- Email: [email protected]
Diabetic Vitrectomy
Diabetic Vitrectomy Influence of Lens Status upon Anatomic and Visual Outcomes William M. Schiff, MD,1 Gaetano R. Barile, MD,1 John C. Hwang, MD, MBA,1 Joseph J. Tseng, MD,1 Osman Çekiç, MD, PhD,2 Lucian V. Del Priore, MD, PhD,1 Stanley Chang, MD1 Objective: To determine the effect of lens status upon the anatomic and visual results in primary diabetic vitrectomy. Design: Retrospective, comparative, consecutive case series. Participants: One hundred two eyes of 85 patients with proliferative diabetic retinopathy and its complications that underwent primary vitrectomy. Methods: The eyes that remained phakic after vitrectomy were compared with the eyes that were either aphakic or pseudophakic (nonphakic) postoperatively. Main Outcome Measures: Intraoperative and postoperative complications, vitreoretinal reoperation rate, and ultimate anatomic and visual success with at least 6 months’ follow-up. Results: Preoperatively, 72 eyes were phakic, and 30 were aphakic (n ⫽ 1) or pseudophakic (n ⫽ 29). During vitrectomy, 1 eye underwent lensectomy and 12 eyes underwent phacoemulsification with lens implantation. Postoperatively, 59 eyes were phakic and 43 eyes were nonphakic. The vitreoretinal reoperation rate was significantly higher (P ⫽ 0.04) for the phakic group (28.8%) than for the nonphakic group (11.6%). Rubeosis iridis developed in 3 phakic eyes and no nonphakic eyes (P ⫽ 0.26). Intraoperative complications were similar in the phakic and nonphakic groups (P ⫽ 0.40). Postoperative complications such as rhegmatogenous retinal detachment (P ⫽ 0.39), nonclearing vitreous hemorrhage (P ⫽ 0.07), and anterior chamber complications (P ⫽ 0.60) were also similar. Visual acuity improved by at least 0.2 logarithm of the minimum angle of resolution units in 76.2% of the phakic eyes and 86.0% of the nonphakic eyes (P ⫽ 0.22). Conclusions: Eyes that were phakic after primary diabetic vitrectomy had a significantly higher subsequent vitreoretinal reoperation rate when compared with nonphakic eyes, suggesting that diabetic eyes are less likely to require additional vitreoretinal surgery if they are rendered nonphakic before or during vitrectomy. Ophthalmology 2007;114:544 –550 © 2007 by the American Academy of Ophthalmology.
The surgical management of complications due to diabetic retinopathy often requires vitrectomy. The most common complication of pars plana vitrectomy (PPV) is the development of nuclear sclerotic cataract.1,2 Visually significant cataract development in the phakic eye after PPV is quite common, with rates reported between 17% and 100%.1–7 When intraocular postoperative temporary vitreous tamponade is used in conjunction with PPV, this rate increases; in Originally received: January 4, 2006. Accepted: August 11, 2006. Manuscript no. 2006-34. 1 Department of Ophthalmology, Columbia University College of Physicians and Surgeons, Edward S. Harkness Eye Institute and St. Luke’s– Roosevelt Hospital Center, New York, New York. 2 Department of Ophthalmology, Suleyman Demirel University Medical School, Isparta, Turkey. Supported by an unrestricted research grant from Research to Prevent Blindness, New York, New York. The authors have no commercial, proprietary, or financial interest in any of the products or companies described in the article. Correspondence and reprint requests to William M. Schiff, MD, 635 West 165th Street, New York, NY 10032. E-mail: [email protected].
544
© 2007 by the American Academy of Ophthalmology Published by Elsevier Inc.
cases of silicone oil injection, the rate is reported to be as high as 100%.7 Management of the crystalline lens in diabetic patients undergoing vitreous surgery has long been a source of controversy. Initially, due to intraoperative lens opacification and diminished visualization, lens removal via the pars plana was often combined with vitrectomy in the diabetic eye.4,8 Due to a well-recognized increased incidence of postoperative anterior segment neovascularization and observations suggesting a protective barrier effect of the crystalline lens in reducing the risk of progression of diabetic retinopathy and anterior segment neovascularization, the posterior segment surgeon later attempted to spare the crystalline lens (lens-sparing vitrectomy).9 –13 More recently, the anatomic and visual results of vitrectomy for severe proliferative diabetic retinopathy (PDR) have improved due to a better understanding of the pathoanatomy of the disease and improvements in surgical instrumentation and laser delivery.14 In this regard, many surgeons have reported on combined PPV with simultaneous lens removal and intraocular lens (IOL) implantation.15–21 Lens removal at the time ISSN 0161-6420/07/$–see front matter doi:10.1016/j.ophtha.2006.08.017
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status of vitrectomy eliminates the need for subsequent cataract surgery and may allow for more rapid visual recovery.15,16,19,21 Such studies suggest that the use of modern vitreoretinal techniques, including intraoperative laser photocoagulation, may have rendered the rationale for preserving the native lens during diabetic vitrectomy less compelling. In addition to becoming cataractous and impeding visual recovery after vitrectomy, the crystalline lens may serve as an anatomic impediment to intraoperative surgical maneuvers during PPV.15 Adequate epiretinal membrane dissection and/or thorough and careful removal of the basal vitreous gel in the vicinity of the sclerotomy ports may be compromised in the phakic eye. The nonphakic eye may facilitate further the recognition and management of intraoperative vitreoretinal complications such as retinal tears, dialysis, and rhegmatogenous retinal detachment (RD). Additionally, postoperative persistent or recurrent vitreous hemorrhage may clear less efficiently in the phakic eye. For these reasons, we studied the influence of lens status at the end of vitreous surgery on anatomic and visual outcomes in diabetic vitrectomy. Specifically, we performed a retrospective analysis of complication rates, rates of reoperation, anatomic success, and visual outcomes in the diabetic eye with complications of PDR after lens-sparing versus nonphakic (or combined) primary PPV.
Materials and Methods The study is a retrospective chart review of diabetic retinopathy patients who underwent primary PPV for complications relating to diabetic retinopathy between April 1999 and March 2005 at Columbia University Medical Center’s Edward S. Harkness Eye Institute and St. Luke’s–Roosevelt Hospital Center. The institutional review boards of New York–Presbyterian Hospital and St. Luke’s–Roosevelt Hospital Center approved this study. Patients included in this study had no prior retinal or vitreous surgery aside from office laser photocoagulation for clinically significant diabetic macular edema or PDR. Indications for vitrectomy included recurrent or persistent nonclearing vitreous hemorrhage, traction or combined traction/rhegmatogenous RD, and adherent posterior hyaloid causing excessive macular traction or macular edema. Patients were excluded from the study if they were younger than 18 years, had a history of uveitis in the operated eye before vitrectomy surgery, had preexisting rubeosis iridis, had previous retinal or vitreous surgery, or had less than 6 months’ follow-up after surgical intervention. The following preoperative information was obtained for each patient enrolled: age; gender; duration of diabetes; Snellen visual acuity (VA); applanation tonometry; lens status and grading of crystalline lens; presence or absence of rubeosis iridis; anterior segment abnormalities (i.e., posterior synechia and optic/iris capture); and indication for vitreoretinal surgery, including preoperative retinal drawings. The eyes that remained phakic after PPV were compared with the eyes that were nonphakic (aphakic or pseudophakic) with respect to severity of retinopathy, intraoperative complications, postoperative complications, postoperative anatomic and visual success, and total number of vitreoretinal reoperations. Patients were operated on by 1 of 4 surgeons (GRB, SC, LVDP, WMS). All patients underwent standard 3-port PPV on their affected eye. For patients who underwent phacoemulsification and IOL implanta-
tion, an experienced anterior segment surgeon performed the procedure, and a foldable IOL with a 6.0-mm optic was placed in the capsular bag via a clear corneal incision. A 10-0 nylon suture was used to secure the phacoemulsification wound. During PPV, membrane dissection and segmentation was performed when necessary to remove all tangential traction. Under conventional plano-convex and wide-field panoramic contact lenses, standard bimanual delamination, en bloc dissection, and segmentation techniques were performed. A variety of vitreoretinal scissors, forceps, and a tissue manipulator were utilized when appropriate to remove as much fibrovascular tissue as deemed necessary. Retinal breaks were marked with internal bipolar cautery. Hemostasis was achieved with elevation of intraocular pressure (IOP) and/or cautery. Peripheral vitrectomy using panoramic visualization was performed in all cases. When appropriate, 360° scleral indentation was performed by a trained assistant in both phakic and nonphakic eyes to remove peripheral cortical gel, relieve anterior vitreous traction, and release epicenters of anterior fibrovascular proliferation. Panretinal endophotocoagulation was performed or augmented and retinal breaks were treated, if present, at that time. Air–fluid exchange then was performed if necessary with placement of perfluorocarbon gases when an iatrogenic retinal tear and/or rhegmatogenous RD was identified intraoperatively. Patients who had undergone gas injection were instructed to remain face down, typically for 7 to 10 days. Postoperatively, the patients were treated with a standard regimen of antibiotic and antiinflammatory drops. Cycloplegics were utilized if necessary, but in the combined cases only to maintain iris mobility. Intraoperative vitreoretinal traction was described by the surgeon in the operative notes, and its status was graded later based on the amount of fibrovascular tissue and time and instrumentation used to remove it, as previously described.19 Intraoperative complications such as peripheral retinal tears and dialysis were categorized as those occurring adjacent to, in relation to, or as a result of a functioning sclerotomy wound. Intraocular hemorrhage was defined as extensive hemorrhage resulting from sclerotomy wound complications or fibrovascular membrane excision or removal that led to a significant increase in operative time. Postoperative rhegmatogenous RD was defined as RD developing from iatrogenic retinal breaks during the postoperative period. Sclerotomy-related postoperative RDs were defined as those detachments resulting from anterior breaks in close approximation to a functioning surgical pars plana sclerotomy. Postoperative hypotony was defined as an IOP ⬍ 5 mmHg (applanation tonometry) on the final visit, with a history of sustained reduction of IOP. Nonclearing postoperative vitreous hemorrhage necessitating reoperation involved those cases in which visual rehabilitation warranted reoperation to clear the ocular media and those cases in which recurrent or persistent vitreous hemorrhage in the postoperative period led to elevation of IOP (hemolytic glaucoma). Postoperative best-corrected Snellen vision was converted to logarithm of the minimum angle of resolution (logMAR) units to facilitate statistical comparison. The documented study final vision was at last follow-up (at least 6 months from surgical intervention). Visual improvement was defined as an increase of at least 0.2 logMAR units. Preoperative and postoperative nuclear scleroses were graded at the slit lamp and defined as follows: grade 0, clear lens; grade 1, early nuclear sclerosis with mild yellow posterior lens in the slit beam; grade 2, yellow color change throughout the lens; grade 3, yellow– brown coloration throughout the lens; and grade 4, brown lens.19 Cortical and posterior subcapsular changes were noted. Where appropriate, Student’s t test and the Fisher exact test were used to compare the differences in the outcomes for the groups. Statistical analyses were performed using STATA for
545
Ophthalmology Volume 114, Number 3, March 2007 Table 1. Patient Demographics and Baseline Ocular Findings Characteristic Age (yrs) Mean Range Gender (%) Female Male Follow-up time (mos) Mean Median Range Preoperative retinal traction (0–3)‡ (%) Mean 0 (none) 1⫹ (mild) 2⫹ (moderate) 3⫹ (extensive) Preoperative best-corrected visual acuity Mean Snellen Range Surgical indication (%) VH Tractional retinal detachment ⫾ VH Macular edema
Overall (n ⴝ 102)
Phakic* (n ⴝ 59)
Nonphakic* (n ⴝ 43)
P Value†
54.7 23–87
50.9 23–75
60.0 29–87
0.001
50 (49.0) 52 (50.9)
26 (44.0) 33 (55.9)
24 (55.8) 19 (44.1)
9.6 10.0 6.0–13.3
9.7 10.5 6.0–13.3
9.4 9.8 6.0–12.9
0.69
1.5 8 (7.9) 44 (43.5) 36 (35.6) 13 (12.8)
1.4 5 (8.6) 29 (50) 18 (31.0) 6 (10.3)
1.7 3 (6.9) 15 (34.8) 18 (41.8) 7 (16.2)
0.14
20/246 LP–20/25
20/207 HM–20/25
20/331 LP–20/40
38 (37.2) 59 (57.8) 5 (4.9)
22 (37.2) 32 (54.2) 5 (8.4)
16 (37.2) 27 (62.7) 0 (0)
0.24
0.41
0.20
0.16
HM ⫽ hand movements; LP ⫽ light perception; VH ⫽ vitreous hemorrhage. *Describes postoperative lens status. † P⬍0.05 was considered significant. ‡ As defined by Lahey et al.19
Windows (version 8.0, StataCorp, Inc., College Station, TX). A P value of ⬍0.05 was considered statistically significant.
Distributions of preoperative and postoperative VAs are highlighted and are similar for each group (Table 3).
Lens and Anterior Segment Complications
Results A total of 228 consecutive PDR patients who underwent PPV for complications due to diabetic retinopathy were reviewed. A subset of 102 eyes from 85 patients met eligibility inclusion criteria and was included in the study (Table 1). The surgical indications were similar in phakic and nonphakic groups, and mean intraoperative traction scores were similar. The age of the patients in the phakic group (mean, 50.9⫾14.8 years) was significantly lower than that of the nonphakic group (mean, 60.0⫾12.5) (P ⫽ 0.001). Mean follow-up time was 9.6⫾2.7 months (range, 6.0 –13.3), with no significant difference between the 2 groups (P ⫽ 0.69). Preoperatively, 72 eyes were phakic, and the remaining 30 eyes were either aphakic (n ⫽ 1) or pseudophakic (n ⫽ 29). At the time of primary vitrectomy, 1 eye underwent lensectomy and was rendered aphakic, and 12 eyes underwent planned phacoemulsification with IOL implantation. Complete patient demographics and baseline ocular findings are listed in Table 1.
Vision Preoperative vision ranged from 20/25 to light perception (mean, 20/246), with no significant difference between the phakic and nonphakic groups (Table 1). Postoperative vision ranged from 20/20 to counting fingers at 1 foot (mean, 20/57), with no difference noted between the 2 groups (P ⫽ 0.13) (Table 2). Visual acuity improved by at least 0.2 logMAR units in 76.2% of the phakic eyes and 86.0% of the nonphakic eyes (P ⫽ 0.22) (Table 2).
546
None of the eyes experienced crystalline lens complications such as inadvertent instrument touch and secondary traumatic cataract. All phakic eyes demonstrated some degree of cataract progression after vitrectomy. Twenty-nine phakic eyes (49.1%) developed visually significant nuclear sclerotic (n ⫽ 28) or cortical (n ⫽ 1) cataracts with at least 6 months’ follow-up at study conclusion (Table 2). Fourteen phakic eyes (23.7%) underwent cataract extraction (CE) and IOL implantation during the follow-up period. Opacification of the posterior capsule occurred in 3 nonphakic eyes (6.9%). Neodymium:yttrium–aluminum– garnet (Nd:YAG) laser posterior capsulotomy was performed postoperatively in all 3 eyes because of deterioration of best-corrected VA. All patients recovered their previous best postoperative VA after Nd:YAG laser. Postoperative hyphema developed in 3 phakic (5.0%) and 2 nonphakic eyes (4.6%). Anterior chamber (AC) fibrin formation occurred in 3 nonphakic (6.9%) and 1 phakic eye (1.6%). Lens and iris complications such as posterior synechia (3 phakic, 1 nonphakic), iris capture (1 phakic, 0 nonphakic), and pupillary block (none) were uncommon (Table 2). Rubeosis iridis developed in 3 phakic eyes only (5.0%). The eyes that developed anterior segment neovascularization did so after an additional surgical procedure. One eye developed rubeosis iridis after undergoing complicated CE; another eye, after repeat vitrectomy with scleral buckling for a sclerotomy-related rhegmatogenous RD; and a third eye, after recurrent vitrectomy for posterior reproliferation and macular pucker formation. Rubeosis iridis was reversed in all eyes in the postoperative period with office panretinal laser augmentation.
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status Table 2. Complications and Postoperative Ocular Findings
Characteristic Best-corrected visual acuity Mean Snellen Improvement of 0.2 logMAR units Intraoperative complications Total no. of patients with complications Patients with retinal break(s) and/or dialysis Retinal dialysis Retinal tear Anterior Posterior Hemorrhage Postoperative complications Rhegmatogenous retinal detachment Nonclearing vitreous hemorrhage Patients with anterior chamber complications Anterior chamber fibrin formation Iris capture Posterior synechiae Rubeosis iridis Hyphema Cataract progression Nuclear sclerosis Cortical Posterior capsular opacification Progressive fibrovascular proliferation Vitreoretinal reoperation Final anatomic success
Overall (n ⴝ 102) [n (%)]
Phakic* (n ⴝ 59) [n (%)]
Nonphakic* (n ⴝ 43) [n (%)]
P Value†
20/57 82 (80.3)
20/52 45 (76.2)
20/65 37 (86.0)
0.13 0.22
22 (21.5) 19 (18.6) 1 (0.9) 19 (18.6) 3 (2.9) 16 (15.6) 5 (4.9)
11 (18.6) 9 (15.2) 0 (0) 9 (15.2) 2 (3.3) 7 (11.8) 3 (5.0)
11 (25.5) 10 (23.2) 1 (2.3) 10 (23.2) 1 (2.3) 9 (20.9) 2 (4.6)
0.40 0.31
5 (4.9) 12 (11.7) 14 (13.7) 4 (3.9) 1 (0.9) 4 (3.9) 3 (2.9) 5 (4.9) 32 (31.3) 28 (27.4) 1 (0.9) 3 (2.9) 17 (16.56) 22 (21.5) 101 (99.0)
4 (6.7) 10 (16.9) 9 (15.2) 1 (1.6) 1 (1.6) 3 (5.0) 3 (5.0) 3 (5.0) 29 (49.1) 28 (47.4) 1 (1.6) 0 (0) 10 (16.9) 17 (28.8) 59 (100)
1 (2.3) 2 (4.6) 5 (11.6) 3 (6.9) 0 (0) 1 (2.3) 0 (0) 2 (4.6) 3 (6.9) 0 (0) 0 (0) 3 (6.9) 7 (16.2) 5 (11.6) 42 (97.6)
0.39 0.07 0.60
0.99 0.99
0.26
0.93 0.04 0.42
logMAR ⫽ logarithm of the minimum angle of resolution. *Describes postoperative lens status. † P⬍0.05 was considered significant.
Intraoperative/Postoperative Retinal Complications Overall, the incidence of intraoperative retinal complications such as sclerotomy-related retinal tear, dialysis, and extensive intraoperative hemorrhage requiring delay and prolongation of surgery did not differ significantly between the phakic and nonphakic groups (P ⫽ 0.40). The vitreoretinal reoperation rate for phakic eyes (28.8%) was significantly greater than that for nonphakic eyes (11.6%) (P ⫽ 0.04). In the phakic group, vitreoretinal reoperations were due to nonclearing vitreous hemorrhage (n ⫽ 10), rhegmatogenous RD (n ⫽ 3), progressive reproliferation (n ⫽ 2), hemolytic glaucoma 2 days after surgery (n ⫽ 1), and combined reproliferation and rhegmatogenous RD (n ⫽ 1). Two of the 4 cases of rhegmatogenous RD were secondary to sclerotomy-related retinal breaks. In the nonphakic group, vitreoretinal reoperations resulted from progressive reproliferation (n ⫽ 2), nonclearing vitreous hemorrhage (n ⫽ 2), and sclerotomy-related rhegmatogenous RD (n ⫽ 1). Mean times to reoperation for nonclearing vitreous hemorrhage
were similar in the phakic (6.2 months) and nonphakic groups (6.4 months). Overall, however, the individual complications of rhegmatogenous RD (P ⫽ 0.39), persistent nonclearing vitreous hemorrhage requiring reoperation (P ⫽ 0.07), and recurrent or progressive fibrovascular proliferation (P ⫽ 0.93) did not differ significantly between the 2 groups (Table 2). Postoperative hypotony occurred in one nonphakic (2.3%) and none of the phakic eyes. Ocular hypotony occurred after vitreoretinal reoperation for progressive reproliferation and traction RD, and the eye ultimately experienced phthisis. Intraocular pressure increased temporarily for several days after surgery in 20 phakic (33.9%) and 13 nonphakic eyes (30.2%). All cases were controlled with topical antiglaucoma medication and/or systemic carbonic anhydrase inhibitors. Longstanding or extended elevation of IOP unresponsive to topical and/or systemic ocular antihypertensives occurred in 2 phakic eyes that developed hemolytic glaucoma requiring repeat vitrectomy.
Table 3. Distribution of Visual Acuity of Subjects in Phakic and Nonphakic Groups Preoperative Visual Acuity
Postoperative Visual Acuity
Visual Acuity
Phakic (n ⫽ 59)
Nonphakic (n ⫽ 43)
P Value
Phakic (n ⫽ 59)
Nonphakic (n ⫽ 43)
P Value
ⱖ20/40 ⬍20/40–⬎20/100 ⱕ20/200
3 (5.0%) 10 (16.9%) 46 (77.9%)
1 (2.3%) 5 (11.6%) 37 (86.0%)
0.64
20 (33.8%) 28 (47.4%) 11 (18.6%)
10 (23.2%) 22 (51.1%) 11 (25.5%)
0.45
547
Ophthalmology Volume 114, Number 3, March 2007
Discussion In the early vitrectomy era, reports suggested that integrity of the crystalline lens could provide a barrier to the angiogenic stimulus in PDR and reduce the risk of postoperative anterior segment neovascularization and rubeosis iridis.9 –13 Intraoperative lens opacification during vitreous surgery became less problematic with the introduction of infusion additives such as glucose.22 These 2 specific factors prompted many vitreoretinal surgeons to attempt to spare the crystalline lens during diabetic vitrectomy. However, the crystalline lens can be an impediment to adequate vitrectomy and access to epiretinal proliferation in advanced cases of PDR.15 In addition, complete visualization of the anterior retina can be more difficult in the phakic eye, potentially increasing the rate of unrecognized sclerotomyrelated complications. Because vitreous incarceration, fibrovascular ingrowth through a sclerotomy wound, and persistent nonclearing vitreous hemorrhage may be more frequent in the phakic eye after diabetic vitrectomy, potentially increasing anatomic complications and affecting the reoperation rate and ultimate vision,10,11,23 we examined the effect of lens status on anatomic and visual outcomes of primary diabetic vitrectomy in this retrospective study. The major finding of this study was a significantly higher subsequent vitreoretinal reoperation rate in eyes that remained phakic at the conclusion of vitreous surgery, primarily due to higher rates of postoperative rhegmatogenous RD (6.7% vs. 2.3%) and nonclearing vitreous hemorrhage requiring reoperation (16.9% vs. 4.6%) (P ⫽ 0.07) in phakic eyes. This outcome is consistent with prior studies showing that phakic eyes exhibited a longer duration of postoperative persistent nonclearing hemorrhage than nonphakic eyes.11,24 Removal of the crystalline lens, which serves as an anatomic barrier between the anterior and posterior segments, probably facilitates more rapid clearance of red cells from the vitreous cavity through the trabecular meshwork. The relative rates of postoperative rhegmatogenous RD in each group were quite low and secondary to either sclerotomy-related or posterior breaks. Only a limited number, 3 of 5 cases, were due to sclerotomy-related complications. In principle, excision of the peripheral vitreous, particularly around the sclerotomies, is easier in the nonphakic eye, and we therefore anticipated a higher complication rate in the phakic group. However, in this review of 102 cases of diabetic vitrectomy, sclerotomy-related breaks were distributed evenly between the groups (2 phakic, 1 nonphakic). The low incidence of sclerotomy complications in this series limits our ability to determine the effect of the lens status on the associated complication rate. There may be several explanations for the limited number of sclerotomyrelated intraoperative and postoperative complications. Advances in vitrectomy surgery, particularly the use of panoramic visualization, may improve modern sclerotomy care in the diabetic eye, limiting the development and improving detection of intraoperative complications such as retinal tear and dialysis, vitreous incarceration, and hemorrhage due to inadvertent trauma to the ciliary processes. In our experience, the instrumentation utilized during diabetic vitrectomy is often capable of and culpable in creating these
548
complications, but current sclerotomy care may now be sufficient to avoid and/or limit postoperative complications such as late vitrectomy-related RD due to persistent vitreous traction, vitreous incarceration with late fibrovascular ingrowth, and anterior hyaloidal reproliferation. However, we also believe that heightened diligence in the diabetic eye, with meticulous care and removal of anterior gel even in the phakic eye, may have limited sclerotomy-related complications in this series. Anterior segment neovascularization was an uncommon event in our study, as only 3 phakic eyes developed postoperative rubeosis iridis. These 3 eyes developed rubeosis after undergoing reoperation, and the neovascularization regressed after additional office panretinal laser photocoagulation in each case. Previous authors have noted that CE with or without IOL implantation in eyes with PDR may increase the incidence of anterior segment neovascularization,25–28 but we could discern no difference in our study due to the small number of eyes that developed rubeosis. The low incidence of rubeosis in this current study suggests that the evolution of modern vitreoretinal surgical techniques has decreased the incidence of postoperative rubeosis significantly over the years, likely in part due to the development and improvement of intraocular endophotocoagulation.15,19 –21 In our cases, aggressive intraoperative endophotocoagulation was performed under panoramic visualization, facilitating treatment of the anterior retina and possibly contributing to the low incidence of postoperative rubeosis iridis. We cannot exclude the possibility that additional factors, such as advances in systemic diabetes control and less severity of proliferative disease compared with historical studies, also may have contributed to the low incidence of rubeosis iridis in this series. If the lens is not removed during vitreous surgery, cataract progression will often occur.1– 6,29 The rapidity with which it occurs in the diabetic eye has been questioned recently.30 In our series, all phakic eyes exhibited some degree of cataract progression after surgery. Twenty-nine phakic eyes (49.1%) developed visually significant lens changes with at least 6 months’ follow-up. Fourteen phakic eyes (23.7%) underwent CE with IOL implantation during the follow-up period. In previous series,15,31 at least 75% of patients who remained phakic after diabetic vitrectomy developed postoperative cataract with extended follow-up. Cataract surgery performed after PPV may be complicated by the presence of posterior synechia, a small pupil requiring the use of flexible iris hooks, and technical difficulties due to weakened zonules and the loss of posterior vitreous support, producing pronounced variations in AC depth during phacoemulsification.32–34 Disadvantages to concurrent CE include optic/iris capture and posterior synechia to either the lens optic or the anterior capsule, particularly when gas tamponade is used. This development may be due to displacement of the lens optic anteriorly, possibly in combination with an exaggerated fibrinoid response often seen in diabetic eyes after vitrectomy and endophotocoagulation.15,19,21 It is possible that these complications may occur more frequently in the diabetic eye. Lens optic/iris capture may be reduced by the use of postoperative miotics, but this may increase the risk of posterior synechia. Previtrectomy
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status cataract surgery, performed a few weeks before planned vitreous surgery, may diminish the risk of these anterior segment/IOL complications, particularly in gas-filled eyes. In the current study design, we were unable to discern any difference between patients who underwent cataract surgery before vitrectomy and those who underwent it at the time of vitrectomy. However, for younger patients, many of whom have no preexisting cataract and may have a slower progression of cataractous lens changes after vitrectomy, the benefits of previtrectomy cataract surgery may be less compelling. There are several inherent limitations to the current series, including its retrospective nature. For example, those eyes in this series that were rendered nonphakic during vitrectomy presumably had sufficient lens changes that, if not addressed, could affect surgical outcomes adversely. Second, the mean age of the nonphakic group was greater than that of the phakic group (P ⫽ 0.001). This is not surprising, because cataract development is strongly correlated with age, but we cannot exclude the possibility that patient age may be a risk factor for subsequent vitreoretinal reoperation after primary vitrectomy. Third, the results reported are those of 4 surgeons with differing surgical approaches to lens management during primary vitrectomy. The strength of this series, however, is the relatively large number of cases with more than 6 months of follow-up, providing some insight into the surgical management of the diabetic eye. In conclusion, our study suggests that the vitreoretinal reoperation rate is lower if the operated eye is nonphakic after primary diabetic vitrectomy. It appears that a major reason for this outcome is the less frequent need for reoperation for persistent or recurrent vitreous hemorrhage relative to the phakic eye. Importantly, in the modern era of diabetic vitrectomy, rubeosis iridis is uncommon and lens preservation does not increase the rate of peripheral retinal breaks, RD, and other sclerotomy-related complications. This study also suggests that the lens status of the eye does not affect ultimate anatomic and visual outcomes of primary diabetic vitrectomy. With these findings in mind, the vitreoretinal surgeon can manage the crystalline lens as deemed appropriate, rendering a diabetic eye nonphakic before or during primary vitrectomy to facilitate intraocular management and reduce the reoperation rate, or retaining the native lens to preserve accommodation in younger patients. The relative merits of rendering the diabetic eye undergoing vitrectomy nonphakic should be considered on an individual basis.
References 1. de Bustros S, Thompson JT, Michels RG, et al. Nuclear sclerosis after vitrectomy for idiopathic epiretinal membranes. Am J Ophthalmol 1998;105:160 – 4. 2. Cheng L, Azen SP, El-Bradey MH, et al. Duration of vitrectomy and postoperative cataract in the vitrectomy for macular hole study. Am J Ophthalmol 2001;132:881–7. 3. Melberg NS, Thomas MA. Nuclear sclerotic cataract after vitrectomy in patients younger than 50 years of age. Ophthalmology 1995;102:1466 –71.
4. Faulborn J, Conway BP, Machemer R. Surgical complications of pars plana vitreous surgery. Ophthalmology 1978;85:116 –25. 5. Koch F, Kloss KM, Hockwin O, Spitznas M. Lens changes following intraocular tamponade in vitrectomy. Linear densitometric image analysis of Scheimpflug photographs 6 months after operation. Klin Monatsbl Augenheilkd 1991;199:8 –11. 6. Novak MA, Rice TA, Michels RG, et al. The crystalline lens after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1480 – 4. 7. Federman JL, Schubert HD. Complications associated with the use of silicone oil in 150 eyes after retina-vitreous surgery. Ophthalmology 1988;95:870 – 6. 8. Rice TA, Michels RG. Long-term anatomic and functional results of vitrectomy for diabetic retinopathy. Am J Ophthalmol 1980;90:297–303. 9. Blankenship GW. The lens influence on diabetic vitrectomy results: report of a prospective randomized study. Arch Ophthalmol 1980;98:2196 – 8. 10. Blankenship G, Cortez R, Machemer R. The lens and pars plana vitrectomy for diabetic retinopathy complications. Arch Ophthalmol 1979;97:1263–7. 11. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Complications of vitreous surgery for diabetic retinopathy: Postoperative complications. Ophthalmology 1983;90:522–9. 12. Thompson JT, Glaser BM. Effect of lensectomy on the movement of tracers from vitreous to aqueous. Arch Ophthalmol 1984;102:1077– 8. 13. Thompson JT, Glaser BM. Role of lensectomy and posterior capsule in movement of tracers from vitreous to aqueous. Arch Ophthalmol 1985;103:420 –1. 14. McLeod D, James CR. Viscodelamination at the vitreoretinal juncture in severe diabetic eye disease. Br J Ophthalmol 1988;72:413–9. 15. Kadonosono K, Matsumoto S, Uchio E, et al. Iris neovascularization after vitrectomy combined with phacoemulsification and intraocular lens implantation for proliferative diabetic retinopathy. Ophthalmic Surg Lasers 2001;32:19 –24. 16. Scharwey K, Pavlovic S, Jacobi KW. Combined clear corneal phacoemulsification, vitreoretinal surgery, and intraocular lens implantation. J Cataract Refract Surg 1999;25:693– 8. 17. Benson WE, Brown GC, Tasman W, et al. Extracapsular cataract extraction with placement of a posterior chamber lens in patients with diabetic retinopathy. Ophthalmology 1993;100:730 – 8. 18. Chung TY, Chung H, Lee JH. Combined surgery and sequential surgery comprising phacoemulsification, pars plana vitrectomy, and intraocular lens implantation: comparison of clinical outcomes. J Cataract Refract Surg 2002;28:2001–5. 19. Lahey JM, Francis RR, Kearney JJ. Combining phacoemulsification with pars plana vitrectomy in patients with proliferative diabetic retinopathy: a series of 223 cases. Ophthalmology 2003;110:1335–9. 20. Douglas MJ, Scott IU, Flynn Jr HW. Pars plana lensectomy, pars plana vitrectomy, and silicone oil tamponade as initial management of cataract and combined traction/rhegmatogenous retinal detachment involving the macula associated with severe proliferative diabetic retinopathy. Ophthalmic Surg Lasers Imaging 2003;34:270 – 8. 21. Honjo M, Ogura Y. Surgical results of pars plana vitrectomy combined with phacoemulsification and intraocular lens implantation for complications of proliferative diabetic retinopathy. Ophthalmic Surg Lasers 1998;29:99 –105. 22. Haimann MH, Abrams GW. Prevention of lens opacification during diabetic vitrectomy. Ophthalmology 1984;91: 116 –21. 23. Aaberg TM. Pars plana vitrectomy for diabetic traction retinal detachment. Ophthalmology 1981;88:639 – 42.
549
Ophthalmology Volume 114, Number 3, March 2007 24. Novak MA, Rice TA, Michels RG, Auer C. Vitreous hemorrhage after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1485–9. 25. Poliner LS, Christianson DJ, Escoffery RF, et al. Neovascular glaucoma after intracapsular and extracapsular cataract extraction in diabetic patients. Am J Ophthalmol 1985;100:637– 43. 26. Pavese T, Insler MS. Effects of extracapsular cataract extraction with posterior chamber lens implantation on the development of neovascular glaucoma in diabetics. J Cataract Refract Surg 1987;13:197–201. 27. Scuderi JJ, Blumenkranz MS, Blankenship G. Regression of diabetic rubeosis iridis following successful surgical reattachment of the retina by vitrectomy. Retina 1982;2:193– 6. 28. Sadiq SA, Chatterjee A, Vernon SA. Progression of diabetic retinopathy and rubeotic glaucoma following cataract surgery. Eye 1995;9:728 –32.
550
29. Grewing R, Mester U. Lens opacities after pars plana vitrectomy in diabetic vitreoretinopathy and macular pucker. Fortschr Ophthalmol 1990;87:440 –2. 30. Smiddy WE, Feuer W. Incidence of cataract extraction after diabetic vitrectomy. Retina 2004;24:574 – 81. 31. Blankenship GW, Machemer R. Long-term diabetic vitrectomy results. Report of 10 year follow-up. Ophthalmology 1985;92:503– 6. 32. Schatz H, Atienza D, McDonald HR, Johnson RN. Severe diabetic retinopathy after cataract surgery. Am J Ophthalmol 1994;117:314 –21. 33. Smiddy WE, Stark WJ, Michels RG, et al. Cataract extraction after vitrectomy. Ophthalmology 1987;94:483–7. 34. McDermott ML, Puklin JE, Abrams GW, et al. Phacoemulsification for cataract following pars plana vitrectomy. Ophthalmic Surg Lasers 1997;28:558 – 64.
The surgical management of complications due to diabetic retinopathy often requires vitrectomy. The most common complication of pars plana vitrectomy (PPV) is the development of nuclear sclerotic cataract.1,2 Visually significant cataract development in the phakic eye after PPV is quite common, with rates reported between 17% and 100%.1–7 When intraocular postoperative temporary vitreous tamponade is used in conjunction with PPV, this rate increases; in Originally received: January 4, 2006. Accepted: August 11, 2006. Manuscript no. 2006-34. 1 Department of Ophthalmology, Columbia University College of Physicians and Surgeons, Edward S. Harkness Eye Institute and St. Luke’s– Roosevelt Hospital Center, New York, New York. 2 Department of Ophthalmology, Suleyman Demirel University Medical School, Isparta, Turkey. Supported by an unrestricted research grant from Research to Prevent Blindness, New York, New York. The authors have no commercial, proprietary, or financial interest in any of the products or companies described in the article. Correspondence and reprint requests to William M. Schiff, MD, 635 West 165th Street, New York, NY 10032. E-mail: [email protected].
544
© 2007 by the American Academy of Ophthalmology Published by Elsevier Inc.
cases of silicone oil injection, the rate is reported to be as high as 100%.7 Management of the crystalline lens in diabetic patients undergoing vitreous surgery has long been a source of controversy. Initially, due to intraoperative lens opacification and diminished visualization, lens removal via the pars plana was often combined with vitrectomy in the diabetic eye.4,8 Due to a well-recognized increased incidence of postoperative anterior segment neovascularization and observations suggesting a protective barrier effect of the crystalline lens in reducing the risk of progression of diabetic retinopathy and anterior segment neovascularization, the posterior segment surgeon later attempted to spare the crystalline lens (lens-sparing vitrectomy).9 –13 More recently, the anatomic and visual results of vitrectomy for severe proliferative diabetic retinopathy (PDR) have improved due to a better understanding of the pathoanatomy of the disease and improvements in surgical instrumentation and laser delivery.14 In this regard, many surgeons have reported on combined PPV with simultaneous lens removal and intraocular lens (IOL) implantation.15–21 Lens removal at the time ISSN 0161-6420/07/$–see front matter doi:10.1016/j.ophtha.2006.08.017
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status of vitrectomy eliminates the need for subsequent cataract surgery and may allow for more rapid visual recovery.15,16,19,21 Such studies suggest that the use of modern vitreoretinal techniques, including intraoperative laser photocoagulation, may have rendered the rationale for preserving the native lens during diabetic vitrectomy less compelling. In addition to becoming cataractous and impeding visual recovery after vitrectomy, the crystalline lens may serve as an anatomic impediment to intraoperative surgical maneuvers during PPV.15 Adequate epiretinal membrane dissection and/or thorough and careful removal of the basal vitreous gel in the vicinity of the sclerotomy ports may be compromised in the phakic eye. The nonphakic eye may facilitate further the recognition and management of intraoperative vitreoretinal complications such as retinal tears, dialysis, and rhegmatogenous retinal detachment (RD). Additionally, postoperative persistent or recurrent vitreous hemorrhage may clear less efficiently in the phakic eye. For these reasons, we studied the influence of lens status at the end of vitreous surgery on anatomic and visual outcomes in diabetic vitrectomy. Specifically, we performed a retrospective analysis of complication rates, rates of reoperation, anatomic success, and visual outcomes in the diabetic eye with complications of PDR after lens-sparing versus nonphakic (or combined) primary PPV.
Materials and Methods The study is a retrospective chart review of diabetic retinopathy patients who underwent primary PPV for complications relating to diabetic retinopathy between April 1999 and March 2005 at Columbia University Medical Center’s Edward S. Harkness Eye Institute and St. Luke’s–Roosevelt Hospital Center. The institutional review boards of New York–Presbyterian Hospital and St. Luke’s–Roosevelt Hospital Center approved this study. Patients included in this study had no prior retinal or vitreous surgery aside from office laser photocoagulation for clinically significant diabetic macular edema or PDR. Indications for vitrectomy included recurrent or persistent nonclearing vitreous hemorrhage, traction or combined traction/rhegmatogenous RD, and adherent posterior hyaloid causing excessive macular traction or macular edema. Patients were excluded from the study if they were younger than 18 years, had a history of uveitis in the operated eye before vitrectomy surgery, had preexisting rubeosis iridis, had previous retinal or vitreous surgery, or had less than 6 months’ follow-up after surgical intervention. The following preoperative information was obtained for each patient enrolled: age; gender; duration of diabetes; Snellen visual acuity (VA); applanation tonometry; lens status and grading of crystalline lens; presence or absence of rubeosis iridis; anterior segment abnormalities (i.e., posterior synechia and optic/iris capture); and indication for vitreoretinal surgery, including preoperative retinal drawings. The eyes that remained phakic after PPV were compared with the eyes that were nonphakic (aphakic or pseudophakic) with respect to severity of retinopathy, intraoperative complications, postoperative complications, postoperative anatomic and visual success, and total number of vitreoretinal reoperations. Patients were operated on by 1 of 4 surgeons (GRB, SC, LVDP, WMS). All patients underwent standard 3-port PPV on their affected eye. For patients who underwent phacoemulsification and IOL implanta-
tion, an experienced anterior segment surgeon performed the procedure, and a foldable IOL with a 6.0-mm optic was placed in the capsular bag via a clear corneal incision. A 10-0 nylon suture was used to secure the phacoemulsification wound. During PPV, membrane dissection and segmentation was performed when necessary to remove all tangential traction. Under conventional plano-convex and wide-field panoramic contact lenses, standard bimanual delamination, en bloc dissection, and segmentation techniques were performed. A variety of vitreoretinal scissors, forceps, and a tissue manipulator were utilized when appropriate to remove as much fibrovascular tissue as deemed necessary. Retinal breaks were marked with internal bipolar cautery. Hemostasis was achieved with elevation of intraocular pressure (IOP) and/or cautery. Peripheral vitrectomy using panoramic visualization was performed in all cases. When appropriate, 360° scleral indentation was performed by a trained assistant in both phakic and nonphakic eyes to remove peripheral cortical gel, relieve anterior vitreous traction, and release epicenters of anterior fibrovascular proliferation. Panretinal endophotocoagulation was performed or augmented and retinal breaks were treated, if present, at that time. Air–fluid exchange then was performed if necessary with placement of perfluorocarbon gases when an iatrogenic retinal tear and/or rhegmatogenous RD was identified intraoperatively. Patients who had undergone gas injection were instructed to remain face down, typically for 7 to 10 days. Postoperatively, the patients were treated with a standard regimen of antibiotic and antiinflammatory drops. Cycloplegics were utilized if necessary, but in the combined cases only to maintain iris mobility. Intraoperative vitreoretinal traction was described by the surgeon in the operative notes, and its status was graded later based on the amount of fibrovascular tissue and time and instrumentation used to remove it, as previously described.19 Intraoperative complications such as peripheral retinal tears and dialysis were categorized as those occurring adjacent to, in relation to, or as a result of a functioning sclerotomy wound. Intraocular hemorrhage was defined as extensive hemorrhage resulting from sclerotomy wound complications or fibrovascular membrane excision or removal that led to a significant increase in operative time. Postoperative rhegmatogenous RD was defined as RD developing from iatrogenic retinal breaks during the postoperative period. Sclerotomy-related postoperative RDs were defined as those detachments resulting from anterior breaks in close approximation to a functioning surgical pars plana sclerotomy. Postoperative hypotony was defined as an IOP ⬍ 5 mmHg (applanation tonometry) on the final visit, with a history of sustained reduction of IOP. Nonclearing postoperative vitreous hemorrhage necessitating reoperation involved those cases in which visual rehabilitation warranted reoperation to clear the ocular media and those cases in which recurrent or persistent vitreous hemorrhage in the postoperative period led to elevation of IOP (hemolytic glaucoma). Postoperative best-corrected Snellen vision was converted to logarithm of the minimum angle of resolution (logMAR) units to facilitate statistical comparison. The documented study final vision was at last follow-up (at least 6 months from surgical intervention). Visual improvement was defined as an increase of at least 0.2 logMAR units. Preoperative and postoperative nuclear scleroses were graded at the slit lamp and defined as follows: grade 0, clear lens; grade 1, early nuclear sclerosis with mild yellow posterior lens in the slit beam; grade 2, yellow color change throughout the lens; grade 3, yellow– brown coloration throughout the lens; and grade 4, brown lens.19 Cortical and posterior subcapsular changes were noted. Where appropriate, Student’s t test and the Fisher exact test were used to compare the differences in the outcomes for the groups. Statistical analyses were performed using STATA for
545
Ophthalmology Volume 114, Number 3, March 2007 Table 1. Patient Demographics and Baseline Ocular Findings Characteristic Age (yrs) Mean Range Gender (%) Female Male Follow-up time (mos) Mean Median Range Preoperative retinal traction (0–3)‡ (%) Mean 0 (none) 1⫹ (mild) 2⫹ (moderate) 3⫹ (extensive) Preoperative best-corrected visual acuity Mean Snellen Range Surgical indication (%) VH Tractional retinal detachment ⫾ VH Macular edema
Overall (n ⴝ 102)
Phakic* (n ⴝ 59)
Nonphakic* (n ⴝ 43)
P Value†
54.7 23–87
50.9 23–75
60.0 29–87
0.001
50 (49.0) 52 (50.9)
26 (44.0) 33 (55.9)
24 (55.8) 19 (44.1)
9.6 10.0 6.0–13.3
9.7 10.5 6.0–13.3
9.4 9.8 6.0–12.9
0.69
1.5 8 (7.9) 44 (43.5) 36 (35.6) 13 (12.8)
1.4 5 (8.6) 29 (50) 18 (31.0) 6 (10.3)
1.7 3 (6.9) 15 (34.8) 18 (41.8) 7 (16.2)
0.14
20/246 LP–20/25
20/207 HM–20/25
20/331 LP–20/40
38 (37.2) 59 (57.8) 5 (4.9)
22 (37.2) 32 (54.2) 5 (8.4)
16 (37.2) 27 (62.7) 0 (0)
0.24
0.41
0.20
0.16
HM ⫽ hand movements; LP ⫽ light perception; VH ⫽ vitreous hemorrhage. *Describes postoperative lens status. † P⬍0.05 was considered significant. ‡ As defined by Lahey et al.19
Windows (version 8.0, StataCorp, Inc., College Station, TX). A P value of ⬍0.05 was considered statistically significant.
Distributions of preoperative and postoperative VAs are highlighted and are similar for each group (Table 3).
Lens and Anterior Segment Complications
Results A total of 228 consecutive PDR patients who underwent PPV for complications due to diabetic retinopathy were reviewed. A subset of 102 eyes from 85 patients met eligibility inclusion criteria and was included in the study (Table 1). The surgical indications were similar in phakic and nonphakic groups, and mean intraoperative traction scores were similar. The age of the patients in the phakic group (mean, 50.9⫾14.8 years) was significantly lower than that of the nonphakic group (mean, 60.0⫾12.5) (P ⫽ 0.001). Mean follow-up time was 9.6⫾2.7 months (range, 6.0 –13.3), with no significant difference between the 2 groups (P ⫽ 0.69). Preoperatively, 72 eyes were phakic, and the remaining 30 eyes were either aphakic (n ⫽ 1) or pseudophakic (n ⫽ 29). At the time of primary vitrectomy, 1 eye underwent lensectomy and was rendered aphakic, and 12 eyes underwent planned phacoemulsification with IOL implantation. Complete patient demographics and baseline ocular findings are listed in Table 1.
Vision Preoperative vision ranged from 20/25 to light perception (mean, 20/246), with no significant difference between the phakic and nonphakic groups (Table 1). Postoperative vision ranged from 20/20 to counting fingers at 1 foot (mean, 20/57), with no difference noted between the 2 groups (P ⫽ 0.13) (Table 2). Visual acuity improved by at least 0.2 logMAR units in 76.2% of the phakic eyes and 86.0% of the nonphakic eyes (P ⫽ 0.22) (Table 2).
546
None of the eyes experienced crystalline lens complications such as inadvertent instrument touch and secondary traumatic cataract. All phakic eyes demonstrated some degree of cataract progression after vitrectomy. Twenty-nine phakic eyes (49.1%) developed visually significant nuclear sclerotic (n ⫽ 28) or cortical (n ⫽ 1) cataracts with at least 6 months’ follow-up at study conclusion (Table 2). Fourteen phakic eyes (23.7%) underwent cataract extraction (CE) and IOL implantation during the follow-up period. Opacification of the posterior capsule occurred in 3 nonphakic eyes (6.9%). Neodymium:yttrium–aluminum– garnet (Nd:YAG) laser posterior capsulotomy was performed postoperatively in all 3 eyes because of deterioration of best-corrected VA. All patients recovered their previous best postoperative VA after Nd:YAG laser. Postoperative hyphema developed in 3 phakic (5.0%) and 2 nonphakic eyes (4.6%). Anterior chamber (AC) fibrin formation occurred in 3 nonphakic (6.9%) and 1 phakic eye (1.6%). Lens and iris complications such as posterior synechia (3 phakic, 1 nonphakic), iris capture (1 phakic, 0 nonphakic), and pupillary block (none) were uncommon (Table 2). Rubeosis iridis developed in 3 phakic eyes only (5.0%). The eyes that developed anterior segment neovascularization did so after an additional surgical procedure. One eye developed rubeosis iridis after undergoing complicated CE; another eye, after repeat vitrectomy with scleral buckling for a sclerotomy-related rhegmatogenous RD; and a third eye, after recurrent vitrectomy for posterior reproliferation and macular pucker formation. Rubeosis iridis was reversed in all eyes in the postoperative period with office panretinal laser augmentation.
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status Table 2. Complications and Postoperative Ocular Findings
Characteristic Best-corrected visual acuity Mean Snellen Improvement of 0.2 logMAR units Intraoperative complications Total no. of patients with complications Patients with retinal break(s) and/or dialysis Retinal dialysis Retinal tear Anterior Posterior Hemorrhage Postoperative complications Rhegmatogenous retinal detachment Nonclearing vitreous hemorrhage Patients with anterior chamber complications Anterior chamber fibrin formation Iris capture Posterior synechiae Rubeosis iridis Hyphema Cataract progression Nuclear sclerosis Cortical Posterior capsular opacification Progressive fibrovascular proliferation Vitreoretinal reoperation Final anatomic success
Overall (n ⴝ 102) [n (%)]
Phakic* (n ⴝ 59) [n (%)]
Nonphakic* (n ⴝ 43) [n (%)]
P Value†
20/57 82 (80.3)
20/52 45 (76.2)
20/65 37 (86.0)
0.13 0.22
22 (21.5) 19 (18.6) 1 (0.9) 19 (18.6) 3 (2.9) 16 (15.6) 5 (4.9)
11 (18.6) 9 (15.2) 0 (0) 9 (15.2) 2 (3.3) 7 (11.8) 3 (5.0)
11 (25.5) 10 (23.2) 1 (2.3) 10 (23.2) 1 (2.3) 9 (20.9) 2 (4.6)
0.40 0.31
5 (4.9) 12 (11.7) 14 (13.7) 4 (3.9) 1 (0.9) 4 (3.9) 3 (2.9) 5 (4.9) 32 (31.3) 28 (27.4) 1 (0.9) 3 (2.9) 17 (16.56) 22 (21.5) 101 (99.0)
4 (6.7) 10 (16.9) 9 (15.2) 1 (1.6) 1 (1.6) 3 (5.0) 3 (5.0) 3 (5.0) 29 (49.1) 28 (47.4) 1 (1.6) 0 (0) 10 (16.9) 17 (28.8) 59 (100)
1 (2.3) 2 (4.6) 5 (11.6) 3 (6.9) 0 (0) 1 (2.3) 0 (0) 2 (4.6) 3 (6.9) 0 (0) 0 (0) 3 (6.9) 7 (16.2) 5 (11.6) 42 (97.6)
0.39 0.07 0.60
0.99 0.99
0.26
0.93 0.04 0.42
logMAR ⫽ logarithm of the minimum angle of resolution. *Describes postoperative lens status. † P⬍0.05 was considered significant.
Intraoperative/Postoperative Retinal Complications Overall, the incidence of intraoperative retinal complications such as sclerotomy-related retinal tear, dialysis, and extensive intraoperative hemorrhage requiring delay and prolongation of surgery did not differ significantly between the phakic and nonphakic groups (P ⫽ 0.40). The vitreoretinal reoperation rate for phakic eyes (28.8%) was significantly greater than that for nonphakic eyes (11.6%) (P ⫽ 0.04). In the phakic group, vitreoretinal reoperations were due to nonclearing vitreous hemorrhage (n ⫽ 10), rhegmatogenous RD (n ⫽ 3), progressive reproliferation (n ⫽ 2), hemolytic glaucoma 2 days after surgery (n ⫽ 1), and combined reproliferation and rhegmatogenous RD (n ⫽ 1). Two of the 4 cases of rhegmatogenous RD were secondary to sclerotomy-related retinal breaks. In the nonphakic group, vitreoretinal reoperations resulted from progressive reproliferation (n ⫽ 2), nonclearing vitreous hemorrhage (n ⫽ 2), and sclerotomy-related rhegmatogenous RD (n ⫽ 1). Mean times to reoperation for nonclearing vitreous hemorrhage
were similar in the phakic (6.2 months) and nonphakic groups (6.4 months). Overall, however, the individual complications of rhegmatogenous RD (P ⫽ 0.39), persistent nonclearing vitreous hemorrhage requiring reoperation (P ⫽ 0.07), and recurrent or progressive fibrovascular proliferation (P ⫽ 0.93) did not differ significantly between the 2 groups (Table 2). Postoperative hypotony occurred in one nonphakic (2.3%) and none of the phakic eyes. Ocular hypotony occurred after vitreoretinal reoperation for progressive reproliferation and traction RD, and the eye ultimately experienced phthisis. Intraocular pressure increased temporarily for several days after surgery in 20 phakic (33.9%) and 13 nonphakic eyes (30.2%). All cases were controlled with topical antiglaucoma medication and/or systemic carbonic anhydrase inhibitors. Longstanding or extended elevation of IOP unresponsive to topical and/or systemic ocular antihypertensives occurred in 2 phakic eyes that developed hemolytic glaucoma requiring repeat vitrectomy.
Table 3. Distribution of Visual Acuity of Subjects in Phakic and Nonphakic Groups Preoperative Visual Acuity
Postoperative Visual Acuity
Visual Acuity
Phakic (n ⫽ 59)
Nonphakic (n ⫽ 43)
P Value
Phakic (n ⫽ 59)
Nonphakic (n ⫽ 43)
P Value
ⱖ20/40 ⬍20/40–⬎20/100 ⱕ20/200
3 (5.0%) 10 (16.9%) 46 (77.9%)
1 (2.3%) 5 (11.6%) 37 (86.0%)
0.64
20 (33.8%) 28 (47.4%) 11 (18.6%)
10 (23.2%) 22 (51.1%) 11 (25.5%)
0.45
547
Ophthalmology Volume 114, Number 3, March 2007
Discussion In the early vitrectomy era, reports suggested that integrity of the crystalline lens could provide a barrier to the angiogenic stimulus in PDR and reduce the risk of postoperative anterior segment neovascularization and rubeosis iridis.9 –13 Intraoperative lens opacification during vitreous surgery became less problematic with the introduction of infusion additives such as glucose.22 These 2 specific factors prompted many vitreoretinal surgeons to attempt to spare the crystalline lens during diabetic vitrectomy. However, the crystalline lens can be an impediment to adequate vitrectomy and access to epiretinal proliferation in advanced cases of PDR.15 In addition, complete visualization of the anterior retina can be more difficult in the phakic eye, potentially increasing the rate of unrecognized sclerotomyrelated complications. Because vitreous incarceration, fibrovascular ingrowth through a sclerotomy wound, and persistent nonclearing vitreous hemorrhage may be more frequent in the phakic eye after diabetic vitrectomy, potentially increasing anatomic complications and affecting the reoperation rate and ultimate vision,10,11,23 we examined the effect of lens status on anatomic and visual outcomes of primary diabetic vitrectomy in this retrospective study. The major finding of this study was a significantly higher subsequent vitreoretinal reoperation rate in eyes that remained phakic at the conclusion of vitreous surgery, primarily due to higher rates of postoperative rhegmatogenous RD (6.7% vs. 2.3%) and nonclearing vitreous hemorrhage requiring reoperation (16.9% vs. 4.6%) (P ⫽ 0.07) in phakic eyes. This outcome is consistent with prior studies showing that phakic eyes exhibited a longer duration of postoperative persistent nonclearing hemorrhage than nonphakic eyes.11,24 Removal of the crystalline lens, which serves as an anatomic barrier between the anterior and posterior segments, probably facilitates more rapid clearance of red cells from the vitreous cavity through the trabecular meshwork. The relative rates of postoperative rhegmatogenous RD in each group were quite low and secondary to either sclerotomy-related or posterior breaks. Only a limited number, 3 of 5 cases, were due to sclerotomy-related complications. In principle, excision of the peripheral vitreous, particularly around the sclerotomies, is easier in the nonphakic eye, and we therefore anticipated a higher complication rate in the phakic group. However, in this review of 102 cases of diabetic vitrectomy, sclerotomy-related breaks were distributed evenly between the groups (2 phakic, 1 nonphakic). The low incidence of sclerotomy complications in this series limits our ability to determine the effect of the lens status on the associated complication rate. There may be several explanations for the limited number of sclerotomyrelated intraoperative and postoperative complications. Advances in vitrectomy surgery, particularly the use of panoramic visualization, may improve modern sclerotomy care in the diabetic eye, limiting the development and improving detection of intraoperative complications such as retinal tear and dialysis, vitreous incarceration, and hemorrhage due to inadvertent trauma to the ciliary processes. In our experience, the instrumentation utilized during diabetic vitrectomy is often capable of and culpable in creating these
548
complications, but current sclerotomy care may now be sufficient to avoid and/or limit postoperative complications such as late vitrectomy-related RD due to persistent vitreous traction, vitreous incarceration with late fibrovascular ingrowth, and anterior hyaloidal reproliferation. However, we also believe that heightened diligence in the diabetic eye, with meticulous care and removal of anterior gel even in the phakic eye, may have limited sclerotomy-related complications in this series. Anterior segment neovascularization was an uncommon event in our study, as only 3 phakic eyes developed postoperative rubeosis iridis. These 3 eyes developed rubeosis after undergoing reoperation, and the neovascularization regressed after additional office panretinal laser photocoagulation in each case. Previous authors have noted that CE with or without IOL implantation in eyes with PDR may increase the incidence of anterior segment neovascularization,25–28 but we could discern no difference in our study due to the small number of eyes that developed rubeosis. The low incidence of rubeosis in this current study suggests that the evolution of modern vitreoretinal surgical techniques has decreased the incidence of postoperative rubeosis significantly over the years, likely in part due to the development and improvement of intraocular endophotocoagulation.15,19 –21 In our cases, aggressive intraoperative endophotocoagulation was performed under panoramic visualization, facilitating treatment of the anterior retina and possibly contributing to the low incidence of postoperative rubeosis iridis. We cannot exclude the possibility that additional factors, such as advances in systemic diabetes control and less severity of proliferative disease compared with historical studies, also may have contributed to the low incidence of rubeosis iridis in this series. If the lens is not removed during vitreous surgery, cataract progression will often occur.1– 6,29 The rapidity with which it occurs in the diabetic eye has been questioned recently.30 In our series, all phakic eyes exhibited some degree of cataract progression after surgery. Twenty-nine phakic eyes (49.1%) developed visually significant lens changes with at least 6 months’ follow-up. Fourteen phakic eyes (23.7%) underwent CE with IOL implantation during the follow-up period. In previous series,15,31 at least 75% of patients who remained phakic after diabetic vitrectomy developed postoperative cataract with extended follow-up. Cataract surgery performed after PPV may be complicated by the presence of posterior synechia, a small pupil requiring the use of flexible iris hooks, and technical difficulties due to weakened zonules and the loss of posterior vitreous support, producing pronounced variations in AC depth during phacoemulsification.32–34 Disadvantages to concurrent CE include optic/iris capture and posterior synechia to either the lens optic or the anterior capsule, particularly when gas tamponade is used. This development may be due to displacement of the lens optic anteriorly, possibly in combination with an exaggerated fibrinoid response often seen in diabetic eyes after vitrectomy and endophotocoagulation.15,19,21 It is possible that these complications may occur more frequently in the diabetic eye. Lens optic/iris capture may be reduced by the use of postoperative miotics, but this may increase the risk of posterior synechia. Previtrectomy
Schiff et al 䡠 Diabetic Vitrectomy and Influence of Lens Status cataract surgery, performed a few weeks before planned vitreous surgery, may diminish the risk of these anterior segment/IOL complications, particularly in gas-filled eyes. In the current study design, we were unable to discern any difference between patients who underwent cataract surgery before vitrectomy and those who underwent it at the time of vitrectomy. However, for younger patients, many of whom have no preexisting cataract and may have a slower progression of cataractous lens changes after vitrectomy, the benefits of previtrectomy cataract surgery may be less compelling. There are several inherent limitations to the current series, including its retrospective nature. For example, those eyes in this series that were rendered nonphakic during vitrectomy presumably had sufficient lens changes that, if not addressed, could affect surgical outcomes adversely. Second, the mean age of the nonphakic group was greater than that of the phakic group (P ⫽ 0.001). This is not surprising, because cataract development is strongly correlated with age, but we cannot exclude the possibility that patient age may be a risk factor for subsequent vitreoretinal reoperation after primary vitrectomy. Third, the results reported are those of 4 surgeons with differing surgical approaches to lens management during primary vitrectomy. The strength of this series, however, is the relatively large number of cases with more than 6 months of follow-up, providing some insight into the surgical management of the diabetic eye. In conclusion, our study suggests that the vitreoretinal reoperation rate is lower if the operated eye is nonphakic after primary diabetic vitrectomy. It appears that a major reason for this outcome is the less frequent need for reoperation for persistent or recurrent vitreous hemorrhage relative to the phakic eye. Importantly, in the modern era of diabetic vitrectomy, rubeosis iridis is uncommon and lens preservation does not increase the rate of peripheral retinal breaks, RD, and other sclerotomy-related complications. This study also suggests that the lens status of the eye does not affect ultimate anatomic and visual outcomes of primary diabetic vitrectomy. With these findings in mind, the vitreoretinal surgeon can manage the crystalline lens as deemed appropriate, rendering a diabetic eye nonphakic before or during primary vitrectomy to facilitate intraocular management and reduce the reoperation rate, or retaining the native lens to preserve accommodation in younger patients. The relative merits of rendering the diabetic eye undergoing vitrectomy nonphakic should be considered on an individual basis.
References 1. de Bustros S, Thompson JT, Michels RG, et al. Nuclear sclerosis after vitrectomy for idiopathic epiretinal membranes. Am J Ophthalmol 1998;105:160 – 4. 2. Cheng L, Azen SP, El-Bradey MH, et al. Duration of vitrectomy and postoperative cataract in the vitrectomy for macular hole study. Am J Ophthalmol 2001;132:881–7. 3. Melberg NS, Thomas MA. Nuclear sclerotic cataract after vitrectomy in patients younger than 50 years of age. Ophthalmology 1995;102:1466 –71.
4. Faulborn J, Conway BP, Machemer R. Surgical complications of pars plana vitreous surgery. Ophthalmology 1978;85:116 –25. 5. Koch F, Kloss KM, Hockwin O, Spitznas M. Lens changes following intraocular tamponade in vitrectomy. Linear densitometric image analysis of Scheimpflug photographs 6 months after operation. Klin Monatsbl Augenheilkd 1991;199:8 –11. 6. Novak MA, Rice TA, Michels RG, et al. The crystalline lens after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1480 – 4. 7. Federman JL, Schubert HD. Complications associated with the use of silicone oil in 150 eyes after retina-vitreous surgery. Ophthalmology 1988;95:870 – 6. 8. Rice TA, Michels RG. Long-term anatomic and functional results of vitrectomy for diabetic retinopathy. Am J Ophthalmol 1980;90:297–303. 9. Blankenship GW. The lens influence on diabetic vitrectomy results: report of a prospective randomized study. Arch Ophthalmol 1980;98:2196 – 8. 10. Blankenship G, Cortez R, Machemer R. The lens and pars plana vitrectomy for diabetic retinopathy complications. Arch Ophthalmol 1979;97:1263–7. 11. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Complications of vitreous surgery for diabetic retinopathy: Postoperative complications. Ophthalmology 1983;90:522–9. 12. Thompson JT, Glaser BM. Effect of lensectomy on the movement of tracers from vitreous to aqueous. Arch Ophthalmol 1984;102:1077– 8. 13. Thompson JT, Glaser BM. Role of lensectomy and posterior capsule in movement of tracers from vitreous to aqueous. Arch Ophthalmol 1985;103:420 –1. 14. McLeod D, James CR. Viscodelamination at the vitreoretinal juncture in severe diabetic eye disease. Br J Ophthalmol 1988;72:413–9. 15. Kadonosono K, Matsumoto S, Uchio E, et al. Iris neovascularization after vitrectomy combined with phacoemulsification and intraocular lens implantation for proliferative diabetic retinopathy. Ophthalmic Surg Lasers 2001;32:19 –24. 16. Scharwey K, Pavlovic S, Jacobi KW. Combined clear corneal phacoemulsification, vitreoretinal surgery, and intraocular lens implantation. J Cataract Refract Surg 1999;25:693– 8. 17. Benson WE, Brown GC, Tasman W, et al. Extracapsular cataract extraction with placement of a posterior chamber lens in patients with diabetic retinopathy. Ophthalmology 1993;100:730 – 8. 18. Chung TY, Chung H, Lee JH. Combined surgery and sequential surgery comprising phacoemulsification, pars plana vitrectomy, and intraocular lens implantation: comparison of clinical outcomes. J Cataract Refract Surg 2002;28:2001–5. 19. Lahey JM, Francis RR, Kearney JJ. Combining phacoemulsification with pars plana vitrectomy in patients with proliferative diabetic retinopathy: a series of 223 cases. Ophthalmology 2003;110:1335–9. 20. Douglas MJ, Scott IU, Flynn Jr HW. Pars plana lensectomy, pars plana vitrectomy, and silicone oil tamponade as initial management of cataract and combined traction/rhegmatogenous retinal detachment involving the macula associated with severe proliferative diabetic retinopathy. Ophthalmic Surg Lasers Imaging 2003;34:270 – 8. 21. Honjo M, Ogura Y. Surgical results of pars plana vitrectomy combined with phacoemulsification and intraocular lens implantation for complications of proliferative diabetic retinopathy. Ophthalmic Surg Lasers 1998;29:99 –105. 22. Haimann MH, Abrams GW. Prevention of lens opacification during diabetic vitrectomy. Ophthalmology 1984;91: 116 –21. 23. Aaberg TM. Pars plana vitrectomy for diabetic traction retinal detachment. Ophthalmology 1981;88:639 – 42.
549
Ophthalmology Volume 114, Number 3, March 2007 24. Novak MA, Rice TA, Michels RG, Auer C. Vitreous hemorrhage after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1485–9. 25. Poliner LS, Christianson DJ, Escoffery RF, et al. Neovascular glaucoma after intracapsular and extracapsular cataract extraction in diabetic patients. Am J Ophthalmol 1985;100:637– 43. 26. Pavese T, Insler MS. Effects of extracapsular cataract extraction with posterior chamber lens implantation on the development of neovascular glaucoma in diabetics. J Cataract Refract Surg 1987;13:197–201. 27. Scuderi JJ, Blumenkranz MS, Blankenship G. Regression of diabetic rubeosis iridis following successful surgical reattachment of the retina by vitrectomy. Retina 1982;2:193– 6. 28. Sadiq SA, Chatterjee A, Vernon SA. Progression of diabetic retinopathy and rubeotic glaucoma following cataract surgery. Eye 1995;9:728 –32.
550
29. Grewing R, Mester U. Lens opacities after pars plana vitrectomy in diabetic vitreoretinopathy and macular pucker. Fortschr Ophthalmol 1990;87:440 –2. 30. Smiddy WE, Feuer W. Incidence of cataract extraction after diabetic vitrectomy. Retina 2004;24:574 – 81. 31. Blankenship GW, Machemer R. Long-term diabetic vitrectomy results. Report of 10 year follow-up. Ophthalmology 1985;92:503– 6. 32. Schatz H, Atienza D, McDonald HR, Johnson RN. Severe diabetic retinopathy after cataract surgery. Am J Ophthalmol 1994;117:314 –21. 33. Smiddy WE, Stark WJ, Michels RG, et al. Cataract extraction after vitrectomy. Ophthalmology 1987;94:483–7. 34. McDermott ML, Puklin JE, Abrams GW, et al. Phacoemulsification for cataract following pars plana vitrectomy. Ophthalmic Surg Lasers 1997;28:558 – 64.