Post-vitrectomy Diabetic Vitreous Hemorrhage: Etiology, Clinical Management, and Prevention

YOEL GREENWALD, MD · AYALA POLLACK, MD · GUY KLEINMANN, MD

Pars plana vitrectomy (PPV) is a well-established treatment for certain complications of proliferative diabetic retinopathy (PDR), including vitreous hemorrhage (VH) and traction retinal detachment. Recent advances in surgical techniques have improved the safety of the procedure, anatomical success, and final visual outcome.^1-4 Despite these improvements, persistent or recurrent VH remains a common postoperative complication and patient management dilemma.

Studies performed in the 1980s suggested the rate of VH after vitrectomy in PDR to be as high as 75%; however, more recent studies indicate the risk to be approximately 30%.⁴ In addition to having a significant impact on quality of life through decreased visual acuity, postoperative vitreous hemorrhage (POVH) is also a very strong risk factor for a poor final visual outcome.⁵ In this article, we will summarize our approach to managing this common complication of diabetic vitrectomy and present evidence-based tips that may help reduce the incidence of POVH.

ETIOLOGY AND MANAGEMENT

It is necessary to consider the potential etiology of the hemorrhage when making clinical management decisions. When the cause is suspected to be self-limited, watchful waiting might be appropriate. However, when a new retinal neovascularization is suspected, more aggressive measures should be considered.

The time of onset of the POVH in relation to surgery can provide a clue as to its etiology. For the purposes of this article, the term early POVH will be used when the hemorrhage is diagnosed in the first 4 weeks postoperatively and late onset POVH when the hemorrhage develops after the first 4 weeks postoperatively. Causes of early POVH include residual blood clots in the peripheral vitreous, iatrogenic injury, and incomplete removal of fibrovascular tissue.^6,7 Even with careful and complete intraoperative removal of blood clots and fibrovascular tissue, rebleeding can still occur from blood-vessel breakdown secondary to damage incurred during dissection of fibrovascular membranes.^8,9

Early studies found vitreous hemorrhage to be present in the early postop period in up to 63% of eyes, with clearance of the hemorrhage occurring on average at 9.1 weeks postoperatively in phakic eyes.⁷ Modern surgical equipment and techniques with complete retinal endophotocoagulation have achieved significantly better results. Emphasis on complete removal of fibrovascular tissue, residual blood, and blood clots to the greatest possible extent, along with judicious use of mechanical pressure and endodiathermy to achieve hemostasis, can reduce the incidence of early POVH to an average duration of less than 2 weeks.^8,9

For early POVH, often the best treatment is to reassure the patient that the hemorrhage will likely clear spontaneously and that hemorrhages seldom last longer than 1 month.⁸ However, if early POVH persists to 6 weeks postoperatively, or if a vitreous hemorrhage occurs more than 4 weeks postoperatively following an initial period of clear vitreous, the possibility of fibrovascular ingrowth (FVI) extending from 1 or more of the sclerotomies should be considered.

FIBROVASCULAR INGROWTH

The importance of the diagnosis of FVI in eyes with nonclearing and later onset POVH was illustrated by Bhende et al.¹⁰ This group used the ultrasound biomicroscope (UBM) to prospectively assess changes at the sclerotomy sites in 86 consecutive eyes after primary diabetic vitrectomy. The UBM examination was carried out in all patients between 6 and 8 weeks postoperatively.

At the time of UBM examination, 13 patients were suffering from concurrent POVH. In 5 of these patients, no evidence of FVI was found on UBM. These subjects experienced spontaneous clearing in 3 instances, while 2 required only outpatient fluid-gas exchange to maintain clear media to the end of follow-up. In the remaining 8 POVH patients, FVI was seen on UBM. In this group only 1 patient was successfully treated with outpatient fluid-gas exchange, while the other 7 patients required reoperation, often with membrane peeling and retinal ablation with cryotherapy. Despite this aggressive treatment, 2 eyes with postoperative FVI were deemed inoperable due to retinal neovascular proliferation.

Similar studies have confirmed a strong association between the presence of POVH more than 2 months postoperatively and evidence of FVI either on UBM^11-12 or 20-MHz high-resolution anterior-segment ultrasonography.¹³ The location of the fibrovascular tissue at the sclerotomy site makes diagnosis by ophthalmoscopy technically difficult. Attempts to predict the existence of FVI by the presence of large episcleral vessels entering the original sclerotomy sites have proved unreliable.¹⁴

Therefore, high-resolution ultrasound or UBM of the sclerotomy sites is suggested in all cases of significant nonclearing POVH present more than 4 to 6 weeks after diabetic vitrectomy. In the absence of FVI, outpatient fluid-air exchange or simple observation may suffice. However, the presence of FVI predicts a poorer visual prognosis and indicates the need for more aggressive surgical intervention,^10,12-13 very often including reoperation and extensive cryotherapy,¹⁵ to achieve clear media.

BEVACIZUMAB

There are encouraging reports which suggest that injections of the antiangiogenic agent bevacizumab (Avastin, Genentech) may be considered in treating later onset POVH.¹⁶ Since this condition is often caused by the presence of FVI, and bevacizumab inhibits the action of vascular endothelial growth factor, there may be a rationale to support the use of this antiangiogenic agent. Work by Avery, among others, supports the use of bevacizumab injections to treat neovascular complications of PDR.¹⁷ However, the case studies supporting the use of antiangiogenic agents alone to treat late-onset POVH are small with short follow-up times. Longer-term studies with larger patient groups are necessary before antiangiogenic agents alone can be recommended in the place of more conventional surgery in patients with POVH and ultrasound evidence of FVI.

PREVENTION

Significant morbidity is associated with POVH. In early-onset POVH this is related to patient anxiety from decreased visual acuity, along with the inability to perform a proper clinical examination. The ramifications of later-onset POVH associated with FVI include a poor visual prognosis and the necessity for further surgical intervention.^7,10,11,15 For these reasons among others, recent attention has focused on specific means of reducing the incidence both of POVH and postoperative FVI.

PREVENTING EARLY ONSET POVH

■ Postoperative Gas Injection. One of the causes of early POVH is bleeding from blood vessels, the integrity of which were compromised intraoperatively during excision of ocular fibrovascular membranes.⁷ Postoperative mechanical tamponade of these damaged vessels by injections of air or nonexpansile SF₆-air or C₃F₈-air mixtures has been attempted to prevent early POVH, with various levels of success. The injection of air or 20% SF₆ has been found ineffectual in decreasing the incidence or duration of POVH.^19,20 However, some evidence suggests that intraocular gas tamponade with 10% C₃F₈ may be useful in reducing the impact of early POVH.

Yang et al.⁸ conducted a prospective study on the effect of a postoperative injection of 10% C₃F₈. They reported that 10% C₃F₈ reduced the incidence of visually significant early-onset (<4 weeks) POVH was from 16% to 0. No increased incidence of elevated intraocular pressure was found in this study. The authors theorized that 10% C₃F₈, combined with proper head position, maintained some tamponade effect up to 3 weeks postoperatively — significantly longer than SF₆ or air injections, as reported in previous studies.^19,20 However, despite the potential benefit of gas tamponade, the interference in patient vision from intravitreal gas and the long-term risk of cataract formation should be considered before recommending routine use of postoperative 10% C₃F₈.

■ Bevacizumab Pretreatment. Bevacizumab treatment in patients with proliferative diabetic retinopathy has gained acceptance in decreasing the impact of ocular neovascularization as an adjunct to retinal photocoagulation.²¹ Recent studies have shown that a single injection of 1.25 mg bevacizumab 1 week before diabetic vitrectomy significantly reduces the severity of intraoperative bleeding, with associated improvements in intraoperative visibility.^9,18

One prospective study showed that bevacizumab pretreatment reduced the average time until clear vitreous was seen postoperatively from 2 weeks to 1 week postoperatively.⁹ The authors attributed this finding to the reduced intraoperative bleeding and subsequent blood clot formation at the end of surgery. There is some concern that bevacizumab may increase vitreoretinal traction by inducing the contraction of fibrovascular tissues.²¹ However, possibly because the time between injection and surgery was short, no increased retinal elevation or enlargement of traction detachment was noticed.⁹

■ Intraoperative Triamcinolone Acetate. Triamcinolone acetate has demonstrated effectiveness in preventing early-onset POVH. In a prospective, randomized clinical trial performed on patients with a nonclearing VH secondary to PDR, a single 4-mg injection of triamcinolone acetate was performed on half of the patients at the end of vitrectomy surgery.²⁴ The injection reduced the risk of POVH during the first month postoperatively from 45% to 13% (P = .003). The risk of reoperation was also significantly lower in the study group, but mean follow-up time was only 6 months. Complications included an increase in posterior subcapsular cataract formation in the treated group (P = .01), as well as elevated IOP at 1 day and 1 week postoperatively (P < .01).

This treatment shows significant promise, especially in pseudophakic patients without a history of glaucoma. However, longer follow-up times are necessary before this treatment can be recommended for routine use. Future studies might compare the effectiveness of bevacizumab pretreatment with triamcinolone injection in reducing the risk of early-onset POVH.

■ Other Preventative Measures. Various treatments have been tested in an effort to reduce the risk of early onset POVH. Systemic administration of the antifibrinolytic drugs epsilon-aminocaproic acid²² and transexamic acid²³ did not reduce the rate of early-onset POVH.

A new as yet unproven treatment with intraoperative plasmin might also provide some benefit. Plasmin introduced at the beginning of diabetic vitrectomy surgery might aid in the peeling of fibrovascular tissue.²⁵ This could reduce the risk of intraoperative vessel injury and reduce subsequent bleeding from damaged vessels. Further study is necessary to determine the risks and benefits of this practice.

PREVENTING FIBROVASCULAR INGROWTH

As previously mentioned, evidence suggests that FVI from the sclerotomy sites or vitreous base is a very important source of late-onset POVH.^10-12,14 Even with aggressive therapy, this complication is often associated with a poor final outcome. Therefore recent efforts have focused on the prevention of postoperative FVI.

A case series examined by Yeh et al. compared the rate of FVI and late-onset POVH in 81 eyes of 71 patients undergoing diabetic vitrectomy with different levels of adjunct cryotherapy.¹¹ UBM findings and the incidence of POVH in 3 groups of patients were compared to determine the effectiveness of cryotherapy in inhibiting FVI and preventing late onset POVH.

Panretinal endophotocoagulation extending to or beyond the equator was performed on all patients. In group 1, no additional retinal ablative treatments were carried out. In group 2, anterior peripheral retinal cryotherapy was added to the panretinal endolaser treatment, and in Group 3, patients underwent cryotherapy on the 3 sclerotomy sites, in addition to the procedures performed on patients in group 2. Each cryotherapy application was performed under direct visualization and was stopped as soon as the retina turned white. The POVH rates in the different groups were as follows: 37.5% in group 1, 11.5% in group 2, and 4.3% in group 3 (P = .0004). FVI was noted in 87.5% of all eyes experiencing POVH. In the lone eye in group 3 with a POVH, FVI was not seen on UBM and the hemorrhage cleared spontaneously.

This study suggests that panretinal endophotocoagulation to the equator without additional ablative therapy may not be enough to inhibit the development of FVI and the associated POVH. Additional ablation to the far retinal periphery is required to further reduce the secretion of angiogenic factors and reduce the risk of FVI formation. This ablation could be carried out using cryotherapy, as described in the preceding study, or possibly by other means, such as endolaser or indirect laser,²⁶ at the surgeon's discretion.

SUMMARY

Recurrent VH is the most common complication of vitrectomy surgery performed to treat complications of PDR.⁶ For early POVH, the hemorrhage often clears spontaneously.⁸ However, if early POVH persists to 6 weeks postoperatively, or if a vitreous hemorrhage occurs more than 4 weeks postoperatively following an initial period of clear vitreous, ultrasonography of the sclerotomy sites is recommended to rule out the possibility of FVI extending from 1 or more of the sclerotomies. In the absence of FVI, watchful waiting can be resumed and outpatient fluid-air exchange can be considered. However, if evidence of FVI is found on ultrasound, the surgical intervention with peripheral anterior retinal ablation should be considered.

Many preventative measures have shown promise in preventing the impact of POVH. Pretreatment with bevacizumab 1 week before diabetic vitrectomy seems to reduce the incidence of intraoperative bleeding and improves postoperative vitreous clearing time. There may be a role as well for the injection of triamcinolone or 10% C₃F₈ at the end of surgery to reduce the incidence of POVH in the first 4 weeks postoperatively.

The prevention of FVI-associated late-onset POVH is very important, as this complication is associated with a poor visual outcome. Anterior peripheral retinal ablation is recommended where possible to reduce the incidence of this serious complication. With a focus on prevention and proper treatment algorithms, the impact of POVH can hopefully be significantly reduced. RP

REFERENCES

1. Ho T, Smiddy WE, Flynn HW Jr. Vitrectomy in the management of diabetic eye disease. Surv Ophthalmol. 1992;37:190-202.
2. Helbig H, Sutter FK. Surgical treatment of diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol. 2004;242:704-709.
3. Helbig H. Surgery for diabetic retinopathy. Ophthalmologica. 2007;221:103-111.
4. Mason JO 3rd, Colagross CT, Vail R. Diabetic vitrectomy: risks, prognosis, future trends. Curr Opin Ophthalmol. 2006;17:281-285.
5. Mason JO 3rd, Colagross CT, Haleman T, et al. Visual outcome and risk factors for light perception and no light perception vision after vitrectomy for diabetic retinopathy. Am J Ophthalmol. 2005;140:231-235.
6. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Complications of vitreous surgery for diabetic retinopathy. II. Postoperative complications. Ophthalmology. 1983 May;90:522-530.
7. Novak MA, Rice TA, Michels RG, Auer C. Vitreous hemorrhage after vitrectomy for diabetic retinopathy. Ophthalmology. 1984;91:1485-1489.
8. Yang CM, Yeh PT, Yang CH. Intravitreal long-acting gas in the prevention of early postoperative vitreous hemorrhage in diabetic vitrectomy. Ophthalmology. 2007;114:710-715.
9. Yang CM, Yeh PT, Yang CH, Chen MS. Bevacizumab pretreatment and long-acting gas infusion on vitreous clear-up after diabetic vitrectomy. Am J Ophthalmol. 2008;146:211-217.
10. Bhende M, Agraharam SG, Gopal L, et al. Ultrasound biomicroscopy of sclerotomy sites after pars plana vitrectomy for diabetic vitreous hemorrhage. Ophthalmology. 2000;107:1729-1736.
11. Yeh PT, Yang CM, Yang CH, Huang JS. Cryotherapy of the anterior retina and sclerotomy sites in diabetic vitrectomy to prevent recurrent vitreous hemorrhage: an ultrasound biomicroscopy study. Ophthalmology. 2005;112:2095-2102.
12. Hershberger VS, Augsburger JJ, Hutchins RK, Raymond LA, Krug S. Fibrovascular ingrowth at sclerotomy sites in vitrectomized diabetic eyes with recurrent vitreous hemorrhage: ultrasound biomicroscopy findings. Ophthalmology. 2004;111:1215-1221.
13. Steel DH, Habib MS, Park S, Hildreth AJ, Owen RI. Entry site neovascularization and vitreous cavity hemorrhage after diabetic vitrectomy. The predictive value of inner sclerostomy site ultrasonography. Ophthalmology. 2008;115:525-532.
14. West JF, Gregor ZJ. Fibrovascular ingrowth and recurrent haemorrhage following diabetic vitrectomy. Br J Ophthalmol. 2000;84:822-825.
15. Neely KA, Scroggs MW, McCuen BW 2nd. Peripheral retinal cryotherapy for postvitrectomy diabetic vitreous hemorrhage in phakic eyes. Am J Ophthalmol. 1998|;126:82-90.
16. Ruiz-Moreno JM, Montero JA, Lugo F, Amat P, Staicu C. Intravitreal bevacizumab in recurrent diabetic vitreous haemorrhage after vitrectomy. Acta Ophthalmol. 2008;86:231-232.
17. Avery RL, Pearlman J, Pieramici DJ, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology. 2006;113:e1-e15.
18. Yeoh J, Williams C, Allen et al. Avastin as an adjunct to vitrectomy in the management of severe proliferative diabetic retinopathy: a prospective case series. Clin Experiment Ophthalmol. 2008;36:449-454.
19. Joondeph BC, Blankenship GW. Hemostatic effects of air versus fluid in diabetic vitrectomy. Ophthalmology. 1989;96:1701-1706;
20. Koutsandrea CN, Apostolopoulos MN, Chatzoulis DZ, Parikakis EA, Theodossiadis GP. Hemostatic effects of SF6 after diabetic vitrectomy for vitreous hemorrhage. Acta Ophthalmol Scand. 2001;79:34-38.
21. Moradian S, Ahmadieh H, Malihi M, Soheilian M, Dehghan MH, Azarmina M. Intravitreal bevacizumab in active progressive proliferative diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol. 2008;246:1699-1705.
22. de Bustros S, Glaser BM, Michels RG, Auer C. Effect of epsilon-aminocaproic acid on postvitrectomy hemorrhage. Arch Ophthalmol. 1985;103:219-221.
23. Ramezani AR, Ahmadieh H, Ghaseminejad AK, Yazdani S, Golestan B. Effect of tranexamic acid on early postvitrectomy diabetic haemorrhage; a randomised clinical trial. Br J Ophthalmol. 2005;89:1041-1044.
24. Faghihi H, Taheri A, Farahvash MS, Esfahani MR, Rajabi MT. Intravitreal triamcinolone acetonide injection at the end of vitrectomy for diabetic vitreous hemorrhage: A Randomized, Clinical Trial. Retina. 2008;28:1241-1246.
25. Hirata A, Takano A, Inomata Y, Yonemura N, Sagara N, Tanihara H. Plasmin-assisted vitrectomy for management of proliferative membrane in proliferative diabetic retinopathy: a pilot study. Retina. 2007;27:1074-1078.
26. Steel D, Habib M, Owen R. Preventing recurrent vitreous hemorrhage. Ophthalmology. 2006;113:2373-2374.