To the Editor:

September's Review of Ophthalmology contained a Research Review profile entitled, "Deposits after PK from Gatifloxacin." I am writing in response to this study profile to offer a different point of view.

[Editor's note: The profiled study was Awwad S, Haddad W, Wang M, et al. Corneal Intrastromal Gatifloxacin Crystal Deposits After Penetrating Keratoplasty. Eye & Contact Lens 2004;30(3):169-172.]

I have serious doubts that the deposits represented crystallization of gatifloxacin, as the description does not resemble ciprofloxacin deposits. As I reported in the American Journal of Ophthalmology, ciprofloxacin deposits were located entirely on the surface.  The pH change occurs in the tear film with pH 7.2. If the solubility pH of the medication is below 7.2 there will be a shift of the equilibrium equation towards the solid phase with crystallization occurring at the surface. Ciprofloxacin has a solubility pH of 4.5, considerably less than gatifloxacin at 6.0, and did not develop "large intrastromal deposits" in any reported cases. Furthermore, these deposits were always detectable in any epithelial defect but would become trapped beneath the epithelium following healing. I am not aware of any reports of intrastromal macroscopic deposits with ciprofloxacin (a much lower pH drug).

I just reported at the American Academy of Ophthalmology on a series of 50 corneal transplant specimens that received preoperative gatifloxacin and were measured to determination the intracorneal levels of gatifloxacin. Results showed mean intracorneal gatifloxacin levels were 15.94 mcg/g. Mean aqueous levels were 1.00 mcg/ml. Despite these very high drug levels in the corneal stroma, not a single specimen showed any indication of deposits on a macroscopic scale. Furthermore, half the specimens were also processed for histopathology. No specimens showed any evidence of corneal deposits on a microscopic exam.

Finally, I have investigated the issue of deposits occurring with the new fourth-generation fluoroquinolones (Zymar and Vigamox) in the tear film. I was not able to detect any visible deposits with either drug even when used in conjunction with AcularLS.

Our group has been using Zymar exclusively for our postoperative corneal graft patients since its introduction; this translates into over 200 grafts since spring 2003. We selected Zymar based on its documented lower toxicity to corneal epithelial healing; a key aspect of post-graft management. We have never observed the phenomenon described in the case profiled in the September issue. The location of the deposits entirely within the donor cornea indicates that the deposits are related to the donor cornea itself. Had the deposits been due to the use of gatifloxacin, it should have also appeared in the host stroma.

The gatifloxacin certainly diffuses throughout the corneal tissue. It does not remain confined to the donor graft only. Furthermore, the persistence of the lesions for six months after stopping the drug also presents a strong argument against a drug-related effect. Even the ciprofloxacin deposits trapped beneath an intact epithelium with a solubility pH of 4.5 disappeared within one to two months following removal of the drug.

I believe, for the reasons stated above, that further research is warranted on the topic of crystalline deposits post-PK before we can assign blame to any one factor.

John R. Wittpenn, MD, FAAO
Stony Brook, NY