Parkinsonism: Clinical features and differential diagnosis

Motor parkinsonism

The term parkinsonism is used to describe a syndrome manifested by a combination of the following six cardinal features: (1) tremor-at-rest, (2) bradykinesia, (3) rigidity, (4) loss of postural reflexes, (5) flexed posture, and (6) freezing (motor blocks). A combination of these signs is used to clinically define definite, probable, and possible parkinsonism. Diagnosis of definite parkinsonism requires that at least two of these features must be present, with one of them being resting tremor or bradykinesia; probable parkinsonism consists of resting tremor or bradykinesia alone; and possible parkinsonism includes at least two of the remaining four features. The four major characteristics of parkinsonism—tremor, rigidity, akinesia, and postural disturbances (forming the acronym TRAP)—account for most of the clinical abnormalities. The initial feature of many basal ganglia diseases is slowness of movement (bradykinesia), which also constitutes a primary feature in the United Kingdom Brain Bank diagnostic criteria for PD ( ), but also paucity or absence of movement (akinesias), often associated with rigidity and tremor ( ). Some authors have used the term hypokinesia to describe a reduction in amplitude of movement. Many parkinsonian symptoms are explained by the combination of slowness and poverty of movement and increase in muscle tone (rigidity). These clinical features of parkinsonism will be discussed individually.

Parkinsonism, as a clinically defined feature, may reflect different underlying etiologies, such as PD, atypical parkinsonian syndromes (discussed in other chapters), and similarly appearing age-related features. Many parkinsonian symptoms and signs that could be associated with PD may be mistakenly attributed to “frailty” of the elderly because the disease is age-related ( ). It is worth noting that one hallmark of age-related frailty is sarcopenia, defined as gait speed slower than 1 m/s and measured appendicular muscle mass of 2 standard deviations (SDs) below the muscle mass of a 20- to 30-year-old individual, which occurs in 5% to 13% of adults between the ages 60 and 70 years ( ). Longitudinal studies of older adults from the Religious Orders Study and Memory and Aging Project (n = 791, mean of 6.4 years until death, mean age 88.5 years) followed with annual clinical evaluations and documented physical frailty demonstrate high rates of mixed neuropathologic findings at autopsy, including macroinfarcts, Alzheimer disease (AD), Lewy body pathologic findings, and nigral neuronal loss, all of which probably contributed to the progressive physical frailty ( ). From the same cohort, 964 of 2001 (48.2%) individuals developed parkinsonism during a 5-year follow-up ( ), which was related to age (hazard ratio [HR], 1.09; 95% confidence interval [CI], 1.08–1.10), but not to gender (HR, 1.06; 95% CI, 0.91–1.23) or education (HR, 0.99; 95% CI, 0.97–1.01). The incidence of parkinsonism increased with age: for adults younger than 75 years, the incidence of parkinsonism was 36 per 1000 person-years for adults younger than 75 years, 94.8 per 1000 person-years for those aged 75 to 84 years, and 160.5 per 1000 person-years for those 85 years and older. Most individuals exhibited 2 parkinsonism signs (83%; n = 800)—most commonly parkinsonian gait (n = 843; 87.4%). Incident parkinsonism was associated with depressive symptoms, neuroticism, urinary incontinence, sleep complaints, and chronic illness. Although walking tends to slow with normal aging, slow walking speed correlates with white matter changes only in individuals with low education level ( ). The progression of parkinsonism in older adults relates to the burden of mixed brain pathologic conditions. In a cohort of 1430 individuals followed for a mean of 8.5 years, 53.9% developed parkinsonism proximate to death; those with more rapid progression of parkinsonism had PD pathologic findings at autopsy and the independent presence of AD and cerebrovascular pathologic findings ( ). Interestingly, the association between a higher person-specific weighted pathologic score and more rapidly progressive did not differ between those with or without a clinical diagnosis of PD but related to the underlying pathologic burden.

Bradykinesia

Bradykinesia translates as slowness in movement but is considered the slowness of movement and reduction in frequency and amplitude of repeated movements in the context of PD examination. Bradykinesia is the most characteristic clinical hallmark for the diagnosis of PD. It may be initially manifested by slowness in activities of daily living (ADLs), movements of limbs and/or whole body, and reaction times ( ; ; ; ; ; ; ; ). On examination, bradykinesia manifests as slowness of movements but with a decrement in the amplitude and speed of the movement, such as when examined with the motor items in the International Parkinson and Movement Disorder Society and Unified Parkinson’s Disease Rating Scale (MDS-UDPRS) for fine motor movement or rapid alternating movements. Although speed and amplitude are usually assessed together on the MDS-UPDRS, there is some evidence that amplitude is disproportionally more affected than speed in PD and may be due to different motor mechanisms. Thus, some have proposed that amplitude should probably be assessed separately ( ). Other manifestations of bradykinesia include reduced arm swing when walking (loss of automatic movement) and gesturing when talking, loss of facial expressionxxx (hypomimia) ( ), reduced blinking, drooling because of failure to swallow saliva ( ; ), and monotonic and hypophonic dysarthria. Micrographia may be represent a form of bradykinesia, although it may reflect an abnormal response from reduced motor output or weakness of agonist force coupled to distortions in visual feedback ( ; ). Micrographia has been further characterized as consistent (small handwriting throughout) or progressive (progressive reduction in size); both types were associated with decreased activity and connectivity in the basal ganglia motor circuit, but in the persistent type, with additional disconnections between the rostral supplementary motor area, rostral cingulate motor area, and cerebellum on functional magnetic resonance imaging (fMRI), but levodopa only improved consistent micrographia ( ). Bradykinesia, like other parkinsonian symptoms, is dependent on the emotional state of the patient. With a sudden surge of emotional energy, the immobile patient may catch a ball or make other fast movements. This curious phenomenon, called kinesia paradoxica, demonstrates that the motor programs are intact in PD, but that patients have difficulty using or accessing the programs without the help of an external trigger. Therefore, parkinsonian patients can use prior information to perform an automatic or preprogrammed movement, but not to initiate or select a movement. Although PD represents the most common form of parkinsonism, there are many other causes of bradykinesia, the parkinsonian clinical hallmark ( Table 4.2 ).

The pathophysiology of bradykinesia is not well understood, but it is thought to relate to disruptions in basal ganglia circuity and dopaminergic function. Bradykinesia may result from failure of basal ganglia output to reinforce the cortical mechanisms that prepare and execute the commands to move ( ). Bradykinesia manifests as slowness of self-paced movements and prolongation of reaction and movement time. Reaction and movement times are independently impaired in PD ( ). Reaction time is influenced not only by the degree of motor impairment but also the interaction between cognitive processing and the motor response. Patients with bradykinetic PD have more specific impairment in choice reaction time, which involves stimulus categorization and response selection, thereby reflecting more complex levels of cognitive processing, which can be affected in basal ganglia disorders, including PD. Of the various objective assessments of bradykinesia, movement time correlates best with the total clinical score, but it is not as sensitive an indicator of the overall motor deficit as is the clinical rating ( ). The premovement electroencephalographic (EEG) potential (Bereitschaftspotential) is reduced in PD, probably reflecting inadequate basal ganglia activation of the supplementary motor area ( ). On the basis of electromyographic (EMG) recordings in the antagonistic muscles of parkinsonian patients during a brief ballistic elbow flexion, concluded that the most characteristic feature of bradykinesia was the inability to energize the appropriate muscles to provide a sufficient rate of force required for the initiation and maintenance of a large, fast (ballistic) movement. PD patients need a series of multiple agonist bursts to accomplish a larger movement. Thus, the amount of EMG activity in PD is underscaled ( ). Although many patients with PD complain of “weakness,” this subjective symptom is probably due to a large number of factors, including bradykinesia, rigidity, fatigue, and reduced power as a result of muscle weakness, particularly when lifting heavy objects ( ).

Reduced dopaminergic function has been hypothesized to disrupt normal motor cortex activity, thereby leading to bradykinesia. Decreases in firing rates from single cortical neuronal recordings in free-moving rats correlated with haloperidol-induced bradykinesia, demonstrating that reduced dopamine action impairs the ability to generate movement and causes bradykinesia ( ). Of the various parkinsonian signs, bradykinesia correlates best with a reduction in the striatal fluorodopa uptake measured by positron emission tomography (PET) scans and in turn with nigral damage ( ). This is consistent with the findings that decreased density of SN neurons correlates with parkinsonism in the elderly, even without PD ( ). PET scans in PD patients have demonstrated decreased ¹⁸ F-fluorodeoxyglucose uptake in the striatum and accumbens-caudate complex roughly proportional to the degree of bradykinesia ( ). Nonhuman primate studies with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) monkeys provide evidence that bradykinesia results from excessive activity in the subthalamic nucleus (STN) and the internal segment of globus pallidus (GPi) ( ), a finding that has been identified in PD patients and may contribute to mechanisms underlying deep brain stimulation (DBS) ( ). Thus, there is both functional and biochemical evidence of increased activity in the outflow nuclei, particularly STN and GPi, in PD. Increased iron accumulation in the substantial nigra, putamen, and GPi has been demonstrated in high-pass filtered phase MRI in PD patients compared with controls. The iron deposition in the deep brain nuclei in the SN correlated with motor and bradykinesia-rigidity scores, but not with tremor scores, suggesting an in vivo reflection of dopaminergic neurodeneration ( ). Although bradykinesia correlates well with dopaminergic deficit, slow gait correlates better with cholinergic degeneration as demonstrated by [(11)C]dihydrotetrabenazine dopaminergic and [(11)C]methyl-4-piperidinyl propionate acetylcholinesterase (AChE) PET imaging ( ); the cholinergic system may be implicated in gait at the level of the brainstem with the pedunculopontine nucleus and the prefrontal areas, which may be involved in gait, attention, and cognition ( ).

Tremor

By using the term shaking palsy, James Parkinson in his An Essay on the Shaking Palsy (1817) drew attention to tremor as a characteristic feature of the disease later known as PD. Typical rest tremor has a frequency between 4 and 6 Hz and is almost always most prominent in the distal part of an extremity. In the hand, the tremor has been called a “pill-rolling tremor.” In the head region, tremor occurs most commonly in the lips, chin and jaw. Head tremor, a common manifestation of essential tremor (ET), is rare in PD ( ; ). Some patients with PD complain of an inner, not visible, tremor, sometimes called “internal tremor” ( ). Rest tremor of PD is often exacerbated during potential provocations, such as walking or mental tasks such as counting backward ( ).

Some PD patients may have a postural tremor that is more prominent and disabling than the classic rest tremor. Postural tremor without parkinsonian features and without any other known etiology is often diagnosed as ET, but isolated postural tremor may be the initial presentation of PD, and it may be found with higher-than-expected frequency in relatives of patients with PD ( ; ; ; ). Relatives of patients with tremor-dominant PD have a significantly higher risk for action tremor than do relatives of patients with the postural instability gait disorder (PIGD) form of PD, but it is not yet clear whether the isolated tremor in the relatives is truly ET or whether it represents an isolated manifestation of PD ( ; ). The two forms of postural tremor, ET and PD, can be differentiated by a delay in the onset of tremor when arms assume an outstretched position. Most patients with PD tremor have a latency of a few seconds (up to a minute) before the tremor reemerges during postural holding, hence the term reemergent tremor ( ; ) ( Video 4.1 ). In contrast, the postural tremor of ET usually appears immediately after arms assume an outstretched posture. Because the reemergent tremor has a frequency similar to that of rest tremor and both tremors generally respond to dopaminergic drugs, reemergent tremor most likely represents a variant of the more typical rest tremor in PD. Some PD patients will also have a kinetic tremor, possibly related to enhanced physiologic tremor, that may also impair normal reach-to-grasp movement ( ).

Video 4.1 Reemergent tremor.

Whereas noted that 100% of 30 patients with pathologically proven PD experienced some degree of rest tremor at some time during the course of their disease, other studies demonstrate that not all PD patients have rest tremor, with estimates about 35%. In another clinicopathologic study, only 76% of pathologically proven cases of PD had tremor ( ) and in an expanded series of 100 pathologically proven cases of PD, tremor was present at onset in 69% and in 75% at some time during the course of PD ( ). In some, such as 9% of those studied by , tremor may lessen as disease progresses and some may have more prominent postural or action tremors.

Presence of tremor as the initial symptom, or “tremor-predominant” PD, in contrast to the PIGD subtype of PD, often confers a favorable prognosis with slower progression of the disease. This is supported by a large number of studies. Indeed, one study addressing long-term course showed that the presence of tremor at onset was associated with much lower hazard for both motor and axial complications ( ). Some have suggested the term “benign tremulous parkinsonism” for a subset of patients with minimal progression, frequent family history of tremor, and poor response to levodopa ( ; ; ; ). In 16 pathologically studied cases, those with benign tremulous parkinsonism had significantly less severe neuronal loss in the SN than controls ( ). Of these, 12 had been correctly diagnosed with PD at their first neurologic evaluation, whereas the other 4 were originally thought to have another tremor disorder. Among the five cases that did not have PD pathologic conditions, none developed bradykinesia within the first 10 years, and the likely clinical diagnoses were ET with associated rest tremor or dystonic tremor. Others have argued against the term “benign tremulous parkinsonism” and proposed the term “monosymptomatic tremor at rest” because some symptoms are not experienced as “benign” and may still affect functioning ( ). Furthermore, imaging documents evidence of dopaminergic deficits associated with PD and parkinsonism.

The pathophysiology of PD rest tremor is complex and most likely results from dysfunction of both the striato-pallidal-thalamocortical and the cerebello-dentato-thalamocortical circuits ( ). The pallidum, in particular, appears to play a fundamental role in generation of tremor as suggested by a 4 to 8 Hz GPi neuronal firing in primate models of parkinsonism, correlation of tremor severity with pallidal (but not striatal) dopamine depletion, and complete abolition or a marked improvement of tremor with GPi ablation or DBS ( ). As a result of the abnormal neuronal activity at the level of the GPi, the muscle discharge in patients with PD changes from the normal high (40 Hz) to pulsatile (10 Hz) contractions. These muscle discharges, which may be viewed as another form of PD-associated tremor, can be auscultated with a stethoscope ( ). Evidence from functional neurosurgical interventions for PD tremor also suggests contributions from excessive inhibitory output to the thalamus and altered firing patterns, leading to rhythmic burst activitiy in thalamic cells and role for cerebellar modulation of tremor ( ). These authors propose a model such that the basal ganglia activity induces tremor, the thalamus generates PD tremor oscillations, and the cerebellum modulates tremor as if it were similar to a voluntary movement; in this model the role of the cerebellum is not its participation in the cerebello-thalamic circuit generating tremor ( ) but rather a modulator that compares an efferent copy of a voluntary alternating movement (i.e., tremor) to incoming peripheral sensory information.

Rigidity

Rigidity refers to the resistance detected when passively moving a joint throughout the range of movement. In contrast to spasticity, rigidity is velocity-independent and often called “lead-pipe” in quality. Cogwheeling is often encountered, particularly if there is associated tremor or an underlying, not yet visible, tremor. Historical references to parkinsonian rigidity include the observation of enhanced resistance to passive movement of a limb about a joint detected during voluntary movement of a contralateral limb (“Froment’s maneuver”), which was based on a series initially published by Froment and Gardere in 1926 ( ). Rigidity may occur proximally (e.g., neck, shoulders, and hips) and distally (e.g., wrists and ankles). At times, it can cause discomfort and pain. Painful shoulder, possibly as a result of rigidity but frequently misdiagnosed as arthritis, bursitis, or rotator cuff, is a frequent initial manifestation of PD ( ; ). Subjective complaints of stiffness and imbalance can occur prior before overt parkinsonian signs and symptoms; those with these features in a longitudinal study (n = 6038, mean age 68.5 years) had an increased risk for developing PD with a HR of 2.11, 2.09, and 3.47, respectively ( ).

Dystonia

Although dystonia is not considered one of the cardinal signs of PD, it is often present in patients with PD ( ; ). Dystonia is present in over 30% of PD patients and precedes the onset of other parkinsonian motor features. Foot dystonia associated with flexion of the toes and inversion of the foot may be the initial manifestation of PD. Furthermore, dystonia may be a feature of levodopa-related motor complications and may include wearing off dystonia or cervical, truncal, or hand dystonia as part of peak dose dyskinesia ( ; ). Other examples of dystonia-parkinsonism overlap is in some patients with genetic dystonia and tardive dystonia coupled with drug-induced parkinsonism. Furthermore, DBS used in the treatment of PD can cause dystonia such as blepharospasm ( ).

Flexed posture and other skeletal abnormalities

Although rigidity is often associated with postural deformity resulting in flexed neck and trunk posture and flexed elbows and knees, flexed posture and other skeletal abnormalities can occur independently. Some patients develop “striatal hand” deformity, characterized by ulnar deviation of hands, flexion of the metacarpophalangeal joints, and extension of the interphalangeal joints ( ; ; , ; ) ( Fig. 4.1 ). Striatal changes in the feet can also occur with extension of the big toe (“striatal toe”) or flexion of the other toes, which can be confused with arthritis. Striatal toe was present in 13 of 62 (21%) of patients with clinically diagnosed PD ( ). This phenomenon can be seen not only in PD but also in other parkinsonian disorders. Moreover, it should be differentiated from psychogenic toe, which usually manifests by increased resistance to passive manipulation, pain, and spontaneous toe plantar flexion with forced extension of the second to fifth toes ( ).

Other skeletal abnormalities can occur in the neck and spine. These can be considered by body region (neck and spine) but also in the sagittal plane (e.g., antecollis and camptocormia) or coronal plane (scoliosis and Pisa syndrome). The prevalence of scoliosis, defined as lateral flexion not relieved by voluntary or passive movement and lateral curvature of the spine of at least 10 degrees, is reported in 9% to 91% of PD, whereas the prevalence of antecollis is about 5% to 6%; camptocormia, about 3% to 18%; and Pisa syndrome, about 2% ( ; ). Extreme neck flexion (“dropped head”) has been found in PD but also other parkinsonian disorders such as multiple system atrophy (MSA) ( ; ; ; ; ; ; ; ). When neck flexion posture is present, neck extensor weakness should be differentiated from dystonic anterocollis in which muscle spasms or pulling of anterior neck muscles may be seen. Besides in parkinsonian disorders, head drop or neck flexion can be seen as an isolated syndrome such as neck extension weakness or in neuromuscular disorders with myopathic changes ( ; ). Some parkinsonian patients with dropped head syndrome also can have these features and evidence of myopathy on EMG, blood tests, or imaging ( ). In 7 of 459 patients with parkinsonism with dropped head syndrome because of neck extensor weakness, myopathic changes on EMG were noted in all, with 2 of 5 undergoing muscle biopsy with mitochondrial abnormalities ( ); other studies find variable rates of EMG changes. Isolated neck and trunk extensor myopathy has been reported in other patients with anterocollis and camptocormia associated with parkinsonism ( ; ; ; ; ; ; ). Other series have identified various etiologies in patients with head drop: dystonia, disproportionately increased tone in the anterior neck muscles resulting in fibrotic and myopathic changes, amyotrophic lateral sclerosis, focal myopathy, inclusion body myositis, polymyositis, nemaline myopathy, facioscapulohumeral dystrophy, myasthenia gravis, encephalitis, dopamine agonists ( ), and valproate toxicity ( ; ; ; ; ). Evaluation may include EMG, laboratory tests for muscle damage or inflammation, and cervical spine MRI in which the muscle tissue may appear to have inflammatory changes. Treatments may involve increased dopaminergic medications but also frequently engage physical therapy and bracing support.

Postural abnormalities affecting the trunk can involve anteroposterior and/or lateral trunk flexion. Lateral trunk flexion or scoliosis can be frequently seen in PD, sometimes occurring as part of the disease, in addition to a prior scoliosis, or as a subacute phenomenon associated with medication adjustments ( ; ). Camptocormia is characterized by extreme flexion of the thoracolumbar spine that has a minimum of 45 degrees flexion in the thoracolumbar region and increases during walking but resolves in supine position ( Video 4.2, Video 4.3, Video 4.4 ) ( Fig. 4.2 ) ( ; ; ; ; ; ; ; ; ; ; ; ).

Video 4.2 Camptocormia.

Video 4.3 Camptocormia.

Video 4.4 Camptocormia.

Camptocormia appears to be more common in patients with more severe PD symptoms and prior vertebral surgery ( ). The term was coined during World War I when young soldiers who were apparently attempting to escape the stress of battle developed this peculiar posture, perhaps promoted by a stooped posture when walking in the trenches. Hypothesized mechanisms for camptocormia include dystonia resulting from a central disorder and extensor truncal muscle myopathy ( ; ; ; ; ). Although myopathy may cause camptocormia, this is probably a very rare pathogenic mechanism of abnormal axial posture in PD because there is usually no demonstrable weakness but instead there is often increased muscle tone and activity, and the reported EMG, imaging, and histologic changes are relatively nonspecific ( ; ).

Pisa syndrome refers to lateral flexion with a minimum of 15 degrees alleviated by passive mobilization or supine positioning ( ; ; ; ) ( Fig. 4.3 ). In most patients, the tilt is away from the original symptom onset and is accompanied by greater EMG activity of the paravertebral thoracic muscles, but at the lumbar level there can be more atrophy, ipsilateral to the deviation, possibly because of muscle misuse ( ). The pathophysiology of Pisa syndrome is not well understood but may relate to features underlying the other described postural abnormalities and cholinergic-dopaminergic imbalance. Postural deformities such as striatal hand and foot deformities, antecollis, Pisa syndrome, scoliosis, and camptocormia are most likely of multifactorial pathophysiology that includes rigidity, dystonia, weakness caused by myopathy, body scheme defects resulting from centrally impaired proprioception, and structural changes in the spine ( ; ). In one study, in which Pisa syndrome was detected in 143 patients (8/8%), it was associated with older age, lower body mass index, longer PD duration, worse disease stage, and poorer quality of life ( ).

Treatments for these spinal deformities include adjustments of dopaminergic medications, botuliunum toxin injections in some cases, DBS, and physical therapy. The axial dystonias resulting in scoliosis and camptocormia may improve with botulinum toxin injections into the paraspinal or rectus abdominus muscles ( ; ). Other treatments include physical therapy, orthopedic procedures ( ), and even DBS ( ). In a study of 78 PD patients with camptocormia and 78 PD patients without camptocormia, the camptocormia group had significantly higher prevalence of compression fractures, more severe parkinsonian symptoms, greater incidence of dementia, and higher serum creatine kinase than those without camptocormia ( ).

Gait disorder and freezing

Parkinsonian gaits are characterized chiefly by the combination of shuffling steps, start-hesitation, and freezing (as if the feet were glued to the floor) associated with stooped posture, flexed knees, narrow base, and turning en bloc ( ). In early PD, common features include shortened stride length, asymmetry of stride length, decreased heel strike, decreased toe clearance, and reduced arm swing. PD patients may also have festination of gait in which their center of gravity is shifted forward or ahead of their feet, which leads to propulsion or forward acceleration. Some PD patients exhibit greater PIGD subtypes, in which gait and balance may be particularly affected. As PD advances, gait abnormalities may further affect balance and risk for falls. In the assessment of gait and posture, the examiner should observe the pattern of movement of the whole body when the patient walks and turns, etc. Although arm swing is not a separate item on the UPDRS, arm swing asymmetry may be the earliest feature of PD and has been found to be a potential prodromal marker in otherwise asymptomatic LRRK2 mutation carriers ( ).

One of the most disabling symptoms of PD is freezing of gait, also referred as motor blocks, considered by some as a form of akinesia (loss of movement) ( , ; ; ; ; ; ) ( Videos 4.5 and 4.6 ). Although it most often affects the legs when walking, it can also involve upper limbs and the eyelids (apraxia of eyelid opening or eyelid closure) ( ). Freezing consists of sudden, transient (a few seconds) inability to move. It typically causes start-hesitation when initiating walking and the sudden inability to move feet (as if glued to the ground) when turning or walking through narrow passages (e.g., the door or the elevator) ( ), when crossing streets with heavy traffic, or when approaching a destination (target hesitation). One study showed that the most efficient way to objectively ascertain freezing of gait is to ask patients to repeatedly make rapid 360-degree narrow turns from standstill, on the spot and in both directions ( ). Freezing is a common cause of falls in patients with PD that can result in injuries, including hip fractures. Recognition of freezing is important because it denotes poor prognosis; most patients become wheelchair bound within 5 years after onset of freezing ( ). Based on responses by 6620 patients to a questionnaire sent to 12,000 members of the German Parkinson Association, 47% of patients reported freezing, present more frequently in men than women and less frequently in patients who considered tremor as their main symptom ( ).

Video 4.5 Freezing.

Video 4.6 Freezing.

Freezing of gait can have multiple components to it, such as motor control, medication effects, certain situations, and cognitive and emotional aspects. Situational freezing of gait tends to occur with gait initiation, turning, navigating environmental obstacles, or when dual tasking (or multitasking). Freezing may be a manifestation of the “off” phenomenon in PD patients who fluctuate but may also occur during “on” time (on freezing), independent of bradykinesia and tremor ( ; ). Off gait freezing correlates with dopa-responsive abnormal discriminatory processing as determined by abnormally increased temporal discrimination threshold ( ). On gait freezing can have several different characteristics: (1) a trembling freezing of gate that worsened with increasing levodopa doses; (2) new on FOG without predating “off” freezing of gait; (3) concurrent worsening (freezing) of repetitive alternating hand movements in the on state; (4) good levodopa responsiveness of other parkinsonian features; and (5) preserved cognition, unlike the frontal cognitive impairment associated with “off” freezing of gate ( ). The timing of motor state or certain situations associated with freezing of gate should be considered in the examination and management of freezing.

When freezing occurs early in the course of the disease or is the predominant symptom, a diagnosis other than PD should be considered. Disorders associated with prominent freezing include progressive supranuclear palsy (PSP), MSA, and vascular (lower body) parkinsonism ( ; ; ; ; ). Freezing often correlates with frontal executive deficits and atrophy in the dorsolateral prefrontal, medial, and lateral temporal, inferior parietal, and occipital cortices ( ). Isolated freezing usually suggests a diagnosis other than PD and may be present in atypical forms of parkinsonism or brainstem strokes ( ). The pathologic involvement of brainstem in patients with pure akinesia and gait freezing is suggested by decreased glucose metabolism on PET scans in the midbrain of such patients, similar to the findings in patients with PSP ( ).

The pathophysiology of freezing of gait may relate to neurodegeneration in certain brain areas, disconnection between brain regions involved in motor and cognitive control, and possibly peripheral muscle abnormalities ( ). Gait freezing in PD may relate to noradrenergic deficiency as a result of locus coeruleus (LC) degeneration ( ) and possible response to noradrenergic agents such as l-threo-dihydroxy-phenylserine, or DOPS ( ). Neurophysiologic studies in monkeys treated with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine found that dopamine depletion is associated with impaired selection of proprioceptive inputs in the supplementary motor area, which could interfere with motor planning and may be related to motor freezing ( ). Integrating EMG signals over real time while recording EMG activity from lower extremities before and during freezing, showed significantly abnormal timing in the tibialis anterior and gastrocnemius muscles, although reciprocity is preserved. Thus, before freezing, the tibialis anterior and gastrocnemius contract prematurely, and the duration of contraction is shortened in the tibialis anterior, but the amplitude of the EMG burst is increased (probably a compensatory strategy pulling the leg into swing), whereas the contraction is prolonged in the gastrocnemius during the actual swing phase. Other studies suggest transient overactivation in the inhibitory striatal output nuclei projecting to the motor thalamus and brainstem regions ( ).

Gait freezing has been associated with reduced functional connectivity in the opercular parietal cortex and other brain areas ( ), abnormalities within both “executive-attention” and visual networks ( ), and increased corticopontine projections crossing at the pons are associated with freezing of gait ( ). In a clinicopathologic study of 58 patients with pathologically confirmed PD and clinical evidence of freezing, greater severity of gait freezing correlated with the density of cortical Lewy body–containing neurons ( ). Using a virtual reality paradigm to elicit freezing of gait behavior in PD participants during fMRI, alterations in functional networks that capture different associated elements of gait freezing (cognitive, motor, and affective function) have been detected in PD freezers ( ). Coupling between the cognitive and limbic networks was associated with “worse freezing severity”, whereas anticoupling between the putamen and the cognitive and limbic networks was related to “increased compensation.” Additionally, anticoupling between cognitive cortical regions and the caudate nucleus were “independent of freezing severity” and thus may represent common neural underpinnings of freezing that are unaffected by heterogenous factors.

Various therapeutic strategies have been studied for freezing of gait, including medical, surgical, and rehabilitative management. If gait freezing is related to the off state or suboptimal motor states, dopaminergic medications may be used. However, gait freezing may relate to nondopaminergic systems, and, thus, studies have focused on these systems as targets. Medical treatments with DOPS, selegiline, idazoxan, atomoxetine, and duloxetine (Cymbalta) ( ; ; ) and surgical approaches ( ; ) rarely provide satisfactory control of freezing. Other studies have included rivastigmine (negative results) ( ), amantadine (inconclusive results ( ; ), methyphenidate (no improvement and on one study, reduced gait freezing only in those with PD with STN-DBS) ( ; ), caffeine (insufficient results) ( ), and istradefylline (reduction in small open label study) ). Other surgical interventions that have generated interest include electrical spinal cord stimulation, with open label case reports to date; one study found a 56% improvement in subjective freezing of gait along with improved objective gait measures after high-frequency (300 Hz) spinal cord stimulation at the upper thoracic (T2-T4) level in four PD patients with prior bilateral STN-DBS but gait freezing; beneficial effects of spinal cord stimulation on freezing remained for the 6 months duration of the study ( ). Patients often adopt a variety of cues or tricks to overcome the freezing attacks: marching to command (“left, right, left, right”), stepping over objects (the end of a walking stick, a pavement stone, cracks in the floor, etc.), walking to music or a metronome, shifting body weight, rocking movements, and others ( ; ; ; ; ; ).

Loss of postural reflexes

Loss of postural reflexes is a characteristic feature in PD patients more typically in advanced disease stages but also those who exhibit the PIGD phenotype. Loss of postural reflexes and also of protective reactions can lead to falls and fall-related injuries in PD. Several studies have examined clinical features or risk factors associated with falls in PD. Compared with controls, those PD patients with falls (“PD fallers”) have a tendency to overestimate balance performance on functional reach testing, particularly with worsened disease, are willing to sacrifice motor performance to complete tasks, and make significantly more motor errors when performing a complex motor-cognitive task (e.g., carrying a tray and performing mental arithmetic) ( ). Abnormal axial posture, cognitive impairment, and freezing of gait were independent risk factors for falls and predicted 38 of 51 fallers (75%) and 45 of 62 nonfallers (73%) ( ). Some studies suggest that frontal impairment, poor leaning balance, and leg weakness contribute to falls in PD, whereas other studies cite female gender, symmetric onset, and postural and autonomic instability as the most reliable predictors of falls in PD ( ). Predictors of falls with high sensitivity (78%) and specificity (84%) include UPDRS total score, total gait freezing score, symptomatic postural orthostasis, Tinetti total score, and extent of postural sway in the anteroposterior direction.

The average period from symptom onset to first fall in PD is 108 months, which is longer than reported for PSP (16.8 months), vascular parkinsonism (40.8 months), MSA (42 months), and dementia with Lewy bodies (DLB) (54 months). Using a battery of neurologic and functional tests in 101 patients with early PD and in an optimally medicated state, 48% reported a fall and 24% more than one fall in a prospective follow-up over 6 months ( ). Many PD patients with postural instability, particularly when associated with flexed truncal posture (camptocormia), have festination of gait, which can contribute to falls. The pull test (pulling the patient by the shoulders) is commonly used to determine the degree of patient’s retropulsion or propulsion and is part of the UPDRS and MDS-UPDRS assessment ( ; ; ).

Balance impairment and falls in PD may be related to cholinergic system alterations to a greater extent than dopaminergic neurotransmission ( ; ; ). Evidence supports involvement of the cholinergic system–mediated higher level cortical and subcortical processing, including pedunculopontine nucleus function, in gait control. In a study of 44 nondemented patients with PD and 15 controls who underwent a clinical assessment and [(11)C]methyl-4-piperidinyl propionate (PMP) AChE and [(11)C]dihydrotetrabenazine (DTBZ) vesicular monoamine transporter type 2 (VMAT2) brain PET imaging, PD fallers compared with PD nonfallers and controls had significantly reduced cortical AChE hydrolysis rates (–12.3% versus –6.6%) ( ). Although subcortical and brainstem cholinergic changes have been found in PD, they are greater in those with atypical parkinsonian syndromes; for example, a [(11)C]PMP PET study in PD (n = 12), MSA-parkinsonian type (MSA-P, n = 13), PSP (n = 4), and controls (n = 22) demonstrated a significant decrease in AChE activity in most cerebral cortical regions in PD and MSA-P, but a nonsignificant decrease in PSP, whereas there were greater decreases in subcortical cholinergic activity in MSA-P and PSP than in PD ( ). Greater decreases in subcortical AChE in MSA-P and PSP, such as in the pedunculopontine nucleus, may account for the earlier and more prominent gait disturbances in these two disorders compared with PD ( ). Using conditioning motor evoked responses elicited by transcranial magnetic stimulation of the motor cortex, significant reduction was found in short latency afferent inhibition, a surrogate measure of cholinergic activity, in patients with PD ( ).

Treatment of falls should include prevention and assessment of fall risk, home safety, and medical comorbidities that may contribute to fall risk ( ). A meta-analysis including five short-term studies (n = 620 participants) and six long-term studies (n = 590) showed that exercise-based interventions can provide benefit in both the short term and long term ( ). Exercise interventions in trials demonstrating reduction in fall risk include tai chi and other movement and strengthening strategies. Medication interventions may include adjustments of PD medications. Given the relationship between the cholinergic system and falls, rivastigmine was studied in a controlled trial and demonstrated reduced falls in PD fallers by 45% compared with the placebo group ( ); however, there was no effect on gait freezing, cognition, or executive function. Droxidopa demonstrated reduced fall risk in PD patients with orthostatic hypotension compared with placebo ( ). Thus, the role of orthostatic hypotension in leading to falls, or unexplained falls, in PD should be considered.