Question of the Week
Video Library
Ophthalmology Books & Manuals
Cyber-Sight Atlas of Eye Diseases
The Ophthalmology Minute
Nursing Education
Eye Care Equipment
ORBIS Program Features
Free Online Journals
Ophthalmology Links
Ask a Professor
Print ViewPrint this Page
Diplopia -  

Neuropathic Causes
Lecture 6 of 7  NEXT»

Third Cranial Nerve Paralysis


The oculomotor nerve is the most complex of the three nerves that control ocular movement. It contains synaptic motor fibers for eye movement and visceral parasympathetic fibers that innervate the intrinsic muscles of the eye. The topographic organization of the oculomotor nuclei has been examined by Bernheimer, Brower, Warwick, Bender and, more recently, by BURIC. The oculomotor nerve lies in the mesencephalon in the inferior periaqueductal gray miler and is about 6 to 8 mm long. The neurons that innervate the superior rectus muscle originate in a subnucleus and innervate most of the opposite superior rectus muscle. The superior rectus muscle is innervated by a central caudal nucleus. In the dorsal caudal location is the Edinger-Westphal nucleus.

As the fibers leave the oculomotor nucleus, they form a fascicle. The diagnosis of a fascicle lesion depends on what structure it is near when affected. The syndromes are well known. Rothnagel's syndrome demonstrates an oculomotor palsy and contralateral cerebella ataxia. It is contralateral, since it is above the decussation of the motor tracks. Benedikt's syndrome demonstrates contralateral abnormal movements due to its proximity to the red nucleus, with an oculomotor palsy. As the oculomotor nerve progresses more ventrally in the mesencephalon, it passes the motor tracks in the cerebral peduncle, producing a crossed hemiparesis syndrome called Weber's syndrome. The above three syndromes usually produce a complete oculoinotor palsy, but on rare occasion the pupil is spared. As the oculomotor nerve emerges from the mesencephalon in the interpeduncular fossa it can also produce a Weber's syndrome identical to an intrinsic brainstem Weber's. The oculomotor nerve passes between the posterior cerebral and superior cerebellar artery. This location is important since it is an alternative location for an oculomotor palsy from an aneurysm. It then crosses through the clura and enters the cavernous sinus, passing the posterior cerebral artery and internal carotid artery, which is the more common location for involvement with an aneurysm. Its position in the cavernous sinus Is above the fourth nerve and first, sometimes second, division of the trigeminal nerve. Parkinson described the blood supply to the oculomotor nerve in the cavernous sinus. It is supplied by the dorsal meningeal, tentorial, and inferior hypophyseal branches of the meningohypophyseal branch. In the anterior cavernous sinus it divides into two divisions. The superior division controls the levator superioris and superior rectus muscles. The inferior division supplies all the other muscles of the oculomotor complex and pupillary fibers. Both divisions pass into the orbit through the annulus of Zinn. The superior division runs anteriorly in the superior rectus muscle, while the fibers for the levator pass along the lateral border of the superior rectus and upward to pass through the belly of the levator muscle. The pupil fibers leave the inferior division to synapse in the ciliary ganglion. The pupil fibers in the subarachnoid portion of the oculomotor nerve are located on the superior portion of the nerve.

fig. 6.15

Figure 6.15. Anatomy of oculomotor nerve, part 1. (Courtesy of William Stewart.)

fig. 6.16

Figure 6.16. Anatomy of oculomotor nerve, part 2. (Courtesy of William Stewart.)

fig. 6.17

Figure 6.17. Anatomy of oculomotor nerve, part 3. (Courtesy of William Stewart.)

fig. 6.18

Figure 6.18. Anatomy of oculomotor nerve, part 4. (Courtesy of William Stewart.)


The third cranial nerve (oculomotor nerve) has many functions. It supplies innervation (a) to the levator muscle, which lifts the lid, (b) to the superior rectus and inferior oblique muscles, which elevate the eye, and (c) to the inferior rectus muscle, which depresses the eye. Accommodation and pupillary constrictions also depend on the third cranial nerve.

When a patient is being examined for possible third cranial nerve paralysis, all the areas innervated by the third cranial nerve must be tested. If ptosis is present, the diplopia may not be evident to either patient or physician when the patient is tested in the straight-ahead position; however, when the patient is instructed to look up and laterally in the field of the vertical action of the superior rectus muscle, the imbalance may be obvious to both patient and observer. This subtle finding may also be true of the other muscles as well as the pupillary and accommodation reactions. Owing to their common origin at the annulus of Zinn, a combination of ptosis and superior rectus muscle weakness is the most frequently occurring of all possible associations within the third cranial nerve. Isolated ptosis and isolated superior rectus muscle weakness are usually of traumatic or congenital origin.

THIRD CRANIAL NERVE MISDIRECTION. Although the physician does not often look for third cranial nerve misdirection, it is of real significance because it occurs after recovery from a third cranial nerve paralysis; however, it may not be obvious for 6 months to a year after the onset of, and recovery from, the paralysis. Several signs of this condition must be looked for, any one of which may occur alone and be as significant as multiple signs of misdirection. The most common form of misdirection is the pseudo-von Graefe's sign. When the patient looks down, the upper lid retracts and the superior sclera is exposed (as in thyroid disease). In another form of misdirection, lid gaze dyskinesis, the lid is elevated when the patient moves the eye medially and innervates the, medial rectus muscle on pure horizontal gaze. When the patient looks laterally, the lid returns to normal position. Lid gaze dyskinesis may be missed when the affected eye is examined in the horizontal plane, and is best seen in horizontal movements of the eye in down-gaze (Fig. 6.19). (In a normal eye, no elevation or depression of the lid should exist in pure horizontal gaze excursion.) If lid gaze dyskinesis is present; pupillary signs frequently are seen. The pupil becomes smaller in near gaze on Innervation of the medial rectus muscle as well as in horizontal gaze, when the patient turns the eyes toward the nose, again innervating the medial rectus muscle. This reaction is termed a pseudo-Argyll Robertson pupil, referring to a lack of pupillary response to light but a better reaction to accommodation-convergence. In patients with the misdirection phenomenon, however, the pupil is contracted, not because of the accommodation reflex, but because of some crossover innervation between the medial rectus muscle and the pupil. This phenomenon can be demonstrated by making the medial rectus muscle work on distant horizontal gaze movement and show the same contraction of the pupil as would occur via the accommodation-convergence mechanism. (This phenomenon also differentiates the condition from one related to tertiary syphilis.)

fig. 6.19a

fig. 6.19b
fig. 6.19c fig. 6.19d

Figure 6.19. Patient with a right third-nerve misdirection of the lid. A, B. In moving the eyes from right to left gaze, the lid elevates. C, D. This sign is more easily seen when the test is performed with the eyes in down-gaze rather than straight ahead.

Third cranial nerve misdirection (a) occurs more frequently after third cranial nerve paralysis owing to trauma or after an aneurysm, (b) is infrequently caused by tumor or syphilis, and (c) has not been reported after third cranial nerve paralysis owing to diabetes. Therefore, if a diabetic patient shows third cranial nerve misdirection, it is likely that a subtle aneurysm has been overlooked and should now be considered in the absence of a significant history of trauma.

The mechanism by which misdirection occurs is far from settled. In an experiment on chimpanzees in which misdirection was created, Bender found that the regrowth of fibers involved a random distribution to all extraocular muscles. Lang and Lipschultz first postulated misdirection of facial nerve fibers as a reason for the synkinetic mass movements of the facial nerve. Bielschowsky then applied that concept to the oculomotor nerve. Peripheral orbital traumatic cases would tend to support this theory. Lepore and Glaser argued against that explanation, stating that "neuromuscular junctions formed by ectodermal muscle of the iris sphincter and axons formerly innervating striated muscle probably would not be functional." Instead, they postulated ephaptic transmission as the mechanism for synkinesis. Other explanations for the synkinetic movements are central reorganization and supersensitivity.

Sunderland's work demonstrated that the endoneurium must be interrupted--not just the axon--for misdirection to occur. If the endoneurium is left intact, then the injured axons follow their normal route during the regeneration process. This phenomenon is supported clinically by pathologic examination of an ocular motor palsy with misdirection from an aneurysm and an ocular motor palsy without misdirection from diabetes. In the former case, there is axon degeneration and disruption of the endoneurium; in the latter case, there is axon degeneration but the endoneurium is intact. These findings seem to support the misdirection theory experimentally, pathologically, and clinically.

There are cases, however, of intracavernous lesions that create synkinesis but never develop an oculomotor paresis; these cases seem inconsistent with the misdirection explanation of synidnesis. Sonic other pathologic studies also appeal to cast doubt on the misdirection theory, Kerns excised an oculomotor nerve specimen from a patient who had developed a paresis from an aneurysm and after recovery had demonstrated synkinesis. He found fewer axons on the involved side rather than more, which would he expected if there was misdirection of MCI'S. Another possible interpretation of these findings is that there is an increase in the number of small axons, suggesting regeneration.

The most difficult cases to explain in this anatomic rewiring scheme are the cases of reversible synkinesis. If some fibers grow to the wrong terminal and innervate, causing the signs of synkinesis, why do they stop? One explanation is that the normal and misdirected fibers compete for preferential innervation of a muscle and that the normal fibers eventually win out. There is experimental evidence by Kuffler to support this view.

A second explanation for synkinesis is interaxonal, or ephaptic, transmission. The myelin sheath acts as an insulation between axons. When the axon is injured and the myelin sheath is interrupted or becomes defective, then interaxonal stimulation may occur. This phenomenon was initially demonstrated by Herring. Moruzzi in 1944 observed that action currents can alter the electric activity in nearby axons. If a myelin defect allows electric "cross talk" to occur, then healing or isolation of that defect in the axon should stop it. VonUexkull observed that he could reduce such cross talk by putting saline at the injured site. Normally, saline is an excellent conductor of electrical impulses and thus might be expected to enhance the cross talk. However, in these experiments, the saline may have caused swelling of the tissues at the injury site, thereby isolating the nerve and its electric transmission. Clinical evidence of ephaptic transmission in a case of chronic limo compression was demonstrated by Thomasulo.

Another theoretical explanation of synkinesis was put forth by Fuchs. He postulated that after a peripheral nerve is injured, not only is there regeneration of the nerve, but also there may be some reorganization centrally that would explain synkinesis. Study of pathologic specimens has shown that when a nerve is injured, changes do occur in the cell body. If the nerve regenerates normally, these changes disappear. The more severe the injury, the more changes are seen in the cell body. The question remains whether the nucleus defects cause the peripheral fibers to regenerate in a misdirected fashion. The other side of the argument is that anatomic changes do not necessarily result in functional change. The most convincing argument against this reorganization theory is diabetic ocular motor palsies. If peripheral injury is the initiating factor for central reorganization, then what distinguishes trauma, tumor, and aneurysm from diabetic injury such that the latter does not cause misdirection?

Denervation supersensitivity is a well-known phenomenon. The more peripheral the nerve that is damaged, the more sensitive it becomes to its own effector substance. The pharmacologic tests to demonstrate Adie's or Homer's pupil are based on this phenomenon. The diabetic cases are again the main stumbling block to accepting this mechanism to explain oculomotor synkinesis.

There is also a syndrome called primary misdirection without apparent oculomotor paresis. This has infrequently been seen with an intracavernous lesion such as meningioma or aneurysm. The cases of aneurysms at the junction of the internal carotid and posterior communicating artery usually present with oculomotor palsies. Subsequent to that, third-nerve misdirection can occur. They rarely cause primary misdirection. As mentioned above, a lesion of the anterior cavernous sinus can interfere with only one division of the third nerve, as it separates into the superior and inferior divisions. An isolated lesion of the superior division of the oculomotor nerve is much more commonly seen in the orbit than in the anterior cavernous sinus. However in the face of no orbital disease to account for the superior division paresis, look in the anterior cavernous sinus.

Congenital Oculomotor Palsy

Congenital oculomotor palsies can be due to absent or partial nerve development. Many cases have miotic not mydriatic pupils, which suggests a misdirection phenomenon in a peripheral nerve location.


Once diplopia has been diagnosed as being caused by dysfunction of the third cranial nerve, the next step is to determine what part of the nerve is affected and what disease is the cause. For the purpose of this discussion, disorders of the third cranial nerve have been grouped into six categories according to anatomic location.

fig. 6.20

Figure 6.20. Clinical approach to third-nerve paralysis.

NUCLEAR LESIONS (FIG. 6.15, 1). Even though the two third cranial nerve nuclei are
separate, they show some interaction. Most authors believe that the fibers of the superior rectus muscle are the ones that are interdependent. Therefore, in identifying an isolated third cranial nerve paralysis as being nuclear in location, the following should be helpful. Total third cranial nerve paralysis affecting one eye but with no limitation of up-gaze in the other eye rules out a nuclear lesion as the causative factor. In paralysis owing to a nuclear lesion, the contralateral superior rectus muscle would have to be limited because of interdependence. Vascular disease is the most frequent cause of this type of paralysis.

FASCICULAR LESIONS (FIG. 6.15, 2, 3). Disease in the fascicular fiber bundle, which is also intrinsic to the brainstem, involves two well-recognized syndromes. A disorder in the dorsal fascicular area produces the syndrome of Benedikt, with ipsilateral third cranial nerve paralysis and a contralateral hemitremor that is produced by proximity to the red nucleus. Disorder in the ventral fascicular area produces the syndrome of Weber, with ipsilateral third cranial nerve paralysis and contralateral hemiparesis. The degree of involvement of the third cranial nerve and the severity of the tremor or paralysis vary from mild to severe. Once again, vascular disease, particularly arteriosclerosis, is the main cause of these symptoms. The syndrome of Weber may also occur as an extrinsic brainstem phenomenon owing to lesions in the interpeduncular fossa, where the third cranial nerve exits from the brainstem, but it cannot be differentiated from the intrinsic variety on clinical grounds alone. In addition to vascular disease, venous malformation and tumors may also cause this condition in the interpeduncular fossa.

SUBARACHNOID LESIONS (FIG. 6.15, 3). The subarachnoid space is the most common location for third cranial nerve paralysis, particularly if the pupil is involved, no matter how minimally. When the pupil is involved, the major cause to be considered is an aneurysm at the junction of the internal carotid and posterior communicating arteries. If aneurysm proves to be the cause, other signs of subarachnoid hemorrhage usually occur, such as pain, stiffness of the neck, photophobia, and a change in the level of consciousness. These additional symptoms are not always present, however, and their absence does not rule out an aneurysm if the pupil is involved. Pupillary involvement owing to an aneurysm is usually significant: the pupils usually are large and fixed; however, any change in size and function is equally significant. Absence of pupillary involvement used to be considered rare, occurring (according to Rucker) in about 3% of cases owing to aneurysm. More recent studies by Kissel et al. and Nadeau and Trobe indicate that absence of pupillary signs is not all that rare, particularly in partial third-nerve paresis. In this case, pupillary response may initially be normal, but the patient should be observed for subsequent development of pupillary signs (Fig. 6.16, 1, 2, Fig. 6.21).

fig. 6.21

Figure 6.21. Basilar aneurysm compressing area between posterior cerebral and superior cerebellar artery.

Diabetes is the second most common cause of third cranial nerve paralysis in the subarachnoid space. In third cranial nerve paralysis secondary to diabetes, the pupillary function is usually unimpaired, owing, theoretically, to collateral circulation from the pial vessels in the peripheral part of the nerve, where the pupil fibers are located. On the other hand, studies by Cogan and Goldstein and by Rucker indicate that pupillary involvement may occur in up to 20% of cases of diabetic ophthalmoplegia, and that pain indistinguishable from that associated with aneurysm can occur in up to 50%. Therefore, the physician must consider the possibility of aneurysm even if the patient is a known diabetic. Clearness of spinal fluid does not rule out an aneurysm, since evidence of subarachnoid hemorrhage is not always present. The 2-hour postprandial serum glucose test is not sufficient for identifying diabetes as the underlying cause of third cranial nerve paralysis. In most such cases, it will be found that the patient has occult diabetes that may be suggested only by a formal glucose tolerance test. Since the treatment for third cranial nerve paralysis secondary to diabetes is to do nothing, the physician may be tempted to wait the 6 or 8 weeks it takes almost all cases to clear up. By delaying, the physician can differentiate between a condition owing to diabetes and one owing to an aneurysm, since the latter is unlikely to improve. Such an approach is still foolish and dangerous because the time of peak incidence for spontaneous rebleeding from an aneurysm is within 12 to 14 clays, a period well short of the healing time for diabetic third cranial nerve paralysis. Moreover, some aneurysms do not progress and the third cranial nerve returns to normal--thus, a potential time bomb is left untreated. The best rule is to regard an isolated third cranial nerve paralysis with pupillary involvement as being caused by aneurysm until it is proven otherwise.

Diabetic oculomotor palsy has been identified in a few pathologic specimens. They occur at the interface of separate vascular supplies to the nerve. This usually occurs from recurrent collateral branches of the ophthalmic artery anteriorly and branches of the posterior cerebral, posterior communicating, and basilar artery posteriorly. Pupil sparing is considered possible from the standpoint of several different explanations. If a mass lesion is intrinsic to the oculomotor nerve, it will grow steadily, occasionally displacing and preserving the small pupil fibers that are more pressure resistant. Lesions that compress the nerve from below are away from the pupil fibers, which are in the superior part of the nerve. A lesion in the cavernous sinus can affect the superior division of the nerve, which does not carry pupil fibers. The same lesions that damage pupil function and subsequently develop aberrant regeneration can give a false sense of pupil function. These last two reasons for pupil sparing are not nearly as valid when examined clinically as are the other previous causes. Lesions in the cavernous sinus can affect the parasympathetic and sympathetic fibers' control of the pupil. This gives the impression of pupil sparing. That is only relatively true for size and not true for function. There are rare cases of preganglionic parasympathetic hypersensitivity found in cavernous sinus lesions as recorded by Slamovitz.

Vascular occlusive disease may also result in isolated cranial nerve dysfunction. Some cases present cataclysmically as a subarachnoid hemorrhage or as a compressive lesion without evidence of a subarachnoid hemorrhage suggesting a tumor. Recent studies by Milisavljevic of the microvascular relationships to nerves III, IV, and VI suggest another type of vascular presentation. In these studies, he found that 40% of the oculomotor nerves are penetrated by small arteries such as the long or short circumflex mesencephalic artery or a branch of the interpeduncular perforating arteries. If these vessels dilate, they can compress parts of the oculomotor nerve either in the periphery, causing pupil involvement, or in the intrinsic portion, sparing the pupil. Similar microvascular studies of the abducens nerve found such an arteriolar relationship rare. However, the abducens nerve not uncommonly is penetrated by a contralateral pontine vein or tributary of the anteromedian pontine vein. The pupil fibers run in the superior peripheral part of the oculomotor nerve. In a vascular lesion such as diabetes, the vessels affecting the nerve may be one of the penetrating branches, thus sparing the peripheral part of the nerve, which may also derive some blood supply from the nearby pial vessels.

Compressive lesions may not initially affect the pupil portion of the oculomotor nerve; an example is a basilar aneurysm that compresses the nerve from below. It is also not uncommon for aneurysmal compression of the oculomotor nerve to compress the upper division of the nerve, leaving the inferior division until later. This occurs commonly in the cavernous sinus where the oculomotor nerve divides into superior and inferior divisions. Sometimes there is apparent pupil sparing due to pupil contraction in adduction in the aberrant regeneration syndrome. Another cause for diagnostic error occurs when there is involvement of the sympathetic nerves in the cavernous sinus, giving the impression of a normal-size pupil. This type pupil, however, does not react normally to stimulation by light or near reflex tests.

The rule of thumb that an isolated oculomotor palsy with the pupil spared is diabetes mellitus and that an isolated oculomotor palsy with the pupil involved is an aneurysm at the junction of the internal carotid and posterior communicating arteries is a good general rule. However, it is not perfect and has many exceptions. Even though diabetes usually spares the pupil, it involves the pupil often enough not to be worthy of a case report. Pupil involvement of the oculomotor nerve with an aneurysm is also the rule, but pupil sparing, once thought to be extremely rare, is now being reported more and more commonly in the literature. O'Connor surveyed 646 neurosurgeons and found that 30.89% of them had seen such pupil-sparing cases, further observation over the following several days will reveal eventual pupil involvement. Since we now know that oculomotor palsy secondary to both diabetes and aneurysm can present with and without pupillary involvement, the general rule outlined above is sometimes difficult to apply in an individual patient. However, because of the potential lethal consequences of an aneurysm, the differential diagnosis is particularly critical in such cases.

Aneurysms are a deadly disease because of the severe consequences of bleeding and the lack of warning to prevent it. The frequency of an aneurysm is estimated at between 1 and 8% of the population, with subarachnoid hemorrhage occurring in 10 to 15 people per 100,000 population. These figures are the result of the combined aneurysm study. Up to 60% of these people die before they are hospitalized. If they make it to the hospital 37% of them will die and 17% will have a severe disability.

Aneurysms are usually symptomatic later in life but develop from inherent weakness at branching of major arteries. Since they become symptomatic decades later, other hemodynamic factors play a role in causing them to rupture, such as hypertension and arteriosclerosis. There are medical diseases that are commonly associated with aneurysm. These include polycystic kidney disease, Moya Moya disease, and fibromuscular dysplasia. Diseases that weaken the arterial wall such as Ehlers-Danlos syndrome, Marfan syndrome, and pseudoxanthoma elasticum also contribute to aneurysm formation. The cooperative study noted that 26% of symptomatic aneurysms rupture within 5 years, while the asymptomatic ones have a rate of 2.6 percent. The cooperative study also noted that between 40 and 50% of patients have some warning, such as a minor hemorrhage or expansion of they aneurysm. It is the latter that confronts the neuro-ophthalmologists more frequently with compression of the third nerve and occasionally the optic nerve.

One of the more serious factors that influence recovery is the arterial spasm. Early operation can cause an increase in that vascular spasm. However, delayed operation can allow rupture. Therefore, controlling factors that increase vascular spasm are important in the proper management of subarachnoid hemorrhage and the secondary vascular spasm.

The description of an aneurysm depends on the size of the dome. A saccular aneurysm in less than 1.5 cm. A globular one is 1.5 to 2.5 cm, and a giant aneurysm is larger than 2.5 cm. The smaller the aneurysm, the less likely it is to bleed. In view of that, magnetic resonance imaging (MRI) can usually visualize an aneurysm 5 mm or larger. Unfortunately this is only a good step forward in diagnosis and is not as good as arteriography. Treatment options will not be discussed here. Some of the newer modalities include interventional neuroradiology as a treatment for some aneurysms. An overview is presented in a chapter on neuroradiology below.

As noted above, diabetic oculomotor palsies tend to get better in less than 3 months. Although it is tempting to delay an arteriogram for only those patients who do not improve after 3 months, the incidence of spontaneous rebleeding of an aneurysm is highest in the first 2 weeks after presentation of a subarachnoid hemorrhage, and the mortality from rebleeding has been as high as 90% in some series. Some English workers believe that the incidence of rebleeding may be overestimated. They do not, however, disagree with the mortality figures for such rebleeding. After reviewing 110 consecutive patients presenting with a ruptured intracranial aneurysm, Maurice-Williams concluded that most of those patients admitted into the hospital in good condition deteriorated from causes other than rebleeding. Of 106 of the 110 cases considered in this relatively good condition category, 49 deteriorated in the next 2 weeks. Only one-third of these cases were attributed to aneurysmal rebleeding; the rest were attributed to delayed vasospasm. Deterioration in the nonhemorrhage, or vasospasm, group peaked between 4 and 12 days and rarely later, whereas rebleeding was evenly distributed over the first 21 days and could occur later. This study is encouraging in the prognosis for recovery of such patients as a group, but unfortunately does not help in the initial neuro-ophthalmologic management of a single patient who may have ophtha lmoplegla secondary to an aneurysm.

In the past, arteriography carried with it a significant morbidity and some mortality. The incidence of complications has been significantly reduced, but not eliminated, by use of femoral catheter studies rather than direct carotid injection and selective arteriography. Therefore, some guidelines must be formulated about patient selection for arteriography. Trobe has formulated a reasonable set of such guidelines. If the patient is over 50 and has a complete oculomotor nerve palsy and a totally normal pupil, then arteriography is not recommended. Complete is the important concept here. If the oculomotor palsy is incomplete and particularly if the inferior division of the oculomotor nerve, which carries the pupil fibers, is not involved, then an arteriogram is recommended. Since the oculomotor nerve divides into the superior and inferior division before it leaves the cavernous sinus, an aneurysm in the anterior cavernous sinus frequently affects the inferior division. Error in pupil evaluation because of misdirection and coexisting sympathetic paresis may give a false impression of a normal functioning pupil on casual examination in such cases. Finally, if there is an abnormal pupil and a complete oculomotor palsy, an arteriogram is also recommended.

An isolated oculomotor nerve palsy in patients under age 20 is another matter. The overwhelming majority of patients under 10 years of age with an aneurysm present with a subarachnoid hemorrhage. At least three cases of teenage patients presenting with third-nerve palsies, headache, and stiff neck have been reported. Of the more than 6000 aneurysms reviewed by Patel, Laitinen, Matson, and Locksley, only 12 cases (0.2%) were children under 10 years of age, and only 42 were under age 20. We certainly cannot ignore the diagnosis of aneurysms in younger patients with oculomotor palsy, but other causes such as ophthalmoplegic migraine, inflammation, postinfectious ophthalmoplegia, and missed trauma should be kept in mind. All of these statistics change if there is evidence of subarachnoid hemorrhage or other associated neurologic signs, whether in a child or in an adult. In a review of childhood aneurysms, Matson, Patel, and Harwood-Nash found only one instance of an aneurysm arising from the posterior communicating artery, which is the typical site for adult aneurysms producing third-nerve palsies.

Infrequently, subarachnoid involvement of the third cranial nerve can be caused by cranial arteritis. I have rarely found this situation, but Meadows has found subarachnoid involvement of the third nerve in 12 to 15% of the cases of temporal arteritis he has seen.

In 40% of cases involving patients with third cranial nerve ophthalmoplegia owing to tumor, the tumor is metastatic. In two large series on ophthalmoplegia, nasopharyngeal carcinoma was found to be the most common metastatic tumor.

Infection accompanying third cranial nerve paralysis is not common; however, varicella and botulism have been reported as causing isolated internal ophthalmoplegia involving accommodation and pupillary function and thus deserve special mention. At one time, diphtheria was a frequent cause of isolated pupillary paralysis, but today diphtheria is rarely reported.

CAVERNOUS SINUS INVOLVEMENT (FIG. 6.17:3). Another form of paralysis can occur where the third cranial nerve runs through the cavernous sinus--the point at which it is associated with the fourth and sixth cranial nerves, the first and second divisions of the fifth cranial nerve, and the ocular sympathetic nerves. An isolated third cranial nerve paralysis in this location is rare, but it can occur; however, signs of cavernous sinus involvement soon become evident on repeated examination.

An intracavernous aneurysm underlying third cranial nerve paralysis is readily differentiated from other cerebral aneurysms. An important distinction is that even if an intracavernous aneurysm ruptures, bleeding occurs into another vascular compartment rather than into the subarachnoid space;thus the signs and symptoms of a subarachnoid hemorrhage are absent. However this is not 100% true. Kuppersmith found that in 79 cases, 1.3% ruptured into the subarachnoid space. The usual complication was optic and cranial nerve neuropathy. Pain is usually intermittent and prominent feature of an intracavernous aneurysm, which frequently develops a fibrous coat as it enlarges and progressively compresses the cavernous sinus nerves. One of the earliest signs associated with third cranial nerve paralysis owing to a disorder in the cavernous sinus is depression of the ipsilateral corneal reflex. Presence of a cephalic murmur that can be altered by carotid compression may help in differential diagnosis. Symptoms of an intracavernous aneurysm generally occur when the patient is over age 50, whereas aneurysms located on the internal carotid artery produce symptoms at least 10 years earlier. Intracavernous aneurysms occur later in life than aneurysms in other parts of the internal carotid artery; the secondary exophthalmos usually is moderate, and chemosis frequently is absent. All of these symptoms differentiate intracavernous aneurysm from arteriovenous cavernous sinus fistula. Optic atrophy owing to enlargement of the aneurysm and compression of the optic nerve is of major concern.

Arteriovenous fistulas are of two varieties--those caused by trauma and those that are spontaneous and arteriosclerotic. Nowadays, more than 75% of such cases are caused by trauma suffered in automobile and motorcycle accidents and are seen mostly in young men, as might be expected. The remainder are spontaneous and arteriosclerotic in origin, occurring primarily in elderly women.

Arteriovenous fistulas can affect the third cranial nerve as well as the other nerves of the cavernous sinus. Exophthalmos and chemosis may be much more prominent than in cavernous sinus aneurysm. Dilated veins around the cornea give the appearance of a caput medusae. Bruits are easily heard by both physician and patient; the ability to change their character by carotid compression varies and depends on the degree of collateral circulation.

Schiotz's tonometer can be a valuable tool in diagnosing arteriovenous cavernous sinus listula. Normal pulsations in the eye cause the arm of this instrument to move 2 or 3 scale readings on the average, but the wider pulse swings of an arteriovenous cavernous sinus fistula may cause the arm to swing 10 scale readings or more. (This wide swing may occur without the physician's having noted any obvious pulsation on simple inspection.) Abnormal pulsation of the globe can be seen with applanation tonometry as well, although not as easily as with Schiotz's tonometer.

Another subtle feature of pulsation may be seen during direct ophthalmoscopic examination of the fundus. As the fundus is observed, the disc and vessels go in and out or focus synchronously with the pulse at the wrist. This phenomenon is distinct from the variation in focus that occurs with variation in accommodation. The pulsations just described are more easily perceived with the direct, than the indirect, ophthalmoscope. Both types of ocular pulsation may be seen when gross movement of the globe is not apparent.

Both secondary glaucoma and ischemic optic atrophy are of major concern in patients suffering from arteriovenous fistula and are the usual reasons for a decision to operate. The results of an operation can be disastrous, however, and the desirability of surgical intervention is presently being widely debated. The conservative view expressed by Spencer, Thompson, and Hoyt and by Sanders and Hoyt is that present surgical techniques tend only to increase the ischemia in many cases and thus make matters worse. The detachable balloon technique has become popular for treatment of arteriovenous fistulas and does not produce the severe consequences seen with previous surgical approaches. Miller has even used the superior ophthalmic vein as a route of insertion. The newer interventional neuroralologic techniques have improved the results significantly.

Carotid cavernous sinus fistula (CCSF), which is an uncommon medical occurrence, occurs in two forms. The first involves a direct flow from the intracavernous carotid artery into the cavernous sinus and exhibits high flow and marked signs. The second involves a connection between a smaller dural artery and the cavernous sinus; it exhibits minimal signs. If fistulas of the second type are low grade, they are frequently missed or at the least treated as chronic conjunctivitis because of a persistent red eye. The red eyes in these patients show dilated vessels with clear conjunctiva between the vessels, with or without chemosis. In true infectious conjunctivitis, the vessels are prominent, but the intervening conjunctiva is pink or red and not clear. Neither type of fistula is usually life threatening, but the high-flow type may threaten vision. The most common complication of the smaller fistula is open-angle glaucoma, which is the most likely cause for the increase in episcleral venous pressure seen in such cases.

Tumor invasion of the cavernous sinus occurs in two different ways: (a) by intracavernous growth of a metastatic tumor and (b) by lateral extension of a pituitary tumor into the cavernous sinus with or without the bitemporal field loss usually associated with this condition. In the latter disorder, termed pituitary apoplexy, the onset is usually sudden, and the patient appears quite ill. The third cranial nerve is rarely solely involved, and the optic nerve on the same side may also be involved owing to extension of the tumor forward into the optic nerve sheath. The fact that most persons exhibiting this syndrome have not had a known pituitary tumor prior to the onset of the apoplexy may delay a proper diagnosis.

SUPERIOR ORBITAL FISSURE INVOLVE-MENT (FIG. 6.18, 3). Tumors impinging on the superior orbital fissure can also cause third cranial nerve paralysis. Because the nerves involved are the same as those involved in disorders located in the cavernous sinus, differentiation usually cannot be made on clinical grounds alone; however, indications of bony changes owing to tumor provide a clue to diagnosis. The most common tumor in this area is a sphenoid ridge meningioma that frequently causes prominent changes readily discernible on even the plain roentgenogram.

Tolosa-Hunt syndrome is a condition entirely unrelated to other disorders causing ophthalmoplegia. Pain usually precedes the ophthalmoplegia and is usually steady rather than throbbing or episodic. The ophthalmoplegia may involve the third, fourth, and sixth cranial nerves separately or in combination. Symptoms may last for several weeks or longer and then usually remit spontaneously, although some residual ophthalmoplegia may continue. It is not unusual to have a remission of the syndrome involving the third cranial nerve, followed by a quick relapse with involvement of the sixth or fourth cranial nerve. Bilateral involvement is rare but not unheard of in these cases.

In diagnosing Tolosa-Hunt syndrome, all other possible causes of this symptom complex must be ruled out. The diagnostic approach usually includes roentgenograms of the sinuses (to exclude infection) and the sphenoid ridge, MRI, or computerized tomograms of the retro-orbital structures, and perhaps cerebral arteriograms. Diagnosis of Tolosa-Hunt syndrome is by exclusion and should therefore be made only after exhaustive evaluation.

ORBITAL LESIONS (FIG. 6.18, 3). Ophthalmoplegia affecting the third cranial nerve in the orbit involves not only the same nerves as in the superior orbital fissure and cavernous sinus but also the optic nerve. As a consequence, the patient may experience loss of vision and perhaps exhibit exophthalmos. In the orbit, the third cranial nerve separates into two divisions, so a partial third cranial nerve paralysis of one or the other division is possible. The upper division innervates the lid and the superior rectus muscles; all the other third cranial nerve functions are in the inferior division.

Involvement of the superior division of the oculomotor nerve usually is located in the orbit or anterior cavernous sinus. Although it used to be thought that such involvement could not occur in the nucleus, nuclear involvement producing isolated third-nerve palsy without other neurologic signs has been reported by Keane. Cases of palsy of isolated muscles innervated by the oculomotor nerve have also been reported by Pusateri as occurring due to selected lesions in the oculomotor nucleus. If the superior division of the oculomotor nerve is involved, there is ptosis and difficulty in elevation in abduction. If this elevation disorder results from a supranuclear lesion, it should be equal in abduction and adduction and orthophoric. The studies of Warwick have provided information on the detailed organization of the oculomotor nucleus. Miller has suggested that the division of the oculomotor nucleus into a superior and inferior division may be actually functionally organized in the nerve all the way back to its exit from the brainstem. There is no direct anatomic evidence for this, but strong inferential evidence. Such an anatomic organization of function from the brainstem has been demonstrated for the facial nerve.

A decrease in vision or merely an afferent pupillary defect on the affected side may be enough of a clue to lead to a diagnosis of third cranial nerve paralysis in the orbit. Among possible causative factors, orbital tumor and contiguous sinus disease, such as mucocele, tumor, or infection, should receive major consideration. Infection may not be immediately obvious if the patient has no sinus symptoms and does not appear toxic; however, progressive pain and exophthalmos should suggest this possibility, and appropriate roentgenographic evaluation is indicated. Since roentgenograms of the sinus area are frequently difficult to read, expert opinion should be sought (a) as to the views to be taken and (b) in the interpretation of the roentgenograms or MRI.


MIGRAINE. Ophthalmoplegic migraine is an uncommon phenomenon that presents in children. Miller found only two admissions to Johns Hopkins Hospital over 25 years. Pried man, Harter, and Merritt, in a review of 5000 admissions for migraine over 30 years, found only eight cases of ophthalmoplegic migraine. It is helpful if the patient presents with the ophthalmoplegia during a typical migraine attack or at least has a strong history for migraine. The fact that the ophthalmoplegia improves is strong support for, but no a guarantee of, a nonaneurysmal cause. If the ophthalmoplegia involves an isolated fourth or sixth nerve, then an aneurysm is extremely unlikely. These considerations and the fact that aneurysms in children usually present as a subarachnoid hemorrhage make it much less necessary to perform arteriography in childhood cases of third-nerve paralysis. Most aneurysms in children are larger than those in adults and have a better chance of being imaged on MRI if done in the appropriate area. However, the presence of any associated neurologic signs and lack of previous history of migraine, particularly with onset between 10 and 20 years of age, changes the circumstances, and arteriography should be performed.

Experience indicates that migraine patients have a significantly higher rate of complications from arteriography than do other types of patients. However, most of the pertinent studies were done in the days of direct carotid arteriography; present techniques using the femoral catheter route might change the statistics significantly. Nevertheless, whether an arteriogram should be done on a patient who presents with a third cranial nerve paralysis, questionably migraine in origin, with pupillary involvement, even though it clears, remains a problem for the physician.

The main diagnostic features of migrainous third cranial nerve paralysis are the transient nature of the ophthalmoplegia, recurrence, rapid clearing of the paralysis, and, frequently, a personal or family history of migraine.

The cause of ophthalmoplegic migraine is unknown as is the true mechanism for migraine itself. Earlier theories by Walsh and O'Doherty suggested compression of adjacent nerves by a swollen carotid artery secondary to migrainous ischemia. Vijayan suggested that the ophthalmoplegia resulted from an ischemic neuropathy. The work of Milisavljevic, mentioned earlier, suggests that both theories can be applied. Thus, penetration of the third cranial nerve by small arteries, which is common, can cause compression or ischemia of nearby structures.

KERNOHAN NOTCH SYNDROME. The Kernohan notch syndrome is a variation of a frequently occurring condition. Ordinarily, supratentorial pressure from a tumor or subdural hematoma compresses the third cranial nerve as it crosses the tentorial edge. Since the pupillary fibers are in the peripheral superior medial part of the nerve, these fibers are frequently affected first, without significant involvement of other parts of the third cranial nerve. The ipsilateral cerebral peduncle is compressed at the same time, and signs of ipsilateral pupillary enlargement and contralateral hemiparesis can be observed.

In the Kernohan notch variation, compression of the ipsilateral third cranial nerve is associated with cross-compression of the contralateral cerebral peduncle, resulting in ipsilateral third cranial nerve paralysis and ipsilateral hemiplegia. This situation leads to confusion as to which side of the head the expanding lesion is on, since these two signs are normally crossed. In determining the location of the lesion, a rule of thumb is to rely on the pupillary sign rather than on the hemiparesis.

ORBITAL FRACTURES. Orbital bone fractures can trap muscles and result in diplopia. The most common fracture is one that occurs in the orbital floor, trapping the inferior oblique and inferior rectus muscles as they form into Lockwood's ligament. Rare cases also exist of simulated superior oblique tendon sheath syndrome with a floor fracture. Medial wall fractures can trap the medial rectus muscle and may look like a lateral rectus palsy; however, the globe retracts with attempts at lateral gaze. This entrapment and pseudo-lateral rectus palsy can be diagnosed by a forced duction test and by saccadic velocity testing.

Orbital roof fractures are less common but have their own set of complications. These complications may be insidious and delayed in onset and are therefore often missed. The most serious of these complications are meningitis and brain abscess secondary to a connection between the intracranial space and the frontal sinuses.

MISCELLANEOUS CAUSES. Herpes zoster ophthalmoplegia occurs when the skin lesions over the first division of the fifth cranial nerve are almost healed. The diagnosis is obvious, since all the signs and symptoms of the herpetic disease are present. The ophthalmoplegia will clear slowly without therapy.

Congenital third nerve palsies are generally considered to be traumatic in origin because of the high incidence of misdirection that occurs in the peripheral nerve. They tend to be an isolated neurologic event without other brainstem signs.

Third cranial nerve paralysis occurs infrequently in connection with lupus erythematosus, encephalitis, amyloidosis, Hodgkin's disease, temporal arteritis, tetanus, sarcoidosis, and rarely, after dental anesthesia. The mechanism in dental anesthesia may be retrograde flow of the anesthetic. More rarely, fat emboli can cause blindness owing to arterial obstruction. Multiple sclerosis rarely causes ophthalmoplegia, and when it does, it is usually in the form of a sixth cranial nerve paralysis. The acquired immune deficiency syndrome has also produced oculomotor paresis, occasionally as a presenting sign. Facial nerve involvement, however, is more common.

The cause of an unresolved oculomotor palsy in children that is not identified by initial studies should be followed up at intervals by repeat imaging studies. Since high-resolution computed tomography (CT) and MRI often allow adequate review of the brainstem without invasive studies, there is little excuse for not pursuing such cases at repeated intervals. Abdul-Rahim and Savino reported on five such cases that revealed tumors of the oculomotor nerve, probably schwannomas, on subsequent studies years later.

Convergence Insufficiency

Convergence insufficiency is a very common problem causing reading complaints in students. Acquired convergence insufficiency or convergence paralysis has been associated occasionally with a diverse group of causes. I have found a significant relationship between this condition and head trauma. In many of my cases, patients who were prepresbyopic before trauma, had convergence insufficiency or paralysis after cerebral trauma, but no medial rectus paralysis on ductions. Many also had a marked presbyopia and no pupillaiy involvement. Some of these cases resolve with time, but most do not. The exact location of the lesion is unknown. Recent studies by Buttney-Ennever enlarged on the anatomic division of the subnuclei of the oculomotor nerve. He believes that the rostral portion of the medial rectus group of cells controls convergence movements. So far no lesion has been identified in this area on imaging studies, but MRI may pinpoint the lesion in the near future.

Fourth Cranial Nerve Paralysis


The trochlear nerve nucleus develops adjacent to the oculomotor nerve in the floor of the fourth ventricle. The trochlear nerve is the only crossed cranial nerve (Fig. 6.22, 1). The fascicle then travels dorsally before crossing in the anterior medullary velum (Fig. 6.22, 2). This is a site often suggested in bilateral trochlear nerve palsies. The fourth nerve emerges from the dorsal surface of the brainstem and travels between the posterior cerebral and superior cerebellar artery as does the oculomotor nerve (Fig. 6.16, 2). The nerve then passes the cerebral peduncle and enters the cavernous sinus where it lies in the substance of the lateral wall adjacent to the oculomotor nerve (Fig. 6.17, 6). It then passes out of the cavernous sinus, into the superior orbital fissure, and into the orbit but outside the annulus of Zinn (Fig. 6.18, 3). The nerve then crosses over the superior rectus muscle and moves immediately to innervate the superior oblique muscle.

fig. 6.22

Figure 6.22. Anil(Onay of trochlear nerve. (Courtesy of William Stewart.)

Unlike the third cranial nerve, the fourth cranial nerve (trochlear nerve) innervates only one muscle, the superior oblique, which causes the eye to intort and turn down. Minimal weakness in this muscle may result in symptoms, since the eye is much less able to overcome vertical than horizontal imbalance. In paralysis of the oblique muscle, the versions are mostly affected; the ductions are frequently full.

A frequent error in testing for a minimal defect in motility is permitting the patient's head to be tilted during testing, which masks the defect. Testing should be done with the patient's head erect, even if it has to be supported manually.


In general, the head-tilt test is useful in distinguishing paralysis of the superior oblique muscle in one eye from paralysis of the superior rectus muscle in the other, since, for all practical purposes, isolated paralysis of the inferior oblique and inferior rectus muscles is uncommon. As noted earlier, the Bielschowsky head-tilt test is positive if further separation of the images occurs when the head is tilted to the side of the affected superior oblique muscle. For instance, in paralysis of the left superior oblique muscle that causes a left hypertropia, the diplopia increases with left head tilt.

In describing a compensatory head position, it is important to describe three facets of the position rather than just the shoulder toward which the head is tilted: (a) whether the chin is elevated or is depressed in an attempt to overcome the vertical aspect of the imbalance, (b) whether the face is turned to the right or turned to the left to overcome weakness in adduction or abduction, and (c) whether the head is tilted toward the left shoulder or toward the right shoulder to overcome torsional weakness. A description of these three head positions represents the attempt to compensate for the three functions of the vertically acting muscles (Table 6.11). For example, in paralysis of the left superior oblique muscle, the chin is tilted down to overcome the weak depression effect of the left superior oblique muscle; the face is turned to the right to overcome the weak abductions; and (last but not least) a right head tilt causes the left eye to extort, thereby overcoming the weak intorsion effect of the paralysis of the left superior oblique muscle.

Exceptions to the foregoing rules are rare, but they do occur. Although a patient presumably chooses a head position to maintain fusion, on rare occasions the patient may elect the reverse position to further separate the images, thus making it easier to suppress the more distant image. Actual testing will show that the head tilt assumed makes the images move farther away rather than nearer. Thus the head position initially seen should not be considered absolutely indicative of the muscle involved.

Sandifer syndrome, an unusual form of head tilting, is seen in children with a short esophagus. In such cases, the head can be easily lifted manually (suggesting no contraction of neck muscles), but the tilt to one shoulder will be resumed as soon as the supporting hand is removed. No muscle imbalance is found on repeated motility testing, which rules out the two common causes of head tilting. A child who is old enough may he able to describe gastrointestinal disturbances that are relieved when the head is tilted and depressed, thus shortening the esophagus. Patients with this condition are best referred to a pediatric surgeon, who can readily diagnose hiatus hernia by use of a barium swallow and treat it.


In evaluating any third cranial nerve paralysis, it is important to detect any associated defects in the fourth and sixth cranial nerves, which may place the lesions in the cavernous sinus or superior orbital fissure. Since third cranial nerve paralysis prevents the eye from being adducted, another form of testing must be chosen to test the vertical action of the superior oblique muscle. The patient should be instructed to attempt to look down and in with the paretic eye while the examiner confirms the effort by observing the fellow eye, which should move down and out. A small intorting maneuver of the affected eye, not vertical depression, is an indication of good fourth cranial nerve function in an eye that cannot be adducted because of third cranial nerve paralysis. The intorsion, which reflects the action of the superior oblique muscle when it is not in the abducted position, is small, and it must be looked for specifically. Watching the movement of a horizontally located conjunctival vessel makes the intorsion easier to see.


Congenital paralysis, a fairly common defect, frequently goes undetected. Many patients' features are cosmetically acceptable, and congenital paralysis of the fourth cranial nerve does not usually involve amblyopia; thus, such patients usually remain undiscovered until they have a routine eye examination. Types of congenital paralysis diagnosed in children of preschool age are either cosmetically obvious or accompanied by a head tilt that suggests a muscle imbalance requiring evaluation.

Brown's syndrome, which involves restriction of the muscle and sheath as it slides through the trochlea, is not innervational but mechanical, and it is therefore an entirely different problem. In a typical case, as the eye is adducted in upgaze, it turns down and in as if an inferior oblique muscle paralysis existecl. The forced duction test easily differentiates this movement from that occurring in inferior oblique muscle paralysis; it is positive when the examiner attempts to move the eye up and in.


Trauma is the leading cause of acquired suporlor oblique muscle paralysis. The trauma can be local, with damage to the trochlea, or more severe, with intracranial damage.

fig. 6.23

Figure 6.23. Differential analysis of fourth-nerve paralysis.

A form of trauma that does not appear to be severe at first glance occurs as the result of a rear-end automobile collision. The history has been similar in the cases I have seen. The stopped car in which the patient was seated was hit from behind, causing a typical whiplash movement, with sudden hyperextension of the head and neck. In some instances, the head struck the steering wheel or dashboard. Usually, the patient was neither rendered unconscious nor sustained more than a bump on the head. Onset of diplopia was immediate. On examination, the eye was neither red nor swollen, indicating trochlea or to the superior oblique muscle in the orbit. The muscle imbalance was minimal and usually cleared over a period of 3 to 6 months.

If the patient has large fusional amplitudes, then consider a congenital palsy that has now decompensated. This can happen as a result of minor trauma, from illness, and for no obvious reason. When the measurements vary from examination to examination or no adequate reason is forthcoming, then consider myasthenia gravis and do a Tensilon test. A ptosis is an extremely common presentation of myasthenia gravis, although it is not 100%, and its absence does not rule out myasthenia gravis. Vasculopathic causes such as diabetes and hypertension are quite common; less common are infectious diseases such as herpes zoster. If none of these causes are present and there are not very wide fusional amplitudes or other neurologic signs develop with worsening of the fourth-nerve palsy, then imaging studies such as MRI are indicated. Imaging is not required for the isolated palsy.

As many as 30% of cases of acquired trochlear nerve paralysis are bilateral. The second trochlear nerve palsy may be more subtle than the first and is frequently missed until it shows up after surgery for the first palsy. To properly identify the bilaterality of this condition, first neutralize the horizontal deviation with prisms during the examination. Then a right hypertropia in left gaze will become a left hypertropia in right gaze. If it is subtle, be sure to examine the motility in the oblique fields.

Superior oblique myokymia is a rare phenomenon. The cyclovertical torsion of the eye causes oscillopsia. This condition was first reported by Hoyt, and all cases appear to be benign. The cause is unknown, and treatment if any is Tegretol. Surgical correction has not been particularly satisfactory and may even cause diplopia.

Lesions in the cavernous sinus or superior orbital fissure can cause a combination of trochlear nerve palsy and ipsilateral Horner syndrome. However, lesions in the brainstem can cause ipsilateral Horner syndrome and contralateral trochlear palsy. A skew deviation is an alternative reason for the diplopia, but appropriate measurements of the hypertropia and head tilt confirm a trochlear palsy and brainstem location.

In cases of acquired paralysis of the fourth cranial nerve, the exact location of the injury to the nerve is uncertain. However, a clue may be gleaned from three bilateral cases that I have seen in which the head injury was slightly more serious. Two of these patients had experienced a period of unconsciousness, but in all three, superior oblique muscle paralysis was the only sign of intracranial trauma, despite the severity of the head blow. Since both fourth cranial nerves come together only in the anterior medullary velum where they cross, it seems likely that the injury occurred in this area. Of further note, the bilateral cases that I have seen have not cleared spontaneously but have required a surgical procedure to resolve the problem. The sign that suggests bilateral fourth nerve involvement is the alternating hypertropia. There is a left hypertropia in right gaze and a right hypertropia in left gaze. The Bielschowsky head-tilt test also changes from side to side if one looks to the right or to the left when performing the test.

In certain surgical approaches to frontal sinus disease, the trochlea is moved and may be damaged; however, these procedures are now rarely used. Other causes of superior oblique muscle paralysis, such as vascular disease and diabetes, have been diagnosed, but trauma is by far the most frequent cause. Although isolated fourth cranial nerve paralysis owing to diabetes is uncommon, a glucose tolerance test is indicated in certain patients.

A form of intermittent Brown's tendon sheath syndrome occurs in adults. The usual complaint is intermittent diplopia immediately preceded by what is described as a popping sensation in the area of the trochlea, sometimes accompanied by mild pain and tenderness. The diplopia may last from minutes to days or weeks. The same forced duction findings seen in a true Brown's syndrome are found while the patient is symptomatic. Although the cause of this disorder is unknown, systemic steroids seemed to help in two cases that I have treated. I have not tried local injection of steroids.

Other inflammatory causes such as rheumatoid arthritis, herpes zoster, cysticercosis, and Tolosa-Hunt syndrome have been infrequently responsible. If the vertical deviation does not fit a trochlear nerve paresis or there is no cyclotorsion or it varies from examination to examination, consider a skew deviation.

Sixth Cranial Nerve Paralysis

ANATOMY (FIGS. 6.24, 6.25)

Differentiation of the sixth nerve nucleus occurs between the fourth and sixth week of gestation. This is coincidental with the development of the fourth and third nerve nuclei.

fig. 6.24

Figure 6.24. Anatomy of sixth nerve, part 1. (Courtesy of William Stewart.)

The sixth nerve nucleus develops in the floor of the fourth ventricle (Fig. 6.24, 1). It appears somewhat elevated because of the passage of the seventh nerve around it. The neighboring structures are the medial longitudinal bundle, which lies medial, The sixth nerve nucleus also lies dorsal and medial to the vestibular nuclei. These are important structures since lesions of the sixth nerve in this location will cause signs that identify the anatomic lesion as intrinsic to the brainstem. The fascicle travels ventrally and exits at the pontomedullary junction. The nerve then passes vertically in close approximation to the anterior inferior cerebellar artery (Fig. 6.16, 2). It continues through the subarachnoid space to pierce the dura. The dura is folded over the petrous bone and is called Gruber's ligament (Fig. 6.25, 4). The nerve passes beneath it through a space called Dorello's canal and enters the cavernous si-nus (Fig. 6.17, 6). The sixth nerve lies in the cavernous sinus near the carotid artery rather than in the wall of the sinus like the oculomotor and the trochlear nerves. It then travels through the superior orbital fissure and through the annulus of Zinn to innervate the lateral rectus muscle.

fig. 6.25

Figure 6.25. Anatomy of sixth nerve, part 2. (Courtesy of William Stewart.)

Like the fourth cranial nerve, the sixth cranial nerve (abducens nerve) innervates only one muscle, the lateral rectus, which moves the eye laterally. Unlike the superior oblique muscle, the lateral rectus muscle has no secondary or tertiary function. It comes into play primarily when the eye is fixed on a distant object. Many a case of minimal sixth cranial nerve paralysis has been missed because the patient was asked to look at a light 3 feet distant rather than 20 feet distant during the alternate-cover test.

Just as head positioning can mask a fourth cranial nerve paralysis, so a head turn can mask a minimal sixth cranial nerve paralysis if the patient is examined only in the straight-ahead or near position. The diplopia that occurs with lateral rectus muscle paralysis is homonymous owing to the esotropiathat is created. That is, when two images are seen, the ipsilateral image disappears when the ipsilateral eye is covered.

The measurement of horizontal vergences reveals the ability of a muscle to overcome stress. Some people can overcome 30 prism diopters of muscle imbalance, which permits certain muscle weaknesses to be overcome before diplopia is experienced. This compensation is different from that of muscles innervated by the fourth cranial nerve, which work primarily in a vertical direction and may overcome only 1 or 2 prism diopters of imbalance. If a patient has a progressive weakening of the lateral rectus muscle, as may occur with increasing intracranial pressure, it will show up in the distance muscle balance. In successive examinations (either days or weeks apart) of the distance phoria, the measurements increase significantly toward the esophoric side. The first readings may be an esophoria of 2 diopters, indicating a small degree of esophoria at distance that will become moderate as the intracranial pressure increases. The increase in measurement will take place before the occurrence of frank diplopia, owing to the vergence reserve of the horizontal fusional mechanism. This testing technique is particularly useful in deciding whether a blurred disc is true papilledema. While the disc problem is being evaluated, over several days, the change in the measurements of the distance phoria may be significant. By the time a frank sixth cranial nerve paralysis has developed or both sixth cranial nerves are shown to be involved, the diagnosis of increased intracranial pressure is obvious.

Synkinetic phenomena of the third cranial nerve are well known and not rare, whereas mynkinetic phenomena involving the sixth cranial nerve are rare. I have not seen a case, but Spaeth reported on a small series in 1950.

Although isolated sixth cranial nerve paralysis is not uncommon, it is almost impossible to adequately identify its cause or specific anatomic location. Therefore, the physician must be familiar with the syndromes and associated signs, both major and minor, that may help to identify either the location or the cause of this disorder.

The refinement of imaging techniques now permits identification of lesions--even in the abducens fascicle--that previously could not be detected. MRI has proven particularly useful in examining for lesions in the brainstem and parenchyma. If a chordoma, chondrosarcoma, or meningioma is suspected, then a computed tomogram with hone windows should be ordered to look for bone erosion or hyperostosis. The best way to identify the anatomic location and possible causes of a sixth nerve paralysis is to identify the neurologic company it keeps. Isolated cases are mostly due to vascular causes such as hypertension, arteriosclerosis, and diabetes. Tumors are a much more common diagnosis in children, unless they occur after upper respiratory tract infections. Therefore in isolated cases, the initial workup should include a CBC, sedimentation rate, collagen vascular battery of tests, and thyroid studies. If the diagnosis is not clear, then a Tensilon test should be performed. A CT or MRI study may be done initially or postponed for 2 to 3 months, by which time most vascular cases will be recovered. If no improvement occurs within 3 months, imaging studies with views along the course of the sixth nerve are necessary. Bone scans at the base of the brain are also indicated as well as a lumbar puncture and nasopharynx investigation. If the sixth nerve palsy is progressive and no cause can be identified, a forced duction test should be done. The progression may be due to contraction of the medial rectus muscle rather than to some progressive disease of the sixth nerve. If surgical correction of the sixth nerve palsy is not contemplated, botulinum toxin injection of the medial rectus muscle should be tried.

If the cause of the sixth nerve paresis remains unknown, the diagnostic studies should be repeated at 6 months. Experience teaches us that tumors progress and cause increasing compression and unrelenting signs and symptoms. However, there are several explanations for sixth nerve palsies that spontaneously recover in such a clinical setting: remyelination, axonal regeneration, displacement of the nerve off a growing mass, and restoration of blood flow to the nerve.


Möbius syndrome involves complete bilateral paralysis of the sixth and seventh cranial nerves. Its cause is unknown, but a carefully taken drug history may be revealing. I had two patients whose mothers had taken thalidomide during pregnancy. In most cases, however, the cause is unknown.

fig. 6.26

Figure 6.26. Differential diagnosis of sixth-nerve paralysis.

Duane syndrome is one form of congenital sixth cranial nerve paralysis. There are three types with different electromyographic responses. Type 1 reveals limited or absent abduction and normal adduction. Type 2 has limited or absent adduction and normal abduction. The third type has limited abduction and limited adduction. Since most patients with this condition are not esotropic in the primary straight-ahead position, the condition is frequently not discovered early in life. When it is finally diagnosed, whether the sixth cranial nerve paralysis is old or of recent origin becomes a serious question. Bilateral paralysis of recent onset may be caused by increased intracranial pressure or brainstem glioma. Bilateral Duane syndrome is rare (Fig. 6.27 A, B, and C).

fig. 6.27a
fig. 6.27b
fig. 6.27c

Figure 6.27. Patient with bilateral Duane syndrome. A.'Straight and primary position. B. Right Duane with decrease in right lateral function and left superior oblique overaction. C. Similar picture in patient with left Duane. (Courtesy of Dr. Caleb Gonzalez, Strabismus and Ocular Motility, Williams and Wilkins, 1984)

Examination for the additional signs of Duane syndrome assists the observer in differentiating between this condition and an acquired sixth cranial nerve paralysis. Patients with Duane syndrome are not usually esotropic in the primary position, and they do not develop amblyopia as a rule. If they look into the field of the paretic muscle, no diplopia results, because this movement brings into play a mechanism called facultative amblyopia rather than the amblyopia exanopsia usually seen with esotropia. The key sign, however, is a narrowing of the lid fissure on adduction and return of the fissure to normal when the eye reverts to the straight-ahead position. In at least some cases, lid narrowing is caused by simultaneous innervation of the lateral and medial rectus muscles. This has been confirmed by electromyographic studies and autopsy material reported by Hoyt and Nachtigaller and again by Hotchkiss, Miller, Clark, and Green. Auditory evoked material, as reported by Jay and Hoyt, further suggests a primary brain-stem malfunction rather than a peripheral cause. On adduction, therefore, the medial and lateral rectus muscles cocontract, caus-ing enophthalmos and narrowing of the lid fissure. Once lid narrowing is observed, the malfunction can be diagnosed as congenital, and further diagnostic procedures are unnecessary.

In rare circumstances, Duane syndrome can be acquired. I have seen one case caused by trauma, and at least one case has been reported in a patient with rheumatoid arthritis. The electromyographic studies are different in these cases from those in the congenital form, which shows cocontraction of the lateral and medial rectus muscles. In addition, superior or inferior oblique muscles may overact.


VASCULAR OCCLUSION. Vascular accidents involving the brainstem occur frequently. The two syndromes that usually accompany such accidents are (a) Foville syndrome, which combines sixth and peripheral seventh cranial nerve paralysis with homolateral  gaze paralysis, and (b) the Millard-Gubler syndrome, which consists of sixth and peripheral seventh cranial nerve paralysis and contralateral hemiplegia.

In elderly patients, spontaneous isolated sixth cranial nerve paralysis can occur and frequently disappears within several months. One explanation of this malfunction, proposed as far back as Cushing's time, holds that it is caused by compression of the sixth cranial nerve on the anterior surface of the pons by the lateral branches of the basilar artery, particularly the anterior inferior cerebellar artery.

Sixth cranial nerve palsies can also be caused by temporal arteries and polyarteritis nodosa. Both these conditions can be diagnosed by a temporal artery biopsy. The sedimentation rate is also elevated in both diseases. The pathology is the differentiating feature.

GRADENIGO SYNDROME. Gradenigo syndrome is rarely seen in this age of antibiotics. The inflammation often involves severe otitis media and mastoiditis with secondary petrositis, in the course of which the area of the petrous bone called Dorello's canal is affected, and the sixth cranial nerve becomes paretic. In more extensive cases, the gasserian ganglion is affected, causing severe pain lit the temporoparietal area. Further extension of the infection may also involve the seventh and eighth cranial nerves. The most serious complications of unchecked infection the meningitis, epidural abscess, or dural sinus thrombosis with secondary pseudotumor cerebri.

LATERAL SINUS THROMBOSIS. Lateral minus thrombosis is frequently idiopathic. The resulting inflammation extends to the Inferior petrosal sinus, which is adjacent to Ili sixth cranial nerve, thus causing malfunction of this nerve. Thrombosis may occur postpartum or from distant emboli, as, for example, in venous stasis of the legs.

SUPERIOR ORBITAL FISSURE SYNDROME. Also called Tolosa-Hunt syndrome, superior orbital fissure syndrome may begin as an isolated sixth cranial nerve paralysis. The course, diagnosis, and treatment of this syndrome are discussed above.

In treating patients with symptoms located al the superior orbital fissure, the physician must be careful not to err by failing to rule out all the diseases possible in the cavernous sinus, superior orbital fissure, and posterior orbit. The same rule applies to suspected lesions in the cavernous sinus, such as arterio-venous fistulas and tumors. It is unusual to see only an isolated sixth cranial nerve paralysis without symptoms relating to the other cavernous sinus nerves; however the literature includes several cases of a sixth cranial nerve paralysis in the cavernous sinus, with further signs developing only 6 Months after onset of the paralysis.

CEREBELLOPONTINE ANGLE TUMORS. Although cerebellopontine angle tumors cause paralysis of the sixth cranial nerve, other prominent signs frequently precede the onset of this paralysis. The seventh and eighth cranial nerves are affected, and frequently, anesthesia of the cornea is present. All three nerves should be tested when an obscure sixth cranial nerve paralysis is seen. Sixth cranial nerve paralysis is not the presenting sign of an angle tumor, but it may be the complaint that leads the patient to seek medical attention. Selective audiograms, and a CT/MRI study outlining the cerebellopontine angle will demonstrate an acoustic neuroma, the type of tumor most commonly found in that area. Of cerebellopontine angle tumors, 90% are acoustic neuromas. The most common causes of the remaining 10% are meningioma and cholesteatoma. In the case of meningiomas, radiographic studies may show osteoblastic changes in adjacent bones. In the case of cholesteatomas, contrast studies demonstrate a scalloped edge to the tumor unlike the smooth surface of an acoustic neuroma. A facial tic is also more common with cholesteatoma. Although not 100% accurate, spinal fluid protein tends to be normal rather than markedly elevated as it is in other tumors located in that area of the angle, which obstruct the flow of cerebrospinal fluid as do acoustic neuromas (Fig. 6.28).

fig. 6.28a
fig. 6.28b

Figure 6.28. A. Patient presents with sixth-nerve paresis from cholesteatoma. B. Same patient with extensive bone erosion.

SPINAL ANESTHESIA. Transient sixth cranial nerve paralysis following spinal anesthesia is a rare but definite occurrence. In a review of 10,400 cases involving spinal anesthesia, Phillips et al. found that transient sixth cranial nerve paralysis had occurred in eight patients. Thorsen felt it was more common, with an incidence of 1 in 400 procedures. This condition is much less common in my experience. Although this transient paralysis appears to be more common with the injection of contrast material than with a simple spinal tap, the incidence is too small to demonstrate this conclusively.

One explanation for this paralysis is that it represents a toxic response to the injected material. This is unlikely, since there is a delay of 1 or 2 weeks before the symptoms develop. A second explanation is more plausible. The patients that have experienced lateral rectus paresis have other post-spinal tap symptoms, such as headache in the erect position. These symptoms have generally been considered to be caused by chronic leakage of spinal fluid from the tap site, with displacement of the brain and traction on pain-sensitive structures. It is not hard to extrapolate this theory to include traction on the sixth nerve with compression over firm structures such as the petrous bone.
Spinal fluid examinations to determine the presence of a cellular reaction indicative of arachnoiditis are not usually performed. Paralysis owing to spinal anesthesia clears rapidly (within days or a week or two) and requires no further investigation if the rest of the neurologic examination is normal. In summary, the cause of this type of nerve malfunction is obscure, and the treatment is to do nothing.

WERNICICE-KORSAKOFF SYNDROME. When associated with chronic alcoholism, Wernicke-Korsakoff syndrome includes several eye signs, among which are nystagmus of a nonspecific character, horizontal gaze paralysis, sixth cranial nerve paralysis, and, rarely, vertical gaze paralysis. If a patient with sixth cranial nerve paralysis has a history of alcoholism, Wernicke-Korsakoff syndrome is frequently the diagnosis; however, the physician should not fail to rule out the possibility that an alcoholic patient may have sustained some intracranial trauma, and that the sixth cranial nerve paralysis could be the result of increasing intracranial pressure from a subdural hematoma. In addition, the patient may suffer from ataxia of gait and from somnolence, the other two symptoms in the triad described originally by Wernicke. The confabulation symptom of the Korsakoff syndrome was added later, and although it is frequently seen, it is not an essential feature of the Wernicke syndrome.

TOXIC DRUG REACTIONS. In cases of sixth cranial nerve paralysis, always consider toxic drug reactions and take a careful drug history. Lateral rectus muscle paralysis has been reported in connection with such drugs as furaltadone and iodochlorhydroxyquin. Optic neuritis is a much more common problem with iodochlorhydroxyquin, but sixth cranial nerve paralysis has also been reported.

INCREASED INTRACRANIAL PRESSURE. Sixth cranial nerve paralysis secondary to increased intracranial pressure is well known. What produces paralysis is not completely understood --stretching of the nerve with bony impingement owing to downward displacement of the brainstem? compression by branches of the basilar artery? The early sign of this condition, increasing esophoria, is discussed above.

An unusual but dramatic form of sixth cranial nerve paralysis is one that is transient but occurs suddenly and repeatedly. Patients with this disorder develop sudden headaches that are accompanied by paralysis of one or both lateral rectus muscles. All symptoms disappear within minutes or hours. While the signs are present, patients are usually active rather than lethargic, and they frequently shake or hit their heads as if to push out whatever is causing the symptoms. This maneuver suggests that some ball-valve mechanism may be causing sudden changes in intracranial pressure. A colloid cyst of the third ventricle that is intermittently closing off the aqueduct can do this.

Compared with tumor, pseudotumor cerebri is a relatively infrequent cause of increased intracranial pressure. If this condition is to be ruled out as a causative factor, however, (a) diagnostic tests should reveal signs of increased intracranial pressure; (b) the results of a spinal fluid examination should be negative except for increased opening pressure; (c) the neurologic examination should be totally negative, except for the possible presence of a sixth cranial nerve paralysis if the intracranial pressure is high enough; and (d) a CT examination should reveal a normal size ventricular system. If there were a small lesion near to and compressing the aqueduct, then the ventricles would be enlarged, suggesting a non-communicating hydrocephalus.

Most cases of pseudotumor cerebri are idiopathic. Some are related to specific causes such as chronic vitamin A intoxication, tetracycline overdose, postpartum dural sinus thrombosis, and the institution or withdrawal of steroids in treating nephrosis or are a complication of Addison's or hypoparathyroid disease. Treatment, if any, varies, but it usually consists of steroid therapy over a period of several weeks.

Pseudotumor cerebri is self-limited, but it may persist for months. The real danger of a prolonged course is secondary atrophy of the optic nerve. Thus, vision and fields should be checked frequently, and the disc should be observed for beginning gliosis-a II of which indicate optic nerve decompensation and suggest the need for more vigorous treatment of the intracranial pressure.

NASOPHARYNGEAL MALIGNANCY. Nasopharyngeal malignancy (Godtfredsen syndrome) is an uncommon disease that often affects the sixth cranial nerve. For example, in a series of 53 cases reviewed by Smith and Wheliss, sixth cranial nerve paralysis was present in 29. A significant feature of this type of sixth cranial nerve paralysis is associated pain or paresthesia over the second division of the fifth cranial nerve. This combination of signs should impel the physician to look for other signs that suggest nasopharyngeal malignancy.

The main presenting complaints usually include cervical lymphadenopathy, pain in the ear, pain in the face, and symptoms of nasal obstruction. Most patients have also experienced unexplained weight loss prior to seeking medical attention, at which time a mild or frank sixth cranial nerve paralysis may be present. The third and fourth cranial nerves may also be affected initially, but not as frequently as is the sixth cranial nerve. Usually, by the time multiple orbital nerves are involved, exophthalmos is also present. Mention of serous otitis media, another significant symptom of this disorder, may be omitted by the patient as of minor importance compared with the diplopia and pain being experienced. Intermittent blockage of the eustachian tube, causing a popping sensation or a blocked ear, is another and more subtle symptom of this disease.

Even when a nasopharyngeal malignancy is suspected, it may be difficult to establish. Examination is frequently misleading, since the tumor initially arises submucosally in the nasopharynx. It then grows intracranially but extradurally, affecting one cranial nerve after another before becoming a significant space-occupying lesion. Therefore, a biopsy is indicated even if the nasopharyngeal mucosa looks normal. A good location for biopsy is in the area called Rosentniiller's fossa, which is a common site for this type of tumor.

Special roentgenographic views may reveal another feature of a nasopharyngeal malignancy. A nasopharyngeal tumor usually gains entrance to the cranium by way of the basilar foramina. The usual skull series does not include these openings; therefore, roentgenographic views of the basilar foramina should be ordered specifically, since they may reveal significant erosion of one or more foramina.

On the basis of cell type, lymphoepithelioma is the most common form of tumor found in the nasopharyngeal area. Such tumors are relatively radiosensitive, and radio-therapy may result in a temporary amelioration of symptoms; however, the 5-year survival rate for patients with this type of malignancy is only about 25%. Although malignant nasopharyngeal tumors are rarely seen in the United States, this condition is the leading cause of cancer among males in mainland China. It is not common among people of Chinese descent living in enclaves in Hawaii and San Francisco.

Dental Anesthesia

Sixth-nerve palsies have infrequently incurred with local dental anesthesia. One explanation is injection of the anesthetic into the superior or inferior alveolar artery. The anesthetic travels by retrograde flow or from the pressure of the injection into the maxillary artery and then into the middle meningeal artery and its orbital branch, which anastomoses with the lacrimal branch of the ophthalmic artery.

Sixth Cranial Nerve Paralysis in Children

When sixth cranial nerve paralysis develops suddenly in an otherwise normal child, it is certainly cause for concern about the possible presence of a serious disorder. Such disorders as increased intracranial pressure or a pontine glioma are always a possibility.

In 1967, Knox, Clark, and Schuster reported on a series of children with spontaneous sixth cranial nerve paralysis that occurred about 7 to 21 days after a nonspecific illness. Spinal fluid tests were usually negative, and no attempt was made to establish that the paralysis was infectious in origin. The patients developed no other symptoms, and the paralysis cleared within several weeks. I have seen two such cases in my practice, and I believe that it is proper to wait if no other signs of increased intracranial pressure exist and the remainder of the neurologic examination is normal. Some noninvasive studies such as CT and MRI may be performed, but the invasive studies may be deferred a week or two, provided that the foregoing criteria are met. A spontaneous sixth cranial nerve paralysis begins to clear in a week or two, whereas a pontine glioma does not. Nevertheless, a week's delay probably does not affect the outcome of a pontine glioma.

Spasm of the Near Reflex

Spasm of the near reflex is not a common syndrome but will be seen by all ophthalmologists from time to time. The patient presents with convergence of the eyes, small pupils, and increased accommodation. The patient reports diplopia and blurring of vision at distance. These are signs of an exaggerated expression of the near reflex, which is convergence, increased accommodation, and miosis. The clue to a functional cause is found with observation. Functional patients cannot maintain this condition for more than 1 or 2 minutes. Just look at them, ancl they will then stop doing it. They will usually cover their eyes as if in pain, because they cannot maintain the convergence spasm and do not to want to reveal that fact. This may be a conscious or subconscious effort. For confirmation, retinoscope and refract to good vision without cycloplegia, and you will find that these patients are quite myopic. After cycloplegia, the same vision can be attained with considerably less myopic correction or even a plano prescription.

Pathologic cases have been reported with intercranial infection, labyrinthine disease, brain tumors, and trauma. The clinical circumstances of these cases are quite different and are not usually confused.

Lecture 6 of 7 «Previous Lecture   1 2 3 4 5 6 7    Next»
footer logo