Recurrent vitreous hemorrhage after vitrectomy for complications of diabetic retinopathy is a common occurrence. The hemorrhage may appear within the first few weeks after surgery or months later. This complication may delay visual rehabilitation significantly and sometimes requires additional procedures or surgery, jeopardizing previous successful operation. The causes of bleeding are diverse. While evidence suggests fibrovascular proliferation from the sclerotomy sites or in the vitreous base may be an important source of recurrent vitreous hemorrhage, other origins of hemorrhage exist including lysed clot from residual vitreous skirt, injured retinal vessels from surgery, and incompletely removed fibrovascular tissues. The latter three conditions may be the major sources of early postoperative vitreous hemorrhage. We have shown that peripheral retinal cryotherapy along with cryo treatment at the sclerotomy sites may effectively reduce the incidence of fibrovascular proliferation at the inner surface of sclerotomy sites and prevent the late-onset recurrent vitreous hemorrhage. However, many patients still experience disturbing vitreous hemorrhage within the first two to three weeks after post-operative transient clear-up of the vitreous. We hypothesize that gas bubble within the vitreous cavity may mechanically temponade the fragile retinal vessels, and concentrate the coagulation factors in the vitreous cavity, allowing the integrity of vessel walls gradually recovers and thus preventing the occurrence of early postoperative recurrent vitreous hemorrhage.

To test this hypothesis, a clinical study was undertaken to investigate the effect of long-acting gas infused into the vitreous cavity at the end of diabetic vitrectomy in the prevention of recurrent vitreous hemorrhage.

The study consisted of two parts. Part one was prospective clinical observational study; part two was case control study.

From January 2005 to May 2005, consecutive patients undergoing primary pars plana vitrectomy for complications of proliferative diabetic retinopathy were recruited for the prospective study. The selection criteria were: 1) anticoagulant therapy should not have been used prior to surgery or during post-operative follow-up period; 2) ther should be no medical history of blood diseases associated with abnormal blood coagulation. During operation, removal of the fibrovascular tissues as well as old and fresh blood was attempted as completely as could be safely done in all cases; blood clots adherent to the peripheral vitreous skirt was removed as much as possible; peripheral retinal cryotherapy (10 to 12 spots in one row), and sclerotomy sites cryo (2 spots, each 6 seconds, for 3 sclerotomy sites) were performed. For those cases without retinal detachment and breaks, fluid-gas exchange with 7.5% C3F8 was done at the end of the surgery; For those with preexisting breaks or iatrogenic breaks induced retinal detachment, fluid-gas exchange with 15% C3F8 was performed before cryotherapy. After surgery, patients without breaks were kept in a prone position overnight, and maintained head down during waking hours and lied on either side, during sleep for 3 weeks. Other patients assumed head positions depending on the location of breaks. Ophthalmological examinations were performed in the first 4 days after surgery, then weekly for 6 weeks and then monthly for 4 months. The preoperative, intraoperative and postoperative data were collected for each patient. These included age, gender, study eye, types and duration of diabetes mellitus, intraoperative diagnosis, and the use of scleral buckle, silicone oil temponade; data regarding the time, duration, frequency and treatment of recurrent vitreous hemorrhage and the duration of post-operation follow-up were also compiled. Results of ophthalmological examinations, including best corrected visual acuity, intraocular pressures, corneal conditions, anterior chamber reactions, lens status, intravitreal gas amount, and retina conditions were recorded. The amount of vitreous blood was assessed by indirect ophthalmoscopy with patients in both sitting and face-up position. The degree of vitreous opacity was recorded according to criteria set for grading vitreous opacity in uveitis.

Patients who had silicone oil infusion, failed to achieve retinal attachment for at least 3 months, or followed for less than a minimum of 5-month were excluded from the study.

Using the same inclusion and exclusion criteria, chart review from September 2004 to May 2005 for diabetic cases operated by the same surgeon with only cryo treatment but without long-acting gas infusion at the end of vitrectomy was done. Similar data as for the prospective study (except intravitreal gas amount) were collected. These cases served as control to compare the rate of recurrent vitreous hemorrhage and the change of best-corrected visual acuity between cases with intravitreal long-acting gas and those without.

Inclusion Criteria:

• patients undergoing primary pars plana vitrectomy for complications of proliferative diabetic retinopathy were recruited for the prospective study

Exclusion Criteria:

• (1) anticoagulant therapy had been used prior to surgery or during post-operative follow-up period; (2) positive medical history of blood diseases associated with abnormal blood coagulation.