Peripapillary Nerve Fiber Layer Thickening, Telangiectasia, and Retinal Hemorrhages in Wernicke Encephalopathy

Abstract

A young woman who underwent gastric bypass surgery for morbid obesity had intractable nausea and vomiting for several weeks postoperatively, leading to poor intake and excessive weight loss. In the ninth postoperative week, she became confused and off balance and reported blurred and double vision. Examination disclosed slow saccades, nystagmus, and impaired abduction of both eyes as well as memory loss and ataxia. Visual acuity was slightly subnormal, and ophthalmoscopy disclosed a thickened and telangiectatic peripapillary nerve fiber layer with retinal hemorrhages. MRI showed high T2 and FLAIR signal in the dorsomedial thalamus and mamillary bodies bilaterally, substantiating a clinical diagnosis of Wernicke encephalopathy (WE). After thiamine treatment, visual acuity returned to normal and eye movements and alignment almost completely normalized. Fundus abnormalities eventually regressed. Although the ocular motor findings of WE have been well documented, the ophthalmoscopic findings have not. Resembling the findings in Leber hereditary and toxic optic neuropathies, they may represent manifestations of impaired mitochondrial function in retinal ganglion cells and capillaries. Recognition that these ophthalmoscopic findings may occur in WE is important to avoid procedures such as lumbar puncture that may delay urgent treatment with thiamine.

A 22-year-old woman underwent Roux-en-Y gastric bypass surgery for morbid obesity (BMI 63 kg/m²). After the surgery, she had intractable nausea and vomiting with poor oral intake, leading to an 80-lb weight loss over 2 months. She did not take vitamin supplements as instructed. In the ninth postoperative week, she became confused, developed difficulty with balance, and noted blurred and double vision. Apart from myopia, she had no ophthalmic problems and there were no preexisting neurologic problems.

On our bedside examination, best-corrected visual acuity at near was 20/40 in the right eye and 20/25 in the left eye. Pupils, confrontation visual fields, and intraocular pressures were normal. She displayed saccadic smooth pursuit and slow saccades with bilateral abduction deficits and substituted convergence movements on attempted lateral gaze (Video, Supplemental Digital Content 1,https://links.lww.com/WNO/A5. Pretreatment clinical ocular motility examination, impaired by poor cooperation, shows normal alignment in primary gaze position with impaired abduction bilaterally, substituted convergence movements on attempted side gaze, and upbeat nystagmus.). Sidebeat, upbeat, and downbeat nystagmus were present in extremes of gaze. The eyes were aligned in primary gaze position. Ophthalmoscopic examination disclosed thickening and telangiectasia of the peripapillary nerve fiber layer in both eyes. Superficial retinal hemorrhages were present near the optic discs in both eyes (Fig. 1A). She was oriented to time and place, had no language deficits but had difficulty with recall of recent events. Her gait was wide-based.

FIG. 1: A. At presentation, fundus photographs show thickening and telangiectasia of the nerve fiber layer and mirror-image superficial retinal hemorrhages. B. Four weeks after presentation, fundus photographs show partial resolution of the findings in A.

Optical coherence tomography (OCT) demonstrated a thickened retinal nerve fiber layer superiorly and inferiorly in both eyes (Fig. 2). Brain MRI showed T2 and FLAIR hyperintensity in the dorsomedial thalami and mamillary bodies (Fig. 3). The thiamine level was 6.0 nmol/L (normal 4.5-15.1 nmol/L). Based on the clinical and imaging findings, a diagnosis of Wernicke encephalopathy (WE) was made.

FIG. 2: At presentation, optical coherence tomography of the retinal nerve fiber layer shows thickening of the superior and inferior nerve fiber layers bilaterally.

FIG. 3: At presentation, axial MRI FLAIR images show hyperintensity of the mamillary bodies (arrow, A) and medial thalami (arrow, B).

The patient was treated with 100 mg/day thiamine intravenously for 7 days followed by 100 mg/day thiamine orally. Within 24 hours, she had marked improvement in mentation and vision. After 48 hours of treatment, distance visual acuity was 20/20 in both eyes. Saccadic and pursuit eye movements were now normal (Video, Supplemental Digital Content 2,https://links.lww.com/WNO/A6. After 48 hours of treatment with 100 mg/day thiamine intravenously, the patient has more normal gaze but still has sidebeat and upbeat nystagmus.). The upbeat nystagmus had resolved, but she still had sidebeat nystagmus on left and right gaze. Formal visual fields (Humphrey 24-2 protocol) were normal in both eyes. Ophthalmoscopic examination showed no change. The patient continued to have difficulty ambulating due to ataxia. Two weeks after initiation of treatment, the patient's ophthalmologic and neurologic examinations were essentially unchanged. Four weeks afterwards, eye movements were essentially normal (Video, Supplemental Digital Content 3,https://links.lww.com/WNO/A7. Four weeks after onset of the clinical manifestations of Wernicke encephalopathy and treatment with thiamine, the patient's eye movements are normal.) and ophthalmoscopic abnormalities were resolving (Fig. 1B).

WE consists of a triad of altered mental status, ataxia, and eye movement abnormalities (nystagmus and ophthalmoplegia) (¹). The pathologic substrate is neuronal loss, hemorrhage, and capillary proliferation principally in mamillary bodies, dorsomedial thalamus, and periaqueductal gray matter of the midbrain and pons. (Fig. 4). These phenomena are attributed to thiamine (vitamin B₁) deficiency. In the past, this deficiency most often arose in chronic alcoholism (¹). More recently it has been described in the nutritional deficiency that may follow bariatric surgery (^2-8). A rare complication in this setting, WE was documented in 84 cases in a 2008 review (⁵). Its manifestations are typically noted 4-12 weeks postoperatively. Risk factors include poor oral intake due to nausea and vomiting, rapid weight loss, and glucose administration without thiamine supplementation (^5,7).

FIG. 4: Alcoholic Wernicke encephalopathy in autopsied cases. A. Whole-mount section of walls of the third ventricle shows necrosis of medial thalamic nuclei (pale areas, arrows). B. Histologic section from a patient with acute WE shows ring of hemorrhages (arrows) around a blood vessel. C. In the mamillary body of a patient with subacute WE, there are dilated blood vessels (white arrow), hypertrophied vascular endothelium (arrowhead), and preserved neurons (black arrows). (Modified from Graham DI, Lantos PL, eds. Greenfield's Neuropathology. Vol 1, 7th ed. London: Arnold; 2002:613.)

Although the ocular motor abnormalities of WE are well documented, the optic fundus findings are not (^9-12). In his original description of 3 alcoholic patients, Wernicke described “bilateral optic neuritis with massive swelling and many streaky hemorrhages” in 1 patient and “redness of the optic discs” in the other (^1,13). In their 1954 report on the ocular signs in 5 patients with WE (4 from alcohol and 1 from malnutrition), Cogan and Victor (⁹) found no optic fundus abnormalities. In an authoritative 1971 review of 245 alcoholic patients with WE by Victor et al (¹), 6 (2%) patients were described as having retinal hemorrhages but none as having optic disc edema. No fundus photographs were published in these reports.

There have been 3 single case reports of the optic fundus findings in WE associated with bariatric surgery (^2-4). Mumford (²) described a 24 year-old woman with hyperemesis gravidarum who presented with findings virtually identical to those of our patient, namely impaired consciousness, ataxia, vertical and horizontal nystagmus, and reduced abduction in both eyes. Ophthalmoscopy revealed “marked bilateral papilledema with capillary dilatation and peripapillary flame hemorrhages.” No photographs of the fundus findings were published. Because the optic fundus findings were initially interpreted as consistent with increased intracranial pressure, she underwent brain CT, which was normal, and a lumbar puncture, which showed a normal opening pressure and constituents. It was only after these studies returned normal results that a diagnosis of WE was made. In the meantime, the patient was receiving an intravenous glucose infusion without thiamine and becoming comatose.

Kramer and Locke (³) reported “peripapillary hemorrhage and blurred disc margins” in a patient who had developed visual difficulty 2 months after bariatric surgery. She was also described as having “coarse horizontal nystagmus on lateral gaze and coarse vertical nystagmus varying with direction of gaze.” The neurologic examination was otherwise affected by psychiatric problems and therefore was not reliable. No visual function measures were reported, and no fundus photographs were published.

Kulkarni et al (⁴) described a 24-year-old patient who reported blurred vision 2 months after bariatric surgery. Visual acuity deteriorated to hand movements in both eyes, and she had sluggishly reactive pupils, bilateral optic disc edema, and peripapillary hemorrhages. Black and white fundus photographs in that report suggested a considerable amount of optic disc edema. There was also sidebeat nystagmus, dysmetria, and gait ataxia but evidently normal mental status. The authors pointed out that this presentation was bound to suggest papilledema, optic neuritis, or bilateral optic neuritis. Brain MRI showed high T2 and FLAIR signal in medial thalamic regions, and lumbar puncture showed a normal opening pressure and constituents, findings consistent with a diagnosis of WE. Methylprednisolone treatment was stopped, and after 24 hours of treatment with 100 mg thiamine intravenously, the patient's visual acuity improved to 20/30 in both eyes and eventually to 20/20. The degree of visual deficit and optic disc edema in this patient is much greater than that described in other reports of WE, but details in other reports are sparse.

The thickened and telangiectatic peripapillary nerve fiber layer and retinal hemorrhages displayed by our patient are probably similar to the findings originally described by Wernicke and by most subsequent authors. They are strikingly similar to the pathologic characteristics noted in the mamillary bodies in WE (^1,14) (Fig. 4).

Thiamine is an important cofactor for enzymes responsible for energy production via the Krebs cycle in mitochondria. Deficiency of thiamine results in a buildup of free radicals and neuronal cell death (¹⁴). The fundus findings seen in our patient bear a resemblance to those seen in methanol-induced optic neuropathy (¹⁵) and in Leber hereditary optic neuropathy (¹⁶), in which mitochondrial abnormalities have been demonstrated on transmission electronic microscopy in the retinal capillary endothelium (¹⁷). We postulate that mitochondrial dysfunction first gives rise to the swelling, telangiectasia, and hemorrhage of the retinal nerve fiber layer, which is shown here, and that actual optic disc swelling appears only if mitochondrial damage is severe and prolonged.

Curiously, the slowly progressive optic neuropathy noted in chronic alcoholism produces severe persistent optic neuropathy with pale optic discs but not the fundus findings we describe here. Thiamine deficiency is implicated in that setting, but evidently its manifestations are different from those in patients with reduced nutrient intake after gastric bypass surgery (^18-21).

Although not the focus of this report, the eye movements in this patient merit comment. She had a deficit of abduction to both sides that we attribute to convergence substitution movements on attempted lateral gaze rather than impairment of sixth cranial nerve function. Were sixth cranial nerve palsies the cause of the impaired abduction, we would expect considerable esotropia in primary gaze position, yet our patient's eyes were aligned in that gaze position. Although sixth cranial nerve palsy is conventionally considered the cause of impaired abduction in WE, none of the reported patients was described as having esotropia in the primary gaze position. Instead, the reports described impaired or slow volitional horizontal gaze or gaze-evoked nystagmus. Convergence substitution movements on attempted side gaze are a common phenomenon in patients with impaired horizontal gaze caused by pontine dysfunction (¹⁵). Cogan and Victor (⁹) also pointed out that in a pathologic study of the brains in patients with WE, “destructive lesions of the ocular motor nuclei were infrequent.” Perhaps the mechanism of eccentric gaze misalignment in patients with WE is substituted convergence rather than sixth cranial nerve palsy.