Fluoxetine does not impair motor function in patients with Parkinson's disease: correlation between mood and motor functions with plasma concentrations of fluoxetine/norfluoxetine Fluoksetin ne remeti motornu funkciju kod bolesnika sa Parkinsonovom boleš u: korelacija raspoloženja i motorne funkci

Background/Aim. Selective serotonin reuptake inhibitors are the most commonly chosen antidepressants in patients with Parkinson's disease (PD). The aim of our study was to assess the influence of fluoxetine (Flu) on motor functions in patients with PD. Methods. In this prospective, controlled, open-label study, 18 patients with PD and mild depression [(10 Hamilton Rating Scale for Depression (HDRS) 23)] without dementia [(25 Mini-Mental State Examination (MMSE)] were treated with Flu. Both single and repeated dose effects of Flu were assessed on days 1–80. Plasma concentrations of Flu and norfluoxetine (NORFlu) were correlated with the results of selected motor function performance scores: The Unified Parkinsons Disease Rating Score (UPDRS), Finger Tapping Test (FTT) and Purdue Pegboard Test (PPT). Severity of PD, depression and dementia were evaluated using standard tests [(Hoehn and Yahr stages (HY), activity of daily living (ADL), UPDRS, HDRS, MMSE)]. Results. Steady-state for Flu/NORFlu was reached after 18 days of treatment. Such a plateau correlated with significant improvements in both scores of depression and Parkinson's disability (HDRS, UPDRS and ADL, respectively). In addition, FTT and PPT scores also increased until day 18, with further slight fluctuations around the plateau. Optimal motor performances correlated with Flu concentrations of approximately 60–110 g/L. Conclusion. Flu (20 mg/day) significantly reduced depression in PD patients while it did not impair their motor performances. Because substantial placebo effects may arise in studies of PD and depression, large, prospective, randomized, placebo-controlled clinical trials are warranted.


Introduction
Depression is the most common and frequently disabling psychiatric condition in patients with Parkinson's disease (PD). Prevalence of depression in patients with PD varies from 7% to 76% depending on the assessment method 1 . Such a depression is mostly persistent or recurrent. It may be accompanied with anxiety, cognitive impairment and may reduce effectiveness of antiparkinson's therapy [2][3][4][5] . Depression increases PD patients' disability and significantly reduces their quality of life. Consequently, approximately 50% of patients with PD receive antidepressant therapy [4][5][6][7] .
Optimal treatment for depression in PD patients has not been established. Several antidepressants were tested in randomized clinical trials without sufficient statistical power (e.g. citalopram, sertraline, fluoxetine, amitriptyline and nortriptyline). Amitriptyline seems to be more effective than fluoxetine in PD patients with severe depression. However, it is not necessarily the first choice for treatment of depression in PD patients, according to the recommendations of the American Academy of Neurology 8 . In addition, the adverse effects of amitriptyline such as orthostatic hypotension, sedation, cognitive and anticholinergic effects might preclude its use and increase the dropout rate in parkinsonians 1,9,10 .
On the other hand, selective serotonin reuptake inhibitors (SSRIs) are used as a first line treatment of depression 51% of the time 1,9,10 . In postmortem studies of patients with PD depletion of 5-hydroxytryptamine  in the caudate as well as hypothalamus and frontal cortex was reported [11][12][13][14] , with preferential loss of 5-HT in the caudate compared with the putamen, but with relatively less loss of 5-HT (66%) than dopamine (98%) 15 . Imaging studies in vivo have also suggested depletion of 5-HT innervation to the striatum as measured via decreased 5-HT transporter binding [16][17][18] . The loss of striatal 5-HT in PD may be secondary to neurodegeneration within the raphe nuclei as Lewy bodies are seen in the raphe nuclei 19,20 , associated with cell loss 21,22 . Tauscher et al. 23 , 1999, were the first to demonstrate the pharmacodynamic action of the selective 5-HT transporter blocker fluoxetine in the human brain in vivo. Meyer et al. 24 , 2004, showed that 80% 5-HT transporter occupancy was achievable with SSRI at therapeutic doses in a study on patients with mood and anxiety disorders. Apart from these drugeffects studies, it has been shown that recovery of central serotonergic system after SSRI therapy was associated with reduction of clinical symptoms in 18 depressive subjects using [ 123| ]-CIT and SPECT 25 . All these findings of SSRIs-5-HT transporter occupancy in PET/SPECT studies clearly reflect the pharmacologically induced changes in serotonergic transmission 5,26 .
However, data on the efficacy and safety of SSRIs in PD are still lacking and sufficiently large scale randomised controlled trials are required. Although the introduction of SSRIs offers new opportunities for the treatment of depression in PD, these agents could produce extrapyramidal adverse reactions aggravating parkinsonism 1,10 . While epidemiological studies have not suggested increased risk of worsening PD using SSRIs for depression 27 , almost one hundred detailed reports on extrapyramidal adverse effects linked to SSRIs antidepressants have been published 28,29 .
The influence of Flu on motor performances in PD patients still remains to be clarified. Extrapiramidal side effects of Flu seem to be related to the exacerbation of Parkinson's disability 30 . However, it was also reported that Flu did not increase Parkinson's disability either in retrospective 31 or in prospective studies 32 . Therefore, the authors argue for more systemic and controlled research examining the treatment of depression in patients with PD 1,33,34 .
The aim of this study was to determine motor performances of PD patients treated with antidepressant Flu and to assess a possible correlation between mood and motor performance scores with plasma concentrations of Flu and its active metabolite, norfluoxetine (NORFlu).

Methods
Efficacy and tolerability of Flu was assessed in the prospective, 80 25)]. These 18 patients were either de novo PD patients (PD 0 group, N = 9), or PD patients who were on the stable antiparkinsonian treatment (PD t group, N = 9), without selegiline, rasagiline and/or dopamine agonists for at least two months prior to Flu.
Patients with secondary parkinsonism, those with the MMSE score 25 36 , history of stroke, neurological disorder other than PD, or any concomitant serious medical illness, and drug toxicity causing hallucinations, confusional episodes or delirium, were not included in the study. During the study, patients were not allowed to use neuroleptics, seda- tives, hypnotics or other antidepressants, as well as drugs with potential extrapyramidal adverse effects.
The study was approved by the Ethical Committee of the Faculty of Medicine, University of Belgrade, Serbia. Before entering the study patients gave written informed consent.
All the tests were performed in 18 out of 18 patients on days 11 and 18. Afterwards, 9 out of 18 patients were tested on day 50, and 8 out of 18 patients on day 80 (dropout rates of 50% and 56%, respectively). Therefore results were showed only until day 50.
All the patients were treated with two consecutive dosing regimens.
First, acute treatment with Flu -first day, the patients received Flu, 20 mg per day, at 8 a.m. Evaluation of motor performances and blood sampling for Flu/NORFlu plasma concentration measurement were carried out immediately before the Flu treatment (day 1, 0 h), and 4 h, 6 h and 8 h after the administration of the drug. Flu was than withdrawn for three consecutive days. On the fifth day, patients received 40 mg of Flu at 8 a.m. and all the tests and blood sampling were repeated in the same order (day 5, 0-8 h after administration of the drug). The pattern of blood sampling depends on T max for Flu, ranging from 4 to 8 h after the single dose administration 37 (Figure 1, panel A).
Second, chronic treatment with Flu -in the same patients, regular Flu treatment was initiated (20 mg per day, at 8 a.m.) on day 6 after the beginning of such a therapy, and the motor performances were evaluated on days 11, 18, 50 (steady state for Flu was reached after 18 days of Flu treatment) (Figure 1, panel B).
Two blinded refers evaluated severity of motor impairment using the Unified Parkinson's Disease Rating Scale (UPDRS) -motor score 38 , ADL (Schwab and England Activities of Daily Living Score) and computerized version of the quantitative motor test Finger Tapping Test (FTT) 39 and the Purdue Pegboard Test (PPT) 40 . The current severity of depression was evaluated using the 17-item HDRS 41 .
Bioanalytical method used for determination of plasma Flu and NORFlu concentrations was high performance liquid chromatography (HPLC) coupled with mass spectrometry (MS). The method used a liquid chromatograph Therm Separation Products Spectra System (Autosampler AS3000, HPLC binary pump P 2000, Degasser SCM 1000), mass spectrometer with electro spray ionization source (Finnigan MAT SSQ 7000 LC/MS -ESI System), Computer Digital UNIX Alpha Station 255. Recovery was very high, not less than 90.8% for Flu and 80.2% for NORFlu. Limit of quantification was 2.5 g/L for Flu and 10 g/L for NORFlu, and limit of detection was 1 g/L for Flu and 5 g/L for NOR-Flu. Correlation coefficient was 0.9993 (concentration range of 2.5-250 g/L), and 0.9989 (concentration range of 10-250 g/L), for Flu and NORFlu, respectively. Coefficient of variation, calculated for precision, was not higher than 8.33% and 8.83% for Flu and NORFlu, respectively.
The results are expressed as the mean ± standard error of the mean (S.E.M.) of N observations (descriptive statistics). Comparisons between groups were analyzed using the Fisher's exact test, t-test, and one-way analysis of variance (ANOVA), when appropriate. In addition, correlation analysis, factor analysis, extraction method (principal component analysis), rotation method (Oblimin with Kaiser normalization) and trend analysis (fitting or least square method) were used.

Results
All the patients were right-handed. Both groups, PD 0 and PD t , had similar laterality of Parkinson's symptoms (affected right side/affected left side = 6/3).
Among 12/18 patients with the affected right side, there was no significant difference between FFT for the right hand (FTTr) and FTT for the left hand (FTTl) scores, as well as between PPT for the right hand (PPTr) and PPT for the left hand (PPTl) scores (p = 0.66, and 0.89, respectively).  Among 6/18 patients with affected left side, FTTr was significantly better than FTTl (p = 0.03) and PPTr was significantly better than PPTl (p = 0.02). In addition, only PPTr score was significantly higher in the left side-affected PD patients comparing to the right side-affected PD patients (p = 0.03).
Age, gender and main clinical scores of PD 0 -and PD tpatients are shown in Tables 1 and 2. Depressive symptoms were similarly reduced after 18 days of Flu treatment in both PD 0 and PD t patients (Table 2, HDRS scores, p < 0.05). At the same time, Parkinson's disability was remarkably improved, especially in PD t patients ( Table 2, UPDRS and ADL, p < 0.05, both).
Acute treatment with Flu: there were no remarkable changes in motor function scores (FTT, PPT) after the administration of 20 mg of Flu (day 1), or 40 mg of Flu (day 5) ( Table 3).
The groups PD 0 and PD t differ only in FTTr scores at 0 h and 8 h after the administration of 40 mg of Flu (day 5).
Chronic, treatment with Flu: plasma concentrations of Flu and NORFlu increased in a time-related manner (C Flu , and C NORFlu , respectively) ( Figure 2). Table 4 shows plasma concentrations of Flu and NOR-Flu, as well as motor performance scores for each group assessed during chronic treatment with Flu.
Different patterns of changes were observed in the PD 0 and PD t patients. In the former case, a sustained increase in both C Flu , and C NORFlu was observed until day 18, i.e. the plateau was reached after 18 days of treatment. In the latter case, plasma concentrations continuously raised until the end of the observation period (day 50) ( Table 4). C Flu was significantly higher in PD 0 than in PD t group after 18 days of treatment ( Figure 2A, Table 4).
During chronic treatment with Flu, FTTr scores in the group PD 0 were continuously higher than in the group PD t , reaching the significance on days 11 and 50 (P = 0.03 and 0.04, respectively) ( Table 4). Such a difference was less pronounced regarding FTTl, PPTr and PPTl scores, never reaching statistical significance.

PD -Parkinson's disease; PD 0 -de novo PD patients; PD t -PD patients with stable antiparkinson's therapy
Of note, the raise in C Flu between days 0 and 18 (the plateau) coincided with the increase in FTT and especially in PPT scores (Tables 3 and 4).
Factor analysis reveals that influence of Flu/NORFlu concentrations increased over time (cummulative data from both PD 0 and PD t patients; plasma samples were taken on days 0, 5, 11, and 18, six hours after Flu administration). The variance explained by the concentrations of Flu and NORFlu permanently increased from 13.9% (day 5) to 29.9% (day 11) and 37.6% (day 18) of cumulative variance (values of 89.4%, 84.9% and 91.8%, respectively). At the same time, influence of motor function scores decreased over time: variance explained by PPT and FTT scores of 75.5%, 55%, and 54.1% (days 5, 11, and 18, respectively).
PPT and FTT scores significantly correlated on day 11 (r = 0.62; p < 0.01). In addition, an inverse correlation was found between Flu/NORFlu concentrations and PPT-, but not with FTT scores, on day 18 (r = -0.70 and 0.48, respectively).
Gastrointestinal, cardiovascular side effects and/or insomnia, somnolence and excessive daytime sleepness as adverse reactions to Flu were not reported in the PD patients considered in the study.

Discussion
The major results of our pilot study show that Flu treatment may alleviate depression in PD patients without deterioration of motor function scores. FTT, PPT and UPDRS-motor scores were even improved despite the parallel increase in plasma concentrations of Flu/NORFlu during the first 18 days of the study.
Depression in PD must be properly diagnosed and treated 42 . However, rare reports on the use of various antidepressants in PD patients offer controversial data on their safety regarding motor adverse reactions. Controlled clinical studies confirming the efficacy of Flu in PD patiens and assessing the risk-benefit ratio of such a therapy are still lacking 43 .
The broad therapeutic window for Flu is due to its highly variable pharmacokinetics 5,[44][45][46] . Flu steady state is achieved approximately after 3 weeks (concentrations of approximately 110 g/L). If plasma concentrations increase above 110 g/L, the dosage should be adjusted accordingly. Factor analyses indicates that mean Flu concentrations of approximately 60-110 g/L have the most powerful effect on both PPT and FTT scores, which were significantly improved within that concentration range.  The PPT and FTT are quantitative motor tests. While FTT more reflects motor speed, the PPT is a test for fine motor functions and coordination 40,47 . Since all the patients were right-handed only among 6/18 patients with affected left side FTTr and PPTr were better than FTTl and PPTl, respectively, pointing to more efficient compensatory mechanisms in dominant hand 48,49 .
The pharmacological profile of fluoxetine is unique among the antidepressants used in PD patients. Fluoxetine is both SSRI agent and a 5HT 2C antagonist 50 . A recent investigation confirmed that 5HT 1A agonists and 5HT 2C antagonists could be important features in treatment of PD. In particular, 5HT 2c receptors seem to tonically inhibit dopamine release from all three major dopaminergic pathways. Accordingly, 5HT 2c antagonists could block such an inhibiton, especially in the terminal regions of the nigrostriatal and mesolimbic pathways 51 .
Additionally, 5-HT 2c receptors are selectively located within substantia nigra pars reticulata (SNr) and medial globus pallidus (GPm) and 5-HT via 5-HT 2c receptors is excitatory in the SNr 52-55 , which may contribute to the increased activity of these regions in PD. Systemic administration of selective 5-HT 2c antagonists to 6-hydroxydopamine-lesioned rodents potentates the antiparkinsonian action of dopamine D 1 and D 2 agonists 56, 57 , which is an action mediated via 5-HT 2c receptors in the SNr 56 . Thus, 5-HT 2c receptor antagonists may improve parkinsonism and drugs with 5-HT 2c receptor antagonist action, such as fluoxetine, are unlikely to worsen PD 57 .
The pathophysiological mechanisms involved in mood disturbances in PD remain complex. Serotonergic dysfunction has been postulated as such systems are involved in mood disorders in non-PD and the raphe nuclei, as well as hippocampus and prefrontal cortex, appear to be the primary sites affected 58,59 . Moreover, transcranial ultrasound studies have suggested an association with reduced brainstem raphe echogenicity and nigral hyperechogenicity in patients with depression preceding PD onset compared with nondepressed patients with PD 60 . As the PD disease progresses, Lewy bodies occur with the rostral raphe, thalamus and limbic and cortical regions [15][16][17][18][19][20][21][22]61 , which may result in the mediating of mood disturbances in PD [23][24][25][26] .
In depression associated with PD, PD-specific pathology, with multiple transmitter deficiencies in mesocortical monoaminergic systems, plays a major role. This includes the mesocorticolimbic dopaminergic projection as well as mesocortical noradrenergic and serotonergic projections. Corticolimbic noradrenergic denervation through cell loss in the locus coeruleus and serotonergic deneravtion via serotonergic cell loss in the raphe nucleus are also likely to be important 11-15, 22-26, 62 . Postmortem evidence showed lower density of neurons in the dorsal raphe nuclei in depressed versus nondepressed patients with PD 22 and cerebro-spinal fluid measurement in vivo showed reduced serotonin metabolite (5-HIAA) levels in depressed patients with PD 63,64 . A [11C]-DASB PET study in seven patients with PD with untreated depression showed elevated serotonin transporter binding in the prefrontal cortex compared with non-PD-matched controls 65 . Recently, Politis et al. 66 have reported that the patients with PD with the highest scores for depressive symptoms showed significantly increased [11C]-DSAB binding in the amigdala, hypothalamus, caudal raphe nuclei and posterior cingytlate cortex compared with those patients with low depression scores, though not compared with healthy controls. The [11C]-DSAB binding values in other regions, including the anterior cingulate cortex, caudate, insula, prefrontal cortex, putamen rostral raphe nuclei, thalamus and ventral striatum, were similarly decreased in patients with PD, irrespective of their depressive symptoms scores, compared with the healthy controls. This study demonstrates that depressive symptoms in antidepressant-naïve patients with PD are associated with relatively higher serotonin binding in raphe nuclei and limbic structures. A relative increase in serotonin transporter binding in these regions could reflect either lower extracellular serotonin levels or a disease-related loss of presinaptic serotonergic neurotransmission in contributing to the pathophysiology of PD depression 62,66 .
The phenomenology of depression in PD is also different from that in patints with non-PD with less anhedonia and feeling of guilt 67 . While etiology of depression in Parkinson's disease is unclear (biochemical changes, psychosocial factors and situational stressors have all been implicated), it has an adverse effect on the quality of patients' lives and doctors should ensure that it is diagnosed and properly treated 1,4,5,68 .
Therefore, along with improvement on parkinsonian quality of life due to antidepressant activity of SSRI, symptoms such as bradikinesia, hypomima, hypophonia that overlap between depression and parkinsonism could ameliorate because an improvement of mood symptoms 1,9,10 . Evenmore, Suzuki et al 69 , 2010, suggested that SSRIs such as fluoxetine potentially are therapeutic drugs for non-motor symptoms as well as motor symptoms in patients with PD, since fluoxetine can reverse the downregulation of cell proliferation in the subgranular zone by the unilateral 6hydroxydopamine lesion.
All these various mechanisms could explain why the improvement in Parkinson's disability scores in our patients coincided with an increase in plasma Flu and NORFlu concentrations during the first 18 days of antidepressive treatment.
Another question is to assess the possible difference between PD 0 and PD t patients' response to Flu treatment. The beneficial effects of Flu on motor symptoms of PD patients seem to be more pronounced in PD t group (UPDRS and ADL scores). In addition, PPT scores were mostly higher in PD t patients during chronic treatment with Flu increasing continuously by the end of the study (day 50). However, the antidepressive efficacy of Flu was similar in both PD groups (HDRS). Also, the statistical significance was rarely observed between those groups regarding motor function scores; FTT values were even somewhat higher in PD 0 patients on days 11 and 50.
According to Taylor et al. 70 , depressive symptoms precede those of motor dysfunction in 12-37% of patients with PD. On the other hand, algorithms for treating depression in PD suggest that optimal antiparkinsonian treatment should precede administration of antidepressants 1,71 . Our results support such an approach only partially: PD 0 and PD t groups did not differ in their response to antidepressive therapy, while the influence of Flu on motor functions scores was not consistently related to the pretreatment with antiparkinsonian drugs. Nevertheless, successfull treatment of PD before the administration of antidepressants may diminish overlapping of depressive symptoms and core Parkinson's diasease symptoms 1 .
In the present study, we failed to observe any deterioration in motor performance scores of patients with PD that was related to the increase in plasma Flu and NORFlu concentrations. A slight improvement was even observed in all the scores (UPDRS, ADL, FTT and PPT). Similar results were obtained with citalopram, which improved mood but did not decrease motor performance scores in PD treated with levodopa; at the same time, citalopram improved the parkinsonian dysability, bradykinesia and finger taps after one and four months of treatment, both in patients with and without depression 72,73 . Also, Weintraub et al. 44 , 2006, reported that escitalopram was well tolerated, but produced only a partial response in the treatment of major depression in elderly PD patients (mean age of 72.1 years). Two openlabel studies suggested that sertraline reduced depression in PD patients, with additional beneficial effect on anxiety, without influencing motor function 74,75 . Additionaly, Ilic et al. 76 showed that the treatment with sertraline exerts complex modulatory effects on human motor cortex with potential behavioural usefulness. In another open-label study with paroxetine (20 mg/day) given to 33 nondemented depressed PD patients during 6 months, Ceravolo et al. 77 , in 2000, reported a significant improvement of depression, as evaluated by HDRS, without influence on parkinsonian symptoms. In only one patient fully reversible worsening of tremor was observed. However, paroxetine frequently may induce tremor as an adverse effect, with a prevalence of 1% to 2%. Chung et al. 78 in 2005, reported that the short-term paroxetine treatment did not alter the motor response to levodopa in patients with PD.
On the other hand, in two retrospective studies worsening of motor symptoms was observed in only small number of PD patients treated with SSRIs 79, 80 . In a prospective study comprising 65 depressed PD out-patients treated with paroxetine (10-20 mg/day) for at least 3 months, two out of 52 patients who completed the study (3%) experienced worsening of parkinsonian symptoms 79 . However, van de Vijver et al. 80 , in 2002, observed that the start of SSRI therapy in levodopa users was followed by a faster increase of antiparkinsonian drug treatment. Gony et al. 81 , in 2003, failed to find any significant difference in the occurrence of serious extrapyramidal symptoms between different classes of SSRI antidepressant drugs in patients with PD treated with dopaminergic antiparkinsonian drugs. According to the results of several studies [82][83][84] , including our results with Flu, it seems that the benefit of SSRIs outweigh the potential problems due to adverse effects and that they may be considered to be the rational choice in the treatment of depression in PD.
There are several limitations of the study: it was an open-label study without randomization including a small number of patients. As with all nonradomized, open-label trials at tertiary research centers, many non-specific factors, such as relatively long duration of symptoms in de novo PD patients, may have influenced the results. However, the quantitative evaluations of motor functions using FTT and PPT significantly improved objectivity and validity of our findings. The observed dropout rates (50% and 56% on days 50 and 80, respectively) are high but fit to the range observed in clinical trials to depression 83 .

Conclusion
This pilot study suggests that Flu 20 mg is effective and well tolerated antidepressant in patients with Parkinson's disease. In addition, Flu improved motor function scores in PD patients and such improvement was observed in parallel with the increase in plasma Flu and NORFlu concentrations. Also, the effects of Flu were similar in de novo PD patients and in those already treated with antiparkinsonian medications.
Therefore, our results would allow an optimal design for further large, prospective, randomized, placebocontrolled clinical trials that are necessary to evaluate the efficacy and safety of SSRI antidepressants and allow the development of evidence-based guidelines.