The use of inhibitors of the synthesis of vitamin K-dependent clotting factors (II, VII, IX, X) (e.g., warfarin; Coumadin) and the antithrombin III activators (e.g., heparin) for the treatment of ocular disease is largely of historical interest. These agents were prescribed for prophylaxis of “impending” retinal vascular occlusions, treatment of retinal artery occlusions, and to halt the progression of macular degeneration¹. The results were disappointing.^2,3,4 Further, retinal vascular occlusions and macular hemorrhages have occurred despite ongoing anticoagulant treatment for nonocular disease.^5,6,7 And even with the most conscientious efforts, uncontrolled extraocular bleeding can occur, as through epistaxis, hematuria, hemarthrosis, hypermenorrhea, and intracranial hemorrhage. Duff and co-workers treated 33 patients with occlusive retinal vascular disease for mean periods of 7 days (heparin) and 3 months (dicumarol).² Twenty-eight percent had unexpected bleeding.

Contradictory evidence has been presented about the risks of bleeding during and after ocular surgery in anticoagulated patients. Patients using warfarin had an increased risk of hyphema after intraocular surgery.⁸ Short-term discontinuation of the drug did not entirely prevent this problem, despite normal laboratory prothrombin times. However, cataract surgery, using retrobulbar or general anesthesia, has been performed on patients with elevated INRs.⁹ No significant ocular complications were reported. In another study, cataract surgery was performed on 734 patients treated with warfarin;¹⁰ 526 continued warfarin use. Of these, 348 had retrobulbar or peribulbar anesthesia and 178 had topical anesthesia. Of the original 734 patients, 208 discontinued warfarin 4 days before surgery; of these, 174 had injection anesthesia and 34 had topical anesthesia. No instances of retrobulbar hemorrhages, hyphemas, or vitreous hemorrhages occurred during the first postoperative week. One myocardial infarction (MI) occurred in the group discontinuing warfarin, and one MI, one stroke, and one deep-vein thrombosis occurred in the group continuing warfarin. The risks of adverse ocular hemorrhages or systemic vaso-occlusive events did not seem altered by changing warfarin usage for the 4 days.

Ligneous conjunctivitis may have multiple etiologies, but it has been associated with severe plasminogen deficiency and impaired fibrinolysis.¹¹ Ligneous conjunctivitis has occurred after treatment with tranexamic acid, which impairs fibrinolysis.¹² The rationale for heparin use to prevent recurrences following membrane removal rested on the observation that ligneous membranes had subepithelial deposits of fibrin. Topical heparin drops, 1,000 or 5,000 IU/mL, were applied immediately after excision of ligneous conjunctivitis membranes in 17 patients.¹³ Drops were applied every 30 to 60 minutes in combination with topical corticosteroids. Twelve of the patients also received α-chymotrypsin drops. Drops were continued until the conjunctival surfaces re-epithelized, and were then tapered until all inflammation was gone. Thirteen of the 17 patients responded (follow-up 6 months or more), although five required one or more additional membrane excisions.

Heparin has been given intravenously or intravitreally during surgery to prevent postvitrectomy fibrin membranes.¹⁴ No significant effect was noted using either an intravenous bolus, 10,000 IU heparin, or an intravitreal infusion solution, 5 IU heparin/mL. An intravitreal solution containing 10 IU heparin/mL produced significantly less fibrin formation but significantly more bleeding. Overall, heparin appeared to be of little value.

Aspirin has been used to prevent amaurosis fugax and retinal arteriolar occlusions due to platelet-fibrin emboli; these emboli often are associated with carotid artery atherosclerotic plaques. Six factors were associated with a higher incidence of stroke in patients with uniocular amaurosis fugax:¹⁵ age 75 years or older, ≥80% stenosis of the ipsilateral carotid artery, male gender, history of cerebral hemispheric transient ischemic attacks or stroke, intermittent claudication, and absence of collateral circulation. The first step in emboli formation is platelet aggregation. This is initiated by the release of the platelet-formed prostaglandin, thromboxane A₂. Platelet aggregation has been reported as abnormally increased in patients with retinal artery occlusions.¹⁶ Aspirin blocks thromboxane A₂ synthesis.¹⁷ This effect is permanent for the life of the platelet. Aspirin also increases blood fibrinolytic activity.¹⁸ Unfortunately, aspirin also inhibits the formation of another prostaglandin, prostacyclin.¹⁹ Prostacyclin (PGI₂) is synthesized in blood vessel walls, and it inhibits platelet aggregation.²⁰ Both thromboxane A₂ and prostacyclin synthesis depend on the activity of cyclo-oxygenase, the enzyme irreversibly inhibited by aspirin acetylation.

The Food and Drug Administration (FDA) approved aspirin, 650 mg twice a day or 325 mg four times a day, as was an effective anti-platelet aggregation agent. However, because doses as low as 20 mg/day effectively inhibit platelet prostaglandin synthesis,²¹ the dose may be unnecessarily high. The most effective aspirin administration may be the enteric-coated preparations, taken as one 325-mg tablet every day.^22,23,24,25 Gastric irritation is avoided and smaller amounts of aspirin are absorbed from the alkaline gastrointestinal tract than from the acidic stomach. These small amounts of aspirin perfuse the portal circulation and permanently inactivate the cyclo-oxygenase of the platelets, but cannot escape beyond the liver into the peripheral circulation. This protects the peripheral vasculature cyclo-oxygenase from aspirin inactivation. The result is that platelet thromboxane synthesis is suppressed whereas peripheral arteriolar prostacyclin synthesis is not.

The Canadian cooperative study considered only stroke or death and not retinal artery occlusion.²⁶ It found that men, but not women, benefited from aspirin therapy. An American study failed to show significant benefit from aspirin, 1,300 mg/day, when the incidences of retinal infarction, continued ischemic attacks, stroke, and death were considered separately.²⁷ However, when the data from all four categories were pooled, aspirin appeared to exert a significant beneficial effect. By the end of 6 months of therapy, the group receiving aspirin had a 19.2% incidence of “unfavorable outcomes” and the placebo group, 44.2%. This need to pool different categories to reach significance has been criticized.²⁸ A subsequent controlled study indicated that patients with amaurosis fugax and an ipsilateral partially occluded carotid artery benefited from aspirin therapy; the frequencies of amaurotic attacks, retinal infarction, cerebral infarction, and death were reduced.²⁹ Although heparin and warfarin may initially be more effective than enteric-coated aspirin in reducing the frequency of amaurotic attacks, they do not appear any more effective in preventing infarctions or death.³⁰ When warfarin (INR 1.4 to 2.8) was compared to aspirin (325 mg daily) for the prevention of recurrent stroke or death, neither drug showed significant superiority;³¹ the size of the study—more than 2,200 patients—allowed a 95% probability that warfarin was no more than 8% more effective than aspirin. Higher doses of warfarin, to achieve INRs of 3.0 to 4.5, may be more effective, but the incidence of major bleeds more than offsets this advantage.³²

Prostacyclin not only prevents platelet aggregation, but in vitro it also disperses platelets that have already aggregated. For this reason, Zygulska-Mach and co-workers treated three patients with partial central vein occlusions and visions of 20/200 (6/60) or less with prostacyclin infusions, 5 ng/kg/min for 72 hours.³³ In one patient, significant visual improvement occurred, to 20/40 (6/12). Little change occurred in the second patient, and the third patient progressed to blindness.

Aspirin and other cyclo-oxygenase inhibitors would appear contraindicated in the treatment of pain in patients with traumatic hyphema or in those undergoing ocular surgery. Re-bleeding after traumatic hyphema has been reported to occur more frequently if aspirin had been ingested before or after the initial injury.³⁴ Ganley and co-workers found that seven of 12 patients who took aspirin after the initial trauma had a recurrent hyphema, whereas only one of 13 patients who did not take aspirin rebled.³⁵ However, when Marcus and co-workers performed a randomized, placebo-controlled prospective study of 51 patients, 23 of whom received aspirin, 1,500 mg a day for 5 days, beginning when admitted for traumatic hyphema, no significant increase in rebleeding occurred.³⁶ With regard to cataract surgery, a large multicentered prospective study was performed.³⁷ Of 3,181 subjects using aspirin, 774 stopped the drug 2 weeks before surgery; of 522 subjects using warfarin, 174 discontinued use 4 days before surgery. In those aspirin users and those warfarin users who discontinued their drug, no instances of intraocular or intraorbital hemorrhages occurred, either intra-operatively or up to 7 days postoperatively. In the aspirin users who continued the drug (2,407 patients), one vitreous and one retrobulbar hemorrhage occurred. In the warfarin users who continued the drug (348 patients), no intraocular or intraorbital hemorrhages occurred. It appeared that continued use of these drugs did not put the eye at risk. With regard to ischemic cardiac or central nervous system events, two occurred in those patients who discontinued aspirin versus 22 in the patients who continued aspirin and one in the patients who discontinued warfarin, versus five in the patients who continued warfarin. It appeared that discontinuing these drugs for the length of time used in the study did not put patients at unacceptable risk.

Clopidogrel bisulfate, a thienopyridine derivative, has largely replaced the chemically related drug, ticlopidine because of fewer side-effects, especially bone morrow suppression. These anti-platelet drugs have an action dissimilar to aspirins: They irreversibly bind to the ADP receptor sites on platelets, preventing secondary activation of the major platelet surface glycoprotein receptor for fibrinogen and leukocytes.³⁸ Unfortunately, their action is limited by some, but not all, statins.³⁹ By 7 days after discontinuation of clopidogrel, platelet function returns to normal.⁴⁰

In 220 patients, ticlopidine, 500 mg daily for 3 years was compared to placebo in 215 patients for effectiveness in reducing the progression of nonproliferative diabetic retinopathy (41). The development of microaneurysms (primary endpoint), as well as new vessels and overall retinopathy progression (secondary endpoints) on red-free photographs and fluorescein angiography (FA) was significantly reduced in insulin-treated patients receiving ticlopidine but not in non–insulin treated subjects. Seven ticlopidine-treated patients had to be permanently discontinued from therapy because of hematologic effects. In total, 29 patients (13%) on ticlopidine versus seven placebo patients (3%) had side effects requiring permanent discontinuation from the study.

Acute retinal necrosis, a disease associated with infection by the herpes family of viruses, especially herpes zoster, has been reported to be associated with platelet hyperaggregation.⁴² Aspirin was found of some limited value in one report but not in another.⁴³

The plasminogen activators include streptokinase, streptodornase, urokinase, and tissue plasminogen activator (tPA). These agents convert the proenzyme plasminogen (profibrinolysin) to the active enzyme, plasmin (fibrinolysin). Plasmin breaks up thrombi by cleaving fibrinogen. Streptokinase and streptodornase are derived from β-hemolytic streptococci. Urokinase is obtained from urine and is less antigenic; about 400 units/day are excreted by the adult male. Streptokinase, streptodornase, and tissue plasminogen activator are enzymatically active. However, urokinase is an inactive pro-enzyme, converted from a single chain to a two-chain active form.⁴⁴ The administration of plasminogen activators may cause bleeding; aminocaproic acid is a specific antidote.

Endogenous plasminogen activators have been identified in human lacrimal fluid⁴⁵ and aqueous humor,⁴⁶ and in primate trabecular endothelium, corneal endothelium, and iris.⁴⁷ Recombinant gene techniques have allowed the commercial production of human tPA. This human tPA has several theoretical advantages over streptokinase and urokinase. Human tPA is markedly active in the presence of fibrin (i.e., in clots) and is relatively inactive in the presence of circulating fibrinogen. Major bleeding complications should therefore be less with tPA. tPA has a shorter half-life, less than 10 minutes, which would allow a more rapid reversal of any bleeding complications. However, these potential advantages of tPA were not convincingly demonstrated in the treatment of acute MI; tPA did not reduce the incidence of major bleeding complications, possibly because of the concomitant use of heparin. Heparin was administered with all the fibrinogen-lysing enzymes and could have masked any advantage of tPA. The intracerebral hemorrhages reported with tPA usewere in patients receiving heparin.^48,49 tPA use alone has been associated with both choroidal and retrobulbar hemorrhages.^50,51 The short life of tPA could be considered a therapeutic disadvantage, leading to reocclusion, and tPA is far more expensive than the other thrombolytic agents. The cost of tPA can be reduced by saving unused portions and storing them at -70° C. These frozen aliquots will remain active for at least 1 year.⁵² The intravenous administration of streptokinase, a much less expensive agent, was shown as effective and safe as tPA in approximately 900 diabetic patients treated for acute MI. Neither streptokinase nor tPA caused intraocular bleeds.⁵³ Diabetic retinopathy should not be considered a contraindication to thrombolysis. Serum sickness can occur after a streptococcosis infection. A possibly under-reported ocular complication of systemic streptokinase administration is serum sickness manifesting itself as a bilateral uveitis.⁵⁴

Human aqueous humor contains more tPA than urokinase.⁵⁵ Although these enzymes are in lower concentrations than in plasma, disproportionately even lower levels of tPA-inhibitors are present.⁵⁶ The result is that normal aqueous humor is shifted toward fibrinolysis.

tPA has been used to clear subconjunctival and anterior chamber clots after glaucoma filtering surgery.⁵⁷ Intracameral injections of 6 to 12.5 μg tPA have been effective in re-establishing bleb function in retrospective reports.^58,59 When 25 μg tPA was used, there was a 36% incidence of hyphema and 21% incidence of hypotony. Corneal edema, as well as hyphema, has been reported at this dosage.⁶⁰ This corneal edema did not seem due to drug toxicity, because in vitro perfusing of the posterior surface of human corneas did not produce permanent endothelial damage.⁶¹ Intracameral urokinase, 500 IU in 0.1 mL diluent, also has dissolved anterior chamber clots following glaucoma surgery.⁶²

Traumatic hyphemas due to nonsurgical injuries have responded to thrombolytic agents. Scheie and co-workers prepared plasmin by exposing human serum plasminogen to streptokinase.⁶³ Anterior chamber clots were lysed. These authors stressed, as have others, that it was not necessary for the entire clot to be removed. Urokinase has been injected directly into the anterior chamber after traumatic hyphemas.⁶⁴ Although the clots were dissolved, the visual results were often poor because of the nature of the injuries. Rakusin advocated initial use of saline irrigations for anterior chamber hyphemas; these are successful more than 60% of the time.⁶⁵ However, in 20 patients in whom saline irrigation failed, clot lysis was achieved using urokinase solutions of 2,500 to 5,000 units/mL water. The anterior chambers were slowly irrigated with 0.3 mL enzyme solution. After 3 minutes, the urokinase was washed out with saline. This sequence was repeated five times. If any clot remained after the last infusion, the enzyme was not irrigated out. No complications from therapy were noted, and all clots were removed within 48 hours.