VITRECTOMY FOR NONDIABETIC VITREOUS HEMORRHAGE

RAY T. OYAKAWA, M.D. , R O N A L D G. M I C H E L S , M.D. , AND W I L L I A M P. BLASE, M.D.

Baltimore, Maryland

In a group of 94 eyes with nondiabetic vitreous hemorrhage that underwent pars plana vitrectomy between March 1974 and September 1982, the causes of the hemorrhages were retinal branch vein obstruc­tion (36 eyes), blunt trauma (11 eyes), cataract extraction (ten eyes), subretinal neovascularization (nine eyes), Eales' disease (eight eyes), Terson's syndrome (four eyes), and idiopathic (five eyes) and miscella­neous (11 eyes) conditions.

Vision was improved postoperatively in 88 of the 94 eyes, including all of those that underwent blunt trauma or cataract extraction, those with Eales' disease and Terson's syndrome, and those in the idiopathic and miscellaneous groups. Final visual acuities, which depended pri­marily on the underlying condition and its effect on the macula, were 20/20 or better in ten eyes, 20/25 to 20/40 in 37 eyes, 20/50 to 20/200 in 26 eyes, 20/300 to 20/400 in 11 eyes, 9/200 to 5/200 in three eyes, and hand movements or light perception in seven eyes.

Retinal tears, the most common surgical complication, occurred in 18 eyes. The incidence of anterior retinal tears was reduced from 11% (11 of 38 eyes) to 4% (two of 56 eyes) after we began using a vitrectomy probe with a smaller diameter. Some postoperative lens opacification occurred in 16 of 50 phaldc eyes, and the incidence of later lens opacification increased as the follow-up lengthened.

Removal of nonclearing vitreous hem­orrhage is an established indication for vitrectomy. Most eyes undergoing vitrec­tomy for vitreous hemorrhage have pro-liferative diabetic retinopathy, and the results of surgery in these cases have been reported.1"4 Vitreous hemorrhage is easier to treat by vitrectomy and has a better prognosis in nondiabetic patients than in diabetic patients. We reviewed

Accepted for publication Aug. 1, 1983. From the Department of Ophthalmology, Johns

Hopkins University School of Medicine, Baltimore, Maryland.

Reprint requests to Ronald G. Michels, M.D., 127 Maumenee Bldg., Johns Hopkins Hospital, 600 N. Wolfe St., Baltimore, MD 21205.

our experience with vitrectomy for non­diabetic vitreous hemorrhage in 94 cases.

SUBJECTS AND M E T H O D S

The review included all vitrectomies for nondiabetic vitreous hemorrhage per­formed by one of us (R.G.M.) at the Wilmer Institute between March 1974 and September 1982. We excluded cases of vitreous hemorrhage from penetrating injuries because of the extensive intraoc­ular scarring often associated with pene­trating wounds. We-also excluded cases of preoperative rhegmatogenous retinal de­tachment or traction detachment involv­ing the macula. Of the 94 cases included in the review, 36 resulted from retinal branch vein occlusion, 11 from blunt

©AMERICAN JOURNAL OF OPHTHALMOLOGY 96:517-525, 1983 517

518 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER, 1983

trauma, ten from cataract extraction, nine from subretinal neovascularization, eight from Eales' disease, four from Terson's syndrome, five from idiopathic causes, and 11 from miscellaneous causes.

The patients included 56 males and 35 females, ranging in age from 8 to 96 years (mean age, 56 years). Three patients un­derwent surgery on both eyes at different times. The series included 50 right eyes and 44 left eyes. Follow-up periods ranged from six months to 84 months, except for two patients who died less than six months after the operation. The mean follow-up interval was 19.5 months.

We reviewed each patient's chart to identify preoperative, surgical, and post­operative features. Data were recorded regarding the preoperative intraocular pressure and the status of the cornea, iris, lens, and vitreous. The preoperative con­dition of the retina was often ascertained during the operation. Exact features of the surgical technique were recorded, including the occurrence of any complica­tions. Postoperative data included the best corrected final visual acuity, the status of the cornea, iris, and lens, the intraocular pressure, the clarity of the media, and the condition of the reti­na. Any patient who had not been exam­ined at least six months after the opera­tion was requested to return for a follow-up examination. We obtained cur­rent information from the referring oph­thalmologist if the patient could not re­turn for examination.

A total of 38 operations were per­formed with the Douvas Roto-extractor from 1974 through 1976. The O'Malley Ocutome was used in 56 operations from 1977 until September 1982. Sixteen of the 94 eyes were aphakic before vitrec-tomy. Lensectomy was performed at the time of vitrectomy in 28 of 78 phakic eyes. Peripheral retinal cryotherapy was applied for 360 degrees as a prophylactic measure in 25 cases. This was combined

with prophylactic use of an encircling band to support the posterior aspect of the vitreous base in 22 eyes. Most of the eyes that had prophylactic retinal cryo­therapy and scierai buckling were operat­ed on during the early part of this series; these two treatment methods were later discontinued for prophylactic use. In 18 additional cases, a scierai buckling proce­dure was done for a specific indication. Transvitreal endophotocoagulation was used to treat neovascularization in three cases toward the end of the series.

R E S U L T S

The final visual acuities improved in 88 of the 94 eyes. We defined improvement as a gain of five lines or more on the Snellen chart. Of the 94 eyes, 47 achieved final visual acuities of 20/40 or better. An additional 26 eyes obtained final visual acuities between 20/50 and 20/200 and 87 eyes obtained final visual acuities of 5/200 or better. The postopera­tive visual acuity often depended on the underlying abnormality (Table 1). For ex­ample, eight of ten eyes with vitreous hemorrhage after cataract extraction and each of the four eyes with Terson's syn­drome had final visual acuities of 20/40 or better, whereas seven of nine eyes with vitreous hemorrhage from subretinal neovascularization in the macula or cho-roidal hemorrhage, or both, had final visual acuities of 20/300 or worse. The postoperative visual acuity tended to re­main stable unless a late complication developed.

Retinal branch vein obstruction—In 34 patients (36 eyes) retinal branch vein obstruction accounted for the vitreous hemorrhage. The patients ranged in age from 32 to 75 years (mean age, 63 years). In the 33 eyes for which the duration of vitreous hemorrhage was known, it ranged from one to 60 months (mean duration, 20 months).

Preoperative visual acuities were light

VOL. 96, NO. 4 NONDIABETIC VITREOUS HEMORRHAGE 519

TABLE 1 BEST POSTOPERATIVE VISUAL ACUITIES

Preoperative Condition

All eyes Retinal branch vein occlusion Blunt trauma Cataract extraction Subretinal neovascularization Eales' disease Terson's syndrome Idiopathic Miscellaneous

20/15 to 20/20

10 2 1 3 0 2 1 0 1

Postoperative 20/25 to

20/40

37 11 5 5 1 4 3 4 4

20/50 to 20/200

26 13 2 1 1 2 0 1 6

Visual Acuities 20/300 to

20/400

11 5 2 1 3 0 0 0 0

9/200 to 5/200

3 1 1 0 1 0 0 0 0

H. M. or L. P.*

7 4 0 0 3 0 0 0 0

No. of Eyes Improved

88 of 94 33 of 36 11 of 11 10 of 10 6 of 9 8 of 8 4 of 4 5 of 5

11 of 11

*H. M., hand movements; L. P., light perception.

perception in 12 eyes, hand movements in 18 eyes, 2/200 to 20/400 in four eyes, 20/200 in one eye, and 20/70 in one eye. The lens was clear preoperatively in 14 of the 36 eyes. Fourteen eyes had mild preoperative cataracts, three eyes had moderate cataracts, and two eyes had mature cataracts. Three eyes were aphakic preoperatively. Seven eyes had preoperative glaucoma, two eyes had hetero-chromia, and one eye had rubeo-sis iridis without increased intraocular pressure.

The Douvas Roto-Extractor was used in 14 of the 36 cases and the O'Malley Ocutome in the other 22 cases. Twelve eyes underwent pars plana lensectomy at the time of vitrectomy. Prophylactic pe­ripheral retinal cryotherapy (360 degrees in extent) combined with an encircling No. 240 silicone band was used in seven eyes. Cryotherapy was applied to the peripheral retina just posterior to the sclerotomies in two other cases. In five cases, an encircling band with or without a grooved silicone exoplant was used be­cause of iatrogenic retinal tears near the sclerotomy. In three recent cases, trans-vitreal endophotocoagulation was used to treat areas of retinal neovascularization. Follow-up periods ranged from six to 84

months (mean, 24 months). Final visual acuities were improved in 33 of the 36 eyes.

Blunt trauma—Eleven eyes had vitre­ous hemorrhage from blunt trauma. The patients ranged in age from 8 to 48 years (mean age, 31 years). The duration of vitreous hemorrhage ranged from three to 18 months (mean duration, nine months). Preoperative visual acuities were light perception in four eyes, hand movements in five eyes, 20/100 in one eye, and 20/60 in one eye. The lens was clear in nine eyes and moderate cataracts were present in two eyes.

The Douvas Roto-Extractor was used in three eyes and the O'Malley Ocutome in the other eight eyes. A pars plana lensectomy was done in the two eyes with cataracts. Four eyes underwent prophy­lactic peripheral retinal cryotherapy (360 degrees in extent) with an encircling No. 240 silicone rubber band. One eye had an iatrogenic retinal dialysis near the scle­rotomy; this was treated with cryothera­py, an encircling band, and a broader exoplant in the region of the retinal break. Follow-up periods ranged from six to 53 months (mean, 26 months). Final visual acuities were improved in all 11 eyes.

520 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER, 1983

Hemorrhage during cataract surgery Ten eyes had vitreous hemorrhage

complicating cataract extraction. The pa­tients ranged in age from 34 to 69 years (mean age, 68 years). The duration of vitreous hemorrhage ranged from one to 21 months (mean, seven months). Preop-erative visual acuities were light percep­tion in four eyes, hand movements in five eyes, and 20/60 in one eye. Four eyes had blood-induced glaucoma.5 One eye had an intraocular lens.

The Douvas Roto-Extractor was used in five eyes and the O'Malley Ocutome in the other five eyes. Prophylactic periph­eral retinal cryotherapy (360 degrees in extent) was applied in one case. Retinal cryotherapy, a No. 240 silicone band with a grooved exoplant, and intraocular air were used in four eyes to treat iatrogenic anterior retinal breaks. An iridodialysis that had occurred at the time of cataract extraction was repaired in one eye. Follow-up periods ranged from four to 44 months (mean, 17 months). The single patient with less than six months of follow-up died four months after the op­eration. Final visual acuities were im­proved in all ten eyes.

Subretinal neovascularization—Nine eyes had vitreous hemorrhage from sub-retinal neovascularization or choroidal hemorrhage. In seven eyes the subretinal neovascularization was the result of se­nile macular degeneration. Neovascular­ization resulted from presumed ocular histoplasmosis in one eye, and the re­maining case had had a choroidal hemor­rhage during cataract surgery. The pa­tients ranged in age from 56 to 96 years (mean age, 73 years). The duration of vitreous hemorrhage before vitrectomy ranged from six to 72 months (mean, 30 months). The preoperative visual acuities were light perception in five eyes and hand movements in the other four. Three eyes had mild or moderate cataracts pre-operatively, two eyes were aphakic, and

the lenses were clear in the other four eyes.

The Douvas Roto-Extractor was used in two eyes and the O'Malley Ocutome in the other seven eyes. Two eyes had pars plana lensectomies. Prophylactic periph­eral retinal cryotherapy (360 degrees in extent) was used in one eye. Another eye was treated with peripheral retinal cryo­therapy, an encircling No. 240 silicone rubber band, and a localized broader exoplant to treat a retinal dialysis posteri­or to a sclerotomy site. Follow-up periods in this group ranged from two to 12 months (mean, ten months). The one patient with less than six months of follow-up died two ir/onths after the oper­ation. Final visual acuities were im­proved in six of the nine eyes.

Eales' disease—Eight eyes had vitreous hemorrhage from Eales' disease. The pa­tients ranged in age from 24 to 56 years (mean age, 41 years). In the seven cases in which the duration of vitreous hemor­rhage was known, it ranged from 12 to 103 months (mean duration, 59 months). The preoperative visual acuities were hand movements in five eyes, 20/400 in two eyes, and 5/200 in one eye. Four eyes had moderate cataracts preoperatively, and the lenses were clear in the other four eyes.

The Douvas Roto-Extractor was used in six cases and the O'Malley Ocutome in the other two. Six eyes underwent pars plana lensectomy. Five eyes had prophyl­actic peripheral retinal cryotherapy (360 degrees in extent), combined with an encircling No. 240 silicone rubber band in four eyes. Follow-up periods ranged from six to 43 months (mean, 13 months). Final visual acuities were improved in all eight eyes.

Idiopathic vitreous hemorrhage—Five eyes had idiopathic vitreous hemorrhage preoperatively. The patients ranged in age from 56 to 85 years (mean age, 71 years). The duration of hemorrhage

VOL. 96, NO. 4 NONDIABETIC VITREOUS HEMORRHAGE 521

ranged from nine to 19 months (mean, 13 months). Preoperative visual acuities were hand movements in four eyes and 2/200 in one eye. Two eyes had moderate cataracts preoperatively, one eye had a mild cataract, and the lenses were clear in the remaining two eyes.

The O'Malley Ocutome was used in all five cases. A pars plana lensectomy was performed in two cases. Follow-up peri­ods ranged from 11 to 19 months (mean, 16 months). Final visual acuities were improved in all five eyes and four had visual acuities of 20/40 or better.

Terson's syndrome—Four eyes (three patients) had vitreous hemorrhage from Terson's syndrome. The patients were 48, 49, and 53 years old. The duration of vitreous hemorrhage ranged from two to 13 months (mean, seven months). The preoperative visual acuities were hand movements in three eyes and 20/60 in one eye. The lens was clear in each case.

The Douvas Roto-Extractor was used in three eyes and the O'Malley Ocutome in one eye. Two eyes were treated with prophylactic peripheral retinal cryothera-py (360 degrees in extent) and an encir­cling No. 240 silicone rubber band. An­other eye was treated with cryotherapy, a No. 240 silicone band, and a localized broader exoplant to treat an iatrogenic anterior retinal tear near the sclerotomy site. Follow-up periods ranged from seven to 18 months (mean, 14 months). Final visual acuities were improved in all four eyes.

Miscellaneous conditions—In 11 other cases, vitreous hemorrhage was caused by a variety of conditions (Table 2). Follow-up periods ranged from six to 40 months (mean, 12 months). Final visual acuities were improved in all 11 eyes.

Complications—The only surgical com­plications were iatrogenic retinal tears or detachment, or both. Thirteen of the 94 eyes had anterior retinal tears near the sclerotomy site. Five eyes developed

posterior retinal tears. The incidence of anterior retinal tears decreased markedly when we changed from the Douvas Roto-Extractor to the O'Malley Ocutome. Anterior retinal tears occurred in 11 of 38 eyes operated on with the Douvas Roto-Extractor, and in two of 56 eyes operated on with the O'Malley Ocutome (P = .01). All of these retinal tears were recognized during surgery and were treated success­fully, although one eye required a second retinal reattachment operation.

Cataract formation was the most impor­tant postoperative complication. Some postoperative lens opacities occurred in 16 of the 50 phakic eyes. The incidence and severity of lens opacification in­creased with increasing length of follow-up. Four of 23 phakic eyes followed up for six to 12 months had some postoperative cataract formation; this incidence in­creased to six of 14 eyes followed up for 13 to 24 months and to six of 13 eyes followed up for more than two years.

Other postoperative complications were less frequent. Late retinal detach­ment occurred in three of the 94 eyes. These detachments were successfully treated by additional surgery in each case. Recurrent vitreous hemorrhage and epiretinal membrane formation each oc­curred in five of the 94 eyes. Phthisis bulbi and rubeosis iridis with secondary glaucoma each occurred in three of the eyes.

DISCUSSION

Removal of nonclearing vitreous hem­orrhage was the first widely accepted indication for pars plana vitreous surgery. Most patients undergoing vitrectomy for vitreous hemorrhage have proliferative diabetic retinopathy, and all other causes of vitreous hemorrhage combined are an infrequent indication for surgery.

The objectives of surgery include re­moval of all axial opacities and excision of the posterior vitreous surface. The diffi-

TABL

E 2

CLI

NIC

AL

SUM

MA

RIES

OF

11 C

ASE

S O

F V

ITR

EOU

S H

EMO

RR

HA

GE

FRO

M M

ISC

ELLA

NEO

US

CAU

SES

Patie

nt N

o.,

Sex,

Age

(yrs)

1, M

, 78

2, M

, 59

3, M

, 32

4, M

, 67

g 5,

M, 6

2 to

6, M

, 53

7, F

, 59

8, M

, 66

9, M

, 24

10, F

, 63

11

, F,

60

Vitr

eous

Hem

orrh

age

Dur

atio

n (m

os)

2 37

10

17

— 9 36

43 6 12 2

Cau

se

Cen

tral r

etin

al v

ein

obst

ruct

ion

Seve

re s

yste

mic

hy

perte

nsio

n D

rug

abus

e A

ntic

oagu

lant

s O

paqu

e sh

eets

of

cond

ense

d vi

treou

s

Leuk

emia

Cy

clod

ialy

sis

surg

ery

Avu

lsed

retin

al

vess

el

Beh

çet's

dis

ease

U

veiti

s Re

tinal

tea

r

Oth

er c

ondi

tions

—

Mod

erat

e ca

tara

ct

—

—

Mod

erat

e ca

tara

ct

Parti

al a

mbl

yopi

a G

lauc

oma

Mod

erat

e ca

tara

ct

—

Mod

erat

e ca

tara

ct

"

Surg

ical

Pro

cedu

re

in A

dditi

on t

o V

itrec

tom

y

—

Pars

pla

na l

ense

ctom

y

—

—

Pars

pla

na l

ense

ctom

y

—

—

Pars

pla

na le

nsec

tom

y

—

Pars

pla

na l

ense

ctom

y C

ryot

hera

py a

nd

scie

rai b

uckl

e

Visu

al

Preo

pera

tive

L. P

.

L. P

.

L. P

. L.

P.

4/20

0

H.

M.

H.

M.

20/2

00

4/20

0 H

. M

. H

. M

.

Acu

ity*

Posto

pera

tive

20/2

00

20/1

00

20/2

00

20/1

5 20

/200

20/7

0 20

/20

20/3

0

20/4

0 20

/50

20/4

0

Rem

arks

Die

d 2

mos

afte

r su

rger

y

—

—

—

Orig

inal

vitr

eous

hem

orrh

age

was

follo

wed

by

a re

tinal

de

tach

men

t th

at w

as r

epai

red

by s

cier

ai b

uckl

e —

—

—

—

—

*L. P

., lig

ht p

erce

ptio

n; H

. M

., ha

nd m

ovem

ents

.

VOL. 96, NO. 4 NONDIABETIC VITREOUS HEMORRHAGE 523

culty of the operation depends largely on the complexity of the posterior vitreoreti­nal anatomy. Eyes with proliferative dia­betic retinopathy often have extensive posterior vitreoretinal adhesions, where­as eyes with vitreous hemorrhage from other causes usually have less complicat­ed vitreoretinal relationships. Sometimes posterior vitreous detachment is com­plete. In eyes with visually significant cataractous changes, vitrectomy is rou­tinely combined with pars plana lens removal with ultrasonic emulsification. Nonclotted blood in the preretinal space is removed by careful suction after exci­sion of most of the vitreous gel, including the posterior vitreous surface.

The results obtained in this series indi­cated that visual improvement can be achieved by vitrectomy in about 94% of eyes with nondiabetic vitreous hemor­rhage. The final level of vision, however, depends in part on the functional capabil­ity of the macula; and this is often affected by factors related to the underlying dis­ease. In this series, eyes with subretinal neovascular membranes affecting the macula had the least improvement in vision; eyes with idiopathic hemorrhage without previous retinal disease and eyes with Terson's syndrome or hemorrhage associated with cataract extraction had the best results.

The primary reason for vitrectomy is to improve vision by removing opacities of the ocular media. It is common practice to wait at least six months for possible spontaneous clearing before recommend­ing surgery. However, in some eyes blood-induced glaucoma that cannot be controlled by medical treatment may be an indication for earlier surgery.5 In still other cases, the severity of the vitreous hemorrhage or the formation of dense blood-stained sheets of condensed vitre­ous may be so great that the chance of spontaneous clearing is remote, and vit­reous surgery may be considered less

than six months after the hemorrhage occurred.

We use ultrasonography to study eyes with vitreous hemorrhage to judge the severity of the hemorrhage and to deter­mine whether there is evidence of poste­rior retinal detachment. We consider ret­inal detachment involving the posterior pole to be an indication for prompt vitrec­tomy, whereas we do not perform surgery as a prophylactic measure to prevent pos­sible vitreous changes that could later result in vitreoretinal traction, retinal break formation, and retinal detachment. However, all our patients with vitreous hemorrhage are instructed to return promptly for repeat ultrasonographic ex­amination if further visual loss occurs, because this may be a symptom of retinal detachment.

A major concern with eyes undergoing vitrectomy is the possibility of postopera­tive rubeosis iridis; this is a serious risk in eyes with proliferative diabetic retinopa­thy.6"9 This complication is limited almost exclusively to eyes with a retinal vascular disorder characterized by extensive areas of capillary nonperfusion, such as may be present in eyes with an extensive retinal branch vein obstruction or Eales' disease. In this series, postoperative rubeosis iridis occurred in three of 34 eyes with branch vein obstruction. The risk of rube­osis iridis is reduced substantially if the crystalline lens is not removed.6"9

Another concern is the possibility of surgical or postoperative complications. The major surgical complications are iat-rogenic retinal breaks; these occurred in 18 of the 94 eyes in this series. Posterior retinal breaks occurred in five of the 94 cases; these were limited to eyes with complicated vitreoretinal anatomy in which breaks occurred from excessive traction on posterior vitreoretinal attach­ments or during epiretinal membrane segmentation or removal. The incidence of posterior retinal breaks in nondiabetic

524 AMERICAN JOURNAL OF OPHTHALMOLOGY OCTOBER, 1983

patients is lower than that in diabetic patients, although the overall incidence of retinal breaks is about the same.10

Anterior retinal breaks posterior to the pars plana sclerotomy occurred in 13 of the 94 eyes in this series. However, the incidence of anterior retinal breaks was reduced substantially after we changed from the larger-diameter Douvas Roto-Extractor to the smaller O'Malley Ocu-tome. The 4% (two of 56 eyes) incidence of peripheral retinal breaks in eyes with nondiabetic vitreous hemorrhage operat­ed on with the O'Malley Ocutome was somewhat lower than the 8% (15 of 179 eyes) incidence of peripheral retinal breaks in diabetic eyes with the same instrument reported by Oyakawa and as­sociates.10 Still, in all cases it is necessary to examine the peripheral retina by indi­rect ophthalmoscopy (combined with scierai indentation) to identify any ante­rior retinal breaks and to permit prompt treatment if any are present.

The incidence of major postoperative complications was low in this series, ex­cept for the occurrence of postoperative lens opacities in 16 of 50 phakic eyes. The patients underwent surgery during a per­iod when important improvements were being made in instruments and surgical technique; with these improvements, the damaging effects on the lens may now be lower. However, others have also noted postoperative lens opacification after vit-rectomy for diabetic retinopathy11 and macular pucker.12 The exact cause of post­operative cataractous changes is un­known, but they may result partly from ocular hemosiderosis caused by the dam­aging metabolic effects of the previous vitreous hemorrhage. Still, the possibili­ty of progressive postoperative cataract­ous changes must be taken into account in phakic eyes being considered for vitreous surgery.

Despite the chance of surgical or post­

operative complications, vitreous surgery produces visual improvement in most eyes blinded by nonclearing vitreous hemorrhage. Our results were compara­ble to those of previous reports.13"19 How­ever, the final level of vision is likely to be below normal, and postoperative visu­al acuity is affected by any macular damage from the underlying condition or the effects of the intraocular blood. Aphakic eyes, or eyes that are undergo­ing combined pars plana lens removal and vitrectomy, are especially good can­didates for surgery, because the most common postoperative complication is progressive lens opacification. However, aphakic eyes may also have a higher risk of postoperative rubeosis iridis if there is an underlying retinal vascular disorder characterized by extensive capillary non-perfusion.

REFERENCES 1. Peyman, G. A., Huamonte, F. U., Goldberg,

M. F., Sanders, D. R., Nagpal, K. C., and Raichand, M.: Four hundred consecutive pars plana vitrectomies with the vitrophage. Arch. Ophthalmol. 96:45, 1978.

2. Blankenship, G. : Pars plana vitrectomy for dia­betic retinopathy. A report of eight years' experi­ence. Mod. Probl. Ophthalmol. 20:376, 1979.

3. Machemer, R., and Blankenship, G.: Vitrec­tomy for proliferative diabetic retinopathy associated with vitreous hemorrhage. Ophthalmology 88:643, 1981.

4. Michels, R. G., Rice, T. A., and Rice, E. F.: Vitrectomy for diabetic vitreous hemorrhage. Am. J. Ophthalmol. 95:12, 1983.

5. Brucker, A. J., Michels, R. G., and Green, W. R. : Pars plana vitrectomy in the management of blood-induced glaucoma with vitreous hemorrhage. Ann. Ophthalmol. 10:1427, 1978.

6. Blankenship, G., Cortez, R., and Machemer, R. : The lens and pars plana vitrectomy for diabetic retinopathy complications. Arch. Ophthalmol. 97:1263, 1979.

7. Blankenship. G. M.: The lens influence on diabetic vitrectomy results. Report of a prospective randomized study. Arch. Ophthalmol. 98:1296, 1980.

8. Aaberg, T. M.: Pars plana vitrectomy for dia­betic traction retinal detachment. Ophthalmology 88:639, 1981.

VOL. 96, NO. 4 NONDIABETIC VITREOUS HEMORRHAGE 525

9. Rice, T. A., Michels, R. G., Maguire, M. G., and Rice, E. F. : The effect of lensectomy on the incidence of iris neovascularization and neovascular glaucoma after vitrectomy for diabetic retinopathy. Am. J. Ophthalmol. 95:1, 1983.

10. Oyakawa, R. T., Schachat, A. P., Michels, R. G., and Rice, T. A. : Complications of vitreous surgery for diabetic retinopathy. I. Intraoperative complications. Ophthalmology 90:517, 1983.

11. Schachat, A. P., Oyakawa, R. T., Michels, R. G., and Rice, T. A.: Complications of vitreous surgery for diabetic retinopathy. II. Postoperative complications. Ophthalmology 90:522, 1983.

12. Michels, R. G.: Vitreous surgery for macular pucker. Am. J. Ophthalmol. 92:628, 1981.

13. Eifrig, D. E., Lockhart, D. L., Berglund, R. D., and Knobloch, W. H.: Pars plana vitrectomy. Ophthalmic Surg. 9:76, 1978.

14. Aaberg, T. M.: Results of 100 consecutive vitrectomy procedures. In McPherson, A. (ed.): New

and Controversial Aspects of Vitreoretinal Surgery. St. Louis, C. V. Mosby, 1977, p. 245.

15. Young, P. M., and Shea, M.: Pars plana vitrec­tomy. Can. J. Ophthalmol. 12:106, 1977.

16. Peyman, G. A., Huamonte, F. U., and Gold­berg, M. F.: One hundred consecutive pars plana vitrectomies using the vitrophage. Am. J. Ophthal­mol. 81:263, 1976.

17. Machemer, R., and Norton, E. W. D.: A new concept of vitreous surgery. 3. Indications and re­sults. Am. J. Ophthalmol. 74:1034, 1972.

18. Ryan, S. J., and Michels, R. G.: Pars plana vitrectomy. Indications and results in 100 cases. In McPherson, A. (ed.): New and Controversial Aspects of Vitreoretinal Surgery. St. Louis, C. V. Mosby, 1977, p. 250.

19. Kloti, R.: Indications for vitrectomy and re­sults in 115 cases. In McPherson, A. (ed.): New and Controversial Aspects of Vitreoretinal Surgery. St. Louis, C. V. Mosby, 1977, p. 237.