Use of antidepressants for depression in patients with Parkinson's disease

Bonuccelli et al. [1] conducted the first and largest study of the effectiveness and tolerability of duloxetine [serotonin–norepinephrine reuptake inhibitors (SNRIs)] for the treatment of depression in patients with Parkinson's disease (PD). That study showed that a 12-week course of duloxetine 60 mg/day was well tolerated and effective for treating depression in patients with PD and that it had no detrimental effects on the signs and symptoms of PD. The use of antidepressants for PD patients has been a common practice not only in psychiatry but also in neurology and primary practice. Hence, clinical issues in the use of antidepressants in PD patients should be more addressed.

Although the etiology of depression in patients with PD is not yet fully understood [2,3], it has consistently been proposed that alterations in neurotransmitters other than dopamine – such as serotonin, noradrenalin, glutamate and acetylcholine – are involved in the development of depression in this population [4]. According to a recent positron emission tomography (PET) study [5], PD patients with significant depressive symptoms showed significantly raised ¹¹C-DASB binding in the amygdala, hypothalamus, caudal raphe nuclei and posterior cingulate cortex compared with those without these symptoms, whereas the ¹¹C-DASB binding values in other regions were similarly decreased in depressed and non-depressed PD patients compared with healthy controls. These data are consistent with the fact that abnormal raphe and limbic areas may contribute, in part, to the pathophysiology of depression and PD as well as justify the use of serotonergic agents. Additionally, pramipexole, a dopamine D2/3 receptor agonist that preferentially affects brain activity in the prefrontal and limbic cortices according to a PET study [6], was also found to be efficacious and safe for the treatment of depression by a number of placebo-controlled clinical trials [7]. Thus, currently available PET studies support the existence of both serotonergic and dopaminergic deficits in the specific brain pathways of such patients.

Based on currently available data [8], tricyclic antidepressants (TCAs) have been the most popular agents for the treatment of depression in PD patients, with efficacy that is superior to that of newer antidepressants, such as selective serotonin reuptake inhibitors (SSRIs). However, the significant tolerability issues of TCAs compared with those of SSRIs [8] should not be neglected in clinical practice, given that PD patients usually have comorbid medical conditions and are especially vulnerable to unwanted side effects due to polypharmacy [2,3,9-11]. SNRIs have weak effects on dopamine reuptake, whereas they have simultaneous effects on serotonin and norepinephrine 5-HT reuptake transporters, theoretically resulting in faster onset and relatively higher efficacy compared with SSRIs. However, the putative superiority of SNRIs over other antidepressants has not yet been confirmed. Indeed, duloxetine and venlafaxine were found to have lower acceptability and tolerability than SSRIs [12]. Thus, clinicians treating depression in PD patients must carefully select the proper agent from among the numerous antidepressants given that these medications have differential safety and partially differential efficacy profiles [12,13].

A recent meta-analysis [13] involving 11 randomized, placebo-controlled trials (RCTs) and 3,304 subjects found no differences in terms of efficacy, even though duloxetine was worse in terms of acceptability and tolerability than some SSRIs (mainly escitalopram) and newer antidepressants (venlafaxine) for treating depression. Adverse events followed a trend toward a dose–response relationship (increasing in the order 30 vs 60 vs 120 mg and by twice daily vs once daily), whereas efficacy was not dose-related (60 vs 120 mg) in trials of duloxetine for treatment of depression. Antidepressant-associated extrapyramidal symptoms (A-EPS) may be underreported but are nonetheless clinically important, particularly in PD patients [14]. An analysis of a French database including 916 patients treated with at least one antiparkinsonian drug (199 of whom received at least one antidepressant) found that 4.5% of patients with PD treated with at least one antidepressant showed EPS. The odds ratio for EPS was 2.18 for SSRIs, 1.17 for TCAs, and 0.74 for other antidepressants, indicating that EPS may develop irrespective of antidepressant class [14]. According to the US Food and Drug Administration Adverse Event Reporting System, 89 cases met the criteria for A-EPS in Madhusoodanan et al.'s study. In that study, duloxetine was implicated in 66% of cases, sertraline in 10%, escitalopram in 7%, and bupropion in 6% of cases [15].

Response and remission rates at the endpoint were 60.4 and 45.6% in the last-observation-carried-forward analysis performed by Bonuccelli et al. [1]. However, such high response and remission rates may include an efficacy bias attributable to the inherent limitations of an open-label design. In fact, the same criteria for response and remission rates were followed by the first placebo-controlled, comparison clinical trial (12 weeks) [16] comparing an SSRI (paroxetine) with an SNRI (venlafaxine XR) for the treatment of depression in PD patients. This study found response and remission rates of 55 and 36% in the paroxetine group, 47 and 32% in the venlafaxine XR group and 38 and 28% in the placebo group, respectively. Additionally, at the endpoint, paroxetine showed a 47.6% greater reduction in depressive symptoms than did venlafaxine XR [16]. Interestingly, according to the study conducted by Bonuccelli et al. [1], the improvement in depressive symptoms by week 4 accounted for 73% of the entire improvement found at the end of the treatment, indicating that further improvement with duloxetine after 4 weeks is limited. Although this trend was similar in the venlaxine XR group, the improvement in depressive symptoms was maintained at week 8 in the paroxetine group [16]. In terms of the improvement associated with duloxetine by week 4, if earlier improvers (e.g., ≥ 25 or 50% reduction on the Hamilton Depression Rating Scale within 2 or 4 weeks) responded differently or remission rates were changed at the end of treatment, such information would be useful for early decisions regarding the strategies to use in prescribing antidepressants (e.g., switch to different antidepressant or choice of different treatment options) when such medically ill patients do not respond to the initial antidepressant after adequate duration of treatment.

Adequately powered and better-designed RCTs are clearly needed to validate the results of Bonuccelli et al.'s study. Currently, three SNRIs and their new formulations are available in the market: venlafaxine, venlafaxine extended release and desvenlafaxine succinate and duloxetine and milnacipran. Although it appears that the efficacy of different SNRIs for the treatment of depression is similar, different tolerability profiles likely exist due to the different pharmacokinetic and pharmacodynamic properties of these medications, leading to different levels of acceptability. Direct comparisons among SNRIs as well as between SNRIs and other classes of antidepressants should provide more advanced and useful clinical information regarding the treatment of depression in PD patients.

Declaration of interest

C-U Pae states no conflict of interest and has received no payment in preparation of this manuscript. This study was supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (A120004).