To compare the anatomic and functional results of 3 different epimacular dissection techniques on macular hole surgery.

Vitrectomy was performed in 123 eyes for macular hole in a retrospective, consecutive case series by one surgeon. The surgical technique was identical except for the method of epimacular dissection, which was performed using 1 of 3 techniques sequentially. The first group of 29 eyes had abrasion of the macula with a sharp pick (epiretinal dissection [ERD]), the second group of 32 eyes had partial or complete removal of the internal limiting membrane (ILM), without ILM staining. The third group of 62 eyes had complete removal of the ILM using indocyanine green (ICG) to stain the ILM.

The mean preoperative visual acuity was 20/125 -2 in the ERD and ILM groups and 20/160 -2 in the ICG groups (P = .167). Long-term successful closure of the macular hole was achieved with one operation in 22 of 29 ERD eyes (75.9%), 31 of 32 ILM eyes (96.7%), and 62 of 62 ICG eyes (100%). The mean visual acuity at 3 months was 20/63 +2 (ERD), 20/80 (ILM), and 20/125 (ICG) (P < .001), with significantly poorer visual acuity in the ICG group. The final visual acuity (mean, 2.27 years) was 20/63 +1 (ERD), 20/50 -2 (ILM), and 20/80 -1 (ICG, P = .073), with no significant differences in mean visual acuity, visual acuity gain, gain of ≥3 lines, or percentage ≥20/40.

Removal of the ILM decreases persistent and recurrent macular holes. ICG staining and complete removal of the ILM slows the rate of visual recovery but does not appear to have any long-term deleterious effect on the results of macular hole surgery.