Depression is the most common neuropsychiatric disturbance in Parkinson's disease (PD), with a prevalence rate of almost 50%.1 There is a range of depressive disorders seen in PD, all of which can contribute to functional impairment.2 There are two major categories of mood disturbances in PD: depressive disorders (e.g., major depression, dysthymia), which last from weeks to years and can occur at any stage of illness, and mood fluctuations, which have been reported in up to two-thirds of advanced PD patients who experience motor fluctuations. These mood fluctuations can be frequent (occurring many times a day) and dramatic (patients can shift from extremely dysphoric and suicidal to euphoric) and can be more distressing to the patients than the motor fluctuations.
Research involving the phenomenology and underlying mechanisms of mood fluctuations in PD has been relatively limited. Most researchers who have described this phenomenon indicate that mood fluctuations tend to be associated with motor fluctuations in that patients experience decreased mood when "off" (immobile, parkinsonian) and elevated mood when "on" (mobile).3—7 Such reports have led to the hypotheses that these mood fluctuations are either a psychological reaction to motor dysfunction or an independent result of changing brain dopamine levels. Menza et al.8 found that in the majority of patients, mood improved from the "off" state (lowest levels of brain dopamine) to the "on" state but then worsened again in the "on with dyskinesia" state (highest levels of brain dopamine). On the basis of these findings, and the authors' belief that patients were most functional when "on" without dyskinesias, Menza et al. concluded that a reactive process was most likely. In their pilot study, Maricle et al.9 carried out intravenous infusions of levodopa in 8 PD patients with motor fluctuations and demonstrated that mood fluctuations are not simply a reaction to motor function since they generally precede motor changes by several minutes. This study also suggests that mood changes, like motor changes, may be a direct or indirect effect of levodopa therapy itself.
Although mood fluctuations have been noted mostly in patients who experience motor fluctuations, there is evidence that mood fluctuations may predate motor fluctuations in some patients. For example, Nissenbaum et al.5 described one patient who had been on carbidopa/levodopa for 17 months whose only evidence of motor fluctuations had been intermittent dyskinesias. She developed feelings of panic, anxiety, and depression if she forgot to take one of her three daily doses of carbidopa/levodopa on time. These authors also noted that mood fluctuations can disappear as the duration of disease and treatment progress. This example also highlights the issue of anxiety in PD. Like mood disturbances, both anxiety disorders and fluctuations in levels of anxiety have been reported in this illness. Although anxiety and depression often co-occur in PD, the relationship between the two remains unclear.
In an attempt to better understand mood fluctuations, including their relationship to motor function and level of anxiety, we conducted a pilot study in which 16 patients with PD and motor fluctuations completed hourly diaries for seven consecutive days documenting their mood, anxiety, and motor state using visual analogue scales.
Subjects included 11 men and 5 women with a mean age of 62 years. Subjects were recruited from our PD clinic and selected after a careful chart review demonstrated evidence of the following inclusion criteria: 1) idiopathic PD and 2) motor fluctuations. The diagnosis of idiopathic PD was based on a careful history and neurologic exam by a neurologist as well as diagnostic studies as indicated (e.g., MRI). Subjects had documented evidence of at least two of the three cardinal signs: rigidity, bradykinesia, and rest tremor. There were no diagnostic alternatives (i.e., parkinsonism was not due to trauma, brain tumor, infection, cerebrovascular disease, drugs, chemicals, or toxins). All subjects had motor fluctuations. Five subjects had severe wearing-off periods, and 11 subjects had discrete "on," "off," and "on with dyskinesia" periods.
Twenty-five subjects were invited to participate during a phone call in which the study procedures were explained. Eighteen subjects agreed to participate and were provided with a seven-day diary containing visual analogue scales for mood, anxiety, and motor function. Visual analogue scales provide a simple technique for measuring subjective experience and have been established as valid and reliable in a range of clinical and research applications.10,11 These scales were used by both Maricle et al.9 and Menza et al.8 for their evaluation of mood, anxiety and motor states in PD patients. The diary consisted of seven pages, one for each day. On each page (front and back) were three columns with the following headings: Mood, Anxiety, and Motor State. Under each heading were 18 horizontal 10-cm lines. Each line represented a one-hour interval starting at 6:00 a.m. and ending at midnight. Subjects were instructed to complete scales during each of their waking hours. The lines on each of the visual analogue scales contained descriptive anchors to help guide subjects. The anchors for mood were "extremely sad" and "extremely happy," those for anxiety were "extremely calm" and "extremely anxious," and those for motor were "extremely immobile (off)" and "excessively mobile (on with dyskinesia)." Subjects were instructed to make a vertical line through the point on the scale that best represented their state for that hour. They were provided with an example page. Subjects were also provided with the Geriatric Depression Rating Scale (GDS),12 the Beck Depression Inventory (BDI),13,14 and the Zung Anxiety Scale (Zung).15 They were advised to complete these self-assessments one day before starting the diary. Subjects were also provided with a history form inquiring about their past psychiatric histories and current medications. This information was verified based on careful chart review, but subjects were not specifically interviewed about the presence or absence of psychiatric symptoms. Sixteen PD patients returned completed diaries.
Spearman correlations were used to quantify associations between mood and motor scores, mood and anxiety scores, and motor and anxiety scores. Correlations (positive or negative) were considered to be significant if P-values were less than or equal to 0.05. Correlations between mood and motor, mood and anxiety, and motor and anxiety scores were determined for each subject for every day. A subject was considered to have fluctuations on a given day if the range in visual analogue scores for that day (mood, anxiety, and motor considered separately) was greater than 1 cm.
Subject characteristics, including age, psychiatric history, rating scale scores, medications, and number of days on which significant correlations between emotional and motor states were present are summarized in t1.
Psychiatric History and Current Medications
A history of depressive symptoms, as indicated by self-report or chart documentation, was present in 50% of cases, and anxiety symptoms were present in 56%. Thirteen percent of the subjects were taking tricyclic antidepressants, 44% were receiving selective serotonin reuptake inhibitors, and 44% were taking benzodiazepines.
The mean score on the GDS was 11. Seven subjects (44%) scored 10 or greater, signifying possible depression. However, the mean score on the BDI was 6.3, with only one subject scoring in the depressed range (≥16). The mean score on the Zung Anxiety Scale was 32.3. Only one subject had a score consistent with significant anxiety (≥45).
Presence or Absence of Fluctuations
As above, for the purposes of this study we considered subjects to have "fluctuations" on a given day if they had a range of at least one 1 cm on their visual analogue scores for that day. Mood, anxiety, and motor fluctuations were considered separately. One subject experienced motor fluctuations on 6 of 7 days. A total of seven subjects (44%) experienced daily mood fluctuations, and seven subjects experienced daily anxiety fluctuations. These were not the same seven subjects, however (i.e., some subjects experienced mood but not anxiety fluctuations and vice versa). The degree to which mood and anxiety fluctuated varied among subjects and even within subjects from day to day.
Correlations were determined for each subject for each day. A correlation was considered significant if the P-value was less than or equal to 0.05. In general, the magnitude of correlation when P was less than or equal to 0.05 was 0.5 or greater for positive correlations and —0.5 or less for negative correlations. This information is presented in t2.
Significant positive correlations between mood and motor scores were present on all 7 days in one subject, 5 of 7 days in one subject, 4 of 7 days in one subject, 3 of 7 days in one subject, 2 of 7 days in three subjects, 1 of 7 days in five subjects, and none of the days in four subjects.
Significant negative correlations between anxiety and motor were present on all 7 days in one subject, 4 of 7 days in one subject, 3 of 7 days in one subject, 1 of 7 days in four subjects, and none of the days in nine subjects.
Significant negative correlations between mood and anxiety were present on all 7 days in one subject, 4 of 7 days in one subject, 3 of 7 days in one subject, 2 of 7 days in three subjects, 1 of 7 days in three subjects, and none of the days in seven subjects.
With our diary study, we have discovered that although preliminary reports suggested that mood and motor states are closely correlated in PD, this is often not the case. Prospective assessment failed to confirm a consistent temporal relationship between mood and motor states in many of the PD subjects we examined. If subjects experienced better mood at times of better mobility, one would expect these measures to be positively correlated. When a correlation between mood and motor was present, it was more likely to be positive. There were, however, examples of subjects in whom the correlation was negative. There were also many instances in which mood and motor fluctuated but appeared to be independent. F1 through F4 provide graphic representations of four subjects who demonstrated different patterns of correlations.
Our findings also suggest that a consistent relationship between anxiety and motor states is not the rule. If patients with fluctuating motor and anxiety states experienced less anxiety during periods of better mobility (as has been suggested in the literature), one would predict that these measures would be negatively correlated. When present, correlations between motor and anxiety states were negative twice as often as they were positive. More often, however, motor and anxiety did not appear to be directly correlated. It is possible that different neurobiologic mechanisms for emotional and motor fluctuations are involved. It has been noted in the literature that patients generally experience increased anxiety and decreased mood together. Our data indicate that when the two emotional states were indeed correlated, this was the typical pattern. As in the case of motor and emotional correlations, however, mood and anxiety were not consistently correlated.
While initially surprising, the lack of correlation between emotional state and motor function may not be unique to PD. Even in primary psychiatric disorders where psychological (mood) and motor symptoms tend to be associated, the precise nature of their relationship is unclear. Preliminary evidence suggests that different mechanisms underlie the psychological and the motor disturbances in major depressive and bipolar illness. There appears to be, for example, a differential time course for the onset and resolution of psychological and nonpsychological symptoms. Because "somatic" symptoms (e.g., anergia, slowness) tend to respond more quickly than the mood disorder to antidepressant medication, there is a well-recognized increased risk of suicide when patients are in the very early recovery phase of a major depression. They have regained their energy and motor function, so they are able to more readily act upon the persistent depressed mood and suicidal ideation. As evidence that the thought disorder and psychomotor disorder in bipolar illness have different underlying mechanisms, Carroll16 noted that manic patients exhibited a behavioral response to physostigmine consisting of psychomotor slowing, but that the quality of their thought disorder was unchanged. Referring to reports of levodopa-induced mania, he questioned whether levodopa does in fact induce a true manic state in depressed bipolar patients, or whether it simply induces "behavioral mania" with hyperactivity but without the characteristic thought disorder.
Our study failed to reveal a consistent relationship between emotional fluctuations and a history of depressive or anxiety disorders, current mood and anxiety states as measured by self-assessment scales, or current medications. This result must be interpreted with caution, however, because subjects did not undergo structured psychiatric interviews. It is interesting to note that the one subject (#9) who experienced fluctuations of mood, anxiety, and motor states on a daily basis and demonstrated very consistent and dramatic correlations among all three states reported no previous psychiatric history, scored within the normal range on mood and anxiety rating scales, and was on no psychotropic medications. One might speculate that this individual experienced changes in anxiety and mood as a psychological reaction to motor status, rather than as an independent neuropathological phenomenon. The relationships between these factors may best be established in a larger study that employs formal psychiatric interviews. It should be noted that the present study included a relatively small number of subjects. It is, however, the most intensive prospective study of mood fluctuations in PD patients to date.
Potential limitations of the study include our inability to verify that subjects filled out the diaries each hour rather than going back and filling in portions at a later time. We also cannot guarantee that subjects fully understood and followed all of the directions. We did not perform any cognitive screening measures. We did feel that subjects understood the directions based on our telephone conversation with them. Furthermore, all subjects had undergone routine clinical evaluations within the past 6 months. Five subjects took the Mini-Mental State Examination and had scores of 27 or greater. Mental status exams were deemed "normal" in eight of the other subjects. There was no documentation of mental status testing in the remaining three.
The interpretation of presence or absence of fluctuations is somewhat limited in this study. Our definition was somewhat arbitrary, since it is not clear what constitutes "fluctuations" without normal control data. The only way to truly know whether or not subjects "fluctuate" to an abnormal degree would be to obtain diary information from a control group. A future study would ideally involve age-matched control subjects without PD in order to determine what constitutes a "normal" degree of mood fluctuation in this age group.
Although the optimal treatment for depressive disorders has not been clearly established, the best treatment for mood fluctuations is even more uncertain. It is possible that a better understanding of mood fluctuations and their relationship to depressive disorders in PD will provide us with further insights into the underlying mechanisms of both forms of mood disturbance and lead to better treatment approaches.