A set of 18 eyes (12 right eyes and 6 left eyes) from 16 patients (2 males and 14 females) were included in our study. Among these subjects, 9 eyes from 9 patients (July 2015-August 2016) were in the ILM insertion group and 9 from 7 patients (September 2016-October 2017) in the NGF group. Subjects receiving different treatment were followed for 6 months. Before conducting the administration of NGF, the clinical data of subjects in ILM insertion group including BCVA and the ELM and EZ defect size were obtained from patient medical diaries and graphical clinical readings throughout the follow-up period and analyzed. After that, we enrolled another 9 eyes to determine the effect of ILM insertion plus NGF on large MHs. Similarly, the follow-up period was 6 months after surgery and the same clinical data were collected and analyzed. Among the 18 recruited eyes, all had large idiopathic MHs (minimum diameter > 400 μm and maximum diameter <900μm with average diameter of 593.1±34.3 μm). All patients were phakic at baseline. The mean age of patients in the ILM insertion group was 63.7 ± 1.5 years, and it was 58.1 ± 3.3 years (95% CI: -2.100 to 13.21) in the NGF group. No between-group difference in age, EZ or ELM defect size, optic axial length, cataract status, BCVA at baseline, or gender was observed (Table 1, chi-squared test). No patient had an open MH at the end of follow-up, showing 100% closure rate in both groups. All patients were pseudophakic at the end of surgery.
Comparison between the ILM insertion versus NGF plus ILM insertion eyes
The BCVA of all patients were improved after surgery. The NGF group exhibited significant improvements in BCVA at both 3rd and 6th months (P=0.0081, 95% CI: 2.159 to 18.29 and P=0.0276, 95% CI: -0.2269 to 18.23 compared to baseline, respectively, one-tailed unpaired t test), but not at 1st month postoperation. The insertion group exhibited significant improvement at 6th month (P=0.0255, 95% CI: -0.05218 to 20.27 compared to baseline, one-tailed unpaired t test). The NGF group exhibited a better BCVA at 3rd month than the insertion group (Figure 1, P=0.0301, 95% CI: -17.01 to -1.224, one-tailed unpaired t test); at other times, although statistical significance was not attained, the BCVA of the NGF group was better than that of the insertion group at 6th month.
OCT revealed significant decreases in the sizes of ELM and EZ defects over time in both groups (Figure 2) (ELM: P<0.0001, <0.0001, and <0.0001, =0.0008, <0.0001, and <0.0001, compared to baseline; 95% CI: -1292 to -653.3, -1508 to -854.1, -1643 to -990.3, -1063 to -299.1, -1260 to -547.4 and -1366 to -641.8, respectively; EZ: P<0.0001, <0.0001, and <0.0001, =0.0031, <0.0001, and <0.0001, compared to baseline, 95% CI: -1317 to -628.4, -1521 to -820.7, -1684 to -994.2, -1091 to -213.8, -1372 to -597.6 and -1420 to -621.2, one-tailed unpaired t test). At 6th month, the mean ELM defect size was 422.2 ± 96 μm in the NGF group and 674.9 ± 103.6 μm in the insertion group (95% CI: -47.26 to 552.8), whereas the mean EZ defect sizes were 496.7 ± 101.6 and 766.7 ± 111.8 μm (95% CI: -50.29 to 590.4), respectively. Significant improvements in EZ and ELM defects were observed in both groups (P<0.0001 and <0.0001 compared to baseline, respectively, as mentioned above). At 6th month postoperation, the mean sizes of ELM and EZ defects were significantly smaller in the NGF than the insertion group alone (P=0.0465 and 0.0464, 95% CI: -47.26 to 552.8 and -50.29 to 590.4). Since the mean sizes of ELM and EZ defects at 1st and 3rd months postoperation were smaller in NGF group, there was no statistical significance between them (ELM: P=0.0704 and 0.0685, 95% CI: -85.12 to 548.0 and -76.90 to 510.9; EZ: P=0.0682 and 0.1804, 95% CI: -95.64 to 639.2 and -172.4 to 447.5, respectively, one-tailed unpaired t test). Two eyes in NGF group and one eye in insertion-only group exhibited complete ELM recovery; one eye in the NGF group and one eye in the insertion group exhibited complete EZ recovery at 6th month; no between-group differences were observed.
All MHs closed (as confirmed by OCT) by 1 month after surgery and remained closed at 6th month (Figures 3 and 4). At 6th month, five of nine foveae in NGF group were U-shaped (55.6%), three were V-shaped (33.3%), and one was W-shaped (11.1%). In the insertion-only group, three of nine foveae were U-shaped (33.3%), four were V-shaped (44.5%) and two were W-shaped (22.2%); there was no significant difference between the two groups (95% CI: -11.49 to 274.8).