Corneal confocal microscopy after CXL for keratoconus

Take-home:

Corneal confocal microscopy may be a useful tool for physicians performing corneal collagen crosslinking.

Dr. Touboul

By Nancy Groves; Reviewed by David Touboul, MD

Bordeaux, France-Conventional corneal collagen crosslinking (C-CXL) has become the gold standard to halt keratoconus. Several new CXL procedures, however, have been developed.

Corneal confocal microscopy (CCM) appears to be a useful tool for noninvasive CXL titration and follow-up in patients with keratoconus treated with these protocols, said David Touboul, MD, French National Reference Center for Keratoconus, Bordeaux, France.

Based on a study Dr. Touboul and colleagues performed at the center, he concluded that keratocyte loss is probably a relevant parameter to make comparisons between different CXL protocols.

He also suggested that incorporation of data from biomechanics and topography could lead to better optimization of the compromise between safety and efficacy in CXL.

The new trends in CXL include improving safety with transepithelial protocols (T-CXL) and decreasing operating time with accelerated protocols (A-CXL), Dr. Touboul said.

T-CXL is similar to C-CXL, but in the newer protocol the epithelium is not removed and 0.1 tromethamine is used to enhance riboflavin uptake in the cornea, while dextran is not used.

The main differences between the accelerated protocol and the others are the times and fluency. In A-CXL, riboflavin is administered for 10 minutes instead of 30, and UVA time is only 3 minutes versus 30, reducing the overall time of the procedure to less than one-quarter of that required for the conventional protocol, 13 minutes versus 60. The fluency in A-CXL is 30 mW/cm² compared to 3 mW/cm² in C-CXL and T-CXL.

The study used CCM to compare conventional CXL with the two newer protocols.

Dr. Touboul and colleagues evaluated 24 eyes of 24 patients with progressive keratoconus and corneal thickness > 400 µm. They were divided into three groups of eight each: Group 1, C-CXL; Group 2, A-CXL; group 3, T-CXL.

In vivo CCM was performed on each patient preoperatively and at 1, 3, and 6 months postoperatively.

“The main confocal findings after C-CXL were nerve plexus loss, keratocyte loss with a decrease of keratocyte density and decrease of nuclear reflectivity, and also a stromal honeycomb-like pattern,” Dr. Touboul said. “All of these signs decreased with stromal depth and with time.”

At 1 month, the control eyes (no CXL) and the T-CXL eyes were similar, while there was a huge loss of keratocytes in the stroma in the C-CXL and A-CXL eyes. In eyes treated with A-CXL, the loss of keratocytes in the anterior stroma was even more pronounced than in the eyes treated with conventional CXL.

At month 3, there were signs of regression of the keratocyte loss, and at month 6 the trend was confirmed with minimal loss of keratocytes in the entire population.

There was no significant endothelial cell loss in any of the CXL protocols at any point during the follow-up (p > 0.05) The preoperative and 6-month counts were, respectively: C-CXL, 2995 ± 367 and 3013 ± 366; T-CXL, 3445 ± 250 and 3594 ± 260; A-CXL, 3591 ± 483 and 3577 ± 516.

Finally, the investigators found that the epi-off CXL protocols exhibited long-term anterior nerve plexus loss.

“For the same amount of photons, based on confocal findings there were very different reactions to the crosslinking in the stroma,” Dr. Touboul said.

He hypothesized that T-CXL did not exhibit changes with CCM. One theory to explain this is that the epithelium was soaked with riboflavin, preventing stomal crosslinking by acting as a UVA light shield, but this is probably wrong because riboflavin cannot enter the epithelial cells and natural UVA epithelial absorption is very low.

The more plausible option to explain the lack of visible changes on CCM is that there was insufficient riboflavin in the stroma at the end of the soaking time. With no riboflavin, there is no effect, Dr. Touboul added.

Another question is why A-CXL was located more anteriorly. Among several possibilities, the most likely is that the shorter soaking and UVA time caused the difference. The less likely choices are that higher fluency was used (30 versus 3 mW/cm²), which is unlikely because at the end, the same dose of photons was used in this study as in others, or that the riboflavin was somehow different.

The study also raised questions about the relationship between keratocyte loss, corneal stiffening, and CXL efficacy, Dr. Touboul said. Two key issues are whether collagen bonding remains possible and effective without killing keratocytes and whether the keratocyte, epithelium, and nerve plexus renewal play a role in keratoconus stabilization.

“Today, nobody knows,” he concluded.

The study was published in the Journal of Refractive Surgery (2012;28:769-776).

David Touboul, MD

E: david.touboul@chu-bordeaux.fr

Dr. Touboul did not report any disclosures.

Subscribe to Ophthalmology Times to receive the latest clinical news and updates for ophthalmologists.