Vitrectomy is one of the major treatment methods for complications of diabetic retinopathy. As surgical techniques and instruments improve, high anatomical success may be achieved; however, functional results are less favorable. Despite attached retina, postoperative visual function may be affected by various macular and disc abnormalities. Major changes include optic atrophy, atrophic retina, poor macular perfusion, and structural alternations of the macula. Funduscopy, fluorescein angiography, and, more recently, optical coherence tomography (OCT) can be used to detect and document these alterations. Although fluorescein angiography is valuable in the assessment of various macular lesions as well as retinal vascular abnormalities, its usefulness in evaluating macular changes is limited in eyes that have undergone diabetic vitrectomy. The quality of the images may be compromised by mild diffuse hyperfluorescence in the posterior pole in middle and late phase angiography, which obscures capillary detail and subtle macular changes. Furthermore, quantification of changes may not be possible. Recently, optical coherence tomography has become very useful for the detection of vitreomacular abnormalities. The examination is noninvasive, can qualify the changes and detect subtle abnormalities not evident with other imaging studies.

In this study, OCT was applied to investigate how common macular structure abnormalities are present after surgery for complications of diabetic retinopathy, and how they may affect visual prognosis.

From June 2008 to December 2008 consecutive patients undergo pars plana vitrectomy for diabetic fibrovascular proliferation were included in the study.

All cases have fibrovascular proliferation with vitreo-retinal adhesions in 3 or more sites but not extending beyond equator in more than one quadrant. Complete removal of the visible posterior preretinal membrane was done during surgery in every case. All cases included in the study had attached retina postoperatively. Cases with rhegmatogenous detachment, silicone oil infusion, or post-operative moderate to dense cataract were not included. Surgical indications included progressive fibrovascular proliferation (FVP), and traction retinal detachment. Postoperative standard 6 mm 6-line OCT (Stratus OCTTM optical coherence tomography, Carl Zeiss Meditec, Inc., Dublin, CA) study centered on the macula and fundus photography were carried out between 7 months to 15 months after surgery. All cases were enrolled from the Department of Ophthalmology, National Taiwan University Hospital, Taipei, Taiwan.

Data on patients' age, gender, systemic diseases, study eye, types and duration of diabetes mellitus, intraoperative diagnosis, and additional surgical procedures such as cryotherapy, scleral buckling [segmental sponge or 360º encircling buckle (Gore-Tex®, polytetrafluoroethylene, W. L. Gore & Associates, Inc., Flagstaff, AZ)], or internal gas (C3F8 or SF6) are recorded. Further data about OCT findings, best-corrected visual acuity at the time of OCT images, and the duration of post-operative follow-up also are compiled. Different morphological appearances of the macula detected by OCT are categorized. These changes are compared between the two surgical subgroups. Correlations of the morphological changes with postoperative visual outcome are conducted.

For the purpose of OCT categorization, measured thickness of more than 300um in the central foveal area and more than 400um in the surrounding macular area in at least 3 examination lines with or without intraretinal cystic spaces or epiretinal membrane is defined as increase of macular thickness. Visible cystic spaces in and immediate adjacent to the fovea is categorized as foveal cysts. The presence of clear optical empty space between pigment epithelium and neurosensory retina are defined foveal detachment.

Statistical analysis In the first step of the statistical analyses, Chi-square test (or Fisher exact test, if n<5) and Kruskal-Wallis test, followed by the Mann-Whitney U test were used to analyze each variant individually for univariate analysis. In the second step, multivariate regression models were used for elucidating the independent influence of each morphological change on postoperative visual outcome. Levels of statistical significance were set at P<0.05.

Inclusion Criteria:

• From June 2008 to December 2008 consecutive patients undergo pars plana vitrectomy for diabetic fibrovascular proliferation are included in our study.
• All cases have fibrovascular proliferation with vitreo-retinal adhesions in 3 or more sites but not extending beyond equator in more than one quadrant.
• Complete removal of the visible posterior preretinal membrane must be done during surgery in every case.
• All cases are included in the study have attached retina postoperatively.

Exclusion Criteria:

• Cases with rhegmatogenous detachment, silicone oil infusion, or post-operative moderate to dense cataract are excluded.
• Patients are unable to follow-up regularly are not included.
• Patient refuses to join our study.

Proliferative diabetic retinopathy