9
Peripheral Retinal Cryotherapy for Postvitrectomy Diabetic Vitreous Hemorrhage in Phakic Eyes KIMBERLY A. NEELY, MD, PHD, MARK W. SCROGGS, MD, AND BROOKS W. MCCUEN II, MD PURPOSE: To review the anatomic and visual outcomes of a consecutive series of phakic patients with postoperative diabetic vitreous hemorrhage (PDVH) who underwent revision vitrectomy with peripheral retinal cryotherapy. METHODS: We performed a retrospective chart review of consecutive phakic patients who under- went revision vitrectomy for PDVH who also received peripheral retinal cryotherapy. Final cor- rected visual acuities after revision vitrectomy with peripheral retinal cryotherapy were compared to corrected visual acuities before and at the time of PDVH. Anatomic outcomes such as retinal attachment, vitreous hemorrhage, iris neovascu- larization, lens opacity, and anterior hyaloidal neovascularization were considered. RESULTS: Nineteen (86%) of 22 eyes (21 patients) that underwent revision of vitrectomy and trans- scleral peripheral retinal cryotherapy for PDVH also received supplementary endolaser photocoagulation in the posterior pole. In 16 eyes (73%), no further vitreous hemorrhaging occurred after this procedure. In six eyes (27%), vitreous hemorrhage recurred after revision of vitrectomy and peripheral retinal cryotherapy but cleared spontaneously in three of these eyes. Of the three eyes with nonclearing recur- rent vitreous hemorrhage after revision of vitrectomy and peripheral retinal cryotherapy, the cause for the vitreous hemorrhage was known for two: severe, progressive anterior hyaloidal neovascularization. With a mean follow-up 6 SD of 6.8 6 5.1 months (range, 0.5 to 19.5 months), final corrected visual acuity after revision of vitrectomy and peripheral retinal cryotherapy for PDVH improved over preop- erative visual acuity (at which time vitreous hemor- rhage was present) in 18 eyes (82%) because of removal of vitreous hemorrhage from the visual axis. However, final visual acuity reached or exceeded pre-PDVH visual acuity in only five of the 15 eyes for which pre-PDVH visual acuity was known. CONCLUSION: For phakic eyes with nonclearing PDVH, peripheral retinal cryotherapy (often aug- mented, when possible, by additional posterior pole endolaser photocoagulation) may be used to supple- ment previous retinal ablative therapy during revi- sion of vitrectomy. This procedure leads to anatomic stabilization and visual improvement in the majority of eyes. Transscleral peripheral retinal cryotherapy is often feasible in situations (such as media opacity) that preclude use of peripheral retinal endolaser or indirect laser photocoagulation. (Am J Ophthalmol 1998;126:82–90. © 1998 by Elsevier Science Inc. All rights reserved.) P OSTOPERATIVE DIABETIC VITREOUS HEMOR- rhage (PDVH) occurs frequently in diabetic patients who undergo pars plana vitrectomy for complications of proliferative diabetic retinopa- thy. In some reported series of patients who under- Accepted for publication Nov 5, 1997. Department of Ophthalmology, The Pennsylvania State University College of Medicine, Milton S. Hershey Medical Center, Hershey, Pennsylvania (Dr Neely); and Washington Eye Clinic, Washington (Dr Scroggs) and Department of Ophthalmology, Duke University Eye Center, Durham (Dr McCuen), North Carolina. Reprint requests to Brooks W. McCuen II, MD, Duke University Eye Center, Box 3802, Durham, NC 27710; fax: (919) 681-6474; e-mail: [email protected] © 1998 BY ELSEVIER SCIENCE INC.ALL RIGHTS RESERVED. 82 0002-9394/98/$19.00 PII S0002-9394(98)00066-X

Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

Embed Size (px)

Citation preview

Page 1: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

Peripheral Retinal Cryotherapyfor Postvitrectomy Diabetic Vitreous

Hemorrhage in Phakic Eyes

KIMBERLY A. NEELY, MD, PHD, MARK W. SCROGGS, MD,AND BROOKS W. MCCUEN II, MD

● PURPOSE: To review the anatomic and visualoutcomes of a consecutive series of phakic patientswith postoperative diabetic vitreous hemorrhage(PDVH) who underwent revision vitrectomy withperipheral retinal cryotherapy.● METHODS: We performed a retrospective chartreview of consecutive phakic patients who under-went revision vitrectomy for PDVH who alsoreceived peripheral retinal cryotherapy. Final cor-rected visual acuities after revision vitrectomywith peripheral retinal cryotherapy were comparedto corrected visual acuities before and at the timeof PDVH. Anatomic outcomes such as retinalattachment, vitreous hemorrhage, iris neovascu-larization, lens opacity, and anterior hyaloidalneovascularization were considered.● RESULTS: Nineteen (86%) of 22 eyes (21 patients)that underwent revision of vitrectomy and trans-scleral peripheral retinal cryotherapy for PDVH alsoreceived supplementary endolaser photocoagulationin the posterior pole. In 16 eyes (73%), no furthervitreous hemorrhaging occurred after this procedure.In six eyes (27%), vitreous hemorrhage recurredafter revision of vitrectomy and peripheral retinalcryotherapy but cleared spontaneously in three ofthese eyes. Of the three eyes with nonclearing recur-

rent vitreous hemorrhage after revision of vitrectomyand peripheral retinal cryotherapy, the cause for thevitreous hemorrhage was known for two: severe,progressive anterior hyaloidal neovascularization.With a mean follow-up 6 SD of 6.8 6 5.1 months(range, 0.5 to 19.5 months), final corrected visualacuity after revision of vitrectomy and peripheralretinal cryotherapy for PDVH improved over preop-erative visual acuity (at which time vitreous hemor-rhage was present) in 18 eyes (82%) because ofremoval of vitreous hemorrhage from the visual axis.However, final visual acuity reached or exceededpre-PDVH visual acuity in only five of the 15 eyesfor which pre-PDVH visual acuity was known.● CONCLUSION: For phakic eyes with nonclearingPDVH, peripheral retinal cryotherapy (often aug-mented, when possible, by additional posterior poleendolaser photocoagulation) may be used to supple-ment previous retinal ablative therapy during revi-sion of vitrectomy. This procedure leads to anatomicstabilization and visual improvement in the majorityof eyes. Transscleral peripheral retinal cryotherapy isoften feasible in situations (such as media opacity)that preclude use of peripheral retinal endolaser orindirect laser photocoagulation. (Am J Ophthalmol1998;126:82–90. © 1998 by Elsevier Science Inc.All rights reserved.)

P OSTOPERATIVE DIABETIC VITREOUS HEMOR-

rhage (PDVH) occurs frequently in diabeticpatients who undergo pars plana vitrectomy

for complications of proliferative diabetic retinopa-thy. In some reported series of patients who under-

Accepted for publication Nov 5, 1997.Department of Ophthalmology, The Pennsylvania State University

College of Medicine, Milton S. Hershey Medical Center, Hershey,Pennsylvania (Dr Neely); and Washington Eye Clinic, Washington (DrScroggs) and Department of Ophthalmology, Duke University EyeCenter, Durham (Dr McCuen), North Carolina.

Reprint requests to Brooks W. McCuen II, MD, Duke University EyeCenter, Box 3802, Durham, NC 27710; fax: (919) 681-6474; e-mail:[email protected]

© 1998 BY ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED.82 0002-9394/98/$19.00PII S0002-9394(98)00066-X

Page 2: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

went vitrectomy for proliferative diabetic retinopa-thy, PDVH (either persistent or recurrent) occurredin over 50% of patients.1–3 In other series, theprevalence of PDVH was lower, 13% to 23%, butnonetheless substantial.4–7

Diabetic patients who develop PDVH may re-quire further surgery, such as revision of vitrectomy.At the time of revision vitrectomy for PDVH, thesurgeon generally adds further retinal ablative ther-apy to reduce the stimulus for retinal neovascular-ization. Additional endolaser may be applied in theposterior pole, depending on the extent of previouslaser treatment. In pseudophakic and aphakic pa-tients, it is possible to apply endolaser photocoagu-lation to the far peripheral retina also, especiallywith the aid of wide-field viewing systems and/orscleral indentation by the surgical assistant. Supple-mentation of peripheral retinal photocoagulation isa more difficult problem in phakic patients becauseof the risk of lens damage by the endolaser probe aslaser treatment is extended into the anterior periph-ery. Therefore, one of us (B.W.M.) has employedtransscleral peripheral retinal cryotherapy as themeans of achieving retinal ablation anterior to theequator in diabetic patients undergoing repeat vit-rectomy for PDVH.

This report summarizes the visual and anatomicresults of transscleral peripheral retinal cryotherapyapplied during revision of vitrectomy in phakicdiabetic patients with postoperative vitreous hem-orrhage following vitrectomy for proliferative dia-betic retinopathy.

PATIENTS AND METHODS

WE RETROSPECTIVELY REVIEWED THE RECORDS OF ALL

phakic diabetic patients who underwent pars planavitrectomy and transscleral peripheral retinal cryo-therapy by one of us (B.W.M.) between July 1989and November 1994. Twenty-four eyes of 22 con-secutive phakic patients underwent this procedure.Two eyes of two patients were excluded becauseperipheral retinal cryotherapy was performed duringthe initial, or primary, vitrectomy for proliferativediabetic retinopathy rather than during revision ofvitrectomy. The remaining 22 eyes of 21 patientswere strictly those eyes in which peripheral retinal

cryotherapy was applied during revision of vitrec-tomy, the indication for which was PDVH (eitherpersistent or recurrent) in all cases. Eyes wereexcluded when they were aphakic or pseudophakic,had a traction retinal detachment involving themacula at the time of revision of vitrectomy, or hadreceived peripheral retinal cryotherapy before therevision vitrectomy. Some patients underwent atleast one outpatient fluid-air exchange withoutsuccessful clearance of the postoperative vitreoushemorrhage before undergoing revision of vitrec-tomy with peripheral retinal cryotherapy (Table 1).

Our technique for transscleral peripheral retinalcryotherapy was as follows. After revision of vitrec-tomy, scleral plugs were placed. The retinal cryo-probe was applied to the exposed sclera, and theretina was viewed with either the indirect ophthal-moscope or the operating microscope in combina-tion with endoillumination. Confluent retinalfreezes were applied in two or three rows beginningat the ora serrata and extending posteriorly to theanterior equatorial region in all four quadrants.Approximately eight to 12 retinal freezes wereplaced in each quadrant. The endpoint for each280 C freeze was a frosted white lesion in the retinaestimated to be roughly 1,200 to 1,500 mm indiameter.

Some patients had undergone one or more priorvitrectomies before presenting to our clinic. There-fore, the first vitrectomy at our hospital may havefollowed one or two previous vitrectomies per-formed elsewhere on that eye. The revision vitrec-tomy at our hospital, during which transscleralperipheral retinal cryotherapy was applied for treat-ment of PDVH, will be referred to as the “indexvitrectomy” in the remainder of this report.

Baseline data included age and duration of dia-betes mellitus at the time of index vitrectomy,duration of PDVH before index vitrectomy, andnumber of previous vitrectomies and fluid-air ex-change procedures before index vitrectomy. Visualacuities with the patient’s current correction or withbest refraction were recorded before the vitreoushemorrhage that led to index vitrectomy, at thetime of vitreous hemorrhage just before index vit-rectomy, at 1 week after index vitrectomy, and atthe last recorded follow-up visit. The extent ofnuclear sclerosis (0 to 41) or posterior subcapsular

PERIPHERAL RETINAL CRYOTHERAPYVOL. 126, NO. 1 83

Page 3: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

opacification (0 to 41) before index vitrectomy andat last follow-up was recorded.

Snellen visual acuities were converted to decimalequivalents for statistical descriptions. For purposes ofanalysis, visual acuities of no light perception, lightperception, hand motions, and counting fingers wereassigned decimal values of 0.00002, 0.0002, 0.002, and0.04, respectively. Visual acuities were expressed aslogMAR equivalents for statistical comparisons.

Data analysis was performed with Abacus Concepts:Statview software (Abacus Concepts, Inc, Berkeley,California). Visual acuity comparisons were performedusing the Wilcoxon signed rank test. Alpha 5 0.05was the chosen significance level.

RESULTS

TWENTY-TWO REVISION VITRECTOMIES WITH TRANS-

scleral peripheral retinal cryotherapy (“index vitrec-

tomies”) were performed on 22 eyes of 21 patientsfor postoperative diabetic vitreous hemorrhage. Onepatient had blood-induced glaucoma in addition topostoperative vitreous hemorrhage.

Table 1 provides preoperative clinical informa-tion for the patients included in this study. Tenpatients were men. In one of the female patients,both eyes qualified for this study. The average age ofthe patients was 48 6 14 years (mean 6 SD; range,27 to 72 years). The average duration of diabetesmellitus at the time of index vitrectomy was 21 6 7years (range, 5 to 35 years). Nineteen patients wereon insulin, and half of the patients had a history ofhypertension.

All 22 eyes had undergone panretinal photoco-agulation either with the slit-lamp laser beforevitrectomy or with the endolaser at the time ofprevious vitrectomy. Nineteen eyes received addi-tional fill-in endolaser photocoagulation in theposterior pole during the index vitrectomy. For

TABLE 1. Preoperative Characteristics of Patients Who Underwent Vitrectomy and Peripheral Retinal Cryotherapy for PDVH

Patient, Age (yrs), Sex

Duration of

Diabetes Before

Index PPV (yrs)*

Duration of PDVH

Before Index PPV

(mos)

PRP Before

Index PPV

No. of PPVs

Before Index

PPV

No. of Fluid-Air

Exchanges Before

Index PPV

1, 44, F 34 13 Yes 2 N/A

2, 31, F 22 1.25 Yes 2 N/A

3, 61, M 21 1.5 Yes 1 2

4, 27, M 19 1.67 Yes 1 1

5, 70, F 5 1 Yes 2 N/A

6, 31, M 14 3.5 Yes 1 1

7, 27, M 23 1.0 Yes 2 1

8, 32, F 19 1.5 Yes 1 1

9, 37, M 35 2.5 Yes 1 3

10, 51, F 21 12.5 Yes 2 N/A

11, 32, M 17 3.0 Yes 1 1

12, 54, F 17 4.5 Yes 1 N/A

13, 57, M 30 2.0 Yes 4 N/A

14, 56, F (RE) 20 2.25 Yes 1 1

— (LE) 20 4.0 Yes 1 N/A

15, 57, M 10 1.25 Yes 1 N/A

16, 37, M 24 1.0 Yes 1 1

17, 62, F 25 4.25 Yes 1 N/A

18, 59, F 27 0.25 Yes 1 N/A

19, 46, F 15 6.0 Yes 1 1

20, 72, F 27 1.75 Yes 1 N/A

21, 53, M 26 5.0 Yes 1 2

PDVH 5 postvitrectomy diabetic vitreous hemorrhage; PPV 5 pars plana vitrectomy; PRP 5 panretinal photocoagulation; N/A 5 not

available.

*“Index PPV” refers to the revision vitrectomy during which panretinal exocryotherapy was applied.

AMERICAN JOURNAL OF OPHTHALMOLOGY84 JULY 1998

Page 4: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

these 19 patients, the average number of laser spotsplaced was 516 6 201 (range, 198 to 956) andestimated spot size was 700 mm. No eyes hadreceived peripheral retinal cryotherapy at any timebefore index vitrectomy.

The median number of vitrectomies each eye hadundergone before index vitrectomy was 1.0 (range,1 to 4). Sixteen eyes had undergone only one priorvitrectomy at the time of index vitrectomy. Of 11eyes known to have undergone outpatient fluid-airexchange for PDVH before index vitrectomy, themedian number of fluid–air exchange procedures pereye was 1.0 (range, 1 to 3). Eight eyes had only onefluid-air exchange before index vitrectomy.

The median duration of PDVH before indexvitrectomy was 2.1 months (range, 0.25 to 13months). The cause for PDVH was not establishedwith certainty before or during the index vitrectomyfor 19 eyes. Anterior hyaloidal fibrovascular prolif-eration was suspected as the cause of PDVH beforeindex vitrectomy in most eyes, especially whenfluid-gas exchange failed to clear the vitreous cavity.In two eyes, anterior fibrovascular proliferationworsened after index vitrectomy manifested asgrowth of fibrovascular tissue on the posterior lenscapsule.

Table 2 lists visual acuity data and postoperativeresults for the 21 patients in this study. The meanperiod of follow-up after index vitrectomy was 6.8 65.1 months (range, 0.5 to 19.5 months). Ten eyeshad 6 or more months of follow-up.

Six eyes of six patients had recurrent vitreoushemorrhage after index vitrectomy. These vitreoushemorrhages developed from 0.5 to 17 months(median, 1.3 months) after the index vitrectomy. Inthree of these eyes, vitreous hemorrhages clearedspontaneously. In a fourth eye, vitreous hemorrhagepersisted but was decreasing gradually at last follow-up. In the fifth and sixth eyes, recurrent vitreoushemorrhage was eventually followed by a cycliticmembrane with no light perception vision and bytotal retinal detachment with hand motions vision,respectively. Retinal detachment recurred in thislast eye despite two further vitrectomies, one withsilicone oil tamponade. Both of these latter two eyeshad severe anterior hyaloidal fibrovascular prolifer-ation.

Ten eyes of 10 patients had neovascularization of

the iris before index vitrectomy. The iris neovascu-larization resolved in nine of these 10 eyes afterindex vitrectomy but persisted in one eye. In one ofthe nine eyes in which iris neovascularization dis-appeared after index vitrectomy, anterior hyaloidalneovascularization later caused recurrent vitreoushemorrhage and iris retraction. Of the 12 eyes withno iris neovascularization preoperatively, new irisneovascularization appeared after index vitrectomyin five. One of these five eyes was the eye thatdeveloped a cyclitic membrane with hypotony 3months after index vitrectomy secondary to anteriorhyaloidal fibrovascular proliferation. No patient de-veloped neovascular glaucoma before or after indexvitrectomy.

Fourteen eyes of 13 patients developed increasedlens opacity after the index vitrectomy. Of these 14eyes, eight had increased nuclear sclerosis alone(increasing, on average, from trace before indexvitrectomy to 21 afterward), four had increasedposterior subcapsular opacity alone (increasing, onaverage, from trace before index vitrectomy to 11afterward), and two had an increase in both nuclearsclerosis and posterior subcapsular opacity. Three ofthese 14 eyes underwent lens extraction after indexvitrectomy, two because of lens opacity and onebecause of anterior hyaloidal fibrovascular prolifer-ation with cyclitic membrane formation.

Records of corrected visual acuity before thepostoperative vitreous hemorrhage that led to indexvitrectomy were available for 15 of the 22 eyes(Figure 1). Records of corrected visual acuity at thetime of vitreous hemorrhage leading to index vit-rectomy, at 1 week after index vitrectomy, and atlast follow-up after index vitrectomy were availablefor all 22 eyes.

Median corrected visual acuity before the postop-erative vitreous hemorrhage that led to index vit-rectomy was 20/200 (range, 3/200 to 20/15). At thetime of the postoperative vitreous hemorrhage thatled to index vitrectomy, all eyes had corrected visualacuities in the range of counting fingers to lightperception (Figure 1). One week after index vitrec-tomy, median corrected visual acuity was countingfingers (range, hand motions to 20/50). At finalfollow-up after index vitrectomy, median correctedvisual acuity was 9/200 (range, no light perceptionto 20/25).

PERIPHERAL RETINAL CRYOTHERAPYVOL. 126, NO. 1 85

Page 5: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

Comparison of corrected visual acuity at lastfollow-up to corrected visual acuity before PDVHgave the following results: four eyes had improvedvisual acuity, one eye had the same visual acuity,and 10 eyes had decreased visual acuity (Figure 2).This comparison could not be made for the seveneyes in which visual acuity before PDVH was notavailable. Of the 10 eyes with decreased visualacuity, nine had increased lens opacity after indexvitrectomy, but the degree of lens opacity was severeenough to account for the poor final visual acuity in

only one. For the 15 eyes in which both observa-tions were recorded, the difference between cor-rected visual acuity (expressed in logMAR form)before PDVH and corrected visual acuity at lastfollow-up after index vitrectomy was statisticallysignificant (Wilcoxon signed rank test, P 5 .028).

Comparison of corrected visual acuity at lastfollow-up to corrected visual acuity at the time ofPDVH just before index vitrectomy gave the follow-ing results: 18 eyes had improved visual acuity, oneeye had the same visual acuity, and three eyes had

TABLE 2. Results of Vitrectomy With Peripheral Retinal Cryotherapy for PDVH

Patient

VA

Before

Index

PPV

VA at Last

Follow-up

After

Index PPV

Additional

Endolaser

at Time of

Index

PPV (No.

of Burns)

Length of

Follow-up

After Index

PPV (mos)

Further VH After

Index PPV and

Postoperative

Time of VH

(mos) Final Occular Status

1 HM 3/200 Yes (350) 3.25 No Retina attached, no active NV

2 4/200 20/70 Yes (461) 13 No Retina attached, no active NV

3 HM HM Yes (780) 4.5 No Retina attached, no active NV

4 HM 20/40 Yes (690) 15.5 Yes, 5.25 Retina attached, VH after index PPV

cleared spontaneously

5 HM CF Yes (318) 4.5 No Retina attached, no active NV

6 HM LP Yes (483) 2.25 Yes, 0.5 Severe AHFVP; went on to total RD after

2 more PPVs, lensectomy, silicone oil

7 HM 20/50 Yes (692) 11 No Retina attached, no active NV

8 HM 20/70 Yes (844) 6.5 No Retina flat, no active NV

9 HM 20/25 Yes (258) 6.0 No Retina attached, no active NV

10 HM 20/80 Yes (451) 6.25 No Retina attached, no active NV, PCIOL

11 HM 20/30 Yes (577) 6.0 No Retina attached, no active NV, fine ERM

12 HM NLP Yes (326) 4.0 Yes, 1.5 Severe AHFVP, cyclitic membrane,

traction RD by ultrasound

13 CF 20/60 Yes (477) 2.0 No Retina attached, no active NV

14 (RE) HM 20/400 Yes (956) 7.0 No Retina attached, no active NV

— (LE) HM 1/200 Yes (527) 5.25 Yes, 1.0 Retina attached, VH after index PPV

cleared spontaneously, mild NVI

15 HM 4/200 No 16.5 No Retina attached, no active NV

16 HM 20/70 Yes (428) 1.0 No Retina attached, no active NV

17 HM 1/200 No 0.5 No Retina attached, no active NV

18 LP 4/200 No 4.25 Yes, 0.5 Retina attached, VH after index PPV

slowly decreased

19 HM 20/300 Yes (460) 5.0 No Retina attached, no active NV

20 HM LP Yes (198) 5.4 No Retina attached, dense cataract

21 1/200 5/200 Yes (545) 19.5 Yes, 17 Retina attached, no active NV, VH

occurred after PCIOL and then

resolved

PDVH 5 postoperative diabetic vitreous hemorrhage; VA 5 visual acuity; HM 5 hand motions; CF 5 counting fingers; LP 5 light

perception; NLP 5 no light perception; NV 5 neovascularization; VH 5 vitreous hemorrhage; AHFVP 5 anterior hyaloidal fibrovascular

proliferation; RD 5 retinal detachment; ERM 5 epiretinal membrane; NVI 5 neovascularization of the iris; PPV 5 pars plana vitrectomy;

PCIOL 5 posterior chamber intraocular lens.

AMERICAN JOURNAL OF OPHTHALMOLOGY86 JULY 1998

Page 6: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

decreased visual acuity (Figure 3). Data for thiscomparison was available for all 22 eyes. Thedifference between visual acuity at the time ofPDVH and visual acuity at last follow-up after indexvitrectomy was statistically significant (Wilcoxonsigned rank test, P 5 .002).

There was no statistically significant differencebetween corrected visual acuity 1 week after indexvitrectomy and corrected visual acuity at final fol-low-up (P 5 .30).

DISCUSSION

POSTOPERATIVE VITREOUS HEMORRHAGE OCCURS

frequently in patients who undergo pars plana vit-rectomy for complications of proliferative diabeticretinopathy.1–7 Postvitrectomy diabetic vitreoushemorrhage may be noted as early as the firstpostoperative day (persistent PDVH) or may occurin a delayed fashion weeks or months after surgery(recurrent PDVH). The prevalence of vitreous

hemorrhage on the first postoperative day amongreported series of diabetic patients undergoing vit-rectomy ranges up to 75%.1–3 The reported preva-lence of delayed postoperative vitreous hemorrhage(occurring 1 week or more after surgery) is between20% and 30%.1–3

Roughly one half of postoperative diabetic vitre-ous hemorrhages clear spontaneously.3 Managementof PDVH that does not clear includes outpatientfluid-air exchange, washout of the vitreous cavity, orrevision of vitrectomy.7–10 Martin and McCuen8

found that fluid-air exchange was effective in clear-ing postoperative diabetic vitreous hemorrhage inhalf of their patients. However, the remaining 50%of patients in their series required revision of vit-rectomy because of vitreous hemorrhage unrespon-sive to one or more outpatient fluid-air exchangeprocedures.

FIGURE 1. Corrected visual acuity measurements be-fore postoperative diabetic vitreous hemorrhage(PDVH), at the time of PDVH before index vitrec-tomy, 1 week after index vitrectomy, and at last fol-low-up after index vitrectomy. CF 5 counting fingers,HM 5 hand motions, LP 5 light perception, NLP 5 nolight perception, and PPV 5 pars plana vitrectomy.

FIGURE 2. The difference between last follow-upvisual acuity after index vitrectomy and visual acuitybefore vitreous hemorrhage that led to index vitrectomyfor 15 eyes. Visual acuities, measured with the patient’scurrent correction or with best correction by refraction,are expressed as decimal equivalents. Negative differ-ences indicate worsened visual acuity, and positivedifferences indicate improved visual acuity after indexvitrectomy. For seven of the 22 eyes in the study, thiscomparison could not be made because visual acuitiesbefore vitreous hemorrhage were not available.

PERIPHERAL RETINAL CRYOTHERAPYVOL. 126, NO. 1 87

Page 7: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

During revision of vitrectomy, the source ofpostoperative diabetic vitreous hemorrhage is rarelydiscovered.1–3,10 Proposed causes include residualretinal neovascularization, anterior hyaloidal neo-vascularization, bleeding from sclerotomies, fibro-vascular ingrowth at sclerotomies, or retinal tears.

To reduce the neovascular stimulus, the surgeonoften performs further retinal ablative therapy at thetime of revision vitrectomy. Usually, this is donewith the endolaser. Endolaser photocoagulationmay be added to the posterior pole or to retinaanterior to the equator. In phakic eyes, however, farperipheral retina is difficult to treat safely becausethe endolaser probe shaft may damage the lens. Thisis true even when scleral indentation or wide-fieldviewing systems are used for visualization of theperipheral retina.

In such situations, transscleral peripheral retinalcryotherapy may be used to treat peripheral reti-na.11,12 Because the cryoprobe remains external to

the eye, there is no risk of lens damage. Cryotherapymay be applied under direct visualization using theoperating microscope with illumination supplied bythe fiberoptic light source. When corneal edema,peripheral vitreous base opacification, or lens opac-ity prevents adequate viewing of the retina with theoperating microscope, both sclerotomies may beplugged and cryotherapy applied while the retina isviewed with the indirect ophthalmoscope. Witheither method of viewing, two to three rows ofconfluent cryotherapy may be applied starting at theora serrata in each quadrant.

Among our patients with PDVH, transscleralperipheral retinal cryotherapy with revision of vit-rectomy was successful in stopping further vitreoushemorrhaging in the majority of eyes (16/22 [73%]).Six eyes developed recurrent vitreous hemorrhageafter revision of vitrectomy with peripheral retinalcryotherapy, three of which cleared spontaneouslyso that the vitreous cavity was ultimately cleared in86% (19 of 22 eyes). Anterior hyaloidal neovascu-larization was present in two of the remaining threeeyes that did not clear spontaneously.

Although peripheral retinal cryotherapy addedsignificantly to the extent of retinal ablation, it didnot prevent further intraocular neovascularizationin all eyes. As noted above, two of 22 eyes (9%) hadcontinued active anterior hyaloidal neovasculariza-tion after peripheral retinal cryotherapy. Iris neo-vascularization present before peripheral retinalcryotherapy resolved in eight of 10 eyes. However,five of 12 eyes (42%) without iris neovascularizationbefore index vitrectomy developed iris neovascular-ization after peripheral retinal cryotherapy.

In this study, revision of vitrectomy with periph-eral retinal cryotherapy improved final correctedvisual acuity in 18 of 22 eyes (82%) compared to thepreoperative visual acuity when vitreous hemor-rhage was present. However, final corrected visualacuity measured 6.8 6 5.1 months after indexvitrectomy was worse than corrected visual acuitybefore PDVH in 10 of 15 eyes (67%). Increased lensopacity after index vitrectomy was common butrarely marked enough to account for the final visualacuity. We speculate that the decrease in final visualacuity compared to visual acuity before PDVH wasrelated to worsening retinal ischemia.

Fourteen of 22 (64%) eyes in this study developed

FIGURE 3. The difference between last follow-upvisual acuity after index vitrectomy and visual acuity atthe time of vitreous hemorrhage leading to index vit-rectomy for 22 eyes. Visual acuities, measured with thepatient’s current correction or with best correction byrefraction, are expressed as decimal equivalents. Nega-tive differences indicate worsened visual acuity andpositive differences indicate improved visual acuityafter index vitrectomy.

AMERICAN JOURNAL OF OPHTHALMOLOGY88 JULY 1998

Page 8: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

increased lens opacity after index vitrectomy. Thiswas vision-limiting in two eyes (9%), which thenunderwent cataract extraction. It is impossible toseparate the effects on the lens of fluid-air exchangebefore index vitrectomy from the effects of indexvitrectomy itself. Both procedures may increase lensopacity. Martin and McCuen8 reported increasedcataract in one of 18 eyes (6%) after fluid-airexchange for PDVH, whereas Han and associates9

reported increased cataract in 10 of 17 eyes (59%).Cataract formation is a well-known complication ofpars plana vitrectomy, with reported rates of approx-imately 30% after vitrectomy for proliferative dia-betic retinopathy.1 The likelihood of cataract for-mation after fluid-air exchange or vitrectomy maydepend on patient age also.

An alternative to transscleral peripheral retinalcryotherapy at the time of repeat vitrectomy forrecurrent PDVH is laser delivered with the indirectophthalmoscope. This may be used to extend lasertreatment anterior to the equator and even out to theora serrata, particularly when scleral depression is used.In our earlier cases, this technology was not yetavailable. Even now, however, use of the indirect lasermay be impossible in certain situations that preventadequate visualization of the peripheral retina. Suchsituations include dense peripheral lens opacity, cor-neal stromal edema, or residual hemorrhage in thevitreous base (as may occur in phakic patients). Trans-scleral peripheral retinal cryotherapy is slightly moreflexible than indirect laser photocoagulation in that itmay be applied even when visualization of the periph-eral retina is not adequate for indirect laser photoco-agulation.13 We expect that in situations where exten-sive far-peripheral indirect laser treatment is possible,the rate of successful prevention of further PDVHwould be similar to that resulting from peripheralretinal cryotherapy.

Our study has no untreated control group withwhich to draw comparisons, and there are no strictlycomparable historical controls in the literature. It is ofinterest, however, to review the results of early vitrec-tomy for severe diabetic vitreous hemorrhage reportedby the Diabetic Retinopathy Vitrectomy Study Re-search Group.14 In the Diabetic Retinopathy Vitrec-tomy Study (DRVS), panretinal endophotocoagula-tion was not performed during vitrectomy. This,therefore, represents a group of patients with diabetic

vitreous hemorrhage in whom no further retinal abla-tive therapy was applied at the time of vitrectomy fordiabetic vitreous hemorrhage. The reported rates ofsevere postoperative neovascular complications (inDRVS patients neovascular glaucoma) ranged from14% to 22% among eyes that had undergone earlyvitrectomy. In our study, severe neovascular compli-cations (anterior hyaloidal neovascularization but noneovascular glaucoma) developed after index vitrec-tomy in 9% of eyes. In the DRVS, 23% of eyes thathad early vitrectomy required one or more subsequentvitreoretinal procedures, often vitreous washouts fornonclearing postoperative vitreous hemorrhage. In ourstudy, six eyes developed postoperative vitreous hem-orrhage after index vitrectomy, although only three(14% of the total of 22 eyes) did not clear spontane-ously.

Eyes studied in the DRVS differed from eyes re-ported here in two important ways: (1) all eyes in thepresent study had panretinal photocoagulation beforeindex vitrectomy, but not all eyes in the DRVS hadphotocoagulation before vitrectomy for vitreous hem-orrhage; and (2) eyes in this study underwent revisionof vitrectomy for postoperative vitreous hemorrhage,whereas eyes in the DRVS underwent primary vitrec-tomy for nonclearing vitreous hemorrhage.

In summary, revision vitrectomy with transscleralperipheral retinal cryopexy (generally augmented withfill-in posterior pole endolaser photocoagulation) re-sulted in visual improvement and anatomic stabiliza-tion in the majority of our phakic patients withPDVH. Peripheral retinal cryotherapy is a widelyapplicable method for supplementation of retinal ab-lative therapy in phakic eyes during revision of vitrec-tomy for PDVH. In phakic eyes, endolaser photoco-agulation of the far peripheral retina is not safe becauseof the possibility of lens damage by the shaft of theendolaser probe. Transscleral peripheral retinal cryo-therapy does not carry this risk. In addition, peripheralretinal cryotherapy can be used in situations (such aslens or corneal opacity) in which indirect laser pho-tocoagulation is not possible.

REFERENCES

1. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Com-plications of vitreous surgery for diabetic retinopathy, II:

PERIPHERAL RETINAL CRYOTHERAPYVOL. 126, NO. 1 89

Page 9: Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes

postoperative complications. Ophthalmology 1983;90:522–530.

2. Novak MA, Rice TA, Michels RG, Auer C. Vitreoushemorrhage after vitrectomy for diabetic retinopathy. Oph-thalmology 1984;91:1485–1489.

3. Tolentino FI, Cajita VN, Gancayco T, Skates S. Vitreoushemorrhage after closed vitrectomy for proliferative dia-betic retinopathy. Ophthalmology 1989;96:1495–1500.

4. Rice TA, Michels RG, Rice EF. Vitrectomy for diabeticrhegmatogenous retinal detachment. Am J Ophthalmol1983;95:34–44.

5. Rice TA, Michels RG, Rice EF. Vitrectomy for diabetictraction retinal detachment involving the macula. Am JOphthalmol 1983;95:22–33.

6. Michels RG, Rice TS, Rice EF. Vitrectomy for diabeticvitreous hemorrhage. Am J Ophthalmol 1983;95:12–21.

7. Blankenship GW. Management of vitreous cavity hemor-rhage following pars plana vitrectomy for diabetic retinop-athy. Ophthalmology 1986;93:39–44.

8. Martin DF, McCuen BW. Efficacy of fluid-air exchange forpostvitrectomy diabetic vitreous hemorrhage. Am J Oph-thalmol 1992;114:457–463.

9. Han DP, Murphy ML, Mieler WF, Abrams GW. Outpa-tient fluid-air exchange for severe postvitrectomy diabeticvitreous hemorrhage: long-term results and complications.Retina 1991;11:309–314.

10. Brown GC, Tasman WS, Benson WE, McNamara A, EagleRC. Reoperation following diabetic vitrectomy. Arch Oph-thalmol 1992;110:506–510.

11. Daily MJ, Gieser RG. Treatment of proliferative diabeticretinopathy with panretinal cryotherapy. Ophthalmic Surg1984;15:741–745.

12. Benedett R, Olk RJ, Arribas NP, et al. Transconjunctivalanterior retinal cryotherapy for proliferative diabetic reti-nopathy. Ophthalmology 1987;94:612–619.

13. Ramsay RC, Cantrill HL, Knobloch WH. Cryoretinopexyfor proliferative diabetic retinopathy. Can J Ophthalmol1982;17:17–20.

14. Diabetic Retinopathy Vitrectomy Study Research Group.Early vitrectomy for severe vitreous hemorrhage in diabeticretinopathy: four-year results of a randomized trial. Diabeticretinopathy study report 5. Arch Ophthalmol 1990;108:958–964.

Authors InteractivetWe encourage questions and comments regarding this article via theInternet on Authors Interactivet at http://www.ajo.com/ Questions, com-ments, and author responses are posted.

AMERICAN JOURNAL OF OPHTHALMOLOGY90 JULY 1998