We thank Vazirani and associates for their interest in our series documenting persistent corneal edema after collagen cross-linking (CXL) for keratoconus. Gokhale was the first to report corneal edema after CXL for keratoconus. Bagga and associates reported a similar case and documented generalized endothelial cell loss on histopathologic analysis. Koppen and associates reported keratitis and corneal scarring after CXL for keratoconus. We thought it was necessary to report this sight-threatening complication to make clinicians aware of it. After this series, we treated 4 additional patients with similar corneal changes referred to us for consultation. A few of our colleagues have shared that they, too, have seen this problem occurring in their patients.

Vazirani and associates wanted more details on temporal summation to ascertain incidence by changing the denominator to the number of patients operated on between the first and the last reported case. The first case included in our series underwent surgery on July 17, 2008, and the last case included in this series underwent surgery on August 21, 2010. We performed 268 CXL procedures during this period. Thus, incidence of persistent corneal edema during this period was 3.72%. We have monitored follow-up records of the patients, and a few of the patients had consultations at other centers, but all of them reported back to us. The follow-up was 18 months, and not 8 months.

Vazirani and associates wanted details regarding patients with corneal haze who were excluded to ascertain subclinical endothelial damage. Six patients had stromal haze, including 1 at the pre-Descemet level persisting beyond 3 weeks. Corneal haze disappeared in these patients within 5.5 ± 2.3 weeks (range, 4 to 8 weeks). The patient with pre-Descemet haze had an endothelial cell count of 1524 cells/mm ² in the operated eye and 2642 cells/mm ² in the nonoperated eye. The patient reported by Bagga and associates had an endothelial cell count after CXL of 1571 cells/mm ² at 3 weeks and demonstrated massive corneal edema at 6 months necessitating penetrating keratoplasty. This indicates that endothelial cell count alone may not detect the extent of damage and that there seems to be generalized functional damage to endothelial cells as well.

Vazirani and associates have raised an important concern of inadvertent excessive irradiance causing endothelial damage. We would like to emphasize that calibration of the equipment is of paramount importance to prevent inadvertent delivery of excessive energy. We were aware of this fact and checked the irradiance using a calibrated ultraviolet meter to confirm 3.0 mW/cm ² emissions before each treatment session. The calibrated ultraviolet meter also was checked by the manufacturer and was found to give correct readings. We were extremely careful in focusing the light source and monitored it through the entire treatment session. Further, there were no errors in the instillations of riboflavin because we instilled drops after exact intervals. After a recent report of corneal thinning during ultraviolet A exposure, intraoperative pachymetry is a routine. We do agree that newer innovations including hypotonus riboflavin, transepithelial CXL, and accelerated CXL may be useful.

We thank once again Vazirani and associates for their interest and compliments.