From Stepping Hill Hospital, Stockport, United Kingdom (Tu); Emmetropia Mediterranean Eye Institute, Heraklion, Crete, Greece (Tu, Aslanides); and Weill-Cornell Medical College, New York, NY (Aslanides).
The authors have no financial or proprietary interest in the materials presented herein.
Presented at the 3rd International Congress of Corneal Cross Linking; December 7–8, 2007; Zurich, Switzerland.
Study concept and design (K.L.T., I.M.A.); data collection (K.L.T., I.M.A.); interpretation and analysis of data (K.L.T.); drafting of the manuscript (K.L.T.); critical revision of the manuscript (K.L.T., I.M.A.); statistical expertise (K.L.T.); administrative, technical, or material support (I.M.A.); supervision (I.M.A.)
Correspondence: Kyaw Lin Tu, FRCSE, Stepping Hill Hospital, Poplar Grove, Hazel Grove, Stockport, United Kingdom SK2 7JE. E-mail: email@example.com
Click here to read the Letter to the Editor.
Although corneal collagen cross-linking (CXL) appears to be a promising new technique to arrest progressive keratoconus and, in a majority of patients, to reverse it,1,2 a question remains: what exactly underlies the observed changes in the uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), and keratometric (K) values? Changes in UCVA have been ascribed to corneal flattening, and changes in BSCVA to the improved corneal symmetry, both produced by the stiffening and shrinkage of corneal collagen.3 A way in which to understand how CXL brings about those changes in individual corneas is needed so that the changes can be monitored and guidelines created for recommending and advising prospective patients about this treatment. This article presents the use of pre- and postoperative anterior elevation difference maps for such a purpose.
Patients and Methods
This retrospective study included 8 consecutive patients (14 eyes) who underwent CXL with riboflavin for keratoconus. Preoperative corneal topography was performed after patients had discontinued contact lens wear for at least 2 weeks. Patients who underwent other refractive procedures were excluded. Patients had discontinued contact lens wear for at least 1 week prior to topography at last postoperative follow-up. All eyes had a minimum preoperative central corneal thickness of 400 μm.
Pre- and postoperative evaluation included measurement of BSCVA, corneal topography using a videokeratoscope (Orbscan IIz; Bausch & Lomb, Rochester, NY), intraocular pressure by Goldmann applanation tonometry, and slit-lamp and fundus examination.
The CXL procedure was conducted under sterile conditions in the operating room. Proxymetacaine hydrochloride 0.5% eye drops were applied for preoperative local anesthesia. The central 8 mm of the corneal epithelium was removed using a blunt Beaver blade assisted by application of 20% ethanol for 30 seconds beforehand. As a photosensitizer, riboflavin 0.1% solution (10 mg riboflavin-5-phosphate in 10 mL dextran-T-500 20% solution) was applied every 3 minutes for 30 minutes at the end of which the patient was examined at the slit-lamp microscope to check for good riboflavin penetration. The patient’s cornea was then irradiated with ultraviolet A (UVA) -light diodes (IROC, Geneva, Switzerland) for 30 minutes using 3 mW/cm2 irradiance, which corresponds to a dose of 5.4 J/cm2. During this irradiation, riboflavin solution application every 5 minutes was continued. After the treatment, antibiotic eye drops (oflaxocin) were applied and a bandage contact lens placed until re-epithelialization was complete. Fluorometholone eye drops were administered four times a day for 1 month.
All patients were evaluated postoperatively on day 1, day 4 (day 5 if epithelial healing was incomplete the day before), and subsequent follow-up was scheduled at 1, 3, 6, and 9 months.
Pre- and postoperative anterior best fit sphere maps (the latter from examination at last follow-up) were used to obtain the anterior elevation difference maps for each patient. Patterns of change, in particular the relationship between the central and paracentral cornea, were visually identified. The two patterns identified were then related to the preoperative Orbscan eye metrics distances.
From each preoperative anterior elevation best fit sphere map, “Eye Metrics” was chosen from the “Tools” pull-down menu. “Options” was selected and the boxes for “Overlap map” and “Pupil” were checked. A line was drawn from the pupil center to maximum anterior elevation and the values of 1) total length of the line (mm) and distances between the two points, maximum anterior elevation and pupil center on the 2) x axis and 3) y axis were determined (Fig 1A). A similar line was also drawn from the topographic geometric center, defined as the center of the Placido rings on the same overlap map, to the maximum anterior elevation and distances measured in the same fashion (Fig 1B). The preoperative eye metrics differences between the patterns were compared, and 95% confidence interval analyses of the mean values were determined.
Figure 1. A) Maximum Anterior Elevation to Pupil Center Measurement and B) Maximum Anterior Elevation to Geometric Center Measurement on Orbscan Overlay Map.
Pre- and postoperative topographic maximum keratometry and irregularity indices, and clinical parameters (eg, spherical equivalent refraction and BSCVA) were also compared using the Student t test (significance level set at P≤.05).
Male to female ratio was 1:1. There were seven right eyes and seven left eyes. Mean age was 29.5 years (range: 16 to 49 years). Mean follow-up was 7 months (range: 5 to 10 months). Preoperative Krumeich stage was I in seven eyes, II in three eyes, III in one eye, and IV in three eyes. Of the four eyes with stage III and IV, three had high mean central keratometry values and one had a mild central scar.
Two main patterns of anterior elevation change were visually identified: 1) paracentral steepening, no change, or flattening centrally (Fig 2); and 2) paracentral flattening with central steepening (Fig 3).
Figure 2. Orbscan II Anterior Elevation (pre- to Postoperative) Difference Maps in Six Eyes Showing Pattern 1 Change—Paracentral Steepening and Central Flattening.
Figure 3. Orbscan II Anterior Elevation (pre- to Postoperative) Difference Maps in Eight Eyes Showing Pattern 2 Change—Central Steepening.
The preoperative maps of eyes that manifested pattern 1 (6/14 eyes) had shorter mean distances for maximum anterior elevation to pupil center (1.70 vs 2.27 mm) and maximum anterior elevation to geometric center (1.45 vs 1.99 mm) than those that resulted in pattern 2 (8/14 eyes). Mean maximum anterior elevation to pupil center distance on the x axis in pattern 1 eyes (0.67 mm [range: 0.47 to 0.83 mm]) was shorter than in pattern 2 eyes (mean distance 1.10 mm [range: 0.67 to 1.66 mm]). The same holds true for mean maximum anterior elevation to geometric center distance on the x axis (0.74 mm [range: 0.52 to 0.99 mm] vs 1.17 mm [range: 0.66 to 1.84 mm]). The corresponding mean distances on the y axis were also shorter in the eyes ending up in the first pattern although the differences were smaller (maximum anterior elevation to pupil center distance 1.53 vs 1.88 mm and maximum anterior elevation to geometric center distance 1.19 vs 1.41 mm). Preoperative topographic distances are compared in Table 1.
Table 1: Comparison of Preoperative Topographic Distances Measured Using Orbscan II in Eyes that Underwent CXL for Treatment of Keratoconus
Mean reduction in maximum topographic simulated keratometry following CXL in pattern 1 eyes was 1.37 diopters (D) (P=.004); the value actually showed an increase of 0.64 D in pattern 2 eyes. In pattern 1 eyes, mean irregularity indices at 3 mm (P=.03) and 5 mm (P=.04) decreased postoperatively but were increased in pattern 2 eyes. Mean BCSVA (decimal) improved slightly for both patterns postoperatively (in pattern 1 eyes by 0.039, in pattern 2 eyes by 0.05). Mean spherical equivalent refractive error decreased (less myopia) by 0.44 D in pattern 1 eyes and increased (more myopia) in pattern 2 eyes by 1.83 D (Table 2).
Table 2: Comparison of Pre- and Postoperative Parameters in Eyes that Underwent CXL for Treatment of Keratoconus
Three hundred to five hundred corneal lamellae extend from limbus to limbus. The lamellae are more closely packed in the center,4 and in the central 7-mm zone, a preferred orientation of the lamellae (50%5 to 66%6) in inferior-superior and nasal-temporal directions has been demonstrated. Significant differences have also been measured in some eyes between the two preferential fibril populations, with some corneas exhibiting as much as 25% more collagen in one direction than the other.7 Apart from this normal and variant corneal lamellar anisotropy, it has been shown that compared to normal corneas, the gross organization of the stromal lamellae was dramatically changed, and the collagen fibrillar mass was unevenly distributed, particularly around the presumed apex of the cone in keratoconic corneas.8 It would be reasonable to assume that this anisotropy in corneal lamellae distribution would affect corneal response to CXL.
Various parameters have been used to monitor and quantify these changes in CXL-treated patient groups (mean keratometry readings, asymmetry indices, hemimeridian differences, and anterior surface elevation,9,10 among others). However, it remains unclear as to how CXL works in individual eyes, leading to a need to monitor the changes and create guidelines for recommending and advising prospective patients about this treatment.
We believe examining anterior surface elevation difference maps of individual patients/eyes on Orbscan can help explain the mechanism for the observed changes.
The procedure as it is currently performed is aimed at the central 7 to 8 mm cornea, so it would be reasonable to look for the anterior elevation changes in the central and paracentral cornea. The anterior elevation maps were used because CXL has been found to affect mainly the anterior half of porcine corneal stroma when tested by hydration behavior.11 Treatment of the cornea with riboflavin and UVA significantly stiffened the cornea only in the anterior 200 μm. This depth-dependent stiffening effect may be explained by the absorption behavior for UVA in the riboflavin-treated cornea. Sixty-five to 70% of UVA irradiation was absorbed within the anterior 200 μm and only 20% in the next 200 μm.12
Yoshida et al13 have shown that after excimer keratorefractive surgery, the anteroposterior movement of the cornea cannot be assessed on a single color-coded postoperative elevation map because the radius of curvature of the reference plane is changed by the surgery. Instead the elevation changes should be evaluated based on the difference map generated from pre- and postoperative maps. Similar consideration led us to use the difference maps in our study.
Two main patterns of difference maps were identified. Pattern 1 (see Fig 2) revealed paracentral steepening with or without central flattening leading to reductions in spherical equivalent refractive error, central (simulated) keratometric values, and irregularity indices at 3 and 5 mm. This pattern is associated with a more central preoperative cone apex position, especially on the x axis. The changes are roughly analogous to those seen after myopic refractive photoablation where structural disruption of the central corneal lamellae occurs.14 In keratoconic eyes, collagen fibril slippage and thinning near the cone apex6 and stiffening (and shortening)15 of the surrounding collagen fibrils, of roughly equal length in all corresponding hemimeridians around the pericentral cone, would result in paracentral steepening and central flattening (Fig 4).
Figure 4. Stiffening (and Shortening) of the Surrounding Collagen Fibrils, of Roughly Equal Length in All Corresponding Hemimeridians Around the Pericentral Cone Would Result in Paracentral Steepening and Central Flattening.
Pattern 2 (see Fig 3) had peripheral flattening with central steepening. This pattern is associated with a more peripheral preoperative cone apex mainly on the x axis. Corneal lamellae on one side of a meridian to the displaced cone apex have more length to shrink than fibrils on the opposite side. Collagen cross-linking treatment, given that the cornea is anchored at the limbus, may induce a relative change in elevation between the corneal center and cone apex (from an effect not unlike water particles comprising a wave moving up and down in the same vertical plane) leading to an apparent pulling up of the cone apex to the center (Fig 5). This relative central steepening may increase the myopic refractive error and could increase both the central keratometry values and irregularity indices. This could explain some of the cases that appear to have increased myopia and/or reduced BSCVA after CXL, a finding that was noted previously.1
Figure 5. Corneal Lamellae on One Side of a Meridian to the Displaced Cone Apex Have More Length to Shrink than Fibrils on the Opposite Side. Following CXL Treatment, a Relative Change in Elevation Between the Corneal Center and the Cone Apex Would Lead to an Apparent Pulling up of the Cone Apex to the Center.
Shorter preoperative distances on the x axis, both for maximum anterior elevation to pupil center and maximum anterior elevation to geometric center, favored pattern 1, but for the y-axis distances, the mean differences between the two patterns were smaller. It may mean that the displacement of the cone inferiorly is less predictive of what pattern of anterior elevation change a particular keratoconic cornea is going to manifest after CXL. One can speculate whether inferior displacement of the cone is exaggerated, possibly influenced by gravity, and does not actually represent the true condition of the corneal lamellae, at least in some corneas.
As always, there are issues with data acquisition by the instrument used, in this case Orbscan II, including poor fixation (especially in eyes with poor visual acuity), involuntary eye movements, and misalignment by the technician during the acquisition process. All of our Orbscan measurements were performed by an experienced technician using an acquisition protocol recommended by the manufacturer. In a study on misalignment during Orbscan topography, rotational misalignment of the eye during acquisition influenced topography more than the translational misalignment during focusing by the technician.16 Avoiding or minimizing the effects of both types of misalignment is critical if valid conclusions are to be drawn.
In a confocal microscopic study of photorefractive keratectomy patients with 6-mm optical zones, the corneal epithelium from initial thinning (50±8 mm at 3 months) is restored to 52±6 mm at 12 months, compared with 51±4 mm before surgery. The authors concluded that the epithelium appeared to restore fully its preoperative thickness without transient or persistent hyperplasia. We believe that with a treatment zone of 8 mm, and without any stromal ablation, our cohort of eyes is also unlikely to exhibit significant epithelial thickness variability that would affect anterior corneal elevation.17
We used two reference points on the preoperative Orbscan maps to measure distances from maximum anterior elevation. The pupil center is an Orbscancalculated center of the pupillary area. Its accuracy is dependent on adequate palpebral opening. Pupil center would also be affected especially by the rotational misalignments in the vertical plane16 given the common inferior displacement of their cones. The other reference point used was the topographic geometric center, a point determination, with its advantage of allowing direct visual examination of the anterior elevation map from the standard Orbscan display.
Areas of weakness are noted in this study starting with its small sample size. Although the use of Orbscan II anterior elevation maps to evaluate the changes that occur after CXL has not been attempted before and therefore is unique, there is no evidence to date that this information is a valid measurement of the true corneal remodeling taking place after treatment. As noted, there are also potential limitations of this technique based on challenges in various data collecting aspects. Longer follow-up would demonstrate whether the changes noted persist or evolve further with time.
Corneal shape change influenced by anisotropy of corneal collagen distribution is a factor in the outcome of CXL treatment for keratoconus. Other factors are undoubtedly involved as well. The methodology developed in this pilot study provides a tool to help analyze corneal changes taking place following CXL treatment, which could lead to more tailored, customized treatment based on individual corneal characteristics mainly for, but not limited to, keratoconic corneas.
- Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-A-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135:620–627. doi:10.1016/S0002-9394(02)02220-1 [CrossRef]
- Raiskup-Wolf F, Hoyer A, Spoerl E, Pillunat LE. Collagen cross-linking with riboflavin and UVA-light in keratoconus: long-term results. J Cataract Refract Surg. 2008;34:796–801. doi:10.1016/j.jcrs.2007.12.039 [CrossRef]
- Kymionis G, Portaliou D. Corneal crosslinking with riboflavin and ultraviolet-A for the treatment of keratoconus. J Cataract Refract Surg. 2007;33:1143–1144. doi:10.1016/j.jcrs.2007.03.056 [CrossRef]
- Boote C, Dennis S, Newton RH, Puri H, Meek KM. Collagen fibrils appear more closely packed in the prepupillary cornea: optical and biomechanical implications. Invest Ophthalmol Vis Sci. 2003;44:2941–2948. doi:10.1167/iovs.03-0131 [CrossRef]
- Meek KM, Newton RH. Organization of collagen fibrils in the cornea stroma in relation to mechanical properties and surgical practice. J Refract Surg. 1999;15:695–699.
- Daxer A, Fratzl P. Collagen fibril orientation in the human corneal stroma and its implication in keratoconus. Invest Ophthalmol Vis Sci. 1997;38:121–129.
- Boote C, Dennis S, Huang Y, Quantock AJ, Meek KM. Lamellar orientation in human cornea in relation to mechanical properties. J Struct Biol. 2005;149:1–6. doi:10.1016/j.jsb.2004.08.009 [CrossRef]
- Meek KM, Tuft SJ, Huang Y, Gill PS, Hayes S, Newton RH, Bron AJ. Changes in collagen orientation and distribution in keratoconus corneas. Invest Ophthalmol Vis Sci. 2005;46:1948–1956. doi:10.1167/iovs.04-1253 [CrossRef]
- Caporossi A, Baiocchi S, Mazzota C, Traversi C, Caporossi T. Parasurgical therapy for keratoconus by riboflavin-ultraviolet type A rays induced cross-linking of corneal collagen: preliminary refractive results in an Italian study. J Cataract Refract Surg. 2006;32:837–845. doi:10.1016/j.jcrs.2006.01.091 [CrossRef]
- Chan CC, Sharma M, Wachler BS. Effect of inferior-segment Intacs with and without C3-R on keratoconus. J Cataract Refract Surg. 2007;33:75–80. doi:10.1016/j.jcrs.2006.09.012 [CrossRef]
- Wollensak G, Aurich H, Pham DT, Wirbelauer C. Hydration behavior of porcine cornea crosslinked with riboflavin and ultraviolet A. J Cataract Refract Surg. 2007;33:516–521. doi:10.1016/j.jcrs.2006.11.015 [CrossRef]
- Kohlhaas M, Spoerl E, Schilde T, Unger G, Wittig C, Pillunat LE. Biomechanical evidence of the distribution of cross-links in corneas treated with riboflavin and ultraviolet A light. J Cataract Refract Surg. 2006;32:279–283. doi:10.1016/j.jcrs.2005.12.092 [CrossRef]
- Yoshida T, Miyata K, Tokunaga T, Tanabe T, Oshika T. Difference map or single elevation map in the evaluation of corneal forward shift after LASIK. Ophthalmology. 2003;110:1926–1930. doi:10.1016/S0161-6420(03)00621-3 [CrossRef]
- Roberts C. Biomechanics of the cornea and wavefront-guided laser refractive surgery. J Refract Surg. 2002;18:S589–S592.
- Wollensak G, Iomdina E, Dittert DD, Herbst H. Wound healing in the rabbit cornea after corneal collagen cross-linking with riboflavin and UVA. Cornea. 2007;26:600–605.
- Hick S, Laliberté JF, Meunier J, Chagnon M, Brunette I. Effects of misalignment during corneal topography. J Cataract Refract Surg. 2007;33:1522–1529. doi:10.1016/j.jcrs.2007.05.029 [CrossRef]
- Möller-Pedersen T, Cavanagh HD, Petroll WM, Jester JV. Stromal wound healing explains refractive instability and haze development after photorefractive keratectomy: a 1-year confocal microscopic study. Ophthalmology. 2000;107:1235–1245. doi:10.1016/S0161-6420(00)00142-1 [CrossRef]
Comparison of Preoperative Topographic Distances Measured Using Orbscan II in Eyes that Underwent CXL for Treatment of Keratoconus
||Pattern 1 Eyes (n=6)
|Pattern 2 Eyes (n=8)
|Mean (Range) (mm)
||Mean (Range) (mm)
|Mean MAE to PC
||1.70 (0.98 to 2.32)
||2.27 (0.52 to 3.56)
|Mean MAE to PC on X axis
||0.67 (0.47 to 0.83)
||1.10 (0.67 to 1.66)
|Mean MAE to PC on Y axis
||1.53 (0.67 to 2.23)
||1.88 (0.21 to 3.46)
|Mean MAE to GC
||1.45 (1.16 to 1.84)
||1.99 (0.69 to 2.61)
|Mean MAE to GC on X axis
||0.74 (0.52 to 0.99)
||1.17 (0.66 to 1.84)
|Mean MAE to GC on Y axis
||1.19 (0.68 to 1.68)
||1.41 (0.09 to 2.52)
Comparison of Pre- and Postoperative Parameters in Eyes that Underwent CXL for Treatment of Keratoconus
||Pattern 1 Eyes (n=6)
|Pattern 2 Eyes (n=8*)
|Maximum topographic Sim K (D)
||52.55 (SD 4.29)
||51.18 (SD 4.10)
||49.10 (SD 5.79)
||49.74 (SD 6.30)
|Mean irregularity index 3 mm
|Mean irregularity index 5 mm
|Mean BSCVA (decimal)
|Mean SEQ (D)
||−7.21 (SD 5.15)
||−6.67 (SD 4.98)
||−9.13 (SD 6.49)
||−10.96 (SD 4.33)