How to spot ocular abnormalities in progressive supranuclear palsy? A practical review

Square wave jerks (SWJs) represent the most common type of saccadic intrusions, characterized by conjugated couplets of horizontal back-to-back microsaccades (0.5–5°), taking the eye from the fixation point and back onto it in an involuntary manner, with a normal intersaccadic interval of about 200 ms [25]. Saccadic intrusions are involuntary conjugate saccades that interrupt fixation, and can be sub-classified by the presence (SWJs, square wave pulses, and macrosaccadic oscillations) or absence (ocular flutter and opsoclonus) of an intersaccadic interval [25].

SWJs can be identified by asking the patients to maintain their gaze fixation to a static object for the duration of 10 to 30 s. Alternatively, SWJs can be observed with Frenzel’s goggles or an ophthalmoscope, by asking the patients either to hold their gaze steadily at the primary position or to shift their gaze towards eccentric positions. On clinical inspection, SWJs are visualized as saccade pairs that appear purely horizontal: the first saccade moves the eye away from the fixation target, and, after a short interval, the second saccade brings it back to the target [26]. In addition to SWJs, microsaccades may be observed, but they are likely to be too small (< 0.5°) to be evident during clinical examinations in normal subjects [27]. While SWJs appear to be purely horizontal on clinical inspection, microsaccades often display an oblique trajectory [28]. Recent studies suggested that both microsaccades and SWJs are generated by the same neural circuit, and that larger saccades are more likely to show SWJs [26, 29].

It is important to understand that occasional SWJs can occur in preschool children and healthy elderly subjects [30, 31]. The typical frequency of SWJs is less than 9/min in healthy subjects, but the frequency seems to increase with age [32]. However, certain observations, such as large (> 5°), frequent (> 16 per minute during fixation or > 20 per minute in the dark), multiplanar, or disconjugate SWJs, should alert physicians of an underlying neurological dysfunction [33]. While SWJs have been reported in patients with PSP, they are not specific to this disorder, and have been observed in a number of disorders affecting the functions of brainstem circuits, particularly the cerebellum and its output through the superior cerebellar peduncle. However, certain characteristics of SWJs such as macro SWJs were included in the core clinical features suggestive of SPS from the new MDS clinical diagnostic criteria of PSP, which may be useful in the differentiation of individuals with PSP from normal subjects or those with other parkinsonian disorders [6]. These disorders include Friedreich’s ataxia, Huntington’s disease, oculomotor apraxia type 2, PD, multiple system atrophy (MSA), corticobasal degeneration (CBD), and recessive spinocerebellar ataxias [34, 35].

Out of 45 articles included in this systematic review, nine articles studied the presence of SWJs in patients with PSP using different eye recording techniques [26, 34, 36‐42].

Among various forms of parkinsonian syndromes, SWJs were reported in most patients with PSP (60–100%), followed by MSA (40–60%), CBD (33%), and, to a lesser extent), PD (0–15% of patients) [34, 36, 38]. The frequency of SWJs reported in the literature may vary depending on the respective recording technique. Moreover, SWJs in patients with PSP were identified to be larger than 1°, which is a useful clinical sign for differentiation of PSP from PD, since large SWJs or macro SWJs are considered uncommon among patients with PD [34]. When compared to healthy normal subjects, SWJs in patients with PSP were more frequent, larger, and more markedly horizontal [36, 41]. Microsaccades also lost their vertical component in patients with PSP [26]. The latter finding has been identified as the most distinctive characteristic sign of PSP [26].

Other studies looked further into fixational characteristics among patients with parkinsonian disorders [36, 39, 42]. A high proportion of abnormal ocular fixations was observed in patients with PSP, MSA, or CBD [34, 38]. Patients with PSP exhibited larger SWJs than controls when fixation targets were visible, and this difference was reversed in the absence of a fixation target [39, 40]. In contrast, patients with MSA demonstrated increased SWJs in conditions with and without a visible target [36]. Whether this observation represents a useful tool to discriminate PSP from other parkinsonian disorders in the early stage is still unclear, and these findings need to be replicated in subsequent studies. Another study described the distinguished patterns of SWJs in PSP, which showed larger sizes, higher frequencies, but less vertical components than SWJs in control subjects. The authors noted that this finding is consistent with the typical feature of PSP, vertical supranuclear palsy, and suggested that SWJs should be considered an early eye sign in PSP, due to the short disease duration and early symptom onset (median duration of 4 years) [26].

SWJs can result from either (1) disruption to areas that influence saccadic control, including the cortex, cerebellum, superior colliculus, and basal ganglia, or (2) direct injury to omnipause neurons located in the brainstem [37]. One study presented similar explanations for SWJ mechanisms, and proposed the hypothesis that the abnormality in supranuclear control might interrupt omnipause cell activity in paramedian pontine reticular formation, resulting in the release of saccadic burst units [43]. Another proposed mechanism for SWJs might be reflected in the degeneration involving the cerebellum and its connections, such as to the superior cerebellar peduncle, and result in abnormal to-and-fro eye movements during attempted steady fixation [19, 44]. This hypothesis is supported by the proposed relevance of SWJs in the impairment of visual scanning during attempt fixation among patients with cerebellar ataxia [45, 46].

Saccades are rapid eye movements that redirect the foveal line of sight toward features of interest so that they can be seen optimally [25]. The speed and range of saccades represent important components of saccadic examination, and their abnormalities reflect the dysfunction of the burst neurons (riMLF for vertical saccades and PPRF for horizontal saccades) that are frequently affected by the disease process in PSP [22, 23]. Since saccades are the fastest eye movements (up to 500°/s) with very brief duration (< 100 ms), the examiner should not expect to be able follow the full trajectory of normal saccades with their own eye [25]. The normal duration of a saccade is 30–100 ms for amplitudes ranging from 0.5 to 40° [47]. The larger the saccade, the higher the peal velocity. Saccades are slower when made in darkness to a remembered target location and when made in the opposite direction to a visual stimulus (the ‘antisaccade test’) [48].

In general terms, gaze palsy refers to deficits in conjugate eye movements that affect both eyes. Therefore, gaze palsy, when present, is usually indicative of supranuclear lesions resulting from interruption of the neural pathways that carry commands for voluntary saccades and pursuit before they reach the eye movement ‘generators’ in the brainstem. Dysfunction of the generators, namely the riMLF, results in vertical supranuclear gaze palsy, while dysfunction of the PPRF can lead to horizontal gaze palsy [22].

Testing for the velocity of saccades can be performed when the subject is asked to glance back and forth between two horizontal and vertical targets. Not only the velocity, but also the accuracy and conjugacy of the saccades should be observed. Normal individuals can reach the target with a single fast movement or with one small corrective saccade. It is important for the examiner to ensure that the subjects pay full attention during the performance of saccades, since patients who are drowsy, inattentive, or heavily medicated may produce normal saccades with slow velocity [49]. Moreover, there is a very strong age effect in the performance of saccades with a tendency for elderly subjects (60 years and over) to produce slower and delayed saccades, compared to younger individuals [50].

In order to determine slow saccades, researchers have utilized different quantitative methods of eye movement recordings, ranging from electrooculography, scleral search coil systems, and video-oculography. However, the identification of slow saccades on clinical examination is more challenging, particularly in patients with mildly slow saccades. If the examiner can follow the full trajectory of the patient’s saccades with their own eye, the saccades of that particular patient are likely to be slow [51]. Another useful clinical pearl is that patients with delayed saccadic initiation and slow saccades often blink or orient their head towards the target before making a saccade [52]. Different methods have been proposed to enhance the recognition of slow saccades, including the use of an optokinetic drum, or the use of oblique targets that will result in an L-shape saccade [52]. However, these methods have not been tested in patients with PSP to demonstrate their sensitivity and specificity. In suspected cases of PSP, particular attention should be paid to the velocity of vertical and horizontal saccades independently, since selective slow vertical saccades are observed in the early stage, followed by slow horizontal saccades when the disease is more advanced [23].

The range of motion of both horizontal and vertical saccades can be tested in a self-paced or verbally guided manner. Before determining that gaze palsy is pathological, it is important for physicians to be aware that a limitation of the vertical range of eye movements, especially upgaze, can occur in elderly subjects; the velocity should however not be affected in such cases [53]. In addition to gaze palsy, features suggesting supranuclear lesions should be specifically looked for. This can be accomplished by assessing improvements in the gaze palsy with a vestibuloocular reflex, thereby bypassing volitional pathways.

Out of 45 articles included in this systematic review, 10 articles studied slow saccades in patients with PSP [10, 23, 37, 38, 40, 41, 54‐57], and 19 articles studied VSP in patients with PSP [1, 4, 10, 15, 18, 19, 36, 37, 41, 54, 58‐66], identified using different eye recording techniques.

In the setting of parkinsonian syndromes, the identification of VSP is highly specific for the diagnosis of PSP, with the exception of a few case reports of dementia with Lewy bodies and CBD [67‐70]. While VSP constitutes one of the mandatory inclusion criteria for the clinical diagnosis of probable PSP, proposed by the NIND-SPSP [3], a real clinical problem is that not all patients with PSP exhibit the full range of eye movement abnormalities, particularly in the early stage [37, 42, 55, 56]. VSP should, however, not be used as an early sign in PSP, as it usually presents in the advanced disease course (range of follow-up periods: 1–10 years) [5, 19, 54, 60‐62, 66, 71]. In addition, some patients with PSP, especially those with variants who meet the clinic-pathologic criteria, might be diagnosed based on the pathological confirmation of the diagnosis without existing VSP or in spite of a lack of other eye movement abnormalities [4, 5, 59, 60, 63, 72‐75].

Therefore, the typically slow saccades with curved trajectories are giving a strong indication when observed in patients with early onset parkinsonism, but vertical saccades are more affected than horizontal saccades in PSP [10, 23, 38, 40, 54]. Additionally, both horizontal and vertical saccades are slow and have irregular trajectories and velocity profiles, but deficits have been found to be particularly severe in the vertical axis [57]. A number of clinicopathological studies have been able to correlate the presence of midbrain atrophy with the degeneration of burst neurons in the riMLF, clinically manifested as selective slow vertical saccades rather than horizontal saccades [22, 23]. Several studies of pathologically proven cases of PSP reported a range of 40–90% of patients that were documented to have vertical gaze palsy, implying that vertical gaze palsy was not evident in all patients [4, 5, 15, 58‐60, 66, 71, 76‐80]. For example, some variants of PSP including PSP-frontal, PSP-PLS, and PSP-SL show no eye movement abnormalities during standard clinical examination at an early stage of the disease [80‐82]. In an autopsy-proven case of PSP-PLS, vertical gaze palsy was absent throughout the 10-year disease course [81].

These pathologically studied cases provide strong evidence that, although vertical supranuclear palsy is relatively specific to PSP in the setting of parkinsonian syndromes, it is not a universal finding in this disorder. Additional involvement of cerebellar signs has been proposed for a new subtype of PSP called ‘PSP-cerebellum’ (PSP-C) [63, 66]. In this group of patients, eye signs can be under-recognized during the first 2 years [63, 66]. The real question to clinicians is what types of oculomotor abnormalities can help increase the diagnostic acumen for different variants and stages of PSP. The typical VSP in PSP can usually be partially improved with vestibule-ocular reflex (VOR) manoeuvres, which show an abnormal reciprocal inhibition of antagonist vertical muscles during attempt voluntary movements, while the reciprocal innervation is well preserved during reflex movements [64]. The supranuclear degeneration of the pathway from the substantia nigra pars reticulata to the superior colliculus may be an important cause of gaze palsy in PSP [83]. Our systematic review provides more evidence that slow saccades and vertical supranuclear palsy do not occur at once in patients with PSP. Rather, they tend to evolve as the disease progresses. One longitudinal study of oculomotor function in patients with PSP suggested that slow saccades occur throughout the disease course with preserved saccadic latency [84].

Indicating an incapability to make pure vertical saccades along a straight line in the midline and lateral arc, which creates a curve course of oblique saccades, the ‘round the houses’ sign for PSP was mentioned for the first time by Quinn in 1996 [85]. Using the analogy that each eye makes a round excursion in its own ‘house’ (orbit), the sign employs the plural ‘houses’. In addition to midbrain atrophy, the degeneration of the burst neurons in the riMLF likely explains the mechanism of early selective slow vertical saccades rather than horizontal saccades, and may also explain the occurrence of the ‘round the houses’ sign [23]. Furthermore, the paramedian pontine reticular construction is mainly responsible for creating horizontal saccades [22, 85, 86].

By prompting a patient to carry out quick eye movements to a moving target between a completely upright and downwards position while keeping the head stationary, a ‘round the houses’ sign can be detected. Eye motion following a curved course (oblique saccades), rather than a straight line, could also be an indication for a ‘round the houses’ sign [85, 86].

Two articles (one commentary and one case report) reporting on ‘round the houses’ signs in patients with PSP were identified [85, 86]. The signs were mentioned by Quinn in an old comment through his experience with several patients with PSP [85]. The curved course of oblique saccades might be a result of the viewer conducting vertical upward and downward saccades, resulting in an oval-like trajectory by inconsistent reduction of the vertical saccades compared to the horizontal saccades, evenly full gaze existed [85]. A ‘round the houses’ sign has thus been defined as an early eye sign to look for in patients with PSP [85]. It may also occur together with other signs as a late indicator in patients with advanced PSP, unless it results from vertical supranuclear gaze palsy [85]. The sign was recently described in Niemann-Pick disease type C (NPC) [87, 88], while dystonia, hyperreflexia, and hepatosplenomegaly are the usual indicators of adult-onset NPC [87].

By eliminating irritants from the surface of the cornea and conjunctiva as well as aiding to spread tears covering the globe surface to sustain a normally moist state, blinking and secretion of tears are basic protective mechanisms of the eyes [89]. Consisting of a swift lowering of an upper eyelid mediated by a short contraction of the orbicularis oculi (OO) muscle, eye blinking is linked to the reciprocal inactivation of the levator palpebrae superioris (LPS) muscle of the upper eyelid [89, 90].

Bilateral paroxysmal closing of the eyelids (duration: less than 1 s) is the definition of a blink that occurs without being triggered by an outside stimulus. By counting how many times a person blinks per minute during a 5-min conversation, a decreased eye blink rate can be identified [91]. Otherwise, digitally-recorded video images concentrated on the eyes may be utilized. When at least 50% closure of the eye happens during a quick closing and subsequent reopening of the eye, a blink is classified as a ‘full blink’. As prolonged eye closure is typically an intentional act in response to deep thought, it is usually excluded from assessments of blinks. Spontaneous blinks with under 50% of eyelid closure are not considered for the total blink rate [92].

Our methodical analysis showed that four studies found a noteworthy reduction in blink rates among patients with PSP in comparison to healthy controls in terms of unprompted as well as trained blink responses [91, 93‐95]. Additionally, patients with PSP exhibited much longer periods of closed eyes during unprompted blinking compared to the healthy controls [95]. Further, patients with PSP showed a significant reduction in their conditioned eye blink response and more verbal memory impairment, resulting in the proposal of a dissociated pattern of learning abilities. In a serial reaction time task, meanwhile, compared to controls, obvious sequence patterns were comparatively sustained [93]. A neurophysiological device employed to gauge brainstem excitability could be advantageous for separating patients with PSP from others with CBS, as found in a study on the benefits of the R2 blink reflex recovery cycle (R2BRRC) [96]. In this study, patients with PSP showed earlier employment of R2BRRC than patients with CBS [96].

Striatal and mesolimbic dopaminergic performance can be assessed via blinking rates. Alterations in rates have been found in various neuropsychiatric disorders, and are thought to come from irregular central dopaminergic functions [97]. Patients with schizophrenia, Tardive dyskinesia, Tourette’s syndrome, and Meige’s syndrome have shown increased blink rates, while those with Parkinson’s disease and PSP exhibit decreased blink rates [95, 97‐99].

Spontaneous, erratic, and strong contractions of the orbital, preseptal, and pretarsal segments of the orbicularis oculi are characteristic of blepharospasm (BSP) [94]. They are typical dystonic indicators of PSP, often related to visual disabilities including functional blindness [100, 101]. Idiopathic BSP and secondary BSP represent the two major forms. Considering the cooccurrence of BSP and other uncharacteristic physical findings including parkinsonism and dementia may aid physicians to make a conclusive diagnosis of PSP [102]. Limb and axial dystonia, oromandibular dystonia, torticollis, and dysfluency (dystonia of the articulation muscle) represent additional focal dystonias other than BSP in PSP [101, 103]. Botulinum toxin usually results in good responses in patients [102]. BSP can be differentiated from apraxia of eyelid opening (ALO) on the basis of dissimilar responses after botulinum neurotoxin treatment [104].

Usually resulting from a lack of supranuclear control of the levator palpebrae and orbicularis oculi muscles, ALO is marked by the incapacity to open one’s eyelid [105, 106]. The majority of patients with this condition stemming from a lack of eyelid elevation control tend to contract their frontalis muscles more strongly to open their eyes instead of using their orbicularis oculi muscles, causing a noticeable forehead wrinkle. In PSP, ALO is a fairly common form of eyelid disorder [37, 94]. If botulinum neurotoxin cannot improve BSP notwithstanding the construction of eyelid weakness, ALO is a significant factor [104]. Due to the selective contraction of the pretarsal section of the orbicularis oculi potentially being responsible for eyelid closure, apraxia is often mistaken [104].

A physician, typically a neurologist or ophthalmologist, is responsible for making a BSP diagnosis during a clinical assessment. While physical and neurological check-ups help offer a diagnosis, it is usually compulsory for the physician to verify the information given by the patient. By prompting a patient to move his/her face into both spontaneous and trained positions based on the physician’s directives, BSP can be recognized. By asking patients to shut their eyes forcefully, physicians can assess the bilateral contractions of the orbicularis oculi muscles. After patients open their eyes again, BSP may be observed as a continuing contraction of the orbicularis oculi muscles, related to a sinking of the eyebrows under the superior orbital rim (Charcot’s sign) [107].

By asking patients to close their eyes, ALO can be recognized. Patients are asked to open their eyes again quickly after closing them. Any incapacity to lift the eyelid without major orbicularis oculi muscle contraction may also be classified as ALO [105, 108]. While a patient is starting to open their eyes, excessive frontalis muscle contraction is typically observed as a forehead wrinkle.

More prevalent in atypical parkinsonism than PD, BSP is a significant characteristic eye sign in PSP [37, 102, 109]. In varying levels of acuteness, nine articles have described BSP in patients with PSP as follows [37, 59, 61, 94, 100‐102, 109, 110].

BSP in PSP could be serious and cause acute visual disability [94, 100]. The intensity and occurrence of BSP may also increase under certain conditions, as found in these studies. These conditions include exposure to bright light, watching television, reading, stress, and driving. Prior to the emergence of typical blepharospasm, more than one-third of patients with PSP suffer from excessive eye blinking [100]. Basic visual tasks could improve BSP, while direct commands commonly have no effect, as found by past research [94]. BSP in PSP leads to deteriorating changes involving the upper brainstem, without a true clinicopathological link to particular sites, based on a clinicopathological study [59]. As the result of dopaminergic medication, however, dystonia and BSP have been recounted in PSP [101, 109]. Levodopa administration was found to be linked to roughly 24% of cases with BSP in patients with PSP [101]. Further, previous research has found that BSP incidences occur late in the disease as well as at the same time with ALO [37, 61, 94, 102, 109]. Thus, the overlap of BSP and ALO could be an indication of PSP [109]. Similarly to BSP stemming from other causes, BSP in PSP typically responds to botulinum toxin injections [37]. Nonetheless, botulinum toxin injections may not achieve a full response in patients with PSP with isolated BSP when considering who had both BSP and ALO conditions [110].

Concerning research on ALO in patients with PSP, 14 articles have been identified as follows [18, 37, 58, 59, 62, 94, 105, 106, 108, 111‐115].

ALO initially gained focus in 1965 as a non-paralytic abnormality characterized by difficulty in performing eyelid elevation [105]. An interruption in supranuclear control of the levator palpebrae and orbicularis oculi muscles may be a main contributor of ALO, causing difficulty in opening the eyes [106]. The benchmarks of ALO can be described as follows: (a) temporary incapability to raise the eyelids; (b) lack of progressive orbicularis oculi muscle contractions; (c) vigorous frontalis muscle contraction combined with incapacity to raise the eyelids; and (d) lack of oculomotor or ocular sympathetic nerve dysfunction as well as lack of ocular myopathy [94]. The most frequent eye sign in patients with PSP who fulfil the criteria of NINDS-SPSP for identification of PSP continues to be ALO, as long established in pathologically established cases of PSP [18, 58, 59, 62, 111]. Although ALO is a belated occurrence in PSP, in one out of six PSP cases it may represent the initial clinical presentation of PSP, at an average of 3.5 years after disease onset [62]. Further, ALO resulting from dopaminergic medications has been revealed in patients with parkinsonism [114, 115]. In a group of 32 patients with ALO of assorted aetiologies who reacted well to botulinum toxin injections, seven were patients with PSP [108]. This study found that ALO may be caused by focal eyelid dystonia instead of true apraxia, because of positive electrophysiologic findings with remote spasms of the pretarsal segment absent but complete contribution of the orbicularis oculi [108]. An incapability to constrain the voluntary contraction of the pretarsal portion of the orbicularis oculi muscle without clinically ongoing contraction of the orbicularis oculi muscle was detected in three patients with eyelid opening apraxia, termed ‘motor persistence of orbicularis oculi muscle’, comparable to the findings in a case study involving three patients with PSP. Subsequent to botulinum toxin injections into their eyelids, all three patients recovered [112]. The hypothesized eyelid dystonia in a report of two patients with PSP with ALO using a suggested injection site was restricted to the joint of the preseptal and pretarsal part of the orbicularis oculi muscles, as verified by the effective treatment response [113]. However, the benefits of botulinum toxin injections were identified in patients with a combination of ALO and BSP only, as proposed by a contrary study [37].