Sir,

I am grateful to Spitzer et al1 for pointing out this error in the Cochrane review of interventions for recurrent corneal erosions.2 This review was able to cite only one study3 that looked at a ‘prophylactic regime to avert the development of recurrent corneal erosion’ following traumatic corneal abrasion. The abstract includes a correct interpretation of our paper, but the body of the publication misrepresents our findings. The reviewers correctly state that we found ‘the addition of lubricating ointment at night [for 2 months] to the standard therapy… resulted in significantly fewer patients with no or minimal symptoms of recurrent corneal erosion at three months’.2 However, they fail to mention that this was because there were significantly more patients who had mild or moderate symptoms at this time (50% in the additional ointment group, 10% in the standard therapy group),3 and thereby their Summary draws the opposite conclusion to our own. We had concluded that there was a ‘significantly higher prevalence of recurrent symptoms in the ‘additional nightly ointment’ group (P=0.016)’.3

In our paper,3 we stated that we were surprised by the higher prevalence of recurrent symptoms in the ‘additional nightly ointment’ group, as we had expected ointment to reduce symptoms. We speculated that ointment might actually interfere with healing of corneal abrasions. We had intended to carry out a further prospective study, to compare ointment, drops, and bandage lenses in the initial management of traumatic corneal abrasion. This never happened, mainly because I moved to a hospital that does not have an open-access eye casualty. I encourage colleagues who do work in such units to carry out this simple study: the results would be of great help to patients who suffer from this common and disabling condition.

In my experience, it is common for authors to mis-quote other papers, and I always encourage my trainees to read an original source in full. Spitzer has highlighted a significant misquotation, in that a Cochrane review has found only one paper to cite, but erroneously draws the opposite conclusion to that of the original researchers. I agree that, in this case, a published clarification would be desirable.