This study was a non-randomized, retrospective review of the medical records of 40 patients diagnosed with SO palsy at Kim's Eye Hospital between 2010 and 2015. SO palsy was defined as positive in the three-step test. Patients with a history of blow-out fracture and surgeries of the strabismus, simultaneous horizontal muscle, or vertical rectus muscle were not considered for the study. Definition of congenital SO palsy was as follows: manifested in childhood, larger than normal vertical fusion amplitude, and/or presence of facial asymmetry. Cyclotropia, head tilt toward the involved side, or absence of abnormal head posture limits the diagnostic criteria of acquired SO palsy. The research was approved by the institutional review board at Kim's Eye Hospital (2017-05-002-001), and according to the institutional review board standard operating procedures on retrospective study, informed consent was not required for this study. This study also conducted in accordance with the principles of the Declaration of Helsinki.

Demographic characteristics of patients included in the study are presented in Table 1. Patients who underwent IO recession were divided into two groups for comparison studies: group 1 (10-mm IO recession, 15 patients) and group 2 (14-mm IO recession, 25 patients). One surgeon performed all surgeries. Preoperative and postoperative horizontal deviations were measured from a distance by the alternating prism cover test with correction. Postoperative horizontal deviation was measured at 6 months after surgery, and the resulting horizontal deviation from the10- and 14-mm IO recession surgeries were compared. To analyze the effect of combined exodeviation, we conducted subgroup analysis of the patients with preoperative exodeviation greater than 8 PD. In addition, the effects of superior oblique underaction and IO overaction on postoperative horizontal deviation were investigated.

Data were analyzed using PASW Statistics ver. 18.0 (SPSS Inc., Chicago, IL, USA). Statistical analyses of demographics and horizontal deviations between the two groups were performed by the chi-square and Mann-Whitney test, respectively. Wilcoxon's signed rank test was used to evaluate the change between preoperative and postoperative deviation.

Forty patients (22 men and 18 women) with a mean age of 24.0 ± 18.7 years were enrolled in the present study. Preoperative horizontal exodeviation was recorded as 6.0 ± 8.7 and 3.0 ± 4.5 PD in group 1 and group 2 (p = 0.159) patients, respectively. Postoperative horizontal exodeviation was found to be 4.1 ± 5.8 and 0.8 ± 2.1 PD in group 1 and group 2 (p = 0.012) patients, respectively (Table 2). The total change of horizontal exodeviation after surgery was noted as 1.9 ± 4.5 PD (p = 0.452) in group 1 patients and 2.2 ± 3.8 PD (p = 0.022) in group 2 patients (Fig. 1A). As a result, 14-mm IO recession demonstrated a significant esodrift after surgery.

Among the 40 patients that participated in the study, 10 had a preoperative exodeviation over 8 PD. The preoperative horizontal deviation in these patients was determined to be 15.0 ± 4.5 and 12.0 ± 2.3 PD in group 1 and group 2 (p = 0.205), respectively. The postoperative horizontal deviation in these 10 patients was calculated to be 9.0 ± 5.2 and 3.0 ± 3.8 PD for group 1 and group 2 (p = 0.06), respectively (Table 3). Both groups showed a significant change in horizontal deviation (group 1, p = 0.004; group 2, p = 0.015). The mean change in horizontal deviation was 6.0 ± 5.4 PD for group 1 patients and 9.0 ± 5.0 PD for group 2 patients (Fig. 1B). Although the amount of SO underaction did not affect the extent of change in horizontal deviation, patients with severe IO overaction showed a significant change in horizontal deviation after 14-mm IO recession (Fig. 2A–2D).

The study demonstrated that a 14-mm IO recession could initiate a significant esodrift, especially in patients with exotropia. In contrast to 10-mm IO recession, the 14-mm IO recession resulted in a statistically significant change in horizontal deviation after surgery. In addition, patients with a preoperative exodeviation greater than 8 PD showed a significant postoperative esodrift.

Esoshift resulting from IO weakening surgery can be caused by different surgical procedures. Chang and Yang [8] presented a 4.4-PD esoshift after IO myectomy or Z-myotomy. Stager and Parks [6] reported that only 16% of patients with straight eyes undergoing weakening procedures for IO overaction showed a postoperative horizontal alignment change of 4 PD or higher, and no patient reported a change greater than 8 PD. Based on the findings of these two reports, the mean esoshift value might be estimated to be around 4 PD, slightly lower than those recorded in the current study (i.e., 1.9 and 2.2 PD). The studies by Stager and Parks [6] and Chang and Yang [8] included IO myectomy and Z-myotomy and IO myectomy and 8-mm IO recession, respectively. However, they did not compare each effect according to surgical procedure. As a result, we could not conduct a direct comparison between the results of these studies and those obtained here.

However, a relationship could be established between the state of oblique function and surgical outcomes. SO underaction might affect the results of horizontal deviation after IO weakening surgery [6]. However, no significant difference was found based on SO action. A notable observation in the current study was that the patients with severe IO overaction demonstrated a significant esoshift after surgery. Therefore, the state of IO might affect the esoshift after IO weakening surgery.

We showed a significant esodrift in patients with exodeviation greater than 8 PD. Tommila and Valle [9] reported a significant esodrift after IO recession in approximately 50% of patients in their study. However, several reports have suggested that horizontal deviation is not affected when IO weakening and exotropia surgeries are performed in combination [1011]. This discrepancy might result from different inclusion criteria and surgical techniques.

We acknowledge that this study has certain limitations. First, it was a retrospective review and not a randomized controlled analysis, which could have resulted in a bias in patient selection. Congenital SO palsy and acquired SO palsy have different characteristics, and this could affect the postoperative horizontal deviation. In addition, the 10- and 14-mm recession were not randomly assigned but were applied according to the status of patients. Thus, a randomized, controlled study is needed. Second, the small number of patients enrolled in the study could impact the precision of subgroup analysis. Finally, as in Fig. 1, although the 14-mm IO recession group had significant postoperative horizontal changes, a change of 2 PD may not have much clinical impact. Thus, when preoperative exodeviation is small, the postoperative effect on horizontal deviation might be meaningless. However, the results of this study highlight a consistent decrease in exodeviation after IO weakening surgery in both study groups.

In conclusion, surgeons should consider changes in horizontal deviation when opting for 14-mm IO recession in the treatment of superior oblique palsy, especially in patients with preoperative horizontal deviation greater than 8 PD or severe IO overaction.