Vitrectomy for Proliferative Diabetic Retinopathy With Severe Equatorial Fibrovascular Proliferation DENNIS P. HAN, MD., JOSE S. PULIDO, MD., WILLIAM F. MIELER, MD., AND MARK W. JOHNSON, M.D.
• PURPOSE: We studied the surgical treatment and visual outcome in a consecutive series of eyes with an unusual syndrome of diabetic retinopathy and severe peripheral fibrovascular proliferation in volving the equatorial and pre-equatorial fundus. • METHODS: In a retrospective study of 2 7 6 eyes (245 patients) that underwent pars plana vitréetomy for diabetic retinopathy between November 1988 and February 1 9 9 3 , nine eyes of eight patients (3.3% of eyes and 3.3% of patients) had severe equatorial fibrovascular proliferation. The condition occurred primarily in previously unoperated-on eyes (except for panretinal photoco agulation) and resulted in peripheral traction or traction-rhegmatogenous retinal detachment (six eyes), severe vitreous hemorrhage (two eyes), and severe hypotony (one eye). Relief of traction from peripheral fibrovascular membranes was obtained with an encircling scierai buckle (nine eyes) and limited delamination and segmentation (five eyes) or relaxing retinectomy (two eyes). Lensectomy was required for adequate membrane dissection in three eyes. • RESULTS: After follow-up of six to 5 2 months (mean, 2 0 . 4 months), the visual acuity was 2 0 / 200 or better in seven of nine eyes, with complete retinal attachment in seven of nine eyes and postequatorial attachment in all eyes (100%). Accepted for publication Oct. 21, 1994. From the Eye Institute, Medical College of Wisconsin, Milwaukee, Wisconsin (Drs. Han, Pulido, and Mieler); and the W. K. Kellogg Eye Center, University of Michigan, Ann Arbor, Michigan (Dr. Johnson). This study was supported in part by an unrestricted grant from Re search to Prevent Blindness, Inc., New York, New York. Reprint requests to Dennis P. Han, M.D., Eye Institute, 8700 W. Wisconsin Ave., Milwaukee, WI 53226; fax: (414) 257-7353. VOL. I 19,
No. 5
Poor outcome resulted from a persistent response resembling Coats' disease in one eye and pré existent long-standing retinal detachment in one eye. • CONCLUSIONS: Vitrectomy for severe equatori al fibrovascular proliferation differs from conven tional approaches to diabetic retinopathy in that relief of retinal traction must be attained by scierai buckling and adequate dissection of peripheral fibrovascular tissue. In advanced cases, lensecto my and relaxing retinotomy may be required.
T
HE FIBROVASCULAR PROLIFERATION OF DIABETIC retinopathy characteristically occurs in the postequatorial fundus, resulting in tractional complications that predominantly involve the posteri or pole.1 In contrast, we observed cases of unusually severe fibrovascular proliferation limited to the equa torial or pre-equatorial fundus, resulting in traction or traction-rhegmatogenous retinal detachment, vit reous hemorrhage, and hypotony. This condition appeared to be distinct from anterior hyaloidal fibro vascular proliferation,2,3 occurring most often in pre viously unoperated-on eyes (except for panretinal photocoagulation) and involving the equatorial retina and not the anterior hyaloid and ciliary body region. Our surgical approach differed from conventional methods of diabetic vitrectomy because of the periph eral location of the disease.
PATIENTS AND METHODS BETWEEN NOVEMBER 1988 AND FEBRUARY 1993, WE PER-
formed pars plana vitrectomy on 276 eyes of 245
© AMERICAN JOURNAL OF OPHTHALMOLOGY ii9:563-570
563
patients for complications of diabetic retinopathy. Of this consecutive series of surgical cases, nine eyes of eight patients had developed severe fibrovascular proliferation in the equatorial or pre-equatorial fundus, resulting in recurrent or severe vitreous hemor rhage, traction or traction-rhegmatogenous retinal detachment, or severe hypotony. Surgical treatment consisted of a three-port pars plana vitrectomy and excision of the posterior hyaloid membrane, beginning in the posterior pole and extending into the periphery as far as possible. Relief of peripheral traction was obtained by excision or segmentation of the peripheral fibrovascular tissue, if possible, and by scierai buckling. During membrane dissection, hemostasis was obtained by brief elevation of the infusion bottles, to increase intraocular pres sure, or by endodiathermy. Relaxing retinectomy was performed if scierai buckling and membrane dissec tion were insufficient to relieve traction. Pars plana lensectomy was performed if membrane dissection was required in the far periphery. Fluid-gas exchange, with injection of a mixture of sulfur hexafluoride (SF6) 20% or perfluoropropane (C3F8) 10% to 20% in air, and photocoagulation were performed for intra operative treatment of retinal breaks or retinectomies. For photocoagulation, either the diode laser indirect ophthalmoscope (810 nm) or the argon blue-green endolaser was used. In selected cases, supplemental scatter photocoagulation was also performed with these lasers.
disclosed heavy ethanol intake, marijuana smoking, and poor control of blood glucose level. There was no history of cigarette smoking, hypertension, or renal failure. Visual acuity was 2/200 in the left eye. Results of slit-lamp examination were normal and showed no rubeosis iridis. No anterior hyaloidal fibrovascular proliferation was noted in the retrolenticular or ciliary body region. Intraocular pressure by applanation tonometry was 7 mm Hg. Fundus examination dem onstrated extensive equatorial and pre-equatorial fi brovascular proliferation for 360 degrees with associ ated vitreous hemorrhage and peripheral traction detachment extending posteriorly and detaching the macula. B-scan ultrasound demonstrated subretinal hemorrhage in the posterior pole (Fig. 1) and con firmed the ophthalmoscopic findings (Fig. 2). A pars plana vitrectomy and lensectomy were performed. During the vitrectomy, perpendicular traction on the peripheral fibrovascular tissue could be appreciated, which was partially relieved by remov al of vitreous in the anterior vitreous cavity. The posterior hyaloid was dissected from the posterior pole
Anatomic success was defined as attachment of the posterior retina (including the macula) with an intraocular pressure of 8 mm Hg or more. Functional success was defined as anatomic success with bestcorrected visual acuity of 20/200 or better. A case report describes the clinical manifestation and surgi cal treatment of a patient with severe equatorial fibrovascular proliferation.
CASE REPORT • CASE 7: A 25-year-old white man with a 22-year history of type I diabetes mellitus had decreased vision in his left eye. He previously received panretinal photocoagulation in both eyes and received a total of 3,270 laser applications in the left eye. He other wise had no previous ocular surgery. Medical history
Fig. 1 (Han and associates). Case 7, left eye. Longitudi nal view of the 3 o'clock meridian by preoperative B-scan echography demonstrated near-total retinal de tachment (straight arrow). Subretinal hemorrhage was also present (curved arrows).
564
OPHTHALMOLOGY
AMERICAN JOURNAL
MAY 1995
Fig. 2 (Han and associates). Case 7, left eye. Preoperative fundus photomontage shows marked equatorial fi brovascular proliferation and total retinal detachment. Evidence of previous panretinal scatter photocoagulation was present. Visual acuity was 2/200.
Fig. 3 (Han and associates). Case 7, left eye. A fundus photomontage obtained 11 months postoperatively shows residual fibrous tissue supported by peripheral scierai buckle. The retina was completely attached. Visual acuity was 20/100. VOL.119,
No. 5
Fig. 4 (Han and associates). Case 5, right eye. Preoperative fundus appearance demonstrates extensive equatori al fibrovascular proliferation and dense retrohyaloid hemorrhage in the superior quadrant. Fibrovascular proliferative tissue throughout 360 degrees was present, associated with a nasal and inferior traction retinal detachment (not shown). Visual acuity was 4/200.
Fig. 5 (Han and associates). Case 5, right eye. Fundus photograph obtained one month postoperatively demon strates residual nondissectable fibrous tissue in the superior quadrant supported by a peripheral scierai buckle. Photocoagulation scars were present on the scierai buckle. Visual acuity was 20/80, and the retina was completely attached after 14 months of follow-up.
VlTRECTOMY FOR PROLIFERATIVE DIABETIC RETINOPATHY
565
and excised anteriorly to its attachments at the peripheralfibrovasculartissue. Delamination and seg mentation of the fibrovascular tissue at its posterior aspect were performed but were limited by a marked adhesion between the more peripheral tissue and the underlying retina. Two iatrogenic breaks were created inferonasally and just posterior to the residual fibro vascular tissue. A broad (7-mm wide), encircling scierai buckle was placed to support the residual fibrovascular tissue for 360 degrees. A fluid-gas ex change with C3F8 20% was performed. During the exchange, the pre-existent subretinal blood spontane ously evacuated into the vitreous cavity through the inferonasal retinal breaks. Three hundred forty-one applications of supplemental argon laser endophotocoagulation were placed focally around the retinal breaks and in a scatter pattern for 360 degrees. Thirty months after the operation, the visual acuity with aphakic correction in the left eye was
20/100. The retina was completely attached with a 360-degree scierai buckling effect and residual white fibrous tissue in the periphery (Fig. 3). Granular depigmentation in the macula and moderate disk pallor were present. A similar course of equatorial fibrovascular proliferation developed in the right eye (Tables 1 and 2).
RESULTS THE PATIENTS IN THE PRESENT STUDY REPRESENT 3.3%
(eight of 245) of the patients and 3.3% (nine of 276) of the eyes on which we performed vitrectomy for complications of diabetic retinopathy during the study period. The clinical findings, surgical proce dures, and visual outcome in these patients are summarized in Tables 1 and 2. The preoperative and postoperative fundus appearances of an additional
TABLE 1 CLINICAL FINDINGS AND SURGICAL PROCEDURES IN NINE EYES OF EIGHT PATIENTS WITH SEVERE EQUATORIAL FIBROVASCULAR PROLIFERATION NO. OF CLOCK HOURS OF PATIENT NO-
EQUATORIAL
AGE (VRS},
PROLIFERATIVE TISSUE
GENDER
1,31, M 2, 55, F
39. M 36, F 5. 24, F
6, 55, F
7, 25, M, R.E.
LE.
INDICATIONS FOR SURGERY
Traction-rhegmatogenous retinal detachment Traction-rhegmatogenous retinal detachment, vitreous hemorrhage Traction retinal detachment Traction-rhegmatogenous retinal detachment Severe vitreous hemorrhage. traction retinal detachment Severe vitreous hemorrhage, traction retinal detachment Hypotony, traction retinal detachment, vitreous hemorrhage Traction retinal detachment,
MACULAR
SCLERAL
MEMBRANE
STATUS
BUCKLE
DISSECTION
LENSECTOMY
RETINECTOMY
3
Attached
Yes
No
No
No
3
Attached
Yes
No
No
No
6
Detached
Yes
No
No
No
6
Attached
Yes
Yes
No
No
12
Attached
Yes
Yes
No
No
12
Attached
Yes
Yes
No
No
2
Attached
Yes
No
Yes
Yes
12
Detached
Yes
Yes
Yes
No
9
Detached
Yes
Yes
Yes
Yes
RELAXING
vitreous hemorrhage 8, 31, F
Traction-rhegmatogenous retinal detachment
566
AMERICAN JOURNAL OF OPHTHALMOLOGY
M A Y 1995
TABLE 2 PREOPERATIVE AND POSTOPERATIVE VISUAL ACUITY, DURATION OF FOLLOW-UP, AND CAUSES FOR VISUAL LOSS
PATIENT NO, 1
2 3 4 5 6 7, R.E.
L.E. 8
VISUAL ACUITY PREOPERATIVE
POSTOPERATIVE
(MOS)
20/50 20/100 Hand motions 20/40 20/80 20/200 20/200
39 52 6 7 14 6 24
2/200 Hand motions
20/100 Light perception
30 6
Primary indications for surgery, which coexisted with the severe equatorial fibrovascular proliferation in the nine study eyes, were traction-rhegmatogenous retinal detachment (four eyes), traction retinal de tachment with the macula involved (two eyes), severe vitreous hemorrhage (two eyes), and severe hypotony (one eye). One of the eyes with traction-rhegmato genous detachment had the macula detached. The No. s
FOLLOW-UP
20/40 5/200 20/100 20/40 4/200 Hand motions 5/200
patient (Case 5) are also shown in Figures 4 and 5. The mean patient age ± S.D. was 37 ± 12.2 years (range, 25 to 55 years). All but two were 39 years of age or younger. All had long-standing type I diabetes mellitus (duration, 18 to 48 years), with the possible exception of one 5 5-year-old patient who had devel oped insulin-requiring diabetes mellitus at age 35 years. Other systemic factors noted to be present included systemic hypertension (four patients), ciga rette smoking (three patients), nephropathy, alcohol consumption and marijuana smoking, coronary artery disease, and transient ischémie attacks (one patient each). There was no history to suggest ocular inflam matory disease, retinal vein occlusion, hemoglobinopathy, or collagen vascular disease in any of our patients. The duration of follow-up after the opera tion was six to 52 months (mean ± S.D., 20.4 ± 16.9 months). All patients but one (Case 5) had under gone previous panretinal scatter photocoagulation in their affected eyes.
VOL.119,
DURATION OF CAUSE OF REDUCED VISUAL ACUITY 320/200
Response resembling Coats' disease
Posterior subcapsular cataract Intraocular pressure: preoperative, 0 mm Hg; postoperative, 8 mm Hg; cystoid macular edema and epiretinal membrane History of chronic traction retinal detachment
eye with severe hypotony had a preoperative intraocu lar pressure of 0 mm Hg associated with peripheral traction detachment and vitreous hemorrhage. Two eyes with traction or traction-rhegmatogenous retinal detachment as a primary indication had coexisting vitreous hemorrhage, and another had a response resembling Coats' disease from peripheral traction retinal detachment. Although the primary indica tions listed above were thought to be compelling reasons for surgical intervention, the basis for surgery in nearly all patients included other factors, such as the density of vitreous hemorrhage, difficulty in applying scatter photocoagulation through hemor rhage, and the progressive nature of the proliferative process. Intraoperative complications consisted entirely of iatrogenic retinal break formation in four eyes (ex cluding those with relaxing retinectomy), which were usually associated with membrane dissection (Cases 4 through 7, left eyes). Two of these eyes also had retinal breaks that may have been related to instru ment insertion through the sclerotomy sites. Postop erative complications and related secondary proce dures are listed in Table 3. Anatomic success was obtained in nine eyes (100%), and complete retinal reattachment was ob tained in seven eyes (78%). Functional success (visu al acuity of 20/200 or better) was obtained in seven
VlTRECTOMY FOR PROLIFERATIVE DIABETIC RETINOPATHY
567
TABLE 3 POSTOPERATIVE COMPLICATIONS AND PROCEDURES PATIENT
NO. OF
NO.
COMPLICATION
EYES
2,3
Localized peripheral traction retinal detachment R.E.: Recurrent epiretinal membrane and cystoid macular edema Pupillary membrane Residual vitreous hemorrhage
2
7 8
e
eyes (78%). In the eye with severe preoperative hypotony (0 mm Hg), the intraocular pressure re turned to 8 mm Hg and has remained at that level for two years postoperatively. Poor visual outcome in the two eyes with functional failure was attributed to a persistent response resembling Coats' disease (one eye) and pre-existent long-standing retinal detach ment (one eye).
DISCUSSION IN 245 PATIENTS (276 EYES) UNDERGOING DIABETIC VIT-
rectomy, the presence of equatorialfibrovascularpro liferation sufficient to produce visually threatening complications in the absence of marked posterior traction was relatively rare, occurring in nine eyes of eight patients (3.3% of eyes and 3.3% of patients). The relative infrequency of this condition in our series may be explained by the preponderance of posterior neovascularization known to occur in proliferative diabetic retinopathy1,4,5 and by a decreased likelihood of visually threatening posterior retinal traction with peripheral disease. Equatorial fibrovascular proliferation appears to be clinically distinct from anterior hyaloidal fibrovascular proliferation,2'3 in which neovascularization occurs in the residual anterior vitreous and along the retrolenticular surface after previous diabetic vitrectomy. In most of our patients, the neovascular process ap peared to originate from the equatorial fundus and occurred before any ocular procedure, except for fundus photocoagulation. Similar pathophysiologic characteristics may be shared between these two conditions, as the neovascular process appears to 568
1 1 1
POSTOPERATIVE PROCEDURES
VISUAL ACUITY
None
20/100, hand motions Vitrectomy, membrane 20/200 stripping YAG laser Light perception Fluid-gas exchange Light perception
originate from the anterior retina in the presence of vitreous attachments in both conditions. The clinical course and anatomic alterations in equatorial fibrovascular proliferation differ from most diabetic surgical cases and pose certain dilemmas regarding the timing, risk-benefit ratio, and indica tions for surgery. The proliferative process occurs in the peripheral fundus where surgical access or visuali zation may be difficult and membrane dissection may more easily result in iatrogenic retinal tears. The surgical methods required to relieve such traction, such as scierai buckling, retinectomy, and lens remov al, may result in a higher degree of morbidity. Despite these considerations, surgical intervention appears to be indicated for the visually threatening complica tions of traction-rhegmatogenous retinal detachment, macula-threatening traction retinal detachment, and intractable hypotony. Surgical intervention should probably also be considered for progressive fibrovascu lar proliferation in the periphery, recurrent or severe vitreous hemorrhage, or anterior segment neovascu larization not responsive to panretinal photocoagulation. For all patients, the risks of surgery should be weighed against those of subsequent visual loss with out surgery and the possibility of stabilization with scatter photocoagulation therapy. In the Diabetic Retinopathy Vitrectomy Study, a multicenter randomized trial, early vitrectomy was shown to have a beneficial effect in selected eyes with severe vitreous hemorrhage6 or advanced fibrovascu lar proliferation.7 The Diabetic Retinopathy Vitrec tomy Study observed increased rates of visual loss during conventional, as opposed to early surgical, treatment of extramacular traction retinal detach ment with increased size of detachment and activity
AMERICAN JOURNAL OF OPHTHALMOLOGY
MAY
1995
of neovascularization.8 These factors may be useful in determining the potential benefit of earlier vitrectomy in this condition. However, direct application of findings from the Diabetic Retinopathy Vitrectomy Study to equatorialfibrovascularproliferation may not be appropriate, since that study did not specifically address disease limited to the periphery. Thus, we currently recommend an individualized approach to surgical intervention in cases of equatorial fibrovascu lar proliferation. In early cases of equatorial fibrovascular prolifera tion without immediate threat to visual function, scatter photocoagulation may reduce the activity of the proliferative process and should be considered before vitrectomy. Peripheral scatter photocoagulation may also create a generalized chorioretinal adhe sion, which may help to stabilize the peripheral retina if vitrectomy is eventually required. Notably, previous scatter photocoagulation has been associated with a better prognosis after vitrectomy for diabetic traction retinal detachment.9 In recalcitrant cases, the results of our study indicate that subsequent, visually threat ening complications of equatorialfibrovascularprolif eration may be treated effectively with the surgical modalities described herein. The need for scierai buckling in these cases is a notable deviation from the conventional approach to the tractional complications of diabetic retinopathy that involve primarily the posterior pole.10'15 After a difficult peripheral membrane dissection, scierai buck ling may relieve additional traction and close retinal breaks. It was performed in all of our patients and likely contributed to a high anatomic success rate. Scierai buckling alone may not be sufficient to relieve traction, however. Other modalities, such as mem brane dissection, retinectomy, and lensectomy, may be required to achieve retinal reattachment. After diabetic vitrectomy, an increased risk of anterior hyaloidal fibrovascular proliferation or ante rior segment neovascularization may occur, especially in markedly ischémie or lensectomized16 eyes. Intra operative panretinal photocoagulation may possibly reduce this risk and stabilize the proliferative process as well. Nevertheless, lensectomy should be per formed only if necessary to allow for adequate relief of retinal traction. Notably, in none of our patients did anterior hyaloidal fibrovascular proliferation or anterior segment neovascularization develop after VOL.119, N O . 5
vitrectomy, possibly because all patients underwent extensive panretinal photocoagulation previously or concomitantly with the vitrectomy procedure. It is not clear to us why the equatorial proliferation was so prominent in these patients. Both systemic and anatomic considerations may have been involved. Most patients were young, had long-standing type I diabetes, and had either generalized vascular disease or its risk factors. Systemic hypertension was present in one half of the patients. In all patients, vitreous traction and attachment of the posterior hyaloid to the peripheral retina were present. In at least one patient (Case 7) there was also a thin sheet of residual posterior cortical vitreous attached throughout the posterior pole, such that posterior detachment of the vitreous did not appear to be required for the periph eral traction to occur. Factors that might be hypothe sized to influence the location of neovascular prolife ration in diabetic retinopathy include previous photocoagulation, extent and location of retinal or ocular ischemia, midperipheral microvascular chang es,17 location of arteriovenous crossings,4 and varia tions in the posterior cortical vitreous pocket.18 In summary, we studied a subgroup of patients with diabetic retinopathy whose neovascular proliferation was limited primarily to the equatorial fundus. We have designated this condition "equatorial fibrovascu lar proliferation" and prefer the term "equatorial" to equally valid descriptive terms such as "anterior," "preretinal," or "peripheral," to avoid confusion with anterior hyaloidal fibrovascular proliferation, a dis tinctly different condition with a much less favorable prognosis.2,3 The surgical approach to equatorial fibro vascular proliferation differed from conventional ap proaches to diabetic retinopathy. In these difficult cases, our techniques for relief of peripheral traction resulted in rates of anatomic and functional success of 100% and 78%, respectively, which are comparable to those of vitrectomy for diabetic traction retinal de tachments in general.9'15'19,20
REFERENCES 1. Michels RG. Proliferative diabetic retinopathy: pathophysiology of extraretinal complications and principles of vitreous surgery. Retina 1981;1:1-17. 2. Lewis H, Abrams GW, Foos RY. Clinicopathologic findings
VITRECTOMY FOR PROLIFERATIVE DIABETIC RETINOPATHY
569
3. 4. 5. 6.
7.
8.
9.
10.
in anterior hyaloidal fibrovascular proliferation after diabetic vitrectomy. Am ] Ophthalmol 1987;104:614-8. Lewis H, Abrams GW, Williams GA. Anterior hyaloidal fibrovascular proliferation after diabetic vitrectomy. Am ] Ophthalmol 1987;104:607-13. Taylor E, Dobree JH. Proliferative diabetic retinopathy: site and size of initial lesions. Br J Ophthalmol 1970;54:11-8. Davis MD. Vitreous contraction in proliferative diabetic retinopathy. Arch Ophthalmol 1965;74:741-51. Diabetic Retinopathy Vitrectomy Study Research Group. Early vitrectomy for severe vitreous hemorrhage in diabetic retinopathy: two year results of a randomized trial: Diabetic Retinopathy Vitrectomy Study report 2. Arch Ophthalmol 1985;103:1644-52. . Early vitrectomy for severe proliferative diabetic retinopathy in eyes with useful vision: results of a randomized trial: Diabetic Retinopathy Vitrectomy Study report 3. Oph thalmology 1988;95:1307-20. . Two-year course of visual acuity in severe prolifera tive diabetic retinopathy with conventional management: Diabetic Retinopathy Vitrectomy Study (DRVS) report 1. Ophthalmology 1985;92:492-502. Thompson JT, deBustros S, Michels RG, Rice TA. Results and prognostic factors in vitrectomy for diabetic traction retinal detachment of the macula. Arch Ophthalmol 1987;105:497-502. Meredith TA, Kaplan HJ, Aaberg TM. Pars plana vitrectomy techniques for relief of epiretinal traction by membrane segmentation. Am ] Ophthalmol 1980;89:408-13.
570
11. Aaberg TM. Pars plana vitrectomy for diabetic traction retinal detachment. Ophthalmology 1981;88:639-42. 12. Charles S. Vitreous microsurgery. Baltimore: Williams & Wilkins, 1981:107-20. 13. Abrams GW, Williams GA. "En bloc" excision of diabetic membranes. Am J Ophthalmol 1987;103:302-8. 14. Tolentino FI, Freeman HM, Tolentino FL. Closed vitrectomy in the management of diabetic traction retinal detachment. Ophthalmology 1980;87:1078-89. 15. Aaberg TM. Clinical results in vitrectomy for diabetic traction retinal detachment. Am J Ophthalmol 1979; 88:246-53. 16. Blankenship G, Cortez R, Machemer R. The lens and pars plana vitrectomy for diabetic retinopathy complications. Arch Ophthalmol 1979;97:1263-7. 17. Shimizu K, Kobayashi Y, Muraoka K. Midperipheral fundus involvement in diabetic retinopathy. Ophthalmology 1981;88:601-12. 18. Kishi S, Shimizu K. Clinical manifestations of posterior precortical vitreous pocket in proliferative diabetic retinopa thy. Ophthalmology 1993;100:225-9. 19. Han DP, Murphy ML, Mieler WF. A modified en bloc excision technique during vitrectomy for diabetic traction retinal detachment: results and complications. Ophthalmolo gy 1994;101:803-8. 20. Williams DF, Williams GA, Hartz A, Mieler WF, Abrams GW, Aaberg TM. Results of vitrectomy for diabetic traction retinal detachments using the en bloc excision technique. Ophthalmology 1989;96:752-8.
AMERICAN JOURNAL OF OPHTHALMOLOGY
MAY
1995