Clinical manifestations

Presentation and course

PSP-Richardson syndrome. In the classic form of progressive supranuclear palsy (PSP-Richardson syndrome) there is unsteady gait, postural instability with falls, vertical gaze palsy, parkinsonism, cognitive slowing, and executive dysfunction (191). Research criteria for diagnosis of Richardson syndrome require that onset is at least 40 years of age, progression is gradual, and vertical gaze palsy is present; postural instability and repeated falls begin within the first two years (191; 81). In one series, the mean age of onset was 65.9 years (08). Cognitive and behavioral problems were identified in 89% of these patients at the time of presentation (3.0 years, on average). Cases of PSP-Richardson syndrome made up about 54% of all patients with supranuclear palsy in one large series (188), 24% in a second series (145), and 40% in a more recent series of PSP cases (138).

PSP-parkinsonism. The variant called “PSP-parkinsonism” is characterized by asymmetrical bradykinesia and rigidity (sometimes with rest tremors) that is responsive to levodopa for the first one to two years. It accounted for 32% of all supranuclear palsy cases in one early series (188) and 36% in a recent series of cases (138). PSP-parkinsonism includes dysarthria and gait disorder, but there is usually an absence of gaze palsy in the early years. PSP-parkinsonism has a more slowly progressing course than PSP-Richardson syndrome; at autopsy there is less cortical pathology and a reduced tau burden (189; 41; 159; 167). At their initial office visits, cognitive test scores are significantly higher in patients with PSP-parkinsonism than in patients with PSP-Richardson (144). In a prospective natural history study, the estimated mean survival time was 11.2 years for PSP-parkinsonism and 6.8 years for Richardson syndrome (87).

Table 1. Features That Distinguish the Common PSP Variants

	PSP-Richardson syndrome	PSP-Parkinsonism	PSP-Pure akinesia	PSP-Speech/language	PSP-Corticobasal syndrome
Early signs and symptoms	Falls, cognitive problems, postural instability, and gaze palsy	Dysarthria, gait disorder	Slowness, gait freezing	Non-fluent, agrammatic speech	Apraxia and parietal lobe signs before gaze palsy
% of all PSP cases	24% to 54%	32% to 36%	4%	5%	12%
Response to levodopa	-	+	-	-	-
Mean survival time	6.8 yrs	11.2 yrs	11 yrs	-	-
(181; 188; 89; 87; 145; 138)

Other PSP variants. Patients with the PSP-pure akinesia with gait freezing variant present with slowness and gait freezing for many years before they show any signs of gaze palsy or cognitive impairment; their mean survival was 11 years, more like that seen in PSP-parkinsonism (190; 132). They are different from patients with PSP-parkinsonism in that they have no response to levodopa, and they are very rare (4% of all PSP cases) (138). In a pathologic series of 22 patients with supranuclear palsy, 10 had clinical signs of PSP-Richardson syndrome, eight had PSP-parkinsonism, and four presented with PSP-cerebellar syndrome (92; 93). Although the latter is a rare variant, it may result in misdiagnosis as multiple systems atrophy (98). There is yet another subgroup of patients with supranuclear palsy, PSP-speech/language variant, which presents with a non-fluent language disorder that is characterized by slow, hesitant, agrammatic speech with phonemic errors, as well as apraxia of speech (89; 156; 186). This speech/language variant of PSP is also very rare, with only 5% of cases being represented in a recent series of PSP (138). Patients with PSP-behavioral variant (rare at 4%) present with personality changes, dementia, and loss of executive function before the signs of supranuclear palsy appear (138). Patients with PSP-corticobasal syndrome (12% of all PSP cases) manifest initially with parietal lobe signs, such as asymmetric motor apraxia or the alien limb phenomenon, before they develop falls or supranuclear gaze palsy (181; 138). When the new Movement Disorder Society-PSP diagnostic criteria of Hoglinger and others were applied to a group of 222 patients with atypical parkinsonian syndromes (81), the number of patients meeting criteria for progressive supranuclear palsy doubled, compared to those who had previously received the diagnosis (05; 85).

Executive dysfunction. The main cognitive deficits in supranuclear palsy are those of attention and executive function (148). Dysfunction of the dorsolateral prefrontal circuits appears particularly early in PSP-Richardson syndrome and is manifested by decreased verbal fluency, difficulty with planning and organization, increased central processing time, and loss of mental flexibility (147; 155). Patients answer questions and solve the simplest problems only after long delays. Cognitive slowing can be documented by using Trail Making Test B, Victoria Stroop, or verbal fluency tests (147; 106). Generally speaking, the executive dysfunction of supranuclear palsy is much more severe than that observed in other subcortical movement disorders, such as Parkinson disease, dementia with Lewy bodies, corticobasal degeneration, or multiple system atrophy (155). Executive function and processing speed are more impaired in PSP-Richardson syndrome than they are in PSP-parkinsonism or PSP-pure akinesia with freezing gait (106).

Giordano and colleagues have shown a significant correlation between frontocerebellar gray matter atrophy and executive impairment in patients with progressive supranuclear palsy, suggesting that cerebellar atrophy plays an important role in the pathogenesis of cognitive dysfunction in this disease (63). In a related study using MRI with diffusion tensor (DT) imaging with tract-based spatial statistical analysis, a significant association was found between frontocerebellar white matter loss and executive cognitive impairment in progressive supranuclear palsy (174). In contrast, another study using DT MRI showed that metrics of the corpus callosum, right superior longitudinal and inferior longitudinal fasciculus, and left uncinate were the best predictors of executive dysfunction in progressive supranuclear palsy (02). Kim and colleagues found that recurrent falls in patients with supranuclear palsy were significantly associated with executive dysfunction, as measured clinically by alternating hand movements and the Stroop color interference test (97).

Memory storage. Aarsland and colleagues used the Mattis dementia rating scale to show that patients with supranuclear palsy had better memory function than those with Parkinson disease or dementia with Lewy bodies (01). Donnelly and colleagues compared patients with early supranuclear palsy to those with early Parkinson disease using the logical memory test of the Weschler scale, as well as the New Dot test (45). They found similar memory functions in these two groups of patients. These findings were consistent with those of Pillon and colleagues, who argued that memory storage was only minimally impaired due to damage to fronto-striatal circuits, rather than hippocampal-limbic circuits (139).

Memory retrieval. Several groups have identified memory retrieval problems in patients with supranuclear palsy. Donnelly’s group was able to document this retrieval problem in patients with PSP-Richardson syndrome by using the Selective Reminding Test (45). Pillon and others demonstrated that when memory encoding is facilitated by semantic category cues, memory retrieval is improved (139). By using this cuing procedure, recall performance of patients with supranuclear palsy improves, and the total recall scores become similar to those of age-matched controls. On the other hand, memory performance remains severely impaired in most patients with Alzheimer disease, as few of them are sensitive to semantic cues. Thus, there is no genuine amnesia in supranuclear palsy, but rather difficulty in accessing stored information. The dissociation between normal storage and impaired retrieval has also been reported in Parkinson and Huntington diseases (139).

Procedural learning. Procedural learning, which is measured with eye blink classical conditioning, is impaired in progressive supranuclear palsy (165). The main circuit involved in procedural learning is the cortex-basal ganglia-thalamus-cortex loop. In the Serial Reaction Time Task, no significant learning was observed in response to a repeated sequence in patients with supranuclear palsy (PSP-Richardson syndrome), in contrast to patients with Alzheimer disease. It should be noted, however, that patients with progressive supranuclear palsy were able to press the correct key in only 50% of the trials; their impaired performance in the Serial Reaction Time Task might result from too great a complexity of this motor task, as well as from impaired procedural learning ability.

Behavioral and emotional changes. Behavioral abnormalities are common in patients with supranuclear palsy, as more than half experience apathy, depression, and sleeping disturbances; approximately one third exhibit agitation, irritability, disinhibition, and eating difficulties (58; 144; 155). Frontal lobe behavioral symptoms, such as apathy, loss of spontaneity, and inflexibility, are more prevalent among patients with supranuclear palsy than among those with Parkinson disease or corticobasal degeneration (155). Apathy is often paradoxically accompanied by impulsivity (28; 102) and is associated with atrophy of the ventromedial frontal cortex, fronto-striatal cortex, and the left insula in supranuclear palsy, corticobasal syndrome, and Alzheimer disease (172; 102). Apathy in progressive supranuclear palsy has also been related to damage in white matter tracts, specifically the corpus callosum, right superior longitudinal, and uncinate fasciculi, as assessed by diffusion tensor MRI (02). Stereotypies (repetitive, ritualistic movements) are more common in behavioral variant frontotemporal dementia, supranuclear palsy, and Parkinson disease with dementia than in Alzheimer disease, probably because of the dysfunctioning of fronto-striatal circuits in these three diseases (142). Alexithymia, the inability to recognize emotions in others, is identifiable early in the course of both PSP-Richardson syndrome and PSP-parkinsonism, and it differentiates progressive supranuclear palsy from Parkinson disease within the first 2 years (10). Disorganized behaviors can be seen in patients with progressive supranuclear palsy in the mild cognitive impairment stage (about the same time that these behaviors are observed in Alzheimer patients, but slightly later in the disease course than they are observed in those with behavioral variant frontotemporal dementia) (178).

Depression and apathy. Some authors report that depression is the most common psychiatric disturbance in patients with progressive supranuclear palsy (20), whereas others have found that apathy is the most prevalent (144; 102). A study showed that both apathy and depression were the most common psychiatric symptoms in a series of 59 patients with supranuclear palsy (88). Even though apathy and depression may coexist, they are two distinct constructs (147). A 1-year prospective study of 46 patients with supranuclear palsy showed that apathy had significant negative predictive value for health-related quality of life (135). Higginson and colleagues reported that in a late-stage group of patients with supranuclear palsy, about half of the sample had severe depression, or anxiety, or both (78). When frontotemporal lobar syndromes were compared with one another (n=124 patients), apathy scores were shown to be higher among those with supranuclear palsy, compared to those with behavioral variant frontotemporal dementia or primary progressive aphasia (104).

Agitation and irritability. Agitation and irritability occur at the same rate in supranuclear palsy as in Parkinson disease, but they are less frequent than in Huntington or Alzheimer diseases. Nevertheless, agitation, irritability, and other behavioral changes may be the earliest clinical manifestations of PSP-Richardson syndrome (72). Overall, patients with PSP-Richardson syndrome show more neurobehavioral symptoms than those who have the PSP-parkinsonism variant (167). The sundown syndrome is a descriptive term that includes agitation, as well as confusion, pacing, and wandering. It occurs in the afternoon or evening in patients who have supranuclear palsy, as well as a variety of other neurodegenerative diseases (30).

Obsessive compulsive disorder. Excessive thoughts (obsessions) and repetitive behaviors (compulsions) were observed in 24% in one series of 74 patients with supranuclear palsy (55). In another study, about 33% of patients with supranuclear palsy met criteria for obsessive-compulsive personality disorder, compared to 13% of those with multiple systems atrophy (126).

Pseudobulbar affect. Emotional lability is a common symptom in patients with supranuclear palsy. It is characterized by relatively brief outbursts of emotion, such as laughing or crying (147). Pseudobulbar affect is a disorder of affect, whereas depression is a disorder of mood (depressed patients usually have a flattened affect, and they hide feelings of hopelessness and helplessness). Patients with supranuclear palsy can have both depression and pseudobulbar affect.

Speech and language disorders. Progressive apraxia of speech, nonfluent aphasia, speech timing disruptions, and agrammatic speech are all clinical features seen early on in the PSP-speech/language variant (89; 81; 186; 57). Word-finding difficulty and slowed articulation rate may be a problem for some patients with supranuclear palsy, but it is less severe than that seen in Alzheimer disease or corticobasal syndrome. Within the first 2 years of following an older cohort of 52 patients with progressive apraxia of speech, motor signs of PSP-Richardson syndrome began to develop in 38% of the cohort (186). Those who developed motor signs of Richardson syndrome were older (mean age =73 years), and falls were observed more frequently than vertical gaze palsy.

Sleep disorders. Rapid eye movement (REM) sleep behavior disorder is a parasomnia that is an important preclinical sign of synuclein-mediated neurodegenerative diseases, such as Parkinson disease and dementia with Lewy bodies (141), but REM sleep disorder is also present in 33% to 37% of patients with supranuclear palsy (09; 122). The patients with supranuclear palsy had lower values for total sleep time and sleep efficiency on polysomnogram. Differences in brainstem pathology were thought to influence these findings.

Social cognitive deficits. Patients with supranuclear palsy show impairments on voice emotion recognition and theory of mind tests, deficits of which correlate with grey matter atrophy in the inferior frontal gyrus and anterior rostral medial frontal cortex (60; 162). Both patients with frontotemporal dementia and those with supranuclear palsy have impaired ability to represent others’ opinions and intentions (theory of mind) (162), and changes in the salience network occur earlier in these two diseases than they do in other neurodegenerative diseases (177). Pontieri and colleagues used the Penn emotion recognition test to assess facial emotional recognition abilities in patients with supranuclear palsy, compared to an age-matched group with Parkinson disease (140). Patients with supranuclear palsy were significantly impaired in their ability to recognize sad or happy faces, compared to those with Parkinson disease. This fits with the previous literature suggesting social cognitive deficits in this disease. Patients with supranuclear palsy also have difficulty in verbally describing their own feelings (alexithymia), a finding that is identifiable very early on, occurs at a higher frequency than is seen in Parkinson disease, and seems to be predicted by the presence of depression (10).

Prognosis and complications

Progression of cognitive impairment to dementia. Cognitive progression is more dramatic and rapid in PSP-Richardson syndrome than in Parkinson disease or multiple system atrophy. In one study, executive, language, and visuospatial abilities declined over a period of 15 months in PSP-Richardson patients, but not in patients with Parkinson disease or multiple system atrophy (52). In that study, 16% of patients with PSP-Richardson had converted to dementia after 18 months from the time of diagnosis.

Progression of disease and predictors of mortality. Supranuclear palsy is an aggressive neurologic condition that commonly leads to death within 6 to 11 years of symptom onset (34; 87). Progression of illness is slower in those who have PSP-parkinsonism and PSP-pure akinesia with gait freezing (188; 87). In PSP-Richardson syndrome, older age of onset, male gender, early falls, early dysphagia, and early dementia are all clinical predictors of shorter disease duration (34; 39; 111; 85). The presence of sleep disturbances and hallucinations have been associated with an earlier death in some patients with supranuclear palsy (08). Others have shown that higher apathy scores are associated with earlier mortality in supranuclear palsy and other frontal lobe syndromes (104). Patients with supranuclear palsy and higher plasma neurofilament light chain levels have more severe neurologic, functional, and neuropsychological deterioration over one year, suggesting this could be used as a biomarker in clinical trials (151). Severe depression, poor baseline cognitive scores, earlier urinary symptoms, greater baseline flortaucipir (PET) uptake, and smaller midbrain volume on MRI are all predictors for a more rapid clinical decline in patients with progressive supranuclear palsy (13; 180; 129; 134). Another study showed that progression was more rapid in patients who had more damage to subcortical white matter tracts, as measured by diffusion tensor imaging (196).

Clinical vignette

A 69-year-old married man and former building inspector was evaluated in an outpatient clinic. When he was first examined, balance and gait problems had been slowly progressing over the last three years. The patient noted that his mind had become “cloudy” over the same period of time. He lacked initiative and was unable to engage in any new projects, according to his wife. His only remaining hobby was television. His wife assumed the responsibilities of driving, financial management, meal preparation, and shopping. He eventually needed help with bathing, dressing, feeding, and climbing stairs. A previous provider had given him a trial of levodopa, but it had not helped with postural instability or gait disorder. He denied prior exposure to neuroleptic drugs, encephalitis, head injury, or alcohol abuse. He denied stroke, and there was no evidence of cerebral infarction on CT scan of the head (bilateral frontal atrophy was the only significant CT finding).

The patient’s speech was so slow and slurred that he sometimes had to repeat himself to be understood. His dysarthria also had a monotonous quality (loss of prosody). The patient seemed apathetic, but he denied feelings of sadness. He admitted to having slowed thoughts (bradyphrenia). He was alert and fully oriented in all domains, except he did not know the floor of the clinic. He could perform serial 7s accurately, but he was very slow in completing this task. He could recall three items at the end of three minutes. He was able to name, repeat, follow a 3-step command, and read, but he could not write a legible sentence. When trying to copy intersecting pentagons, he perseverated and copied a single pentagon over and over again (his total Mini-Mental State Examination score was 27/30).

He had a fixed stare (eye blink frequency was markedly diminished). Voluntary downgaze was severely limited (he looked to the left when attempting to look down). He was unable to initiate upgaze on command. When he fixated his eyes on a spot on the wall, the full range of vertical eye movements could be elicited with passive movement of the patient’s head. Voluntary horizontal gaze to the right was limited to 30 degrees of arc, with no limitation of voluntary gaze to the left. In addition to vertical gaze palsy, there were jerky pursuit eye movements, square wave jerks on fixation, hypometric saccades, slowing of saccades, loss of the Bell reflex (when the eyes are tightly closed, the eyes should move up and outward), loss of optokinetic nystagmus and faulty suppression of the vestibular-ocular reflex. There was no evidence for ocular dysmetria and only rare instances of apraxia of eyelid opening.

There was marked masking of facial expression. Moderate rigidity (grade 2/4) was detected symmetrically in the limbs. Slowing of finger- and foot-tap frequency was also moderate and bilaterally symmetric. The patient was unable to stand from a chair, even by using both arms. He had grade 2/4 postural instability, becoming unsteady in the Romberg position with eyes closed. He walked slowly with a mildly kyphotic posture on a widened base, but he needed no assistance for ambulation.

Biological basis

Etiology and pathogenesis

Microtubule-associated protein tau. Studies have shown a link between progressive supranuclear palsy and the gene for the microtubule-associated protein tau (MAPT). Tau is involved in axonal transport, and the aggregation of abnormal tau protein in supranuclear palsy results in the development of neurofibrillary (globose) tangles. Hoglinger and colleagues studied 1114 autopsied cases of supranuclear palsy and 3247 controls to find that three genes could modify expression of the MAPT gene in supranuclear palsy, including the MOBP (myelin-associated oligodendrocyte basic protein) gene (80). Other studies have verified the association of the MOBP gene with both supranuclear palsy and corticobasal degeneration (99; 33). In normal brains, there are equal numbers of the 3- and 4-repeat isoforms of the tau protein. In brains of patients with supranuclear palsy and corticobasal degeneration, there is preferential accumulation of the 4-repeat isoform of tau. In both of these diseases, there is overrepresentation of the H1 haplotype of MAPT whereas the H2 haplotype appears to be neuroprotective (24; 189; 33). The effect of the H1 haplotype on the likelihood of a diagnosis of PSP-Richardson syndrome (OR=13.2) was much higher than it was on the likelihood of a diagnosis of PSP-parkinsonism (OR=4.5). Within H1, a subhaplotype (H1c) appears to be a risk factor for both supranuclear palsy and corticobasal degeneration (33). A genome-wide association study on autopsy-proven supranuclear palsy cases has revealed additional supranuclear palsy risk alleles in STX6, EIF2AK3, and SLCO1A2 (84; 06). One study has shown that a mutation in a gene for another structural protein, dynactin, can result in progressive supranuclear palsy (71). A variation in the tripartite motif-containing protein 11 (TRIM11) gene, which is a component of the ubiquitin proteasome system, has been shown to be a modifier in the phenotypic expression of progressive supranuclear palsy (whether a patient has signs of PSP-Richardson or PSP-parkinsonism) (86). A study by Tipton and colleagues demonstrated that among familial cases of frontotemporal lobar degeneration, the youngest age of motor symptom onset was associated with the MAPT gene (different than the GRN gene, which was more often associated with corticobasal degeneration) (175). In the study, the MAPT gene was more likely to be associated with vertical gaze palsy than the GRN gene.

Cortical and subcortical atrophy. Macroscopic examination of the brain in PSP-Richardson syndrome has demonstrated mild atrophy of the frontal lobes and marked atrophy of the midbrain (41). Studies with structural MRI have shown less cortical atrophy in the PSP-parkinsonism variant and in PSP-pure akinesia with gait freezing, as compared to PSP-Richardson syndrome (159; 187); this would explain why cognitive and behavioral changes occur later in the PSP-parkinsonism variant compared to the classical Richardson syndrome (106). Lansdall and others found disruption of fronto-striatal networks in patients with progressive supranuclear palsy as an explanation for their lack of motivation (102). They suggested that atrophy in the brainstem (and the reticular activating system) could account for their altered attention.

Neurofibrillary tangles. Neurofibrillary (globose) tangle formation in the brainstem and basal ganglia is the most prominent microscopic pathological sign of supranuclear palsy (41). These “globose” tangles consist of clusters of straight filaments of tau, which are different from the paired helical filaments of tau seen in the brains of patients with Alzheimer disease. Other pathologic changes (cell loss, gliosis, and granulovacuolar degeneration) develop in several other cortical and subcortical structures, such as prefrontal cortex, basal forebrain, and cerebellum (191; 41). When the neurofibrillary tangles, tufted astrocytes, and coiled bodies were quantitated in each region of the brain, a grading system was developed to compare PSP-Richardson syndrome with PSP-parkinsonism and PSP-pure akinesia with freezing gait (189). Thus, a “PSP-tau score” was developed to grade the overall tau load in each autopsy case. These authors found that the overall tau load of patients with PSP-Richardson was significantly higher than that of PSP-parkinsonism patients. Those carrying the H2 haplotype protective allele in that study had the lowest PSP-tau scores. In a study, 14 experts tested a new set of criteria for the neuropathological diagnosis for PSP (150). Each of the experts scored digital slides from 10 brain regions from 15 PSP cases and 10 cases with other tauopathies. The criteria relied on identifying neurofibrillary tangles or pretangles in two of three regions (substantia nigra, subthalamic nucleus, or globus pallidus), as well as tufted astrocytes in one of two regions (peri-Rolandic cortices or putamen). These criteria showed high sensitivity (0.97) and specificity (0.91), as well as good inter-rater reliability for making a neuropathological diagnosis of PSP (Fleiss-kappa = 0.826).

Pathologic overlap with other diseases. Pathologic studies of brains from those with supranuclear palsy can be confusing because of the overlap with other neurodegenerative illnesses, such as Alzheimer disease (131) and corticobasal degeneration (184). Tufted astrocytes are unique to supranuclear palsy, whereas astrocytic plaques can also be seen in corticobasal degeneration (41; 150). Patients presenting with corticobasal syndrome have supranuclear palsy at autopsy in one of five cases (184). In one pathologic study, midbrain atrophy was only identified in supranuclear palsy, whereas cell loss in the substantia nigra was equal in severity in supranuclear palsy, multiple system atrophy, and Parkinson disease (166). In an autopsy study, TAR DNA-binding protein 43 (TDP-43) co-pathology was present and correlated with age in 14% of supranuclear palsy patients (149). TDP-43 is the cardinal protein associated with one of the frontotemporal dementias (semantic dementia). This study also demonstrated that Lewy body co-pathology was present in 9% to 15% of supranuclear palsy patients, but they found no age-association with this co-pathology.

Neurotransmitters. The pathophysiology of progressive supranuclear palsy involves all five of the recognized basal ganglia-thalamocortical circuits (107). A number of neurotransmitters and neuromodulators are involved in the organization and modulation of these circuits, including dopamine, GABA, and norepinephrine. Cholinergic dysfunction in the brains of patients with supranuclear palsy is related to loss of cholinergic interneurons in the striatum and in the nucleus basalis of Meynert. Two groups have used (11)C-PET to verify the reduction of paracentral and thalamic cholinergic activity in patients in the early stages of progressive supranuclear palsy (65; 79). Neuropathological studies have shown 49% fewer noradrenergic (pigmented) neurons in the locus coeruleus of PSP patients, compared to the number in age-matched controls (90). This loss of pigmented neurons correlated directly with disease severity and inversely with tau inclusions.

Epidemiology

Prevalence and incidence. The estimated prevalence for progressive supranuclear palsy in the United States and the United Kingdom ranges from five to seven per 100,000 inhabitants, similar to that of myasthenia gravis (124). The incidence rate for supranuclear palsy in the Rochester Epidemiology Project was reported to be 1.1 cases per 100,000 person-years (157).

Age at onset. One study compared two supranuclear palsy subtypes and showed that presentation was often later among those with PSP-Richardson syndrome (mean = 74.3 years) than among those with PSP-parkinsonism (mean = 71.0 years).

Survival. Mean survival after symptom onset for PSP-Richardson syndrome is now 5.9 to 6.8 years, compared to 9.1 to 11.2 years for PSP-parkinsonism (188; 87). Survival for PSP-pure akinesia with gait freezing is about 11 years, similar to that seen in PSP-parkinsonism (190).

Male:female ratio. The male:female ratio in two series was 1.6:1 (157). In another study, two thirds of the patients with PSP-Richardson syndrome were men, whereas the sex distribution among patients with PSP-parkinsonism was even (188). In a study of 242 patients with progressive supranuclear palsy (all phenotypes included), 58% of participants were men (138).

Risk factors. Supranuclear palsy is generally thought to be a sporadic disorder, but one Rotterdam study found that 33% of patients with supranuclear palsy had at least one first degree relative with either dementia or parkinsonism, compared to 25% of controls (44). A Japanese study showed that 15% of patients with supranuclear palsy had a family history of progressive supranuclear palsy, parkinsonism, or dementia (54). Twelve cases in that study (7% of the total) had an autosomal dominant mode of familial transmission. The familial and sporadic cases did not differ with respect to age of onset, duration of disease, or presence of dementia. APOE-e4 genotype was lower in supranuclear palsy cases, compared to controls. Familial supranuclear palsy cases had a stronger association with the MAPT-H1 haplotype than with the controls or with the sporadic supranuclear palsy cases (54). A large case-control study has shown a potential role of the environment on the etiology of supranuclear palsy because a higher number of years of drinking well water appears to be a significant risk factor after adjustment for possible confounders (112).

Differential diagnosis

Confusing conditions

Parkinson disease. This is the most common of the synucleinopathies. Parkinson disease is manifested early on by unilateral rest tremor in 82% of cases, whereas rest tremor is seen in only 19% of those with early progressive supranuclear palsy. Hypokinesia without decrement (the finger-tap test) distinguishes PSP-parkinsonism from Parkinson disease (110). Gait disorder with falls is the most common early motor symptom of PSP-Richardson syndrome (22). Hershey and others (unpublished) found that 83% of patients with early PSP-Richardson syndrome had instability of visual fixation (square wave oscillation), whereas only 18% of patients with early Parkinson disease had this finding. Jerky pursuit eye movements were more commonly seen in supranuclear palsy (67%) than in Parkinson disease (18%). neurofilament light chain levels are higher in patients with supranuclear palsy than in those with Parkinson disease (85).

Dementia with Lewy bodies. Dementia with Lewy bodies is the second most common of the synucleinopathies. Besides dementia, it presents with a variety of core clinical features, including parkinsonism, visual hallucinations, fluctuating levels of alertness, and REM sleep behavior disorder (117). Although patients with Lewy body disease may have similar amounts of rigidity and bradykinesia as those with Parkinson disease and supranuclear palsy, they have less resting tremor and fewer falls. Patients with Lewy body disease and Parkinson disease both respond to levodopa in the early years (76), whereas only about 26% of patients with supranuclear palsy respond well to this drug (191). Genetic risks for dementia with Lewy bodies include the APOE-e4 allele, the alpha-synuclein gene, and the glucocerebrosidase gene (69).

Multiple system atrophy. This group of neurodegenerative diseases constitutes a less common form of synucleinopathy. It is divided into those with autonomic features (MSA-autonomic), those with parkinsonian features (MSA-parkinsonism), and those with cerebellar features (MSA-cerebellar). The first group has symptomatic postural hypotension, early urinary incontinence, and syncope (Shy-Drager syndrome, or MSA-autonomic). Their signs of parkinsonism are usually very mild, compared to those seen in supranuclear palsy. Koga and colleagues reported that when 134 clinically diagnosed multiple systems atrophy cases came to autopsy, only 62% had the correct diagnosis (98). Of the 15 cases in this series with supranuclear palsy at autopsy, the age at onset was significantly older; 64% were more cognitively impaired (vs. 37% for multiple system atrophy), and 57% had vertical gaze palsy (vs. only 14% for multiple system atrophy). Most patients with supranuclear palsy masquerading as multiple system atrophy either presented with cerebellar ataxia (MSA-cerebellar) or developed it along the way. Apathy scores are higher in patients with supranuclear palsy than in those with multiple systems atrophy (104).

Neuronal intermediate filament inclusion disease. This is a neuropathologically distinct, clinically heterogeneous variant of frontotemporal dementia with an early age of onset (mean=40 years) and an early age at death (mean=54 years). The neuropathological hallmark is intracellular inclusions that stain positively for antibodies to neurofilaments, alpha-internexin and fused in sarcoma (FUS) (119; 17). The inclusions stain variably with antibodies to ubiquitin and negatively with antibodies to tau, alpha-synuclein, or TDP-43. The common clinical presentation is like frontotemporal dementia, but patients can also have truncal ataxia, falls, parkinsonism that is unresponsive to levodopa, saccadic eye movements, and supranuclear gaze palsy (29; 176). Neuroimaging shows atrophy involving the frontal lobes more than the temporal or parietal lobes. Sometimes paravermal hyperintensities are evident in the cerebellum (176). Progression is usually more rapid than that seen with PSP.

Perry syndrome. These patients have a rare dominantly inherited form of parkinsonism that is associated with IgLON-5 antibodies (75). The abnormal psychiatric behaviors have been reported to improve after steroids and intravenous immunoglobulins.

Associated or underlying disorders

Frontotemporal dementia. The most common frontotemporal lobar degeneration syndromes, like behavioral variant frontotemporal dementia, nonfluent primary progressive aphasia, and progressive supranuclear palsy, are tauopathies (23). In one series, 3 of 66 autopsy-proven cases of supranuclear palsy presented as the PSP-behavioral variant with prominent behavioral and personality changes before motor features of supranuclear palsy appeared (73). In these cases, at least one cardinal symptom or sign of supranuclear palsy emerged up to 5 years after the original presentation. MRI showed reduced midbrain volume, comparable to typical supranuclear palsy cases. The frontal assessment battery (FAB) is not useful to discriminate progressive supranuclear palsy from frontotemporal dementia. However, the sum of only two FAB subscores (verbal fluency and the Luria assessment of motor apraxia) were as good as the total score in differentiating supranuclear palsy from both Parkinson disease and MSA-parkinsonism (170). This is because patients with supranuclear palsy have more problems with verbal fluency and motor apraxia due to the involvement of frontocerebellar pathways. In both supranuclear palsy and frontotemporal dementia, behavioral changes are often apparent to family members but are rarely reported by patients themselves (147).

Corticobasal syndrome. Corticobasal syndrome is another tauopathy that is one of the frontotemporal lobar degeneration syndromes (102; 23). These patients have frontal and parietal lobe atrophy on brain imaging studies, whereas patients with supranuclear palsy have pronounced midbrain atrophy (25; 120; 47). Patients with corticobasal syndrome demonstrate parietal lobe signs, such as motor apraxia, alien hand syndrome, extinction to double simultaneous stimulation, and neglect (25). Apathy scores are high in patients with both supranuclear palsy and corticobasal syndrome. Patients with both diseases can show similar amounts of logopenia and lack of spontaneity (184). There are a few patients who have been found to have signs of progressive supranuclear palsy at autopsy who had clinically presented with corticobasal-like symptoms (109) and a shift in tau burden away from the basal ganglia towards the cortical regions, arguing in favor of a cortical variant of supranuclear palsy. This is called the PSP-cortical basal syndrome variant.

Diagnostic workup

Neuropsychological tests. The following neuropsychological tests are helpful in documenting memory disturbance, global cognitive dysfunction, cognitive slowing, loss of executive function, loss of social cognitive deficits, and neuropsychiatric symptoms in patients with suspected supranuclear palsy:

1. Global cognitive function. The Mattis Dementia Rating Scale, the Montreal Cognitive Assessment Scale (MoCA), and the Mini-Mental State Examination (MMSE) assess global cognitive abilities. MMSE is a relatively insensitive tool to assess executive function and attention, which are the two main cognitive deficits in patients with supranuclear palsy (148).
2. Verbal fluency. This can be measured by the number of words generated in one minute either within a category (eg, animals) or with letters. For example, the ability to produce less than seven p-words in one minute was found more likely to be associated with a diagnosis of supranuclear palsy than with Parkinson disease (148). The positive predictive value of this test was 0.81, and the negative predictive value was 0.93.
3. Frontal lobe dysfunction. The Frontal Assessment Battery (FAB) and Trail Making Test B are commonly used to assess frontal lobe dysfunction (106). The FAB was shown not to be useful for monitoring the decline in executive dysfunction in patients with supranuclear palsy over a 1- to 2-year period, probably because frontal cognitive impairments are already fully developed early in the disease process (61; 111). Stamelou and others demonstrated a lack of specificity for the FAB because it could not distinguish patients with supranuclear palsy from those with frontotemporal dementia (170). The Stroop color interference test is also used to test frontal lobe dysfunction in supranuclear palsy (97; 106).
4. Social cognitive deficits. These can be measured with several tests, including pictures of faces (95), voice emotion recognition tests (14), and the Penn emotion recognition test (140). When the latter test was used, patients with supranuclear palsy were found to be significantly impaired in their ability to recognize sad or happy faces compared to patients with Parkinson disease or age-matched controls. When impairment of voice emotion recognition is recognized in supranuclear palsy patients, it has been correlated with dysfunction of the right inferior frontal gyrus (60).
5. Socioemotional sensitivity. Higher functional connectivity in the salience network between the right anterior insula and both cortical and subcortical nodes predicts socioemotional sensitivity in normal adults (177). Impairment in the salience network occurs earlier in the behavioral variant of frontotemporal dementia and progressive supranuclear palsy, compared to other neurodegenerative diseases. Using the Revised Self-Monitoring Scale (RSMS), measurement of connectivity in the salience network requires resting state functional MRI imaging as well as a measure of informant described responsiveness to emotional stimuli.
6. Neuropsychiatric symptoms. The Neuropsychiatric Inventory (NPI) of Cummings and colleagues is useful to identify apathy, depression, and anxiety in patients with supranuclear palsy (38; 111; 144; 88). Apathy and depression were the most common early neuropsychiatric symptoms, and apathy appeared to worsen most dramatically during the first year after study enrollment (88). Alexithymia can be measured with the 20-item Toronto Alexithymia Scale (TAS-20), and anhedonia can be measured with the Snaith-Hamilton Pleasure Scale (10).
7. Identifying fall risk. Kim and colleagues found that the Stroop test, along with the Rey Complex Figure Test and alternating hand movements, were useful to distinguish patients with supranuclear palsy who fell frequently from those who did not (97). Another study reported that slowed toe tapping, abnormal voluntary horizontal saccades, and slowed rising from a chair were predictive signs for frequent falls among those with supranuclear palsy (21).
8. Assessing multiple cognitive domains. Bang and colleagues used the Repeatable Battery for Assessing Neuropsychological Status (RBANS) in patients with supranuclear palsy who were enrolled in a 52-week phase III randomized placebo-controlled trial of a new drug that was designed to slow progression of the disease (13). This test had the advantage of being able to assess multiple domains of cognitive function and of being applied at multiple follow-up visits during the trial.

MRI studies: midbrain atrophy. Brain imaging with MRI can now distinguish patients with supranuclear palsy from those with other basal ganglia disorders. Midbrain atrophy is the imaging hallmark of supranuclear palsy (83; 120; 47). The “hummingbird sign” is a sign of midbrain atrophy that is used in everyday practice to identify patients with progressive supranuclear palsy (67). The addition of the midbrain/pons ratio improves the specificity of the MRI as a way to distinguish supranuclear palsy from multiple system atrophy, Parkinson disease, and corticobasal degeneration (116; 85). Owens and colleagues showed that a midbrain-to-pons ratio of less than or equal to 0.52 was 100% specific for progressive supranuclear palsy among a group of 75 cases of parkinsonism syndromes (133). Among an autopsy-confirmed cohort of six patients with supranuclear palsy and 23 patients with nonsupranuclear palsy, all patients with nonsupranuclear palsy had midbrain-to-pons ratios higher than 0.50, whereas all but one patient with supranuclear palsy had a ratio lower than 0.50. The positive predictive value of the ratio (< 0.50) was 100%, and the negative predictive value was 95.8% (91). The pooled sensitivity of the midbrain-to-pons ratio for the diagnosis of supranuclear palsy (vs. Parkinson disease) is 0.98 and the pooled specificity is 0.99 (195). The Movement Disorder Society-PSP subtypes cannot be differentiated by the available MRI-based midbrain-pons ratios (137; 68). Following the MR Parkinson index (MRPI) over 1 to 2 years appears to be a more powerful method than monitoring clinical motor scores in evaluating the progression of progressive supranuclear palsy over time (143). The MRPI is sensitive (83%) and specific (85%) for distinguishing PSP from other parkinsonism syndromes (68). The MRPI can also be used to provide estimates for sample sizes needed in prospective clinical trials. However, despite the successful application of the MRPI in grouped analyses, a study showed that the positive predictive value of this metric was only 79% when applied to individual patients (40).

MRI studies: whole brain atrophy. Annualized rates of whole brain atrophy are significantly higher in progressive supranuclear palsy (1.26%) and multiple systems atrophy (1.65%), compared to Parkinson disease (0.54%) or controls (0.37%) (70). When brain atrophy over 6 to 12 months was compared in 55 patients with progressive supranuclear palsy to 35 with corticobasal syndrome, there were no differences in rates of change in midbrain and pontine volume (47). When the progressive supranuclear palsy variants were compared to one another, volume loss in the frontal lobes was found to be particularly characteristic of the speech/language variant, the corticobasal variant, and the frontal behavioral variant of progressive supranuclear palsy (187). The least amount of volume loss was seen in PSP-parkinsonism and PSP-pure akinesia with gait freezing. Patients with progressive supranuclear palsy with depression have shown greater reduction in cortical thickness in temporal-parietal regions (182).

Diffusion tensor imaging (DTI) studies. Piatella and colleagues performed diffusion tensor imaging on 16 patients with supranuclear palsy and 16 age-matched healthy subjects and found widespread changes in subcortical white matter, mainly affecting cerebellar peduncles and thalamic radiations (136). They found a significant correlation between MMSE scores and mean brain fractional anisotropy (r=0.66; p< 0.001). These observations of extensive myelin changes in patients with supranuclear palsy fit with genetic evidence that the myelin basic protein gene (MOBP) is involved in this disease’s pathogenesis (80). These findings also fit with plasma neurofilament light chain studies, where levels are significantly higher in patients with supranuclear palsy than in those with Parkinson disease or age-matched controls (neurofilament light chains are markers of degeneration of large myelinated axons) (74; 85). Diffusion tensor imaging studies of 70 patients with progressive supranuclear palsy (vs. 30 controls) showed that impulsive and apathetic behaviors correlated best with white matter changes in the ventral fronto-tegmental tracts (103).

FDG-PET studies. Several authors have shown that metabolic activity, as measured by FDG-PET, is reduced in the frontal lobes of patients with progressive supranuclear palsy (48), especially in those who have the PSP-Richardson syndrome variant (168). When compared to those with Parkinson disease and controls, patients with PSP-Richardson syndrome show pronounced thalamic hypometabolism whereas those with PSP-parkinsonism have putaminal hypometabolism (168). Patients with the nonfluent aphasia variant of progressive supranuclear palsy have decreased metabolism in the left fronto-insular areas on FDG-PET early in the course of their disease (42).

CIT-PET studies. Dopamine transporter loss can now be measured with F-18-CIT PET, and these studies have shown earlier transporter loss in the anterior caudate and ventral putamen in patients with supranuclear palsy compared to those with Parkinson disease (128).

Tau-PET studies. A number of tau PET radio tracers have demonstrated promising preliminary results in progressive supranuclear palsy. Thus, binding of the tracer [18F]-FDDNP is significantly increased in the subthalamic area, midbrain, and cerebellar white matter of patients with supranuclear palsy, which is consistent with the reported distribution of tau. There was a positive correlation between cortical [18F}-FDDNP binding and the severity of the supranuclear palsy (96). PET studies with the [11C]-PBB3 ligand have shown superiority over the [18F]-AV-1451 ligand in labeling tau lesions in progressive supranuclear palsy brains (130). Endo and colleagues used the [11C]-PBB3 ligand in 13 patients with progressive supranuclear palsy and 13 controls (all were negative for amyloid) and found patients with mild progressive supranuclear palsy had increased tau ligand uptake in the parietal grey matter and midbrain (49). In moderate-severe progressive supranuclear palsy cases, the [11C]-PBB3 uptake was more marked in the fronto-parietal white matter and parietal grey matter. Studies with the [18F]-flortaucipir ligand have shown that patients with progressive supranuclear palsy have increased uptake of the ligand in the striatum, globus pallidus, cerebellar dentate, red nucleus, and thalamus (59; 187). Regional uptake of this ligand was able to distinguish patients with corticobasal syndrome from those with PSP (59).

Synaptic-PET studies. The PET tracer [11C]-UCB-J binds to synaptic vesicle glycoprotein 2A and is a marker of synaptic density. Binding density for this ligand was reduced in both patients with PSP-Richardson syndrome and corticobasal syndrome in frontal, temporal, parietal, occipital, cingulate, and hippocampal areas, compared to controls (82). There was a negative correlation between global [11C]-UCB-J binding and the PSP-rating scale score (R=-0.61), suggesting that synapse loss was an early feature of both of these primary tauopathies.

Neuroinflammation-PET studies. A group examined the neuroinflammatory ligand [11C]-PK11195 along with the tau ligand [18F]-AV-1451 in 17 patients with PSP-Richardson syndrome (115). Their hypothesis was that the neuroinflammatory and tau markers would co-localize. The correlation between binding of the two ligands was positive in both subcortical (R=0.769) and cortical (R=0.836) regions. Positive correlations were also noted between disease severity and binding of the tau ligand (R=0.667) and the neuroinflammatory ligand (R=0.788).

MRS studies. Magnetic resonance spectroscopy studies have demonstrated decreased ATP levels in the basal ganglia and frontal lobes of patients with early supranuclear palsy, suggesting that mitochondrial dysfunction could be an upstream event in the pathway to cell death in this disease (169).

SPECT perfusion studies. Brain perfusion, as measured by HMPAO-SPECT, is reduced in the frontal lobes of patients with supranuclear palsy. When SPECT scans of patients with early supranuclear palsy were compared to those of early Parkinson disease using a double-blind method, bilateral frontal hypoperfusion was found on SPECT in 50% of patients with early progressive supranuclear palsy but in none of the patients with Parkinson disease (77). SPECT studies have shown significant reductions in orbitofrontal perfusion in those patients with supranuclear palsy and obsessive-compulsive symptoms as compared to those without behavioral symptoms (55).

SPECT transporter studies. Studies of the dopamine transporter have demonstrated that patients with PSP-Richardson syndrome are more likely to have reduced dopamine transporter activity in the striatum than those with PSP-parkinsonism (108). The putamen-to-caudate ratios of transporter activity are significantly different when patients with Parkinson disease were compared to all those with supranuclear palsy, making dopamine transporter scans a useful diagnostic tool (108). Dopamine transporter imaging has been shown to be helpful in discriminating among the various parkinsonism syndromes. Thus, patients with Parkinson disease, multiple systems atrophy, supranuclear palsy, and corticobasal syndrome all have distinct patterns of dopaminergic depletion on [123-I] ioflupane SPECT, although patients with supranuclear palsy and multiple systems atrophy show similar signal reduction in the head of the caudate relative to patients with Parkinson disease (12).

Biomarker studies. Although patients with Alzheimer disease and neuronal intranuclear inclusion disease have elevated CSF levels of p-Tau181, the CSF concentrations of this biomarker are within the normal range in patients with PSP (101). Skin biopsies that identify alpha-synuclein deposits are able to distinguish patients with Parkinson disease from those with PSP and corticobasal syndrome (62).

Management

Cholinesterase inhibitors. Even though central cholinergic neurons are known to die in patients with progressive supranuclear palsy, placebo-controlled trials with cholinesterase inhibitors, such as physostigmine, have been disappointing (19; 185). In a double-blind, crossover trial of donepezil in 21 patients with supranuclear palsy, modest improvements in memory were observed, but these changes came at the expense of worsening motor and functional abilities (113).

Antidepressants. In one study of depression in supranuclear palsy, only 31% of patients were being appropriately treated with antidepressant drugs (193). The authors concluded that increased attention should be paid to the detection and treatment of depressive symptoms in patients with supranuclear palsy. Antidepressant drugs are also useful in treating obsessive-compulsive symptoms, which are common (24%) in patients with the Richardson syndrome subtype of progressive supranuclear palsy (55). Mendez and others showed that sertraline was useful to manage the verbal and motor stereotypies of patients with frontotemporal dementia (118). Because stereotypies have been described in supranuclear palsy, it might be good to consider using this drug for supranuclear palsy as well (142). Trazodone, another serotonin agent used to treat depression and anxiety in frontotemporal dementia, can be used to treat these symptoms in progressive supranuclear palsy (160). Low doses of citalopram (10 mg once daily) can be effective in treating pseudobulbar affect in patients with supranuclear palsy (147).

Dextromethorphan/quinidine. Pseudobulbar affect can sometimes be an embarrassing problem for patients with progressive supranuclear palsy, but there are few trials to guide therapy (147). Dextromethorphan/quinidine has been shown to be effective in treating pseudobulbar affect in other neurodegenerative conditions, such as Alzheimer and Parkinson disease (158). The quinidine component inhibits the metabolic conversion of dextromethorphan to its active metabolite, so that adverse effects are minimized. Caution is advised for patients who are at risk for torsade de pointes and those who are on drugs that might interact with dextromethorphan or quinidine.

Coenzyme Q10. The inhibition of complex I in experimental animals by rotenone can produce pathologic changes that are similar to those seen in supranuclear palsy. Supplementation with coenzyme Q10 (CoQ10), the electron recipient of complex I, has been shown to reduce the neurotoxicity of rotenone in rats. Stamelou and colleagues found small but significant improvements in motor and cognitive function in a small group of patients with supranuclear palsy when they were treated with CoQ10 (171). However, in a multicenter, randomized, placebo-controlled, double-blind study, high doses of CoQ10 (2400 mg/day) did not show improvement of progressive supranuclear palsy symptoms or disease progression over a 12-month period (07).

Behavioral management. Nonpharmacologic interventions should be considered for managing the behavioral complications of progressive supranuclear palsy (verbal aggressiveness, restlessness, repetitive behaviors). One approach, the A-B-C method, asks the caregiver to keep a diary that identifies the antecedent event for the change in behavior, along with the behavior itself and its consequences (164). Behavioral strategies can be developed, based on these diary entries, to change the patient’s schedule, or to modify his environment. Another approach for managing behavioral complications is the introduction of purposeful activities (37).

Balance and eye movement training. One study showed that gaze control can be improved after five weeks of balance and eye movement training in patients with progressive supranuclear palsy (194). Nineteen ambulatory patients were randomly assigned to receive balance training and visual awareness exercises, or balance training alone. At five weeks, gaze control improved significantly in those assigned to the balance training and eye movement exercises. Another small trial with eight patients with supranuclear palsy showed that audio-biofeedback training produced significant improvement on the Berg Balance Scale, which remained significant at the 4-week follow-up visit (125). These studies seem promising, but they need to be repeated in a blinded study design, and the studies need to be extended for longer periods of time.

Levodopa. Because pathologic changes in supranuclear palsy extend far beyond the substantia nigra, it is not surprising that dopaminergic therapies are limited in their ability to improve motor function (123; 41). Levodopa is only effective in about 26% of patients with supranuclear palsy; those who improve are most likely to have the PSP-parkinsonism variant of the disease (191). One study of pain in patients with supranuclear palsy showed that dopaminergic therapy provided benefit in only 25%, compared to 50% on patients with Parkinson disease and 47% in those with multiple systems atrophy (94). The benefit of levodopa rarely lasts longer than a few years in supranuclear palsy. Nevertheless, some caregivers continue the drug for longer periods of time, as it helps them to perform activities of daily living, such as bathing and toileting.

Valproic acid. Occasionally, a supranuclear palsy patient may present to an inpatient setting after a fall/head injury and develop agitation or acute confusion. Some of these agitated patients may respond to low doses of valproic acid when it is used as an adjunct to a low dose of an atypical antipsychotic drug (43; 56). Systematic reviews have shown that use of valproic acid as monotherapy is less effective than use in combination with other psychoactive drugs. Nevertheless, a double-blind, placebo-controlled trial of valproic acid in supranuclear palsy showed lack of efficacy and poor tolerability (105).

Amantadine. One group described benefit from amantadine, levodopa, anticholinergics, dopamine agonists, and selegiline in differing combinations in nine of 15 autopsy-verified cases of progressive supranuclear palsy (18). Another argued that amantadine provides benefit in supranuclear palsy for treating akinetic rigidity, fatigue, lack of motivation, and balance problems (147). Dose limitations include hallucinations, aggressive behavior, insomnia, and livedo reticularis. Patients with progressive supranuclear palsy in their late 60s and early 70s are more likely to benefit from low doses of amantadine than those aged 75 and older, primarily due to fewer adverse effects in the younger age groups (152).

Rasagiline. The MAO inhibitor rasagiline has shown neuroprotective effects in preclinical models of neurodegeneration. A 1-year randomized, double-blind placebo-controlled trial in 44 patients with supranuclear palsy showed no effect on symptom progression as measured by the PSP rating scale, whereas there were mild increases in known side effects, such as hallucinations and ventricular extra systoles (127).

Melatonin. REM sleep behavior disorder has been shown to be present in 33% to 37% of patients with supranuclear palsy (09; 122). Clonazepam was the first medication described for use in REM sleep disorder and remains the favored therapy for many physicians, but melatonin 3 mg/d has been demonstrated to be effective in reducing both sleep disorder symptoms and tonic EMG activity in a double-blind placebo-controlled trial (100). Melatonin has also been used to prevent the symptoms of the sundown syndrome (30).

Riluzole. This glutaminergic modulator is currently approved for the treatment of amyotrophic lateral sclerosis. When it was compared to placebo in a large 3-year trial designed to test its neuroprotective effects in supranuclear palsy, it showed no benefit with respect to either survival or disease progression (15).

Botulinum toxin. In some advanced cases of supranuclear palsy, focused injections of botulinum toxin have been shown to relieve rigidity and spasticity in the legs so that caregivers can more easily assist with bathing (11). Botox injections can also be given into the cricopharyngeal muscle for neurogenic dysphagia if the problem is hyperactivity of that muscle (04). Other uses for botulinum toxin in the PSP-Richardson syndrome include the treatment of apraxia of eyelid opening and blepharospasm, both of which can coexist (50).

Adaptive equipment. Several rehabilitation interventions can prevent falls in those with progressive supranuclear palsy: exercise programs, weighted walkers, wheelchairs, bathroom safety equipment, and occupational therapy. In a pilot observational trial, robot-assisted gait training was found to be a feasible and safe form of rehabilitation in five cognitively intact patients with progressive supranuclear palsy, with all subjects showing an improvement in the gait spatiotemporal index (mean velocity, cadence, step length, and step width) (154). Another study showed that a robotic device was no more effective than treadmill training in patients with supranuclear palsy undergoing a multidisciplinary goal-based rehabilitation treatment program (36). A wheelchair is the safest option for the patient when falling becomes a regular occurrence.

Discontinuation of medications. Another important intervention to prevent falls is to discontinue medications that put elderly patients at risk for delirium: opioids (OR=2.5), benzodiazepines (OR=3.0), and antihistamines (OR=1.8), according to a meta-analysis (35; 192). Antispasmodic drugs (oxybutynin) also have anticholinergic activity that can impair memory and concentration, so their doses should be reduced or the drugs discontinued altogether.

Speech therapy. Speech pathologists can help patients with supranuclear palsy who suffer from dysarthria or dysphagia. For example, a device called a ChatterVox can be used to amplify the patient’s voice. In a longitudinal study in a rehabilitation inpatient unit, 16 sessions of speech therapy using the Lee Silverman Voice Treatment protocol resulted in similar increases in maximum phonation duration and volume of voice in reading in supranuclear palsy compared to Parkinson disease patients, although improvements in the quality of voice and articulation were more significant in the Parkinson group (153). A swallowing evaluation is helpful to teach patients how to safely swallow food to prevent aspiration. Family members need to learn how to perform the Heimlich maneuver in case the PSP patient chokes.

Exercise programs. The most recent Cochrane Database Review on exercise programs for patients with dementia showed no evidence of benefit for improving cognitive function or neuropsychiatric symptoms, but six trials (n=289 patients) showed improvement in activities of daily living (53). A systematic review specifically addressing patients with supranuclear palsy did not show robust evidence for benefit of therapeutic exercise (163).

Modulators of tau phosphorylation. Glycogen synthase kinase-3 (GSK-3) is believed to hyperphosphorylate tau protein in progressive supranuclear palsy. A phase 2, double-blind, placebo-controlled, randomized trial of the GSK-inhibitor tideglusib, dosed orally at 600 mg or 800 mg/day over 52 weeks, failed to show clinical efficacy in patients with mild moderate progressive supranuclear palsy (179). Danuvetide is a neuroprotective peptide that modulates cytoskeletal structures within neurons and glia, in part due to a reduction in tau phosphorylation (66; 64). A randomized, double-blind, placebo-controlled, phase 2/3 trial of davunetide in patients with progressive supranuclear palsy also failed to show an effect of this drug on progression over 52 weeks in scores on the PSP Rating Scale or on the Schwab and England ADL scale (26).

Deep brain stimulation. Although few data on deep brain stimulation exist, there are a few encouraging reports suggesting that patients with supranuclear palsy who experience frequent falls might benefit from either unilateral or bilateral stimulation of the pedunculopontine nucleus (161; 46). The latter study showed that the PSP-parkinsonism phenotype experienced the most improvement in their gait.

Transcranial direct current stimulation. In a sham-controlled, double-blind crossover design study, 12 patients with supranuclear palsy who were treated with transcranial direct current stimulation over the dorsolateral prefrontal cortex showed improved performance on some language tasks, but no improvement on executive function (183). In a sham-controlled case study of a progressive supranuclear palsy patient who had reduced verbal fluency, active sessions of transcranial direct current stimulation improved the patient’s performance on tests of phonemic fluency and action naming, compared to sham stimulation sessions (114).

Palliative care. A stepwise approach for the initiation of palliative care for patients with progressive supranuclear palsy using clinical milestones at various stages as triggers for the initiation of care has been proposed by Bessemer and colleagues (16). The goal of palliative care is to improve the management of pain, mobility, and psychological distress for the patient and to ease the burden for caregivers. Most patients with progressive supranuclear palsy eventually require care from several different healthcare providers (about four on average, depending on the phenotype of PSP and the stage of disease) (138). Meaning in life should also be considered in patients with supranuclear palsy who, compared with healthy individuals, seem to focus on supportive relationships and leisure more than health status (51). The most common cause of death in supranuclear palsy (as in other forms of parkinsonism) is pneumonia (121).

Outcomes

Patients with the two milder variants of supranuclear palsy (PSP-parkinsonism and PSP-pure akinesia with gait freezing) have a life expectancy of 9.1 to 11.2 years (188; 190; 87). This is in contrast to the natural history of patients with PSP-Richardson syndrome, where there are persistent falls, no response to levodopa, and death within 5.9 to 6.8 years. In PSP-Richardson syndrome, male gender and older age at onset are predictors for shorter disease duration. Litvan and Kong showed that worsening of ADL scores over two years was positively correlated with progression of falls, eye movement sub-item scores, and executive dysfunction (111). The PSP tau score is higher in those who have PSP-Richardson syndrome than in those with PSP-parkinsonism (189). The PSP-tau score correlated negatively with disease duration. According to a prospective longitudinal study by Arena and colleagues, the presence of sleep disturbances and hallucinations was associated with an increased risk of death in supranuclear palsy (08).

Special considerations

Anesthesia

Anticholinergic agents should be avoided by patients with supranuclear palsy because they can cause increased confusion. Several common agents have anticholinergic activity, including oxybutynin, doxepin, amitriptyline, disopyramide, dicyclomine, and diphenhydramine (35).