Reshaping procedures for the surgical management of corneal ectasia Mohammed Ziaei, MBChB (Hons), FRCOphth, Allon Barsam, MD, FRCOphth, Neda Shamie, MD, David Vroman, MD, Terry Kim, MD, Eric D. Donnenfeld, MD, Edward J. Holland, MD, John Kanellopoulos, MD, Francis S. Mah, MD, J. Bradley Randleman, MD, Sheraz Daya, MD, FACP, FACS, Jose G€ uell, MD, PhD, for the ASCRS Cornea Clinical Committee Corneal ectasia is a progressive, degenerative, and noninflammatory thinning disorder of the cornea. Recently developed corneal reshaping techniques have expanded the treatment armamen- tarium available to the corneal specialist by offering effective nontransplant options. This review summarizes the current evidence base for corneal collagen crosslinking, topography-guided pho- torefractive keratectomy, and intrastromal corneal ring segment implantation for the treatment of corneal ectasia by analyzing the data published between the years 2000 and 2014. Financial Disclosure: No author has a financial or proprietary interest in any material or method mentioned. J Cataract Refract Surg 2015; 41:842–872 Q 2015 ASCRS and ESCRS Corneal ectasia is a progressive, noninflammatory thinning disorder of the cornea that compromises the stromal collagen matrix, resulting in protrusion and alteration of corneal shape. The types of ectasia include keratoconus, pellucid marginal degeneration (PMD), keratoglobus, posterior keratoconus, and post laser refractive–surgery ectasia. The incidence of keratoconus, the most common type of ectasia, is 1.3 to 25 per 100 000 1 with a reported prevalence of 50 to 230 cases per 100 000. 2 There is a strong association of ectasia with atopy, 3 and a growing body of evidence suggests genetic factors play an important role in its pathogenesis. 4 Although the total corneal collagen content of keratoconic eyes is normal, changes in epithelial basement membrane structure, stromal collagen fiber and extracellular matrix components, keratocyte morphology, and cell matrix interactions result in lamellar/fibrillar slip- page, leading to a reduction in corneal stiffness by a factor of 0.7. 5,6 This in turn leads to protrusion of the corneal apex and an increase in corneal power. 7 Pellucid marginal degeneration is a rare, idiopathic, progressive, nonhereditary, and noninflammatory ectatic disease classically affecting the inferior aspect of the cornea. 8 Histopathologic evaluation of PMD shows a thin or irregular Bowman layer, 9 while elec- tron microscopy studies reveal fibrous long collagen with a periodicity of 100 to 110 nm scattered among re- gions of otherwise normal collagen fibers. 10 Post laser refractive–surgery ectasia is an uncommon and complex entity that has been reported following both laser in situ keratomileusis (LASIK) 11 and photore- fractive keratectomy (PRK), 12,13 with a reported prevalence of 1/2500 14 and 1/3000, 15 respectively. Iat- rogenic ectasia has been associated with laser refractive Submitted: May 1, 2014. Final revision submitted: July 15, 2014. Accepted: July 23, 2014. From Moorfields Eye Hospital (Ziaei), Luton and Dunstable Univer- sity Hospital (Barsam), and Centre for Sight (Daya), East Grinstead, London, United Kingdom; Department of Ophthalmology (Kim), Duke University Eye Center, Durham, North Carolina, Carolina Cataract and Laser Center, Ladson, South Carolina (Vroman), Ophthalmic Consultants of Long Island (Donnenfeld), Rockville Centre, and Department of Ophthalmology (Kanellopoulos), New York University Medical School, New York, New York, Cincinnati Eye Institute (Holland), Cincinnati, Ohio, University of Pittsburgh School of Medicine (Mah), Pittsburgh, Pennsylvania, and Emory University School of Medicine (Randleman), Atlanta, Georgia; Laservision.gr Eye Institute (Kanellopoulos), Athens, Greece; Insti- tuto de Microcirug ıa Ocular (G € uell), Barcelona, Spain. Corresponding author: Mohammed Ziaei, MBChB (Hons), FRCOphth, Moorfields Eye Hospital, 162 City Road, London, United Kingdom. E-mail: [email protected]. Q 2015 ASCRS and ESCRS Published by Elsevier Inc. 842 http://dx.doi.org/10.1016/j.jcrs.2015.03.010 0886-3350 REVIEW/UPDATE
31
Embed
Reshaping procedures for the surgical management of ...sites.cardiff.ac.uk/ukcxl/files/2013/11/Reshaping-procedures-for... · Reshaping procedures for the surgical management of corneal
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
REVIEW/UPDATE
SubmittedFinal revisAccepted:
From Moosity HospiLondon,Duke UniCataractOphthalmCentre, anYork UnivEye InstitSchool ofUniversityLaservisiotuto de M
CorresponFRCOphthKingdom.
Q
Pub
842
Reshaping procedures for the surgicalmanagement of corneal ectasia
Mohammed Ziaei, MBChB (Hons), FRCOphth, Allon Barsam, MD, FRCOphth, Neda Shamie, MD,David Vroman, MD, Terry Kim, MD, Eric D. Donnenfeld, MD, Edward J. Holland, MD,
John Kanellopoulos, MD, Francis S. Mah, MD, J. Bradley Randleman, MD,Sheraz Daya, MD, FACP, FACS, Jose G€uell, MD, PhD, for the
Corneal ectasia is a progressive, degenerative, and noninflammatory thinning disorder of thecornea. Recently developed corneal reshaping techniques have expanded the treatment armamen-tarium available to the corneal specialist by offering effective nontransplant options. This reviewsummarizes the current evidence base for corneal collagen crosslinking, topography-guided pho-torefractive keratectomy, and intrastromal corneal ring segment implantation for the treatment ofcorneal ectasia by analyzing the data published between the years 2000 and 2014.
Financial Disclosure: No author has a financial or proprietary interest in any material or methodmentioned.
Corneal ectasia is a progressive, noninflammatorythinning disorder of the cornea that compromises thestromal collagen matrix, resulting in protrusion andalteration of corneal shape. The types of ectasia includekeratoconus, pellucid marginal degeneration (PMD),keratoglobus, posterior keratoconus, and post laserrefractive–surgery ectasia.
y 1, 2014.ubmitted: July 15, 2014.23, 2014.
s Eye Hospital (Ziaei), Luton and Dunstable Univer-arsam), and Centre for Sight (Daya), East Grinstead,d Kingdom; Department of Ophthalmology (Kim),ty Eye Center, Durham, North Carolina, CarolinaLaser Center, Ladson, South Carolina (Vroman),nsultants of Long Island (Donnenfeld), Rockvillepartment of Ophthalmology (Kanellopoulos), NewMedical School, New York, New York, Cincinnatiolland), Cincinnati, Ohio, University of Pittsburghicine (Mah), Pittsburgh, Pennsylvania, and Emoryool of Medicine (Randleman), Atlanta, Georgia;Eye Institute (Kanellopoulos), Athens, Greece; Insti-irug�ıa Ocular (G€uell), Barcelona, Spain.
author: Mohammed Ziaei, MBChB (Hons),orfields Eye Hospital, 162 City Road, London, Unitedail: [email protected].
SCRS and ESCRS
by Elsevier Inc.
The incidence of keratoconus, the most commontype of ectasia, is 1.3 to 25 per 100 0001 with a reportedprevalence of 50 to 230 cases per 100 000.2 There is astrong association of ectasia with atopy,3 and agrowing body of evidence suggests genetic factorsplay an important role in its pathogenesis.4 Althoughthe total corneal collagen content of keratoconic eyesis normal, changes in epithelial basement membranestructure, stromal collagen fiber and extracellularmatrix components, keratocyte morphology, and cellmatrix interactions result in lamellar/fibrillar slip-page, leading to a reduction in corneal stiffness by afactor of 0.7.5,6 This in turn leads to protrusion of thecorneal apex and an increase in corneal power.7
Pellucid marginal degeneration is a rare, idiopathic,progressive, nonhereditary, and noninflammatoryectatic disease classically affecting the inferior aspectof the cornea.8 Histopathologic evaluation of PMDshows a thin or irregular Bowman layer,9 while elec-tron microscopy studies reveal fibrous long collagenwith a periodicity of 100 to 110 nm scattered among re-gions of otherwise normal collagen fibers.10
Post laser refractive–surgery ectasia is an uncommonand complex entity that has been reported followingboth laser in situ keratomileusis (LASIK)11 and photore-fractive keratectomy (PRK),12,13 with a reportedprevalence of 1/250014 and 1/3000,15 respectively. Iat-rogenic ectasia has been associated with laser refractive
843REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
correction in preexisting subclinical (forme fruste) kera-toconus, excessive tissue ablation, enhancements, andattempted high myopic correction.14
Conservative management of keratectasia initiallyinvolves spectacle correction followed by contactlenses. Contact lenses represent the treatment of choicefor most keratoconic patients.16 However, surgicalintervention is required in patients who become con-tact lens intolerant or are not capable of visual rehabil-itation with contact lenses. For advanced disease withcentral corneal scarring, corneal transplantation,preferably with deep anterior lamellar keratoplasty(DALK) or alternatively penetrating keratoplasty, re-mains the only viable treatment option. Large prospec-tive cohort studies report a keratoplasty rate of 11.8%after an 8-year period of follow-up17 and 18.8% over a20-year period.18 However, the high financial burdenof corneal transplantation,19 limited patient satisfac-tion with postoperative vision, and contact lens intol-erance,20 coupled with an almost 50% failure rate atthe 20-year timepoint after keratoplasty,21 has led tothe development of many new nontransplant surgicaltreatment options to both stabilize the ectatic processand improve vision.
There has been particular interest in the newer treat-ment modalities as they offer the potential to treat thedisease at an earlier stage, prevent the morbidity asso-ciatedwith disease progression, and preclude the needfor the more invasive keratoplasty procedure. For thisreason, the Corneal Clinical Committee of the Amer-ican Society of Cataract and Refractive Surgerythought it important to review corneal collagen cross-linking (CXL) and intrastromal corneal ring segments(ICRS) as treatment modalities in patients withkeratectasia.
METHODS
This systematic review adopted a search strategy designedto identify all relevant published English articles using theOvid Medline and PubMed databases. A literature searchwas performed to identify studies of photochemical cornealCXL with riboflavin (vitamin B2) and ultraviolet-A (UVA)and ICRS for the management of corneal ectasia publishedbefore January 1, 2014.
CORNEAL COLLAGEN CROSSLINKING
Crosslinking can be defined as the creation of bondsthat connect 1 polymer chain to another. CornealCXL is a technique to strengthen corneal tissue. Ituses riboflavin as a photosensitizer and UVA to in-crease the formation of intrafibrillar and interfibrillarcarbonyl-based covalent bonds through the processof photopolymerization.22,23 Although the exact mo-lecular response to CXL is not fully understood, exper-imental data suggest that it occurs via 2 mechanisms.
J CATARACT REFRACT SURG
After the riboflavin molecule is excited to its singleor triplet state and in the early aerobic phase of CXL,photosensitized oxidation of stromal proteins occursvia interaction with reactive oxygen species (type IImechanism).24 After the oxygen supply is depletedand in anaerobic conditions, reactive species of radicalions are produced and interact with numerous mole-cules in the corneal stroma (type I mechanism).22 Theliberation of these molecules in the corneal stromacoupled with the resulting wound-healing responseto the procedure’s mechanical and photochemicalinsult is thought to induce an increase in corneal rigid-ity in the anterior 200 mm of the stroma,25–28 modulusof elasticity, collagen fiber thickness,29 resistance tostretching,28 and enzymatic degradation,30 with anassociated decrease in swelling and permeability.31–33
Although the major indication for CXL is to arrestthe progression of keratoconus,34 CXL has been usedin the treatment and prophylaxis35 of iatrogenic ecta-sia resulting from LASIK,35–37 PRK,36,38 and radial ker-atotomy.39 It has also been used in combination withother treatments, such as ICRS implantation40,41 andlimited topography-guided photoablation.42,43
As the CXL process leads to compaction of thecorneal stroma and reduces the potential space forfluid in edematous corneas,44,45 it has been used effec-tively in short-term palliative treatment of patientswith pseudophakic bullous keratopathy with a re-ported reduction in patient pain scores and centralcorneal thickness (CCT) but no significant improve-ment in visual acuity.46
Initial laboratory studies indicated potential antimi-crobial properties, and subsequently small case serieshave reported successful treatment of bacterial (Pseu-domonas aeruginosa, Staphylococcus aureus, Escherichiacoli), fungal (Fusarium, Aspergillus), protozoal (Acan-thamoeba), and atypical (nontuberculous mycobacte-ria) keratitis using CXL.47
Dresden Protocol
The standard protocol, or what is now referred to asthe Dresden protocol for CXL, as initially reported byWollensak et al.,34 requires epithelial removal, applica-tion of riboflavin 0.1% solution for 30 minutes beforeUV exposure, with a wavelength of 370 nm and homo-genous irradiance of 3 mW/cm2 for 30 minutes(5.4 J/cm2). Collateral endothelial, lenticular, or retinaldamage is not expected using this protocol.48
Alternative Treatment Protocols
More recent CXL techniques have departed from theoriginal Dresden protocol. These newer techniques usedifferent formulations and delivery methods of ribo-flavin as well as altered UV exposures to shorten
- VOL 41, APRIL 2015
844 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
duration times, reduce patient discomfort, and mini-mize postoperative complications.
High-Fluency Collagen Crosslinking The theoreticalfoundation of high-fluency CXL is the Bunsen-Roscoe law of photochemical reciprocity, which statesthat the same photochemical effect can be achievedwith a reduced irradiation interval provided the totalenergy level is kept constant through a correspondingincrease in irradiation intensity. Commercially avail-able ultrafast devices are now able to provide an irra-diance intensity of 43 mW/cm2. Using this setting, atotal treatment time of 2 minutes would achieve thestandard Dresden protocol energy dose of 3.4 J or aradiant exposure of 5.4 J/cm2.49
Ex vivo human cornea studies using scanningacoustic microscopes have failed to demonstrate a sig-nificant difference in corneal stiffness between corneastreated with low and high intensity.50 Early in vivosafety studies using high-fluency crosslinking haveshown the procedure to be safe with no adverse effectson epithelial healing times or corneal limbalmorphology.51 However, a recent human safety studysuggests transient but significant differences in endo-thelial cell density, percentages of hexagonality, andcoefficient of variation of the endothelial cell area inthe first month after treatment with high-fluencyCXL.52
Pulsed Collagen Crosslinking There are currently nopeer-reviewed publications on the use of this CXLmodality.
Transepithelial Collagen Crosslinking The diffusion pro-cess of riboflavin 0.1% in the stroma is a time-dependent 1-dimensional process that is limited bycorneal epithelial tight junctions.53,54 Epithelialdebridement is thought to be a major contributorto the postoperative complications of CXL, such asinfective keratitis and an abnormal wound-healingresponse. This issue has perpetuated interest in the po-tential application of an epithelium-on technique.55,56
The 2 main challenges associated with this techniqueare the limited diffusion of the riboflavin moleculewith a molecular weight 376 g/mol through the lipo-philic cornea and the molecule’s epithelial tightjunctions,57,58 as well as the high UV absorption coef-ficients of the corneal epithelium and Bowman layer.59
The diffusion of riboflavin across an intact cornealepithelium can be achieved through a number of tech-niques, such as modifying the corneal epithelialpermeability, changing the physicochemical proper-ties of the riboflavin molecule, and directly deliveringthe riboflavin molecule into the corneal stroma. Modi-fication of epithelial permeability can be an alternativeto epithelial removal. In this approach, chemical
J CATARACT REFRACT SURG
enhancers such as benzalkonium chloride (BAK), eth-ylenediaminetetraacetic acid (EDTA), trometamol, orgentamicin are added individually or concurrentlywith the riboflavin solution to loosen epithelial tightjunctions and facilitate the diffusion of riboflavin intothe stroma.31,55 A preparation of riboflavin 0.1% con-taining EDTA and trometamol is currently availablecommercially (Ricorlin TE).
Although human studies report a low complicationrate of 0% to 3.9% (transient haze) in patients treatedwith transepithelial CXL,60,61 the effectiveness of thetechnique remains uncertain and its use is still the sub-ject of debate and controversy. Experimental data haveshown that performing BAK 0.0005%-enhanced trans-epithelial CXL results in an increase in biomechanicalrigidity of only one fifth the level expected with stan-dard CXL in rabbit corneas (21.3% versus 102.4%).62
Other studies have shown the technique to affect thecollagen pattern profile and not the collagen fiberdiameter.63 Furthermore, limited keratocyte apoptosishas been demonstrated in transepithelial CXL,64 aswell as a more superficial CXL demarcation line at100 to 140 mm of depth when 1 enhancer is used64,65
and 250 mm when 2 enhancers are used.66 This is incontrast to standard CXL in which a demarcation lineis visible at a depth of 280 to 330 mm.67,68
To improve the diffusion of the negatively chargedwater-soluble riboflavin molecule through an intactepithelial layer, iontophoretic delivery using a mildelectrical current has been investigated. Studies showthat while this technique allows greater and deeperriboflavin penetration in the corneal stroma than theconventional epithelium-on technique, this concentra-tion is half that seen with the epithelium-off method.69
Confocal studies of eyes treated using this techniqueagain show a keratocyte apoptotic effect at a 60 to70 mm more superficial level than standard CXL, sug-gesting suboptimal penetration of the riboflavin intothe corneal stroma.70
In another attempt to avoid disturbing the cornealepithelial integrity and to hasten postoperative heal-ing, clinical investigators have used a grid-like patternof deepithelialization,71 riboflavin soaking of cornealpockets created for ICRS,72 and superficial intrastro-mal administration of riboflavin using a femtosecondlaser–created central corneal pocket.73,74
Early clinical and experimental data show that thegrid-like pattern of deepithelialization is ineffectivein increasing the absorption of riboflavin,71,75 whilethe femtosecond pocket procedure has shown somepromise, with 1 small human study demonstratingstabilization of the ectatic process as well as improvedcorrected distance visual acuity (CDVA) andmaximum keratometry (K) values over a follow-upof up to 26 months.73
- VOL 41, APRIL 2015
845REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
Corneal Crosslinking in Thin Corneas
Several alternative strategies have been used to pro-tect the intraocular structures from UVA irradiance inpatients with progressive ectasia who do not meet theminimum corneal thickness criteria for CXL treatment(440 mm including the epithelium). These includetransepithelial CXL, pachymetry-guided epithelialdebridement,76 use of hypoosmolar preparations ofriboflavin to produce corneal swelling,77 decreasingthe UVA irradiance dose, reducing the duration ofriboflavin soaking, increasing the riboflavin concen-tration, or a combination of the above.31 Althoughthese modifications may protect the endothelium inthin corneas, they are yet to be standardized and it re-mains unclear whether theywill yield results similar tothose of conventional CXL.
Typical Postoperative Findings
Anterior corneal stromal haze often occurs in thefirst month after treatment and typically resolves after12 to 20 weeks.56 The posterior aspect of this haze isdemarcated by an indistinct hyperreflective demarca-tion line seen in the midstroma. While no consensusabout its etiology has been reached, it is in all probabil-ity a clinical manifestation of the depth of CXL andoccurs as a result of a change in the stromal refractiveindex, an increase in collagen fiber diameter, andfibrillar spacing.29,78
Cellular and Ultrastructural Findings After CornealCollagen Crosslinking
Corneal CXL induces a wound-healing responsewith associated changes in the structure and cellularityof the cornea evident on confocal microscopy from theearly postoperative period up to 36 months after treat-ment.79 Confocal microscopy studies performedimmediately after CXL reveal stromal edema, thinningof the epithelial layer, keratocyte apoptosis, and areduction in nerve fiber density in the anterior stroma(250 to 300 mm).31 A human histopathologic study24 hours after CXL in a normal cornea confirmed com-plete loss of anterior stromal keratocytes to a depth of250 to 280 mm, whereas keratocyte densities werefound to be normal in the posterior aspect of thestroma.80 Three to 6 months after treatment, confocalmicroscopy shows a reduction in edema, increaseddensity of the stromal extracellular matrix, andgradual repopulation of the stromawith activated ker-atocytes from the peripheral cornea.81,82
At the 1-year timepoint, posterior stromal andendothelial cell morphology has been shown to beunchanged.83 Thirty-six months postoperatively, a“bridge-and-needle” shaped hyperreflective densityin the anterior extracellular corneal matrix can be
J CATARACT REFRACT SURG
visualized and this has been interpreted as newly re-placed collagen.67,84
In Messmer et al.’s histopathologic study of kerato-conic eyes,6 long-term changes in the anterior/middlecorneal stroma with central and peripheral keratocyteloss were observed up to 30 months after CXL. Whilethis study contradicts reported in vivo confocal studiesof human corneas and animal histological studies, it isa worrisome report and shows the possible long-termconsequences of this ever more popular procedure.
Clinical Outcomes
Keratoconus Conventional Crosslinking (Dresden Proto-col) Case Control StudiesMany studies have evaluatedthe role of CXL as a treatment modality for keratoco-nus since Wollensak et al.34 reported the firstcontrolled clinical study in 2003. In that ground-breaking study, CXL was performed in 23 eyes of 22patients with progressive keratoconus. At the lastfollow-up (range 3 months to 4 years), there was noobservable progression of keratoconus in any of thepatients. A reduction in maximum K was observedin 16 eyes (70%): a mean reduction of 2.01 diopters(D) accompanied by a 1.14 D reduction in refractive er-ror. A progression in the K readings (mean 1.48 D)occurred in 22% of untreated contralateral eyes.
Since the Wollensak et al. report, more than 20cohort studies of the effect of standard CXL usingdifferent outcome measures have been reported. Theresults of these investigations are summarized inTable 1.34,40,68,85–105
The majority of the studies are prospective longitu-dinal studies (84%), whereas a small proportion areretrospective (16%). Of the 24 prospective longitudinalstudies, 11 report the outcomes of the treated eye (usu-ally the worse eye) with those in the untreated felloweye serving as a control.
The studies report the effect of CXL in a patientgroup with a mean age between 16 years and 35 yearsand a follow-up duration between 3 months and60 months. Half the studies have a mean follow-upperiod of up to 1 year; only 3 studies followed their pa-tient cohort for more than 3 years.
The studies show an efficacy rate of 62% to 100% inhalting keratoconus progression as assessed bymaximum/mean K changes. All the studies (wheredata are available) report a reduction in keratometricmeasurements, while half the studies show the reduc-tion to be statistically significant. The degree of flat-tening ranges between 0.01 D and 1.0 D (43% ofstudies), 1.0 D and 2.0 D (33% of studies), and 2.0and 3.0 D (24% of studies). The flattening effect ofCXL appears to continue up to 6 years104 with no sig-nificant late complications.96
- VOL 41, APRIL 2015
Table 1. Cohort studies of CXL in progressive keratoconus.
(d) Z data not available;C Z deterioration in parameters; CCT Z central corneal thickness; CDVAZ corrected distance visual acuity;CXLZ crossinking; FUZ follow-up; KmaxZmaximumkeratometry; KmeanZmean keratometry; SEZ spherical equivalent; UDVAZuncorrected distance visual acuity*First author†Statistically significant resultzHypoosmolar solution used if !400
846 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
J CATARACT REFRACT SURG - VOL 41, APRIL 2015
Table 1. (Cont.)
Mean Change
CCT (mm) Permanent Complications Study Design CXL DeviceCDVA
0.07 logMAR O400z Loss of CDVA (7%) Retrospective pre–postlongitudinal cohort
847REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
J CATARACT REFRACT SURG - VOL 41, APRIL 2015
Table 2. Randomized controlled trials of CXL in progressive keratoconus.
Author* (Year) Eyes (n) Mean Age (Y) Mean FU (Mo)
Stabilization(Improvement) in Kmax/
Kmean (%)
Mean Change
Kmax/Kmean (D)
Wittig-Silva112 (2008) 66 26.9 12 100 (O50) 1.45†
Hersh113 (2011) 49 (d) 12 90 (51) 2.0†
O’Brart111 (2011) 24 29.6 18 100 (23) 0.62†
Wittig-Silva114 (2014) 46 25.6 36 98 (d) 1.03†
(d) Z data not available;C Z deterioration in parameters; CCT Z central corneal thickness; CDVAZ corrected distance visual acuity;CXLZ crossinking; FUZ follow-up; KmaxZmaximumkeratometry; KmeanZmean keratometry; SEZ spherical equivalent; UDVAZuncorrected distance visual acuity*First author†Statistically significant resultszHypoosmolar solution used if !400
848 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
The improvement in corneal shape is accompaniedby an improvement in uncorrected distance visualacuity (UDVA) of between 0.01 logMAR and 0.72logMAR and an improvement in CDVA of 0.01 to0.55 logMAR in all the studies (where data are avail-able). The improvement in visual acuity after treatmenthas been attributed to a decrease in astigmatism, areduction in corneal curvature, and topographicalhomogenization resulting from increased cornealrigidity. Coma-like aberrations, which are the mostprevalent higher-order aberrations (HOAs) in kerato-conic eyes,106,107 have also been shown to be signifi-cantly reduced following CXL.96,101,103,105
The efficacy of the standard CXL procedure is accom-panied by no significant changes in endothelial cellcounts (ECCs),92,94,96,105 intraocular pressure (IOP),96,105
lens density,98 or foveal thickness.92,94,98 Although mostof the studies report long-term stability of corneal thick-ness,40,91,92,100,105 a long-term prospective randomizedcontrol trial (RCT) has shown that CXL induces an initialthinning of the central cornea that is followed by agradual recovery toward baseline levels,108 with thechange in corneal thickness hypothesized to be due tochanges in the spacing between individual collagen fi-brils rather than to change in their diameter.109,110
It is worth noting that while the aforementionedstudies tend to have well-defined inclusion and exclu-sion criteria and trial designs, they are largely single-center, nonrandomized, and nonconsecutive. Assuch, they are prone to a number of error sourcessuch as selection and reporting bias. The studieshave also used different topography devices anddifferent measures, such as corneal surface curvatures,asymmetry indices, and higher-order corneal aberra-tions, to assess progression in ectatic eyes. Therefore,they represent low-quality evidence of the efficacy ofCXL in treating keratectasia.Randomized Controlled Trials To date, 4 RCTs have re-ported the clinical outcomes of conventional CXL
J CATARACT REFRACT SURG
in the treatment of keratoconus.111–114 These RCTsare summarized in Table 2. Although small in size,the studies are of moderate quality based on the Ja-dad scale. They report a 90% to 100% success rate inhalting the progression of the disease, and in all ofthem, significant corneal flattening was observed,with a reduction in the maximum K value between0.62 D and 2.0 D. Three of the 4 studies also showevidence of disease progression in the control groupas indicated by progressive steepening of thecornea.112–114
The impact of the CXL procedure on CDVA isencouraging, with all studies showing a statisticallysignificant improvement in CDVA (0.09 to 0.14 log-MAR) after treatment.111,113,114 Although there was atrend toward improved UDVA and reduced meanspherical equivalent (SE) after CXL in the 4 studies,the trend did not reach statistical significance in mostof the trials. O’Brart et al.111 also report an improve-ment in cone apex power and wavefront measure-ments (root mean square [RMS], coma, and pentafoil)after the CXL procedure. No permanent vision-threatening complications were reported in any trial.
It should be noted that these RCTs are small withrelatively short follow-up periods (12 to 36 months).It is also noteworthy that none had a sham controlledgroup of more than 3 months, nor were they double-blinded, and it is conceivable that the reportedchanges in corneal topography and visual acuitywere affected by this error source.
Transepithelial Collagen Crosslinking A number of smallstudies have evaluated the role of transepithelial CXLin keratoconic patients. Most of the studies have a pro-spective longitudinal cohort design with a follow-upperiod between 1 year and 2 years. The results in thesestudies are summarized in Table 3.60,61,65,66,70,115–119
The studies report the effect of transepithelial CXL ina patient group with a mean age between 14 years and
849REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
34 years and a follow-up period between 6months and24 months. Seven of the 9 studies (where dataare available) report an overall reduction in maximumK: between 0.01 D and 1.0 D (3 studies), 1.0 D and 2.0 D(1 study), and 2.0D and 3.0D (3 studies). This improve-ment in corneal shape is accompanied by an improve-ment inUDVAof 0.13 to 0.27 logMAR in all but 1 studyand an improvement in CDVA of 0.02 to 0.11 logMARin all the studies (where data are available).
Although transepithelial CXL appears to be a safe pro-cedure with no significant adverse effects and stablepostoperative ECCs,61,66,70,115,116,119 3 of the studiesreport significant corneal steepening despite treat-ment60,61,117 and in 1 study nearly one fifth of patientsrequired a repeat CXL procedure using theDresden pro-tocol.117 There are also inconsistent results in pachymet-ric measurements after transepithelial CXL, with 3studies reporting continued corneal thinning in treatedpatients,60,70,117 3 studies reporting stable CCTmeasure-ments,65,115,119 and 1 study reporting increasing thick-ness after treatment.118 The effect of treatment onHOAs is also debatable, with 2 studies reportingimproved RMS, coma, and spherical aberration mea-surements65,66 and 1 study reporting an increase inspherical aberration and no significant change in comameasurements after 2 years of follow-up.117 Further-more, the efficacy of transepithelial CXL in a pediatricpopulation is debatable as Caporossi et al.117 reported a50% retreatment rate in this subgroup due to demon-strable progression after 1 year. This is in contrast tothe only head-to-head study available in whichMagli et al.120 report that transepithelial CXL has safetyand efficacy indices similar to those of conventionalCXL in pediatric patients.
The diversity of surgical protocols and reported resultsmake overall assessment of transepithelial CXL as a pro-cedure to halt the progression of keratoconus problem-atic. While it seems likely that transepithelial CXL has aless tangible therapeutic effect than standard
J CATARACT REFRACT SURG
CXL,60,61,116,117 it could be beneficial to patients withthin corneas, uncooperative individuals, or those withquestionable progression, as it is expected to have a lowercomplication rate than standard CXL. A well-designedRCT comparing a standardized transepithelial CXL pro-cedure and conventional CXL is required to establish itsnoninferiority, confirm its superiority in complicationprofile, andultimately clarify theconsiderableuncertaintythat surrounds the use of transepithelial CXL today.
High-Fluency Collagen Crosslinking In a long-term ran-domized contralateral eye comparative study byKanellopoulos,121 CXL using 7 mW/cm2 irradiation for15 minutes had a similar efficacy to 30 minutes of stan-dard 3 mW/cm2 treatment. In both groups, postopera-tive clinical outcomes such as UDVA, CDVA,maximum K, and SE were similar after a follow-up of46months. If these results are replicated in other random-ized studies, it would be a major advantage to patientswho currently have to have a longer, more cumbersomeconventional CXL procedure.
Pellucid Marginal Degeneration
Two isolated reports demonstrate the efficacy ofCXL in improving the CDVA,122 keratometric astig-matism,122 and corneal biomechanical parameterssuch as corneal hysteresis (CH) and corneal resis-tance factor123 in patients with PMD. Topography-guided PRK combined with CXL has also been re-ported to result in improved visual, refractive, andtopographic indices 1 year after treatment.124,125
Although further studies are required to establishthe role of CXL in PMD, treatment appears to besafe with no disease-specific adverse outcomes re-ported to date.
Iatrogenic Ectasia
Collagen crosslinking has successfully treated iatro-genic ectasia after LASIK and PRK. In 2005, Kohlhaas
- VOL 41, APRIL 2015
Table 3. Cohort studies of transepithelial CXL in progressive keratoconus.
(d) Z data not available;C Z deterioration in parameters; CCTZ central corneal thickness; BAC Z benzalkonium chloride; CDVAZcorrected distance visual acuity; Trans CXL Z transepithelial crossinking; EDTA Z ethylenediaminetetraacetic acid; FU Z follow-up;Kmax Z maximum keratometry; Kmean Z mean keratometry; NA Z data not available; SE Z spherical equivalent; UDVA Zuncorrected distance visual acuity*First author†Statistically significant result
850 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
et al.126 reported the first successful treatment of a pa-tient with post-LASIK ectasia using CXL, with stabilityreported at 18 months. Later, Hafezi et al.127 reported aseries of 10 eyes with follow-up of up to 25 months inwhich there was a reduction in maximum K at 12months in all cases and a gain of more than 2 lines ofCDVAin40%of cases. These resultsmirror thoseofVin-ciguerra et al.,128 who reported the results of CXL in 13eyes of 9 consecutive patients with iatrogenic ectasia.The group reported a statistically significant reductionin mean SE as well as improved CDVA at the 6-monthtimepoint. A statistically insignificant improvement inHOAs (coma, spherical, and higher-order astigmatism)was also observed. Salgado et al.129 also reported stabil-ity of the ectatic process at 1 year in 22 eyes of 15patientswith post-LASIK ectasia treated with CXL.
In a prospective RCT performed by Hersh et al.,113 aless notable reduction in maximum K and associatedimprovement in CDVA was seen in iatrogenic ectasiacases than in keratoconic eyes. This phenomenon ispoorly understood as confocal microscopy studieshave shown comparable morphologic changes afterCXL in cases of iatrogenic ectasia and keratoconus.82
J CATARACT REFRACT SURG
Theoretically, CXL in post-LASIK eyes is likely to beless effective as CXL typically treats the anterior300 mm of the cornea, which includes the LASIK flap.The flap affords no biomechanical stability, leaving200 mm or less of treatment in the residual stromal bed.A reduced riboflavin diffusion in corneas that havehad LASIK, aswell as an intrinsic difference in the path-ophysiology of iatrogenic ectasia and keratoconus, mayalso account for the less pronounced CXL effect.130
The published reports, although few and relativelyshort in terms of follow-up, provide some evidencethat CXL may arrest or even improve iatrogenic ecta-sia, with only 1 reported case of keratectasia progres-sion after CXL.131
Combined Treatment
Collagen Crosslinking Combined with Topography-GuidedPhotorefractive Keratectomy Keratoconus There issome rationale for the use of PRK instead of LASIK inthinner corneas to reduce the risk for ectasia, especiallyas PRK results in less corneal biomechanical destabiliza-tion. Keratoconus with a preexisting structural
- VOL 41, APRIL 2015
Table 3. (Cont.)
Mean ChangePermanent
Complications Study Design CXL Device (Trans CXL Method)CDVA
851REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
weakness was traditionally considered to be a contrain-dication to any photoablative procedure. However,with the almost 3-fold increase in the corneal rigidityof patients treated with CXL,22 it is hypothesized thatthe crosslinked corneamay be able towithstand limitedtissue ablation without an adverse biomechanicalresponse. Various clinical investigators have demon-strated stability in corneas that have had a combinationof CXL and PRK, either sequentially or combined.
Kanellopoulos and Binder53 describe a case in whicha keratoconic patient who had CXL a year earlierwas treated with PRK. The cornea remained stablefor 1 year after mitomycin-C (MMC)-enhancedtopography-guided PRK. At the 18-month follow-up,the patient retained a CDVA of 20/20 with no clinicalor topographic evidence of progression. Speculatingthat increased rigidity would result in a different rateof ablation, the authors advocate undercorrecting thetreatment to 70% to 80% of the manifest refraction.Furthermore, the topographic ablation profile with acompensatory superior steepening reduced the needfor a large amount of tissue removal centrally, furtherimproving the safety of the procedure.
Favorable results have since been reported in ker-atoconic patients having simultaneous topography-
J CATARACT REFRACT SURG
guided PRK and CXL with an improvement in SE,defocus equivalent, UDVA, CDVA, and maximumK and no evidence of progression after a follow-up period of up to 12 months.125,132,133 It appearsthat visual results are superior following photother-apeutic keratectomy removal of the corneal epithe-lium compared with mechanical debridement,132
while a transepithelial nontopographic asphericastigmatic ablation profile is proposed to minimizetissue removal and obtain a corrective effect basedon the differential epithelial profile around thecone.133
In a prospective nonrandomized trial comparingthe effects of CXL alone and combined CXL andPRK, Alessio et al.134 report improved UDVA, SE,and corneal HOAs with the combined procedure,with no evidence of progression at 24 months.
The timing of the ablation treatment and CXL aswell as the interval between the 2 procedures hasbecome a topic of interest. Kanellopoulos and Binder53
compared sequential versus same-day PRK in kerato-conic patients having CXL. The sequential group hadtopography-guided PRK after aminimum observationperiod of 6months after the CXL procedure. The same-day group had a CXL procedure immediately after
- VOL 41, APRIL 2015
Table 4. Intrastromal corneal ring design specifications.
Parameter Intacs Intacs SK Ferrara Ring Keraring Myoring
(d) Z data not available; C Z deterioration in parameters; PMMA Z poly(methyl methacrylate)
852 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
MMC-enhanced topography-guided PRK. In bothgroups, the ablation was adjusted to correct only70% of the refractive error and the ablations wereless than 50 mm. The simultaneous group demon-strated statistically superior results in CDVA, SEreduction, maximum K reduction, and corneal hazescore at the final follow-up. No progression of ectasiawas reported in either group after a mean follow-up of36 months.
The simultaneous approach appears to be a superiorprocedure as the ablation occurs in “virgin” cornealtissue with well-established and clinically tested abla-tion nomograms. When a consecutive approach isused, unpredictable refractive outcomes may beachieved given that the ablation rate in crosslinkedcorneal tissue is currently uncertain. It is also counter-intuitive to ablate already strengthened corneal stro-mal tissue, and it may lead to a loss of efficacy. Afurther theoretical advantage of the simultaneousapproach is that the photopolymerization processwill lead to lower densities of keratocytes in the ante-rior 300 mm of stroma compared with the sequentialtechnique, which may reduce the risk for post-PRKhaze. However, there have been reports of stromalhaze after the simultaneous procedure, which maybe a result of keratocyte activation following the pho-toablative procedure.135
Iatrogenic Ectasia In the only study available to date,Kanellopoulos and Binder43 used MMC-enhancedtopography-guided transepithelial PRK followed by
J CATARACT REFRACT SURG
CXL with hypotonic riboflavin to treat patients withpost-LASIK ectasia. The authors used an effective op-tical zone of 5.5 mmwhile aiming to correct 70% of therefractive error to minimize tissue ablation. Theyreport topographic stability in all but 2 eyes and animprovement in UDVA in 27 of 32 cases.
There is no doubt that the combined surface ablationand CXL procedure has provided an exciting noninva-sive treatment option for keratoectasia as it has theability to offer patients functional vision and halt theprogression of the ectatic process. There is, however,a current a lack of RCT evidence to confirm its efficacyand U.S. Food and Drug Administration (FDA) trialdata are keenly anticipated.
Complications
Corneal CXL has become a ubiquitous treatment inearly keratoconus, with a growing perception of theprocedure as being routine. However, it is worthremembering that there are several complicationsassociated with this technique and no long-term safetyoutcomes are available to date.
Reported adverse effects of CXL can be broadly cate-gorized as treatment failures in which stabilization isnot achieved as well as infectious, inflammatory, andsuboptimal healing complications leading to loss ofCDVA. Failure of treatment is usually defined ascontinued progression of the disease with an increasein maximumK readings of 1.0 D over the preoperativevalue. Reports indicate that this occurs in 7.6% to 9.8%
- VOL 41, APRIL 2015
853REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
of patients in the first year following treat-ment.100,113,136 Topographic progression after 3 yearsof stability has also been reported, suggesting thatcorneal stromal remodeling may lead to loss of treat-ment effect in some individuals.137
Koller et al.100 evaluated 117 keratoconic eyeshaving conventional CXL and report a 7.6% treat-ment failure rate and a 2.9% rate of loss of 2 ormore Snellen lines of CDVA. In that study, a preop-erative maximum K value of 58.0 D or more was asignificant risk factor for failure, while patients olderthan 35 years with a preoperative CDVA of 20/25 orbetter were thought to carry a higher risk forcomplications.
Some of the most common complications seen afterthe CXL procedure include sterile infiltrates (7.6%)and central stromal scars (2.8%).100 Stromal haze forma-tion has also been described after CXL84,100 and isthought to be due to transient corneal fibroblast gener-ation.138 This phenomenon is most apparent at 1 monthpostoperatively, plateaus for 3 months, and usuallyfades 12 months after treatment.56 Permanent cornealhaze is reported to occur in 8.6% of treated eyes.139
Rarer adverse incidents following treatmenthave been reported. These include bacterial, protozoal,herpetic, and fungal keratitis140; corneal burns95;diffuse lamellar keratitis at the interface of a LASIKflap141; iritis; significantly elevated IOP142; and cornealmelting requiring deep lamellar keratoplasty.143
Also of concern are reports of serious complicationssuch as persistent corneal edema,144 endothelial dam-age,145,146 and conjunctival intraepithelial neoplasia147
in the keratocyte-voided bed of the recipient cornea ina patient who had DALK following CXL failure.
INTRASTROMAL CORNEAL RING SEGMENTS
Intrastromal corneal ring segments were developedand FDA-approved in 1999 for the treatment of lowmyopia, but they were quickly overshadowed by laserrefractive procedures. Interest in ICRS resurfacedwhenColin et al.148 reported their usefulness in treating mildto moderate keratoconus in 2000. The concept behindICRS insertion in keratectasia is not to eliminate the ec-tatic process, but to modify the corneal shape withoutremoving tissue or manipulating the central cornea.The goal of this treatment is to improve spectacle-corrected vision or contact lens tolerance, thus delayingor eliminating the need for a corneal transplant.149 In-trastromal corneal ring segments as a treatment optionoffer the added advantage that the rings are removableif not effective or poorly tolerated, with eyes returningto their preoperative refractive status within 3 monthsof explantation.150–152 Intrastromal corneal ring seg-ments appear to be highly biocompatible,153 with
J CATARACT REFRACT SURG
confocal studies showing normal central corneal im-ages in all layers with normal epithelial cells, subepithe-lial nerve plexus, keratocyte scattering, and endothelialmorphology in most individuals.154,155
Intrastromal corneal ring segments act as passivespacing elements that shorten the arc length of theanterior corneal surface and cause displacement ofthe collagen fibers, resulting in flattening of the corneain myopic eyes.A As a consequence of this effect, thecentral portion of the anterior corneal surface tendsto flatten and the peripheral area adjacent to the ringis displaced forward.156,157 The Barraquer thicknesslaw158 can roughly predict the induced change incorneal shape after additive device implantation,whereby the outcome achieved is directly propor-tional to the thickness of the ICRS and inversely pro-portional to its diameter.159–163
Although the mechanism of action of ICRS is welldocumented in myopic eyes, the structural alterationsresponsible for topographic homogenization of ectaticcorneas after ICRS implantation are not fully under-stood.164 The response of the ectatic cornea to pros-thesis implantation appears to be more complex, andthis variable response is hypothesized to be due tothe nonorthogonal lamellar architecture of these cor-neas.165,166 The refractive results achieved with ICRSimplantation in ectatic eyes are independent of patientage167 and are thought to be due to a reduction inHOAs,168–170 central and peripheral corneal flattening,anterior chamber depth shortening, and corneal apexdisplacement to a more physiologic location via areduction in paracentral ectasia.164 Intrastromalcorneal ring segments do not appear to alter cornealviscoelastic biomechanical properties.171,172
Design
Currently, 2 main types of ICRS devices are avail-able commercially. Incomplete ring segments,including Intacs (Addition Technology, Inc.) and Fer-rara ring (Mediphacos, Inc.), have been available forsome time, while the complete intrastromal Myoring(Dioptex GmbH) became available more recently.The Keraring (Mediphacos Ltda.) has been specificallydeveloped for keratoconus patients and shares thesame design, thickness, and composition as the Ferrararings but includes different arc lengths for bettermanagement of astigmatism. Intacs SK (severe kerato-conus or steep K) was recently introduced with signif-icant design modifications to better suit the higherrefractive errors associated with more advanced casesof keratectasia. The modifications include a smaller in-ner diameter of 6.0mm and an elliptical cross-sectionaldesign. These segments are available in 2 thicknessprofiles of 400 mm (steep K value of 57.0 to 62.0 D
- VOL 41, APRIL 2015
Table 5. Cohort studies of Intacs/Intacs SK implantation in keratoconic eyes.
Author*/Device (Year) Eyes (n) FU (Mo) Surgery Incision
(d)Z data not available;CZ deterioration in parameters; CDVAZ corrected distance visual acuity; FUZ follow-up; KZ keratometry;SE Z spherical equivalent*First author†Statistically significant result
854 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
J CATARACT REFRACT SURG - VOL 41, APRIL 2015
Table 5. (Cont.)
Mean Change% Gaining CDVA, Mean
Change Study Design ComplicationsCylinder (D) K (D)
(d)Z data not available;CZ deterioration in parameters; CDVAZ corrected distance visual acuity; FUZ follow-up; KZ keratometry;SE Z spherical equivalent*First author†Statistically significant result
856 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
and cylinder !5.0 D) and 450 mm (steep K O62.0 Dand cylinder O5.0 D).149 The smaller diameter of In-tacs SK and its associated elliptical shape may reducehalos and glare.173,174 The design features of these seg-ments are summarized in Table 4.
Implantation
Two surgical methods have been described forICRS implantation: mechanical and femtosecondlaser assisted. The mechanical procedure is done un-der sterile conditions. For the centration point, the
J CATARACT REFRACT SURG
surgeon has various options: the pupil center, thegeometric corneal center, or the visual axis. The deci-sion will depend on several factors such as the anglekappa and the surgeon’s criteria. This reference pointwill determine the final position of the segment, thelocation of the incision, as well as the center of intra-stromal dissection. After the axis of incision is defined(usually the steepest K reading), an intraoperative ul-trasonic pachymetric measurement is recommendedbefore the incisions are created. With a calibrated dia-mond knife, a 1.0 mm radial incision is made at 70%
- VOL 41, APRIL 2015
Table 6. (Cont.)
Mean Change
% Gaining CDVA, Mean Change Study Design ComplicationsCylinder (D) K (D)
857REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
of the pachymetric value. A blunt dissector is used tocreate the stromal tunnels after the application of avacuum trephine, with dissection done in a clockwiseand counterclockwise fashion. After the vacuumapplication is completed, the intrastromal segmentsare implanted, the incision site is closed with a single10-0 nylon suture, and a therapeutic contact lens isapplied.175–177
The other ICRS implantation method is the creationof a tunnel with the assistance of the femtosecondlaser. The patient is positioned under the laser, and
J CATARACT REFRACT SURG
the globe is fixated with a disposable suction device.The cornea is marked at the pupil center,173,178 and adisposable glass applanation cone affixed to the laseris lowered onto the eye to flatten the corneal surfaceand create a reference plane for the laser focus andapplication. The recommended settings to programthe laser are as follows: channel depth of 400 mm, innerchannel diameter of 6.6 mm, outer channel diameter of7.4 mm, and incision length of 1.4 mm. After the chan-nel is created, the suction device is removed and thesegments are inserted using a technique similar to
- VOL 41, APRIL 2015
Table 7. Cohort studies of Myoring implantation in keratoconic eyes.
Author* (Year) Eyes (n) FU (Mo) Surgical Procedure Incision Location
Mean Change
SE (D) Sphere (D)
Jabbarvand244 (2014) 42 12 FemtosecondGroup I, 250 mm depthGroup II, 300 mm depth
(d)Z data not available;CZ deterioration in parameters; CDVAZ corrected distance visual acuity; FUZ follow-up; KZ keratometry;SE Z spherical equivalent*First author†Statistically significant results
858 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
that in themechanical procedure. Incision closure aftersurgery is as described above.178
Although the femtosecond laser–assisted procedureis easier and faster170 and would theoretically create amore uniform and accurate stromal dissection plane,it has not resulted in superior visual/refractiveoutcomes compared with the manual tech-nique.170,179–183 In 1 study, however, the femtosecondlaser–assisted technique was superior in aberrometriccorrection.180
Nomograms
The planning for ICRS insertion is a crucial step indetermining the outcome of surgery. Some earlierstudies describe uniform placement of symmetricalor asymmetric ICRS for all patients,154 but since thenthere have been several nomograms that aim toimprove the predictability of surgical intervention.The various parameters used to improve predictabilityinclude preoperative SE,149,184,185 the location andmorphology of the cone,163 the magnitude of associ-ated asymmetric astigmatism,163,186,187 and preopera-tive CDVA levels.188,189 These parameters are used todetermine the thickness of the appropriate ringsegment for a given patient.
The decision for asymmetric placement of segmentsis made after evaluation of the topographic maps. Thethicker segment is usually placed to correspond to thesteeper half of the cone (predominantly inferiorlyplaced) to lift the cone and produce the maximum flat-tening effect. The thinner segment is placed in theopposite half of the cornea to counterbalance the thick-er segment and flatten the rest of the corneal sur-face.190–192
Even though the majority of proposed nomogramsare based on empirical data, the results obtained are
J CATARACT REFRACT SURG
often satisfactory. To date, there has been no directcomparison of the various nomograms demonstratingthe superiority of 1 particular technique. An inter-esting further development in this field would beincorporation of the biomechanical and aberrometriccharacteristics of the virgin corneal tissue in nomo-grams to further improve topographic predictabilityafter ring insertion.192
Single- Versus Double-Segment Implantation
While Sharma and Boxer Wachler193 reported supe-rior postoperative results and a more physiologiccorneal surface after single-ring implantation,others have found no significant difference betweensingle- and double-segment implantation.194 Largerprospective studies of this topic are anticipated.
Incision Location
Most surgeons make a temporal incision at or justsuperior to the horizontal meridian; others choosethe 12 o’clock or 1 o’clock position. Another option isto place the incision in the axis of coma, in the positivecylinder axis (if it is not 90 degrees from thetopographic axis), or in the steepest topographicmeridian.149,170,190,194 No comparative studies haveevaluated the role of incision location on clinicaloutcomes.
Tunnel Size
No significant differences in visual outcomes havebeen reported when the ICRS segment is implanted in2 different tunnel sizes of 6.7 mm � 8.2 mm (wide)and 6.6 mm � 7.6 mm (narrow). However, the narrowtunnel results in higher levels of complications after 6months of follow-up (42% versus 15%).195
- VOL 41, APRIL 2015
Table 7. (Cont.)
Mean Change
Study Design ComplicationsCylinder (D) K(D) CDVA
2.6 Group I†
2.53 Group II†5.43 Group I†
4.96 Group II†0.02 logMAR Group I0.12 logMAR Group II
859REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
Clinical Outcomes in Keratoconus
Intacs To date, more than 20 studies have evaluatedthe role of Intacs implantation in the management ofkeratoconic patients (Table 5).148,154,169,173,181,184,185,187,190,191,193,196–210 Most of these are prospective non-randomized studies with some reporting their resultsretrospectively.
The results show an impressive reduction in SE be-tween 1.3 D and 6.5 D; almost 70% of studies report areduction between 2.0 D and 4.0 D. Intacs implantationappears to effectively achieve corneal flattening,with allstudies (where data are available) reporting a reductionin maximum K readings between 1.5 D and 7.8 D, withmore than two thirds reporting a 3.0 to 5.0 D reductionin K readings. This improvement in corneal shape isassociated with an improvement in CDVA in 12% to86% of treated patients. However, the visual outcomesare variable, with the improvement in visual functionattributable to an improvement in corneal shape and areduction in HOAs after segment implantation.168–170
A growing body of evidence suggests that better vi-sual acuity outcomes are to be expected in patientswith less advanced keratoconus,186 with visual out-comes becoming more unpredictable in advancedcases.184,191 This issue is not universally accepted assome authors report an improvement in corneal shapeand visual function in advanced cases of ectasia191,196
and even in patients with corneal scarring.193
Restoration of contact lens tolerance is one of themain objectives of ICRS implantation and has beenachieved in 35% to 100% of treated patients.154,197 Ina cohort of 100 patients who had Intacs implantation,84% of the 44 patients requiring contact lens correctionafter surgery were able to tolerate their lenses at the24-month follow-up.198
Intacs insertion has been found to positively affectvision-specific patient satisfaction in up to 87% of
J CATARACT REFRACT SURG
treated patients,199,200 a phenomenon that graduallyincreases after the procedure191 and is maintained upto 5 years after treatment.154
The impact of Intacs insertion on the thinning cen-tral to the ectatic process has been evaluated in severalstudies looking at CCT measurements after ringsegment placement. Most studies show a nonsignifi-cant change in CCT up to 24 months aftersurgery.169,183,184,187,191,198,201 These findings coupledwith the fact that long-term studies with up to 9 yearsof follow-up demonstrate stability and safety (92.9% ofpatients being stable after ICRS implantation in 1study210) have led to suggestions that Intacs implanta-tion might be considered a therapeutic option to haltthe progression of the ectatic process.211 Supplemen-tary evidence in this domain is required, and furtherstudies are awaited.
Iatrogenic Ectasia Several studies have shown that In-tacs can be an effective option in secondary corneal ec-tasia, especially in patients with severe disease and inthose who have lost 2 or more lines of CDVA after thedevelopment of the ectatic process.212 According tothese studies, more than 70% of post-LASIK ectaticcases can expect an improvement in CDVA afterICRS implantation.194 Small case series show animprovement in visual acuity and topographic param-eters after Intacs177,179,210,213–217 and Intacs SK implan-tation174,218 with stability reported up to 5 years aftersurgery.154 Moreover, ICRS implantation has beenshown to defer the need for corneal transplantationand may even help halt disease progression.213
Pellucid Marginal Degeneration Intacs can also be effec-tive in reducing corneal astigmatism and 2nd-orderaberrations andHOAs219 in contact lens–intolerant pa-tients with early to moderate PMD.220–223 Several tech-niques have been proposed for treating such cases by
- VOL 41, APRIL 2015
860 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
placing the segments to bisect the steepest part of thecornea224 or implanting 1 segment above and 1 belowthe horizontal meridian.208 It is generally advisable toplace a thinner segment in the inferior cornea to avoidinadvertent corneal perforation.223
Ferrara and Kerarings In 1 of the first studies of the useof Ferrara rings in keratoconic patients, Siganoset al.164 reported a significant improvement inUDVA, CDVA, SE, and topography in all treated pa-tients with 6 months of follow-up. Since then, 17studies have reported the effects of these rings in ker-atoconic eyes (Table 6).163,164,182,201,225–238
Although most of the studies are retrospective, theydo demonstrate a significant reduction in SE of be-tween 0.06 D and 5.8 D, with nearly 80% (wheredata are available) showing a mean reduction ofmore than 2.0 D. This is accompanied by a reductionin keratometry between 2.0 D and 6.0 D in approxi-mately 80% of studies where data are available (range1.21 to 5.98 D). This improvement in corneal shape isaccompanied by an improvement in CDVA in all thestudies in which data are available, with 48% to 97%of treated patients experiencing an improvement in vi-sual function.
Overall, the studies suggest a good refractive and vi-sual outcome, even in cases that require adjustment orreinsertion after removal.239 The largest series to date,which includes more than 1000 consecutive cases, alsoconfirms a low complication rate of less than 4%.225
This low rate of ring-related complications is attribut-able to the “pachymetry law” in which the thickestportion of a pair of segments in the stromal bed cannotexceed half the thickness of the cornea. This large se-ries also demonstrates that the 210-degree ring seg-ments are more effective than their 160-degreecounterpart in reducing K readings and asphericity.225
There appears to be no difference in topographicchanges166 or final visual outcomes182,225,226 betweenthe mechanical and femtosecond methods of creatingthe intrastromal channels. Furthermore, a statisticallysignificant reduction in HOAs in patients with highcoma aberration values (RMS greater than 3 mm)has been reported.238 Keraring implantation hasbeen shown to lead to a significant reduction inmaximum elevation on both the anterior and poste-rior corneal surfaces,166 with no significant changesin mean CCT values postoperatively.225,227 Althoughthese studies support the claim of efficacy, revers-ibility, and adjustability of the Ferrara and Keraringsin keratoconus, no randomized studies are availableto date.
Pellucid Marginal Degeneration To date, 2 small retro-spective case series have evaluated the effect of Kerar-ing implantation in patients with PMD. They show
J CATARACT REFRACT SURG
an improvement in CDVA, SE, and keratometricpower219,234 aswell as a reduction in 2nd-order aberra-tions and HOAs.219
There are isolated case reports of successful visualoutcomes after ICRS implantation coupled withArtisan,240 Artiflex,240 or monofocal posterior cham-ber intraocular lens implantation.241
Iatrogenic Ectasia Regarding ectasia after refractivelaser surgery (mainly LASIK), Torquetti and Ferrara178
report statistically significant improvements in visualand refractive outcomes 6 months after manual im-plantation of a single Ferrara ring segment. They alsoreport a significant reduction in keratometric valuesand corneal asphericity.
The study by Pi~nero et al.242 shows that moreadvanced cases of iatrogenic ectasia with pronouncedconical protrusion are poor candidates for ICRS im-plantation. While the group reported a significantimprovement in visual acuity and an associated reduc-tion in coma-like aberrations after ring implantation,there was some regression of effect after 12 months,suggesting that the procedure did not stop the ectaticprocess. This finding is in contrast to findings in otherstudies that show a stable CCT after ringimplantation.243
Myoring
Since the first implantation of theMyoring in 2008, 5studies have evaluated its role in the management ofkeratoconic eyes. The results of these studies are sum-marized in Table 7.72,244–247
It is apparent that this device is effective in reducingrefractive error (SE reduction 5.7 to 7.3 D), flatteningthe cornea (maximum K reduction 4.9 to 8.0 D), andimproving CDVA (logMAR improvement ranging be-tween 0.02 and 0.35). The improvement in CDVA ap-pears to be more profound than the improvementwith incomplete ring segments. This phenomenon isthought to be because the complete continuous ringdesign may induce a more powerful arc-shorteningeffect for modifying the corneal curvature thanincomplete ring segment designs.245 The age of thepatient245 and the depth of ring implantation246 didnot significantly alter the clinical outcomes.
The studies also report a trend toward a reduction incoma,244,247 an increase in CCT,245,247 and stability inCH measurements after ring implantation.244,245,247
Although most of the reported studies are prospec-tive, they are small with 3 reports originating from asingle center. Further studies are therefore necessaryto confirm the efficacy and long-term stability of thisnew procedure.
- VOL 41, APRIL 2015
861REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
Pellucid Marginal Degenernation and Iatrogenic Ectasia Todate, no published studies have evaluated the role ofthe Myoring in the management of these 2 disorders.
Comparison of Intrastromal Corneal Ring SegmentOptions
While all ICRS studies have demonstrated goodvisual and refractive outcomes, the commerciallyavailable segments have fundamental differences indesign and various studies have evaluated whether aparticular design is more efficacious in treatingpatients with keratectasia.
Intacs Versus Keraring Two retrospective studies showthat while both these ICRS models are safe and effec-tive, the Keraring is superior in controlling astigma-tism and improving visual outcomes.183,248 This isprobably because the smaller optical zone of the Kerar-ing has a more profound effect on visual performance,while the Intacs segment has been shown to induce asignificant degree of negative primary spherical aber-ration.248 A slight nonsignificant regression of spher-ical correction has been reported with both ICRSdesigns.248
Intacs Versus Ferrara Kaya et al.249 report that bothICRS designs result in comparable refractive, topo-graphic, and optical quality outcomes, whereas eyesimplanted with the Ferrara segment experience agreater pupil-dependent reduction in scotopic contrastsensitivity.
Intacs SK Versus Keraring In the only retrospectivestudy available, Haddad et al.201 report a similar andsignificant improvement in refractive error, kerato-metric curvature, visual function, anterior and poste-rior corneal elevation, and coma values with bothring designs at the 1-year timepoint.
Although a smaller ring diameter is more effectivein controlling astigmatism, it may also lead to scat-tering of light rays that can adversely affect visualfunction. Therefore, it seems logical that an optimaldesign can achieve effective control of refractive errorwith minimal induction of adverse visual phenomena.However, no prospective data endorsing the superior-ity of 1 ring design are available to date. Prospectiverandomized studies with longer follow-up are needed.
Combined Treatments
The future of complete visual rehabilitation in ecta-sia may well lie in combined treatments dedicated toresolving both the structural and refractive issuesseen in this patient population. The most promisingoption in this arena is the combined use of ICRS,CXL, and advanced surface ablation techniques.
J CATARACT REFRACT SURG
Collagen Crosslinking and Intrastromal Corneal Ring Seg-ments Several studies show improved UDVA,CDVA, refraction, and keratometry 7 to 12 months af-ter combined treatment in keratoconic eyes.168,250–252
In a retrospective study,253 Vicente and Boxer Wachlerconclude that the patients with worse preoperativeCDVA and SE had a better postoperative gain inCDVA after simultaneous ICRS and transepithelialCXL. Although the synergistic effect of the 2 treat-ments has not been universally confirmed,Chan et al.252 report that the combined treatmentgroup showed a significantly greater reduction incylinder and maximum K than the group havingICRS implantation alone. Cakir et al.254 could notdemonstrate superiority of combined ICRS and CXLtreatment over ICRS treatment alone.
In a prospective randomized study, Coskunsevenet al.255 show that ICRS placement followed by CXLled to statistically superior keratometric, refractive,and visual outcomes than CXL followed by ring im-plantation. The authors also demonstrate the safetyof this combined approach, with no significant changereported in pachymetry or ECCs. Furthermore, therefractive changes after simultaneous treatmentappear to be reversible after ring explanation whilethe topographic outcomes are maintained.256
It is worthwhile to note that femtosecond laser po-wer must be modified in eyes that have had CXL tocompensate for the increase in corneal rigidity andreduction in corneal clarity.102
Collagen crosslinking can be performed before, dur-ing, or after ICRS implantation, but the ideal sequenceand timing of combined treatment is still uncertain.
Triple Procedure (Collagen Crosslinking, Intrastromal CornealRing Segments, and Photorefractive Keratectomy) Twosmall case series257,258 show the efficacy of simulta-neous Intacs implantation with same-day PRK andCXL, as well as sequential Intacs implantation followedbyCXL and PRK 6months later. After 6 to 12months offollow-up, significant improvements were noted inUDVA, CDVA, refraction, and topography in bothgroups, with no eyes losing CDVA. Triple therapyhas also been shown to result in a reduction in totalaberrations,257 as well as stable pachymetry, ECCs,and corneal viscoelastic properties.258
Sequential Intrastromal Corneal Ring Segment Im-plantation A new technique of double ICRS implanta-tion has been adopted by several research groups,with 2 reports indicating successful refractive out-comes after implantation of a Myoring259 and Kerar-ing260 in eyes with previous Intacs implantation.
- VOL 41, APRIL 2015
862 REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
Intrastromal Corneal Ring Segment Complications
Intacs-related intraoperative complications are veryrare and often occur as isolated events related to a sub-optimal surgical technique. Adverse events includesuperficial implantation of the segment, decentration,implant asymmetry,148,209 and perforation into theanterior chamber.149,261
Postoperative complications also appear to be infre-quent262 but include microbial keratitis (1.4% ofcases),263 corneal thinning in the area over thesegment, extrusion, reduced corneal sensation, deepneovascularization at the incision site,264,265 persistentepithelial defects, corneal haze around the segments,corneal melting, iritis, and uveitis.194 Other rare com-plications such as acute corneal hydrops,266 cornealedema,267 central stromal opacification,268 stromal ne-crosis,269 aberrant corneal nerve regeneration,270 andchronic pain secondary to direct contact between thesegment and corneal nerves271 have been reported.Furthermore, histopathologic changes have beendocumented after ICRS implantation, with 1 studyfinding a reversible induction in keratocyteapoptosis.272
Explantation of segments is required in approxi-mately 6.2% of cases; it is frequently due to segmentmigration.262 Adjustment of the ring segment may berequired in up to 10% of cases.273
A common occurrence after ring implantation is theformation of finewhite deposits within the intrastromalchannels and around the segments, a phenomenon thatis almost always visually insignificant.274–276 However,visually significant symptoms have been reported tooccur after ICRS implantation and include diplopia, ha-los, difficulty in night vision, vision fluctuation, oculardiscomfort, and photophobia.149
CONCLUSION
Keratectasia is a debilitating condition with adisproportionately high impact on public healthresources and vision-specific quality of life.277,278 Itseems likely that early management of keratectasiais useful in preventing disease progression, preser-ving visual function, as well as deferring, reducing,and even precluding the need for cornealtransplantation.
One of the most promising treatment modalities inthe field today is CXL, which has revolutionized themanagement of keratectasia in recent years by effec-tively stabilizing the underlying ectatic process andin some cases reversing the disease as quantified bykey topographic, refractive, and visual outcomes.The literature suggests that the results obtained aremore convincing for keratoconus and less remarkablein eyes with other forms of ectasia.113 This has not
J CATARACT REFRACT SURG
curbed the excitement regarding this new technology,with some predicting that it has the potential to pre-vent up to 50% of the corneal transplantations per-formed in the United States.279
The efficacy of standard CXL as a means of haltingthe ectatic process is backed by nearly 15 years oflaboratory and clinical studies. The procedureappears to be safe with a reported complicationrate of 1% to 3%,280 most of which is attributableto corneal epithelial debridement. This has led toimmense interest in the development of other inno-vative methods of treatment such as transepithelialcrosslinking. While the effects of transepithelialcrosslinking are still not fully understood, theyappear to be less pronounced than that of “epithe-lium-off” crosslinking. However the noninvasive na-ture of this technique makes it potentially useful incases in which epithelial debridement is ideallyavoided, such as in patients with other ocular sur-face pathology, uncooperative patients, and in pa-tients with very thin corneas.
Although there is no lack of clinical studies in thefield of CXL, to date only a small number of RCTsexist. It is apparent that robust evidence will onlybe forthcoming from well-designed multicenter, pro-spective, sham-controlled RCTs with standardizedcase descriptions, progression definitions, outcomemeasures, and sufficient follow-up to clarify the un-certainty that surrounds the long-term effects of pho-topolymerization. A search for trials on CXL on theU.S. National Institutes of Health Clinical Trials data-base shows the substantial interest in this domain,with multiple ongoing studies of the various formsof CXL and its synergistic use with other treatmentmodalities.
One such synergistic treatment option that can beused concurrently with CXL is ICRS implantation.Although the evidence base for ICRS implantationconsists solely of uncontrolled longitudinal follow-up studies, the introduction of this corneal reshapingtechnique for the correction of refractive errors in ec-tatic corneas has been a valuable and welcome eventas it is a simple outpatient-based procedure with ahigh technical success rate and a complication rateof approximately 4.6%.262 The procedure is alsoreversible and adjustable and in the event of failure,does not preclude subsequent corneal transplanta-tion. Furthermore, multiple short-term andmedium-term studies have shown statistically signif-icant improvements in corneal topography, refractiveerror, and visual acuity after ICRS implantation.While these ring devices are a useful alternative treat-ment, further basic research is required to fully un-derstand their impact on a cellular level, which willhelp determine the optimal size, shape, and design
- VOL 41, APRIL 2015
863REVIEW/UPDATE: RESHAPING PROCEDURES FOR KERATECTASIA
of segments and further refine existing surgicalnomograms.
So what does the future hold? It is evident thatfurther progress in the field of keratectasia manage-ment will occur only with a better understanding ofthe basic pathophysiology of the disease process andthe exact biomechanical response following CXL andICRS implantation. Corneal hysteresis, elastography,waveform analysis, noncontact tonometry, multi-photon fluorescence, photon microscopy, and inversecomputational analysis studies may well lead to newdiscoveries in this rapidly evolving field.
We envisage that the future norm of visual rehabil-itation in keratectasia will lie in the field of combinedtreatments aiming to correct both the biomechanicalinstability of the underlying disease process and theresultant refractive error. One such approach is theso-called triple procedure in which ICRS implantationis combined with sequential/simultaneous cornealCXL and corneal reshaping using excimer laser abla-tion. Although the triple procedure is promising,further detailed in vitro and in vivo studies arerequired to better define the potential impact of ICRSimplantation on riboflavin distribution, UVA irradi-ance, and the potential for an altered healing response.Further collaboration is also needed to adjust currentablation nomograms for combined treatments toreduce the potential for postoperative refractive sur-prises and the need for retreatment.
We foresee that CXLwill continue to play an increas-ingly prominent role in the management of the patientwith ectasia, as it has the capacity to prevent diseaseprogression and its associated visual morbidity. Aswe gain more experience with this new treatment mo-dality, the question of prophylactic treatment of all indi-viduals at the time of diagnosis will undoubtedly arise.The answer to this fundamental question will only bedefinitively answered through a better understandingof the procedure’s long-term safety and efficacy, notleast because our current, albeit limited, understandingof keratocyte and collagen turnover suggests that thestability gained after crosslinking is likely to be imper-manent if the renewed collagen does not exhibit thesame increased tensile and resistance properties.34 Iflong-term studies do show a regression of the crosslink-ing effect, further research will have to clearly defineand clarify the efficacy and safety profile of retreatment.
Another important gap in our current understandingwhich requires further investigation is whether the fullstiffening effects of the Dresden protocol is essential inhalting the ectatic process in all individuals or whetherthe stiffening effect seen after newer treatments such astransepithelial, high-fluency, or pulsed crosslinking issufficient to effectively prevent disease progression.
J CATARACT REFRACT SURG
Indeed, in this rapidly evolving field we may soon seemore widespread use of customized crosslinking pro-files while topical riboflavin treatment with exposureto natural sunlight may lead to the birth of a medicalmanagement option in the field of keratectasia.
Finally, if the promise of the antibacterial, antienzy-molytic, antiedematous,281 and neuroprotective prop-erties282 of CXL, as well as its potential to be used as anadjuvant treatment option in refractive surgery283 andprogressivemyopia,284 stand the test of time, photopo-lymerization will undoubtedly lead to a significantparadigm shift in the management of many otherocular pathologies.
SUMMARY
Collagen crosslinking and ICRS implantation aretechniques that can be used as standalone treat-ments or synergistically to strengthen and reshapethe ectatic cornea to improve visual outcomes andpossibly circumvent the need for more invasive sur-gical interventions such as corneal transplantation.Multiple studies report patient outcomes after thesereshaping procedures, but the optimal order, timing,and long-term effects of the interventions remainelusive.
Although reported adverse events are generally rareand often reversible, we encourage strict adherence tostandardized guidelines and protocols to ensure pa-tient safety. As always, the potential risks and benefitsof intervention must be clearly discussed with poten-tial patients, with the ophthalmologist acting as anenabler in the process of informed consent.
REFERENCES1. Georgiou T, Funnell CL, Cassels-Brown A, O’Conor R. Influ-
ence of ethnic origin on the incidence of keratoconus and asso-
ciated atopic disease in Asian and white patients. Eye 2004;
18:379–383. Available at: http://www.nature.com/eye/journal/
v18/n4/pdf/6700652a.pdf. Accessed December 15, 2014