Main

Massive intraocular haemorrhage is a devastating and well-known complication of anticoagulation treatment in eyes with exudative age-related macular degeneration.1 Thromboembolism, prosthetic heart valves, previous myocardial infarction, and strokes are some of the most common conditions where anticoagulation treatment is indicated.2 However, intraocular bleed is not common in normal nonpredisposed eyes (without any neovascularization) even with anticoagulation treatment. Background diabetic retionopathy alone is not known to be related to massive retinal bleed. Here we report a case of massive subretinal bleed in a patient with background diabetic retinopathy who was taking warfarin.

Case report

A 76-year-old woman presented to the eye A&E with sudden painless loss of vision in her right eye of 7 days duration. In the past ocular history, she had an exaggerated subconjunctival haemorrhage and haemotoma in her left eye 3 years ago after a minor blunt trauma, which resolved without any sequlae. She had been diagnosed with NIDDM a year before and blood sugar was well controlled on diet alone. She had undergone aortic valve replacement surgery in 1990 and was on a maintainance dose of warfarin, 2 mg/day. She was taking no other medication at the time and was not suffering from any other systemic diseases.

On examination, her VA was perception of light in the right eye and 6/12 with full correction in the left eye, which had early cataract. There was no relative afferent pupillary defect. The intraocular pressures were 16 mmHg in both eyes. The anterior chambers were of average depth and the anterior vitreous was clear in both eyes. A dilated fundus examination of the right eye revealed a slate grey and dark red mass, lying under the major blood vessels of the retina and extending from the macula up to the temporal disc margin. It was slightly elevated from the surface and there was no vitreous haemorrhage (Figure 1a). A clinical diagnosis of right massive subretinal bleed was made.

Figure 1
figure 1

(a, b) Massive subretinal bleed in the right eye, involving the macula and the peripapillary areas and (c) B scan ultrasonography showing subretinal bleed.

Full size image

An FFA was performed, which showed blocked fluorescence in the area with the subretinal bleed with no evidence of leakage (Figure 1b). B-scan ultrasonography did not show any changes attributible to intraocular tumours and also reconfirmed the clinical diagonosis of subretinal bleed (Figure 1c).

A full-blood count (FBC), international normalized ratio (INR) and bleeding, and clotting profiles were urgently carried out. The only positive findings of note were an INR of 3.8 and 4.5 carried out in a span of a week. Her platelet count, WBC count, RBC count, haemoglobin level, and other bleeding and clotting profiles were all within normal limits.

Comment

Warfarin is the most commonly prescribed oral anticoagulant both for prophylaxis as well as treatment of cerebro-vascular and cardio-vascular disorders. Up to 3% of patients on long-term warfarin treatment develop retinal haemorrhage, mostly visually insignificant.3 Patients with hypertension and older than 75 years are at increased risk of retinal bleed because of warfarin.3 Clinical indications for warfarin and the therapeutic range of anticoagulation, monitored by the INR is given in Table 1.2,5

Table 1
Full size table

Warfarin acts by inhibiting the vitamin-K-dependent carboxylation of factors II, VII, IX and X in the liver,2 and induces a state analogous to vitamin K deficiency.4 Most patients require a daily maintainence dose of 2.0–7.5 mg of warfarin to remain anticoagulated. About 10% of patients on an oral anticoagulant for 1 year have a serious complication requiring medical supervision and 0.5–1% have a fatal haemorrhagic event (eg cerebral haemorrhage) despite careful medical management.4

Superstein et al3 reported the mean duration of warfarin therapy to retinal bleeding to be 1.9 years (range 2 months to 5.5 years). Additional treatment with aspirin in patients receiving warfarin double the risk of bleeding,6 our patient was on no other medication. Several studies have examined the minimum effective dose of warfarin in a variety of conditions. Optimal risk–benefit appears to be at an INR of 2–3 for almost all conditions, with the exception of prosthetic valves requiring an INR of 3–4.5.2,5

El Baba et al7 reported that 19% of the patients with ARMD who developed massive intraocular haemorrhages were taking warfarin or aspirin. Other ocular lesions that can also bleed and present similarly, that is, intraocular tumours particularly malignant choroidal melanoma, need consideration and must be ruled out.8,9

Our patient had an INR of 3.8 and 4.5, which although high is an acceptable level of anticoagulation, considering the fact that she had a prosthetic aortic valve. We took advice from haematology and cardiology departments regarding her further management and were advised not to stop or reduce warfarin as these could put her into major risk of thrombo-embolism. She had a few microaneurysms in either eye. We propose that the subretinal bleed was a complication of the long-term warfarin treatment and high level of anticoagulation. We feel it may have been compounded by the coexisting background diabetic changes in the retina through the vascular damages attributible to diabetes.

There are few effective theraupeutic options once the haemorrhage has occurred. The surgical removal (Pars plana vitrectomy with internal drainage) of a massive subretinal haemorrhage has been described by several authors, but the visual outcome is poor because of delays between the onset of the haemorrhage and the actual vitrectomy and the amount of damage to the photoreceptors.10 Among the newer treatment modalities, low-dose intravitreal tissue plasminogen activator injection and pneumatic displacement using an expansile gas bubble is promising. However, this needs validation with prospective randomized controlled trials.11 We managed our patient conservatively.