Abstract
Purpose :
There is currently a lack of guidance on how to determine whether patients with diabetic tractional retinal detachment (TRD) should be managed with surgical intervention or observation to optimize visual outcomes. Our study tests the hypothesis that specific clinical characteristics can predict stable versus poor nonsurgical outcomes and good versus poor surgical outcomes.
Methods :
We retrospectively reviewed the data of 560 eyes diagnosed with diabetic TRD at Emory University Hospital, Atlanta, Georgia, USA, between 2012 and 2018. 101 eyes were managed non-surgically (observation, intravitreal avastin [IVA] or panretinal photocoagulation [PRP]) and 136 eyes with pars plana vitrectomy (PPV). 145 eyes were excluded due to loss to follow up. 178 eyes were excluded for TRDs not caused by diabetes. The following data were retrieved: age, A1C, pre-existing DME, history of intravitreal Avastin or PRP treatment. For eyes managed non-surgically, presenting BCVA, and subsequent BCVA at 3, 6, 9 and 12-month follow-up were recorded. For eyes that underwent PPV, presenting and preoperative BCVA as well as postoperative BCVA at 3, 6, 9, and 12-month follow-up were recorded. Dilated fundus exam findings, including presence of vitreous hemorrhage (VH), subretinal fluid (SRF), macular holes (MH), worsening TRD, retinal thinning, iris neovascularization (NVI) and simultaneous rhegmatogenous retinal detachment (RDD), were recorded for both observation and surgery cohorts.
Results :
Out of the 237 eyes that met inclusion criteria, 101 eyes (median age; 56 (interquartile range [IQR]: 48-65), female; 62%, A1C; 8.1) were managed without surgery (75% observation, 21% PRP, 4% IVA) and 136 eyes (median age; 48 (interquartile range [IQR]: 41-55) female; 45%, A1C; 7.9) underwent PPV. 74 eyes of the non-surgical cohort had macula-involving TRDs, but 36% of these eyes subsequently regained 2 or more lines of visual acuity at 1-year follow-up (78% observation only, 15% PRP, 7% IVA). Within the non-surgical cohort, ESRD is more prevalent in patients who lose 2 or more lines of visual acuity at 1-year follow-up (p=0.038). In the PPV cohort, concomitant RDD at presentation is more prevalent in patients with poor visual outcomes post-operatively at 1-year follow-up (p=0.036).
Conclusions :
Observation is indicated for a select portion of patients with diabetic TRD. Concomitant RDD may be a poor prognostic indicator in eyes undergoing PPV.
This is a 2020 ARVO Annual Meeting abstract.