Introduction

Abducens or sixth cranial nerve palsy is the most common extraocular muscle palsy, causing horizontal diplopia that worsens in ipsilateral gaze and head turn towards the affected side [1]. Treatment in the initial stages of this paralysis is aimed at avoiding diplopia by means of patching, prism glasses or botulinum toxin (BT) injection [2]. When the symptoms do not disappear and the eye misalignment remains stable over time, performing a surgical intervention is considered. In patients with residual action of a partially palsied muscle, recession and resection surgery of the horizontal recti will have greater effect with fewer complications than a transposition. Hence, transposition techniques should be reserved for cases with poor or absent lateral rectus function [3,4,5,6,7,8].

Various transposition procedures have been described for abducens palsy, classical transposition techniques including partial tendon [9], full tendon [10], and muscle union procedures [11]. Vertical rectus transposition surgery for the treatment of paralytic strabismus was first described by Hummelshein in 1907 [12], it was not until 1959 that full tendon vertical rectus transposition was described by Schillinger [13]. The augmented modalities to increase these procedures’ effect include posterior augmentation suture (Foster suture) [6], resection of the transposed muscle [14], and additional medial rectus recession (MRc) or BT injection of the ipsilateral medial rectus muscle [15]. Possible complications following full tendon vertical rectus transposition are residual horizontal deviation, overcorrection, vertical deviation, ocular torsion, and anterior segment ischemia [16], especially if MRc is required or in patients who have undergone previous recess/resect surgery on the horizontal rectus muscles.

In 1964, Jensen introduced a different concept with splitting of the muscle bellies of both the vertical and the lateral rectus. He split the superior and inferior rectus muscles into medial and lateral halves and then united them with the upper and lower halves of the lateral rectus muscle in the equator region [11]. This technique was initially described to protect against anterior segment ischemia but, unfortunately, a case of acute anterior segment ischemia in the immediate postoperative period has been described [17].

More recently, full tendon transposition (FTT) of either the superior rectus or the inferior rectus to the lateral rectus with posterior fixation suture, with or without MRc, has been proposed [18,19,20,21,22,23].

In order to reduce complications associated with the previously described transposition techniques, mainly ischemia, Nishida described a new transposition technique in 2003. In this technique, the halves of the vertical rectus muscle bellies were divided and sutured onto the sclera without either tenotomy of the vertical rectus muscles, as in Hummelsheim’s procedure, or surgical treatment of the lateral rectus muscle, as in Jensen’s procedure. Nine patients were evaluated with this technique, obtaining comparable results to those published with other muscle transposition techniques [24]. This technique was then further modified and Nishida et al. reported a muscle transposition procedure in which the vertical rectus were not divided, instead the lateral margin of each vertical rectus muscle was transposed superotemporally or inferotemporally and were sutured onto the sclera without tenotomy or muscle splitting (Fig. 1). This is commonly considered as the modified Nishida technique [25].

Fig. 1: Modified Nishida procedure.
figure 1

LR lateral rectus.

Full size image

The objective of the current study is to compare the effectiveness of three procedures: modified Nishida procedure alone vs modified Nishida procedure combined with medial rectus recession (MRc) vs modified Nishida procedure combined with MRc and botulinum toxin (BT) in a large series of patients with severe unilateral sixth nerve palsy who underwent modified Nishida procedure.

Methods

In this interventional case series, the preoperative, surgical and postoperative data of patients diagnosed with severe abducens nerve palsy who underwent modified Nishida procedure between January 2017 and December 2019 were reviewed retrospectively and a standardized data collection form was completed by each author for every patient. The following centres were included: SÖ´s private Clinic in Aydin (Turkey), All India Institute of Medical Sciences in New Delhi (India), Instituto de Oftalmología Fundacion Conde de Valenciana in Ciudad de Mexico (Mexico), Hospital Parmenio Piñero in Buenos Aires (Argentina), Kaohsiung Medical University Hospital in Kaohsiung (Taiwan), Omni Hospital in Guayaquil (Ecuador), Hospital Metropolitano in Quito (Ecuador) and Hospital Clínico San Carlos in Madrid (Spain).

Patients with abduction defects due to severe acquired unilateral sixth nerve paralysis and no recovery within six months after onset were included in the study. Those patients who had undergone previous surgery, presented with bilateral or moderate sixth nerve paralysis and patients with under 6 months postoperative follow-up were excluded.

Data collected included age, sex, aetiology of the abducens nerve palsy, laterality (specifically whether the surgery was performed on the affected or the unaffected eye), preoperative deviation in the primary position (horizontal and vertical deviations considered separately, for both far and near), preoperative abduction deficit and diplopia. The ophthalmologists measured the horizontal and vertical deviations in prism dioptres (PD) by alternate prism cover test. Limitation of ductions was measured on a scale from 0 to −8, with 0 indicating full rotation up to canthus, −4 for abduction as far as primary position and −8 for globe immobilized in extreme adduction.

Patients with Nishida modified procedure were divided into three groups according to the surgical technique: those with Nishida modified procedure alone, those with a MRc added to the Nishida procedure and a group of patients in which BT was injected to the recessed medial rectus muscle. In cases with muscle recession, the amount in millimetres was recorded.

Briefly, the surgical procedure performed, modified Nishida procedure, consisted of a 10 mm transposition of the superior and inferior rectus temporally at 12 mm from the limbus, equidistant to the lateral rectus and superior or inferior rectus. The amount of MRc was based on the amount of deviation in primary position, tightness of the medial rectus as judged subjectively by preoperative forced duction testing, and standard tables were used to calculate dosage. When BT was added, a standard dose of 5 units in 0.1 mL was injected in all cases.

Postoperative follow-up examinations were performed at 1 and 6 months. Postoperative deviation in the primary position (horizontal and vertical deviations considered separately) and postoperative abduction deficit were recorded in the same manner as the preoperative examinations.

Outcome measures studied included change in horizontal esodeviation in primary position fixing on a distant target (in PD and percentage of preoperative deviation corrected), diplopia, complications and the need for further surgery. Surgical success was defined as the absence of diplopia in primary position with postoperative primary position deviation of 10PD or less. Overcorrection and undercorrection were defined as deviations of more than 10PD in the appropriate directions.

Statistical analysis was performed by the Hospital Clinico San Carlos team using SPSS software (Version 22.0 SPSS Inc., Chicago, IL) to summarize baseline characteristics and outcomes and to compare pre- and postoperative deviations. Continuous variables are presented as mean and standard deviation (SD) with range, while categorical variables are presented as numbers and percentages. The paired t test was used to analyse the preoperative and postoperative changes in esotropia in primary position. P < 0.05 was considered statistically significant.

Results

A total of 43 patients with modified Nishida operated at the centres included were initially selected. However, 3 were excluded due to previous surgery, 4 to incomplete follow-up and 4 because they were bilateral. Hence, the final study population comprised 32 patients of mean age 38.6 ± 19.8 years (range 3–82); 56% male and 44% female patients.

Table 1 shows the demographic, preoperative and postoperative features of the patients. Most participants (15 patients, 47%) presented with idiopathic abducens nerve palsy. Mean preoperative deviation was 63.0 ± 27.3 PD (range 24–125) and mean limitation of abduction −4.5 ± 1.2 (range −8 to −4), 88% of patients referring diplopia.

Table 1 Demographic, clinical, preoperative and postoperative characteristics of the patients included in the study.
Full size table

Regarding surgical technique, 5 patients underwent a modified Nishida procedure alone, 24 patients had an additional MRc, and 3 patients were also injected with BT. Surgical technique was decided preoperatively at the surgeon’s discretion. In all of the patients, the Nishida modified procedure, MRc and BT injection were all performed in the affected eye, except for 3 patients who underwent Nishida modified procedure combined with MRc in whom the procedure was performed on the unaffected eye. The patients with Nishida modified procedure and MRc are detailed in Table 2, mean MRc being 6.1 ± 1.3 mm (range 5–10) in this group. In patients with modified Nishida procedure, MRc and BT the mean MRc was 8.0 ± 3.6 mm (range 5–13).

Table 2 Characteristics of the patients who underwent modified Nishida procedure and medial rectus recession.
Full size table

Mean horizontal deviation preoperative, at 1 month and 6 months with the different techniques are illustrated in Fig. 2. All postoperative results were statistically significant (p < 0.05). The changes in horizontal deviation in primary position stratified by technique and preoperative deviation (<40, 41–79 and >80 PD preoperatively) are shown in Fig. 3. Patients with BT were followed up to 12 months, obtaining similar results to those at 6 months.

Fig. 2: Long-term outcomes of horizontal deviation in primary position with the different surgical techniques.
figure 2

Mean measurements with standard deviation are shown. PD prism dioptres, MRc medial rectus recession, BT botulinum toxin.

Full size image
Fig. 3: Long-term outcomes of mean horizontal deviation in primary position with the Nishida modified procedure and medial rectus recession stratified by technique and preoperative deviation.
figure 3

PD prism dioptres, MRc medial rectus recession, BT botulinum toxin.

Full size image

Overall, the modified Nishida technique by itself corrected 29.4 ± 6.6 PD (range 20–36), whereas adding a MRc corrected 62.6 ± 23.8 PD (range 24–120). Modified Nishida procedure, MRc and BT altogether corrected 95.0 ± 18.0 PD (range 75–110). Surgical success was overall 75%; 100% with modified Nishida procedure, 75% with an additional rectus muscle recession and 33% in the group with BT. 16% of patients required further surgery, but no surgical complications were noted. At the 6-month follow-up 22% of all the included patients presented diplopia: none from the modified Nishida procedure alone, 25% of the augmented technique and 33% of the patients with BT. None of the patients lost vision or had anterior chamber signs of ischemia. Other results from similar studies using the Nishida procedure, along with those with an additional rectus muscle recession are provided in a Supplemental Table.

Discussion

Various vertical rectus muscle transposition procedures have been described for abducens palsy. We hereby present the results of the modified Nishida technique in a considerably large sample of patients. Before it was described, other techniques with higher associated risks were used.

One of the most effective surgical methods available for large-angle strabismus with minimal or absent muscle force is FTT augmented with posterior fixation [6, 26, 27]. González et al. [28] compared three different transposition procedures without simultaneous medial rectus weakening: FTT, FTT with 4 mm resections before reinsertion, and FTT with myopexy sutures. FTT with 4 mm resections before reinsertion corrected the most esotropia (mean correction 46.4 PD; range 33–64) and improved abduction to the greatest degree. Complications following full tendon vertical rectus transposition included residual horizontal deviation, vertical deviation, ocular torsion, and anterior segment ischemia [16].

Until Jensen’s technique was described [11], transposition techniques required the performing of tenotomy with the subsequent risk of ischemia. Jensen’s procedure is thus considered safer because it reduces the risk of anterior segment ischemia. However splitting of 3 rectus muscles is required and cases of anterior segment ischemia have been described [17].

In 2003 Nishida et al. [24] reported ten cases of abducens palsy treated using a muscle transposition procedure in which the split halves of the vertical rectus muscle bellies were sutured onto the sclera without performing a tenotomy of the vertical rectus muscles. They concluded that it achieved similar postoperative results as vertical rectus transposition. In 2005, the authors reported the results of a series of three cases with this procedure but without splitting the vertical recti [29] and later described the outcomes of nine more patients [25]. This modified Nishida technique was performed by suturing the lateral muscle belly margins of vertical recti onto the superotemporal or inferotemporal sclera to transfer the lateral muscle bellies temporally and create additional muscle insertions as new points of action which can generate abductional forces.

In Nishida’s initial study, data from patients with unilateral and bilateral paralysis was discussed together, as well as that of patients with and without previous surgeries and with and without recession of the middle rectus muscle. No separate analysis was performed so as to reach specific recommendations depending on preoperative deviation and the technique used was the initially described with vertical recti splitting. If we exclude bilateral cases reported in Nishida’s first paper and analyse the results of the technique by itself in five patients, preoperative mean deviation was 42 PD with a postoperative deviation of 3.4 PD (average correction 38.6 PD).

Using the modified Nishida technique, Nishida et al. reported nine cases of severe abducens palsy, six unilateral and three bilateral. In cases where muscle transposition alone was performed, a correction of 24–36 PD was noted. Our results show a mean correction of 29.4 ± 6.6 PD (range 20–36), being thus comparable. In the group of augmented modified Nishida with MRc (patients with the same selection criteria as ours), deviation improved by 50–56 PD. Abduction improved beyond midline in all cases and none had anterior segment ischemia. Our results demonstrate the resolution of diplopia in most patients, as well as the improvement of the limitation of abduction. Therefore, this procedure seems to achieve similar results to other published procedures, and it is simple to perform as it requires only a suture from muscle to sclera, with tenotomy or muscle splitting being unnecessary.

In addition, Sabermoghadam et al. [30] provided a series of ten patients undergoing Nishida’s transposition combined with MRc with a mean preoperative deviation of 49.5 ± 9 PD (range 40–65). Postoperative deviation ranged from orthotropia to 12 PD of esotropia and all the patients obtained abduction at least beyond the midline. No vertical ductional disturbances or deviations were developed, adduction was not compromised in any patient and no patient developed anterior segment ischemia. In all of the patients, this transposition was combined with MRc of 5–6.5 mm regardless of result of forced duction test. The amount of MR recession was 5 mm in 40 PD esotropia and increased in steps of 0.5 mm for each 5 PD increase in esotropia. These results are comparable to those found in patients in our series, but the follow-up time was relatively short and they did not evaluate separately patients with and without BT injection.

Regarding muscle transpositions combined with BT injection, only Flanders et al. [15] reported five patients who were given BT before surgery, correcting a mean deviation of 66 PD when combined with complete transposition of vertical rectus. In our work we present three patients with Nishida’s transposition combined with MRc and BT with an average correction of 95 DP, along with improvement of abduction and diplopia in 2 of the 3 patients.

Based on Nishida’s technique, Marcon et al. [31] described in 2013 a muscle union procedure in which a single non adsorbable suture was used to unify the muscle belly of two recti muscles supplying the action of the paretic muscle without scleral pass nor muscle splitting. Its mean effect in isolation was 35 PD (range 25–40) in 39 patients and results seemed to be satisfactory. However, Marcon’s series is heterogeneous with joint analysis of all patients including those with previous surgeries, BT injections, and with various aetiologies.

Limitations of our study include that multiple surgeons were involved and there may have been minor variations in the surgical procedure; measurement of binocular field of vision would have also been useful in these patients and the procedure was not randomly selected, but instead decided by the surgeon depending on preoperative deviation. Notwithstanding, in most published works, the number of cases described is low and the analysis of the data is very heterogeneous without results being grouped according to surgical technique. To the best of our knowledge, this is the largest and most homogeneous series of patients receiving Nishida’s procedure for severe sixth nerve palsy. In addition, patients were subdivided into three groups for a more accurate analysis of the results.

In our series, high success rate was achieved in patients with the modified Nishida procedure, regardless of baseline deviation magnitude, although the deviation amount corrected varies with the procedures. There was also an improvement in the limitation of abduction and resolution of diplopia in most patients with no complications associated with surgery being noted. Therefore, modified Nishida procedure alone is an excellent surgical option for treating severe sixth nerve palsy.

Summary

What was known before

  • Transposition techniques are reserved for cases of sixth nerve palsy with poor or absent lateral rectus function.

What this study adds

  • Excellent outcomes with fewer complications are obtained with modified Nishida procedure.

  • The need for additional procedures such as medial rectus recession and botulinum toxin which increase the effect in primary position can be determined depending on passive duction and preoperative horizontal deviation.