Abstract
Purpose
To investigate factors associated with good visual acuity (VA) following repair of rhegmatogenous retinal detachments (RD) with proliferative vitreoretinopathy (PVR) undergoing retinectomy.
Design
Interventional, retrospective, case-control study.
Methods
This single-institution study evaluated patients who underwent retinectomy during repair of RD with PVR from January 1, 2015 to December 31, 2019. A good VA cohort was identified based on a final VA ≥20/70. A 2:1 age-matched and gender-matched poor VA cohort with VA <20/70 was subsequently identified. Metrics compared between the two cohorts included time from primary and recurrent RD diagnosis to surgery, lens status, initial RD size, macula involvement, PVR grade, and size of retinectomy.
Results
A total of 5355 eyes were diagnosed with primary RD during the study period, of which 345 had PVR and underwent retinectomy. The good VA cohort included 62 eyes with a mean final logMAR VA of 0.32 [Snellen 20/42], while the poor VA cohort included 119 eyes with a mean final logMAR VA of 1.54 [Snellen 20/693; P < .0001]. On multivariate analysis, smaller initial RD size ( P = .0090), fewer surgeries ( P = .0002), shorter time between recurrent RD diagnosis and subsequent surgeries ( P = .0006), better preoperative VA ( P = .0276), and pseudophakia at final visit ( P = .0049) remained significant predictors of good vision.
Conclusion
Eyes undergoing retinectomy during repair of RD with PVR can achieve good VA outcomes. The primary modifiable factor associated with better VA was shorter delay between redetachment diagnosis and surgery, particularly in the absence of silicone oil tamponade.
INTRODUCTION
Proliferative vitreoretinopathy (PVR) is the most common cause of failure of rhegmatogenous retinal detachment (RD) repair and affects 5% to 10% of RD cases. Is thought to be a sequalae of a pathologic wound healing response and is characterized by the growth and contraction of preretinal and/or subretinal cellular membranes related to epithelial mesenchymal transformation of retinal pigment epithelial cells, as well as retinal foreshortening due to activation and contraction of intrinsic glial cells. It most commonly develops following surgical RD repair; however, primary PVR may also be found in the setting of chronic primary RD or RD related to trauma.
Surgical management of RD with PVR typically includes pars plana vitrectomy with or without a scleral buckle, membrane peeling, and long-acting gas or silicone oil (SO) tamponade. A retinectomy is often performed if support of a scleral buckle and/or peeling of the preretinal and/or subretinal membranes is insufficient to reduce the tractional forces on the retina, likely due to intrinsic fibrosis of the retina leading to retinal foreshortening. PVR has been associated with worse functional and anatomic visual outcomes, most likely due to significant disruption of the blood retinal barrier and contractional forces that may create new breaks and recurrent detachments. Studies have shown up to 51% of eyes following RD with PVR repair achieve a final visual acuity (VA) of <20/400, and 12% to 16% of eyes achieve a final VA of ≥20/80. ,
Previous studies have identified factors predictive of better outcomes following RD repair for PVR, including preoperative VA, duration of SO tamponade, macular status, and retinectomy size. However, many of these studies have focused on anatomic success, defined as retinal attachment, rather than functional success, defined as good vision. Additionally, the cutoffs of what constitutes good vision vary across studies, and there are limited data comparing eyes with good visual outcomes to those with poor visual outcomes following retinectomy for RD repair.
This case-control study aimed to investigate the preoperative and intraoperative metrics associated with good VA outcomes following RD repair for PVR that included retinectomy at the time of surgery.
METHODS
This retrospective study was approved by the Institutional Review Board at Wills Eye Hospital (Philadelphia, PA). It was conducted in compliance with the Health Insurance Portability and Accountability Act and adhered to the tenets of the Declaration of Helsinki. All eyes that underwent retinectomy during the surgical repair of RDs in the setting of baseline PVR or postoperative PVR at Wills Eye Hospital and Mid Atlantic Retina from January 1, 2015 to December 31, 2019 were identified. These charts were manually reviewed for study inclusion. Eyes were excluded with: uveitis; diabetic macular edema; diabetic retinopathy; retinal vein occlusion; age-related macular degeneration; choroidal neovascularization; optic disc pit maculopathy; glaucoma; colobomas; trauma and/or central serous retinopathy; pediatric vitreoretinopathies, including retinopathy of prematurity and familial exudative vitreoretinopathy; and self-injurious behavior. The minimum follow-up for inclusion was 3 months following retinectomy surgery.
All eyes with a best-available VA based on habitual correction or pinhole of ≥20/70 at the final follow-up visit following surgical repair with retinectomy for RD were identified during the study period and assigned to the good VA cohort. A 2:1 age-matched and gender-matched poor VA cohort of eyes with final VA of <20/70 was then identified consecutively from the same group of eyes with PVR and retinectomy for RD during the same study period. To match the 2 cohorts, each patient’s chart was reviewed for age at time of first retinectomy. Two poor VA patients within 2 years of age at time of retinectomy were then matched to each good VA counterpart and included in the study.
The patient demographic data and baseline data collected included age, sex, race, macula status, VA, intraocular pressure, presence of PVR and grade, lens status, number of breaks, RD extent, RD duration, number of RD surgeries, and days between RD diagnosis to RD surgery. For each RD surgery, the following data points were collected: type of surgery, tamponade, extent of retinectomy, macula status, extent of PVR, and preoperative VA. Days between each surgery and time to surgery from primary and recurrent RD diagnosis were also recorded for each operation. Best-available Snellen VA was collected at the following postoperative time points when recorded: 1 month, 3 months, 6 months, 12 months, and final visit. Snellen VA was converted to logarithm of the minimum angle of resolution (logMAR) VA for statistical analysis. For poor VA values not included in the Snellen chart, counting fingers was converted to 1.98 logMAR, and hand motion, light perception, and no light perception were all converted to 2.28 logMAR. Oil at the final visit and lens status at final postoperative visit were also recorded.
Data were entered into Microsoft Excel 2010 (Microsoft Corporation). JMP software version 15.0 (SAS Institute) was used for statistical analysis. A paired t test was used for within-group comparisons, and the Wilcoxon rank-sum test was performed for between-group comparisons. Group comparisons of the categorical data were performed using the Fisher exact test or Pearson likelihood ratio if more than 2 variables were compared. Clinically and statistically significant metrics were selected for multivariate logistic regression analyses, including: age and gender of patients, size of initial RD, extent of initial retinectomy, retinectomy during initial surgery, number of overall surgeries, preoperative VA, macula status at presentation and subsequent surgeries, time to subsequent surgeries, final lens status, and final SO status for both cohorts. A P- value < .05 was considered to be statistically significant.
RESULTS
During the study period, 5355 eyes were diagnosed with primary RDs, and 828 presented with or later developed PVR. Of the eyes with PVR, 345 underwent retinectomy at the time of surgical repair, with 62 eyes (17.9%) meeting criteria for the good VA cohort, with a mean (standard deviation, SD) final logMAR VA of 0.32 (0.16) [Snellen 20/42; P < .001]. The matched poor VA cohort of eyes that had also undergone retinectomy consisted of 119 patients who had a mean (SD) final logMAR VA of 1.54 (0.59) [Snellen 20/693; P < .001]. Due to the lack of eligible age-matched patients within a 2-year age range of the good VA patients, 119 control patients were found. Two patients in the good VA cohort were unable to be matched with control patients, and 1 patient in the good VA cohort was matched with 1 control patient. For the combined cohorts of 181 patients, the mean (SD) age at presentation was 62.7 (10.2) years, with 54 women (29.8%), and the mean (SD) time from diagnosis to the initial surgery was 4.5 (12.1) days.
The mean (SD) age at presentation in the good VA cohort was 62.4 (10.7) years, with 19 women (31%), and the mean (SD) time from diagnosis to the initial surgery was 5.7 (19) days. The mean (SD) extent of RD at presentation was 5.0 (2.1) clock hours and 24 eyes (41%) were macula-on at the initial visit. The mean (SD) preoperative VA was 0.92 (0.85) logMAR (20/166 Snellen). Postoperatively, mean (SD) VA was 0.41 (0.24) logMAR (20/51 Snellen) at the 6-month visit (n = 50), 0.38 (0.24) logMAR (20/48 Snellen) at the 12-month visit (n = 45), and 0.32 (0.16) logMAR (20/42 Snellen) at the final visit (n = 62). The mean (SD) duration of follow-up was 29.8 (16.5) months (median 26 months, range 5-72 months). For eyes requiring multiple RD repair surgeries, the mean (SD) time to subsequent surgeries was 2.9 (5.7) days from diagnosis of recurrent RD, and 20 redetachments (23.8%) were macula-on at diagnosis. Eyes underwent a mean (SD) of 1.2 (0.9) subsequent RD surgeries during the study period. Three eyes (5%) still had SO tamponade in place at the final visit.
There were 119 eyes in the poor VA cohort (VA <20/70 Snellen). Mean (SD) age at presentation was 62.9 (10.0) years, with 35 women (30%), and the mean (SD) time from diagnosis to the initial surgery was 3.8 (6.0) days. The mean (SD) extent of RD at presentation was 7.0 (3.4) clock hours and 25 eyes (24%) were macula-on at the initial visit. Mean (SD) preoperative VA was 1.39 (0.87) logMAR (20/491 Snellen). Postoperatively, mean (SD) VA was 1.50 (0.61) logMAR (20/632 Snellen) at the 6-month visit (n = 103), 1.56 (0.60) logMAR (20/726 Snellen) at the 12-month visit (n = 87), and 1.54 (0.59) logMAR (20/693 Snellen) at the final visit (n = 119). The mean (SD) duration of follow-up was 35.1 (21.5) months (median 30 months, range 3-135 months). For eyes requiring multiple RD repair surgeries, the mean (SD) time to subsequent surgeries was 5.8 (11.0) days after diagnosis of recurrent RD, and 41 of the redetachments (23.6%) were macula-on. Eyes underwent a mean (SD) of 2.2 (1.3) subsequent RD surgeries during the study period. Fifty eyes (43%) still had SO tamponade in place at the final visit.
When comparing the good VA cohort with the matched poor VA cohort in univariate analysis, no difference was found in the mean age, gender, presence of PVR at baseline, mean time to initial surgery, mean time to first reoperation, macula status at subsequent detachments, or surgical approach ( Table 1 ). The good VA cohort had better mean (SD) logMAR VA at presentation [0.92 (0.85), Snellen 20/166 versus 1.39 (0.87), Snellen 20/491; P = .0007)], a faster mean (SD) time to surgical repair after recurrent RD diagnosis [2.9 (5.7) versus 5.8 (11.0) days; P < .0012], a smaller mean (SD) size of the initial retinectomy [5.2 (1.6) versus 6.2 (2.3) clock hours; P = .0103], initially presented with a smaller mean (SD) extent of RD [5.0 (2.1) versus 7.0 (3.4) clock hours; P = 0.0041], fewer mean (SD) subsequent RD repair surgeries [1.2 (0.9) versus 2.2 (1.3); P < .0001], a lower likelihood of retinectomy at the initial RD repair [7 (11.3%) versus 33 (27.7%); P = .0114], and a higher likelihood of macula-on at initial presentation of RD [24 (41%) versus 25 (24%); P = .0407)]. On multivariate analysis accounting for patient age and gender, size of the initial RD, extent of the initial retinectomy, number of overall surgeries, preoperative VA, macula status at presentation, retinectomy during initial surgery, time from recurrent RD diagnosis to subsequent surgeries, final lens status, and final SO status for both cohorts, the size of the initial RD ( P = .0090), number of overall surgeries ( P = .0002), time between diagnosis of recurrent RD and subsequent surgery ( P = .0006), preoperative VA ( P = .0276), and pseudophakia at final visit ( P = .0049) remained statistically significant between the two cohorts ( Table 2 ).
| Eyes with VA ≥20/70 (n = 62) | Eyes with VA <20/70 (n = 119) | P Value | |
|---|---|---|---|
| Mean age (years) | 62.4 ± 10.7 | 62.9 ± 10.0 | .7563 |
| Females | 19 (31%) | 35 (30%) | .8911 |
| Phakic eyes at baseline | 30 (58%) | 54 (49%) | .3183 |
| Extent of RD at presentation (clock hours) | 5.0 ± 2.1 | 7.0 ± 3.4 | .0041 |
| Macula-on at presentation | 24 (41%) | 25 (24%) | .0407 |
| Mean preoperative VA (Snellen) | 0.92 ± 0.85 (20/166) | 1.39 ± 0.87 (20/491) | .0007 |
| Mean 6-month postoperative VA | 0.41 ± 0.24 (20/51) | 1.50 ± 0.61 (20/632) | < .0001 |
| Mean 12-month postoperative VA | 0.38 ± 0.24 (20/48) | 1.56 ± 0.60 (20/726) | < .0001 |
| Mean final VA | 0.32 ± 0.16 (20/42) | 1.54 ± 0.59 (20/693) | < .0001 |
| Eyes with PVR at first surgery | 17 (30%) | 44 (41%) | .1794 |
| Mean time from initial RD diagnosis to first surgery (days) | 5.7 ± 19.0 | 3.8 ± 6.0 | .0932 |
| Mean time from first recurrent RD diagnosis to reoperation (days) | 2.4 ± 5.1 | 5.2 ± 12.1 | .1414 |
| Mean time from all subsequent RD diagnoses to reoperation (days) | 2.9 ± 5.7 | 5.8 ± 11.0 | .0012 |
| Mean time to subsequent surgery when SO in place (days) | 5.9 ± 6.9 | 7.0 ± 13.1 | .7029 |
| Mean time to subsequent surgery without SO in place (days) | 1.2 ± 1.4 | 4.3 ± 7.4 | .0047 |
| Mean number of subsequent surgeries | 1.2 ± 0.9 | 2.2 ± 1.3 | < .0001 |
| Macula on redetachments | 20/84 (23.8%) | 41/174 (23.6%) | .9681 |
| Eyes with first surgery primary PPV | 48 (77%) | 82 (70%) | .5123 |
| Eyes with first surgery retinectomy | 7 (11.3%) | 33 (27.7%) | .0114 |
| Eyes with SB in place during at least one surgery | 29 (46.8%) | 45 (37.8%) | .2460 |
| Size of initial retinectomy (clock hours) | 5.2 ± 1.6 | 6.2 ± 2.3 | .0103 |
| Eyes with SO at final visit | 3 (5%) | 50 (43%) | < .0001 |
| Eyes pseudophakic at final visit | 56 (93%) | 96 (82%) | .0482 |
| Eyes aphakic at final visit | 0 (0%) | 10 (8.4%) | .0188 |
| Mean follow-up (days) | 894 ± 507 | 1086 ± 654 | .0861 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
