Sir,

Bevacizumab (Avastin) is a recombinant humanised monoclonal IgG1 antibody that inhibits human vascular endothelial growth factor (VEGF). It has been administered intravitreally in VEGF-mediated disease processes such as choroidal neovascularisation,1 central retinal vein occlusion,2 proliferative diabetic retinopathy3 and pseudophakic cystoid macula oedema. We report the use of intravitreal bevacizumab as a preoperative adjunct for severe proliferative diabetic retinopathy and highlight a complication that has not been previously documented in the literature.

Case

A 36-year-old type I diabetic was referred for deterioration of left visual acuity to 3/36 within 6 months. Examination revealed a dense fibrovascular membrane on the left posterior pole (Figure 1) and comprehensive bilateral panretinal photocoagulation. Optical coherence tomography documented gross cystoid macular oedema. Two weeks following intravitreal bevacizumab (Avastin) 1.25 mg in 0.05 ml there was a marked reduction in the vascularity of the premacular membrane (Figure 2). Pars plana vitrectomy and membrane peel was performed, revealing a full-thickness macular hole, which was managed conventionally. There was minimal intra-operative bleeding. Three months postoperatively the macula was free of any fibrovascular element (Figure 3) and visual acuity improved to 6/12.

Figure 1
figure 1

Fundus photograph at presentation showing a significantly vascularised premacular membrane.

Figure 2
figure 2

Fundus photograph at 2 weeks post-injection showing a mainly fibrous premacular membrane.

Figure 3
figure 3

Fundus photograph at 3 months post-surgery showing clearance of premacular membrane and a sealed hole.

Discussion

Bevacizumab has proven effective as an adjunctive treatment for reducing the vascularity of membranes before diabetic vitrectomy.3, 4, 5 In our case, the intra-operative finding of a macular hole was unexpected. Contraction of the premacular membrane briskly devascularised by bevacizumab may have resulted in tangential forces on the retina causing macular hole formation.

The marked attenuation of vasculature observed after an intravitreal bevacizumab of 1.25 mg raised concerns of its effects on native retinal vasculature, particularly where there is widespread capillary non-perfusion. In the light of a dose–response relationship reported in proliferative diabetic retinopathy3 and the likely causation of macular hole in our patient, a smaller dose of bevacizumab would have been preferable. Intravitreal bevacizumab is a powerful pre-operative adjunct for extensive diabetic fibrovascular disease but caution should be exercised in titrating the dosage to minimise complications associated with rapid devascularisation.