Van Stavern, Gregory P. MD; Biousse, Valérie MD; Lynn, Michael J. MS; Simon, Deborah J. MD; Newman, Nancy J. MD
Neuro-ophthalmic deficits may commonly follow head trauma (1–13). The afferent and efferent visual systems are susceptible to injury from a variety of mechanisms. These patients can be a diagnostic and therapeutic challenge, in large part secondary to the frequently vague nature of their visual complaints and their coexistent neurologic deficits. Although the association between head trauma and neuro-ophthalmic deficits is clear, there is as yet no definite consensus regarding the relative frequency of specific neuro-ophthalmic deficits, both afferent and efferent, that may accompany head trauma. In this review of all head trauma patients seen in one academic neuro-ophthalmology unit during a 9-year period, we determined the relative frequency of various neuro-ophthalmic deficits incurred after head trauma and their relationship to the nature of the head injury.
A retrospective chart review of all consecutive patients seen between 1991 and 1999 in the neuro-ophthalmology unit at Emory University and given a diagnosis code of head trauma was performed. All patients underwent a standardized neuro-ophthalmic history and examination. Visual fields were tested using a Goldmann perimeter or a Humphrey automated perimeter. Length of time from injury to examination, type of trauma sustained, extent of injuries (including neuroimaging, if available), and loss of consciousness were noted for each patient. For patients diagnosed with optic neuropathy, a distinction among mechanisms of injury (indirect, direct, papilledema) was made on the basis of visual acuity, visual field, and optic disc appearance. Cause of injury, sex, loss of consciousness, and the presence of a neuroimaging abnormality were all evaluated as potential predictors of outcome, with outcome being defined as the presence of a neuro-ophthalmic deficit or a combination of neuro-ophthalmic deficits. All predictors and outcomes were categorical in nature. The association between the predictors and outcomes was evaluated using a chi-squared test. In addition to univariate analyses, multivariate analyses were done using logistic regression; however, these analyses did not reveal any important associations beyond those discovered with the univariate analyses.
Three hundred and twenty-six patients were reviewed. They included 203 (62.3%) men and 123 (37.7%) women. Age ranged from 2 to 86 years, with a median of 30 years. Most patients were admitted in the rehabilitation service at the time of their evaluation in our unit. The neuro-ophthalmic examination was part of a systematic evaluation obtained before theirdischarge from rehabilitation and included both symptomatic and asymptomatic patients. Only patients cooperative enough to perform a complete neuro-ophthalmic examination, including formal visual field testing, were evaluated.
The various causes of head trauma in this series are listed in Table 1. Motor vehicle accident (MVA) was the most common cause of head trauma, occurring in 195 patients (59.8%). Of these, 166 were passengers or drivers, and the remaining patients were pedestrians struck by a moving vehicle. Falls, motorcycle accidents, and projectile injuries were less common, accounting for 31.2% of patients in total. Time from injury to examination ranged from 3 days to 12 years, with a mean of 73.5 days, ± 291.8 days. One hundred and thirty patients (39.9%) experienced a loss of consciousness lasting from several minutes to a prolonged coma of several months' duration (Table 2) at the time of injury. Clinically relevant neuroimaging findings at the time of injury were noted in 153 patients (46.9%) (Table 2). We considered the following abnormalities as clinically relevant: intracranial hemorrhage (ICH) (epidural, subdural, or intraparenchymal), basal skull fracture, and radiographically evident contusion. ICH was the most common finding, occurring in 95 patients (62.1%). In 70 patients (34.3%), neuroimaging was normal.
An abnormal neuro-ophthalmic examination was found in 185 patients (56.7%) (Table 3). Of the patients with abnormal neuro-ophthalmic examinations, 93 (50.2%) had afferent pathway deficits, and 109 patients (58.9%) had efferent pathway deficits. Among afferent deficits, retrochiasmal visual field defects were the most common, occurring in 47 patients (50.5%). Optic neuropathies occurred in 40 patients (43%) and included indirect optic nerve injury (27.9%) and optic nerve injury secondary to previously elevated intracranial pressure with papilledema (15%). Chiasmal injury and Terson syndrome were uncommon, each occurring in three patients (3.2%). Among efferent deficits, ocular motor cranial nerve abnormalities predominated, with 84 patients (77.1%) having at least one ocular motor nerve palsy. Among these patients, 19 (22.6%) had multiple ocular motor nerve palsies. Trochlear nerve palsy was the most common deficit (51.2%), followed closely by oculomotor nerve palsy (46.4%). Twenty-four patients had bilateral ocular motor palsies, with bilateral trochlear nerve palsy occurring most frequently. Supranuclear ocular motor deficits were less common and included convergence insufficiency (16 patients, 14.7%), supranuclear gaze palsy (two patients), and dorsal midbrain syndrome (two patients). Horner syndrome was uncommon, occurring in only three patients (2.7%). Central vestibular nystagmus was documented in only two patients.
Cause of injury was not significantly associated with any outcome. Female sex was significantly associated with trochlear nerve palsy (χ2 = 3.80, P = 0.05) but not with any other outcome. Male sex was not significantly associated with any outcome. The presence of any neuroimaging abnormality was significantly associated with indirect optic nerve injury (χ2 = 3.68, P = 0.05) and oculomotor nerve palsy (χ2 = 3.79, P = 0.05).
The presence of an ICH was significantly associated with oculomotor nerve palsy (χ2 = 8.86, P = 0.003) and bilateral trochlear nerve palsy (Fisher test, two-tail, P = 0.02). The presence of a basilar skull fracture was significantly associated with trochlear nerve palsy (Fisher test, two-tail, P = 0.05). The presence of a radiographically evident contusion was not significantly associated with any outcome.
Head trauma is common. In one survey, the incidence of traumatic head injury requiring hospitalization ranged from 109 to 322 per 100,000 people (4). The advent of motorized transportation has increased the clinician's exposure to head trauma and its complications. Indeed, MVA is the most frequent cause of head injury in the United States. Young men are most frequently affected (4,5), as noted in our study (Table 4).
The relationship between head trauma and neuro-ophthalmic injury has been recognized for some time. Indeed, Hutchinson's studies (6) in the 1880s documented a clear relation between head trauma and ocular motility deficits. Surprisingly, there have been relatively few large studies assessing the frequency of specific neuro-ophthalmic deficits seen in association with head trauma (Table 4). The relative frequency of neuro-ophthalmic deficits is highly variable, and this finding may be explained by differences in the patient population. However, a few studies reviewed patient populations that were homogeneous in age or clinical characteristics; for example, Lepore (11) studied motility deficits in patients with known heterodeviations and subjective diplopia, thereby perhaps missing patients with only afferent, or afferent and efferent, deficits (Table 4). Furthermore, in some reviews, certain aspects of the neuro-ophthalmic examination (such as visual field defects) were not addressed. Finally, referral bias may play a role in that patients with only mild neuro-ophthalmic deficits or patients whose neurologic status renders them unable to communicate symptoms may not be referred to the neuro-ophthalmologist. It is very likely that we had such a bias, because not all patients with head trauma were evaluated by us; however, we think that our series is most likely representative of the population with severe head trauma. Indeed, almost all our patients were referred from the rehabilitation service, where they were admitted for at least 3 to 4 weeks. This neuro-ophthalmic evaluation was part of a routine evaluation performed before their discharge. Only 56.7% of our patients had an abnormal neuro-ophthalmic examination, suggesting that not only symptomatic patients were referred to us.
Closed head injury may damage the optic nerve through a variety of mechanisms, including direct injury from a penetrating wound, indirect injury from concussive forces transmitted to the nerve, and disc edema from elevated intracranial pressure (2). Indirect optic nerve injury is the most common form of traumatic optic neuropathy (TON) (2). The incidence of indirect TON ranges from 2.8 to 26.1% (Table 4). This large range is most likely related to several factors: 1) TON may be difficult to diagnose in the acutely ill, head-injured patient; 2) distinction between the mechanisms of TON (direct vs. indirect injury, papilledema) requires adequate historical information and accurate and reliable visual field testing, both of which may be difficult to obtain in neurologically impaired patients; and 3) attention is often directed to efferent pathway deficits because these are more amenable to management. Our study showed a higher incidence of indirect TON than many previous studies, suggesting that TON may be underdiagnosed in the head trauma population. Although the management of indirect TON is controversial, patients with indirect TON seen within the first few hours may benefit from treatment with high-dose corticosteroids (1,2,14,15). Therefore, indirect TON should always be suspected in patients with severe head trauma.
Traumatic chiasmal injury is uncommon and is often associated with frontal head trauma. Typical features include bitemporal hemianopia (often complete), cerebrospinal fluid rhinorrhea, and diabetes insipidus. The mechanism of injury is unclear but may involve physical disruption of the chiasm or diffuse axonal injury (1). Only three of our patients had chiasmal injury, attesting to the rarity of the phenomenon.
Because a relatively large proportion of neural tissue is dedicated to the primary retrochiasmal visual system, it makes intuitive sense that diffuse brain injuries should frequently damage these pathways; however, the relative frequency of traumatic retrochiasmal visual field defects is difficult to ascertain in the literature, because many studies do not specifically address this issue. This observation may again reflect the difficulty in assessing visual fields in neurologically impaired patients, particularly using the automated perimetry. We were able to obtain a reliable visual field test in all our patients, using the Humphrey automated perimeter or Goldmann visual field test performed by a skilled technician in patients with cognitive disorders. More than half of our patients with afferent pathway deficits had retrochiasmal visual field defects, a much higher incidence than that reported in comparable studies. Because these visual field defects may have substantial financial and legal implications (regarding driving and employment), accurate visual field testing is critical in head-injured patients. Because accurate testing is often not possible in the immediate period after the injury and may not be possible for weeks or even months after the injury, it is important that patients be monitored at regular intervals until an accurate assessment can be made.
The ocular motor nerves may be damaged at any point in their course from brainstem nucleus to extraocular muscle. The most likely mechanism of injury is axonal shearing resulting from differential acceleration of the skull and brain (2,3); however, focal lesions such as hemorrhage at the brainstem exit site (16) or avulsion of the nerve root (12) may also occur. In some cases, minor head trauma may unmask a pre-existing, otherwise unrecognized mass lesion (17).
The trochlear nerve is the smallest and longest of the ocular motor nerves and is closely associated with the rigid tentorium, making it susceptible to traumatic injury. It is particularly susceptible to injury as it emerges from the dorsal surface of the brainstem; damage at this location often causes bilateral injury (1). The fourth nerve has been variably reported to be the most and least frequently involved ocular motor nerve after head trauma (Table 4) (1,18,19). Indeed, the diagnosis of trochlear nerve palsy may be difficult, particularly in uncooperative, neurologically impaired patients. In our study, trochlear nerve palsy was the most common ocular motor nerve palsy, occurring in 51.2% of patients with ocular motor nerve palsies. A high proportion (32.5%) was bilateral, reinforcing the concept that a bilateral trochlear nerve palsy should be suspected in any patient with what appears to be a unilateral injury. Further, a certain proportion of posttraumatic trochlear nerve injury may represent decompensation of a pre-existing congenital trochlear nerve palsy. Indeed, in one study, congenital trochlear nerve palsy was the most common cause in children (20). Determination of vertical fusion amplitude and examination of old photographs usually help to distinguish congenital from acquired trochlear nerve palsy.
Trauma is the second or third leading cause of oculomotor nerve palsies in adults and the leading cause of acquired third nerve palsies in children (1,20,21). The site of injury is often difficult to localize unless other neurologic findings are present. Among our patients with ocular motor nerve palsies, oculomotor nerve injury was the second most common deficit, found in 46.4% of patients with ocular motor nerve palsies; only 7.7% were bilateral.
Acquired abducens nerve palsy is the most commonly recognized ocular motor nerve palsy in any age group (1,19,21). This finding probably relates more to the ease of diagnosis rather than frequency of occurrence, because acquired abducens nerve palsies are relatively rare in certain age groups, such as young adults (22). The abducens nerve has a long course and is susceptible to injury at any point; it is particularly vulnerable to the effects of elevated intracranial pressure. The abducens nerve was the least commonly affected ocular motor nerve in the previous studies (Table 4), and this finding is in agreement with our series.
Convergence insufficiency is a supranuclear motility disorder characterized by a remote near point of convergence, poor convergence amplitudes, and an exodeviation at near. Convergence insufficiency may be seen in a variety of clinical settings, but its association with head trauma has been well documented (3). Similar to previous studies (Table 4), convergence insufficiency was the most common supranuclear ocular motility deficit among our patients, seen in 14.7% of patients with efferent deficits. Other supranuclear motility disorders (such as supranuclear gaze palsies, skew deviation, and dorsal midbrain syndrome) are less commonly reported in association with head trauma. This finding may again reflect difficulty in diagnosing these deficits in uncooperative patients. In addition, some of these disorders may be relatively asymptomatic and therefore not prompt referral to a neuro-ophthalmologist. Furthermore, these supranuclear disorders, particularly convergence insufficiency, may be masked by coexistent ocular motor deficits (11).
Knowing if specific neuro-ophthalmic deficits were more common after certain types of head injury might allow the clinician to have a higher index of suspicion for certain abnormalities when examining these patients. Loss of consciousness (LOC) implies head injury severe enough to cause brainstem reticular formation or bihemispheric dysfunction through diffuse axonal injury, parenchymal contusion, or ICH. Several studies have found a correlation between LOC and particular neuro-ophthalmic deficits (9,23); however, LOC was not significantly associated with any outcome in our study. This finding would suggest that although LOC may be an indirect indicator of the severity of head injury, it may not be a reliable predictor of neuro-ophthalmic outcomes. The presence of a neuroimaging abnormality (skull fracture, contusion, or ICH) would suggest a greater severity of head trauma and perhaps a more frequent occurrence of certain neuro-ophthalmic deficits. In our study, the presence of an ICH was significantly, and strongly, correlated to unilateral oculomotor palsy. The presence of a skull fracture was significantly, but weakly, correlated to unilateral abducens nerve palsy. The specific location and extent of the neuroimaging abnormalities were not available in all patients; however, the information available would suggest that the presence of a neuroimaging abnormality, especially ICH, should increase the index of suspicion for an ocular motor nerve injury, particularly of the oculomotor nerve.
Finally, it is worth mentioning that more than one third (34.3%) of our patients with neuro-ophthalmic deficits had normal neuro-imaging. Therefore, even in the absence of any neuroimaging abnormality, the prevalence of neuro-ophthalmic findings is high.
1. McCann JD, Seiff S. Traumatic neuropathies of the optic nerve, optic chiasm, and ocular motor nerves. Curr Opin Ophthalmol 1994; 5:3–10.
2. Steinsapir KD, Goldberg RA. Traumatic optic neuropathy. Surv Ophthalmol 1994; 38:487–517.
3. Baker RS, Epstein AD. Ocular motor abnormalities from head trauma. Surv Ophthalmol 1991; 35:245–67.
4. Thurman DJ, Jeppson L, Burnett CL, et al. Surveillance of traumatic brain injury in Utah. West J Med 1996; 165:192–6.
5. Sosin DM, Sniezek JE, Thurman DJ. Incidence of mild and moderate brain injury in the United States. Brain Inj 1997; 11:649–59.
6. Hutchison J. Four lectures on compression of the brain. Clin Lect Rep Med Surg Staff Lond Hosp
7. Jacobi G, Ritz A, Emrich R. Cranial nerve damage after paediatric head trauma: a long-term follow-up study of 741 cases. Acta Paediatr Hung 1986; 27:173–87.
8. Keane J. Neurologic eye signs following motor vehicle accidents. Arch Neurol 1989; 46:761–2.
9. Sabates N, Gonce M, Farris B. Neuro-ophthalmological findings in closed head trauma. J Clin Neuroophthalmol 1991; 11:273–7.
10. Kowal L. Ophthalmic manifestations of head injury. Aust N Z J Ophthalmol 1992; 20:35–40.
11. Lepore F. Disorders of ocular motility following head trauma. Arch Neurol 1995; 52:924–6.
12. Mariak Z, Mariak Z, Stankiewicz A. Cranial nerve II-VII injuries in fatal closed head trauma. Eur J Ophthalmol 1997; 7:68–72.
13. Moster M, Volpe NJ, Kresloff MS. Neuro-ophthalmic findings in head injury. Neurology 1999; 52(suppl 2):A23.
14. Cook MW, Levin LA, Joseph MP, et al. Traumatic optic neuropathy: A meta-analysis. Arch Otolaryngol Head Neck Surg 1996; 122:389–92.
15. Levin LA, Beck RW, Joseph MP, et al. The treatment of traumatic optic neuropathy: The International Optic Nerve Trauma Study. Ophthalmology 1999; 106:1268–77.
16. Balcer LJ, Galetta SL, Bagley LJ, et al. Localization of traumatic oculomotor nerve palsy to the midbrain exit site by MRI. Am J Ophthalmol 1996; 122:437–9.
17. Jacobson DM, Warner JJ, Choucair AK, et al. Trochlear nerve palsy following minor head trauma. A sign of structural disorder. J Neuroophthalmol 1988; 8:263–8.
18. Keane JR. Fourth nerve palsy: historical review and study of 215 in-patients. Neurology 1993; 43:2439–43.
19. Rucker CW. The causes of paralysis of the III, IV, and VI cranial nerves. Am J Ophthalmol 1966; 61:1293–8.
20. Holmes JM, Mutyala S, Maus TL, et al. Pediatric third, fourth, and sixth nerve palsies: a population-based study. Am J Ophthalmol 1999: 127:388–92.
21. Richards BW, Jones FR, Young BR. Causes and prognosis in 4,278 cases of paralysis of the oculomotor, trochlear, and abducens nerve. Am J Ophthalmol 1992; 113:489–96.
22. Moster ML, Savino PJ, Sergott RC, et al. Isolated sixth nerve palsies in younger adults. Arch Ophthalmol 1984; 102:1328–30.
23. Carlson GS, Svardsudd K, Welin L. Long-term effects of head injuries sustained during life in three male populations. Neurosurgery 1987; 67:197–205.
© 2001 Lippincott Williams & Wilkins, Inc.