Refractory full thickness macular hole: current surgical management

Abstract

This review aims to collect the proposed surgical techniques for treating full thickness macular hole (FTMH) refractory to pars plana vitrectomy and internal limiting membrane (ILM) peeling and to analyse and compare anatomical and functional outcomes in order to evaluate their efficacy. The articles were grouped according to the surgical techniques used. Refractory FTMH closure rate and best-corrected visual acuity (BCVA) gain were the two analysed parameters. Thirty-six articles were selected. Ten surgical technique subgroups were defined: autologous platelet concentrate (APC); lens capsular flap transplantation (LCFT); autologous free ILM flap transplantation (free ILM flap); enlargement of ILM peeling, macular hole hydrodissection (MHH), autologous retinal graft (ARG), silicon oil (SO), human amniotic membrane (hAM), perifoveal relaxing retinotomy, arcuate temporal retinotomy. Refractory FTMH closure rate was similar among subgroups, not significant heterogeneity emerged (p = 0.176). BCVA gain showed a significant dependence on surgical technique (p < 0.0001), significant heterogeneity among subgroups emerged (p < 0.0001). Three sets of surgical technique subgroups with a homogeneous BCVA gain were defined: high BCVA gain (hAM); intermediate BCVA gain (APC, ARG, LCFT, MHH, SO); low BCVA gain (free ILM flap, enlargement of peeling, arcuate temporal retinotomy). In terms of visual recovery, the most efficient technique for treating refractory FTMH is hAM, lens capsular flap and APC that allow to obtain better functional outcomes than free ILM flap. MHH, ARG, perifoveal relaxing and arcuate temporal retinotomy require complex and unjustified surgical manoeuvres in view of the surgical alternatives with overlapping anatomical and functional results.

Fig. 1. Postoperative tomographic foveal patterns.

Tomographic foveal patterns after surgically induced closure of full thickness macular hole (FTMH). a Restoration of foveal contour (dashed line) with integrity of outer retinal layers (ORL) (white arrows) equivalent to Imai’s “type U” and Kang’s “type 1”. b Thinned foveal floor with irregular foveal contour (dashed line) and interruption of ORL (white arrows), equivalent to Imai’s “type V” and Kang’s “type 1”. Patterns of full thickness macular hole (FTMH) refractory to pars plana vitrectomy and internal limiting membrane (ILM) peeling. c Refractory FTMH with elevated edges (dashed line) and bared retinal pigment epithelium (RPE) (white arrow), equivalent to Hillenkamp’s “type with cuff”. d Refractory FTMH with flat edges (dashed line) and bared RPE (white arrow), equivalent to Imai’s “type W”, Kang’s “type 2” and Hillenkamp’s “type without a cuff”. Tomographic images from the database of the department of Ophthalmology of University of Padova, Italy.

Fig. 2. Flow chart shows detailed information on the number of articles screened (No. 412), assessed for eligibility and excluded or included for review.

Thirty-six articles were divided on the basis of the proposed surgical technique into ten subgroups.

Fig. 3. Forest plot from meta-analysis of weighted closure rate of refractory full thickness macular hole (FTMH) for all articles.

No significant difference in refractory FTMH closure rate among the 36 articles was detected.

Fig. 4. Funnel plot of refractory full thickness macular hole (FTMH) closure rate.

Symmetrical funnel plot suggests a low grade of heterogeneity of FTMH closure rate between the 36 analysed articles.

Fig. 5. Forest plot from meta-analysis of best-corrected visual acuity (BCVA) gain among the 32 analysed articles.

Significant difference in BCVA gain among the surgical technique subgroups was demonstrated.

Fig. 6. Funnel plot of best-corrected visual acuity (BCVA) gain.

Asymmetrical funnel plot suggests the high grade of heterogeneity of BCVA gain among the analysed articles. Thirteen articles fall outside the CI95% corresponding to the circles with ID number.