Split-Tendon Medial Transposition of Lateral Rectus for Pediatric Complete Oculomotor Palsy

Abstract

Background: 

Split-tendon medial transposition of lateral rectus (STMTLR) for complete oculomotor palsy can correct large angles of exotropia in adults, but outcomes are variable, and complications are frequent. Only a few pediatric cases have been reported, and further insight is needed to assess the child's alignment outcomes and ability for postsurgical gain of function. The aim of our study is to report the outcomes of this surgical procedure in pediatric cases of complete oculomotor palsy.

Methods: 

A retrospective review of outcomes was conducted on 5 consecutive patients with complete oculomotor palsy treated with STMTLR by a single surgeon (V.S.S.) between 2015 and 2021 at tertiary referral centers. Primary outcome was postoperative horizontal alignment, and secondary outcome was demonstration of gain-of-function activity in the field of action of the paretic medial rectus muscle.

Results: 

Five cases of pediatric complete oculomotor palsy underwent surgical treatment with STMTLR. Subjects averaged 5.3 years old (range 10 months-16 years). Two were female. Etiologies were heterogeneous, and all presented with unilateral (n = 2) or bilateral complete oculomotor palsy with exodeviations ranging from 45 to >120 prism diopters. Two subjects had bilateral disease secondary to military tuberculosis with CNS involvement. A third subject presented iatrogenically with complete bilateral third nerve palsies secondary to removal of a nongerminomatous germ cell tumor (NGGCT) of the pineal gland. The 2 remaining subjects had monocular involvement in their right eye, 1 from compressive neuropathy after a cavernoma midbrain hemorrhage, and 1 from a congenital right oculomotor palsy. All patients were observed to have stable ocular alignment for a period of at least 6 months before surgery. Unilateral STMTLR was performed in all cases except the subject with NGGCT, in which bilateral STMTLR was performed. Measurement of alignment permanence out to 1–3 years postop resulted in an average correction of 40.83 prism diopters (range 37.5–45 prism diopters) per operated eye. Four of 5 subjects regained limited but active adduction eye movements, and the 2 unilateral cases demonstrated improved convergence. None of the subjects experienced significant complications.

Conclusions: 

STMTLR was a safe and effective approach for the surgical correction of complete pediatric oculomotor palsy in our case series. In addition, pediatric patients may benefit from STMTLR with immediate gain-of-function activity in the transposed lateral rectus muscle, which supports the hypothesis that children have a dynamic and adaptive neuroplasticity of visual target selection that predominates established agonist/antagonist neural signaling.

Third cranial nerve palsies in children present a unique and difficult challenge to the ophthalmic surgeon. It occurs rarely in the pediatric population, with an approximate incidence rate of 1.7 per 100,000 children (¹). The etiologies of third nerve palsy in children are heterogeneous, with congenital causes representing almost half (43%) of all cases (²). In adults, however, acquired causes of third nerve palsy predominate, with diabetes, hypertension, vascular aneurysm, and trauma being some of the commonest etiologies.

A third nerve palsy can be classified as partial or complete, determined by the extent of extraocular muscle paralysis. In a previous study of 49 children (53 affected eyes) with diagnosed third nerve palsy, 31 children (32 eyes) were partial, whereas 18 children (21 eyes) were complete (³). A complete third nerve palsy in a pediatric patient poses an especially difficult management challenge because the child's developing visual system is vulnerable to strabismic, anisometropic, or deprivation amblyopia while uncorrected. Thus, surgical management in children should be directed toward restoring fusion and stereoacuity, in addition to achieving orthotropia and eliminating diplopia.

Surgical management of complete third nerve palsies resists conventional maximally dosed recession–resection procedures (⁴). This is primarily because the completely paretic medial rectus cannot benefit from any degree of muscle resection. Any adduction function is therefore impossible without the transposition of a functioning muscle, of which only the lateral rectus and superior oblique remain. Accordingly, alternative techniques such as the transposition of the superior oblique with lateral rectus recession (⁵) and medial transposition of the lateral rectus muscle (⁶) have demonstrated variable success. Herbert Kaufmann pioneered a novel technique where the lateral rectus is split and transposed nasally (Kaufmann, 1991), which was later modified to have the split halves reattached 1 mm posterior to the superior and inferior borders of the medial rectus insertion site (⁷). Further refinements of the split-tendon medial transposition of lateral rectus (STMTLR) technique improved ocular alignment and force augmentation (^8–10) and achieved sustained efficacy in primary position alignment as demonstrated through various cohort studies (¹¹). However, most subjects followed in these studies were adult patients.

Given the major differences in management strategy for pediatric vs adult third nerve palsy, and the unique vulnerabilities and opportunities of a developing visual system, our study sought to describe outcomes in a retrospective pediatric cohort that underwent STMTLR at a tertiary children's hospital. We then review and compare our outcomes with the few pediatric cases collated from the current literature.

METHODS

This prospective cohort study was conducted at Texas Children's Hospital, Houston, Texas, USA, between 2015 and 2019. This study was reviewed and approved by the Baylor College of Medicine Institutional Review Board (H-38264). One additional case (#3) was performed at Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, USA, in 2020–2021. Consent was obtained for all patients presented.

A total of 5 participants with a primary diagnosis of complete third nerve palsy were included in this study and underwent STMTLR. All participants must have demonstrated a stable ocular deviation angle as measured across 2 or more consecutive visits spanning at least 6 months. Exclusion criteria included partial third nerve palsy, thyroid orbitopathy, ocular myasthenia gravis, previous strabismus surgery, and any other mechanical cause of ophthalmoplegia (e.g., muscle entrapment, fibrosis, and orbital wall fracture). Patients with signs of aberrant regeneration/oculomotor synkinesis were also excluded from the study.

Before surgery, all patients underwent a comprehensive ophthalmological examination, which included manifest refraction, slit-lamp biomicroscopy, and dilated fundus examination. The strabismus angle in primary gaze position was measured using Krimsky tests. Primary deviation was measured in all cases. In patients with bilateral disease, secondary deviation on fixation with the affected eye was noted but not compared. All strabismus angles were measured in prism diopters (Δ).

All patients in the study underwent the same procedure and were operated on by the same surgeon (V.S.S.) under general anesthesia. The operative eye was prepped and draped, and a conjunctival limbal peritomy was made in the lateral quadrant. The inferior oblique was identified and isolated with a nylon tie. The lateral rectus muscle was then hooked, and the muscle belly was divided using a small muscle hook into a superior (sLR) and inferior (iLR) division from the insertion site to a distance 13–15 mm posteriorly. A double-armed locking suture (6-0 Vicryl; Ethicon) was placed at the insertion end of each lateral rectus half-muscle. Both lateral rectus half-muscles were then disinserted. Next, the limbal peritomy was extended to the inferior quadrant. The iLR was then passed between the sclera and both the isolated inferior oblique and inferior rectus muscles and nasalized to the inferior border of the medial rectus insertion site. The limbal peritomy was then extended to the superior quadrant. The sLR was passed between the superior rectus and superior oblique tendons and the sclera and advanced anteriorly to the superior border of the medial rectus insertion site. Both lateral rectus halves were reattached to the sclera approximately 4.0 mm posterior and 2.0 mm superior/inferior to the respective poles of the medial rectus insertion. On reattachment, the locking sutures were trimmed, and the limbal peritomy was repaired with a 6-0 Vicryl suture. Stepwise surgical photographs of the STMTLR procedures are provided (Fig. 1).

FIG. 1.: Surgical steps to split-tendon medial transposition of lateral rectus. A. The lateral rectus and inferior oblique muscles are isolated and tied. B. The lateral rectus is hooked. C. The belly of the lateral rectus is split with a small muscle hook. D. The split belly is extended up to 15 mm posteriorly, leaving superior and inferior divisions of the lateral rectus. E. After imbricating both insertion ends of the lateral rectus divisions with a double-armed 6.0 Vicryl suture, the muscle divisions were disinserted. F. The inferior division of the lateral rectus is passed between the sclera and the inferior oblique and (G) inferior rectus muscles. H. The superior rectus is isolated and hooked. I. The superior division of the lateral rectus is passed between the sclera and the superior rectus and superior oblique muscles. J. The medial rectus muscle is isolated and hooked. K, L. The superior division of the lateral rectus is fixed to a new insertion 4.0 mm posterior and 2.0 mm superior to the superior pole of the medial rectus insertion site. M. The inferior division of the lateral rectus is fixed to a new insertion 4.0 mm posterior and 2.0 mm inferior to the inferior pole of the medial rectus insertion site. N. The limbal peritomy was repaired with 6-0 Vicryl suture. O. Patient ocular alignment in primary gaze on postoperative Day 1. iLR indicates inferior division of lateral rectus; IO, inferior oblique; IR, inferior rectus; LR, lateral rectus; MR, medial rectus; sLR, superior division of lateral rectus; SR, superior rectus.

In one bilateral subject (Case 3), STMTLR was performed on both eyes approximately 2 months apart. The primary rationale was the significantly large strabismus angle in primary gaze (+120Δ). The patient tolerated both procedures well without complications.

Study participants were re-examined at postoperative week 1, 3 months, 1 year, and 3 years when available. Postoperative horizontal deviation (Δ) was measured in the operative eye at distance and followed in subsequent visits. The primary outcome was the mean change in postoperative horizontal strabismus angle (Δ) at primary gaze from baseline per procedure. The secondary outcome was the reacquisition of ocular motility function, which was assessed by adduction of the postoperative eye to a near target and scored as present or absent. All data collection and statistical analyses were performed in MATLAB version 9.10.0 (R2021a) (Mathworks, Natick, MA).

RESULTS

A total of 6 eyes from 5 patients (3 male and 2 female) were operated on in this study. The mean age of the study population was 5.3 years (range 10 months–16 years). Comprehensive preoperative details for all subjects, including demographic information, are included in Table 1.

TABLE 1. - Patient demographics and summary of preoperative orthoptic status

Case No.	Demographic				BCVA		Preoperative Assessment
	Age	Gender	Etiology	Laterality	OD	OS	Horizontal Deviation, Δ	EOM Abduction*	EOM Adduction*	Forced Duction Abduction	Forced Duction Adduction	Ptosis
1	20 mo	F	Acquired—infectious	Bilateral	FF	FF	XT 85	0	−4	Neg	Neg	Present
2	10 mo	M	x—infectious	Bilateral	FF	FF	XT 90	0	−4	Neg	Neg	Present
3 †	16 yr	M	Acquired—iatrogenic	Bilateral	2/40	20/30	XT 120+	0	−4	Neg	Neg	Present
4	5 yr	F	Acquired—ischemic	Unilateral (Right)	20/40	20/30	XT 45	0	−4	Neg	Neg	Present
5	3 yr	M	Congenital	Unilateral (Right)	20/130 (Teller)	20/63 (Teller)	XT 45	0	−4	Neg	Neg	Present

^*Preoperative abduction and adduction on paralytic eye.

^†Case 5 underwent split-tendon medial transposition of lateral rectus surgery on both eyes.

BCVA, best-corrected Snellen visual acuity; EOM, extraocular movements; FF, fixate-follow; XT, exotropia.

All subjects presented initially with complete oculomotor palsy, classified by total lack of ocular adductor, supraduction, and infraduction function, presence of ptosis, and mydriasis. Two subjects (Cases 1 and 2) developed bilateral complete oculomotor palsy as a complication of tuberculous meningitis. One subject (Case 3) presented with bilateral disease secondary to a complicated resection of a nongerminomatous germ cell tumor in the pineal region. One subject (Case 4) had a remote history of a ruptured right-sided capillary malformation hemangioma at 7 months of age, which subsequently caused a longstanding unilateral right oculomotor palsy with gradually worsening moderate angle exotropia and ptosis despite patching. Finally, 1 subject (Case 5) presented with a right complete oculomotor palsy that was determined to be congenital in nature after unremarkable neuroimaging studies.

The mean preoperative horizontal deviation for all subjects was 77 ± 32.13Δ (range 45–120Δ), of which the mean for bilateral subjects (n = 3) was 98.33 ± 18.93Δ and the mean for unilateral subjects (n = 2) was 45 ± 0Δ. The mean postoperative horizontal deviation at 1-week follow-up for all subjects was 26 ± 23.89Δ (range 0–47.5Δ). The overall difference between preoperative and postoperative exodeviations was statistically significant (P < 0.05, Fig. 2A, B). The average strabismus angle correction achieved per STMTLR procedure was 40.83 ± 3.42Δ (range 37.5–45Δ) when measured out to 1–3 years postop. All 5 subjects tolerated the procedure without complications. Four of the 5 subjects regained limited adduction ocular motilities in the postoperative period. An example of the gain-of-function in adduction for Case 4 is provided (See Supplemental Digital Content 1, Video S1, https://links.lww.com/WNO/A641). The 1 subject who failed to reacquire adduction (Case 2) eventually developed cerebral visual impairment and was unable to fixate on a visual target in the months after his strabismus procedure. The remaining subjects (Cases 1, 3, 4, and 5) retained their convergence function at the 3-month and 1-year follow-up. The mean follow-up period (per eye) was 22.5 ± 10.5 months (range 12–36 months). Postoperative exodeviations at each follow-up interval, along with reacquisition of convergence, are detailed for each subject in Table 2.

FIG. 2.: Comparison of preoperative and postoperative horizontal deviation. A. Bar graph representing mean preoperative and postoperative horizontal strabismus angle in prism diopters (Δ) across all 5 patients. Error bars represent SE. P value of 0.0286 was derived from a 2-tailed Student t test statistic. B. Per-patient change in exodeviation before and after surgery. C. Preoperative clinical photograph of a successfully treated patient with STMTLR (Case 5) in primary gaze. D. Postoperative clinical photograph 1 day after right STMTLR surgery showing orthotropia in primary gaze. E. Five-gaze ocular motility photographs of the patient 3 years after right STMTLR surgery. STMTLR indicates split-tendon medial transposition of lateral rectus.

TABLE 2. - Postoperative results of split-tendon medial transposition of lateral rectus per procedure and longitudinal follow-up

Case No.	Surgery	Postoperative Horizontal Deviation, Δ				Initial Change From Baseline, Δ	Adduction Function	Total Follow-up (mo)
		1st Week	3 Months	1 Year	3 Years
1	R STMTLR	45–50	45–50	45	—	40	OD regained [−2]	25
2*	R STMTLR	40–45	50	50	—	40	Unable	13
3	R STMTLR L STMTLR	x 50	UA UA	45 45	— —	75 (37.5 per procedure)	OD regained [−3] OS regained [−3]	13 12
4	R STMTLR	Ortho	LTF	LTF	Ortho	45	OD regained [−3]	36
5	R STMTLR	Ortho	Ortho	Ortho	Ortho—(XT) flick	45	OD regained [−1.5]	36

^*Developed CVI.

CVI, cerebral visual impairment; LTF, lost to follow-up; OD, right eye; OS, left eye; STMTLR, split-tendon medial transposition of lateral rectus.

External photographs of a successful case (Case 5) document profound right exotropia before STMTLR (Fig. 2C). One week after surgery, the patient was orthotropic in primary gaze and reacquired convergence that was still demonstrable at the 3-year follow-up visit (Fig. 2D). Full assessment of extraocular movements still revealed restricted supraduction and infraduction (Fig. 2E). This patient experienced gradual improvement in convergence function out to 3 years postop (See Supplemental Digital Content 1, Video S1, https://links.lww.com/WNO/A642).

DISCUSSION

Descriptions of split-tendon medial transposition of the lateral rectus muscle to treat strabismus from oculomotor palsy has been gradually increasing in the literature. Notably, Shah et al (¹¹) found STMTLR to carry a 69% (68/98) success rate, defined as surgically achieving <15Δ postoperative horizontal deviation. In addition, 83% (73/87) patients were satisfied with their postoperative alignment, and 34% (30/87) of patients successfully demonstrated postoperative binocular fusion. However, these studies and previous case reports feature patient populations that are predominantly adult and not pediatric.

Surgical outcomes from pediatric cases reported in the literature are detailed in Table 3. Gokyigit, Aygit, and Sukhija all describe similar outcomes in mean horizontal correction across adult and pediatric patients (^7,12,13). However, Shah, Erbagci, and Saxena uncovered significantly different degrees of horizontal correction between pediatric and adult patients, especially when followed beyond the immediate postoperative period (^8,10,14).

TABLE 3. - Published outcomes from split-tendon medial transposition of lateral rectus with mean reduction in horizontal alignment in pediatric patients

Author	Journal	Year	Region	Total No. Cases	No. of Pediatric Cases	Mean Horizontal Correction, Δ (Adult)	Mean Horizontal Correction, Δ (Pediatric)
Gokyigit et al (7)	JAAPOS	2013	Turkey	10	2	59 ± 12.91	52.5 ± 10.61
Sukhija et al (13)	JAAPOS	2014	India	3	1	50	50
Shah	JAMA Ophthalmol	2014	Boston	4	3	40	71 ± 55.43
Demer	JNO	2015	UCLA	1	1	—	60
Erbagci	JPOS	2016	Turkey	6	3	78.33 ± 16.07	56.67 ± 5.77
Saxena	BJO	2016	India	3	1	45 ± 9.90	80
Aygit	Int Ophthalmol	2019	Turkey	8	1	42.57 ± 2.51	40
Basiakos	Graefe's	2019	Germany	29	Unknown	43.6 ± 14.8	Unknown
Saxena	JAAPOS	2020	India	4	2	93 ± 4.24	77 ± 12.73

Previous procedures for oculomotor palsy have seen a spectrum of events arise postoperatively, ranging from minor undercorrections and overcorrections (¹⁵) and lid swelling (¹⁶) to choroidal effusion and central serous chorioretinopathy (¹⁷). In comparison, STMTLR has lower complication rates in the management of complete oculomotor palsy; however, this procedure still has vision-threatening complications including choroidal effusions, subretinal fluid, and serous retinal detachments have still been reported (^4,11,18). Despite this being a large amplitude transposition surgery, our case cohort reported no adverse postoperative complications from STMTLR, and all patients underwent uneventful recovery. The lack of oculoemetic sequelae (nausea/vomiting) may be due to the previously tight lateral rectus being more amenable to stretch after transposition or the stretch receptors themselves having a dampened response to manipulation. The transposed lateral rectus is also the smallest rectus muscle and, with a split belly, may contribute less to visceral activation.

The variability of longitudinal postoperative primary alignment outcomes seen in adult vs pediatric may be due to plasticity of the developing brain during childhood. Neuroplasticity is much greater in early development, as exemplified by management of pediatric amblyopia, and has been closely investigated in both infants and monkeys as a contributor to alignment instability, regression, and recovery (^19,20). Evidence of adduction reacquisition in the majority of our pediatric patients (4/5) postsurgically both immediately and longitudinally represents gain-of-function activity of a traditionally antagonist transposed muscle that exemplifies dynamic and adaptive plasticity of multiple higher-order neural systems responding to vision target-specific afferent information. These clinical observations may suggest a critical period during early childhood where transposition surgery may best leverage the plasticity of the developing brain to supplant traditional neural agonist/antagonist extraocular function.

STMTLR for complete oculomotor palsy in children remains an attractive surgical option for the experienced strabismus surgeon. Our study outcomes concur with previous adult cases in the literature with respect to the degree of correction and postoperative ocular alignment, but encountered less complications. We also show that STMTLR performed in children has the opportunity to promote partial adduction and gain-of-function activity because of neurodevelopmental plasticity. These findings support the notion that other domains of improvement besides ocular alignment should be investigated in pediatric strabismus surgery.

Limitations

Some limitations inherent in our study include the retrospective design, the lack of a control group, and the small sample size, as well as the inconsistency of follow-up in some of the patients (because of the Coronavirus 2019 pandemic). In our cohort, 1 patient (Case 2) developed cerebral visual impairment as a late sequelae of complicated central nervous system meningitis, abolishing efforts to assess for potential gain-of-function activity of the transposed lateral rectus muscle. Our patients also had various causes for their oculomotor palsy, which may differentially interfere with alignment permanence during and beyond the postoperative period.

STATEMENT OF AUTHORSHIP

Conception and design: K. X. Zhang, H. Varma, V. S. Shah; Acquisition of data: K. X. Zhang, H. Varma, V. S. Shah; Analysis and interpretation of data: K. X. Zhang, H. Varma, Y. Cao, V. S. Shah. Drafting the manuscript: K. X. Zhang, Y. Cao, V. S. Shah; revising the manuscript for intellectual content: K. X. Zhang, H. Varma, Y. Cao, V. S. Shah. Final approval of the completed manuscript: K. X. Zhang, Y. Cao, V. S. Shah.