Special Features of Superior Oblique Hypofunction Due to Tendon Abnormalities

Abstract

Purpose:

While most cases of superior oblique (SO) hypofunction represent contractile weakness due to denervation, sometimes the lesion is exclusively in the tendon. This study sought to distinguish the pattern of incomitant strabismus caused by deficiency of SO oculorotary force caused by tendon abnormalities versus neurogenic palsy.

Methods:

Clinical and magnetic resonance imaging (MRI) findings of 7 cases of unilateral SO tendon interruption or extirpation were compared with 11 cases of age matched unilateral SO palsy having intact tendons. We compared angles of misalignment with high-resolution MRI in central and down gazes.

Results:

Muscle bellies in neurogenic palsy were markedly atrophic with maximal cross sections averaging 6.5±2.7 mm², in contrast with 13.5±3.0 mm² contralesionally (P<0.0001). In contrast, SO muscle bellies ipsilateral to tendon interruption had maximum cross sections averaging 15.1±3.0 mm² occurring more posterior than on the contralesional side whose maximum averaged 12.1±2.4 mm². While cross sections of SO bellies ipsilateral to tendon interruption exhibited normal contractile increase in infraduction (P<0.0005), there was nevertheless strabismus with incomitance similar to that in SO atrophy. Binocular alignment was statistically similar (P> 0.5) in the two groups for all diagnostic positions, including head tilt, except in deorsumversion, where cases with SO tendon abnormalities averaged 20.5±6.9Δ ipsilateral hypertropia, significantly more than 8.5±6.6Δ in neurogenic SO atrophy (P = 0.001). The average difference in hypertropia between central gaze and infraversion was 9Δ in cases of tendon abnormalities, significantly greater than −4.1Δ in SO atrophy (P = 0.019). In contralesional version, average overelevation in adduction was 1.7 (scale of 0 – 4) in tendon abnormalities, and 2.6 in SO atrophy (p = 0.23), while average underdepression in adduction was −2.3 in cases of tendon abnormalities and −1.6 in SO atrophy (p = 0.82). Repair of the SO tendon in three cases was effective, while alternative procedures were performed when repair was infeasible.

Conclusion:

While both denervation and tendon interruption impair SO oculorotary function, interruption causes greater hypertropia in infraversion. Surgical tightening of interrupted SO tendons may have particularly gratifying effects. Posterior SO thickening and large hypertropia in infraversion suggest SO tendon interruption that may guide a surgical strategy of tendon repair.

Introduction

Abundant evidence supports the inference that acute, unilateral superior oblique (SO) muscle paralysis causes ipsilateral hypertropia that increases in contralateral gaze and with head tilt towards the ipsilateral shoulder^1–3. The foregoing three-step test has been the foundation of diagnosis and classification of cyclovertical strabismus for many decades³. Clinicians thus often infer SO hypofunction when the three-step test is an any degree positive, and excuse the large variation in inter-individual alignment to secondary changes^{4, 5} such as inferior oblique (IO) overaction and superior rectus (SR) contracture. However, recent studies using magnetic resonance imaging (MRI) employing neurogenic atrophy of the SO belly as a gold standard for actual deficit of SO force generation have shown only 70% sensitivity and the is 50% specificity for this test^{6, 7}. Intraoperative force measurement during strabismus surgery has in one case demonstrated contractile force generation in the SO despite a motility pattern consistent with SO palsy⁸. Upshoot and downshoot in adduction can be induced in normal people merely by monocular occlusion for a few days, and so do not specifically imply extraocular muscle paralysis⁹. Computer simulation indicates that SO deficiency alone cannot account for the magnitude of the large variations in hypertropia typically associated with a positive three-step test^{10, 11}. On the other hand, numerous alternative conditions can mimic the pattern of incomitant hypertropia classically associated with SO paralysis^{12, 13}. For example, skew deviation can mimic SO palsy on the three-step test¹⁴, as can heterotopy of rectus pulleys¹⁵.

Abnormalities of the SO tendon can attenuate application of otherwise normal SO contractile force to the globe, and thus mimic oculorotary deficiency of the SO muscle. While SO muscle atrophy should logically also be associated with laxity of its tendon, as much as 18% of clinically diagnosed congenital SO palsy may be associated with absence of SO tendon altogether¹⁶.

Ongoing advances in magnetic resonance imaging (MRI) have enabled direct investigation of the functional anatomy of the SO muscle. Contractility of the SO can be radiographically determined by evaluating the change in SO cross-sectional area during gaze shift from supraduction to infraduction^{4, 17, 18}. Intracranial trochlear nerve neurectomy in monkeys has been shown to rapidly produce significant neurogenic atrophy of SO readily demonstrable by MRI¹⁹. This current study was founded on the concept that SO palsy due to neurogenic atrophy can be objectively verified by the MRI finding of SO belly atrophy, but that no such atrophy will be associated with deficient SO function caused by transection of the SO tendon without direct injury to the belly. The study aimed to evaluate patterns of binocular misalignment in central gaze, lateral gaze, vertical gaze, and head tilt, in patients who have head-tilt dependent hypertropia associated with discontinuity of the SO tendon but without SO muscle atrophy, versus patients who have SO muscle atrophy without tendon abnormality. The latter group represents classical cases of neurogenic SO palsy.

METHODS

All patients gave prior written informed consent according to a protocol approved by the Institutional Review Board of the Redacted Institution, and in conformity with the US Health Information Portability and Accountability Act of 1996 and Declaration of Helsinki. Subjects were participants in an ongoing MRI study of strabismus by conducted the same principal investigator at the same institution from 1990 to 2019. This study includes both surface coil MRI and detailed clinical data, including clinical imaging of ocular motility in diagnostic positions and head tilts, alignment in diagnostic gaze positions by prism-cover testing, Hess screen testing where appropriate, and operative records if surgery was performed. From this database were selected 7 patients who had SO hypofunction due to tendon abnormalities, but who did not exhibit MRI evidence of SO atrophy or reduced contractility. A comparison group was selected of 11 age-matched patients in whom MRI demonstrated unilateral neurogenic SO atrophy, defined as maximum SO cross sectional area less than the 2.5th percentile of normal18. In the two groups, we tabulated evaluated symptoms, torticollis, heterotropia in diagnostic positions and head tilts, and fundus torsion.

All MRI in the database was personally performed by the same investigator using a 1.5 T General Electric Signa (Milwaukee, WI) scanner and T1, or T2 fast spin echo pulse sequences, and a surface coil array (Medical Advances, Milwaukee, WI) with fiber optic fixation target^{17, 20}. High-resolution (312 μm), 2-mm slice thickness quasi-coronal images and 256×256 matrix parallel to the long orbital axis were obtained for each eye in a target-controlled central gaze, supraduction, and infraduction.

Image analysis was by published methods using the public domain program ImageJ (Rasband, W. S. ImageJ, U.S. National Health Institutes, Bethesda, MD; http:/rsb.info.nih.gov/ij/, 1997–2009, accessed February 2009)²¹.

Results

All 18 patients with SO hypofunction complained of vertical or oblique, binocular diplopia and exhibited abnormal head posture. All had normal ocular anterior and posterior ocular segments. Patients were divided into two groups: 7 cases (5 males, and 2 females, range 5–72 years) had SO tendon abnormalities; 11 cases (7 males and 4 females, age range 2–80 years) exhibited significant SO muscle atrophy without tendon abnormalities.

Tendon abnormalities varied. Case 1 had undergone SO tendon extirpation during excision of an orbital lymphoma involving the trochlea (Fig. 1). Case 2 had SO tendon interruption as a surgical complication of blepharoplasty (Fig. 2). Cases 3 – 5 sustained direct facial blunt trauma. Case 6 sustained face and head trauma due to motor vehicle accident with loss of consciousness. Case 7 was due to congenitally abnormal SO tendon insertion.

Fig. 1.

Coronal T1-weighted magnetic resonance image demonstrating extirpation of left trochlea and superior oblique (SO) tendon during biopsy of orbital lymphoma involving the trochlea. IO – inferior oblique muscle. LG – orbital lobe of lacrimal gland. MR – medial rectus muscle.

Fig. 2.

Coronal T2-weighted magnetic resonance images demonstrating left superior oblique tendon interruption complicating blepharoplasty surgery. SO – superior oblique.

Causes of neurogenic SO were also varied. Eight cases had histories suggesting that they were congenital. One case was involved severe closed head trauma, one case resulted from failure to recover from microvascular nerve palsy, and one case was idiopathic. Figure 3 illustrates a case of unilateral SO atrophy with cross section markedly smaller than in the unaffected fellow eye.

Fig. 3.

Neurogenic right unilateral superior oblique (SO) atrophy with smaller cross section than in the unaffected left orbit.

It is possible to track the SO muscle from its origin to trochlea in MRI views. The mean cross sectional area of both SO muscles along the length of the orbit is plotted for cases of unilateral tendon abnormalities in Fig. 4 The plot in Fig. 4 shows that the SO muscle tended to be thicker and have a more posterior maximum cross section than normal. While maximum SO cross section was larger a 15.1±3.0 mm² ipsilesional to its abnormal tendon, SO cross section was not significantly larger than contralesional SO cross sectional area at 12.1±2.4 mm² that occurred 2 mm posterior to the globe-optic nerve junction (P = 0.34). Maximum SO cross sectional area ipsilesional to the abnormal tendon occurred 4 mm posterior to the globe-optic nerve junction, 2 mm posterior to the point of contralesional maximum cross section.

Fig. 4.

Mean superior oblique (SO) cross-sectional area in central gaze for 7 cases with unilateral tendon lesion. Cross-sections are plotted as functions of 2-mm thickness image plane number, referenced to image plane 0 at the globe–optic nerve junction in central gaze. ANOVA – analysis of variance.

Figure 5 illustrates mean cross sectional areas of both SO muscles in central gaze for unilateral neurogenic SO palsy. As expected, ipsilesional maximum SO cross sectional area was 6.5±2.7 mm² at 2 mm posterior to the globe-optic nerve junction. Maximum contralesional SO cross sectional area of 13.5±3.0 mm² occurred 6 mm posterior to the globe-optic nerve junction and was significantly greater than maximum ipsilesional SO cross section observed 0 – 4 mm posterior to the globe-optic nerve junction (P <0.050).

Fig. 5.

Mean superior oblique (SO) cross-sectional area in central gaze for 11 cases of unilateral neurogenic SO palsy. Cross-sections are plotted as functions of 2-mm thickness image plane number, referenced to image plane 0 at the globe–optic nerve junction in central gaze. ANOVA – analysis of variance.

Figure 6 shows SO contractility ipsilesional to damaged tendons, palsied SO muscles, and normal fellow muscles, by plotting the difference in SO cross sectional area from contraction in infraversion, minus relaxation in sursumversion. As measured 2 mm posterior to the globe-optic nerve junction, the palsied SO muscle was hypocontractile compared with both the normal contralateral muscle (P < 0.05), and with the muscle ipsilateral to tendon damage (P < 0.001). Ipsilesional to the damaged SO tendon, muscle contractility was comparable that of the normal fellow eye (P >0.15).

Fig. 6.

Mean contractile change in superior oblique (SO) cross-sectional area from infraduction to supraduction for 7 cases of unilateral tendon lesion, and 11 cases of unilateral SO palsy.

A. Contractility of palsied SO was significantly subnormal to that of the contralateral SO 8 and 10 mm posterior to the globe.

B. Contractility of SO muscles with damaged tendons did not differ significantly from that of normal fellow SO muscles where contralateral SO had neurogenic atrophy.

C. Contractility of SO with damaged tendon was greater than that of palsied SO.

D. Contractility of SO contralateral to palsy was similar to that of intact SO contralateral to damaged SO tendon.

Binocular alignment in diagnostic positions is listed in Table 1, and was statistically similar (P> 0.5) in the two groups for all comparisons except those involving deorsumversion. In deorsumversion, cases with SO tendon abnormalities averaged 20.5±6.9Δ ipsilateral hypertropia, significantly more than 8.5±6.6Δ in patients with neurogenic SO atrophy (P = 0.001). The average difference in hypertropia between central gaze and infraversion was 9Δ in cases of tendon abnormalities, significantly greater than −4.1Δ in SO atrophy (P = 0.019). In contralesional version, the average overelevation in adduction was 1.7 (scale of 0 – 4) in cases of tendon abnormalities, and 2.6 in SO atrophy (p = 0.23), while average underdepression in adduction was −2.3 in cases of tendon abnormalities and −1.6 in SO atrophy (p = 0.82).

Table 1.

Ipsilesional Hypertropia in Unilateral Superior Oblique Hypofunction, Δ

Gaze Position	Central	Contralesional	Ipsilesional	Contralesional Head Tilt	Ipsilesional Head Tilt	Sursumversion	Deorsumversion
SO Tendon Abnormality
Mean	11.5	24.3	5.4	4.7	17.9	9.7	20.5
Standard Error	9	13.2	4.2	5.5	11.6	16.5	6.9
Neurogenic SO Palsy
Mean	12.6	20.2	6.6	3.4	19.1	6.9	8.5
Standard Error	11	14.1	7.7	6.2	13	5.1	6.6

SO – superior oblique.

Surgical treatment was aimed at repairing the SO tendon when it was abnormal. Table 2 summarizes cases with tendon abnormalities and procedures used for correction. The damaged SO tendon was repaired by plication (tucking) in four cases. Four patients underwent inferior oblique weakening procedures by myectomy or recession. Three patients underwent contralateral inferior rectus recession. Results were considered successful if post-operative hypertropia was less than 3Δ in central gaze, and less than 4Δ in deorsumversion. Surgical correction was successful for central gaze in all cases of tendon abnormalities, but was unsuccessful in infraversion in one patient. In this patient, ipsilateral hypertropia nevertheless decreased markedly from 30Δ to 12Δ.

Table 2.

Superior Oblique Tendon Abnormality

Case	Preoperative Hypertropia		Postoperative Hypertropia		Surgery
	Central	Deorsumversion	Central	Deorsumversion	First	Second
1	16	20	0	0	Left IO myectomy and right IR recession 3.5mm	2mm right IR Re-recession with temporal transposition of ¾ tendon width; 4 mm left IR resection with ¾ tendon width nasal transposition
2	6	25	0	4	Left IO myectomy.	2.5mm right IR recession with scleral posterior fixation 18mm posterior to limbus.
3	5	20	0	3	8mm plication of posterior half of right SO tendon
4	2	20	0	0	Right IO recession
5	20	30	0	12	6 mm right SO plication, and 2mm left IR recession with scleral posterior fixation
6	20	8	2	2	Right IO recession, and right SO tendon plication 7mm
7	25	25	2	2	7mm left SO plication, and 4mm right IR recession using adjustable suture.

IO – inferior oblique muscle. IR – inferior rectus muscle. SO – superior oblique.

Discussion

This study employed MRI and clinical measurements to distinguish unilateral SO hypofunction due entirely to tendon abnormality, from unilateral hypofunction due to neurogenic atrophy without anatomical tendon interruption or disinsertion. In standard clinical diagnostic positions and head tilt, the ipsilesional hypertropia varied in a manner that was statistically indistinguishable between groups for the most part. The only clinical strabismus measurement that did distinguish the two groups was the amount of ipsilesional hypertropia in deorsumversion, which averaged more than twice as great at 20Δ in SO tendon abnormality as the 8Δ observed in neurogenic SO atrophy.

This research supports and extends an earlier report that the classic three-step clinical test is not unique to SO weakness caused by muscle atrophy that is demonstrable by high-resolution MRI¹². Patients with either SO tendon discontinuity or neurogenic SO muscle atrophy had similar versions, and similar variation in gaze direction and head tilt variation of hypertropia. Importantly, contractility and size of the SO muscle belly was normal in cases of tendon abnormality, while the SO belly was atrophic in neurogenic atrophy. Maximum SO cross-sectional area tended to be slightly larger but more posterior than on the normal, contralesional side in SO tendon abnormality. This trend might be attributed to stretching or re-distribution of muscle fibers, but was statistically insignificant and would therefore not be a reliable clinical clue.

Since experimental trochlear neurectomy in monkeys causes rapid SO atrophy that reduced maximum cross section by about 50 percent like the reduction observed in subjects with SO atrophy here¹⁹, it was reasonable to consider the current group of patients with unilateral SO atrophy as having “actual” neurogenic SO palsy consistent with trochlear denervation. Reduced SO cross section is strongly correlated with reduced cross-sectional area change from sursumduction to deorsumduction, and thus tends to correlate with reduced contractility^{4, 17}. It is likely that surgical observation would have found the tendons “floppy” when in continuity with lax, atrophic SO bellies, and has been reported for cases of SO atrophy confirmed by MRI in Japan²².

On the other hand, the current patients with SO tendon abnormality had affected SO cross sections of at least normal size, and normal contractility as measured by cross-sectional area change from sursumduction to deorsumduction. We regard these cases as having normal SO contractility but failing to transmit SO force to the globe due to tendon abnormality. The detached or absent SO tendon may permit the SO muscle belly to recoil posteriorly, consequently tending to shift the maximum SO cross section farther posterior than normal. Disinserted or traumatically stretched SO tendons that could be observed during surgery were also “floppy, because the tendons were not attached to stiff SO muscle bellies.

The current findings imply that surgical detection of SO laxity^{23 24–28} cannot distinguish SO tendon defect from neurogenic SO atrophy. Forced duction testing of the SO is often subjective, and its clinical reliability has been questioned because it has been insensitive even to profound SO atrophy shown by MR⁶. A prospective study of SO traction testing stated only fair agreement by masked participating clinicians, low correlation between SO tendon laxity and imaging proof of SO atrophy, and frequent observation of SO laxity in clinical situation where the SO palsy was not a consideration²⁶. The torsional forced duction test recently suggested by Jung and Holmes has demonstrated abnormally great maximal excyclorotation during intraoperative SO disinsertion in two cases of SO palsy, but not prior to disinsertion in multiple cases of congenital and acquired SO palsy²⁷. Distinction between neurogenic SO muscle atrophy from tendon damage requires orbital imaging to ascertain SO belly size, recognizing that even high resolution, surface coil MRI with target fixation lacks the resolution required to characterize the state of the thin, reflected SO tendon itself in most cases. However, MRI is reliably able to directly demonstrate pathology of the SO belly.

Despite case accumulation over more than two decades using consistent technique, sample size of this single institution study was relatively small. Some of the details of accidental acquired SO tendon traumas are rendered uncertain by the rapid sequence of traumatic events, and occasional associated loss of consciousness. Nevertheless, the objective MRI findings of this study permit clear and objective definition of pathological changes in the SO belly.

In clinical circumstances involving trauma or surgery near the trochlea, it is clinically appropriate to consider SO tendon abnormalities in patients who have hypertropia that increases in contralateral gaze and ipsilateral head tilt, and is particularly large in deorsumversion. In such cases imaging evidence showing absence of SO atrophy might be considered among the indications to conduct careful surgical exploration of the entire course of the reflected SO tendon. The complex reflected SO tendon and insertion region are nevertheless accessible to surgical repair, particularly if the surgeon anticipates finding pathology there, and such surgical repair sometimes has potential for near-normal restoration of infraduction and incycloduction in a physiologic manner that is likely to be superior to alternatives such as contralateral inferior rectus weakening.