Kushner [51] suggested that increasing the dose of BLR might be beneficial for patients with basic type X(T). Since Lee et al. [99] first reported better surgical results with augmented BLR by 1.5 to 2.5 mm, BLR augmentation by 1.0 to 2.5 mm or increasing the target angle by 5 PD [100] have been shown to improve the long-term surgical results without a significant difference in the overcorrection rate [32,99,100]. Kim et al. [32] found that patients with divergence excess type showed better results than those with the basic type after augmented BLR. They suggested that augmentation should be considered when planning BLR, especially for the divergence excess type X(T). Awadein et al. [101] demonstrated that augmented BLR by 1.5 mm significantly increased the success rate only in older patients above 12 years of age. They suggested that augmentation according to age could improve the surgical outcome of X(T).

Increasing the surgical dosage of MR resection by 1 mm improved the long-term surgical results compared to the original dosage without significant differences in overcorrection [34]. With an increased surgical dosage, the recurrence rate (7.0 %) at postoperative 2 years was remarkably lower compared to the recurrence rate (range, 38%-61%) reported in the previous studies [25,27,62,63].

The relationship between age and surgical results is controversial. Early surgery may help develop better binocular vision [102-104], but may also pose a risk of developing amblyopia to children with immature visual system because of the postoperative overcorrection [105]. The fact that the surgery was performed at a young age could imply that X(T) may be present from an early age consequently disturbing the binocular vision, and that X(T) was severe enough for the caretakers to notice. Late surgery can provide benefits of more accurate measurements and lesser chances of developing amblyopia, but can impose lesser chances of developing binocular vision [102].

Maruo et al. [62] reported that an alignment ≤ ±4 PD was achieved with surgeries that were performed at <3 years or >11 years of age. Abroms et al. [102] reported better sensory and motor outcomes with surgeries performed at <7 years of age, X(T) duration <5 years, or with intermittency rather than constancy. Asjes-Tydeman et al. [103] reported that surgeries performed at <7 years of age could result in better motor (exodeviation of <10 PD) and sensory outcomes as well as fewer instances of reoperations. Tibrewal et al. [104] found an older age at surgery and a larger preoperative deviation to impose greater chances of failure. Recently, a meta-analysis, including 11 retrospective studies revealed that earlier surgery at <4 years of age led to better results after BLR [40]. Awadein et al. [101] reported a negative correlation between age at surgery and response to surgery for all angles. Reduction of the surgical dose in patients younger than 7 years of age resulted in no significant surgical outcomes; however, increasing the surgical dose in older patients >12 years of age significantly increased the success rate [101]. They concluded that modifying the surgical dose according to age can improve the surgical outcome of X(T).

In contrast, Kim et al. [38] reported that a younger age at onset and at surgery were risk factors for recurrence in the largest case series of basic type X(T). Lim et al. [25] found lesser recurrence with older age at onset and at surgery after RR. Lim et al. [26] also found lesser recurrence with older age at surgery after BLR.

Jeon et al. [106] reported no difference in terms of surgical success before versus after 4 years of age at surgery. Chia et al. [17] found no difference in surgical outcome among the different age groups.

Although most studies reported that refractive errors do not affect the final surgical outcomes [15,16,18-22], Gezer et al. [107] found a relationship with a shift toward myopia and less favorable surgical outcomes, such as higher myopia with larger postoperative deviation. In contrast, Zou et al. [108] found that a greater degree of myopia and smaller deviations contributed to higher success rates, and that hyperopia and larger deviations resulted in lower success rates. Kim et al. [38] reported that preoperative hyperopia > +2.00 diopters was one of the risk factors of recurrence. Han et al. [44] revealed that preoperative monocular occlusion increased near deviation angle, especially in hyperopes. Pineles et al. [23] found that anisometropia was associated with a poor surgical outcome.

Kim and Choi [98] reported a larger effect of ULR with a narrower tendon width of LR. Lee et al. [109] revealed that the limbus-insertion distance of LR was significantly correlated with dose-response effects. The greater distance of the LR insertion to limbus, the greater is the BLR/ULR effect; for example, 0.2 PD/mm in ULR caused 0.4 PD/mm in BLR at postoperative 6 months. They suggested that the limbus-insertion distance of LR could be a predictor of the recession effect [109].

Differing surgical results have been associated with the changes in the preoperative deviation angle. A larger preoperative deviation angle may be associated with poor surgical outcomes [29,104,107]. Gezer et al. [107] found that the preoperative deviation angle was the most important factor for determining the surgical outcome. Patients with larger preoperative deviations had a poorer successful outcome after a single surgical intervention, while those with smaller preoperative deviation angle had more favorable surgical outcomes. Yam et al. [29] also reported that a smaller angle of deviation correlated with a smaller postoperative drift after BLR. Yang et al. [110] reported that a larger preoperative near deviation angle >16 PD and a larger initial postoperative exodeviation of >5 PD at distance were predictors of recurrence in 92 patients with X(T) ≤25 PD after ULR of 10 mm at more than postoperative 2 years. Kim et al. [38] found that a larger exodeviation of >5 PD at near than at distance was one of the risk factors of recurrence in 560 children with basic type X(T) with an average follow-up of 9.5 years. In contrast, Lim et al. [25,26] found no relationship between the preoperative deviation angle and surgical outcome in both RR and BLR. Chia et al. [17] found no difference in the surgical outcome according to the preoperative degree of control or deviation angle of X(T) at distance. Kim and Kim [45] found better surgical outcomes with variable preoperative measurements than with consistent preoperative measurements.

Postoperative exodrift is common; therefore, an initial overcorrection could be desirable. The desirable initial overcorrection varies in values and has been reported as: 0-9 PD [13], 0-10 PD [9], 5-7 PD [11], 6-10 PD [10], 10 PD [12], 4-14 PD [5], >10 PD [19], 10-15 PD [6,7], 11-20 PD [4,8], and 1-20 PD [14]. In contrast, Leow et al. [111] reported that the exodrift was greater with an initial esotropia and orthophoria compared to that with exotropia. They found that success was unaffected by initial ocular alignment and suggested that deliberate initial overcorrection may be unnecessary.

An initial overcorrection of 17 to 20 PD might be acceptable [4,7,14]; however, risks of consecutive esotropia and monofixation syndrome exist, which can lead to loss of stereopsis and amblyopia especially in young children [13,23]. Kim et al. [15] reported their findings in 68 consecutive patients with initial overcorrection of ≥20 PD at distance or at near. The patients were managed with an alternate full-time occlusion, echothiophate iodide, or prism glasses for the period of overcorrection. An initial overcorrection ≥20 PD decreased to ≤10 PD at distance and at near within postoperative 4 weeks in 72% (49 of 68) of the patients. Four patients (5.9%) needed a reoperation for consecutive esotropia. Finally, orthotropia to exodeviation and/or esodeviation ≤10 PD was achieved in 71% (48 of 68), and exodeviations >10 PD in 16% (11 of 68) of the patients at distance or at near. They concluded that an initial overcorrection ≥20 PD mostly reduced to ≤10 PD within 4 weeks, and reoperations for consecutive esotropia were necessary in 6% of the patients.

Residual exotropia at postoperative 1 week and 1 month led to recurrence [25,26,104]. However, few patients may manifest delayed consecutive esotropia after showing orthotropia to exotropia at postoperative 1 month, especially after BLR [112,113]. Therefore, surgeons should recognize the possibility of delayed development of consecutive esotropia.

Recurrences increase with a longer follow-up period. Ekdawi et al. [114] reported that a misalignment of ≥10 PD increased from 54% at 5 years to 76% at 10 years, and further increased to 86%, 15 years later. Lim et al. [25] reported that the recurrence rate after RR continuously increased from 5.3% at 1 month, 16.0% at 3 months, 27.6% at 6 months, 41.9% at 12 months, 60.7% at 2 years, 70.8% at 3 years, 74.5% at 4 years, and 77.9% at 5 years.

Yang and Hwang [22] compared the efficacy of BMR to UMR for the treatment of recurrent exotropia ≤25 PD after BLR. After BMR resection, 54% (13 of 44) of the patients showed successful outcomes (exodeviation of <10 PD to esodeviation of <5 PD), 42% (10 of 44) showed overcorrection, and 4% (1 of 44) showed an undercorrection at 22.8 ± 10.9 months (range, 8-43 months). After UMR resection, 80% (16 of 20) of the patients showed successful outcomes, 10% (2 of 20) showed overcorrection, and 10% (2 of 20) showed an undercorrection at 21.6 ± 8.8 months (range, 9-43 months). The success rates and incidence of recurrence were not significantly different between BMR and UMR resection; however, the incidence of overcorrection was significantly higher with BMR resection. They concluded that large UMR resections could be safe and effective in treating recurrent exotropia ≤25 PD after BLR with significantly lower overcorrection rate than that with BMR resection. Later, Sung et al. [36] compared the long-term effects of BMR to UMR resection for recurrent exotropia of 20 to 30 PD with a minimum follow-up of 5 years. At 5 years after surgery, the success rates were not significantly different between the BMR and UMR resection (57% vs. 62%); similar findings were observed in the recurrence rates. The overcorrection rate was still significantly higher with BMR resection (35% vs. 15%, p = 0.039). The effect of UMR resection was significantly greater with a larger amount of BLR. Multivariate analysis showed that a large BLR and an initial overcorrection >10 PD were significant risk factors for overcorrection. They concluded that a large UMR resection is still safe and efficient for recurrent exotropia of 20 to 30 PD; however, reduction in surgical dosage for UMR resection in cases with the previous large BLR is necessary considering the high risk of long-term overcorrection.

Initial overcorrections were managed conservatively with monocular patching, full hyperopic correction, bifocal glasses, topical anticholinesterase, botulinum toxin, or base-out prisms [28,115-118]. Kim and Choi [119] reported that the divergence excess type X(T), amblyopia, BLR, esodeviation of ≥20 PD at postoperative day 1, younger age at diagnosis and surgery, and shorter duration from onset to surgery were risk factors for consecutive esotropia. Baik et al. [112] found that 79% of consecutive esotropia were patients who underwent BLR that were orthotropic after 1 month, then later developed esodeviation within 7.4 ± 6.0 months (range, 2-29 months). Lee and Kim [113] also found that 5% (13 of 284) of the patients developed consecutive esotropia after showing orthotropia or exodeviation at postoperative 1 month. They presented a larger preoperative angle of exodeviation, poorer stereopsis, younger age at the time of surgery, and mostly after BLR rather than RR. Chougule et al. [120] found that constant exotropia and larger lateral rectus recession were associated with abduction limitation in consecutive esotropia. Notably, lateral rectus advancement was effective irrespective of abduction limitation.

Lee and Hwang [28] administered prism glasses in 110 patients with consecutive esodeviation of ≥5 PD from 4 weeks after RR. An average period of prism wearing was 21 months, and the esodeviation decreased at the rate of 2.9 PD in 6 months. Thirty-four patients (32%) were weaned off the prism glasses within 1 year. The prism-weaned group showed a larger preoperative constant deviation and anisometropia, and higher exodrift. No patient demonstrated new-onset amblyopia; however, one patient (0.95%) demonstrated loss of stereoacuity because of overcorrection. Finally, successful motor outcomes (exodeviation of ≤10 PD to esodeviation of ≤5 PD) were achieved in 71% of the patients, and the stereoacuity improved or was maintained in 92% of the patients. They concluded that the prismatic correction could lead to good motor outcomes while maintaining favorable sensory status in consecutive esotropia.

Lee et al. [117] administered prism glasses in 39 patients with esodeviation of ≥5 PD from 4 weeks after BLR. Successful motor outcome (exodeviation of ≤10 PD to esodeviation of ≤5 PD) was achieved in 67% of the patients, and stereoacuity was improved or maintained in 79% of the patients. Notably, no patient lost stereopsis because of overcorrection. Prism glasses were necessary for an average of 24 months, and the esodeviation decreased at the rate of 4.2 ± 6.7 PD/yr. At the last visit, 82% of the patients were weaned off prism glasses. They concluded that after BLR, consecutive esodeviation continuously decreased, and 82% of the patients were weaned off prism glasses after 3 years. Prismatic correction could lead to good motor and sensory status in most patients with consecutive esotropia after BLR.

Another option for managing the persistent consecutive esotropia could be botulinum toxin injection with or without electromyographic guidance. Yang et al. [118] reported successful alignment in 69% of the patients with 1.3 ± 0.7 botulinum injections without electromyographic guidance in patients with consecutive esotropia of 21.8 ± 9.1 PD at distance and 21.3 ± 8.3 PD at near. Multivariate analysis revealed that initial postoperative esodeviation of ≤18 PD at 1 month after exotropia surgery was a predictor for success. Vertical deviation and/or ptosis occurred in 8% of the patients and resolved within 3 months. There was no recurrence of exotropia. They concluded that botulinum toxin injection without electromyographic guidance could be safe and effective in the treatment of consecutive esotropia without causing recurrent exotropia, especially with esodeviation ≤18 PD at 1 month after exotropia surgery.

Long-standing consecutive esotropia could be managed with MR recession or LR advancement. Adamopoulou and Rao [121] reported that UMR recession was successful in 63% (10 of 16) of 16 patients with consecutive esotropia after BLR, RR, or BMR resection. X(T) recurred in 31% (5 of 16) of the patients with RR as the initial surgery, and one patient showed undercorrection.

Kim et al. [122] calculated the ratio of the change in muscle position to the change in visual angle deviation, and the surgical dose-effect relationship of lateral rectus advancement in 11 patients with consecutive esotropia. One week postoperatively, the success rate was 82% and the average correction ratio was 4.31 ± 0.96 PD/mm. Multiple regression analysis showed that the amount of muscle advancement and preoperative deviation angle were positively correlated with the correction ratio. Surgical dose-effect relationship increased with the preoperative deviation angle and amount of muscle advancement; therefore, reducing the amount of muscle advancement with larger angle deviations should be considered. Kim and Lee [123] evaluated the efficacy of ULR advancement based on one-fourth of the angle for consecutive esotropia within 25 PD after BLR. Orthotropia and/or exodeviation of ≤8 PD was achieved in 91% (10 of 11) of the patients. The surgical effects of ULR advancement were 3.3 ± 0.7 PD/mm after 1 day, 3.7 ± 0.6 PD/mm after 1 week, and 3.8 ± 0.7 PD/mm after 6 months.

Park et al. [124] compared ULR advancement with MR recession (AR) to BMR recession for consecutive esotropia after BLR. One year later, the success rates were almost similar (86% after AR vs. 73% after BMR recession). Recurrence of X(T) occurred in 14% of the patients after AR versus 27% after BMR recession.