We thank Dr. De Bernardo et al. for their interest and their insightful comments on our recently published paper on accelerated corneal collagen cross-linking (CXL) for progressive keratoconus [1].

We fully agree that, instead of ultrasound pachymetry (USP), Pentacam (Oculus, Wetzlar, Germany) should be used to assess the corneal thickness in patients with keratoconus. Indeed Pentacam was utilised during our study to evaluate the corneal thickness during the initial presentation, before CXL and the entire follow-up period. In addition Belin-Ambrosio Enhanced Ectasia Display—a unique feature available on Pentacam—was also used to aid the diagnosis in borderline cases (https://www.oculus.de/uploads/media/belin.pdf). In our study USP was only employed during the CXL procedure to ensure the epithelium-off corneal thickness was more than 400 microns before the start of ultraviolet A irradiation. It was not practical to send the patients for Pentacam imaging during the procedure as this facility was not available in our operating theatre. In addition performing corneal imaging on patients with epithelium-off in a non-sterile environment could potentially increase the risk of corneal infection. We believe that assessing the corneal thickness with Pentacam during CXL is not a routine practice in most treatment centres. Although USP may overestimate the corneal thickness in keratoconic eyes, a meta-analysis has shown that such difference was small, albeit statistically significant [mean 6.33 μm; 95% confidence interval (CI): 3.49–9.17] [2]. On a reassuring note, we did not observe any endothelial dysfunction or corneal decompensation following accelerated CXL.

We appreciate that the corneal volume can be assessed with Pentacam and may be used in combination with the optical data for evaluating and monitoring the progression of keratoconus [3]. However, this is not a commonly used parameter in many long-term CXL studies [4], and therefore this was not analysed in our study.

It is true that vector analysis is required to fully assess the astigmatic correction of a refractive procedure. However, like many other long-term studies [4, 5], CXL was employed to stabilise progressive keratoconus—which was demonstrated in our study—and not used as a means to correct astigmatism. As such we placed more emphasis on the magnitude than the axis component of astigmatism since an increase in astigmatism can often be observed in progressive keratoconus. We also thank Dr. De Bernardo et al. for highlighting the potential risk of bias from studying two eyes instead of one eye per patient. Although the number of bilateral cases was small in our study (n = 4), we have performed further analysis using the data of one eye per patient (the first eye was analysed in bilateral cases) for confirmatory purpose and we did not find any significant changes to our published results.