Paediatric Cross-Linking in Keratoconus: A Clinical Audit

  • Post author:Azeem Siddique, Gareth Spence, Michael O’Gallagher, Dimple Patel, Jonathan Jackson
  • DOIDOI:10.48089/jfo7688126

Dr Azeem Siddique, Dr Gareth Spence, Mr Michael O’Gallagher, Miss Dimple Patel, Prof Jonathan Jackson

Belfast Health and Social Care Trust


Keratoconus is an ectasia in which progressive thinning and cone-shaped distortion of the cornea result in worsening vision through increasing myopia and astigmatism. Age at diagnosis is usually between 15 and 30 years however it has been suggested that keratoconus is often more advanced if first diagnosed in childhood and progresses more quickly (1). The estimated prevalence of keratoconus is 1.38 per 1000 population (2). Risk factors include family history of keratoconus,  eye rubbing and atopic conditions (2). Patients may experience a reduction in visual acuity and distortion of the visual field. Keratoconus is diagnosed through keratometry, the measurement of the curvature, thickness and steepness of the cornea with computerised corneal topography (3).

Management of keratoconus has two primary goals: firstly to provide functional visual acuity and secondly to halt the progression of corneal distortion. Refractive correction through spectacles or contact lenses is the mainstay of management of stable keratoconus. If refractive correction fails to provide functional visual acuity, surgical management can be considered. In severe keratoconus, patients may require corneal transplantation. Corneal collagen cross-linking (CXL) is a minimally invasive procedure that aims to slow the progression of keratoconus by strengthening corneal stromal tissue (3). Riboflavin and ultraviolet light is used to induce the formation of additional cross- link bonds with the extracellular matrix of stromal collagen (4). CXL is indicated at the onset of documented progression of keratoconus (3).


The 2021 KERALINK randomised control trial was carried out to examine the efficacy and safety of CXL for stabilisation of progressive keratoconus in paediatric patients (3). Sixty participants were randomised into a standard care control group which involved glasses or contact lens correction, or a cross-linking in addition to standard care treatment group (3).

The primary outcome measured was K2, the mean corneal power at the steepest meridian, in diopters (D) (3). Secondary outcomes included keratoconus progression, defined as a 1.5D increase in K2, visual acuity, apical corneal thickness and quality of life (3). The trial found statistically significant differences between the two treatment arms in the primary outcome, favouring corneal cross-linking over standard care alone, and concluded that CXL arrests progression of keratoconus in the majority of young patients (3). Larkin et al. advised that CXL should be considered as a first-line treatment for paediatric keratoconus with the potential for avoiding later requirements for corneal transplantation (3).

The National Institute for Health and Care Excellence (NICE) in the UK have developed guidance for carrying out CXL and recommend audit and review of clinical outcomes for all patients undergoing the procedure (4). In the guidance, NICE also advise on the outcomes to be  considered (4). The Royal College of Ophthalmologists have also published a data set that advises the clinical information to be collected to facilitate clinical audit of the CXL procedure (5).

The aims of this retrospective clinical audit were to assess outcomes in paediatric patients who had undergone CXL for keratoconus at our tertiary paediatric ophthalmology department, compare outcomes with current literature and make recommendations to improve assessment and practice.


Patients who had undergone CXL within the department between January 2017 and November 2022 were identified. This allowed for the formation of a sample of eyes for that were at least 6 months post-CXL procedure.

Baseline patient demographic data was collected including age, sex, laterality of eye and best corrected visual acuity. NICE guidance, the Royal College of Ophthalmologist data set and the KERALINK randomised control trial (RCT) were used to determine the following outcomes to be considered: K2 (steep keratometry), apical corneal thickness, Kmax (maximal keratometry), progression of keratoconus (defined as an increase in K2 of 1.5 dioptres), best corrected visual acuity (BCVA), and adverse events (3,4,5).

Retrospective data collection was carried out for each of the outcomes. Documentation and patient correspondence was searched to determine baseline and post-CXL best corrected visual acuity at 6 months or later as well as any adverse outcomes. Pentacam databases interrogated to  determine baseline and post-CXL corneal topography data. The mean differences and percentage change in outcomes before and after CXL were determined. Analysis data from out sample was compared with that of the KERALINK RCT.


26 patients aged 10-16 years old had CXL from 2017 until 2022 at our tertiary ophthalmology centre. Unfortunately pentacam corneal topography data was only available for 23 eyes. Of these, 6 months had passed since the CXL procedure in 22 eyes. A sample of these 22 eyes were used for analysis.

Table 1: Mean values for outcomes at baseline and following CXL

Outcome Mean value for patients undergoing CXL
Baseline Post-CXL (6-months)
K2 (D) 51.2 50.4
Kmax (D) 60.6 59.3
ACT (µm) 461.8 450.4
BCVA (logMAR) 0.305 0.193

The mean age of patients at time of procedure was 13.7 years with a range of 11 to 16. 13 right eyes and 9 left eyes were included in analysis. 90.9% (20/22) of the sample were male eyes. Mean values for each outcome including K2, Kmax, apical corneal thickness and BCVA at baseline and  at six or more months following the CXL procedure are shown in table 1.

4.5% (1/22) of sample eyes demonstrated progression of keratoconus, defined as an increase in K2 of 1.5 dioptres or more following the CXL procedure. 77% (17/22) of sample eyes showed an improvement in BCVA and 4.5% (1/22) showed no change. There was an average improvement in BCVA of 0.1 logMAR across all eyes. 90.9% (20/22) of the sample eyes used in analysis had no adverse outcome. Two sample eyes demonstrated some mild, corneal stromal haze during follow up.

Table 2: A comparison of mean values of outcomes between audit data and KERALINK trial data (3)

Outcome (Mean) Belfast Trust
Baseline Post-CXL (6 months) Baseline Post-CXL (6 months)* Post-CXL (18 months) Baseline Post-standard care (6 months)* Post-standard care (18 months)
K2 (D) 51.2 50.4 49.1 49.3 49.7 50.2 50.9 53.4
Kmax (D) 60.6 59.3 56.0 NA 57.0 57.2 NA 60.0
ACT (µm) 461.8 450.4 512.0 512.0 501.8 507.0 490.0 479.9
BCVA (logMAR) 0.305 0.193 0.500 0.400 0.400 0.500 0.500 0.600

*Values interpreted from graphical data


The average K2 at baseline was 51.2D compared to 49.1D in the KERALINK study CXL treatment group and 50.2D in the KERALINK standard care treatment group (3). The mean Kmax at baseline was also higher in our sample eyes than in the KERALINK trial groups (3). The average apical corneal thickness in our sample eyes was 50.2 micrometers less than that of the KERALINK trial CXL group eyes and 45.2 micrometers less than the KERALINK control group eyes (3). These differences suggest our sample of eyes had more advanced keratoconus at baseline with greater corneal distortion and thinning on average.

There was a reduction in mean K2 and Kmax following the CXL procedure in our sample eyes of 0.8D and 1.3D respectively, suggesting there was some reduction in corneal distortion following the CXL procedure (3). Importantly, this supports the KERALINK trial finding that CXL effectively halts progression of keratoconus.

Progression of keratoconus was defined as clinically significant in the KERALINK trial at an increase of 1.5D in K2 – this degree of change was determined to have an effect on vision (3). By the end of the trial at 18 months, 7% (2/30) of patients in the CXL group had experienced keratoconus progression compared with 43%(12/30) in the standard care group (3). In our sample of eyes at 6 months post CXL, 4.5% (1/22) of eyes experienced keratoconus progression. It is important to note that this comparison is limited due to the differing follow-up timeframes.

There were no adverse events in the follow up period for the 30 eyes in KERALINK trail that underwent CXL (3). There were reports of mild stromal haze following CXL in two of our sample eyes, wa grade I adverse event according to the NICE clinical audit tool (4). Despite this, in one of these eyes there was a 0.150 improvement in BCVA over the follow up period. The second eye showed a reduction in BCVA of 0.075.


Our data corroborates the KERALINK trial: CXL appears to be a safe and effective procedure that slows the progression of keratoconus in paediatric patients. Our data also suggests that there may be reduced corneal distortion and improvement in best corrected visual acuity following the procedure.

Following the clinical audit, a number of key recommendations were made for implementation to improve practice. Firstly, the development of an efficient data collection system based on the Royal College of Ophthalmologists and NICE guidance would allow for improvements in the recording of baseline and follow up data. Building this tool into the digital record for CXL patients would allow for ease and standardisation of the data to be collected.

At present, no quality of life data has been collected for our patients. Using a quality of life assessment tool such as the Cardiff Visual Ability Questionnaire or the Child Health Utility 9D questionnaire at baseline and follow-up for new keratoconus patients may be beneficial to identifying differences in quality of life before and after the CXL procedure.

Regular six-monthly follow-up of patients undergoing CXL with Pentacam measurements at each visit would allow for closer comparisons with the data of the KERALINK trial, determining the long- term effects of any adverse events, as well as providing patients with continuity of care. Finally, further development of our cohort of patients will be useful in assessing longer term outcome data including progression to corneal transplant and repeat procedures.


  1. Léoni-Mesplié S, Mortemousque B, Touboul D, Malet F, Praud D, Mesplié N, Colin J. Scalability and severity of keratoconus in children. Am J Ophthalmol. 2012 Jul;154(1):56-62.e1. doi: 10.1016/ j.ajo.2012.01.025. Epub 2012 Apr 24. PMID: 22534107.
  2. Hashemi, Hassan MD*; Heydarian, Samira PhD†; Hooshmand, Elham MSc‡; Saatchi, Mohammad PhD§; Yekta, Abbasali PhD¶; Aghamirsalim, Mohamadreza MD‖; Valadkhan, Mehrnaz M.Sc*; Mortazavi, Mehdi MD**; Hashemi, Alireza MD††; Khabazkhoob, Mehdi PhD. The Prevalence and Risk Factors for Keratoconus: A Systematic Review and Meta-Analysis. Cornea 39(2):p 263-270, February 2020. | DOI: 10.1097/ICO.0000000000002150
  3. Larkin DFP, Chowdhury K, Burr JM, Raynor M, Edwards M, Tuft SJ, Bunce C, Caverly E, Doré C; KERALINK Trial Study Group. Effect of Corneal Cross-linking versus Standard Care on Keratoconus Progression in Young Patients: The KERALINK Randomized Controlled Trial. Ophthalmology. 2021 Nov;128(11):1516-1526. doi: 10.1016/j.ophtha.2021.04.019. Epub 2021 Apr 21. PMID: 33892046.
  4. National Institute for Health and Care Excellence. (2013). Photochemical corneal collagen cross‑linkage   using   riboflavin   and   ultraviolet   A   for   keratoconus   and   keratectasia   (NICE Interventional procedures guidance [IPG466]). Available at: ipg466
  5. Allan B et al., Royal College of Ophthalmologists, 2016, Corneal Cross-linking Data Set 2016: Available at:, Date accessed: 02/02/2023

Leave a Reply