Introduction

Initial postoperative overcorrection could be desirable after exotropia surgery because of a tendency towards postoperative exodrift. However, postoperative overcorrection may remain, or delayed-onset consecutive esotropia may occur. Approximately 5% [1,2,3] after more than 6 months of follow-up, to 20% after 1 year of follow-up [4] of patients after surgical treatment of intermittent exotropia reportedly show consecutive esotropia. The incidence of consecutive esotropia varies because the follow-up period and diagnostic criteria for consecutive esotropia differ for each study. Risk factors include amblyopia [2, 3], high myopia [2], lateral incomitance [2], poorer stereopsis [5], high accommodative convergence/accommodation ratio [6], younger age [7] and types of surgery [2, 3].

Initial postoperative overcorrection may be managed non-surgically with occlusion, hyperopia correction, prism glasses, or Botox injection; however, if it persists, a patient may develop cosmetic problems and binocular vision dysfunction such as loss of stereopsis [2,3,4, 6, 8, 9]. In addition, even if there is orthotropia early postoperatively, consecutive esodeviation could occur later, leading to deterioration of binocular function, especially in children.

Some previous studies have examined the incidence and risk factors of consecutive esotropia. However, to the best of our knowledge, there are few reports in the English literature on the timing of its occurrence or its longitudinal course. Therefore, we investigated the onset, duration and course of esodeviation after surgery for intermittent exotropia.

Materials and methods

Subjects

We conducted a retrospective review of the medical records of patients who underwent surgery for the treatment of basic-type intermittent exotropia between 2011 and 2017 at Seoul National University Children’s Hospital in South Korea. This study was approved by the Institutional Review Board of Seoul National University Hospital in South Korea and the study protocol followed the tenets of the Declaration of Helsinki. The minimum required follow-up period after surgery was 24 months, and patients who required reoperation within 24 months after the primary surgery were also included.

The following patients were excluded from the study: patients older than 17 years of age at the time of surgery, those with a history of prior strabismus surgery, those who underwent simultaneous vertical and/or oblique muscle surgery and those with a systemic anomaly such as a neurologic disorder or developmental delay. Patients with dissociated vertical deviation or oblique muscle overaction that did not require surgery were included. The following patient characteristics were recorded: age at onset of deviation, age at the time of surgery, sex, birth weight, fixation preference, refractive errors, types of exotropia, associated strabismus (vertical deviation or oblique dysfunction), angle of deviation, presence of lateral incomitance, methods of surgery performed (bilateral lateral rectus muscle recession [BLR] or unilateral lateral rectus muscle recession–medial rectus muscle resection [RR]), duration of esodeviation, presence of prismatic correction, stereopsis and fusional ability.

Preoperative ophthalmic examinations

A diagnosis of intermittent exotropia is made using the cover test and by measuring the angle of exodeviation using the alternate prism cover test. These tests are performed with the individual fixing on an accommodative target at distance (6 m) and then at near (0.33 m) fixation. The cover test shows intermittent or constant exotropia either at distance fixation only or also at near fixation. Intermittent exotropia is divided into the following types: basic, pseudo-divergence excess, true divergence excess and convergence insufficiency. This study only included patients with basic type.

The angle of deviation is indicated as prism dioptres (PD), with exodeviation represented as minus (−) numbers and esodeviation (+) as plus numbers. Lateral incomitance was defined as more than 20% decrease in the right or left gaze from the primary position. Preoperative measurements of the angle of deviation were performed on at least three different occasions by one experienced ophthalmologist (S-JK). Based on the maximal preoperative angle of deviation during follow-up, all surgeries were performed under general anaesthesia by a single surgeon (S-JK). The surgeon selected the procedure (BLR or RR) [10] and surgical amount using a surgical formula based on the surgeon’s experience (Table 1).

Table 1 Surgical dosage used for children with basic-type intermittent exotropia in this study.
Full size table

Amblyopia was defined as a difference of at least 2 lines in visual acuity between eyes or best-corrected visual acuity lower than 20/30 on the Snellen visual acuity chart. Refractive errors were determined using cycloplegic refraction, and analysed as mean spherical equivalent values. Anisometropia was defined as a difference of >1.5 dioptres (D). In patients with a fixation preference or amblyopia, part-time occlusion therapy was prescribed with spectacles if needed.

Stereoacuity was tested with the Titmus stereotest (Stereo Optical, Chicago, IL, USA), where cooperation was adequate, before surgery and at the final follow-up. All values were transformed to log arcsec for the purpose of analysis [11], and divided into good (40–100 arcsec) and poor (>100 arcsec) group. When stereoacuity was recorded as nil if the largest disparity could not be passed, and the score for nil stereopsis was 6000 arcsec, it was regarded as total loss of binocular function. Fusional ability was measured using the Worth 4 dot Test at distance and near where cooperation was adequate.

Postoperative managements and evaluations

Postoperative alignment at distant and near fixation in the primary position was measured at week 1, months 1, 6, 12 and 24 or later. Patients with diplopia associated with postoperative esotropia were managed by alternate full-time patching for 1–4 weeks until diplopia was resolved. If the esotropia had not decreased within 4 weeks, cycloplegic refraction was performed and the residual hyperopia was fully corrected and base-out prism glasses were prescribed to allow constant fusion until esotropia resolution. If esodeviation is more than 15 PD, Fresnel (press-on) prism (3M Press-On Optics; 3M Health Care, St Paul, MN, USA) was prescribed. During the follow-up, the PDs of the glasses were reduced if esodeviations decreased without diplopia. In cases where the consecutive esotropia persisted for longer than 6 months postoperatively, the esodeviation increased with prisms, or the patient rejected the conservative treatment option, reoperation was considered. The number of patients with esodeviation at every follow-up was evaluated. When esodeviation occurred later in patients with orthotropia or exodeviation in an early phase after surgery (at postoperative month 1), it was defined as delayed-onset consecutive esotropia.

Statistical analyses were performed using SPSS for Windows version 23.0 (IBM Corp., Armonk, NY). The independent Student’s t-test, chi-square test and Fisher’s exact test were used. A logistic regression analysis was used to investigate factors associated with reoperation rate for consecutive esotropia; first using an univariate model, then with a multivariate model that included variables with P < 0.10 in the univariate model. P values < 0.05 were considered statistically significant.

Results

Patient demographics and clinical characteristics

A total of 336 patients were included in this study (BLR, 236; RR, 100), with a mean age at exotropia surgery, 6.2 ± 2.1 years. The mean maximal preoperative angle of exodeviation was 30.5 ± 5.2 PD at distance and 33.0 ± 5.2 PD at near. After surgery, deviation at postoperative week 1 was 5.9 ± 7.1 PD, −1.8 ± 6.2 PD at postoperative month 1, −3.6 ± 8.4 PD at postoperative month 6, −6.4 ± 9.0 PD at postoperative year 1, −9.8 ± 9.7 PD at postoperative year 2 and −12.2 ± 12.1 PD at the final follow-up. Reoperation was performed in 76 patients (22.6%) with recurred exotropia and in 12 patients (3.6%) with consecutive esotropia during the 48.3 ± 21.5 (range 24.0–120.7) months of postoperative follow-up (Table 2).

Table 2 Demographics and clinical characteristics of patients who underwent surgery for the treatment of basic-type intermittent exotropia.
Full size table

Postoperative exodrift mostly occurred during the 1-month follow-up. In the RR group, the proportion of patients with overcorrection gradually decreased. While patients in the BLR group showed less exodrift after 1 month compared to that within 1 month. The amount of exodrift was significantly decreased after 1 month (Fig. 1).

Fig. 1: Longitudinal changes of angle of deviation in patients following intermittent exotropia according to the type of surgery.
figure 1

Exodrift occurred mostly during 1 month after surgery in both the RR and BLR group. Afterwards, exodrift steadily occurred over time after the surgery in the RR group. In contrast, patients in the BLR group showed an earlier exodrift and a more stable course after surgery. * with exodeviation represented as minus (−) numbers and esodeviation (+) as plus numbers.

Full size image

Longitudinal changes in postoperative deviation

Esodeviation at postoperative week 1 was observed in 218 patients (64.9%) and was considered immediate overcorrection after exotropia surgery. The mean esodeviation angle was 9.6 ± 5.9 PD at distance and 9.0 ± 7.6 PD at near. Afterwards, esodeviation tended to decrease. There were 52 patients (15.5%) with esodeviation at postoperative month 1, 47 (14.0%) at postoperative month 6 and 34 (10.1%) at postoperative year 1. At postoperative year 2, there were 28 patients (8.3%) with consecutive esotropia: six in the RR group, and 22 in the BLR group. Sixteen patients discontinued use of the prismatic glasses at a mean 43.3 ± 21.1 months of follow-up, six patients underwent surgery and six patients continued using prismatic glasses at a mean 31.3 ± 9.1 months of follow-up (Table 3, Fig. 2).

Table 3 Number of patients with esodeviation and postoperative angle of deviation in children with basic-type intermittent exotropia following surgery.
Full size table
Fig. 2: Number of patients with postoperative esodeviation after intermittent exotropia surgery.
figure 2

The angle of esodeviation steadily decreased over time. At postoperative year 2, there were 28 patients (8.3%) with consecutive esotropia: six in the RR group, and 22 in the BLR group.

Full size image

Of the 52 patients with esodeviation at postoperative month 1, except for one patient for whom stereopsis was not measured before surgery, there were 36 patients with good stereopsis and 15 with poor stereopsis. Of the 15 patients with poor stereopsis, three presented with total loss of binocular function, nil stereopsis and underwent surgery for consecutive esotropia; at the final follow-up, one patient had recovered binocular function as gross stereopsis, while loss of binocular function remained in the other two, who did not recover (Fig. 3).

Fig. 3: Changes in stereopsis in patients with esodeviation at postoperative month 1.
figure 3

Of the 52 patients with esodeviation at postoperative month 1, except for one patient for whom stereopsis was not measured before surgery, there were 36 patients with good stereopsis and 15 with poor stereopsis. Three patients presented with total loss of binocular function, nil stereopsis and underwent surgery for consecutive esotropia; at the final follow-up, one patient had recovered binocular function as gross stereopsis, while loss of binocular function remained in the other two, who did not recover.

Full size image

Patients who underwent surgery for consecutive esotropia

Twelve patients underwent surgery for consecutive esotropia during the follow-up period; two patients at postoperative month 6, three at postoperative month 12 and six at postoperative month 24. Two patients had orthotropia at postoperative month 24, but consecutive esotropia occurred at postoperative month 36, so surgery was afterwards performed. The mean age at exotropia surgery in the 12 patients who underwent surgery for consecutive esotropia was 4.0 ± 2.1 years; the angle of esodeviation at postoperative month 1 was 7.8 ± 7.9 PD at distance and 8.6 ± 9.1 PD at near. They were all included in the BLR group, and the postoperative stereoacuity at the final follow-up was 3.26 ± 0.77 log arcsec; six patients lost stereoacuity. In contrast, to those who underwent surgery for consecutive esotropia, in patients without reoperation, the mean age at exotropia surgery was 6.6 ± 2.2 years, angle of deviation at postoperative month 1 was 1.1 ± 6.3 PD at distance and 2.0 ± 7.0 PD at near and the postoperative stereoacuity at the final follow-up was 1.89 ± 0.45 log arcsec (p = 0.003, <0.001, <0.001 and 0.002, respectively, the independent Student’s t-test).

Factors including age, surgical method, stereoacuity and amount of initial overcorrection were significantly associated with the reoperation rate for consecutive esotropia in the logistic regression analysis; younger age at exotropia surgery, BLR surgery, poorer stereoacuity and greater amount of initial overcorrection were associated with the reoperation rate for consecutive esotropia (p = 0.005, <0.001, 0.011 and 0.027, respectively), whereas the preoperative angle of exodeviation and fusional ability was not associated (p = 0.245 and 0.458).

Delayed-onset consecutive esotropia

Among the 284 patients with orthotropia or exodeviation at postoperative month 1, 13 patients subsequently developed delayed-onset consecutive esotropia. Their mean age at surgery was 4.7 ± 2.0 years. Their mean maximal preoperative angle of exodeviation was 36.6 ± 5.4 PD, which was significantly larger than that in the other patients, 30.3 ± 5.0 PD (p = 0.004, chi-square test). It was also associated with the types of surgeries for exotropia (p = 0.005, Fisher’s exact test). All patients, except one, underwent BLR. The surgical dosage of BLR in the delayed-onset esotropia group was 7.17 ± 0.90 mm, and the overall surgical dosage of BLR in the study was 6.30 ± 0.85 mm (p = 0.028, chi-square test).

Five patients who were found to have consecutive esotropia at an earlier phase and were actively treated with prismatic glasses showed gradual decrease of the angle of esodeviation. They finally discontinued use of the prismatic glasses. However, another three patients were still using the prismatic glasses at the final follow-up. The remaining five patients underwent surgery for consecutive esotropia; no evidence of lateral rectus slippage was found at surgery in the three patients in whom the lateral recti were re-operated (Table 4).

Table 4 Clinical characteristics of patients with delayed-onset consecutive esotropia who underwent surgery for the treatment of basic-type intermittent exotropia.
Full size table

Discussion

Persistent consecutive esotropia after surgery disappears in most cases with conservative management. Prismatic glasses following alternate full-time patching help establish fusion and reduce the angle of esodeviation for some patients with consecutive esotropia without need for additional surgery [4, 12]. In the present study, there were 28 patients (8.3%) with consecutive esotropia at postoperative year 2. Among them, 16 patients discontinued use of the prismatic glasses at a mean 43.3 ± 21.1 months of follow-up. In most patients, management with prism spares the patient from the need for reoperation and helps maintain a good sensory status.

We have previously reported that the amount of initial overcorrection was important to achieve favourable surgical outcomes in children with intermittent exotropia [13, 14]. Consecutive esotropia has been reported, which may remain as persistent overcorrection [4, 8, 9]. In the present study, most patients presented with immediate overcorrection in the initial postoperative period, but postoperative esodeviation showed a significant resolution at 1 month postoperatively. Because of the tendency towards exodrift, the patients with orthotropia or small angle of exodeviation at the early phase stopped visiting the clinic or increased the period until the next follow-up visit. However, delayed-onset consecutive esotropia may also occur, and its exact cause remains unclear. Therefore, it would be meaningful to further examine the clinical characteristics of the patients who develop this condition.

In this study, 13 patients presented with delayed-onset postoperative consecutive esotropia. The onset of consecutive esotropia ranged from 6 months to 3 years. It was difficult to predict this course based only on the postoperative angle of deviation at the early phase. The parents of some children who were determined to have orthotropia at the office examination reported that the children intermittently appeared to have esotropia, especially immediately after waking up, after close follow-up, they were found to have esotropia at the office examination. We suppose that these patients barely maintained normal motor alignment at the early postoperative phase. If fusional ranges decrease exceeding the threshold, decompensation occurs leading to consecutive esotropia. It could be regarded as decompensation of convergence–divergence, and unrevealed central factors might be related to this change. Therefore, it is the responsibility of the surgeon to encourage patients to return for regular follow-up visits even if they present with orthotropia at the early phase after exotropia surgery. In addition, as these changes occurred even at 3 years after surgery, postoperative alignment with longer follow-up should be carefully monitored for possible change in esodeviation. However, continuing long-term follow-ups in a tertiary hospital would be overburdened in many countries and healthcare systems. Therefore, regular check-ups may take place at the primary hospital and a re-referral should be provided when changes in the alignment are suspected.

The incidence of consecutive esotropia could vary according to the methods of exotropia surgery. Jang et al. [2] and Lee et al. [15] reported a higher incidence of consecutive esotropia following RR, and Kim and Choi. [3] and Park et al. [16] reported a higher incidence following BLR. In the present study, the total incidence during the follow-up period did not differ between the two groups. However, the patients in the BLR group more frequently had delayed-onset consecutive esotropia. Interestingly, Lee et al. [12, 15] also reported that the average timing of prism glasses’ prescription after BLR recession was 7.4 months due to late overcorrection after 3 months in some patients, which was significantly later than that after RR. The average timing of prism glasses prescription after RR was 2.3 ± 2.9 months. It may be induced by relatively more innervational input to the medial rectus (MR) muscle after BLR recession, which increases MR muscle tone [17]. Strong and variable tonic convergence or a tight MR muscle might be associated with the development of consecutive esotropia [17, 18]. This convergence usually occurs in the midterm phase after surgery. In such cases, botunilum toxin injection could be a helpful intervention without the need to proceed to further surgery, which we did not consider due to the cost problem; Botox is not covered by insurance in our country.

Recent studies have examined changes in the extraocular muscle structure after various types of strabismus surgery and have shown that there was significant adaptation of the operated extraocular muscles [19, 20]. The unfavourable outcomes of strabismus surgery may result from the inability of the oculomotor system to adapt to the significant alteration in eye alignment produced by surgical manipulation. In the present study, the surgical dosage of BLR in the delayed-onset esotropia group was larger than that in the others. Greater changes in the sarcomeres of extraocular muscles in response to surgery in antagonist muscles might contribute to unpredictable changes in eye alignment after surgery [21]. Therefore, the possibility of consecutive esotropia should be considered, especially in patients who undergo BLR recession with a larger angle of exodeviation.

In addition, high AC/A ratios or lateral rectus underaction and/or stronger MR tension following release of tension from recession of the lateral recti could be associated with this phenomenon. Baik et al. [22] recently asserted that the cause of delayed-onset consecutive esotropia is unknown, but an increase in “near work” in the current generation during the unstable postoperative period within several early months after surgery may be playing a role. In addition, as children undergo emmetropisation or myopic shift, changes in the refractive error should be considered during the follow-up period. However, we only analysed the preoperative refractive errors based on the cycloplegic refraction. Further prospective study including the AC/A ratios and changes in refractive errors is necessary to provide further information.

Further, it has not been determined whether the preoperative angle of exodeviation is associated with surgical success in intermittent exotropia; in the present study, it was not associated with the surgical outcome at the final follow-up. However, patients with larger preoperative angle of exodeviation were more prone to present with delayed-onset consecutive esotropia, as were those who were younger; their mean age at exotropia surgery was 4.7 years. There is still no definite consensus on the appropriate timing of surgery for intermittent exotropia. A study reported that it is preferable to perform surgery for intermittent exotropia earlier, before the time when binocular vision function is completed, to achieve optimal sensory function [23]. Others asserted that there was no difference in the incidence of consecutive esotropia between younger and older children after 3–4 years of age [24, 25]. Conversely, there have been reports that delaying surgery to avoid overcorrection in younger children with immature visual function is recommended; it would be better to delay surgery before age 4 years [26, 27].

In functionally and visually immature children, it is more difficult to maintain motor alignment, which may lead to loss of stereopsis and the development of amblyopia [6]. Although surgical treatment resolves this state, younger children with a larger angle of exodeviation are more likely to lose the fusional ability earlier. In addition, the effect of persistent overcorrection can differ by age [3, 23, 28]. Considering the disadvantages of consecutive esotropia, it may be more appropriate to delay the operation for mild or moderate intermittent exotropia in children with immature visual function. In other words, it would be better to consider early surgery only when the angle of deviation increases or fusional ability is poor. As the preoperative angle of deviation was markedly large in younger children in this study, we had to opt for surgery at an early age. We also want to emphasize that younger children should be carefully monitored, and early management such as patching or prismatic correction should be applied to patients [29, 30].

This study included a relatively larger sample that was followed for a longer follow-up period compared to those in previous studies. However, there are some limitations to this study. First, it was a retrospective chart review, and patients lacking documentation were excluded. In addition, detailed information such as for stereopsis during the follow-up period was not available. Second, fusional control was assessed using the Worth 4 dot test, which is prone to strong dissociation, with the possibility of not reflecting the normal environment. Real-life conditions are quite different, and attention should be paid when interpreting these data. Therefore, the fusional range should be investigated to clarify this matter. In addition, we did not follow-up fusional ability and stereoacuity immediately postoperatively. Further investigation of fusional ability and stereopsis would be necessary. Third, despite setting the minimum follow-up at 2 years, the final follow-up period differed by patient. A prospective study with a long-term follow-up exceeding 2 years should be performed in the future.

In conclusion, in patients with postoperative overcorrection, the angle of esodeviation decreased and patching/prismatic correction achieves good surgical outcome. However, delayed-onset of consecutive esotropia occurred in some patients, especially in those who had undergone BLR recession with a larger angle of exodeviation. Therefore, our results indicate that postoperative alignment should be carefully monitored after surgery for intermittent exotropia.

Summary

What was known before

  • Overcorrection after exotropia surgery has been considered to be desirable because of a tendency of exodrift. Initial postoperative overcorrection is managed by non-surgical methods in an early phase. However, postoperative overcorrection may remain, or delayed-onset consecutive esotropia may occur. If consecutive esotropia persists, a patient may develop cosmetic problems and deterioration of binocular function, especially in children. There have been previous studies about the incidence and risk factors of consecutive esotropia. However, to the best of our knowledge, there are few reports in the English literature on the timing of its occurrence or its longitudinal course.

What this study adds

  • We investigated the longitudinal course of esodeviation after surgery for intermittent exotropia, including a relatively larger subject that was followed for a longer follow-up period compared to previous studies. In patients with postoperative overcorrection, the angle of esodeviation decreased and patching/prismatic correction helped achieve the good surgical outcome. However, delayed-onset consecutive esotropia occurred in some patients, especially in those who had undergone BLR surgery with a larger angle of exodeviation. Younger age at exotropia surgery, BLR surgery, poorer stereoacuity, larger preoperative angle of deviation and more amount of initial overcorrection were associated with reoperation rate for consecutive esotropia. Since several strabismus experts have experienced the delayed-onset consecutive esotropia leading to the binocular dysfunction and reoperation, postoperative alignment should be carefully monitored after surgery for intermittent exotropia.