Sir,

We agree with Canning and Lotery1 that a clinical trial of bevacizumab in the treatment of neovascular age-related macular degeneration (AMD) could reveal important safety issues, but for rarer diseases randomised controlled trials are not always appropriate.

Neovascular glaucoma is another disease where bevacizumab is promising. It is, however, uncommon compared with AMD, and fortunately most cases respond to panretinal photocoagulation (PRP) when timely performed.2 For the minority in whom PRP is either not possible or not effective, the prognosis is dire.

Bevacizumab has several possible roles in the management of this vexing condition. Used early it may prevent intractable glaucoma; later when the angle is closed, it may be used as an adjunct to glaucoma drainage surgery reducing the risk of intraocular bleeding. Finally by causing regression of new vessels and reducing leakage of inflammatory mediators, it may help in the treatment of painful blind eyes.3

Faced with the alternatives of a painful blind eye or enucleation, patients are unlikely to be willing to accept the control arm of a trial, nor would we be willing to offer it.

On this basis we have been given approval by our Regional Ethics Committee to use bevacizumab for the treatment of neovascular glaucoma unresponsive to PRP. Where randomised placebo-controlled clinical trials are not appropriate, one should not underestimate the power of a well-conducted, prospective, observational case series.