Functional Motor Symptoms in Parkinson’s Disease and Functional Parkinsonism: A Systematic Review

Parkinson’s disease (PD) and functional neurological disorders (FNDs) are very common and usually considered as distinct and nonoverlapping disorders. Recent studies showed that functional symptoms and PD co-occur in 1.4%−7% of patients with PD (1, 2). The diagnosis of FND in PD may be challenging, because of the potential overlap of symptoms and clinical signs of FND and PD, notably when patients present with pain, resting tremor, or akinesia (3–6). Moreover, FND symptoms, which are fluctuating and sensitive to distractibility, may be difficult to distinguish from fluctuations of PD symptoms related to stress or anxiety (5).

Clinical examination is crucial for an accurate diagnosis of both diseases. Some clinical signs are relatively specific of functional parkinsonism (FP) (7) and can be helpful to distinguish FP from PD. In FP, tremor may be present during all phases (rest, posture, and action), with a variability in frequency, an entrainment effect, and a distractibility (5). The characteristics of rigidity in FP tend to resemble those of oppositional paratonia related to frontal lobe dysfunction, and rigidity is diminished by reinforcement maneuvers. There is no bradykinesia but slowness of voluntary movements, with a nonprogressive decrease in amplitude (no decrement), distractibility, and inconstancy (5). There is sometimes an exaggerated discrepancy between the major apparent difficulty in movement initiation and the actual speed of movement execution when eventually performed—during the finger tapping task for example. (See Table S1 in the online supplement to this article for a detailed comparison of PD and FP.)

The identification of underlying PD in FND is crucial, because accurate dopaminergic treatment might significantly improve PD symptoms. In addition, accurate diagnosis of FND in PD is critical, because functional symptoms may lead to significant disability, escalation of antiparkinsonian treatment, requests for alternative therapy (such as deep brain stimulation), and poorer quality of life (2). Moreover, correct identification of FND in PD might improve knowledge about pathophysiology and potential common mechanisms involved in both diseases (8).

A parkinsonian syndrome can occasionally reflect a pure functional disorder—functional parkinsonism. FP represents 1.7%–25% of all functional movement disorders (5, 7, 9). The diagnosis of pure FP involves the elimination of PD as a differential diagnosis and as an associated diagnosis. There is evidence that patients with FP are often misdiagnosed as having PD. An example of such a situation is the high proportion of patients with “scan without evidence of dopaminergic deficit” (SWEDD) enrolled in PD clinical trials (7, 10–12). Most of the patients with SWEDD are misdiagnosed as having PD, and a proportion of these patients have FP, which tends to be underdiagnosed (10, 13).

To clarify the link between FND and PD, we analyzed all studies assessing FP or functional neurological symptoms in PD.

Methods

We followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) (14) guidelines for the literature search and analysis.

Eligibility Criteria

We included all English-language articles published in peer-reviewed journals that reported original data about case reports or cohorts of PD patients with functional neurological symptoms and patients with FP. The diagnosis of functional symptoms was based on clinical examination and mainly relied on DSM-IV-TR or DSM-5 criteria—and, more specifically, on the Fahn and Williams (15) or Gupta and Lang (16) criteria for FNDs (see Table S2 in the online supplement). Diagnosis of PD was also based on clinical examination, and most studies relied on the UK Parkinson’s Disease Society Brain Bank Clinical Diagnostic Criteria (17).

Studies were excluded if they did not include a description of patients. To ensure an exhaustive review, articles were included even when some data were missing.

Search Strategy

We searched on PubMed and ScienceDirect and attempted to screen all the literature from 1988 (first description of a psychogenic parkinsonism) to December 2018. The following terms were used: (“psychogenic” OR “functional disorder” OR “conversion” OR “hysteria” OR “somatoform”) AND (“Parkinson” OR “parkinsonism”). Relevant studies were identified, and their reference lists were hand-searched.

Study Selection

Two authors (MAA and BG) independently screened for the titles of potentially eligible publications. Some articles were excluded at this stage (see Figure S1 in the online supplement). After review of abstracts and articles, likely inclusions were assessed by MAA and BG, who independently extracted all information on age and gender of patients, symptoms, duration, psychiatric history and identified triggering stressors, imaging data, past treatments, efficacy of treatments, and disability.

In studies about FP, some patients with an initial diagnosis of FP were eventually diagnosed as having an underlying neurodegenerative PD. We classified these patients in the category PD with FND (PD-FND). Patients whom we classified in the FP category are those who had no underlying degenerative parkinsonism.

Results

The initial search strategy yielded 844 articles. After removal of duplicates and review of titles and abstracts and then full text and references, 22 studies were included: 12 studies reported on PD patients with FND (PD-FND), and 16 studies reported on patients with FP. (See Table S2 in the online supplement for detailed information on the diagnostic criteria used in the selected studies.)

Functional Neurological Symptoms in PD (PD-FND)

The 12 studies that reported on PD-FND patients included a total of 121 patients (Table 1). The study samples varied in size, ranging from one to 53 patients. Most studies had small samples. Mean age at examination was 61.3 years (range 37.5–71.2), and 69 women were included (57% of the 121 patients). History of psychiatric disorder was frequent but was not always found. Data about psychiatric history were available for 80 patients, and 48 of them (60%) had a history of a psychiatric disorder. Among these 48 patients, details about psychiatric disorders were available for 15 patients only, depression being the most frequent (N=12/15, 80%). Other psychiatric comorbidities were anxiety, suicidal behavior history, personality disorders, and chronic pain. Stressors were also frequent. Among the 63 patients with available data, a stress factor was identified in 52 (82.5%). Reported stressors included physical injuries, stress at work, or loss of a family member or friend.

Two of the 12 studies assessed the prevalence of functional overlay in PD (1, 2). The first study included 1,360 patients with neurodegenerative disorders. It found a significantly higher frequency of somatic symptom disorders in PD (7%) and in dementia with Lewy body (12%) than in other neurodegenerative diseases (Alzheimer’s, multiple-system atrophy, progressive supranuclear palsy, or frontotemporal dementia, 0%−3%) and also a higher frequency than in populations with psychiatric disorders (1.5%) (1). Among all patients with PD, Wissel et al. (2) calculated that 1.4% had PD-FND.

Little is known about symptom onset. FND preceded the diagnosis or occurred at the same time as PD in a majority of patients (61.6%, N=61/99). Onset was acute for 55% of the patients in Wissel and colleagues’ study (2).

The vast majority of patients with PD-FND described in these studies presented with motor functional symptoms (Table 1). Patients often had more than one phenotype. Indeed, in Wissel and colleagues’ study (2), 60% presented with multiple different functional symptoms. Overall, among all 12 studies, tremor was the most common FND phenotype (N=41, 33.8%), but a large variety of motor functional symptoms were described: FP (N=28), paresis (N=23), gait disorder (N=29), dyskinesia (N=13), abnormal posture (N=12), dystonia (N=9), speech impairment (N=6), myoclonus (N=3), and give-way weakness (N=1). Detailed information was available for 31 patients, all of whom had motor FND: tremor, 45.2% (N=14); and FP, 64.5% (N=20). Two additional studies (N=90 patients) provided information about functional phenotypes, but most of the patients had combined phenotypes, so it was not possible to extract information about the proportion of patients presenting with motor FND (1, 2). In both studies, the motor phenotype was predominant (Table 1). Most of the time, functional symptoms were on the same side as the side most affected by PD, or symmetrical. Data were available for 76 patients, and FND was on the most PD-affected side in 50% (N=38), on the less PD-affected side in 17.1% (N=13), and symmetrical in 32.9% (N=25).

Six studies described patients with FP and found an underlying PD confirmed by striatal dopamine transporter imaging (DaT-SCAN) in 7.1%–66.6% (mean 32.6%) (3–6, 13, 18). One study specifically compared demographic and clinical features as well as treatments in PD patients with FP to patients matched on age, gender, and disease duration (2). PD patients with FP had a longer delay to PD diagnosis, notably when FP preceded PD symptoms, and they had a greater prevalence of dyskinesia, compared with the matched patients (42% versus 18%). They also had worse depression and anxiety symptoms. PD patients with FP had higher doses of dopaminergic treatment (dopa equivalence, 972 mg versus 741 mg, p=0.029) (2). In general, PD patients with FP had more frequent use of health care and were more often hospitalized.

Authors of all 12 studies proposed dopamine transporter imaging to support the diagnosis of underlying PD. A DaT-SCAN (¹²³ioflupane or [99mTc]TRODAT-1; 93.9%, N=93) or a fluoro-dopa PET (6.1%, N=6) was performed for 81.8% (N=99) of the 121 patients and was abnormal in all of them. Wissel and coauthors (2) showed that use of DaT-SCAN was more frequent in PD-FND patients than in PD patients (58.5% versus 30%, p<0.001). DaT-SCAN proved useful to avoid misdiagnosis of underlying neurodegenerative PD in patients presenting with FP (6, 13, 18). Electrophysiological recording is useful to differentiate functional tremor from PD’s tremor (3).

FP

We included 16 studies with a total of 120 patients with FP (Table 2). Mean age was 50.7 years, and 62.1% (N=72) were women (data about gender were not available for four patients). Mean duration of symptoms was 5.03 years, with an abrupt onset in 50% (N=50), and a progressive onset in 50% (N=50) (data were not available for 20 patients).

Data about psychiatric history were available for 108 patients, 73 of whom (67.6%) had a history of a psychiatric disorder. Among these 73 patients, information on the psychiatric diagnosis was available for 55 patients, and depression was the most frequently reported disorder (N=42, 76%). Other reported psychiatric comorbidities included anxiety, bipolar disorder, posttraumatic stress disorder, personality disorders, and substance use disorders. Triggering stressors were identified in 46.8% of patients (N=44/94) and consisted of physical injury, stress at work, or loss of family or friends.

A tremor was described on clinical examination in 72.2% (N=57) of FP patients (available data for 79 patients). Tremor could be of the rest, postural, or action type (13, 19, 20); data were available for 36 patients, of whom two had no tremor at examination, and 70.5% (N=24) had a mixed component of postural, rest, or action tremor). Tremor had a variability over time in amplitude, frequency, and distribution (3, 4) and with distractibility and entrainability (4). Other reported FP signs were oppositional rigidity with distractibility, atypical gait, pseudo-slowness (effortful slowness of the movement), pseudo-hypokinesia with no decrement (fixed reduction of the amplitude of the movement, without decrement), and other incongruent neurological signs.

Some patients with FP had a normal DaT-SCAN that confirmed the absence of dopaminergic denervation (38 patients had normal DaT-SCAN results). Diagnosis of FP was based only on clinical history and examination of 12 patients from three studies (Walters et al. [19], N=1; Lang et al. [5], N=10; and Frasca Polara et al. [4], N=1) who did not undergo DaT-SCAN, and data about DaT-SCAN were not available in three additional studies (N=70) (7, 20, 21).

In one of the above studies, the authors aimed to provide an estimate of the prevalence of FP in the general Swiss population (4). The crude prevalence rate of FP was 0.64 per 100,000, representing 0.24% of all patients with parkinsonism and 1.3% of nondegenerative parkinsonism (drug-induced parkinsonism [44%] and vascular parkinsonism [37%] being the most frequent).

Little data were available regarding response to L-dopa treatment; data were available for 28 patients, and there was no response to L-dopa treatment for 18 of them (64.3%). Other treatments were occasionally reported. One patient underwent a single session of repeated transcranial magnetic stimulation, leading to a complete and immediate recovery from all symptoms, with a persistent remission at 6 months (22). Another patient was initially diagnosed as having multiple-system atrophy, and because of the reported levodopa response, had a unilateral subthalamic nucleus deep brain stimulation surgery but without any improvement (23).

Discussion

Main Results Regarding PD-FND

Most patients in the studies reviewed here presented with motor FND. This raises the question of whether motor FND is more prevalent in PD or whether other FND phenotypes, such as nonepileptic seizure or sensory disorders, are underdiagnosed or underreported. Interestingly, tremor, gait disturbance, and parkinsonism, which are part of the PD phenotypic spectrum, were the most prevalent presentations. PD patients tended to express FND on their most PD-affected side, and FND preceded or was mainly diagnosed at the same time as PD (21). This finding is similar to what is observed in other neurological conditions with functional overlay, such as seizures with functional nonepileptic seizure (24), tics with functional tics (25), or various other hyperkinetic movement disorders with functional hyperkinetic movements (26). In this case setting, functional manifestations tend to be reminiscent of those resulting from the other neurological condition or to involve the same region of the body. A cohort study showed that FND was significantly more frequent in PD and in dementia with Lewy bodies than in other neurodegenerative disorders (1).

These results raise the question of common mechanisms involved in both diseases. The loss of dopamine in PD induces a switch of the control of the motor behavior from a procedural process toward a more conscious and goal-directed mechanism involving attention (27). As proposed by Pareés et al. (8), this increased attention toward the affected limbs, combined with physical symptoms induced by the disease and additional affective and cognitive factors, may produce the required vulnerability to induce FND. Affective and cognitive factors may be induced by PD (28, 29) but may also preexist PD. Moreover, emotion dysregulation may play a role in the development of FND (30). We found that psychiatric comorbidities and exposure to potential triggering stressors were very frequent. Only one study compared patients with PD-FND to patients matched on age, gender, and disease duration (2). In this study, PD-FND patients were more likely than the comparison patients to have a preexisting psychiatric disorder and had worse depression and anxiety. However, several biases may have influenced these results: the study was a retrospective analysis of charts, and psychiatric comorbidities may have been searched more carefully in PD-FND patients than in PD patients. Of note, previous studies that compared patients with FND to those with PD did not find significant differences in psychiatric comorbidities (31, 32). Additional prospective case-control studies are necessary to clarify the prevalence and the potential impact of stressors and psychiatric comorbidities in patients with PD-FND and those with PD.

DaT-SCAN was used in most studies to confirm the loss of dopaminergic neurons and hence the diagnosis of underlying PD. Some authors proposed electromyography and accelerometry to further document the functional movement disorders (3, 8). As highlighted in many case reports, we recommend performing DaT-SCAN because the distinction between PD, PD-FND, and FP can be challenging without DaT-SCAN and electrophysiological recording, and patients may otherwise be misdiagnosed (11, 13, 33, 34).

Finally, the studies reviewed provided limited data about prognosis in PD-FND. It has been shown that compared with PD patients, patients with PD-FND were likely to receive higher doses of antiparkinsonian treatment and more medical attention (number of hospitalizations and calls to their doctor) and that the duration before diagnosis was longer (2). This may lead to more difficulties in the care of these patients, and future prospective research is necessary to better characterize the prognosis of PD-FND patients.

Accurate diagnosis of PD with FND is critical, because functional symptoms in PD may lead to significant disability, escalation of antiparkinsonian treatment, requests for alternative therapy (such as deep brain stimulation), and poorer quality of life (2).

Main Results Regarding FP

Overall, the reported prevalence of FP in functional movement disorder (FMD) series varied between 1.7% and 25% of the patients (21). In the Toronto Western Hospital series of FMD patients (N=64), 6.4% had FP (N=4) (5). Jankovic (7) reported that among more than 530 patients with FMD, 3.2% (N=17) had FP as the predominant presentation.

In this review, we found 120 reported cases of “pure” FP, i.e. with no comorbid PD diagnosis. A tremor was present in most patients and was atypical for PD’s tremor because it had a significant postural or action component in most patients. Other clinical characteristics of functional tremor were distractibility and entrainability, and other reported signs of FP included oppositional rigidity with distractibility, atypical gait, pseudo-slowness, pseudo-hypokinesia without decrement, and other incongruent neurological signs. These positive signs of FP are usually underrecognized. Indeed, only 25% of these cases received a diagnosis of FP at the initial visit in the study by Frasca Polara et al. (4). There is thus an important unmet need for development of specific training for neurologists in FND semiology. The search for positive signs of FP should be systematic when patients present unusual features of PD, such as a sudden change in PD course or unusual evolution of PD, or when there are some unusual features upon examination. Better training and extended evaluations would help avoid misdiagnosis of FP and improve prognosis of patients. Moreover, even for well-trained neurologists, the distinction between FP and PD can be very challenging (18, 21–23, 35, 36).

In this context, a proportion of patients underwent a DaT-SCAN to check for the normality, and hence the absence, of underlying PD. There is evidence that DaT-SCAN is useful for PD diagnosis in doubtful cases (37). Although a normal DaT-SCAN is necessary to rule out underlying PD, it does not discriminate between essential tremor, FND, and other diagnoses (7, 10–12). In the studies reviewed, only a few patients underwent EMG/accelerometry recording to document and confirm the clinical diagnosis of FP (3). Electrophysiological testing is not available in most centers (even in large tertiary referral centers, unless this is a focus of research), and only a few experts have sufficient training to conduct this assessment. We believe, however, that EMG and accelerometry can be of particular interest to further document the diagnosis of FP, notably to improve diagnostic accuracy (38).

Management of functional symptoms can be challenging, and little information was available among all cases about treatment and outcome. Response to L-dopa treatment was available for 28 patients. Most of these patients (64.3%) did not show any improvement with L-dopa. Other management approaches were reported: one patient improved dramatically with transcranial magnetic stimulation (22), another improved spontaneously at follow-up (19), and another did not improve after deep brain stimulation (23). There was no information about the efficacy of physiotherapy, even though it is widely used in FND (39).

FP and Functional Neurological Symptoms in PD

FP and PD present with very similar clinical symptoms and signs. Both diseases are characterized by the association of akinesia, tremor, and/or rigidity. However, several clinical differences can be highlighted: patients with FP, compared with those with PD-FND, are younger, more often women, more likely to have a psychiatric history, and less likely to have identified stress factors (see Table S3 in the online supplement). Also, atypical tremor is more than twice as frequent in FP patients, compared with PD-FND patients.

Among patients with FP, the mean rate of comorbid PD was 32.6%, which is significantly higher than what is usually reported in FND (40). The high rate of comorbidity of both diseases raises the question of shared pathogenic mechanisms. Further prospective studies are necessary to verify the high rate of PD comorbidity in FP and to better characterize the differences and similarities between pure FP, FP with comorbid PD, and PD-FND.

Limitations

Several limitations should be acknowledged. First, we decided a priori to include all trial designs and case reports of patients with PD-FND or FP, even if some data were missing or incomplete. Therefore, our conclusions were drawn in some studies from subgroups of the total sample and might not be generalizable to the entire population. Second, diagnostic criteria differed in the studies we reviewed, mainly in the case reports. Some FP patients did not undergo DaT-SCAN, and the diagnosis was based only on clinical examination. Some of these patients may have had undiagnosed underlying PD and may have been misdiagnosed as “pure” FP patients. Moreover, FND symptoms, which are fluctuating and sensitive to distractibility, may be difficult to distinguish from fluctuations of PD symptoms related to stress or anxiety. Some PD patients with anxiety- or stress-related fluctuations may thus have been incorrectly diagnosed as having FND.

Except for case reports, all studies reviewed were retrospective, implying multiple potential biases. Therefore, to assess the prevalence of functional neurological symptoms in PD and to systematically compare “pure” FP with PD-FND, future research should be conducted with a prospective design and a systematic assessment of a larger number of consecutive PD patients.

Conclusions

The findings of this review about FND and PD (including PD-FND and pure FP) suggest that a relationship may exist between these two diseases, with overlapping and shared clinical characteristics. These observations also highlight the importance of screening for FND symptoms in patients with PD and of considering the possibility of PD in FP patients. Additional studies are needed to understand the pathophysiology of FP and the shared or specific mechanisms involved in both FND and PD.