Tundo et al.'s study [86]. Although the rates of these side effects were lower than that reported in patients with Parkinson's disease (17% and 10%, respectively) (https://www. drugs.com/sfx/pramipexole-side-effects.html (accessed on 31 December 2022), the authors suggest prescribing pramipexole with caution in elderly patients with comorbid chronic cerebrovascular conditions and careful monitoring of delirium or psychotic symptoms.
Notably, the studies included in this systematic review report a very low rate of hypersexuality and no cases of others impulse dyscontrol (gambling and compulsive shopping), which are common in patients with Parkinson's disease receiving pramipexole [96]. This result is in line with those of a previous meta-analysis on pramipexole including observational and RCT studies in which pramipexole is used as monotherapy or as AD augmentation [58]. The different susceptibility to impulse dyscontrol between patients with mood disorders and with Parkinson's disease might be due to different neurobiological dysfunctions underlying these two diseases [97,98], although we cannot exclude that the relatively small sample size of patients with TRD treated with pramipexole precludes the possibility to detect these adverse events.
Overall, the safety of pramipexole augmentation is comparable to that reported in the literature for traditional antidepressants employed for the treatment of unipolar nonresistant depression and for aripiprazole augmentation, the strategy for TRD with the strongest evidence of efficacy.
In fact, the tolerability and acceptability of pramipexole augmentation were quite similar to that of traditional antidepressants for the treatment of unipolar non-resistant depression (drop-out rate due to any type of side effects 11.4% and 10.4%, respectively; drop-out rate for any reason 29.7% and 26.4%, respectively) [91] and tolerability was slightly better than that reported for aripiprazole augmentation (drop-out rate due to any side effect 11.4% and 23.7%, respectively) [99]. We found only six studies reporting the acceptability of aripiprazole augmentation [89] showing that the drop-out rate due to any reasons for this augmentation is lower than that of pramipexole augmentation (14% and 29.7%, respectively), but the difference in follow-up duration, shorter for aripiprazole than for pramipexole (6 and 27.3 weeks median duration, respectively) could explain the difference in acceptability.
The main limitation of the present review is the restricted number of studies and participants. This limitation may have contributed to the lack of statistical significance in subgroup comparisons. Other limitations of the study were the high variability of dosing and the time points at which outcomes were collected, and the lack of a critical appraisal of the articles.
Furthermore, significant heterogeneity between the studies was observed that could not be controlled using covariates in a meta-regression because of the limited information available in the studies.

Conclusions
In conclusion, evidence from real-world studies suggests that adding pramipexole to traditional ADs could be an effective treatment option for patients with TRD. Pramipexole augmentation appeared to be safe, with nausea and somnolence being the most frequent side effects.
Should the findings of the current review be confirmed by high-quality and welldesigned RCTs, clinicians would have a further augmentation strategy for TRD with a different side effects profile from the current strategy with the best evidence of effectiveness (SGA augmentation), without QTc prolongation, metabolic, and akathisia risk. Thus, clinicians may choose the most suitable treatment for each patient according to his/her medical condition and preference.
We reiterate that the use of pramipexole as AD augmentation is off-label and its prescription is currently allowed only to centers specialized in TRD [9] and carefully selected patients without current or previous episodes with psychotic symptoms because pro-dopaminergic agents could induce psychosis in vulnerable individuals [100].
Pramipexole use requires special precautions. The dosage must be gradually increased to reduce the risk of side effects, mostly nausea, and gradually escalated to avoid the risk of dopamine withdrawal syndrome. Patients and their family members must be instructed to report the occurrence of any potentially dangerous side effects such as lethargy, hypersexuality, gambling, compulsive shopping, and, mostly in old patients with chronic cerebrovascular comorbid diseases, psychotic symptoms or delirium. Although the preliminary data suggest the safety of pramipexole in the treatment of bipolar depression, evidence from further studies is needed to confirm the low risk of mixed features, (hypo)manic switch, or rapid cycling course development in patients with bipolar depression.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/life13041043/s1, Table S1: Stratified and total pooled randomeffect estimates of the proportion of drop-outs due to any adverse event and stratification by diagnosis. % W is the percentage weight of each study; Table S2: Stratified and total pooled random-effect estimates of the proportion of drop-outs due to any reasons and stratification by diagnosis; Table S3: Stratified and total pooled random-effect estimates of the proportion of (hypo)mania switch and stratification by diagnosis. Funding: This study was funded by Fondazione dell'Istituto di Psicopatologia Onlus, Rome, Italy, grant number 01/2021. The funding source had no role in the study design, in the collection, analysis, or interpretation of the data, the preparation of the manuscript, or the decision to submit the paper for publication.