Sir,

I would like to thank Drs Anwar and Teichman for their useful comments.

A ‘big bubble’ can be formed in nontrephined corneas and even with their own results there is a chance, although only 10–20% after the second air injection that a ‘big bubble’ does not form.1 In our four cadaver eyes we did not see a ‘big air bubble’2 and perhaps with a larger series we would have achieved better results. However, the question remains what to do if a ‘big air bubble’ does not form. Endoscopic visualisation of the posterior surface of the cornea is a possible aid to confirm or refute the presence of a ‘big air bubble’, the ideal end point. This information may aid the surgeon on how to proceed. Reinjection into opaque cornea in a different site is difficult and may cause perforation and unnecessary if a ‘big air bubble’ had formed but not been recognised. Dissection without a ‘big air bubble’ is time consuming with a higher chance of ‘irregular dissection’ and less than optimum visual results.

It is the thin ectatic corneas that present a surgical challenge to any lamellar technique, and prior trephination in such eyes is hazardous. Drs Anwar's and Teichman's method states the importance of prior trephination to isolate the central cornea and may aid deeper spread of air towards Descemet's membrane, thus helping formation of the ‘big air bubble’.

Excessive air injected into pretrephined eyes escapes from the trephined interface. Air entry into a closed eye would impede air dissection more posteriorly into the cornea as intraocular pressure is raised. However, one could argue that air entry through one of our paracentesis, which we were careful to avoid, would create a softer eye than fluid inside the anterior chamber and possibly aid a ‘big air bubble formation’.

Again I would like to emphasise that our experiment was in cadaver eyes and this could explain the differnce between Drs Anwar's and Teichman's results and ours.

Direct endoscopic visualisation remains an alternative to aid visualisation and surgery affecting the posterior corneal surface particularly in situations where the view is compromised. It may also help future developing techniques such as Descemet's transplantation3, 4, 5 as such tissue is difficult to visualize by its transparent nature and delicate to handle. Reorientation of Descemet's membrane6 may also be aided by direct visualisation with an endoscope.