Surgeon's view of pars plana vitrectomy for the identification and management of primary rhegmatogenous retinal detachment. Peripheral vitreous dissection with dynamic scleral depression led to identification of the primary break (arrow).
Martínez-Castillo V, Boixadera A, García-Arumí J. Pars Plana Vitrectomy Alone With Diffuse Illumination and Vitreous Dissection to Manage Primary Retinal Detachment With Unseen Breaks. Arch Ophthalmol. 2009;127(10):1297-1304. doi:10.1001/archophthalmol.2009.254
To report on pars plana vitrectomy with diffuse illumination, wide-angle viewing, and meticulous vitreous dissection for identifying and managing retinal breaks when no breaks were detected before surgery for primary rhegmatogenous retinal detachment.
Prospective clinical study of 61 of 800 consecutive eyes (7.6%) (61 of 782 patients) seen at a university hospital during the 48-month study for primary rhegmatogenous retinal detachment in whom no break could be identified preoperatively despite thorough examinations. All of the patients underwent pars plana vitrectomy alone with meticulous peripheral vitreous dissection assisted by diffuse illumination, a wide-angle viewing system, perfluorocarbon liquid, triamcinolone acetonide suspension, and balanced salt solution to identify and manage primary retinal breaks.
Retinal breaks were found intraoperatively in 60 eyes (98%). In 51 of 61 eyes (84%), balanced salt solution was left in the vitreous cavity. Best-corrected visual acuity was 20/40 or better in 25 of 61 study eyes (41%). Primary retinal reattachment was attained in 60 study eyes (98%). Final reattachment was achieved in all 61 eyes (100%).
Pars plana vitrectomy alone with diffuse illumination and extensive vitreous dissection led to identification and management of retinal breaks undetectable before surgery, achieving a high primary reattachment rate.
The principles of surgery for managing primary rhegmatogenous retinal detachment (RRD) are to precisely identify and correctly treat all causative retinal breaks.1 Traditionally, the preoperative examination to find the retinal break has been considered as important as the surgical technique for its management.2 However, even a diligent preoperative examination does not disclose the primary break in 2.2% to 22.5% of cases of primary RRD.3,4
Various strategies have been used to identify and manage primary retinal breaks intraoperatively when they have not been found before surgery, including circumferential buckling, broad retinopexy, scleral buckling, and pars plana vitrectomy (PPV).5- 13 When retinal breaks can be identified intraoperatively, primary reattachment rates are significantly higher than when retinal breaks cannot be identified before or during surgery.5- 12
During recent years, advances in the PPV technique have occurred that allow the surgeon to conduct a detailed intraoperative examination of the peripheral retina and, thereby, identify small retinal breaks located at the vitreous base. The aim of the present prospective study is to determine the success of PPV alone with a diffuse illumination system, dynamic scleral depression (DSD), and meticulous vitreous base dissection in cases in which primary retinal breaks had not been detected during extensive preoperative examinations.
Candidates for this study were 782 patients (800 consecutive eyes) who were scheduled for surgical repair of primary RRD at Vall d´Hebrón Hospital, Barcelona, Spain, by 2 of us (V.M.-C. and A.B.) during a 48-month period. The study protocol was approved by the institutional review board of Vall d´Hebrón Hospital. Patients with redetachment, recurrent retinal detachment, giant tears, retinal detachment due to a macular hole, corneal decompensation, dense cataract, or vitreous hemorrhage were excluded from the study.
Preoperative assessments included evaluation of the anterior segment, measurement of pupil size, fundus examination, and evaluation of the peripheral retina by indirect ophthalmoscopy with scleral depression and slitlamp biomicroscopy using a fundus contact lens, all performed by one of us (V.M.-C.). In 61 candidate eyes (61 patients), no retinal break could be detected preoperatively, and the patient/eye was enrolled in the study. Each patient underwent PPV without an encircling buckle and using a series of maneuvers to identify the retinal break(s) intraoperatively. In phakic eyes, at the beginning of the surgical procedure, the lens was removed by means of phacoemulsification, and a posterior chamber intraocular lens was inserted. All of the patients received acrylic lenses via a clear corneal incision.
Under retrobulbar anesthesia, 3- or 4-port PPV was performed using a wide-angle viewing system and lighted infusion, a 25-gauge sutureless xenon chandelier light (Synergetics USA Inc, O'Fallon, Missouri), or both at the surgeon's discretion. When needed, iris hooks were used intraoperatively to increase mydriasis. A fiberoptic light was used as needed to identify and examine retinal breaks. Eyes were classified according to when in the operative procedures the first retinal break was identified (see the “Patient Groups” subsection). A retinal break was defined as primary when it was judged from the contour of the detachment that this break alone could account for the detachment. A break was defined as secondary when the contour of the detachment could not be accounted for by this break alone.2
The first operative procedures used to identify retinal breaks intraoperatively in all eyes were DSD and peripheral vitrectomy. The DSD was started at the most probable location of the retinal break and then continued along the entire retinal periphery. The surgeon performed all DSDs using a scleral depressor or a muscle hook. Simultaneous vitrectomy and depression were accomplished using diffuse illumination.
In some eyes, a break was identified by noting a primary (spontaneous) Schlieren effect eFigure A).14 In other eyes, the Schlieren effect occurred secondary to peripheral vitrectomy with 360° DSD or another maneuver (eFigure B).
Additional procedures were performed to identify the first break or additional retinal breaks as follows: injection of perfluorocarbon liquid (PFO) (DK Line; Bausch & Lomb Inc, Waterford, Ireland) over the posterior pole and 360° DSD, then injection of 0.5 to 1.0 mL of an aqueous suspension of triamcinolone acetate (TA) (Trigon Depot, 40 mg/mL; Bristol-Myers Squibb SL, New York, New York) prepared as described elsewhere15 and dissection of the vitreous base with 360° DSD.
Eyes were assigned to group 1 when the first retinal break was identified by the initial DSD and peripheral vitrectomy with 360° DSD. Eyes were assigned to group 2 when the first retinal break was not seen until the next step, injection of PFO with 360° DSD. If injection of PFO did not result in any indirect signs, the whole retinal periphery was examined by using DSD to look for direct signs of the location of the retinal break. Retinal breaks could sometimes be identified by a flap or operculum on the slope of the DSD (Figure). Eyes were assigned to group 3 when the first retinal break was not identified until injection of TA and peripheral vitreous dissection at the vitreous base with 360° DSD.
Retinal breaks were managed intraoperatively using different maneuvers to completely drain subretinal fluid from the borders of every retinal break.16 When no retinal break could be identified with the vitreous cavity filled with PFO or air, then circumferential retinopexy was performed using a diode laser, and the vitreous cavity was filled with silicone oil.
After retinopexy of breaks identified while the vitreous cavity was filled with air, the air was exchanged for balanced salt solution and the whole periphery was carefully examined for secondary breaks under scleral depression.16 Follow-up postsurgical examinations were performed at 1, 3, and 7 days; 2 weeks; and 1, 3, 6, and 12 months.
Statistical analysis were performed using a software program (SPSS version 15.0; SPSS Inc, Chicago, Illinois). The χ2 and Fisher exact tests were used to compare categorical data, and the t test was used to compare continuous data. P < .05 was considered statistically significant.
Table 1 details findings for each of the 61 patients (61 eyes) in this study. The mean (SD) patient age was 65 (13.7) years (age range, 19-86 years). Thirty-eight patients (62%) had myopia of less than 6 diopters (D), and 23 (38%) had myopia of 6 D or more. In 3 eyes, the duration of symptoms or detachment was unknown. For the remaining 58 eyes, the mean (SD) duration of retinal detachment was 29.3 (41.8) days (range, 1 day to 6 months); in 36 of these 58 eyes (62%), the duration was less than 2 weeks. At the time of PPV, 11 eyes were phakic (18%), 41 were pseudophakic (67%), and 9 were aphakic (15%). The mean (SD) time from cataract surgery to PPV was 76.8 (86.7) months (range, 1-359 months). For all 61 study eyes, the mean (SD) pupil size was 7 (1) mm (range, 4.0-8.5 mm). The retinal detachment involved 2 quadrants in 12 eyes (20%), 3 quadrants in 21 (34%), and 4 quadrants in 27 (44%). Seven of the 61 patients (11%) were first seen with proliferative vitreoretinopathy (PVR) grade B and 3 (5%) with PVR grade C. The macula was attached in 8 eyes (13%) and detached in 53 (87%). Mean (SD) follow-up was 24.8 (11.1) months (range, 6-46 months).
Table 2 details the characteristics of all 85 retinal breaks identified intraoperatively in the 61 patients/eyes in this study. Of the 60 eyes (98%) in which breaks were identified intraoperatively, 42 (70%) had a single break. Seventy breaks (82%) were horseshoe tears and 15 (18%) were atrophic holes. Of these 85 breaks, 82 (96%) were anterior to the equator, 1 (1%) was at the equator, and 2 (2%) were posterior to the equator. Sixty-four (75%) of the 85 breaks were superior, and the other 21 (25%) were inferior. Of the 85 breaks, 29 (34%) were less than disc diameter, 37 (44%) were ¼ disc diameter, 18 (21%) were ½ disc diameter, and 1 (1%) was greater than disc diameter.
Table 3 summarizes the characteristics of the 42 single retinal breaks identified intraoperatively in 42 eyes with only 1 break according to group (defined by the number of intraoperative maneuvers needed to identify the first break). For the 42 eyes with a single break, group number (the number of procedures needed) was related to the break's size (P = .04) (Table 3).
Table 4 summarizes the characteristics of the 85 retinal breaks identified intraoperatively in the 60 eyes in this study. Of the 35 eyes in group 1, 23 (66%) had a single break and 12 (34%) had multiple breaks. Of the 12 eyes in group 1 with multiple breaks, in 7 (58%) the first retinal break identified was the primary break and in 5 (42%) it was a secondary break.
Of the 60 first retinal breaks identified intraoperatively, 46 (77%) were identified directly and 14 (23%) were discovered through indirect signs. The schlieren effect was observed in 34 eyes (57%); it was primary in 7 eyes (21%) and secondary in 27 (79%).
Of 18 eyes with multiple breaks, in 9 (50%) the first break identified was the primary break and in 9 (50%) the first break identified was a secondary break. In 6 of 18 eyes with multiple breaks (33%), the secondary breaks were located within 2 clock hour positions of the primary retinal break, and in the other 12 (67%), they were not.
When the vitreous cavity was filled with balanced salt solution, in 3 (5%) of 60 eyes a secondary break was identified. At the end of surgery in the 61 eyes, balanced salt solution was left in the vitreous cavity in 51 (84%), air was left in 4, 12% of perfluoropropane gas and air mixture was left in 5, and silicone oil was left in 1. Four of 61 eyes (7%) had vitreous incarceration at the sclerotomy site (patients 4, 25, 44, and 51), and in 1 of the 61 eyes (2%), a retinal break developed at the sclerotomy site.
For all 61 study eyes, the mean preoperative best-corrected visual acuity (BCVA) was 20/125 (range, hand movements to 20/20), and the mean final BCVA was 20/50 (range, hand movements to 20/20) (Table 1). No significant differences were noted among the 3 groups in preoperative BCVA (P = .68) and final BCVA (P = .61). Visual acuity was 20/40 or better in 25 of 61 study eyes (41%). The mean final BCVA was significantly better relative to the mean preoperative BCVA (P < .007).
There were 8 eyes with macula-attached RRD (Table 4), with a mean preoperative BCVA of 20/26 (range, 20/100 to 20/20) and a mean final BCVA of 20/22 (range, 20/40 to 20/20). For the 53 eyes with macula-detached RRD, the mean preoperative BCVA was 20/300 (range, hand movements to 20/25), and the mean final BCVA was 20/60 (range, hand movements to 20/25).
Primary retinal reattachment, defined as complete reabsorption of subretinal fluid at 3 months, was attained in 60 of 61 study eyes (98%; 95% confidence interval, 91.2%-99.9%). The single reattachment failure was the only case in which the retinal break could not be identified intraoperatively. This patient underwent PPV combined with scleral buckling and photocoagulation of the posterior border of the buckle and silicone oil tamponade. This treatment was successful. Thus, we achieved final reattachment in all 61 eyes by the 12-month follow-up visit.
The surgical management of primary RRD when no retinal break can be seen preoperatively has been controversial, particularly when no break can be found despite the presence of clear medium in the vitreous during the surgical technique.5,12 Wong et al5 reported in 1987 on the results of PPV combined with scleral buckling; the causative breaks were identified in 18 of 47 cases (38%), with a reattachment rate of 60%. The authors stated that the undiscovered holes were probably located in the pre-equatorial retina, where the view was poor at vitrectomy. Salicone et al,11 in a retrospective comparative study, did not find the causative breaks in 18 cases treated with PPV combined with scleral buckling. Table 5 summarizes findings in these and other major studies.5- 11 In the present prospective clinical study, we identified the causative breaks in 60 of 61 eyes, and primary reattachment occurred in all 60 of these eyes (98%).
Several factors (described in the following subsections) contribute to the success of this protocol for identifying and managing eyes with primary RRD in which retinal breaks could not be detected preoperatively.
During vitrectomy, peripheral vitreous incarceration at the sclerotomy site is a risk factor for a retinal break. Despite advances in instrumentation, the rate of iatrogenic retinal breaks during PPV has been reported to be 11.6%.17 Using the protocol described herein for the recognition and treatment of vitreous incarceration and meticulous examination of the retinal periphery at the sclerotomy site, we achieved a low incidence of sclerotomy site tears (1.6%).
The RRDs form in a predictable manner around the hole of origin, and the shape of the detachment indicates the position of the primary break 96% of the time.2 The extension of subretinal fluid is governed by the position of the break, the effect of gravity, and anatomical limits. In the present study, the use of independent diffused illumination systems combined with DSD allowed us to accurately establish the limits of detachment and to identify at least 1 retinal break in 35 patients. In some cases, the first break identified was secondary, and in many of these we identified the primary break by following the extension of subretinal fluid to the primary break.
The PFO permits stabilization of the posterior retina and elevation of the peripheral detached vitreous and promotes the Schlieren effect.7 However, despite a meticulous search with DSD after injection of PFO, we identified a retinal break in only 11 of 25 eyes that had a retinal break not identified preoperatively or intraoperatively by means of peripheral vitrectomy and DSD before injection of PFO. We attribute the relatively low proportion of retinal breaks identified using PFO to the small size of the retinal breaks and the small amount of subretinal fluid streaming out through the breaks.
The goal of dissection of the peripheral vitreous after injection of PFO and TA, with DSD, is to create drainage of subretinal fluid through the break. Subretinal fluid drainage promotes identification of previously unseen breaks by means of the Schlieren effect. The DSD may also lead to direct identification of retinal breaks on the slope of the depression or indirect identification through the Schlieren effect (video). Subretinal fluid drainage from the borders of the retinal break permits the surgeon to seal retinal breaks intraoperatively and, thus, to avoid the use of a postoperative tamponade agent.16
As in most studies of how to identify and manage primary retinal breaks not seen before surgery, we limited the study population to eyes with clear media. However, not all previous studies describe the exclusion criteria5,6 or describe other or additional exclusion criteria, such as a preoperative PVR grade of B or C7- 10 or exclusion of macula-attached RRD treated with PPV alone.11
The present study population had a higher proportion of pseudophakic and aphakic eyes compared with populations in previous studies5- 12 of unseen retinal breaks. Small breaks located at the posterior border of the vitreous base are more prevalent in pseudophakic eyes compared with phakic eyes with retinal detachment.18 In these cases, lens status, pupil size, position of the intraocular lens, and status of the posterior capsule make preoperative identification of breaks more difficult, but they did not hinder intraoperative identification of retinal breaks using the present protocol, as shown by the lack of a significant difference in results for lens status, intraocular lens, and posterior capsule status in Table 4.
This study has several findings of importance to surgeons who treat primary RRD in which no retinal break can be seen preoperatively. Most important, the protocol we used to identify retinal breaks intraoperatively that had not been seen preoperatively was effective in finding causative breaks in 60 of 61 eyes and in achieving a final reattachment rate of better than 98%. It is also important that, after we found a single break in 42 of 60 eyes (70%), continuing the search led to the discovery of multiple breaks (2-4) in the remaining 30%. We performed meticulous peripheral vitrectomy to identify retinal breaks when no break could be seen before surgery for retinal detachment, and we found a significant relationship between the need for peripheral vitrectomy with TA (more procedures) to identify a break and smaller size of the break as it appeared to the surgeon when it was finally detected (Table 3).
We found that, when the primary retinal break was undetected at the most probable position according to the distribution of subretinal fluid, it could still be located by means of injection of PFO and DSD (group 2) or by the use of TA suspension to identify the vitreous and peripheral vitrectomy with DSD (group 3) in 98% of eyes (video). We found that the first retinal break identified was the primary break in 44% of eyes. Regarding the location of secondary breaks, in 6 eyes, the secondary break was located within 2 clock hour positions of the primary break, but, in 12 eyes, the secondary break was located farther from the primary break.
In summary, this prospective study showed that retinal breaks that cause primary RRD that were not seen preoperatively can be identified intraoperatively in approximately 98% of patients, achieving a high primary reattachment rate.
Correspondence: Vicente Martínez-Castillo, MD, Calle Londres n 54 4 1 B, Barcelona, Spain (firstname.lastname@example.org).
Submitted for Publication: December 13, 2008; final revision received April 28, 2009; accepted April 29, 2009.
Financial Disclosure: None reported.