We read with interest the paper entitled “Corneal collagen cross-linking with riboflavin and ultraviolet-A irradiation in patients with thin corneas,” and would like to offer comments and raise our concerns when considering collagen cross-linking (CXL) procedure for patients with a thin cornea.

The authors documented in this paper that corneal endothelial damage occurred in a series of patients with corneal thickness below the cutoff limit of 400 μm after a standard CXL procedure. This study, therefore, further adds a line of evidence to the conclusion that a corneal thickness of at least 400 μm (without epithelium) should be a minimum safety limit when performing the standard CXL and that an alternative treatment protocol where hypo-osmolar riboflavin solution is used to induce iatrogenic corneal swelling before CXL application should be considered. However, the criteria for procedure selection may be complicated by the fact that corneal thickness significantly decreases by approximately 19% during the CXL treatment because of water evaporation from the corneal stroma even with the standard application of isotonic or hypotonic riboflavin (0.1%) drops to saturate the cornea. So, the authors might have performed ultraviolet irradiation on the thinner corneas and corneal thickness should be intraoperatively monitored during CXL treatment. It is suggested that cornea should be swelled in the standard procedure or reswelled in the modified procedure by administering a hypotonic solution if corneal thickness falls below 350 μm at any time point during the treatment. However, the therapeutic outcome of the modified CXL procedure may not be optimal, especially in the cases whose corneal thickness is below 330 μm.

Since CXL now is mainly applied to keratectasia patients who commonly have thin corneas and the modified CXL procedure often is recommended when corneal thickness is below 400 μm, many of them will potentially have to be confronted with penetrating keratoplasty if CXL cannot effect the arrest of the disease progression. As it was demonstrated in this study, CXL in corneas with central pachymetry less than 400 μm seems to only result in a significant endothelial cell count loss postoperatively. This finding, however, was not related to other intraoperative or postoperative complications, nor were there any signs of intraocular toxicity in any of the eyes. The standard CXL procedure, therefore, may be a selection for those who probably will not acquire an anticipated optimal therapeutic outcome with the modified CXL procedure. Nevertheless, further studies with a larger sample and longer follow-up are needed to assess the safety before the procedure is applied on a larger scale.