We are intrigued by the case series by Sharma and associates documenting corneal edema after collagen cross-linking. Our group has reported a similar case. An account by Gokhale and 4 cases reported by Koppen and associates constitute other similar reports. This sight-threatening complication seems to be rare, and a series of cases from a single center raises a few questions.

There is an incongruity in the Table, which gives central corneal thickness values at 12 months after surgery, even for the patient with a follow-up of just 8 months. The authors mention 350 consecutive procedures performed from June 2008 through July 2011. The temporal distribution of the complications interests us. The dates of surgery of the first and the last reported cases would help us to understand if there were a cluster. The number of procedures performed in the interim would be a more appropriate denominator for an approximate incidence of this complication during this period. Also, our case report illustrates the fact that not all patients with such a complication may have followed up with the primary surgeon.

Patients with persistent stromal haze were excluded from the present series. It would be interesting to know how many eyes had persistent stromal haze and how long the haze persisted. If available, specular microscopy details of these eyes and other eyes that underwent the procedure during the same period but did not demonstrate corneal edema would clarify if a larger number of eyes sustained subclinical endothelial damage. Likewise, more detailed histopathologic results of the cases that underwent keratoplasty, with particular emphasis on the endothelium, would be desirable.

A meticulous effort has been made by the authors to explore putative causes of the endothelial decompensation. We concur on importance of performing intraoperative pachymetry, which is standard practice at our institution. In our experience, there can be marked and unexpected corneal thinning during the procedure. In this context, innovations such as use of hypotonic riboflavin, transepithelial cross-linking, and accelerated cross-linking may prove to be useful. To us, the clinical features in this series closely resemble those of ultraviolet light-mediated damage. The factors responsible may be errors in focusing or calibration of the light source or inadequate riboflavin instillation leading to greater energy delivery to the cornea. Interestingly, Koppen and associates used the same light source. The system is designed to provide maximal irradiance at only one focal plane, and focusing errors actually would decrease the amount of energy delivered. Repeated errors in riboflavin instillation also are unlikely. This leaves the possibility of greater irradiance delivered by the light source than displayed, despite the calibration checks. It is important to know whether there have been no further similar cases since the last reported case, and if so, whether any modification has been made to either the light source or riboflavin used.

Finally, we compliment the authors for reporting a tragic, but unforeseen, complication in their patients. Such exemplary honesty in scientific communication can only be applauded, because it furthers the collective understanding of such problems.