FocalPoints

Clinical Modules for Ophthalmologists

VO L U ME X X X III NUMBER 11
N OV EMBER 2 01 5

Corneal Collagen Crosslinking

Aaron Chan, OD
Clara C. Chan MD, FRCSC, FACS

a
a

b
b

REVIEWERS AND CONTRIBUTING EDITOR CONSULTANTS

Elmer Y. Tu, MD, Editor for Cornea and External Disease José L. Güell, MD, PhD
Stephen E. Orlin, MD, Basic and Clinical Science Course Faculty, Section 8 R. Doyle Stulting, MD, PhD
Dasa V. Gangadhar, MD, Practicing Ophthalmologists Advisory Committee for Education
 FocalPoints™ Editorial Review Board
Eric Paul Purdy, MD, Bluffton, IN
Editor in Chief; Oculoplastic, Lacrimal, and Orbital Surgery
CONTENTS
Lisa B. Arbisser, MD, Bettendorf, IA
Cataract Surgery Introduction 1
Syndee J. Givre, MD, PhD, Raleigh, NC Keratoconus and Other Ectatic Conditions 1
Neuro-Ophthalmology
Deeba Husain, MD, Boston, MA Corneal Collagen Crosslinking 3
Retina and Vitreous Safety Considerations in CXL 4
Katherine A. Lee, MD, PhD, Boise, ID Indications 4
Pediatric Ophthalmology and Strabismus
Contraindications 4
W. Barry Lee, MD, FACS, Atlanta, GA
Refractive Surgery; Optics and Refraction
Standard Surgical Technique 4
Ramana S. Moorthy, MD, FACS, Indianapolis, IN
Ocular Inflammation and Tumors Postoperative Management
Sarwat Salim, MD, FACS, Milwaukee, WI After Standard Crosslinking 6
Glaucoma
Elmer Y. Tu, MD, Chicago, IL Complications 7
Cornea and External Disease
Results of Major Clinical Trials 7
FocalPoints Staff
Surgical Technique Variations 7
Susan R. Keller, Acquisitions Editor
Epithelial-On Versus Epithelial-Off
Kim Torgerson, Publications Editor Crosslinking 7
D. Jean Ray, Production Manager
Techniques for the Thin Cornea 8
Debra Marchi, CCOA, Administrative Assistant
Techniques for Shorter Light Treatments 9
Clinical Education Secretaries and Staff Combining Crosslinking With
Louis B. Cantor, MD, Senior Secretary for Clinical Education, Other Modalities 9
Indianapolis, IN
Future Directions 11
Christopher J. Rapuano, MD, Secretary for Ophthalmic
Knowledge, Philadelphia, PA Conclusion 11
Dale Fajardo, Vice President for Education
Clinicians’ Corner 12

Suggested Reading 15

WWW
READ MORE ONLINE This icon in text denotes additional content
accessible in your digital module.

Focal Points Cumulative Index 1983–2014

ONLINE VIDEO This icon in text denotes video content www.aao.org/fpindex
accessible in your digital module.

The American Academy of Ophthalmology is ac-
credited by the Accreditation Council for Continuing
Medical Education to provide continuing medical
education for physicians.
The American Academy of Ophthalmology des-
ignates this enduring material for a maximum
of 2  AMA PRA Category  1  Credits™. Physicians
should claim only the credit commensurate with
the extent of their participation in the activity.
Claiming CME Credit: The American Medical
Association requires that all learners participating in
activities involving enduring medical materials com-
plete a formal assessment before claiming continu-
ing medical education (CME) credit. To assess your
achievement in this activity and ensure that a speci-
fied level of knowledge has been reached, a post
test for this module is provided. A minimum score
of 80% must be obtained to pass the test and claim
CME credit.
To claim Focal Points CME credits, visit the Acad-
emy web site (one.aao.org/CME-Central) to report
CME credit you have earned. You can claim up to
2 AMA PRA Category 1 Credits™ per module. CME
credit may be claimed for up to three (3) years from
date of issue.
Focal Points (ISSN 0891-8260) is published quar-
terly by the American Academy of Ophthalmology
at 655 Beach St., San Francisco, CA 94109-1336.
For domestic subscribers, print with digital 1‑year
subscription is $205 for Academy members and
$245 for nonmembers. International subscribers,
please visit www.aao.org/focalpoints for informa-
tion about shipping charges. Digital-only 1‑year sub-
scription is $165 for Academy members and $205
for nonmembers. Periodicals postage paid at San
Francisco, CA, and additional mailing offices. POST-
MASTER: Send address changes to Focal Points,
P.O. Box 7424, San Francisco, CA 94120-7424.
The Academy provides this material for educa-
tional purposes only. It is not intended to represent
the only or best method or procedure in every case,
nor to replace a physician’s own judgment or give
specific advice for case management. Including
all indications, contraindications, side effects, and
alternative agents for each drug or treatment is
beyond the scope of this material. All information
and recommendations should be verified, prior to
use, with current information included in the manu-
facturers’ package inserts or other independent
sources and considered in light of the patient’s
condition and history. Reference to certain drugs,
instruments, and other products in this publication
is made for illustrative purposes only and is not
intended to constitute an endorsement of such.
Some material may include information on applica-
tions that are not considered community standard,
that reflect indications not included in approved
FDA labeling, or that are approved for use only in re-
stricted research settings. The FDA has stated that
it is the responsibility of the physician to determine
the FDA status of each drug or device he or she
wishes to use, and to use them with appropriate in-
formed patient consent in compliance with applica-
ble law. The Academy specifically disclaims any and
all liability for injury or other damages of any kind,
from negligence or otherwise, for any and all claims
that may arise out of the use of any recommen-
dations or other information contained herein. The
author(s) listed made a major contribution to this
module. Substantive editorial revisions may have
been made based on reviewer recommendations.
Subscribers requesting replacement copies
6 months and later from the cover date of the issue
being requested will be charged the current module
replacement rate.
©2015 American Academy of Ophthalmology®.
All rights reserved.

ii FocalPoints Module 11, 2015

 LEARNING OBJECTIVES
Upon completion of this module, the reader
should be able to:
 Describe the common features of
keratoconus and other corneal ectatic
conditions
 Describe the mechanism of collagen
crosslinking as well as indications and
contraindications for its use
 Describe the surgical technique for
crosslinking and appreciate variations in
surgical technique
Introduction
The concept of crosslinking as a means to stiffen tissue
and materials has been known for some time, but it was
not until 1998 when Eberhard Spoerl and Theo Seiler,
researchers from the University of Dresden, began con-
sidering crosslinking as a therapeutic treatment for cor-
neal diseases. The idea was based on the observation that
the cornea naturally crosslinks itself to a more rigid and
stronger state with age, suggesting that simulated cross-
linking with ultraviolet‑A (UVA) light and riboflavin
(vitamin B2) may have a similar effect. Since then, cor-
neal collagen crosslinking (CXL) has become a widely
accepted treatment for keratoconus (KC) and related
ectatic conditions. While conventional management op-
tions (contact lenses and glasses) for KC are effective
in improving vision, they do not address the underlying
weakness in the cornea. Corneal crosslinking, however,
is effective because of its ability to stiffen the cornea,
which has been demonstrated to slow progression, and
in many cases, halt the disease. Though many trials
have proven successful, the practicing ophthalmologist
should be mindful of the relatively nascent nature of this
technology and the many controversies and questions,
such as the long-term effect of CXL, that remain to be
answered in this rapidly evolving procedure.
KERATOCONUS AND OTHER ECTATIC
CONDITIONS
Keratoconus is a degenerative disorder that is character-
ized by paracentral thinning and ectasia. It is the most
common corneal dystrophy, affecting approximately
1 in 2000 in the general population with no conclusive
predilection for gender or race. The condition is typi-
cally bilateral but asymmetric. As the cornea thins, it
protrudes anteriorly, forming a conical cross section that
most commonly decenters inferotemporally. The protru-
sion leads to a steepened curvature and scarring that may
manifest as striae at the level of Descemet membrane.
The resultant changes in shape cause significant irregu-
lar astigmatism, myopia, and reduced best-­ corrected
visual acuity (BCVA). Figure 1 shows a corneal topog-
raphy that one might expect in a patient with keratoco-
nus. KC is initially managed with spectacle correction
alone; however, as the disease progresses, the patient of-
ten requires rigid contact lenses to overcome the optical
aberrations. In its advanced stages, it is estimated that
10%–20% of patients require corneal transplantation.
The etiology of KC is not well understood as it in-
cludes genetic, biochemical, and physical factors. There
is no sole theory elucidating its range of clinical pre-
sentation, and KC is likely the eventual manifestation
of several different conditions. The condition usually
appears in isolation, but there have been reports of a
number of systemic and ocular associations, including
connective tissue disorders and vernal disease. Ocular
trauma associated with eye rubbing, contact lens wear,
and allergic eye disease also appears to be related. In-
creased activity of protein enzymes has been identified

Financial Disclosures
The authors, reviewers, and consultants dis-
close the following financial relationships:
Lisa  B. Arbisser, MD: (C) Bausch  + Lomb
(2013, 2014). (C, L, S) OptiMedica (2013).
(L) Abbott Medical Optics (2014). Clara
C. Chan, MD, FRCSC, FACS: (C) Bausch  +
Lomb (2013–2015), Tearscience (2015). (C,
L) Allergan (2013–2015). (L) Alcon Laborato-
ries (2013–2015). (S) Allergan (2015). José L.
Güell, MD, PhD: (C) Alcon Laboratories, Carl
Zeiss, OPHTEC BV, RVO Raindrop, Thea
(2015). (O) Calhoun Vision, Orca Surgical
(2014, 2015). W.  Barry Lee, MD, FACS: (C)
Shire (2015). (L) Allergan, Bausch  + Lomb
(2013–2015). Eric Paul Purdy, MD: (L) Al-
con Laboratories (2014). Sarwat Salim, MD,
FACS: (L) Alcon Laboratories (2013, 2014),
Merck & Co (2013). R. Doyle Stulting, MD,
PhD: (C) Abbott Medical Optics, Calhoun Vi-
sion, Hoya, NuLens, Optovue (2014, 2015);
OPHTEC (2015). (C, L) Alcon Laboratories
(2014, 2015). (C, O) Cambium Medical Tech-
nologies, EyeYon, TearLab (2014, 2015). (L)
Allergan (2014, 2015). Elmer  Y. Tu, MD:
(C) Bausch  + Lomb (2013), Seattle Genetics
(2014, 2015).
The following contributors state that they
have no financial interest or other relation-
ship with the manufacturer of any commer-
cial product discussed in their contributions
to this module or with the manufacturer of
any competing commercial product: Syn-
dee  J. Givre, MD, PhD; Deeba Husain, MD;
Susan R. Keller; Katherine A. Lee, MD, PhD;
Ramana S. Moorthy, MD, FACS; Kim Torger-
son (2013–2015). Eric Paul Purdy, MD (2013,
2015). Dasa Gangadhar, MD; Stephen  E.
Orlin, MD (2014). Aaron Chan, OD (2014,
2015). Lisa  B. Arbisser, MD; Sarwat Salim,
MD, FACS (2015).
C = Consultant fee, paid advisory boards or fees for at-
tending a meeting (for the past 1 year)
E = Employed by a commercial entity
L = Lecture fees (honoraria), travel fees or reimburse-
ments when speaking at the invitation of a com-
mercial sponsor (for the past 1 year)
O = Equity ownership/stock options (publicly or pri-
vately traded firms, excluding mutual funds)
P = Patents and/or royalties that might be viewed as
creating a potential conflict of interest
S = Grant support for the past year (all sources) and all
sources used for this project if this form is an up-
date for a specific talk or manuscript with no time
limitation

FocalPoints Module 11, 2015 1

 Figure 1  Topographic image showing inferior steepening in the right eye, worse than the left eye in a patient with
keratoconus.

Figure 2  Topographic image showing high against-the-rule astigmatism from pellucid marginal degeneration in the
right eye.

in keratoconic corneas, resulting in weakened biome-
chanical stability. Growing evidence suggests that tissue
degradation and corneal thinning involve the expression
of inflammatory mediators. The pathogenesis and pro-
gression of keratoconus is a very active area of research
that warrants a separate discussion, one that is beyond
the scope of this module. It is, however, important to be
aware of some of the ongoing debate about theories of
inflammation and progression in keratoconus.
Pellucid marginal degeneration (PMD) is less com-
mon than keratoconus and usually affects the inferior
peripheral cornea rather than the paracentral cornea
(Figure  2). PMD is characterized by a peripheral band
of thinning from approximately the 4  o’clock to the
8 o’clock position, which gives a classic “crab-claw” ap-
pearance on topography. Consequently, managing PMD
with penetrating keratoplasty is more difficult, as larger
and more eccentric grafts are required, raising the risk for
postoperative astigmatism and graft rejection. Crosslink-
ing for PMD is performed in the same fashion as with
KC. Although the area of steepening is more peripheral
in PMD, the central 8–9 mm treatment zone appears to

2 FocalPoints Module 11, 2015

 Figure 3  Topographic image showing irregular astigmatism and inferior steepening in a patient with post-laser
keratoectasia in the left eye.
be adequate in coverage. Along with keratoglobus, there
seems to be overlap between PMD and keratoconus.
Progressive post-laser keratoectasia (PPLK) is a con-
dition of gradual corneal thinning and steepening fol-
lowing LASIK or photorefractive keratectomy (PRK)
surgery (Figure  3). It generally occurs within 2  years
after refractive surgery and is estimated to occur in 1 in
2500 to 1 in 5000 cases. The underlying cause of PPLK
is not well understood but is thought to be from the
weakening of the anterior stroma from flap creation,
insufficient residual stromal bed thickness after tissue
ablation, or from the uncovering of a previously undi-
agnosed ectasia such as forme fruste keratoconus. Risk
factors for PPLK include thin corneas, repeated en-
hancement treatments, irregular preoperative topogra-
phies, and deep ablations.
Corneal Collagen
Crosslinking
CXL specifically involves the activation of riboflavin
(vitamin B2) with UVA light to enhance covalent bond
formations (or “crosslinks”) between collagen fibers in
the stroma. The process begins with the activation of the
photosensitive riboflavin to an excited state. The excited
riboflavin then reacts with oxygen, creating reactive
oxygen species and radicals. These reactive species theo-
retically induce covalent crosslink bonding between the
protein molecules within the stroma including collagen
and proteoglycans.
Considerable evidence supports the creation of
these crosslink formations following CXL. In a seminal
study, Spoerl and Seiler irradiated porcine and rabbit
corneas presoaked in photosensitizing agents, and re-
sultant corneas were stiffer and more resistant to en-
zymatic degradation. Subsequent in vivo studies have
confirmed increases in corneal rigidity, with an original
study by Wollensak supporting a >320% gain in corneal
biomechanical strength. Later reports claimed that cor-
neal stiffening was dependent on depth, with the effect
confined mostly to the anterior 200 µm of the stroma.
In regard to the crosslinked cornea being more resistant
to enzymatic degradation, it is postulated that changes
to the tertiary structure of the collagen fibers prevent en-
zymes (eg, collagenases) from accessing specific cleavage
sites. This may be helpful in understanding in part the
pathogenesis of keratoconus, as some studies indicate a
disruption in the balance between proteolytic enzymes
and their endogenous inhibitors in KC corneas, imply-
ing more proteolytic activity and fewer protein compo-
nents when compared to normal controls.
FocalPoints Module 11, 2015 3
SAFETY CONSIDERATIONS IN CXL
The use of UVA light in CXL warrants a discussion on
potential negative effects on ocular structures. Exposure
to UVA radiation is cytotoxic to structures that lie deep
to the intended stromal tissue including the endothe-
lium, lens, and retina. In fact, several case series have
reported endothelial and iris damage from UVA irra-
diation in CXL. Standard parameters in thickness, ir-
radiance, wavelength, and exposure time have thus been
chosen to maximize safety while ensuring efficacy.
The current recommendation is that any patient
treated with CXL should have a corneal thickness of
at least 400  µm at the thinnest point after the epithe-
lium has been removed. Treatment is not safe below this
400 µm threshold and an alternative technique should
be considered if such is the case (see “Techniques for the
Thin Cornea” later in this module). This recommenda-
tion is based on the observation that the standard irra-
diance of 3 mW/cm2 is effectively 0.18 mW/cm2 at the
level of the endothelium of a 400 µm riboflavin-­imbibed
stroma, which is two-fold lower than the irradiance
considered cytotoxic to endothelial cells. The chosen
wavelength of 365–370 nm is outside the range that can
cause photokeratitis (270–315 nm) and cataract forma-
tion (290–360 nm), and the standard treatment time of
30 minutes delivers a total dose of 5.4 J/cm2.
The very nascent nature of this technology has re-
sulted in a number of groups modifying the standard
parameters by changing irradiance, duration of UVA
exposure, and riboflavin concentration in order to ca-
ter for thinner corneas and more rapid treatments (see
“Techniques for Shorter Light Treatments” later in this
module). The variety of these new protocols has raised
some controversies as to the safest and most efficacious
way to perform CXL.
INDICATIONS
The primary indication for CXL is progressive kerato-
conus. However, it is critical to emphasize the challenge
in defining “progression” as it relates to indications
for treatment and analysing outcomes. The prevailing
goal of CXL is to stop progression; however, the term
is not standardized and can be arbitrarily set for each
study. Previous studies have defined the progression of
keratoconus with diverse parameters, from the clinical
progression that necessitates transplantation to several
topographic indices (eg, +1.0 D change from baseline in
Kmax). As such, there is little uniformity in the litera-
ture, and this has caused much confusion in comparing
data and interpreting successful outcomes.
Regardless, there is general consensus in the ability
of CXL to stiffen the cornea, and it is thus often rec-
ommended as a first-line treatment option. Despite this,
CXL may not be suitable for every keratoconic patient.
KC typically manifests at around puberty and progresses
at variable rates until the age of 30–40, after which it
often stabilizes. Though the etiology of KC is largely
4 FocalPoints Module 11, 2015
unknown, a part of its stabilization is explained by the
cornea’s inherent ability to stiffen with age. Therefore,
the newly diagnosed and/or rapidly progressing young
patient has the most potential benefit from CXL. Con-
versely, the treatment may be unnecessary for a 40‑year
old keratoconic patient with no evidence of progression.
To date, applicable treatment age and progression crite-
ria are still being studied and remain controversial.
CONTRAINDICATIONS
CXL should be avoided in a number of situations; how-
ever, since there are no official criteria, treatment is de-
pendent on the surgeon’s discretion and management of
patient expectations. Patients very advanced in their dis-
ease tend not to make good candidates. These patients
typically have thin corneas and steep maximal corneal
curvatures. One study indicates that patients with pre-
operative keratometry readings of greater than 58.00 D
have a greater risk of treatment failure and postopera-
tive haze formation. Additionally, this study suggests
that patients who are older than 35  years of age and
present with better than 20/25 preoperative corrected
distance visual acuity are at a higher risk for vision
loss from stromal scarring. Finally, eyes with preopera-
tive corneal thicknesses of less than 400 µm should be
avoided for the concern of damage to deeper structures.
As a result of these findings, the authors recommend
that patients be treated with the requirements of being
less than 35  years old, having greater than 400  µm of
corneal thickness at the thinnest point after epithelial
debridement, and measuring less than 58.00 D on maxi-
mal keratometry readings.
Factors that can potentiate postoperative complica-
tions, such as delayed healing or infection, also need to
be seriously considered prior to treatment. Severe dry
eye, vernal/atopic inflammatory disease, and concurrent
infections, such as herpetic keratitis, should be managed
or treated appropriately before considering CXL. There
is also a case study suggesting that pregnancy limits
the effects of CXL and may exacerbate corneal ectasia.
Though the evidence is limited, it is probably prudent
to avoid CXL during and immediately after pregnancy.
Lastly, severe central scarring with associated decreased
visual acuity is unlikely to be improved after CXL, so
alternate treatment options such as corneal transplanta-
tion should be considered.
Standard Surgical
Technique
The most accepted treatment protocol is based on
the original Dresden protocol, an epithelium-­ off ap-
proach first described by Wollensak et al (2003). It in-
volves 0.1% riboflavin combined with UVA-­irradiation
(370 nm; @ 3mW/cm2), providing 5.4 J/cm2 of energy to
the cornea. The following are the step-by-step instruc-
tions for the Dresden protocol:
1. Under sterile operating conditions, the patient is
draped and the cornea is anaesthetised topically
with 2–3 applications of 0.5% proparacaine or
0.5% tetracaine ophthalmic solution.
2. The orbital area is cleaned with 10% povidone-­
iodine solution (Betadine) and a lid speculum is
applied (Figure 4).
3. The central 7–9 mm of the corneal epithelium is
debrided (Figure 5).
4. A baseline corneal thickness measurement is Figure 5  Removal of corneal epithelium using a
taken with an ultrasound pachymeter to ensure Beaver blade and dry Weck-Cel ophthalmic sponge.
thickness of 400 µm (Figure 6).
5. 0.1% iso-­osmolar riboflavin in 20% dextran
solution is instilled on the corneal surface every
2–3 minutes for 30 minutes (Figure 7).
6. The patient is examined under slit lamp with
blue light to ensure riboflavin has fully infused
the cornea. (Check for riboflavin in the anterior
chamber by presence of a yellow coloration.)
7. Corneal pachymetry is remeasured before UVA
irradiation.
8. If it is safe to continue, a sponge ring is placed
around the limbus of the cornea to protect the
limbal stem cells (Figure 8a).
9. UVA light is focused directly on the area
of absent epithelium to photoactivate the
riboflavin-­imbibed stroma.

Figure 6  Measurement of corneal pachymetry to

ensure 400 µm thickness prior to UVA irradiation.

Figure 4  After application of povidone iodine to the

periorbital region and insertion of a lid speculum, an
8 mm corneal optic zone marker is filled with alcohol
for 10 seconds. (Bores Optic Zone Marker, Katena, Figure 7  Application of 0.1% riboflavin in 20% dextran
Denville, New Jersey.) onto the cornea every 2–3 minutes for 30 minutes.

FocalPoints Module 11, 2015 5

 a
b
Figure 8  UVA treatment of riboflavin-­imbibed
cornea. a. Limbal stem cells are protected with
a Kerocel Chayet Drainage Ring (Beaver Visitec,
Waltham, Massachusetts). b. Irradiation commences
for 30 minutes with continued riboflavin instillation.
10. Irradiation commences for 30 minutes, with
continued riboflavin instillation every 3 minutes
(Figure 8b).
11. Alternating applications of balanced saline
solution (BSS) and topical anaesthetic are given
intermittently to prevent desiccation and to
maintain corneal anesthesia.
12. Corneal pachymetry is measured at 10, 20, and
30 minutes after riboflavin drops have started
to ensure that the stromal thickness remains
greater than 400 µm. Should the pachymetry
measurement fall below 400 µm, hypotonic
riboflavin drops are used instead of the routine
isotonic riboflavin drops.
13. Upon completion of the UVA irradiation
treatment, a broad-­spectrum antibiotic, such as
moxifloxacin, is instilled and a bandage contact
lens is positioned onto the cornea (Figure 9).
6 FocalPoints Module 11, 2015
Figure 9  Application of bandage contact lens after
UVA treatment.
This lens is left until the cornea is fully re-­
epithelized, which usually occurs in 4–7 days.
14. The patient is discharged with a course of
antibiotics and topical steroids. Regular
follow‑up should occur until the epithelium has
fully healed and the contact lens can be removed.
Postoperative
Management After
Standard Crosslinking
The postoperative management of crosslinking is a simi-
lar experience to postoperative management after PRK.
The early healing stage is most symptomatic, and pa-
tients may experience symptoms of pain, tearing, foreign
body sensation, and photophobia that are consistent
with having a corneal abrasion. A bandage contact lens
is used to support re-­epithelialization and reduce pain.
Topical antibiotics and preservative-­free lubricants are
given until re-­epithelialization and pain can be man-
aged with oral opioids and analgesics. The epithelium
typically heals by day 4–7, at which point the bandage
contact lens is removed and a regimen of topical ste-
roids is given to reduce scarring and promote further
healing. (The author uses dexamethasone 4 times daily
for 2–4  weeks, tapering to twice daily for 2–4  weeks
depending on the presence or absence of any haze, but
there is much variation in the literature.) Patients are
followed up at 1 day, 4–7 days for bandage contact lens
removal, 1 month to monitor for haze, then further fol-
low‑up at 3 months, 6 months, and 1 year. Thereafter,
patients are seen every 6–12 months to monitor for pro-
gression via serial manifest refraction and topography
with their optometrist or ophthalmologist.
Complications
CXL, like any surgical procedure, involves potential
risks. Intraoperatively, improper UVA delivery from
excessive energy, incorrect wavelength, or insufficient
amount of riboflavin in the stroma can lead to possible
corneal, iris, lens, and retinal damage. Corneas with a
thickness less than 400 µm after de-­epithelization have
a greater chance for endothelial cell and permanent vi-
sion loss. It has been demonstrated that extreme cor-
neal dehydration following debridement can exacerbate
thinning, thus extra care is warranted to ensure appro-
priate corneal thickness and hydration throughout the
duration of the treatment. Sterile infiltrates have been
reported. In rare cases, severe inflammatory responses
have occurred within 24 hours of UVA treatment, with
patients developing iritis, keratic precipitates, and cor-
neal edema. Research suggests that UVA light induces
apoptosis in keratocytes, which may be linked to an ini-
tial inflammatory response. However, keratocyte dam-
age is of lesser concern as a normal cell population is
typically regained 6 months after treatment.
Postoperatively, there is added risk for haze, scar-
ring, delayed wound healing, corneal melt, and infectious
keratitis, especially when the epithelium is removed.
Though permanent scarring is rare, stromal haze is fre-
quently noted after CXL and usually responds well to
topical steroids, resolving within 6 months to a year. A
transient decrease in corneal sensitivity and innervation
can also be observed up to 6 months postoperatively,
although this has minimal effect on tear film stability
or basic tear secretion. Close observation during the
epithelial healing phase is important to monitor for in-
fection and delayed wound healing, the latter of which
has been linked to limbal stem cell damage or poor epi-
thelial cell migration over a steepened cone. In spite of
this, potential chronic complications such as late lim-
bal stem cell deficiency or relationship to ocular surface
malignancy have yet to be demonstrated in literature.
Finally, it is worthwhile to mention that many of the
earlier studies reporting adverse events lacked explicit
inclusion and exclusion criteria, thus potentially treat-
ing at-risk patients who would be otherwise excluded
today. Regardless, the incidence of complications from
CXL represents an extremely small proportion of eyes,
and patients should be counselled accordingly.
Results of Major
Clinical Trials
The results of major clinical trials are listed in Table 1.
Of the many peer-­reviewed clinical trials, nearly all re-
port halting moderate to advance progressive kerato-
conus, often with some regression of corneal steepness
(Kmax). On average, studies report a Kmax reduction
of approximately 2.00  D, resulting in improved visual
outcomes. Though studies indicate stability for up to
5 years, the exact duration of CXL’s effect on the cornea
is unknown, and longer-­term data remains to be seen.
In examining the data presented in the studies, one
should be aware that there are significant inconsisten-
cies in evaluating the literature. As each study is inde-
pendently conducted, there is variability in the way
Kmax and other parameters are measured. The study
of the disease in itself is confounded by variability from
patient to patient or even between the 2  eyes of a pa-
tient. As aforementioned, factors such as defining the
rate of progression, time of onset, natural course of
disease, genetic components, and physical environment
vary dramatically between patients diagnosed with this
same condition. In this rapidly evolving field, one should
be cognisant of the variability and confounding factors
that have a made it difficult to study this chronic and
progressive disease with relatively small and short-term
studies.
Surgical Technique
Variations
EPITHELIAL-ON VERSUS EPITHELIAL-OFF
CROSSLINKING
The discussion between epithelial-­ on (epi‑on) versus
the standard epithelial-­off (epi‑off) CXL is arguably the
most contentious topic in crosslinking research today. As
the name suggests, epithelial‑on or transepithelial CXL
leaves the epithelium intact as riboflavin penetrates to
the stroma. Riboflavin, a large molecule, is impermeable
to the intact epithelium, thus necessitating its removal in
standard crosslinking. However, the theoretical benefits
of leaving the epithelium on include less risk of infec-
tion, reduced haze, reduced pain, and improved healing
times. Accordingly, various approaches have been tried
clinically and in the laboratory to enhance epithelial
permeability. Riboflavin preparations with trometamol,
benzalkonium chloride (BAC), and/or EDTA, among
other drugs have demonstrated an ability to loosen epi-
thelial tight junctions. “Mechanical disruption” to cre-
ate pockmarks or grid patterns in the epithelium have
been utilized to facilitate diffusion. Unfortunately, at the
expense of leaving the epithelium on, treatments may
take significantly longer than epi‑off CXL to ensure the
riboflavin has adequately infused the stroma. The drugs
used are also toxic to the epithelium, which can be espe-
cially uncomfortable for the patient.
There is much debate to epi‑on CXL’s effectiveness
as it compares to the “gold-­standard” of epi‑off CXL.
Experimental support for epi‑on CXL is controver-
sial. Wollensak experimented with epi‑on in a labora-
tory and found it to be 20% as effective as epi‑off. A
number of subsequent reports have concluded similar
decreases in efficacy. On the contrary, there are a few
FocalPoints Module 11, 2015 7
 Table 1. Results of Major Clinical Studies for Standard CXL

MAXIMUM NO. OF EYES % HALTED OR

AUTHORa FOLLOW-UP START/END IMPROVED IMPROVEMENT
Wollensak, et al (2003) 4 years 23/2 95.5 (70) Kmax 2.01 D,
BCVA 1.26 lines
SE –1.14 D
Caporossi, et al (2006) 3 months 10/10 — Kmax 1.90 D
BCVA 1.66 lines, UCVA 3.6 lines
Raiskup-Wolf, et al (2008) 6 years 241/5 81 (57) Kmax 2.44 D
BCVA –0.18 LogMAR
Jankov, et al (2008) 6 months 25/25 100 (52) Kmax 2.14 D
UVCA –0.11 LogMAR
Wittig-Silva, et al (2008) 12 months 33/9 (>50) Kmax 1.45 D
BCVA–0.12 LogMAR
Vinciguerra, et al (2009) 2 years 28/28 — Kmax 1.35 D
BCVA –0.15 LogMAR, UCVA –0.24 LogMAR
Agrawal (2009) 1 year 37/37 92 (54) Kmax 2.47D
BCVA improved >1 line
Coskunseven, et al (2009) 1 year 19/19 — Kmax 1.57D
BCVA –0.10 LogMAR, UCVA –0.06 LogMAR
Koller, et al (2009) 1 year 192/155 98 (37.7) Kmax 0.89D
BCVA –0.55 LogMAR
Derakhshan, et al (2011) 6 months 31/31 90.3 (77) Kmax 0.65 D
BCVA 1.7 lines, UCVA 2 lines
Hersh, et al (2011) 1 year 49/49 89.8 (51.0) Kmax 2.00 D
BCVA –0.14 LogMAR, UCVA –0.05 LogMAR
Viswanathan, et al (2013) 4 years 51/? — Kmax 0.96 D
BCVA –0.05 LogMAR
Goldich, et al (2012) 2 years 14/14 92.8 Kmax 2.40 D
BCVA –0.07 LogMAR
Vinciguerra, et al (2012) 2 years 40/40 — Kmax 1.27 D
(<18-year-olds studied) BCVA –0.19 LogMAR, UCVA –0.21 logMAR
SE –1.57 D
Vinciguerra, et al (2013) 4 years 400/? — BCVA:
–0.11 LogMAR (<18 years old @ 12 months)
–0.31 LogMAR (18–29 years old @ 36 months)
–0.33 LogMAR (30–39 years old @ 36 months)
–0.25 LogMAR (>40 years old @ 36 months)
a
Refer to Suggested Reading for citations.
Kmax = keratometry value in the steepest meridian, BCVA = best-corrected visual acuity, UCVA = uncorrected visual acuity, SE = spherical equivalent refraction

small, published studies that suggest epi‑on is at least as
efficacious as epi‑off. The results of these epi‑on CXL
studies are listed in Table 2. There is also a very large,
multicenter, nonrandom, prospective, research effort un-
derway in the United States that cites excellent initial re-
sults. The CXL‑USA Study Group supports that epi‑on
CXL provides patients with equivalent or superior out-
comes to epi‑off CXL while causing significantly less
pain and healing time. The study group has treated pa-
tients as young as 9 and has even found benefit in treat-
ing patients more than 35. Readers should be mindful
that the group’s UV light source and riboflavin formula-
tion are proprietary and much of their data is yet to be
published. At present, all published long-term data have
been performed with epi‑off procedures. Further clinical
studies with longer follow-­up and randomized controls
are required to ascertain the efficacy of epi‑on CXL as it
compares to the current standard of epi‑off CXL.
TECHNIQUES FOR THE THIN CORNEA
Many patients with advanced keratoconus have corneas
that are under the threshold for treatment (<400  µm

8 FocalPoints Module 11, 2015

 research needs to be conducted to confirm the safety of
Table 2. Results of Clinical Studies for
such newer techniques.
Epi-On CXL
NO. OF EYES COMBINING CROSSLINKING WITH OTHER
AUTHORa FOLLOW-UP START/END IMPROVEMENT MODALITIES
Filippello, 18 months 20/20 Kmax: 2.97D Although in most cases, CXL halts the ectatic process
et al (2012) UCVA –0.66 and can result in mild to modest gains in vision, the
LogMAR, BCVA:
–0.11 LogMAR procedure alone typically does not provide significant
visual recovery, and patients should be counselled ap-
Koppen, Up to 18 53/18 Sphere, Cyl, Kmax
et al (2012) months and refractive propriately. Investigators therefore, have experimented
power remained with combining CXL with various refractive techniques
stable from with the goal of stabilizing the cornea as well as improv-
baseline without
statistically ing vision. The results for some of these clinical trials are
significant listed in Table 3.
improvements at
18 months. CXL IN COMBINATION WITH INTRASTROMAL
Lessicotti, 12 months 63/51 BCVA: –0.036 RINGS. Intrastromal corneal ring segments (ICRS) are
et al (2012) LogMAR
an effective treatment option for keratoconic patients
SE: 0.35D
who are contact lens intolerant. Intacs (AJL Ophthamic,
(overall favorable
but limited effect)
Vitoria-­Gasteiz, Spain) and the Ferrara Ring/Keraring
(Mediphacos, Belo Horizonte, Brazil) are specific ex-
a
Refer to Suggested Reading for citations.
amples of ICRS that consist of semicircular polymethyl-
methacrylate (PMMA) segments which can be implanted
into a channel deep in the stroma. The segments flatten
after epithelial debridement). Hypo-­ osmolar ribofla- the central cornea and, with strategic positioning, move
vin solutions have been used to swell corneal stromas the decentered cone to a more central position, reduc-
to >400 µm, allowing for the treatment of thin corneas ing optical aberrations such as coma and improving vi-
that would otherwise be excluded. It appears that the sion and/or contact lens fit. Figure 10 depicts an Intacs
effect is transient, so surgeons must carefully monitor segment properly positioned in the stroma. Studies have
pachymetry throughout the entire CXL treatment, in- indicated good visual recovery with more than a 2‑line
stilling additional hypo-­osmotic riboflavin on a regular improvement in BCVA. Though the exact mechanism is
basis. There is good evidence that subsequent CXL treat- unknown, it is believed that the remodelled cornea re-
ment stabilizes the progression of keratoconus with no distributes and dampens the biomechanical strain on the
side effects or damage to the endothelium with results cone to slow progression of the disease. As such, there
up to a year post-­treatment. As with other techniques, are trials, including those of Joseph Colin, that have
additional research is required to see if corneas treated demonstrated significant stability in visual outcomes
with hypo-­osmolar riboflavin have lasting effects. and in halting keratoconus progression over many
years. With the establishment of CXL, a few investiga-
TECHNIQUES FOR SHORTER LIGHT
tors have begun looking at ways to combine the 2 treat-
TREATMENTS
ments in order to synergize their effects. Initial reports
Recent modifications have been made to reduce CXL suggest that it is safe and effective to combine CXL with
surgery time. The current standard protocol takes ap- Intacs insertion.
proximately an hour, which can be considerably uncom-
fortable for the patient. Furthermore, longer exposure CXL IN COMBINATION WITH PRK. Patients with
time puts corneas at risk for desiccation, which can de- keratoconus often suffer from significant irregular astig-
crease corneal thickness during surgery and increase the matism. In healthy corneas with adequate thickness, la-
risk for infection afterwards. By decreasing the exposure ser refractive surgery can be an option. However, KC
time of UVA delivered, the same total energy on the cor- is traditionally a contraindication for such procedures
nea can be maintained if the power is increased. Avedro because of the obvious dangers of exacerbating an al-
(Waltham, Massachusetts), a private medical device ready weakened cornea. Recently, the favorable effects
company, was the first to take advantage of this concept of CXL have allowed some ophthalmologists to recon-
in 2011 with their accelerated KXL system. The device sider limited refractive surgery on ectatic corneas. The
provided up to 30 mW/cm2 of power over 3 minutes of controversial concept was first introduced by Kanello-
exposure, providing the equivalent 5.4 J/cm2 of energy poulos, where he simultaneously combined CXL with
while drastically truncating surgery time. Theoretically, topographically guided photorefractive keratectomy
increasing the power of UVA to the cornea may increase (PRK) in what is now known as the Athens Protocol.
the risk for cytotoxic damage, and therefore more The procedure does not treat the whole refractive error

FocalPoints Module 11, 2015 9

 Table 3. Results of Clinical Studies on CXL in Combination With Other Modalities

AUTHORa MEAN FOLLOW-UP NO. OF EYES IMPROVEMENT

CXL COMBINED SIMULTANEOUSLY WITH PRK
Kanellopoulos (2009) 36 months 198 Kmax: 3.50 D
(simultaneous group) BCVA –0.28 logMAR, UCVA –0.66 LogMAR
SE: 3.20 D
Kymionis, et al (2009) 10.6 months 17 Kmax: 3.07D
BCVA –0.10 LogMAR, UCVA: –0.83 LogMAR
SE: 1.74 D
Stojanovic, et al (2010) 12 months 12 Astigmatism: 2.7D
BCVA: 20/57 to 20/35
UCVA: 20/1000 to 20/100
Kanellopoulos, et al 24 months 32 UCVA & BCVA: –2.25 LogMAR
(2011) Mean RE: 2.50D
CXL COMBINED WITH INTACS/ICRS
Kılıç, et al (2012) 7.07 months 131 Kmax: 4.47D
UCVA –0.26 LogMAR, BCVA: –0.24 LogMAR
Mean Sph Refraction: 2.62D
Mean Cyl: 1.62D
Legare, et al (2013) 9.8 months 27 BCVA –0.20 logMAR, UCVA –0.79 LogMAR
(ICRS + CXL group) Mean Sphere: 1.59D
Mean Cyl: 3.55D
Yeung, et al (2013) 12 months 38/47 (single/paired ICRS) Single ICRS:
UCVA: 3.4 Lines
Mean Cyl: 1.65D
Paired ICRS:
UCVA: 2.7 Lines
Mean Cyl: 0.95D
CXL Combined With PTK
Kymionis, et al (2012) 12 months 19 UCVA –0.36 LogMAR, BCVA: –0.12 LogMAR
Mean Cyl: 1.53D
Kapasi (2012) 12 months 17 Mean Sphere: 1.68D
Mean Cyl: 0.53D
Lines of Improvement: 0.33 lines
a
Refer to Suggested Reading for citations.

but rather, uses minimal laser ablation to resurface and
normalize the irregular cornea by removing no more
than 50 µm of tissue. The goal is to improve BCVA with
a flatter and more regular corneal surface while halt-
ing the ectasia with CXL. A number of case series have
repeated the Athens protocol with encouraging results
and demonstrated safety (see Table 3); however, the un-
orthodox procedure remains quite contentious amongst
leading experts. Larger, longer, prospective, randomized
studies establishing safety and efficacy are needed to
validate the results of these studies. For now, extra cau-
tion and explicit informed consent is warranted when
considering this procedure for patients.
CXL IN COMBINATION WITH PTK. In another tech-
nique variation, laser phototherapeutic keratectomy
(PTK) has been used to remove the epithelium prior
to CXL. By using the patient’s epithelium as a mask-
ing agent and with the understanding that the epithe-
lium is thinnest above the steepest point of the cornea,
the premise is to take advantage of the smoothing and
normalizing effect of PTK to achieve better visual out-
comes. An excimer laser is used to remove the anterior
50 µm of surface tissue, including the epithelium. At a
depth of 50  µm in a keratoconic eye, a small amount
of the stroma and anterior cone is ablated, effectively
flattening and levelling out the cone. Initial studies have

10 FocalPoints Module 11, 2015


Figure 10  Patient with intrastromal corneal ring
segment (Intacs, Addition Technology Inc, Lombard,
Illinois) inserted into a femtosecond laser-created
channel prior to same-day CXL treatment.
reported lowered mean refractive spherical equivalent
(MRSE) and better visual outcomes but longer term,
follow‑up studies with larger patient series are necessary
to properly evaluate outcomes of this combination.
CXL TO TREAT INFECTIOUS KERATITIS. As we enter
an era of antimicrobial resistance, there has been added
emphasis in finding agents that can eradicate microbes
with limited side effects. The known antimicrobial prop-
erties of activated riboflavin have been widely leveraged
in the past to disinfect blood products and inactivate
viruses. In vitro experiments have also demonstrated
the bactericidal effect of CXL. Subsequently, a num-
ber of reports have surfaced in which CXL is used to
treat multi-­ drug resistant infectious keratitis. Of the
published data, nearly all report a reduction in pain,
ulceration, and healing time in a wide array of bacte-
rial, fungal and Acanthamoeba organisms. Moreover,
the stiffening effect of CXL increases resistance to en-
zymatic degradation, which has been demonstrated to
halt corneal melt. Worth noting is that the majority of
these cases are severe, and many eventually require cor-
neal transplant from significant scarring. Though this
potential application is exciting, published literature
is mostly anecdotal and randomized clinical trials are
needed to define appropriate criteria. Until more data is
established, CXL for infectious keratitis should be ap-
proached with caution.
Future Directions
Corneal collagen crosslinking remains a heavily re-
searched field. One novel consideration involves using
other agents in lieu of the riboflavin/UVA combination,
essentially eliminating the risks that arise from exposure
altogether. Aliphatic beta nitro-­alcohols without UVA
have shown potential in crosslinking porcine corneas;
however, studies documenting efficacy and safety in vivo
remain to be seen. There is also a favorable trend toward
using CXL in nonectatic conditions such as for corneal
edema in cases of pseudophakic bullous keratopathy
and Fuchs endothelial dystrophy. It is hypothesized that
CXL increases intracellular attachments, thereby reduc-
ing space for fluid accumulation. Reported cases have
had mixed results, ranging from some with only tempo-
rary reduction in swelling to others with limited effect.
Studies with longer follow-­up and greater patient num-
bers are necessary to assess the efficacy and limitations
of CXL in the treatment of corneal edema.
Conclusion
Corneal collagen crosslinking has ushered fundamen-
tal changes in the treatment of keratoconus and kera-
tectatic conditions. CXL treatment can stabilize these
diseases, allowing patients to preserve or improve their
vision and possibly delay or obviate the need for cor-
neal transplantation. Long-term studies continue to be
carried out to confirm the lasting effects of CXL and
combined treatment modalities such as with PRK, PTK
or intrastromal corneal ring segments. Questions about
rates of re-treatment and definitions of progression in
patients treated for 10 years and beyond still need to be
addressed. Further modifications to the treatment pro-
tocol with epithelial-­on techniques and adjustments in
UVA light delivery will lead to increased patient com-
fort and decreased surgical risks associated with the
procedure.
Aaron Chan, OD, is an MD candidate at the University
of Toronto Medical School and is a Doctor of Optom-
etry at Downtown Eye Associates, Toronto, Canada.
Clara C. Chan MD, FRCSC, FACS, is an assistant pro-
fessor in the Department of Ophthalmology and Vision
Sciences, University of Toronto, Toronto, Canada.
FocalPoints Module 11, 2015 11
Clinicians’ Corner

Clinicians’Corner
Clinicians’ Corner provides additional viewpoints on the subject covered
in this issue of Focal Points. Consultants have been invited by the
Editorial Review Board to respond to questions posed by the Academy’s
Practicing Ophthalmologists Advisory Committee for Education. While
the advisory committee reviews the modules, consultants respond
without reading the module or one another’s response. –Ed.

1. Do you prefer epithelial-­on or epithelial-­off treat-
ment for corneal collagen crosslinking (CXL)?
>>>Dr. Güell: Today, despite the obvious advantages
of leaving the epithelium untouched (postoperative risk
and subjective complaints), I still prefer an epithelial-­
off procedure with the classical protocol in my stan-
dard CXL treatments. The available data of riboflavin
stromal distribution through an intact epithelium is still
quite debatable and inconsistent. On the other hand,
and together with other investigators, we are exploring
the epithelial‑on iontophoresis technique (particularly
in our older patients) because it looks like that, although
slightly reduced, the penetration of riboflavin through
this approach is quite homogeneous throughout the cor-
neal stroma.
>>>Dr. Stulting: We have been performing CXL for
7  years as investigators for 3  different clinical trials,
treating over 400 eyes using both epi‑off and epi‑on pro-
tocols. Although we have not had any cases of severe,
significant visual loss, we have seen sterile infiltrates
and subepithelial scarring due to persistent epithelial de-
fects with epi‑off treatments. Epi‑off protocols result in
greater discomfort, lengthier visual recovery, and more
frequent follow-­up examinations than epi‑on protocols.
On the other hand, epi‑on treatments probably do not
stiffen the cornea as much as epi‑off treatments do.
The key to obtaining good results with epi‑on treat-
ment, I believe, is to obtain good saturation by effective
enhancers in the riboflavin solution, mild mechanical
epithelial disruption during saturation, slit lamp verifi-
cation of adequate riboflavin penetration, not applying
riboflavin during light treatment, and greater fluence.
Nobody knows how much stiffening is “enough” for
any given patient, but I believe the increased safety of
epi‑on CXL more than compensates for any reduced
stiffening effect that occurs. Remember, it is always pos-
sible to offer epi‑off treatment if there is progression af-
ter epi‑on treatment.
2. Can patients with thinner corneas (<400 µm) re-
ceive CXL safely? How do you modify your tech-
nique for a thinner cornea?
>>>Dr. Güell: Again, available data on this item is
quite confusing. I still recommend CXL with the stan-
dard epithelium-­ off protocol in cases with central
pachymetry between 375–400 µm using hyposmotic ri-
boflavin preparations. We should take into account that
these unstable thin corneas would only have, as a sec-
ondary option, a keratoplasty approach. Thus, it makes
sense to test a not-so-­invasive procedure first.
>>>Dr. Stulting: For the past 1  1/2  years, we have
been performing epi‑on CXL under a protocol that al-
lows treatment of corneas as thin as 250 µm. We have
actually treated corneas with a thickness in the low
300 µm range without complications. During our early
years of performing CXL, we instilled artificial tears
and asked patients to sit with their eyes closed when the
corneal thickness was less than 400 µm after saturation
with riboflavin. This method causes rapid swelling of
the cornea so that treatment can be begun on a cornea
thicker than 400 µm typically in only a few minutes.
3. What age limits are recommended for CXL? Do
these limits change depending on the reason for
treatment (eg, keratoconus, postrefractive ectasia)?

12 FocalPoints Module 11, 2015

Clinicians’ Corner
>>>Dr. Güell: I do not consider any age limit nor dif-
ferences between primary or secondary ectasia for my
CXL indication. I only consider refractive-­topographic
instability or progression in primary and secondary
cases. Age might be an issue when recommending CXL
without demonstrating this instability or progression.
In these very young patients, I recommend CXL before
these signs are topographically obvious if I consider the
risk of progression to be high.
>>>Dr. Stulting: Having seen the results of CXL in
my own hands and reading the international literature,
I now believe that CXL is the procedure of choice for
keratoconus at the time of diagnosis for young patients,
regardless of patient age. Because CXL is a relatively
safe treatment, particularly with epithelium on, and
progression of keratoconus so likely, I do not believe
it is in the best interest of a young patient to wait for
demonstrated progression before offering CXL. On the
other end of the spectrum is the older patient, whom we
initially believed might not benefit from CXL because
of the natural stiffening of the cornea that occurs with
age. We have now seen not only stabilization of disease,
but improvement in vision in older patients. Based on
these observations and the safety of epi‑on CXL, I be-
lieve the upper age limit for CXL treatment should be
up to 60 years of age.
4. What major contraindications exist for CXL?
>>>Dr. Güell: From a clinical point of view and beside
the corneal thickness limitation, the only major contra-
indications are a history of recent active ocular surface
herpetic disease or allergy to riboflavin.
>>>Dr. Stulting: Patients with very thin corneas that
might receive endothelial damage from CXL and corneal
scarring that would prevent the patient from obtaining
acceptable vision are contraindications to CXL. I believe
that the original recommendation that CXL with the
Dresden protocol should not be performed on corneas
less than 400 µm in thickness may be overly conserva-
tive. It is certainly overly conservative for epi‑on treat-
ments as we are now performing. We are not certain,
however, what the minimum safe limit of corneal thick-
ness for epi‑on CXL might be. This question will only
be answered when we gain more experience with CXL.
5. What is your experience with CXL in treating
infectious keratitis? What other future indications
might apply for the CXL modality?
>>>Dr. Güell: My limited experience using CXL in
severe cases of infectious keratitis unresponsive to the
conventional topical treatment has been extremely sat-
isfactory. There were 3 cases of infectious keratitis, sev-
eral months apart, 2 caused by Pseudomonas and 1 by
Acanthamoeba. In 1 of the pseudomonas cases, amni-
otic membrane transplantation was applied immediately
after the CXL treatment. All cases responded very well,
stopping the active process in less than 1 week. Several
groups are investigating the effect of CXL early in the
management of infectious keratitis. Although the results
have not yet been published, preliminary results look
quite promising.
>>>Dr. Stulting: I have no experience in treating in-
fectious keratitis with CXL.
6. What is your experience with CXL in combina-
tion with photorefractive keratectomy (PRK) and
intrastromal corneal ring segments (ICRS)?
>>>Dr. Güell: I started relatively late with the combi-
nation of PRK and CXL because I needed to see long-
term results from other groups first. Finally, during
the last 3 years, I have been using such a combination
in 2  groups of patients with very good results: forme
fruste keratoconus with low ametropia (<3.00 D of SE)
in those patients seeking or refractive surgery and in
those progressive keratoconic eyes with low ametropia
(<3.00 D of SE). I recommend doing both procedures at
the same time, starting with the PRK, and my favorite
approach is the topography-­guided transepithelial tech-
nique for PRK, in order to precisely try to correct the
irregular component of the topography.
Regarding the combination with ICRS, I am recom-
mending CXL after ICRS implantation only in kerato-
conic eyes that show progression, accepting the practical
limitations in evaluating progression in these cases.
>>>Dr. Stulting: I have no experience with CXL in
combination with PRK and ICRS.
7. What is the re-treatment rate for CXL? Is the ef-
fect of the surgery stable?
>>>Dr. Güell: After 7  years of standard protocol
epithelium-­off CXL and with more than 300 eyes oper-
ated on, I have re‑treated only because of progression of
the ectasia in one case. In most series, the r­e‑treatment
rate is very low, although I presume that groups us-
ing transepithelial techniques and shorter treatment
FocalPoints Module 11, 2015 13
Clinicians’ Corner
protocols will probably observe a higher incidence when
using them in their young progressive keratoconic cases.
>>>Dr. Stulting: After performing CXL for 7  years,
we finally saw 1 patient who had progressed. We realize
our follow-­up time is not as long as it is internationally,
but we believe the need for re‑treatment will be minimal.
8. How do you define progression of keratoconus?
>>>Dr. Güell: There is not yet an accepted consensus
worldwide about the definition of progression in ker-
ataconus (Gomes et al, 2015). From a practical point
of view, I consider that an already diagnosed primary
or secondary ectasia is progressing if its mean K read-
ings increase >1.00  D, refractive astigmatism increases
>1.00  D, best spectacle-­corrected visual acuity dimin-
ishes >1  Snellen line while best-­corrected visual acuity
with rigid gas permeable contact lenses remains the same
and/or central corneal thickness diminishes >20 µm.
>>>Dr. Stulting: Progression is difficult to determine
because measurements of disease severity are not very
precise in eyes with ectatic corneal disease. Fortunately,
as our experience with CXL has increased and as we
have migrated from epi‑off treatment to epi‑on treat-
ment, proof of progression has become less important to
us in deciding whether to recommend treatment or not.
9. What are the alternatives to treat progressive
keratoconus, and when should crosslinking be
considered?
>>>Dr. Güell: There is today no other treatment but
CXL to treat progression in keratoconic eyes. It has
been described in several published series, including
ours, that ICRS might stabilize in a significant number
of cases of progression but they should not be used for
14 FocalPoints Module 11, 2015
this indication but only to improve irregularity indexes
and to correct low ametropia. For obvious reasons, pen-
etrating and lamellar keratoplasty techniques will sta-
bilize the situation for a number of years but the late
progression of the disease has been described after all
these keratoplasty techniques. Consequently, some of
us are considering the use of CXL several months after
keratoplasty in order to ensure a longer stability. Un-
fortunately, as with most approaches directed toward
long-term stabilization of the disease, only long-term,
well-­designed studies would provide the answer of our
questions.
>>>Dr. Stulting: Traditionally, we have treated kera-
toconus with glasses, rigid contact lenses, and corneal
transplantation, in that order, as the disease progresses.
None of these modalities addresses the underlying pa-
thology. They just aid in improving vision. CXL is unique
because it stiffens the cornea, halting progression of the
disease and affecting the underlying pathology. Other
modalities are now being used internationally in con-
junction with CXL to improve vision—such as ICRS,
thermokeratoplasty, and topography-­guided PRK. Based
on the extremely favorable risk-­benefit analysis of CXL,
I believe it should be considered as soon as the diagnosis
of ectatic corneal disease is made. Properly administered
at the time of diagnosis, along with early disease detec-
tion, CXL can virtually eliminate loss of vision from
ectatic corneal disease and other pathology. It has the
potential to reduce the number of corneal transplants
we perform in this country by up to 50%.
José L. Güell, MD, is director of the Cornea and Re-
fractive Surgery Unit at the “Instituto de Microcirugia
Ocular de Barcelona” and an associate professor of oph-
thalmology at the Autonomous University of Barcelona
in Barcelona, Spain.
R. Doyle Stulting, MD, PhD, is cofounder of the Stult-
ing Research Center at Woolfson Eye Institute and a
practicing ophthalmologist in Atlanta, Georgia.
SUGGESTED READING
Agrawal VB. Corneal collagen cross-­linking with riboflavin and ultra-
violet—a light for keratoconus: results in Indian eyes. Indian J Oph-
thalmol. 2009;57(2):111–4.
Bettis DI, Hsu M, Moshirfar M. Corneal collagen cross-­linking for
nonectatic disorders: a systematic review. J Refract Surg. 2012;28(11):
798–807.
Caporossi A, Baiocchi S, Mazzotta C, Traversi C, Caporossi T. Para-
surgical therapy for keratoconus by riboflavin-­ultraviolet type A rays
induced cross-­linking of corneal collagen: preliminary refractive re-
sults in an Italian study. J Cataract Refract Surg. 2006;32(5):837–45.
Chan CC, Squissato V. Keratconus and crosslinking: pharmacoki-
netic considerations. Expert Opin Drug Metab Toxicol. 2013;9(12):
1613–24.
Coskunseven E, Jankov MR 2nd, Hafezi F. Contralateral eye study of
corneal collagen cross-­linking with riboflavin and UVA irradiation in
patients with keratoconus. J Refract Surg. 2009;25(4):371–6.
Derakhshan A, Shandiz JH, Ahadi M, Daneshvar R, Esmaily H. Short-
term outcomes of collagen crosslinking for early keratoconus. J Oph-
thalmic Vis Res. 2011;6(3):155–9.
Filippello M, Stagni E, O’Brart D. Transepithelial corneal collagen cross-
linking: bilateral study. J Cataract Refract Surg. 2012;38(2):283–91.
Goldich Y, Marcovich AL, Barkana Y, et al. Clinical and corneal bio-
mechanical changes after collagen cross-­linking with riboflavin and
UV irradiation in patients with progressive keratoconus: results after
2 years of follow-­up. Cornea. 2012;31(6):609–14.
Gomes JA, Tan D, Rapuano CJ, et al. Group of Panelists for the Global
Delphi Panel of Keratoconus and Ectatic Diseases. Global consensus
on keratoconus and ectatic diseases. Cornea. 2015;34(4):359–69.
Hersh PS, Greenstein SA, Fry KL. Corneal collagen crosslinking for
keratoconus and corneal ectasia: One-year results. J Cataract Refract
Surg. 2011;37(1):149–60.
Hovakimyan M, Guthoff RF, Stachs O. Collagen cross-­linking: current
status and future directions. J Ophthalmol. 2012;2012:406850. Epub
2012 Jan 12.
Jankov MR 2nd, Hafezi F, Beko M, et al. Corneal cross-linking for
the treatment of keratoconus: preliminary results. Arq Bras Oftalmol.
2008;71(6):813–8.
Kanellopoulos AJ. Comparison of sequential vs same-day simultane-
ous collagen cross-­linking and topography-­guided PRK for treatment
of keratoconus. J Refract Surg. 2009;25(9):S812–8.
Kanellopoulos AJ, Binder PS. Management of corneal ectasia after
LASIK with combined, same-day, topography-guided partial transepi-
thelial PRK and collagen cross-linking: the athens protocol. J Refract
Surg. 2011;27(5):323–31.
Kapasi M, Baath J, Mintsioulis G, Jackson WB, Baig K. Phototherapeu-
tic keratectomy versus mechanical epithelial removal followed by cor-
neal collagen crosslinking for keratoconus. Can J Ophthalmol. 2012;
47(4):344–7.
Khan YA, Kashiwabuchi RT, Martins SA, et al. Riboflavin and ultra-
violet light a therapy as an adjuvant treatment for medically refrac-
tive Acanthamoeba keratitis: report of 3 cases. Ophthalmology. 2011;
118(2):324–331.
Kılıç A, Kamburoglu G, Akıncı A. Riboflavin injection into the cor-
neal channel for combined collagen crosslinking and intrastromal cor-
neal ring segment implantation. J Cataract Refract Surg. 2012;38(5):
878–83.
Koller T, Mrochen M, Seiler T. Complication and failure rates after
corneal crosslinking. J Cataract Refract Surg. 2009;35(8):1358–62.
Kontadakis GA, Kymionis GD, Kankariya VP, Pallikaris AI. Effect of
corneal collagen cross-­linking on corneal innervation, corneal sensitiv-
ity, and tear function of patients with keratoconus. Ophthalmology.
2013;120(5):917–22.
Koppen C, Wouters L, Mathysen D, Rozema J, Tassignon MJ. Refrac-
tive and topographic results of benzalkonium chloride-­assisted trans­
epithelial crosslinking. J Cataract Refract Surg. 2012;38(6):1000–5.
Kymionis GD, Grentzelos MA, Kounis GA, Diakonis VF, Limnopou-
lou AN, Panagopoulou SI. Combined transepithelial phototherapeutic
keratectomy and corneal collagen cross-­linking for progressive kera-
toconus. Ophthalmology. 2012;119(9):1777–84.
Kymionis GD, Kontadakis GA, Kounis GA, et al. Simultaneous
topography-­guided PRK followed by corneal collagen cross-­linking
for keratoconus. J Refract Surg. 2009;25(9):S807–11.
Leccisotti A, Islam T. Transepithelial corneal collagen cross-­linking in
keratoconus. J Refract Surg. 2010;26(12):942–8.
Legare ME, Iovieno A, Yeung SN, et al. Intacs with or without same-
day corneal collagen cross-linking to treat corneal ectasia. Can J Oph-
thalmol. 2013;48(3):173–8.
Lin DT, Holland S, Tan JC, Moloney G. Clinical results of topography-­
based customized ablations in highly aberrated eyes and keratoconus/
ectasia with cross-­linking. J Refract Surg. 2012;28(11 Suppl):S841–8.
Paik DC, Wen Q, Braunstein RE, Airiani S, Trokel SL. Initial stud-
ies using aliphatic beta-­nitro alcohols for therapeutic corneal cross-­
linking. Invest Ophthalmol Vis Sci. 2009;50(3):1098–105.
Raiskup F, Spoerl E. Corneal cross-­linking with hypo-­osmolar ribo-
flavin solution in thin keratoconic corneas. Am J Ophthalmol. 2011;
152(1):28–32.
Raiskup-Wolf F, Hoyer A, Spoerl E, Pillunat LE. Collagen crosslink-
ing with riboflavin and ultraviolet‑A light in keratoconus: long-term
results. J Cataract Refract Surg. 2008;34(5):796–801.
Rechichi M, Daya S, Scorcia V, Meduri A, Scorcia G. Epithelial-­
disruption collagen crosslinking for keratoconus: one-year results.
J Cataract Refract Surg. 2013;39(8):1171–8.
Sharma N, Roy S, Maharana PK, et al. Outcomes of corneal colla-
gen crosslinking in pseudophakic bullous keratopathy. Cornea. 2014;
33(3):243–6.
Stojanovic A, Zhang J, Chen X, Nitter TA, Chen S, Wang Q.
Topography-­guided transepithelial surface ablation followed by cor-
neal collagen cross-­linking performed in a single combined procedure
for the treatment of keratoconus and pellucid marginal degeneration.
J Refract Surg. 2010;26(2):145–52.
Vinciguerra P, Albé E, Frueh BE, Trazza S, Epstein D. Two-year corneal
cross-­linking results in patients younger than 18 years with documented
progressive keratoconus. Am J Ophthalmol. 2012;154(3):520–6.
Vinciguerra P, Albè E, Trazza S, Seiler T, Epstein D. Intraoperative and
postoperative effects of corneal collagen cross-­linking on progressive
keratoconus. Arch Ophthalmol. 2009;127(10):1258–65.
Vinciguerra R, Romano MR, Camesasca FI, et al. Corneal cross-­
linking as a treatment for keratoconus: four-year morphologic and
clinical outcomes with respect to patient age. Ophthalmology. 2013;
120(5):908–16.
FocalPoints Module 11, 2015 15
Viswanathan D, Males J. Prospective longitudinal study of corneal col-
lagen cross-­linking in progressive keratoconus. Clin Experiment Oph-
thalmol. 2013;41(6):531–6.
Wittig-Silva C, Whiting M, Lamoureux E, Lindsay RG, Sullivan LJ,
Snibson GR. A randomized controlled trial of corneal collagen cross-­
linking in progressive keratoconus: preliminary results. J Refract Surg.
2008;24(7):S720–5.
Wollensak G. Crosslinking treatment of progressive keratoconus: new
hope. Curr Opin Ophthalmol. 2006;17(4):356–60.
Wollensak G, Iomdina E. Biomechanical and histological changes after
corneal crosslinking with and without epithelial debridement. J Cata-
ract Refract Surg. 2009;35(3):540–546.
16 FocalPoints Module 11, 2015
Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-a-induced colla-
gen crosslinking for the treatment of keratoconus. Am J Ophthalmol.
2003;135(5):620–7.
Yeung SN, Ku JY, Lichtinger A, et al. Efficacy of single or paired in-
trastromal corneal ring segment implantation combined with collagen
crosslinking in keratoconus. J Cataract Refract Surg. 2013 Aug;39(8):
1146–51.
RELATED ACADEMY RESOURCES
Corneal Collagen Cross-­Linking. Learning Plan. Consolidates peer-­
reviewed, educational content available through the ONE Network.
CME. www.aao.org/learning-plan-detail/collagen-crosslinking. Ac-
cessed September 14, 2015.
029037D