A randomized, controlled trial of corneal collagen cross-linking in progressive keratoconus: three-year results.

Purpose: To report the refractive, topographic, and clinical outcomes 3 years after corneal collagen cross-linking (CXL) in eyes with progressive keratoconus.

Design: Prospective, randomized controlled trial.

Participants: One hundred eyes with progressive keratoconus were randomized into the CXL treatment or control groups.

Methods: Cross-linking was performed by instilling riboflavin 0.1% solution containing 20% dextran for 15 minutes before and during the 30 minutes of ultraviolet A irradiation (3 mW/cm(2)). Follow-up examinations were arranged at 3, 6, 12, 24, and 36 months.

Main Outcome Measures: The primary outcome measure was the maximum simulated keratometry value (Kmax). Other outcome measures were uncorrected visual acuity (UCVA; measured in logarithm of the minimum angle of resolution [logMAR] units), best spectacle-corrected visual acuity (BSCVA; measured in logMAR units), sphere and cylinder on subjective refraction, spherical equivalent, minimum simulated keratometry value, corneal thickness at the thinnest point, endothelial cell density, and intraocular pressure.

Results: The results from 48 control and 46 treated eyes are reported. In control eyes, Kmax increased by a mean of 1.20±0.28 diopters (D), 1.70±0.36 D, and 1.75±0.38 D at 12, 24, and 36 months, respectively (all P <0.001). In treated eyes, Kmax flattened by -0.72±0.15 D, -0.96±0.16 D, and -1.03±0.19 D at 12, 24, and 36 months, respectively (all P <0.001). The mean change in UCVA in the control group was +0.10±0.04 logMAR (P = 0.034) at 36 months. In the treatment group, both UCVA (-0.15±0.06 logMAR; P = 0.009) and BSCVA (-0.09±0.03 logMAR; P = 0.006) improved at 36 months. There was a significant reduction in corneal thickness measured using computerized videokeratography in both groups at 36 months (control group: -17.01±3.63 μm, P <0.001; treatment group: -19.52±5.06 μm, P <0.001) that was not observed in the treatment group using the manual pachymeter (treatment group: +5.86±4.30 μm, P = 0.181). The manifest cylinder increased by 1.17±0.49 D (P = 0.020) in the control group at 36 months. There were 2 eyes with minor complications that did not affect the final visual acuity.

Conclusions: At 36 months, there was a sustained improvement in Kmax, UCVA, and BSCVA after CXL, whereas eyes in the control group demonstrated further progression.