Figure 1: Possible thickness distribution of the corneal epithelium. A uniform thickness (A) requires Bowman's radius to be smaller, while with identical radii (B) the epithelium is thicker at the center. Our measurements of radii at the 2- and 3.6-millimeter optical zones indicated a thinner central epithelium in the cadaver eye (C).
The recent enthusiasm for corneal refractive procedures, most notably radial keratotomy1"4 has been tempered in part by the unpredictability of the results.5'8 The advantage of performing a less invasive and more precise procedure with excimer laser sculpting of the anterior corneal surface9"12 may offer a better method of correcting refractive errors. With this high technology comes the expectation of a more accurate and predictable outcome. Since tissue removal by excimer laser photoablation can be made in steps as small as 0.1 to 0.3 µp?,13 refractive corrections could be made to within 0.03 D, a precision that surpasses both our current expectations and measuring capabilities, and perhaps even the needs of the patient as well.
Most keratorefractive procedures achieve their effect by altering the radius of curvature of the central anterior corneal surface either by flattening, as in radial keratotomy,14 or steepening, as in epikeratoplasty.15 In some of these techniques, the preoperative keratometric readings are performed directly on the epithelial surface while subsequently performing the surgery on the underlying stroma.16·17 This is the case in excimer photorefractive keratectomy.17
Due to large differences between the epithelial and stromal ablation rates,18 the epithelium is usually removed mechanically before photoablation. This assumes that under normal conditions the central epithelial thickness is constant, makes no significant contribution to the ocular refraction, and hence can be ignored in any subsequent calculation of refractive outcome following reepithelialization. However, we contend that it would have an effect on refraction as this procedure causes an alteration in ocular surface shape.
A simple model simulating the optical characteristics of the cornea with a radius of curvature of 7.75 mm gives a dioptric power of 43.55 D (using a keratometric index19 of 1.3375 and formula D = 0.3375/R for the schematic eye). When the radius of curvature of the model is reduced by 70 µp?, the mean value for the quoted range of epithelial thickness,20-21 a dioptric change of 0.40 D occurs (7.68 mm = 43.95 D). Assuming an equal refractive index for both epithelium and stroma, this suggests that removal of the corneal epithelium will increase the refractive power of the cornea due to alteration in the radius of curvature of the corneal surface (Fig IA). If the epithelial thickness is not uniform across the central optical zone, the difference in refracting power might be even greater and will not be constant over the entire cornea (Figs IB-C).
In this study, we measured the contribution of the epithelium to corneal refractive power at both 2.0and 3.6-millimeter diameter central optical zones.
MATERIALS AND METHODS
For this study, we used 10 whole fresh human eye-bank eyes supplied by the Florida Lions Eye Bank which were unsuitable for penetrating keratoplasty. Eyes that showed evidence of previous ocular surgery, substantial astigmatic irregularities, or significant corneal pathology, including clinically apparent postmortem exposure changes, were excluded from the study. The average time from death to keratometric measurement was 20.2 hours (range, 8 to 30 hours). Epithelium removal and all keratometric readings were performed by a trained corneal surgeon.
An identical procedure was followed in the preparation of each of the 10 eyes and their subsequent positioning in the artificial orbit system.22 Using the spiral of Tillaux formed by the extraocular muscles and the orientation of the optic nerve as guides, the eye was placed in the original anatomical position (nasal, temporal, superior, inferior) in relation to the surgeon and the operating microscope. Since the eyes were fresh and the corneas optically clear, no attempt was made to exchange the aqueous with dextran.23·24 The piezoelectric pressure transducer and the infusion needle of our artificial orbit system were inserted into the vitreous cavity. The intraocular pressure throughout the procedure was maintained at a constant level of 22 mm Hg. The operating microscope (OpMi-I, Carl Zeiss Ine, Thornwood, NY) used throughout this procedure to measure the refracting power of the corneal surface was fitted with an automatic digital keratometer (SK-I, Surgical AutoKeratometer, Canon USA Ine, NY). The keratometric readings of the SK-I have a resolution of ±0.1 D and ± Io of arc, and a repeatability of ± 0.3 D and ± 4° in cadaver eyes.25 With the aid of the SK-I, the operating microscope was aligned with the geometrical center of the cornea. We ensured that both the globe and the microscope remained stationary in relation to one another. Using this method, we were able to conduct measurements on the same area of the central cornea throughout the procedure.
The surface of the cornea was irrigated with a hyperosmotic solution of 7% dextran B.26 Unlike BSS, the dextran's high molecular weight (67 000 Dalton) will not induce stromal swelling and resultant artifactious changes in corneal radius of curvature. It is more likely that dextran's dehydrating effect will reduce any fine postmortem epithelial edematous changes.
Twenty seconds after a drop of dextran was placed on the cornea, a keratometric reading was taken. This gave an automatic reading and printout of both the flattest and steepest meridians as well as the astigmatic power and angle with respect to the 12 o'clock position. Immediately after the activation of the SK-I digital frame grabber by a footswitch, the cornea was again irrigated with a drop of dextran. Another reading was taken after an additional 20 seconds. A total of 35 sequential readings were taken of the corneal epithelial surface of each eye in this manner. The readings of the flattest and steepest meridians were later averaged individually for analysis.
Figure 2: Keratometrlc data obtained with the SK-I from a 24-year-old donor comea 5 hours after enucleation show that corneal power is less with the epithelium in place (Epithelium) than with It absent (Bowman). Differences in astigmatism are less pronounced. O.Z. = zone.
Figure 3: Keratometric difference in corneal power induced by removing the epithelium. "K Bowman - K epithelium" for both the flattest (Kl) and the steepest <K2) meridians of each eye di two zones. X = SK-I repeatability.25
Upon completion of these readings, the epithelium was completely removed with the aid of a cotton-tipped applicator. Great care was taken during this procedure to ensure that the globe's position in relation to the keratometer was not altered. After the epithelium was removed, the process of irrigating with dextran and taking readings at 20-second intervals was repeated. A total of 25 readings were made, averaging the flattest and steepest keratometric readings for subsequent analysis and comparison with the preoperative readings.
The procedure described above was followed for all 10 eyes. The cornea was kept moist throughout with topical dextran B. This enabled clear visualization of the keratometric mires. While it is likely that dextran B had some epithelial dehydrating effect, stromal hydration would have been minimally affected, if at all. This is because the agent was only in contact with the cornea for 20 minutes.26 Readings were taken at both the central 2.0- and 3.6-millimeter diameter zones.
The keratometric readings for each eye were plotted using Cricket Graph (Cricket Software, Ine, Malvern, Pa) as shown in Figure 2. A statistical analysis of the keratometric data was obtained through the 512 = Stat View (Brian Power Ine, Calabasas, Calif) software program.
All 10 eyes showed a change in corneal refractive power after removal of the epithelium at both the 2.0- and 3.6-inillimeter diameter zones (p < .001, paired ¿-test). In each case, the keratometric power of the cornea was greater after removal of the epithelium. This change occurred at both meridians, indicating that the spherical equivalent increased after the removal of the epithelium (Fig 3).
Comparing the two optical zones examined, we found a greater keratometric change at the central 2-millimeter diameter zone than at the 3.6-millimeter diameter zone in 8 of the 10 eyes (p = .143). At the former, the mean change was 1.03 D (range, 0.55 D to 1.85 D) while at the 3.6-millimeter diameter zone, it was 0.85 D (range, 0.29 D to 1.60 D), clearly indicative of an asphericity contributed by the corneal epithelium.
Figure A: Centrol corneal asphericity is demonstrated by this plot of the keratometric ratios computed for each meridian at both zones. In 7 out of 10 eyes, Bowman is found more prolate than the epithelium, an indication of a thinner central epithelium.
In Figure 4, the keratometric ratio (K^sub 2.00 mm^/ K^sub 3.6mm^) was plotted for each eye before and after removal of the epithelium. If the areas under examination were spherical, then the ratio should be 1.0. However, we found that the ratio was greater than 1.0 in 19 of the 20 readings (p < .002), suggesting a more prolate shape of the surfaces under examination. Of more significance, however, is the fact that Bowman's layer appeared to be more prolate than the epithelium in 7 of the 10 eyes, as the data in Figure 3 demonstrate.
Figure 5 indicates the change of astigmatic power and angle at both central optical zones, before and after removal of the epithelial surface. In 16 of the 20 total readings, Bowman's layer had less astigmatic power than the overlying epithelium. In a number of the eyes, there was also a noticeable change in the axis of astigmatism after removal of the epithelium.
While there is mention in recent literature regarding the role played by the epithelium in the results of keratorefractive17,18 procedures, we could not find any reference quantifying the contribution of the epithelium to the total power of the normal, unoperated eye. Therefore, we conclude this contribution to normal refraction was not fully appreciated. Our study clearly illustrates that the corneal epithelium plays a direct role in reducing the corneal refractive power by effectively increasing the radius of corneal curvature. The degree of change, up to 1.85 D in the central 2-niillimeter diameter zone of one of the eyes, with a mean of 1.03 D for all 10 eyes, was considerably more than mathematical calculations predict.
Assuming that the corneal surface is spherical and the epithelium is uniform in thickness over the corneal surface, the removal of the epithelial layer (70 µp?) results only in a 0.4-diopter change. This discrepancy indicates that the epithelium is not uniform in thickness over the cornea. It is known that the cornea with epithelium is prolate. We demonstrated in this study that Bowman's layer is more prolate than the epithelial surface (Fig 4). Therefore, a small change in epithelium thickness between the center and periphery can result in a significant dioptric change. This was shown by Munnerlyn and colleagues:27 A 3.36-micrometer difference in thickness between the corneal center and the 4- millimeter zone produces a 0.63-diopter change and therefore explain our findings. Such small change in epithelial thickness is very difficult to be detected histologically. Of potentially greater importance to keratorefractive procedures are the findings of a more aspheric shape of Bowman's layer and the change in astigmatism after removal of the epithelium.
The more aspheric shape of Bowman's layer is shown in Figure 4, illustrating the keratometric ratio with, and without epithelium. In 7 of 10 eyes, Bowman's layer appeared more prolate than the overlying epithelium. This would also indicate that the epithelium is not of uniform thickness across the cornea. It becomes increasingly thinner near the central cornea, indicating a more prolate surface to Bowman's layer, as illustrated in Figure IC.
The changes in astigmatic power and axis shown in Figure 5 were surprising and even harder to explain. These findings indicate that the epithelial layer is not of uniform thickness over Bowman's surface. It may be that the central corneal epithelial thickness is controlled dynamically by a number of factors including underlying irregularities of Bowman's and/or blinking movements of the lids.28
Figure 5: A change In astigmatism axis and power was observed between the epithelium and Bowman's surfaces in most eyes. (3,8 = SK-I dioptric and angular repeatability for low astigmatism value).25
Whatever the factores) responsible for epithelial thickness, the implications for keratorefractive procedures, most notably photorefractive keratectomy, are apparent. In the latter procedure, the epithelium is first removed mechanically prior to photoablation of Bowman's layer and the underlying stroma.9,17,18 Calculations regarding how much tissue should be removed, however, are based on pre- or intraoperative keratometric readings of the corneal epithelium.10,18'27 The computer programs and nomograms indicating how much tissue should be removed in photorefractive keratectomy assume either that the epithelium does not significantly contribute to total corneal refracting power, or at best can be ignored as the bare stroma will eventually be covered with a layer of epithelium of uniform thickness. Preliminary studies reveal that the refractive outcome of photorefractive keratectomy is as yet unpredictable.9,17,29 While some of the problems may be related to the laser delivery systems, it would appear that wound healing factors, particularly reepithelialization, is as yet unpredictable and contributes to the less than desirable outcome.30 Ignoring the refracting role of the epithelium may be a contributing problem.31
Figure 6: Deforming Bowman's surface (eg, with an intrastromal ring35) induces changes in epithelial thickness from normal (A) to hyperplastic (B) in close proximity to the implant.34
Another possibility not addressed by our study, and one that could potentially affect apparent epithelial astigmatism, is epithelial hydration. As the hyperosmotic dextran B solution used prior to keratometric measurement tends to dehydrate the epithelial surface, it may also alter the surface characteristics, thereby unmasking the increased epithelial astigmatism compared to that of the surface of Bowman's layer. This requires further study.
It is worthwhile to point out that all of the measurements presented in this article are based on a specular refraction of an air-liquid interface. The liquid, in this case dextran B, is resting on the surface to be studied, either Bowman's layer or the epithelium. Difference in film thickness distribution may cause an additional source of error. However, no keratometric differences were found using BSS or dextran B in these eyes. Therefore, the error must be smaller than 0.20, the SK-I readout limit.
It would appear from recent reports of epithelial regrowth after photorefractive keratectomy11,17,31 and other refractive procedures32,33 that the epithelium certainly does not have a passive role in resurfacing the cornea. Rather, it may be influenced strongly by changes in curvature of the underlying Bowman's layer or stroma.31
We know from previous work with anterior intrastromal ring implantation34 that there is a tendency with time for the epithelium to increase in thickness in the areas of the cornea surrounding the implant, as illustrated in Figure 6. This epithelial thickening modifies significantly the keratometric change induced by the implant. The epithelium, therefore, appears to have a dynamic role in maintaining a smooth anterior refracting surface, a process alluded to by Binder.32 Other recent work in this field suggests that reepithelialization after keratorefractive procedures does not occur in a readily predictable manner but is influenced strongly by the underlying changes induced by the procedure. Nirankari et al33 have suggested that myopic regression in epikeratoplasty may be secondary to thickening of the epithelium toward the center of the lenticule. A recent paper by Fantes et al17 which examined the question of wound healing after photorefractive keratectomy addressed this problem. They observed a transient thickening of the epithelium during the first few weeks of epithelial healing and felt it was possible that deeper excision might stimulate epithelial hyperplasia, particularly if the contour was not smooth. Similarly, early work by McDonald et al29 indicated a slightly thicker epithelium in the ablated area as compared to the nonablated area. This thickening of the epithelium after excimer laser ablation was further confirmed in work performed by Tuft et al.31 The suggestion by Lemp and Mathers28 that a significant factor in controlling the shedding of surface epithelial cells is the shearing force of the upper lid might help explain some of these findings. By altering the contour of the corneal surface with laser keratectomy, the cells covering the base should be more protected from the forces of the upper Hd, and thus able to increase in thickness. As a result of this, it appears that the epithelium attempts to minimize abrupt changes in surface curvature.
Although our study was small, we can nevertheless draw a number of conclusions regarding the keratometric contribution of the corneal epithelium to the overall refraction of the eye. The corneal epithelium accounts for an average of 1.03 D to the power of the eye at the central 2-millimeter diameter zone. Our findings indicate that Bowman's layer was more aspheric and more prolate than the overlying epithelial surface (Fig IC). The difference in astigmatism suggests that the epithelium does not form a layer of uniform thickness over Bowman's layer but may be influenced by factors such as movement of the lids, the underlying irregularities of Bowman's layer, and its radii of curvature.
Refractive computations based on keratometric measurements assume a unique index of refraction for the epithelium, Bowman, and the stroma. Should the epithelium index be higher, its contribution to the total refractive power of the eyes35 will be much higher than measured by keratometry.
In keratorefractive procedures such as photorefractive keratectomy, the refractive contribution of the epithelium must be taken into account to improve predictability. Further research into both corneal topography and refractive index measurements are required to ascertain the magnitude of this optical contribution.
1. Neumann AC, Osher RH, Fenzl RE. Radial keratotomy. A comprehensive evaluation. Doc Ophthalmol. 1984;56:275-301.
2. Arrowsmith PN, Marks RG. Visual, refractive and keratometric results of radial keratotomy: one-year follow-up. Arch Ophthalmol. 1984;102:1612-1617.
3. Deitz MR, Sanders DR, Marks RG. Radial keratotomy: an overview of the Kansas City Study. Ophthalmology. 1984;91:467-478.
4. Binder PS. The status of radial keratotomy in 1984. Arch Ophthalmol. 1984;102:1601-1603.
5. Schanzlin DJ, Rice Santos V, Waring GO, et al. Diurnal change in refraction, corneal curvature, visual acuity, and intraocular pressure after radial keratotomy in the PERK study. Ophthalmology. 1986;93:167-175.
6. Lynn MJ, Waring GO, Sperduto RD, the PERK Study Group. Factors affecting outcome and predictability of radial keratotomy in the PERK study. Arch Ophthalmol. 1987;105:42-51.
7. Deitz MR, Sanders DR. Progressive hyperopia with long-term follow-up of radial keratotomy. Arch Ophthalmol. 1985;103:782-784.
8. McDonald MB, Klyce SD, Suarez H, Kandarakis A, Friediander MH, Kaufman HE. Epikeratophakia for myopia correction. Ophthalmology. 1985;92:1417-1422.
9. McDonald MB, Frantz JM, Klyce SD, et al. One-year refractive results of central photorefractive keratectomy for myopia in the nonhuman primate cornea. Arch Ophthalmol. 1990;108:40-47.
10. Steinert RF, Puliafito CA. Corneal incisions with the excimer laser. In: Sanders DR, Hofmann RF, Salz JJ, eds. Refractive Corneal Surgery. Thorofàre, NJ: Slack Ine; 1986:401-410.
11. Marshall J, Trokel S, Rothery S, Krueger RR. Photoablative reprofiling of the cornea using an excimer laser: photorefractive keratectomy. Lasers Ophthalmol. 1986;1:21-48.
12. Hanna KD, Chastang JC, Asfar L, Samson J, Pouliquen Y, Waring GO. Scanning slit delivery system. J Cataract Refract Surg. 1989;15:390-396.
13. Puliafito CA, Steinert RF, Deutsch TF, Hillenkamp F, Dehm EJ, Adler CM. Excimer laser ablation of the cornea and lens; experimental studies. Ophtholmology. 1985;92:741-748.
14. Steel D, Jester JV, SaIz J, et al. Modification of corneal curvature following radial keratotomy in primates. Ophthalmology. 1981;88:747-754.
15. McDonald MB, Koenig SB, Safir A, Friedlander MH, Kaufman HE, Granet N. Epikeratophakia: the surgical correction of aphakia. Update: 1982. Ophthalmology. 1983; 90:668.
16. Keates RH, Kelley CG. Epikeratophakia for myopia: preliminary considerations. Journal of Refractive Surgery. 1985;1:25-28.
17. Fantes FE, Hanna KD, Waring GO, Pouliquen Y, Thompson KP, Savoldelli M. Wound healing after excimer laser keratomileusis (photorefractive keratectomy) in monkeys. Arch Ophthalmol. 1990;108:665-675.
18. Seiler T, Kriegerowski M, Schnoy N, Bende T. Ablation rate of human corneal epithelium and Bowman's layer with the excimer laser (193 nm). Refract Corneal Surg. 1990;6:99-102.
19. von Bahr G. Measurements of the thickness of the cornea. Acta Ophthalmol. 1948;26:247-266.
20. Rodrigues MM, Waring GO, Hackett J, Donohoo P. Cornea. In: Tasman W, ed. Duane's Biomedical Foundations of Ophthalmology. Philadelphia, Pa: JB Lippincott; 1989;l(chap 8):1-20.
21. Davson H. Physiology of the Eye. 3rd ed. New York, NY: Academic Press; 1972:68.
22. Lowery JA, Parel J-M, Roussel TJ, Simon G, Lee W, Nose I. Artificial orbit system for experimental surgery with enucleated globes. Ophthalmic Surg. 1990;21:522-528.
23. Swinger CA, Kornmehl EW. Dehydration of post-mortem eyes for practice and experimental surgery. Ophthalmic Surg. 1985;16:182-183.
24. Duffey RJ, Tchah H, Lindstrom RL. Human cadaver corneal thinning for experimental refractive surgery. Refract Corneal Surg. 1989;5:41-42.
25. Simon G, Parel J-M, Nose I, Lee W. Modification, calibration and comparative testing of an automated surgical keratometer. Refract Corneal Surg. 1991;7:151-160.
26. Kim JH, Green K, Martinez M, Paton D. Solute permeability of the corneal endothelium and Descemets membrane. Exp Eye Res. 1971;12:231-238.
27. Munnerlyn R, Kooms SJ, Marshall J. Photorefractive keratectomy: a technique for laser refractive surgery. J Cataract Refract Surg. 1988;14:46-52.
28. Lemp MA, Mathers WD. Corneal epithelial cell movement in humans. Eye. 1989;3:438-445.
29. McDonald MB, Frantz JM, Klyce SD, et al. Central photorefractive keratectomy for myopia; the blind eye study. Arch Ophthalmol. 1990;108:799-808.
30. Corneal epithelial and stromal reactions to excimer laser photorefractive keratectomy. Editorial. Gipson IK: Concerns regarding the response of the corneal epithelium to excimer laser ablation. Cintron C: Unpredictable corneal cicatrization. Binder PS: The excimer laser and radial keratotomy: Two vastly different approaches for myopia correction. Arch Ophthalmol. 1990;108:1539-1542.
31. Tuft SJ, Zabel RW, Marshall J. Corneal repair following keratectomy. A comparison between conventional surgery and laser photoablation. Invest Ophthalmol Vis Sci. 1989;30:1769-1777.
32. Binder PS. What we have learned about corneal wound healing from refractive surgery. Refract Corneal Surg. 1989;5:98-120.
33. Nirankari VS, Rodriguez MM, Bauer SA, Jarmarwala MG, Glaros DS. Effects of epikeratoplasty on the host cornea; an , experimental study. Cornea. 1990;9:211-216.
34. Simon G, Barraquer JI. Queratocricoemfitesis, nuevo proced- ^ imiento de Cirurgia Refractiva (Estudio Experimental). ArcA Soc Oftal Invest. 1989;1:87-94.
35. Rol P, Parel J-M, Hanna K. Computed estimation of visual acuity after laser refractive keratectomy. SPIE Ophthalmic Technologies II Proc Ed. Puliafito, Calif 1991;1423:89-93.