Photochemical corneal collagen cross-linkage using riboflavin and ultraviolet A for keratoconus (interventional procedures consultation)

Interventional procedure consultation document

Photochemical corneal collagen cross linkage using riboflavin and ultraviolet A for keratoconus and keratectasia

Strengthening fibres in the cornea using riboflavin eye drops and ultraviolet A light to treat keratoconus and keratectasia

Keratoconus is a degenerative condition that usually affects both eyes. It changes the shape of the cornea (the clear outer layer of the eye), causing blurred vision and light sensitivity. Keratoconus that occurs as a complication of laser surgery (LASIK) is called keratectasia. Spectacles or contact lenses can help some people with mild to moderate disease.

Photochemical corneal collagen cross linkage is performed as an outpatient procedure using local anaesthetic eye drops. Riboflavin (vitamin B2) drops are applied to the eye before exposing it to ultraviolet light. The procedure aims to stabilise the cornea and prevent further changes to its shape.

The National Institute for Health and Care Excellence (NICE) is examining photochemical corneal collagen cross linkage using riboflavin and ultraviolet A for keratoconus and keratectasia and will publish guidance on its safety and efficacy to the NHS in England, Wales, Scotland and Northern Ireland. NICE’s Interventional Procedures Advisory Committee has considered the available evidence and the views of specialist advisers, who are consultants with knowledge of the procedure. The Advisory Committee has made provisional recommendations about photochemical corneal collagen cross linkage using riboflavin and ultraviolet A for keratoconus and keratectasia.

This document summarises the procedure and sets out the provisional recommendations made by the Advisory Committee. It has been prepared for public consultation. The Advisory Committee particularly welcomes:

  • comments on the provisional recommendations
  • the identification of factual inaccuracies
  • additional relevant evidence, with bibliographic references where possible.

Note that this document is not NICE’s formal guidance on this procedure. The recommendations are provisional and may change after consultation.

The process that NICE will follow after the consultation period ends is as follows.

  • The Advisory Committee will meet again to consider the original evidence and its provisional recommendations in the light of the comments received during consultation.
  • The Advisory Committee will then prepare draft guidance which will be the basis for NICE’s guidance on the use of the procedure in the NHS in England, Wales, Scotland and Northern Ireland.

For further details, see the Interventional Procedures Programme manual, which is available from the NICE website (

Through its guidance, NICE is committed to promoting race and disability equality, equality between men and women, and to eliminating all forms of discrimination. One of the ways we do this is by trying to involve as wide a range of people and interest groups as possible in the development of our interventional procedures guidance. In particular, we aim to encourage people and organisations from groups who might not normally comment on our guidance to do so.

In order to help us promote equality through our guidance, we should be grateful if you would consider the following question:

Are there any issues that require special attention in light of NICE’s duties to have due regard to the need to eliminate unlawful discrimination, advance equality of opportunity, and foster good relations between people with a characteristic protected by the equalities legislation and others?

Please note that NICE reserves the right to summarise and edit comments received during consultations or not to publish them at all where in the reasonable opinion of NICE, the comments are voluminous, publication would be unlawful or publication would otherwise be inappropriate.

Closing date for comments: 21 June 2013

Target date for publication of guidance: 25 September 2013

1                      Provisional recommendations

Most of the published evidence on photochemical corneal collagen cross linkage (CXL) using riboflavin and ultraviolet A (UVA) for keratoconus and keratectasia relates to the technique known as ‘epithelium-off’ CXL. ‘Epithelium-on’ (transepithelial) CXL is a more recent technique and less evidence is available on its safety and efficacy. Either procedure (epithelium-off or epithelium-on CXL) can be combined with other interventions, and the evidence base for these combination procedures (known as ‘CXL-plus’) is also limited. Therefore, different recommendations apply to the variants of this procedure, as follows.

1.1                  Current evidence on the safety and efficacy of epithelium off CXL for keratoconus and keratectasia is adequate in quality and quantity. Therefore, this procedure can be used provided that normal arrangements are in place for clinical governance, consent and audit.

1.2                  Current evidence on the safety and efficacy of epithelium on (transepithelial) CXL, and the combination (CXL plus) procedures for keratoconus and keratectasia is inadequate in quantity and quality. Therefore, these procedures should only be used with special arrangements for clinical governance, consent and audit or research.

1.3                  Clinicians wishing to undertake epithelium on (transepithelial) CXL, or the combination (CXL plus) procedures should take the following actions:

  • Inform the clinical governance leads in their Trusts.
  • Ensure that patients and their parents or carers understand the uncertainty about the efficacy and safety of the procedures in the long term and provide them with clear information. In addition, the use of NICE’s information for the public [URL to be added at publication] is recommended.
  • Audit [URL to audit tool to be added at publication] and review clinical outcomes of all patients having these procedures for keratoconus and keratectasia.

1.4                  The procedures should only be offered to patients with progressive keratoconus or keratectasia and adequate corneal thickness.

1.5                  The procedures should only be carried out by ophthalmologists with expertise in managing corneal disease and specific training in the use of ultraviolet light.

1.6                  NICE encourages further research into CXL using riboflavin and UVA for keratoconus and keratectasia, especially epithelium on (transepithelial) CXL and the combination (CXL plus) procedures. Information reported should include visual acuity, corneal topography and quality of life outcomes, and long-term data for all types of CXL using riboflavin and UVA for keratoconus and keratectasia – specifically data about prevention of progression to corneal transplantation and about repeat procedures and their efficacy.

2                      Indications and current treatments

2.1                  Keratoconus is a natural degeneration of the structure of the cornea in which the corneal surface thins and begins to bulge into a cone shape. This causes refractive error, which is usually a myopic shift and is often associated with astigmatism, leading to visual impairment. It commonly affects children and young adults and may be progressive. Iatrogenic keratoconus (for example, as a result of laser-assisted in situ keratomileusis [LASIK] surgery) is called keratectasia.

2.2                  In mild to moderate keratoconus, visual acuity can be corrected using spectacles or contact lenses.  Keratectasia can be managed by using contact lenses. However, with disease progression, these cease to be of benefit and intracorneal ring segment (ICRS) implantation may be used. Corneal surgery, including deep lamellar keratoplasty or penetrating keratoplasty, may be needed for severe progressive keratoconus or keratectasia.

3                      The procedure

3.1                  The CXL procedures are done as outpatient procedures using topical anaesthesia, and typically take 60–90 minutes.

3.2                  In epithelium-off CXL, the epithelium is first abraded with a blunt spatula to allow penetration of riboflavin into the corneal tissue. Riboflavin eye drops are applied to the corneal surface 5 minutes before the procedure and every 5 minutes during the procedure. The corneal surface is exposed to UVA radiation, usually for about 30 minutes. Postoperatively, topical antibiotics and anti-inflammatory drops are normally prescribed, with topical steroids if necessary. In some cases, a bandage contact lens may also be used for a few days. The procedure is done on 1 eye at a time and may also be repeated if needed.

3.3                  In epithelium-on (transepithelial) CXL, the corneal epithelial surface is left intact and a longer riboflavin loading time is needed.

3.4                  Sometimes the procedure is used in combination with other interventions such as ICRS implantation, photorefractive keratectomy (PRK) or phakic intraocular lens (PIOL) implantation to improve visual acuity. These combination procedures are referred to as ‘CXL-plus’.

3.5                  The mechanism of action of the procedure is to increase the number of ‘anchors’ that bond collagen fibres together and strengthen the cornea. This is expected to stop the progression of the disease but the duration of benefit is uncertain.

4                      Efficacy

This section describes efficacy outcomes from the published literature that the Committee considered as part of the evidence about this procedure. For more detailed information on the evidence, see the systematic review.

4.1                  A systematic review of the published evidence on these procedures was commissioned by NICE.

Epithelium off CXL

4.2                  In the studies included in the systematic review for epithelium‑off CXL, meta-analysis of the change between preoperative and postoperative data for topography reported significant improvements in maximum keratometric values (max K) at 6, 12 and 24 months (−0.8 dioptres [D] at 6 months and −1.0 D at 12 and 24 months), and mean and minimum keratometric values (mean K and min K) at 12 months only (−1.0 D and −0.7 D for mean K and min K respectively).

4.3                  Meta analysis results for visual acuity (VA) reported significant improvement in corrected VA (CVA) (−0.20 LogMAR) but not uncorrected VA (UCVA) between intervention and control eyes on the LogMAR scale at 12 months follow up. Meta analysis of the change between preoperative and postoperative data showed significant improvements in UCVA postoperatively at 6, 12 and 24 months. Improvements on the LogMAR scale were in the order of −0.15 for UCVA and −0.10 for CVA at various time points.

4.4                  Meta-analysis results for astigmatism reported no significant differences between the treatment and control groups at 12 months (-1.42 D [-3.85;1.00]). Differences between preoperative and postoperative data showed significant improvements at 6, 12 and 24 months (−0.4 D at 6 months, −0.7 D at 12 months and −0.5 D at 24 months). Change in spherical equivalence (SE) was only significant at 12 months (there was a reduction of between 0.3 and 0.5 D).

4.5                  Meta-analysis of the change between preoperative and postoperative data showed a significant decrease (-14.4 micrometres) in central corneal thickness and no significant difference in intraocular pressure at 12 months follow-up.

Epithelium-off CXL in combination with other interventions (CXL-plus)

4.6                  A randomised controlled trial in 48 eyes with progressive keratoconus (43 patients) compared 2 sequences of combined CXL and intracorneal ring segment (ICRS) implantation with a mean interval between treatments of 7 months. Group 1 had epithelium-off CXL followed by ICRS implantation and group 2 had ICRS implantation followed by epithelium-off CXL .The mean UCVA and corrected distance VA (CDVA) had improved significantly in both groups 6 months after the procedures (each group gained 1 Snellen line in UCVA, and group 1 gained 1 line in CDVA but group 2 gained only half a line). The same study reported statistically significant improvements in the mean SE, cylinder and mean K values in both groups but there was more improvement in CDVA, SE (2.76 D versus 0.93 D), and mean K (3.3 D versus 1.1 D) in group 2 than in group 1.

4.7                  A randomised comparative case series of 42 eyes (21 patients) with bilateral keratoconus and 50 micrometres of epithelium removed by photorefractive keratectomy (PRK) compared 2 levels of exposure to UVA (both eyes were treated with CXL). In both groups mean UDVA improved, from 20/60 to 20/38 and from 20/62 to 20/40 Snellen lines respectively; CVA improved from 20/30 to 20/25 Snellen lines; mean SE was reduced by 2.5 D and 2.1 D; mean refractive cylinder was reduced by 2.9 D and 2.5 D; max K was reduced by 3.4 D and 2.9 D at 24 months.

4.8                  A case series of 11 eyes (11 patients with progressive keratoconus) treated with CXL 6 months before phakic intraocular lens (PIOL) implantation and followed up 6 months after the PIOL implantation reported statistically significant improvement in mean UCVA (by 0.24 LogMAR 6 months after CXL, and by 1.24 LogMAR 6 months after PIOL) and  CVA (by 0.02 LogMAR and 0.1 LogMAR at the 2 time points), reduction in mean K values (by 1.26 D and 2.14 at the 2 time points), SE (by 0.45 and 5.43 D at the 2 time points) and cylinder values (by 0.16 and 0.55 D at the 2 time points)

Epithelium-on (transepithelial) CXL with or without additional interventions (CXL plus)

4.9                  A comparative case series of 51 eyes (51 patients with progressive keratoconus) treated with epithelium-on (transepithelial) CXL reported improvements in mean CDVA (by 0.036 LogMAR for CXL versus 0.039 LogMAR for control, p<0.05), and reduction in mean K values (by 0.51 D after CXL versus 1.61 D for control, p>0.05) and SE (0.35 D after CXL versus 0.83 D for control, p<0.05) at 12 months follow-up.

4.10               A case series of 14 eyes (10 patients with mild to moderate keratoconus) treated with epithelium-on (transepithelial) CXL and combined same day corneal implants reported improvements in mean best CVA (BCVA) (from 0.24 LogMAR to 0.16 LogMAR, p=0.34), and mean K values (45.83 D to 44.03 D, p=0.0023) at 3 years follow up.

4.11               A case series of 21 eyes (13 patients) treated with ICRS implantation followed by CXL after a mean of 4 months reported improvements in mean UCVA and BCVA 6 months after CXL (UCVA from 0.05 to 0.23 LogMAR, p=0.951; BCVA from 0.18 to 0.41 LogMAR, p=0.08). The study also reported reductions in SE and cylindrical and mean keratometric values after ICRS implantation (2.8 D, p<0.05; 2.1 D, p<0.05; and 2.5 D, p<0.05 respectively) which were maintained 6 months after CXL.

4.12               The Specialist Advisers listed efficacy outcomes as arrest of progression of keratoconus and stabilisation of the corneal shape measured by topography, refraction and keratometry, refractive astigmatism, change in corneal thickness, cone apex power, quality of life and contact lens independence.

5                      Safety

This section describes safety outcomes from the published literature that the Committee considered as part of the evidence about these procedures. This safety evidence is not subdivided by procedure variant because similar adverse events were reported for each. For more detailed information on the evidence, see the systematic review.

5.1                  Infections were reported in 8 single case reports. In 4 of these patients there was no major long-term adverse impact, visual acuity was reduced in 1 patient and no further details were reported in 3 patients.

5.2                  Sterile keratitis associated with scarring or loss of vision or needing keratoplasty was reported in 3% (4/117) of patients in a case series of 117 patients. This was treated with high-dose topical or subconjunctival corticosteroids and 2 patients had a persistent decrease in best spectacle CVA. Deep stromal infiltrates with scarring at 2 months despite treatment with antibiotics was reported in a single case report.

5.3                  Stromal scar developed in 4 patients (3 cases in 1 study); in 3 patients, the UCVA increased significantly despite scars, and in the other patient visual acuity was corrected with a lens.

5.4                  Corneal oedema within 24 hours after CXL and inflammation for 2–3 weeks, iris atrophy and pigment dispersion were reported in a case series of 10 patients. These resolved in 1 patient and improved in 4 patients after treatment. Corneal oedema that developed in a single case report resolved after 6 months treatment but scarring and poor visual acuity remained.

5.5                  Corneal melting was reported in 1 single case report and perforation due to corneal melting was reported in 2 single case reports; 2 patients were treated by penetrating keratoplasty and antibiotics with no long-term adverse impact and 1 patient was initially treated with tissue glue and bandage contact lens application but needed penetrating keratoplasty on day 12.

5.6                  Corneal burn and ulcer were reported in 3 single case reports; 1 patient was treated by phototherapeutic keratectomy (CVA improved from 20/69 to 20/20 at 28 months) and treatment details were not reported for the other 2 patients.

5.7                  Corneal haze with diffuse subepithelial opacification and paracentral corneal thinning associated with scarring was reported in a single case report. This disappeared only gradually despite intensive therapy.

5.8                  Stromal haze was reported in all eyes in a randomised controlled trial of 10 patients comparing same-day ICRS and CXL with ICRS and CXL 6 months later. This resolved eventually in both groups (time period not reported). Mild posterior linear stromal haze at 1 month after CXL with PRK was reported in 46% of eyes (13/28) in a case series of 28 eyes (23 patients). At 12 months follow-up, this had decreased in density but had not completely disappear.

5.9                  The Specialist Advisers listed anecdotal adverse events as delayed epithelial healing, bilateral corneal infection and transient recurrent erosion syndrome.

6                      Committee comments

6.1                  The Committee noted that these procedures may be useful for people who are unable to wear contact lenses, which may include people with disabilities.

7                      Further information

7.1                  This guidance requires that clinicians undertaking the procedure make special arrangements for audit. NICE has identified relevant audit criteria and is developing an audit tool (which is for use at local discretion), which will be available when the guidance is published.

7.2                  This guidance is a review of Photochemical corneal collagen cross linkage using riboflavin and ultraviolet A for keratoconus NICE interventional procedure guidance 320 (2009).

Bruce Campbell

Chairman, Interventional Procedures Advisory Committee

May 2013


This page was last updated: 21 June 2013