Evidence-Based Ophthalmology: Clinical Trials and Beyond Retinal Detachment and Proliferative Vitreoretinopathy
T. Mark Johnson MD FRCSC
I. INTRODUCTION
The management of rhegmatogenous retinal detachment has evolved over the last century as a result of advances made in surgical techniques and surgical instruments and with the advent of vitreous substitutes.
The “Custodis” method of segmental scleral buckle to seal the retinal break, without drainage of subretinal fluid (SRF), allowing for spontaneous resorption of SRF, the earliest technique, was only suitable for isolated small tears. Encircling scleral buckles offered a more effective, though more invasive, alternative for more extensive detachments and breaks. With the introduction of pneumatic retinopexy by Dominguez,1 followed by the popularization of pars plana vitrectomy (PPV) by Machemer, the strategies for tackling retinal detachments have become more sophisticated, especially with the advent of vitreous substitutes, the use of long-acting gases, and more recently wide-angle viewing systems and high-speed state-of-the-art vitreous cutters. Despite the significant advances in techniques, proliferative vitreoretinopathy (PVR) remains the most common cause of failure for primary rhegmatogenous retinal detachment.
Randomized controlled trials (RCTs) have been designed to ask several important questions regarding retinal detachment management. The question of pneumatic retinopexy as a safe and effective alternative to scleral buckle was addressed by the Retinal Detachment Study group. The Silicone Oil study was designed to test the hypothesis that silicone oil offered an advantage over the various gas tamponades in the management of advanced PVR. The question of vitrectomy versus scleral buckle in primary uncomplicated retinal detachment continues to be controversial; however; new evidence is emerging that may clarify this issue. Randomized trials have been conducted to address this question in a group of patients with pseudophakic and aphakic retinal detachment (PARD).
II. PNEUMATIC RETINOPEXY VERSUS SCLERAL BUCKLE: THE RETINAL DETACHMENT STUDY GROUP
Pneumatic retinopexy was first introduced by Dominguez in 1984 and popularized by Hilton and Grizzard in 1985 for a nonincisional repair of retinal detachment.1,2 Pneumatic retinopexy is based on the principle that an inert long-acting gas, when injected into the vitreous cavity, is capable of sealing a retinal break by positioning the gas bubble against the retinal tear to create an internal tamponade. The surface tension of the gas prevents continued ingress of liquid vitreous, thus allowing the SRF to be naturally absorbed by the Retinal pigment epithelium (RPE) pump. Typically, the fluid that has accumulated under the retina will be reabsorbed within 1 to 2 days depending on the chronicity and the extent of the SRF. Given that gas will disappear from the eye within 1.5 to 6 weeks, it is necessary to create a more permanent seal surrounding the retinal tear.2 The choices in performing retinopexy include laser and cryopexy. Transconjunctival cryopexy can be performed prior to the injection of the gas bubble or on a subsequent day after resolution of the SRF. Laser photocoagulation requires attached retina and thus reabsorption of the
SRF in the area of the break and therefore is performed following gas injection.
SRF in the area of the break and therefore is performed following gas injection.
Sulfur hexafluoride (SF6) and perfluoropropane (C3F8) are the gases most frequently used in pneumatic retinopexy.3 Sterile room air can also be used.4 The type of gas selected is based on surgeon preference, the size of the retinal breaks, the number of breaks, the chronicity of detachment, the ability of the patient to position properly, and the duration of tamponade required. A gas bubble of 0.3 ml covers more than 45° of arc of the retina. To cover 80° to 90°, a bubble of 1.2 ml is required.5 Generally, 1.0 ml is sufficient to cover all breaks simultaneously or alternately. This requires an injection of 0.5 ml of pure SF6 and 0.3 ml of pure C3F8, and if sterile room air is injected, 0.6 to 0.8 ml is recommended. Sterile air, because of the requisite large volume of gas injected, will require a large volume paracentesis or multiple paracenteses to normalize intraocular pressure (IOP) following injection.
Patient selection and compliance are essential for success with pneumatic retinopexy. Patients with back or neck problems may not be ideal candidates. The location of the retinal breaks will determine the position that must be maintained. Breaks between 11 and 1 o’clock are easiest to target. Generally, the break should have tamponade maintained for 3 to 5 days.6 This allows for resolution of the SRF and maturation of the chorioretinal adhesion. Restrictions that must be adhered to while gas is present in the eye include no travel above 4,000 ft and no air travel due to decreases in atmospheric pressure leading to bubble expansion and an unsafe rise in IOP. In addition, patients should not have anesthesia that requires the use of nitrous oxide. Nitrous oxide is more soluble in blood and rapidly diffuses into the vitreous gas bubble, also leading to an unsafe rise in IOP. Phakic patients should also be instructed not to lay flat on their back until the bubble dissipates to avoid prolonged contact with the lens that may accelerate formation of cataract.
Prior to pneumatic retinopexy, the primary operation for repair of retinal detachment had been scleral buckling (SB), with single surgery success rates between 75% and 88%.7,8 However, there had been no RCT to compare the two procedures. The controversy concerning the safety, efficacy, and indications for pneumatic retinopexy led to the conduct of the Pneumatic Retinopexy Study.
Study Objectives
The Pneumatic Retinopexy Study was conducted to determine the efficacy of pneumatic retinopexy in comparison with SB for selected retinal detachments.
Inclusion/Exclusion Criteria
Patients were eligible for the study if they had
Single break no larger than 1 clock hour located in the superior 8 clock hours, or a group of small breaks within 1 clock hour of each other.
Media sufficiently clear to rule out other retinal breaks, determine macular attachment, and not significantly reduce visual acuity.
Availability for follow-up for at least 6 months.
History of good vision before retinal detachment.
Macula-on eyes corrected visual acuity of 20/50 or better.
Macula-off eyes corrected visual acuity of 20/50 or worse.
Shortest diameter of detachment at least 6DD.
Exclusion criteria included the following:
PVR, grade C or D.
Uncontrolled glaucoma or cup-disc ratio exceeding 0.6.
Retinal breaks in inferior 4 clock hours.
Inability to maintain required postoperative head position.
Treatment Groups/Trial Design
Prior to randomization, retinal detachments were stratified into two separate groups:
Macula-on
Macula-off
Outcome Measures
The primary outcome measures were anatomic and functional success following surgical intervention:
Single operation success was strictly defined as retinal reattachment at 6 months after one surgical intervention or injection of gas with one laser and/or cryotherapy performed immediately or within 72 hours.
Important Methodologic Aspects
Pneumatic Retinopexy
Pneumatic retinopexy was performed in accordance with a specific protocol (see below). The type and volume of gas injected, the number of cryopexy or laser photocoagulation applications, paracentesis, IOP at 5, 10, 20, 30, and 60 minutes were noted. The patients were not randomized to the type of gas used for the procedure.
Summary of Protocol
Transconjunctival cryotherapy of retinal break.
Eyelid speculum.
Topical Betadine solution with equal parts balanced salt solution, wait 3 minutes.
Dry injection site 3 to 4 mm posterior to limbus with cotton-tipped applicator.
Briskly inject sterile (millipore filter) C3F8 (0.3 ml) or SF6 (0.6 ml) with a 30-gauge needle in the uppermost pars plana (supine patient with head turned 45° to side; Table 10.1).
Cover conjunctival perforation with sterile cotton-tipped applicator as needle is withdrawn and turn head to move gas bubble away from injection site.
Observe central retinal artery: if artery is closed, wait up to 10 minutes; if artery does not pulsate, use paracentesis or vitreous aspiration.
“Steamroller maneuver,” if indicated, is done at this time.
Monitor IOP and central retinal artery for 60 minutes.
Topical antibiotics and eye pad.
Diamox (250 mg four times daily for 3 days) if patient will drive to a higher altitude not exceeding 4,000 ft.
TABLE 10.1 Expansion and Duration of Intraocular Gases | ||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||
Scleral Buckling
When SB was used, the surgeons were asked to perform the surgery using their usual and customary techniques. The surgeon recorded the type of buckling material, number of cryopexy applications, drainage or SRF, paracentesis, and the type and volume of gas injected.
Follow-up
Follow-up was done on days 1, 3, 7, 14, 30, 60, 120, and 180. Visual acuity was obtained using an Early Treatment Diabetic Retinopathy Study (ETDRS) chart in a masked manner. Refractions were performed at 1 and 6 months after surgery. The macula was examined for holes, pucker, and edema. The peripheral retina was examined for new tears, SRF or blood, PVR, and choroidal detachment.
TABLE 10.2 Summary of the Pneumatic Retinopexy Study12 | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Summary of Major Results
A total of 198 eyes were followed for a minimum of 6 months, 145 (81%) were followed for 1 year.
In the scleral buckle group, most cases were managed with the drainage of SRF and an encircling buckle. In one-third, gas was injected.
In the pneumatic group, most cases were managed with C3F8, and approximately one quarter required paracentesis.
Average cryotherapy applications were similar in both groups.
With one operation, retinal reattachment was slightly higher in the SB group (82% with scleral buckle vs. 73% with pneumatic retinopexy), but the difference was not statistically significant. The addition of postoperative laser photocoagulation or cryotherapy resulted in similar reattachment rates (84% with scleral buckle and 81% with pneumatic retinopexy).
With reoperations, the final reattachment rate was 98% in the SB group and 99% in the pneumatic retinopexy group.
If the detachment did not include the macula, the 6-month final visual acuity was similar for both groups.
If the detachment included the macula for 14 days or less, final visual acuity was significantly better in the pneumatic retinopexy group (p = 0.01); 80% of cases treated with pneumatic retinopexy had better than 20/50 visual acuity compared with 56% with scleral buckle.
Phakic eyes had similar cure rates when treated by either procedure. Aphakic/pseudophakic eyes also had similar success rates by both procedures. However, as a group, aphakic/pseudophakic eyes have a lower cure rate than phakic eyes regardless of the procedure used.
New/missed retinal breaks occurred with significantly greater frequency in the pneumatic retinopexy group.
PVR developed in 5% of the SB group and 3% of the pneumatic retinopexy group. This difference was not statistically significant.
Complications were similar in both groups.
III. INTERPRETATION OF RESULTS AND IMPLICATIONS FOR CLINICAL PRACTICE
Multiple methods exist to successfully repair retinal detachments. Success of surgery is measured by anatomic reattachment and final visual outcome. Pneumatic retinopexy, due to its relatively noninvasive nature, is less likely to be associated with complications including anisometropia and diplopia compared to scleral buckle. Pneumatic retinopexy can restore vision more quickly with lower morbidity than other retinal operations; therefore, in selected patients, it offers certain advantages over other more invasive techniques of retinal detachment repair. The Pneumatic Retinopexy Trial demonstrated that patients with a preoperative macular detachment of less than 2 weeks duration had a significantly better chance of achieving 20/50 or better visual acuity when treated with pneumatic retinopexy compared to scleral buckle.6 This finding has not been found in other retrospective, comparative series where no difference in final visual acuity was noted between scleral buckle and pneumatic retinopexy.9,10
Although single operation pneumatic retinopexy success is desirable because it is associated with the highest level of visual acuity return, the evidence suggests that a failed pneumatic attempt does not disadvantage ultimate anatomic correction of the retinal detachment. In the Pneumatic Retinopexy Study, the single procedure success rate was lower with pneumatic retinopexy compared to scleral buckle; however, the final anatomic success rate was similar.6 Similar findings have been observed in retrospective, comparative studies of pneumatic retinopexy versus scleral buckle.9,10,11 Higher single procedure failure rates with pneumatic retinopexy are ascribed to reopening of the original break, missed retinal breaks, and new retinal breaks. In a retrospective study of 213 eyes undergoing pneumatic retinopexy based on the Pneumatic Retinopexy Study Group inclusion criteria, the single procedure success rate was significantly lower in patients with preoperative vitreous hemorrhage or retinal detachment greater than 4.5 clock hours.12 Mean visual acuity was significantly better in patients achieving retinal reattachment with single procedures (mean visual acuity 20/30) compared to those requiring a secondary procedure (visual acuity 20/60).
Success with pneumatic retinopexy depends upon case selection and surgical technique. Most favorable cases include phakic eyes with less extensive detachment, secondary to a superior retinal break less than 1 clock hour in size, and no PVR. Retrospective series of pneumatic retinopexy suggest that patients with a single retinal break and a retinal detachment in the superior two-thirds of the fundus have a single procedure success rate as high as 97%.11 Factors negatively influencing single operation anatomic success include pseudophakia, an increased number of retinal breaks, and a greater area of detached retina. Factors not influencing outcome include the presence of lattice degeneration (less than 3 clock hours), the type of retinal break, the type or volume of gas used, the type of retinopexy (laser or cryotherapy), the sequence of gas insertion versus retinopexy application, the status of the posterior capsule, and gender.
With increased attention to health care costs, the ability to treat retinal detachments with a minimally invasive, office-based procedure may make pneumatic retinopexy increasingly important in management. Estimates have suggested that pneumatic retinopexy may cost 25% to 50% less than SB surgery when operating room and anesthesia costs are considered.11
Controversies and Future Use of Pneumatic Retinopexy
Increased familiarity and comfort with pneumatic retinopexy has led to expanded usage of this technique in the management of retinal detachment. Technique modifications have been suggested to improve the outcomes of pneumatic retinopexy.
Tamponade during pneumatic retinopexy is largely determined by surgeon preference. Options include sterile air, SF6, and C3F8. There are limited data comparing tamponade agents’ effects on procedure success rate. A randomized, noninferiority trial has been conducted comparing sterile air to C3F8.13 This study demonstrated a single procedure success rate of 60% in the sterile air group and 73% in the C3F8 group, which was not a statistically significant difference. The final reattachment rate was 92% in the air group and 96% in the C3F8 group, with similar final visual acuities suggesting that sterile air is a reasonable alternative to C3F8. Of note in this study, 0.3 cc of tamponade was injected in each group, which is a lower volume than what many practitioners have advocated for procedures using sterile air.
Inferior retinal detachment was initially considered to be an exclusion for treatment with pneumatic retinopexy.6 Inverted positioning required to tamponade the inferior retinal breaks was considered to be impractical. In addition, concerns have arisen regarding the practicality of prolonged inverted positioning required to achieve adequate reabsorption of SRF and chorioretinal adhesion. Inverted pneumatic retinopexy had been previously used to successfully reattach the retina following recurrent retinal detachment after scleral buckle.14 Recent case series have revisited and
expanded the role of inverted pneumatic retinopexy. In one series of recurrent inferior retinal detachment following encircling scleral buckle, 17 patients underwent inverted positioning.15 Positioning was achieved with 10° Trendelenburg, 10° neck extension, and 10° ocular supraduction. Tamponade was achieved with injection of 0.3 to 0.8 ml of intraocular gas. Patients maintained strict positioning for 48 hours and part-time positioning for 1 week; 88% of patients achieved lasting retinal reattachment with a median follow-up of 1.3 years (0.1 to 11.5 years). A second case series of 11 patients, including 5 primary inferior retinal detachments, achieved an 82% single procedure success rate with inverted pneumatic retinopexy.16 Patients were positioned with their head dependent in a prone position for 8 hours. No further positioning was required. This series suggests that limited positioning in selected patients may allow inferior retinal reattachment. No comparative trials exist to determine the true efficacy of inverted pneumatic retinopexy in a larger population versus scleral buckle.
expanded the role of inverted pneumatic retinopexy. In one series of recurrent inferior retinal detachment following encircling scleral buckle, 17 patients underwent inverted positioning.15 Positioning was achieved with 10° Trendelenburg, 10° neck extension, and 10° ocular supraduction. Tamponade was achieved with injection of 0.3 to 0.8 ml of intraocular gas. Patients maintained strict positioning for 48 hours and part-time positioning for 1 week; 88% of patients achieved lasting retinal reattachment with a median follow-up of 1.3 years (0.1 to 11.5 years). A second case series of 11 patients, including 5 primary inferior retinal detachments, achieved an 82% single procedure success rate with inverted pneumatic retinopexy.16 Patients were positioned with their head dependent in a prone position for 8 hours. No further positioning was required. This series suggests that limited positioning in selected patients may allow inferior retinal reattachment. No comparative trials exist to determine the true efficacy of inverted pneumatic retinopexy in a larger population versus scleral buckle.
The lower single procedure success rate observed with pneumatic retinopexy has been ascribed to the frequent development of new retinal breaks. The majority of these breaks occur within 1 month of the procedure.6 The majority will occur in the superior fundus, often in relative proximity to the initial retinal break. It is postulated that the gas bubble may shift the vitreous, leading to new areas of vitreo-retinal traction. One attempt to reduce the rate of new retinal breaks is the application of 360° laser retinopexy. In one retrospective case series, prophylactic laser was suggested to reduce the rate of new retinal breaks.17 These findings have not been validated in prospective studies nor have the potential complications of extensive laser been fully explored.
The popularity and comfort of surgeons with PPV has prompted many to treat primary uncomplicated retinal detachment with primary vitrectomy. A RCT to compare pneumatic retinopexy to primary pars plana vitrectomy (PPPV) would help answer these questions further and elucidate whether the risk of an operating room procedure with its added cost is justified by the potential benefits of earlier rehabilitation, enhanced primary success, and faster and more complete visual recovery.
IV. VITRECTOMY VERSUS SCLERAL BUCKLE FOR PRIMARY RHEGMATOGENOUS RETINAL DETACHMENT
Introduction
The surgical choice of treatment for patients with primary retinal detachment uncomplicated by PVR remains controversial. Traditionally, the initial management of retinal detachment has been scleral buckle. Increased experience with vitrectomy, improvements in surgical instrumentation, the advent of high-speed cutters, and the introduction of wide-field viewing systems have led to an increased utilization of vitrectomy in the management of primary retinal detachment. Medicare data in the United States indicates an 80% increase in the use of vitrectomy to repair retinal detachment and a 70% decrease in SB since 1997. Potential advantages of primary vitrectomy include removal of vitreous opacities and capsular remnants, possibly faster and increased rate of foveal reattachment in macula-off retinal detachments, and avoidance of complications associated with SB including refractive shifts, extraocular muscle imbalance, and buckle extrusion.18,19
Data published to date suggest that vitrectomy compares favorably with SB. The two primary outcome measures of success in retinal detachment repair cited in most studies are anatomic retinal reattachment and visual acuity. The overall retinal reattachment rate for PPV in a recent review was 85%, compared to 71% to 95% reattachment rate achieved in retrospective reports of SB procedures.20,21
Evaluation of the literature to determine the true efficacy of primary vitrectomy compared with SB is difficult for several reasons. There is a lack of uniform inclusion criteria in the studies, including different configurations of retinal detachment, duration of detachment, and preoperative lens status that could significantly influence the results.22,23 While the bulk of the literature on the primary repair
of retinal detachment is composed of case series, there are several comparative trials; however, in not all of these studies were the subjects truly randomized, and thus, selection bias may influence the stated results.24 In addition, many randomized studies lack an a priori sample size calculation and thus it is difficult to determine whether the study enrollment had adequate power to detect a true difference in treatment efficacy. Duration of follow-up varies between studies. In some cases, a lack of long-term follow-up makes the results of the studies difficult to compare.
of retinal detachment is composed of case series, there are several comparative trials; however, in not all of these studies were the subjects truly randomized, and thus, selection bias may influence the stated results.24 In addition, many randomized studies lack an a priori sample size calculation and thus it is difficult to determine whether the study enrollment had adequate power to detect a true difference in treatment efficacy. Duration of follow-up varies between studies. In some cases, a lack of long-term follow-up makes the results of the studies difficult to compare.
In general, vitrectomy has been compared to SB in two separate groups: pseudophakic/aphakic retinal detachment and phakic retinal detachment. In addition, some literature has examined the role of primary vitrectomy compared to combination vitrectomy/scleral buckle.
Primary Vitrectomy versus Scleral Buckle in Primary Pseudophakic and Aphakic Retinal Detachment
Capsular opacities, poor dilation, vitreous debris, and the presence of small retinal breaks have been cited as reasons for failure of primary scleral buckle in cases of PARD. Primary vitrectomy has become increasingly popular in the management of these cases due to the ease of improving visualization of small retinal breaks and the lack of induced myopia secondary to the presence of a scleral buckle. In addition, these cases are not subject to the primary complication of vitrectomy, cataract.
Several case series have been conducted examining the role of primary vitrectomy alone in the management of PARD.24,25,26,27,28,29 These series report a primary retinal reattachment success rate ranging from 88% to 94% and a final reattachment rate of 96% to 100%. The rate of final visual acuity better than 20/50 is reported to be 69% to 79%.
Nonrandomized, comparative series have studied vitrectomy versus scleral buckle in patients with pseudophakic retinal detachment.29 Similar primary and final retinal reattachment rates as well as visual acuities were reported. A meta-analysis of 29 published studies on the management of pseudophakic retinal detachment from 1966 to 2004 demonstrated a higher single procedure reattachment rate with vitrectomy alone (odds ratio [OR] 1.69 [1.07-2.68]) or combined PPV/SB (OR 3.54 [1.57-7.97]) compared with scleral buckle alone.26
Two RCTs have compared primary vitrectomy with scleral buckle in the management of pseudophakic retinal detachment.30,31 One study was a single-center trial conducted with a single surgeon.30 The other was conducted as a multicentered trial.31 Both studies benefited from clear inclusion and exclusion criteria, a priori sample size calculations to ensure adequate statistical power, and a defined randomization schedule.
In the single-center RCT, 150 patients with pseudophakic retinal detachment were randomized to an encircling silicone scleral band (240 style, 2.5 mm) versus a conventional 20-gauge three-port vitrectomy, retinal reattachment with perfluoro-n-octane (PFO), and gas tamponade with 20% SF6.30 Results from this study indicated that vitrectomy was associated with a significantly higher single procedure anatomic reattachment rate (94% with PPV vs. 83% with SB). Operative time was significantly shortened with vitrectomy. The number of unidentifiable retinal breaks was significantly higher in the scleral buckle group, which partially accounted for the better observed single procedure success rate of vitrectomy. Final retinal reattachment rates were similar in both groups (95% in the SB group vs. 99% in the PPV group). No difference in final visual acuity was observed. Axial length was significantly increased in the scleral buckle group postoperatively.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
