Neuro-Ophthalmologic Features and Outcomes of Thalamic Infarction: A Single-Institutional 10-Year Experience

Abstract

Background: 

Neuro-ophthalmologic deficit after thalamic infarction has been of great concern to ophthalmologists because of its debilitating impacts on patients' daily living. We aimed to describe the visual and oculomotor features of thalamic infarction and to delineate clinical outcomes and prognostic factors of the oculomotor deficits from an ophthalmologic point of view.

Methods: 

Clinical and neuroimaging data of all participants were retrospectively reviewed. Among the 12,755 patients with first-ever ischemic stroke, who were registered in our Stroke Data Bank between January 2009 and December 2018, 342 were found to have acute thalamic infarcts on MRI, from whom we identified the patients exhibiting neuro-ophthalmologic manifestations including visual, oculomotor, pupillary, and eyelid anomalies.

Results: 

Forty (11.7%) of the 342 patients with thalamic infarction demonstrated neuro-ophthalmologic manifestations, consisting of vertical gaze palsy (n = 19), skew deviation with an invariable hypotropia of the contralesional eye (n = 18), third nerve palsy (n = 11), pseudoabducens palsy (n = 9), visual field defects (n = 7), and other anomalies such as isolated ptosis and miosis (n = 7). Paramedian infarct was the most predominant lesion of neuro-ophthalmologic significance, accounting for 84.8% (n = 28) of all patients sharing the oculomotor features. Although most of the patients with oculomotor abnormalities rapidly improved without sequelae, 6 (18.2%) patients showed permanent oculomotor deficits. Common clinical features of patients with permanent oculomotor deficits included the following: no improvement within 3 months, combined upgaze and downgaze palsy, and the involvement of the paramedian tegmentum of the rostral midbrain.

Conclusions: 

Thalamic infarction, especially in paramedian territory, can cause a wide variety of neuro-ophthalmologic manifestations, including vertical gaze palsy, skew deviation, and third nerve palsy. Although most oculomotor abnormalities resolve spontaneously within a few months, some may persist for years when the deficits remain unimproved for more than 3 months after stroke.

Thalamic infarction accounts for approximately 4% of all cerebral infarctions and 11% of all vertebrobasilar infarctions. It gives rise to a wide range of neurologic syndromes depending on the arterial territories and associated nuclei. In addition to classical sensorimotor and amnestic syndromes, thalamic infarction can lead to the development of numerous visual and eye movement dysfunctions, either in isolation or in combination with infarction involving other structures (^1–8). Because of its immediate proximity to the rostral midbrain and thalamomesencephalic junction, the critical area for supranuclear premotor control of vertical eye movements, thalamic infarction can produce various eye movement dysfunction. However, apart from the effects caused by the anatomical vicinity to the midbrain, the thalamus alone can cause several oculomotor deficits. Another clinical characteristic of patients with thalamic infarction is visual field defects mainly due to the involvement of the lateral geniculate nucleus (^4,9–12).

Visual and oculomotor deficits after thalamic infarction have been a subject of great concern because of the resultant disease burden and serious disability affecting everyday activities. However, great uncertainty is associated with not only the underlying mechanisms of the oculomotor deficits but also its potential future clinical courses and the best method of management. Although several studies have been conducted to investigate clinical features of thalamic infarction, only a few have addressed its neuro-ophthalmologic features in detail, even in case series studies of thalamic infarction with small samples (^13–15).

The purpose of this study was to describe the visual and oculomotor features of patients who presented to our institution with thalamic infarction over a period of 10 years and to delineate the clinical outcome and prognostic factors of the ocular motility deficits from an ophthalmologic point of view.

METHODS

We retrospectively reviewed the clinical and neuroimaging data of all the patients registered in our Stroke Data Bank between January 2009 and December 2018. Among the 12,755 registered patients, 342 (2.68%) were identified to have thalamic infarction: 282 with isolated thalamic lesions and 60 with additional extrathalamic lesions. We finally included the patients with thalamic infarction who had neuro-ophthalmologic features including visual, oculomotor, pupillary, and eyelid abnormalities.

MRI including diffusion-weighted imaging and MRA was performed in all patients with thalamic infarction within 24 hours of the onset of symptoms, and a neuro-radiologist independently read the images. The topography of thalamic infarction was categorized into 4 groups depending on their arterial supply: (1) paramedian infarction in the territory of the paramedian artery, (2) polar infarction in the territory of the polar artery, (3) inferolateral infarction in the territory of the inferolateral artery, previously called thalamogeniculate artery, and (4) posterior choroidal infarction in the territory of the posterior choroidal artery (^1,2,4,7).

The etiology of infarction was assessed according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification criteria, combining risk profile of the patients with the results of the following tests: complete blood cell count, coagulation laboratory test, 12-lead electrocardiography, transthoracic or transesophageal echocardiography, duplex sonography of the carotid arteries, MRI, and MRA. The causes of thalamic infarction were classified into 5 subgroups: (1) large-artery atherosclerosis, (2) cardioembolism, (3) small-vessel occlusion, (4) stroke of other determined etiology, and (5) stroke of undetermined etiology (¹⁶).

At the initial visit, a neurologist performed a thorough neurological examination. Visual field tests were performed by using a Humphrey field analyzer (Swedish Interactive Threshold Algorithm [SITA] 30-2; Carl Zeiss Meditec, Dublin, CA) or a by confrontation method depending on the patient's co-operability. A comprehensive evaluation of external and internal oculomotor functions, including ocular alignment, duction and version, pupillary and eyelid functions, and involuntary eye movement, was performed. A detailed analysis of patient data on neuro-ophthalmologic features and outcomes, including both visual and oculomotor aspects, was performed.

This study adhered to the tenets of the Declaration of Helsinki, and the institutional review board of the Asan Medical Center approved the protocol.

RESULTS

Among the 342 patients with thalamic infarction, 40 (11.7%) patients developed neuro-ophthalmologic manifestations. The mean age was 59.2 ± 14.4 years (range, 21.4–80.4 years), and 16 men were included. Most of the patients had other combined infarcts, and only 8 (20.0%) patients showed isolated thalamic infarcts. The demographic data and distribution of lesions are summarized in Table 1. Among the 40 patients, 11 patients (10 patients with paramedian infarctions and 1 patient with polar infarction) were notified to have consciousness changes, and 2 patients (1 patient with paramedian infarction and 1 patient with polar infarction) showed memory deficits.

TABLE 1. - Demographics, causes of stroke, and distribution of lesions in patients with neuro-ophthalmologic manifestations of thalamic infarction

	Paramedian Infarction (n = 28)	Polar Infarction (n = 2)	Inferolateral Infarction (n = 5)	Posterior Choroidal Infarction (n = 5)
Age (year)
Mean ± SD	58.0 ± 14.9	66.0 ± 6.2	67.9 ± 11.4	54.7 ± 15.0
Range	21.4–78.4	60.6–69.4	55.0–80.4	33.9–70.8
Sex (M: F)	9: 19	1: 1	4: 1	2: 3
Hypertension, no. (%)	12 (42.9)	1 (50.0)	3 (60.0)	2 (40.0)
Diabetes mellitus, no. (%)	5 (17.9)	1 (50.0)	3 (60.0)	0
Hypercholesterolemia, no. (%)	6 (21.4)	0	2 (40.0)	2 (40.0)
Cardiac disease, no. (%)	18 (64.3)	1 (50.0)	0	2 (40.0)
Atrial fibrillation*	10	1		1
Valve disease*	5	0		1
Congenital heart disease	5	0		0
Myocardial infarction	1	0		0
Other	1	0		0
Malignancy	4 (14.3)	0	1 (20.0)	0
Smoking (never: ex-smoker: Yes)	20: 2: 6	0: 0: 1	3: 1: 1	2: 0: 3
Laterality (unilateral: Bilateral)	21: 7	2: 0	4: 1	5: 0
Cause of stroke, no. (%)
Large-artery atherosclerosis	7 (25.0)	1 (50.0)	5 (100.0)	2 (40.0)
Cardioembolism	14 (50.0)	0	0	2 (40.0)
Small-vessel occlusion	3 (10.7)	1 (50.0)	0	0
Other determined	0	0	0	0
Undetermined	4 (14.3)	0	0	1 (20.0)
Infarct location, no. (%)
Thalamus only	7 (25.0)	1 (50.0)	0	0
Midbrain	19 (67.9)	0	1 (20.0)	0
Cerebellum	5 (17.9)	0	2 (40.0)	1 (20.0)
Occipital lobe	1 (3.7)	0	3 (60.0)	3 (60.0)
Temporal lobe	1 (3.7)	0	1 (20.0)	2 (40.0)
Parietal lobe	1 (3.7)	0	1 (20.0)	0
Frontal lobe	1 (3.7)	0	0	0
Basal ganglia and internal capsule	0	0	2 (40.0)	0
Others	0	1† (33.3)	0	1‡ (50.0)

^*Four patients had atrial fibrillation related to cardiac valve disease.

^†Hypothalamus.

^‡Pons.

Among the 342 patients with thalamic infarction, 54 patients with paramedian infarction were identified; among them, 28 (51.9%) were discovered to have oculomotor abnormalities. The most common oculomotor abnormality was vertical gaze palsy, followed by skew deviation with an invariable hypotropia of the contralesional eye. Ipsilateral ptosis without third cranial nerve palsy was found in 3 patients, one of whom had associated ipsilateral miosis and the other 2 had normal pupillary function. Among the 55 patients with polar infarction, 2 (3.6%) demonstrated ipsilateral ptosis without miosis as a sole neuro-ophthalmological feature. Among 217 patients with inferolateral infarction, only 5 (2.3%) presented with significant neuro-ophthalmological signs, and among 38 patients with posterior choroidal infarction, 5 (13.2%) had the following neuro-ophthalmologic manifestations: one-and-a-half syndrome in a patient with multiple infarcts in the thalamus, pons, and temporal lobe and visual field defects in 4 patients (Fig. 1 and Table 2).

FIG. 1.: This flow diagram shows the disposition of patients with ischemic stroke according to the affected vascular territory within the thalamus.

TABLE 2. - Neuro-ophthalmologic features of thalamic infarction according to the affected vascular territory

	Paramedian infarction (n = 28)	Polar Infarction (n = 2)	Inferolateral Infarction (n = 5)	Posterior Choroidal Infarction (n = 5)
Vertical gaze palsy, no. (%)	19 (67.9)	0	0	0
Combined upgaze and downgaze	7
Upgaze	9
Downgaze	3
Skew deviation, no. (%)	18 (64.3)	0	0	0
Ipsilateral III cranial nerve palsy, no. (%)	10 (35.7)	0	1 (20.0)	0
Pseudoabducens palsy, no. (%)	9 (32.1)	0	0	0
Gaze-evoked nystagmus, no. (%)	3 (10.7)	0	1 (20.0)	0
Ipsilateral ptosis,* no. (%)	3 (10.7)	2 (100.0)	0	0
Ipsilateral miosis, no. (%)	2 (7.1)	0	0	0
Contralateral wrong-way deviation, no. (%)	1 (3.6)	0	0	0
One-and-a-half syndrome, no. (%)	0	0	0	1 (20.0)
Visual field defect, no. (%)	0	0	3 (60.0)	4 (80.0)
Hemianopia			2	1
Quadrantanopia			1	3

^*Not associated with III cranial nerve palsy.

According to the range of infarction on diffusion-weighted neuroimaging, we classified our 40 patients into a isolated thalamic infarction group and combined thalamic infarction group. The 8 patients with isolated thalamic infarction (7 with paramedian infarction and 1 with polar infarction) did not fail to show significant neuro-ophthalmologic signs (Table 3).

TABLE 3. - Neuro-ophthalmologic features in patients with isolated thalamic infarction

	Paramedian Infarction (n = 7)	Polar Infarction (n = 1)
Vertical gaze palsy, no. (%)	6 (85.7)	0
Combined upgaze and downgaze	0
Upgaze	4
Downgaze	2
Contralateral skew deviation, no. (%)	4 (57.1)	0
Ipsilateral III cranial nerve palsy, no. (%)	0	0
Pseudoabducens palsy, no. (%)	5 (62.5)	0
Gaze-evoked nystagmus, no. (%)	0	0
Ipsilateral ptosis,* no. (%)	1 (12.5)	1 (100.0)
Ipsilateral miosis, no. (%)	1 (12.5)	0
Contralateral wrong-way deviation, no. (%)	0	0
One-and-a-half syndrome, no. (%)	0	0
Visual field defect, no. (%)	0	0
Hemianopia
Quadrantanopia

^*Not associated with III cranial nerve palsy.

Oculomotor abnormalities, the most prevalent neuro-ophthalmologic manifestations of thalamic infarction, resolved rapidly within 3 months of onset in 22 cases (66.7%). However, if the abnormalities did not spontaneously resolve within 3 months, they remained stable for longer than 1 year. Consequently, except for 5 cases who could not confirm the duration of oculomotor abnormalities, 6 (18.2%) of the 33 patients maintained their oculomotor abnormalities far beyond 1 year after onset (average duration: 26 months). Their clinical characteristics are summarized in Table 4.

TABLE 4. - Detailed clinical characteristics of 6 patients with permanent oculomotor deficits

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6
Sex	M	F	F	F	M	M
Age (yr)	53.2	31.5	61.8	47.2	31.3	64.3
Risk factors	DM, HTN, HL, CHD	CHD	HTN, Afib	Valve disease, Afib	Hypercholesterolemia	Afib
TOAST classification	CE	CE	CE	CE	LAA	CE
Laterality	Unilateral	Bilateral	Unilateral	Unilateral	Unilateral	Bilateral
Topographic lesion	Paramedian artery	Paramedian artery	Paramedian artery	Paramedian artery	Paramedian artery	Paramedian artery
Other infarct lesions	Midbrain	Midbrain	Midbrain, cerebellum	Midbrain, cerebellum, temporal lobe	Midbrain, cerebellum	Midbrain
Oculomotor deficits	Vertical gaze palsy, skew deviation	Vertical gaze palsy, skew deviation, 3rd nerve palsy	Vertical gaze palsy, skew deviation, 3rd nerve palsy, gaze-evoked nystagmus	Vertical gaze palsy, skew deviation, 3rd nerve palsy	Vertical gaze palsy, skew deviation	Vertical gaze palsy, skew deviation, gaze-evoked nystagmus

Afib, atrial fibrillation, CE, cardioembolism, CHD, congenital heart disease, DM, diabetes mellitus, F, female, HTN, hypertension, HL, hypercholesterolemia, LAA, large-artery atherosclerosis, M, male, TOAST, Trial of Org 10,172 in Acute Stroke Treatment.

We investigated the common characteristics shared by these 6 patients with persistent oculomotor deficits. Their one common characteristic was that all were diagnosed to have combined upgaze and downgaze palsy as well as ocular tilt reaction at their initial ophthalmologic evaluation. Among the 7 patients with combined upgaze and downgaze palsy, 6 finally remained symptomatic. By contrast, 9 patients with upgaze palsy and 3 with downgaze palsy recovered fully within 3 months and remained free of ocular symptoms thereafter. Another clinical characteristic feature shared by these 6 patients was the involvement of the rostral midbrain as elucidated by diffusion-weighted imaging. Specifically, the paramedian tegmentum of the rostral midbrain, which is considered an area responsible for coordinating vertical eye movements, was invariably involved in all 6 patients as a visually obvious signal hyperintensity (Fig. 2).

FIG. 2.: Diffusion-weighted MRI scans of 6 patients with persistent oculomotor deficits. The axial view of the rostral midbrain at the same level delineates the extent of hyperintense lesion in each patient, uniformly involving the paramedian tegmentum of the rostral midbrain (white arrows).

DISCUSSION

Of the 12,755 patients registered in our stroke database, 342 (2.7%) were identified to have thalamic infarction, which was consistent with previous studies showing that thalamic infarction constituted 3.1%–4.4% of all infarctions (^4,6). Among those with thalamic infarction, neuro-ophthalmologic features were observed in 40 (11.7%) patients.

Oculomotor abnormalities were most frequently associated with infarctions in the territory of the paramedian artery, and the most common feature was vertical gaze palsy. The frontofugal dorsothalamic bundle transverses the mediodorsal nucleus and internal medullary lamina of the thalamus, and then, it reaches the superior colliculus. The development of vertical gaze palsy can be attributed to a lesion involving this bundle (^4,17). Lesions involving the rostral interstitial nucleus of the medial longitudinal fasciculus and posterior commissure in the rostral midbrain can also lead to the development of vertical gaze palsy (^17–19). In our study, 19 (67.9%) patients with paramedian infarction showed vertical gaze palsy, and 6 of them did not have any lesions in the midbrain on MRI.

A previous study reported that unilateral thalamic infarction mainly causes upgaze palsy and bilateral infarction causes upgaze, downgaze, or combined vertical gaze palsy (^12,20). In our study, however, 6 of the 21 patients with unilateral paramedian infarction showed combined upgaze and downgaze palsy, and only 1 of the 7 patients with bilateral paramedian infarction showed combined upgaze and downgaze palsy. Thus, unilaterality or bilaterality did not seem to affect the extent and range of vertical gaze palsy in this study.

Skew deviation can occur with lesions anywhere in the vestibulo-ocular pathway, which is also believed to traverse the thalamus. As with several previous studies, skew deviation was found in this study in more than half of the patients with paramedian thalamic infarction (^21,22). All of these skew deviations were associated with hypotropia of the contralateral eye, contraversive conjugate ocular torsion, and head tilt in the direction contralateral to the side of the thalamic lesion as a typical presentation of contralateral ocular tilt reaction.

Pseudoabducens palsy, which is known as thalamic esotropia, is a tonic inward ocular deviation (²³). The prefrontal descending cortico-oculomotor pathway for binocular convergence passes through the paramedian portion of the thalamus and the mesodiencephalic junction and terminates in oculomotor complexes. This connection exists bilaterally, and contralateral premotor neurons exert an inhibitory influence on the neurons of the oculomotor nucleus. Acute disinhibition of these neurons would result in the tonic activation of the medial rectus and cause pseudo-abducens palsy (^24–26). Among our 40 patients, 8 patients with paramedian infarction were identified to have pseudoabducens palsy; however, in this retrospective study, it could not be determined whether the esotropia was contralateral or ipsilateral to the side of the lesion.

Gaze-evoked nystagmus was detected in 4 patients. Three of them presented with a typical horizontal gaze-evoked nystagmus occurring in lateral gazes, suggestive of a defective gaze-holding mechanism. The 3 patients were found to have lesions in the cerebellum as well as the thalamus. The remaining one male patient showed a gaze-evoked nystagmus in vertical gaze. He also showed vertical gaze palsy, ocular tilt reaction, and vertical-torsional nystagmus in primary gaze. His brain MRI revealed a bilateral paramedian thalamic infarction. The direction of his ocular tilt reaction and torsional nystagmus was contralateral to the more severe side of the bilateral lesion. All of these findings suggested that the combined lesion involving both the rostral interstitial nucleus of the medial longitudinal fasciculus and the interstitial nucleus of Cajal was the underlying cause of his ocular motility abnormality (^27–29).

In our study, isolated ipsilateral ptosis without ophthalmoplegia was observed in 3 patients with paramedian infarction and 2 patients with polar infarction. All of them had unilateral thalamic infarctions and unilateral mild ptosis. Ipsilateral ptosis has already been reported in patients with thalamic infarction, and some plausible mechanisms for the development of this ptosis without ophthalmoplegia have been proposed (^30,31). First, Horner syndrome, in either complete or incomplete form, can be considered as a cause of ptosis in this clinical situation. Horner syndrome can occur from anterior thalamic lesion, which is explained by the anatomical proximity between the anterior thalamus and the hypothalamus. Furthermore, the paramedian and polar arteries supply not only the anterior thalamus but also the hypothalamus from which the sympathicoexcitatory tract originates (^31,32). In this study, one patient with paramedian infarction showed ipsilateral ptosis and miosis. Although there is uncertainty on whether or not the ptosis was associated with ipsilateral anhidrosis and although pharmacologic tests could not be performed, it seemed reasonable to assume that Horner syndrome was the most likely diagnosis, considering that the infarction involved the hypothalamus and the signs suggesting third nerve palsy were absent. Second, lesions in the thalamus can cause “cerebral ptosis” (^31,33). Several studies have reported that the internal medullary lamina, which is a strip of white matter in the thalamus, is connected to the frontal eye fields, supplementary eye fields, and posterior parietal cortex (^34,35). Thus, lesions involving this connection can induce cerebral ptosis. Cerebral ptosis was reported to be unilateral or bilateral; if bilateral, it may be either symmetric or asymmetric (³⁶). Among our subjects, 4 patients were suspected to have cerebral ptosis, all of whom had no associated pupillary miosis or ophthalmoplegia.

Wrong-way deviation, a conjugate eye deviation directed contralaterally to the affected side, was identified in 1 patient with paramedian infarction. This deviation is known to be associated with supratentorial hemorrhagic or extensive ischemic stroke (^37,38). In our patient, the lesion in the thalamus was rather small, whereas the associated infarction of middle cerebral artery territory was shown to be extensive in diffusion-weighted MRI. Thus, it was considered that the wrong-way deviation in our patient was more likely due to the large hemispheric infarction than the thalamic infarction itself.

Although the oculomotor abnormalities improved early in the course of the disease, it was consistently found that if the abnormalities did not spontaneously resolve within 3 months, they remained stable for longer than 1 year. Therefore, it would be reasonable to suggest that if the oculomotor abnormalities in patients with thalamic infarction persist for more than 3 months, it is likely to become a permanent deficit without further improvement. In addition, it would also be sensible to suggest, based on the findings of this study, that the same outcome can be expected when combined upgaze and downgaze palsy is observed at the initial presentation and when the paramedian rostral midbrain is obviously affected.

Visual field defects were observed in 7 patients in this study. All the 7 patients, except 1, had thalamic and extrathalamic lesions, with the latter being responsible for the visual field defects. The only 1 patient with posterior choroidal infarction and no apparent extrathalamic lesion presented with homonymous sectoral quadrantanopia in the contralateral inferior visual fields. This defect was considered to be induced by a lesion involving the lateral geniculate body, which was consistent with other previous reports (³⁹). Hence, considering the significant number of patients without visual field defects from our data, it can be inferred that visual field defects are not commonly encountered in patients with thalamic infarction, even in those with posterior choroidal infarction.

This study has several limitations that need to be acknowledged. One of the major limitations is its retrospective nature and the incomplete documentation, especially of detailed clinical outcomes after the infarction. Another limitation is related to the fact that visual fields were not always measured using Humphrey automated perimetry, but rather tested by a confrontation method in some of our subjects. Despite these limitations, this study has several noteworthy strengths. First, we used a well-established stroke database with a highly standardized and structured patient assessment protocol, which provided comprehensive data from a homogenous clinical setting over a long period of time. Second, this study included a substantial number of patients. In combination with a long-term study period, the large study population made it possible to explore the incidence, clinical features, and outcomes of neuro-ophthalmologic manifestations. Finally, to the best of our knowledge, this is the first study to investigate the prognostic factors of persistent oculomotor deficits in patients with thalamic infarction.

In conclusion, thalamic infarction can show multiple and various neuro-ophthalmologic manifestations. The most frequent neuro-ophthalmologic finding was vertical gaze palsy, followed by skew deviation. Paramedian infarction was the most common type of infarction causing oculomotor abnormalities. Although in most cases the abnormalities resolve spontaneously within a few months, in some cases, persistence of deficits was noted when the oculomotor abnormalities remained unimproved for more than 3 months.

STATEMENT OF AUTHORSHIP

Category 1: a. Conception and design: H. T. Lim; b. Acquisition of data: Y. Moon, K. S. Eah, E.-J. Lee, D.-W. Kang, S. U. Kwon, and J. S. Kim; c. Analysis and interpretation of data: Y. Moon and H. T. Lim; Category 2: a. Drafting the manuscript: Y. Moon and H. T. Lim; b. Revising it for intellectual content; E.-J. Lee, D.-W. Kang, S. U. Kwon, and J. S. Kim; Category 3: a. Final approval of the completed manuscript: H. T. Lim.