New advances in the understanding of extraocular muscle and orbital anatomy and function are leading to increased accuracy in diagnosis of cranial nerve (CN) palsies and pareses. Multiple conditions may simulate the strabismus pattern of a CN palsy without any innervational abnormality at all. These include muscle displacements, orbital and pulley abnormalities, flap tear, Duane’s syndrome, heavy eye syndrome, and thyroid eye disease. These should be termed “pseudopalsies.”
Superior oblique (fourth CN) palsies and pareses may be congenital or acquired, unilateral or bilateral. There are a number of surgical approaches to these cases, but all aim for slight undercorrection rather than overcorrection, with particular attention to torsion. Approaches include directly strengthening the weak superior oblique, or weakening other ipsilateral or contralateral eye muscles.
Third cranial nerve palsies and pareses create difficult strabismus management issues due to the complex intracranial and distal anatomy of the nerve. The third CN innervates the medial, superior, and inferior rectus muscles, and the inferior oblique. The levator muscle and pupillary fibers also travel with the third CN, so coexisting ptosis and pupillary abnormalities are frequent. Aberrant regeneration is common following third CN palsy, which further complicates strabismus management. Etiologies may be congenital or acquired, and involvement may be complete or partial. Multiple surgical approaches are possible, and the approach is tailored to the degree of muscle function and the number of muscles involved.
Sixth cranial nerve (lateral rectus) palsies and pareses may result from neurologic or vascular abnormalities, but may also be mimicked by lateral rectus displacement. When complete palsy exists, transposition procedures are used, and several differenat approaches are available. When some lateral rectus function remains, then a simple recess-resect procedure may suffice.
18 Cranial Nerve Palsies
This large topic is a work in progress written at a time when knowledge in this area is changing. The new knowledge of the last decade (and possibly the next decade) will change the way we assess and treat cranial nerve (CN) palsies.
This new knowledge includes the following:
Most of the extraocular muscles (EOMs) have a binary nerve supply. 1 , 2 The superior compartment of the horizontal rectus muscles has a different nonoverlapping nerve supply from that of the inferior compartment. In the superior oblique (SO) muscle, the presumed torsionally acting and vertically acting parts of the muscle each have a different nonoverlapping nerve supply. We do not yet know what this means clinically and therapeutically. We do not know, for example, how the clinical presentation of a superior compartment medial rectus (MR) palsy differs from a paresis of both compartments, do not know the natural history, and do not know the best treatment. The new clinical, radiologic, and therapeutic understanding and importance of these findings will probably change the way EOM palsies are evaluated and managed.
Recent radiologic findings in acquired ocular motility defects have shown that long-held diagnostic labels may no longer apply or be accurate. Uncrossed diplopia due to an abduction deficit in the elderly is more likely to be due to a sagging lateral rectus (LR), a mechanical aging change, than to a sixth nerve paresis, though in the office they can look the same. 3 , 4 The label “superior oblique palsy” can now only be used with confidence if a coronal scan shows the muscle to be atrophic. Without this radiologic sign, the label of SO palsy is tentative and often wrong. 1 , 5
Surgical techniques change. Old techniques that had been abandoned are modified and revived. One example is the transposition of the LR to the nasal side of the globe in third nerve palsy. 6 , 7 Another is the use of plication instead of resection; plication lessens the risk of anterior segment ischemia in surgeries requiring multiple muscles (Chapter 24). 8 New techniques are developed that might capture an important place in the repertoire, e.g., the single vertical rectus transposition in sixth nerve palsy. 9 , 10
The literature that guides our clinical practice is necessarily older than these new concepts, and may be wrong. As we learn more about these newer concepts we will apply them with increasing confidence and success of our patients’ outcomes.
In this chapter we will consider the third, fourth, and sixth CN palsies and pareses, their assessment, their differential diagnoses, and principles of treatment. Though typically grouped together in a chapter like this one, they are very disparate clinical entities with different core issues.
Finally, ophthalmologists are often imprecise in their use of the terms paresis and palsy. In this chapter we will try to be precise: palsy refers to a functionally “dead” muscle, paresis to a muscle that still has some, but significantly reduced, function. This difference can usually be determined with some confidence in 3rd and 6th nerve conditions, but is often difficult in 4th nerve pathology.
18.2 Fourth Nerve Palsy and Paresis/Superior Oblique Palsy and Paresis
There has long been a fault line between the “lumpers” and the “splitters” in diagnosis and treatment planning in SO palsy and paresis (SOP).
Splitters make a careful clinical and radiologic diagnosis, take careful measurements, describe the subtype of SOP, and typically choose a treatment plan from a variety of surgical precedents.
Lumpers tend to use the label SOP for any cyclovertical abnormality, especially one with inferior oblique (IO) muscle overaction (IOOA), and will do an IO weakening procedure as a first procedure for all patients. 11
Knapp was a pioneer of the splitters, and his schema of seven subtypes of SOP is still useful today. 12 Other splitters are Guyton, with his intraoperative traction techniques, 13 Jampolsky, who pointed out that superior rectus (SR) contracture was a common outcome of SOP and needed special recognition and different surgery, 14 and Demer, who pointed out the need for radiologic evaluation and more careful labeling of possible SOP. 1 , 5
The splitter will try to segregate “true” SOP from conditions that resemble SOP. The lumper will tend to label as SOP all cases of true SOP and also patients with cyclovertical anomalies that resemble SOP. Even the most careful splitter cannot reliably differentiate between superior oblique paresis and palsy, and ’SOP’ can usually refer to either paresis or palsy.
Both lumpers and splitters will use one muscle for small deviations, two for larger ones. The step-up from one to two muscles is usually in the range of 15 to 20Δ.
Conditions that simulate SOP that the splitter will claim as justifying this approach include oblique dysfunction secondary to orbital abnormality (e.g., plagiocephaly 15 ), thyroid eye disease 16 (Chapter 12 Thyroid Ophthalmopathy), flap tear of inferior rectus (IR) (Chapter 20 Traumatic Strabismus: Direct Orbital and Muscle Trauma and Flap Tear, Chapter Chapter 29 Traumatic Strabismus Repair (Flap Tear, Slipped and Lost Muscles)), skew deviation, 17 and pulley anomalies 5 (Chapter 19 Displaced Muscles (Pulley Heterotopias), Chapter 30 Pulley Surgeries).
In this chapter, the term SOP will henceforth be used as by a splitter. A cyclovertical anomaly that resembles SOP but is not a true SOP will be referred to as a pseudo-SOP.
18.2.1 Congenital and Acquired Superior Oblique Palsy and Paresis
The phenotypes of congenital SOP (C-SOP) and acquired SOP (A-SOP) are easily recognised at their extremes, with a large area of overlap.
C-SOP that presents in the first 5 years of life will often be associated with hemifacial asymmetry, marked IOOA, and a head tilt. C-SOP that presents in the third decade of life typically does so because of neck and shoulder problems from chronic torticollis. Special occasion photos will usually show a head tilt in the past. The angle of primary position is often large (at least >10–15 Δ) and the muscle often shows atrophy on a coronal scan. Sensory testing is sometimes confusing because of associated suppression and/or an expanded range of motor fusion.
The typical A-SOP phenotype is more likely to involve head trauma causing loss of consciousness (sometimes years previously) and presents with diplopia and a torticollis that corrects the diplopia. In an older population, A-SOP may be due to microvascular causes. Tumor is a rare cause of A-SOP, though the frequency of trochlear schwannoma may be proportional to the persistence of the investigator. 18
Skew deviation (skew) resembles SOP, but can be differentiated:
In skew, the head tilt does not decrease diplopia (is not therapeutic), whereas in SOP the head tilt does correct diplopia (is therapeutic).
In skew, changing from a head erect to a head supine position will correct the diplopia. In SOP this has little or no effect. 17
18.2.2 Radiologic and Clinical Correlation
In a Japanese cohort, if the SO muscle was atrophic, the tendon was always floppy when the patient was anesthetised. 19 In Sato’s cohort, the SO tendon did not always insert onto sclera but sometimes into Tenon’s capsule.
In an Australian cohort, if the SO muscle was atrophic, only a minority of patients had a floppy tendon when the patient was anesthetised. 20 These two cohorts had different genetic bases, and there may have been some other unrecognized selection bias as well.
Many surgeons believe that the best treatment for a floppy SO tendon in SOP is to tighten the tendon (Chapter 9 The Superior Oblique, Chapter 24 Plications and Tucks, Chapter 26 Superior Oblique Surgical Techniques). This is not in every surgeon’s repertoire and may have higher morbidity than other SOP surgeries. Recognition of radiologic atrophy in advance is thus worthwhile in order to appreciate a situation of potentially higher surgical complexity and more difficult natural history after surgery.
18.2.3 Bilateral Superior Oblique Palsy and Paresis
Bilateral SOP is seen after closed head injury and after brainstem damage and presents with torsional diplopia on downgaze, with or without incomitant vertical misalignment and downgaze esotropia.
18.2.4 Principles of Surgical Treatment
When an operation is needed to reduce the misalignment, the following principles apply:
If the surgeon is treating diplopia, aim to reduce it to zero tropia, but do not overcorrect. There may be a tiny residual vertical heterophoria.
In treating torsion, aim to reduce it to near zero, but do not overcorrect.
When treating a presumed C-SOP with some suppression or enhanced motor fusion, aim to reduce it to a small, slightly undercorrected misalignment. The patient has every sensory and motor adaptation to handle a small undercorrection, but no mechanism to handle a small overcorrection.
You need to avoid any overcorrection unless it is in a rarely used direction of gaze, such as upgaze in a tall patient.
If a muscle is floppy, typically the SO, it must be tightened.
If a muscle is tight, typically the SR, it must be weakened.
If none of the above apply, this author’s default operation is usually IO weakening.
If fairly comitant, this author prefers recession of the vertical rectus muscles—one muscle if 12Δ or less, two muscles if more than 12Δ.
Ipsilateral SR and IO recessions should be avoided. If done in the one surgery, they are prone to cause an upgaze deficit that may be progressive.
Vertical rectus recessions can induce torsion. The SR is an intorter, and recession of the SR can produce or augment extorsion. Temporal shift of the SR (e.g., draping it from the temporal edge of its insertion for an adjustable recession) will lessen this torsional effect.
22.214.171.124 Inferior Oblique Muscle Weakening
There are many types of IO muscle weakening (IOW) procedures (Chapter 8, Chapter 25). There are four preferred by this author, as follows.
126.96.36.199.1 Standard IOW (Approximates Parks’ Recession)
The anterior corner of the IO is sutured to a point 3 mm temporal to and 3 mm posterior from the lateral edge of the IR insertion (Chapter 25, Video 25.1). The posterior corner of the IO is not sutured and allowed to hang back; this lessens the risk of antielevation. This corrects 10 to 15 Δ of hypertropia in primary position. Some authors claim this can be used for small and even zero primary position deviations without causing a hypotropia. 21 There is some unrecognized selection bias in these authors’ experience or in mine, because I cannot agree with this, having seen problems with my own patients and with others’ patients.
188.8.131.52.2 Anterior Transposition of the IO
The anterior corner of the IO is sutured to the lateral edge of the IR insertion (Chapter 25, Video 25.2, Video 25.3). The posterior corner of the IO is not sutured and allowed to hang back; this lessens the risk of antielevation. This corrects 15 to 20 Δ of hypertropia in primary position and may be more effective in correcting torsion than standard IOW.
184.108.40.206.3 Adjustable IOW (after Alan Scott)
The adjustable IO recession, as modified by Kowal, is detailed in Chapter 25, Section 25.3 Inferior Oblique Weakening Surgeries, text box.
220.127.116.11.4 Anterior Incomplete Disinsertion (for Torsion)
This was popularized by Kushner, and reliably reduces extorsion by about 3 degrees. 22
Disinsert the anterior 75% of the IO. Pass a 6–0 polyglactin suture through the anterior corner of the disinserted muscle, and lasso the bulk of the muscle with this suture, ensuring that the anterior corner will not readhere to the original insertion.
18.104.22.168 Superior Oblique Surgery
There are two types of SO surgeries, whole tendon and partial tendon (Chapter 9, Chapter 26).
22.214.171.124.1 Partial Tendon
Partial tendon surgery was popularized by Harada and Ito, with subsequent variations published by others.
This author prefers the “original” technique. Split the distal tendon into a temporal one-third. Take a suture bite of this temporal part 4 mm from the insertion, and suture it to sclera several millimeters toward the LR. Ensure that you have not produced a Brown syndrome and tie the suture.
Fells removes the temporal one-third of the split tendon from its insertion and resutures it 8 mm behind the LR insertion alongside the upper edge of the LR. Holmes and others have popularized the adjustable variation of Fells’ technique, and have also demonstrated significant regression effects with this variation, which this author has not recognized with the original technique 23 (Chapter 26, Video 26.3, Video 26.4).
126.96.36.199.2 Whole Tendon
Whole tendon SO surgery is particularly indicated if the tendon is floppy, and the tendon should be tightened until the forced duction test is similar to the other side or, if bilateral, approximates one’s estimate of normal. If not sure, always do a little less—improving the patient is better than overcorrecting the patient.
There are two main approaches, plicating the tendon nasal to the SR (there is a tendon tucker designed to make this easy) (Chapter 24) or advancing the insertion (my preferred technique) (Chapter 26, Video 26.1, Video 26.2). The advancing technique allows for easier intraoperative adjustment to get to the desired endpoint.
188.8.131.52 Inferior Rectus Muscle Surgery
IR recessions need to be no more than 4 mm to avoid lower lid retraction and may be performed with an adjustable technique (Chapter 32, Video 32.2). IR recession with an absorbable suture is prone to late progressive overcorrection, typically beginning at week 2 to 4 (Chapter 5, Chapter 27). To lessen the risk of slippage, this author uses 5–0 rather than the usual 6–0 Vicryl if adjusting, in the (unproven) hope that this will lessen the risk of late progressive overcorrection. Using a nonabsorbable suture (such as 6–0 braided polyester suture) largely eliminates the risk of progressive overcorrection. Using a nonabsorbable suture usually limits the ability to use an adjustable suture because the exposed part of the nonabsorbable suture causes intolerable irritation for some patients.
The adjustable IR faden operation popularized by Cruz is valuable for cases that measure out as SO underaction (UA) but do not have a floppy SO. 24 This author uses 5–0 Vicryl, adjustable, and has not (yet) seen the late progressive overcorrection that IR recession is prone to. Possibly the resection excites a more aggressive and secure healing response and that reduces the tendency toward progressive overcorrection.
An IR recession will have a small net intorting effect, lessening any preexisting excyclotropia. This torsional change can be lessened by a temporal shift (e.g., draping the IR from the temporal edge of the insertion) or augmented by a nasal shift (Table 11.1).
184.108.40.206 Superior Rectus Muscle Surgery
SR recessions can be done with an adjustable technique (Chapter 32, Video 32.3). To lessen the risk of slippage, this author prefers 5–0 Vicryl if adjusting, and 6–0 Mersilene or Surgidac if a nonadjustable technique is used. (It is not known whether 5–0 Vicryl reduces the risk of slippage versus 6–0 Vicryl).
There is a consistently low rate of unwanted overcorrection. It is not present on day 1 after surgery, can then be seen early or late, and is probably caused by the frenulum between the SR and SO, or the SO tendon itself. 25
An excellent result after suture adjustment on day 1 can be later undone, due to a (presumed) adhesion between the recessed SR and the frenulum giving way. The next day this adhesion may be broken and the SR recessed several millimeters further with a large hypotropia.
If the SR slips in month 2 it is possibly due to the SO tendon interfering with normal scar formation between SR and sclera. During surgery, when the eye is infraducted by the surgeon, the SO tendon is far posterior to any planned recessed SR insertion. When the eye is no longer infraducted, however, the SO tendon is much closer to the original or planned SR insertion.
To avoid or remediate these problems, there is some benefit in using fixed rather than adjustable recession surgery and using nonabsorbable rather than absorbable sutures.
The SR is an intorter. Recession will lessen intorsion with a net excyclo-shift. Recessing from the temporal pole of the insertion will lessen the excyclo-shift, while recessing from the nasal pole will augment the excyclo-shift (Table 11.1).
18.3 Third Cranial Nerve Palsy
Because of its complex anatomy in the brainstem and distally, clinical management of third CN palsy (3NP) is always challenging. One or more of four EOMs can be involved and create complicated incomitant horizontal, vertical, and torsional strabismus. This can further be complicated by aberrant regeneration, ptosis, pupil involvement, lack of accommodation, and deficient Bell’s phenomenon.
In children congenital onset 3NP is the most frequent presentation, followed by traumatic, neoplastic, vascular, presumed migrainous, and parainfectious etiologies. In recent years we have learned that some congenital partial 3NPs are due to subtle genetic defects—the congenital fibrosis of EOMs (CFEOM) group of 3NP. 26
In adults 3NP is more common than in children. In a large study from Minnesota, the most common causes were presumed microvascular causes (42%), trauma (12%), compression from neoplasm (11%), neurosurgery (10%), and compression from aneurysm (6%). 27
Management of 3NP depends upon the degree of involvement (complete or partial), and in partial cases, which muscle or muscles are involved, and to what degree.
18.3.2 Complete (or Nearly Complete) Third Nerve Palsy
All four EOMs innervated by the third CN may be involved, and the only active muscles remaining are the LR and SO muscles. Since both of these two muscles have no active antagonists, the eye remains in exotropia and hypotropia. During attempted infraduction, intorsion can be seen due to SO activity.
In children, this fixed position of the eye, especially if coupled with loss of accommodation and/or ptosis, is usually a cause of amblyopia.
In adults diplopia is the most disturbing clinical sign, sometimes relieved by the accompanying ptosis. Since eye movements are severely restricted, prisms are rarely useful. Another frequent clinical manifestation is aberrant regeneration, seen in children as well as adults. The occurrence of aberrant innervation requires special attention when planning surgery.
220.127.116.11 Nonsurgical Management
Since in many cases it is advised to wait for some recovery of the nerve function after insult, short-term treatment for the first 6 months is aimed at alleviating diplopia and preventing amblyopia in children.
For adults, diplopia can be avoided by occlusion of one eye with a patch or opaque contact lens. In cases of paresis of just one muscle, prisms can sometime be used. For instance, in cases of isolated IR palsy the use of prisms can facilitate reading. Use of botulinum toxin (Botox) is another nonsurgical option in the acute phase of partial 3NP (Chapter 6, Chapter 31, Video 31.1, Video 31.2). This is especially useful in cases of isolated involvement of MR muscle, when injection to the LR can neutralize the horizontal deviation and prevent contracture. Spontaneous recovery of the MR together with recovery of the LR from the toxin can sometimes prevent the need for surgery. Toxin injection to the SR is contraindicated, as it may cause ptosis, but injection to the IR can be done in the rare cases of isolated SR paresis.
In children, the risk of amblyopia is substantial. Occlusion therapy and close follow-up are warranted. In cases of accommodation palsy the use of bifocal glasses is important.
18.104.22.168 Surgical Treatment
Surgical treatment varies in accordance with the severity of paralysis and number of muscles involved. Usually it is advised to delay surgical treatment for a period of at least 12 months after the insult with hope of some spontaneous recovery. If the alignment measurements are getting worse at an earlier time, surgery can be planned earlier.
Three important questions to answer in planning surgery are whether the MR is dead, whether the SO is tight, and whether there is aberrant regeneration.
Is the MR dead?
Forceps testing will usually tell you, but this is a difficult and sometimes terrifying test for a patient, and even in the most experienced hands forceps testing is not 100% reliable unless it is repeatable and unequivocal. Saccadic velocity tests are not reliable if the saccade is small.
If the MR is dead, you need to cripple the LR to get a result that lasts, either by suturing the LR to the adjacent periosteum or by redirecting its insertion to the nasal side of the globe (Video 33.2).
If the MR is not dead, a large resection or plication will work and will usually last.
For recurrent exotropia sometimes the MR insertion can be tethered to a bony point in the medial orbit (such as the posterior lacrimal crest) using fascia lata or tethered with a periosteal flap. 28 , 29
Consider a presumed left 3NP and we wish to know if the left MR is dead (needs transposition, or periosteal suture for the LR) or weak (recess-resect will be effective).
Anesthetize with topical drops until “I didn’t feel that last drop.”. Then soak a cotton swab in anesthetic eyedrops or lidocaine and massage the limbus at the 6 o’clock position with this for 30 seconds.
Grab the limbus at the 6 o’clock position with a pair of fine-toothed forceps and do the following:
Ask the patient to adduct the eye as much as possible, then see if you can forcibly adduct more. This is the forced duction test. This tells you how much of the adduction deficit is due to the tight left LR, and how much to the weak left MR.
In this adducted position, ask the patient to keep looking right. Abduct the eye with the forceps and feel for any resistance to abduction (indicating paresis) or for free movement (indicating palsy).
Ask the patient to abduct the eye as much as possible. In this position, try to generate a saccade to the right. If you can definitely feel a “tug” through the forceps, this is a paresis, not a palsy.