Journal of Refractive Surgery

A Comparison of Corneal Stromal Edema Induced From the Anterior or the Posterior Surface*

Stephen M Cristol, MD; Henry F Edelhauser, PhD; Michael J Lynn, MS

Abstract

ABSTRACT

BACKGROUND: Many differences between the anterior and posterior corneal stroma have been reported. The physiological and mechanical properties of the cornea are a summation of these properties across each of the corneal regions. This article investigates corneal stromal swelling that is experimentally induced through each surface.

METHODS: Corneal stromal swelling was induced in human and rabbit corneas through either the anterior or posterior surface. The rate of stromal swelling was analyzed with a linear regression model.

RESULTS: Swelling in the rabbit stroma was 3.65 × faster when induced through the posterior surface than through the anterior surface (p < .0001), while the human stroma swelled 13.1 × faster through the posterior surface (p < .0001). The hydration of the stroma increased during swelling through the posterior surface, but paradoxically decreased during swelling through the anterior surface.

CONCLUSIONS: These experiments showed that stromal swelling occurs more rapidly through the posterior corneal surface than through the anterior surface. These results may have implications for the refractive surgeon performing laser ablative procedures on the anterior surface of the cornea. Refract Corneal Surg 1992;8:224-229.)

Abstract

ABSTRACT

BACKGROUND: Many differences between the anterior and posterior corneal stroma have been reported. The physiological and mechanical properties of the cornea are a summation of these properties across each of the corneal regions. This article investigates corneal stromal swelling that is experimentally induced through each surface.

METHODS: Corneal stromal swelling was induced in human and rabbit corneas through either the anterior or posterior surface. The rate of stromal swelling was analyzed with a linear regression model.

RESULTS: Swelling in the rabbit stroma was 3.65 × faster when induced through the posterior surface than through the anterior surface (p < .0001), while the human stroma swelled 13.1 × faster through the posterior surface (p < .0001). The hydration of the stroma increased during swelling through the posterior surface, but paradoxically decreased during swelling through the anterior surface.

CONCLUSIONS: These experiments showed that stromal swelling occurs more rapidly through the posterior corneal surface than through the anterior surface. These results may have implications for the refractive surgeon performing laser ablative procedures on the anterior surface of the cornea. Refract Corneal Surg 1992;8:224-229.)

The visual impairment caused by corneal stromal edema has been of interest to ophthalmic surgeons for many years. Ultrastructural studies of corneal edema have shown that the mean spacing between collagen fibrils increases during edema.1-3 This spreading of the collagen fibrils results in a decrease in the short range order of the fibrils and the consequent clouding of the cornea.4,5 Recently, corneal edema has also been associated with a loss of proteoglycans from the stroma.6 Turss et al7 reported that the anterior stroma is less hydrated (3.04 g of water/g dry cornea) than the posterior stroma (3.85 g of water/g dry cornea). Castoro et al8 found that the average ratio of keratan sulfate to dermatan sulfate increases linearly with corneal depth from 1.59 in the anterior stroma to 2.23 in the posterior stroma. As keratan sulfate proteoglycan has a greater water sorptive and lesser water retentive capacity than dermatan sulfate proteoglycan,9 the difference in hydration between the anterior and posterior stroma may be attributable to the proteoglycan grathent.

In addition to the biochemical differences between the anterior and posterior stroma, there are also anatomic and physiological differences. The anterior stroma has narrower and more interwoven lamellae,10 a greater loss of transparency with increases in hydration,11 and a lesser capacity to swell11,12 than the posterior stroma. These studies suggest that the corneal stroma has at least two different regions with distinct properties.

The purpose of this study was to investigate the rate of corneal stromal swelling induced through the denuded anterior or posterior corneal surface when exposed to deionized water. These experiments used deionized water to achieve a maximal effect, although it is not a physiologic solution. In all experiments, the corneas were perfused at a constant pressure of 18 mm Hg.

Studies of differences between the anterior and posterior stroma may find application due to the increasing use of the excimer laser in refractive surgery. Many of the physiological and mechanical properties of the cornea are a summation of those properties across each corneal region. Ablation of the anterior cornea may significantly alter the swelling characteristics in the remaining stroma. Clinical manifestations of this change in the cornea may occur as the number of patients who have undergone this type of procedure increases.

MATERIALS AND METHODS

Rabbit Eyes

Rabbit corneas were obtained from Dutch-belted rabbits (anterior and posterior swelling experiments used 2.5 kg and 4.0 kg rabbits, respectively). The rabbits are anesthetized with 50 mg of ketamine hydrochloride and 10 mg of xylazine hydrochloride prior to receiving a lethal dose of sodium pentobarbital (324 mg). Excision and mounting of the corneas follows the procedure described by Dikstein and Maurice,13 except that the epithelium is removed with a Gill knife before mounting.

Human Eyes

Donated human eyes, that are unsuitable for transplantation, are obtained from the Georgia Lions Eye Bank. The corneal epithelium is removed with a Gill knife. The globe is entered with a razor and then the anterior segment is removed with scissors. The ciliary body and lens-iris diaphragm are removed en bloc with forceps. The excised corneoscleral tissue is mounted using the same technique and apparatus used to mount rabbit corneas, except that the conjunctiva is not secured to the mounting ring.

Only eyes with clear corneas and an intact anterior chamber were used in this study. The median times from death to enucleation and from enucleation to perfusion were 2.75 hours and 27.5 hours, respectively. The mean initial thickness of the control corneas was 663 µm.

Corneal Swelling

For control corneas, the anterior surface is covered with silicone oü and the posterior surface is perfused at 100 u.L/minute by a Harvard infusion pump with glutathione bicarbonate ringer's (GBR) for rabbit corneas or BSS Plus (Alcon Surgical, Ine, Fort Worth, Tex) for human corneas.14

In the anterior swelling group, stromal edema is experimentally induced through the anterior corneal surface by placing deionized water on the anterior surface while the corneal endothelium is perfused with GBR or BSS Plus. The deionized water on the anterior surface is replaced every 15 minutes during the experiment.

In the posterior swelling group, stromal edema is experimentally induced through the posterior corneal surface by removing the endothelium with a Gill knife during the mounting procedure. After mounting, the anterior corneal surface is covered with silicone oil and the posterior surface is perfused with deionized water.

During a perfusion, control and experimental corneas are mounted within a dual-chamber in vitro specular microscope, which allows sequential measurements of corneal thickness.13'15 Corneas are maintained at a pressure of about 18 mm Hg throughout a perfusion. Perfusions are stopped when a cornea becomes too cloudy to measure or after 2 to 3 hours.

Corneal Hydration

Following the perfusion, the corneas are removed from the mounting ring and placed on gauze to remove free fluid from the anterior and posterior surfaces. The central cornea is cut from the corneoscleral button along the mark left by the mounting ring (approximately 12 mm in diameter). The central cornea is weighed and placed in a drying oven at 105° C. The corneas are dried until they achieve constant weight on successive days. Hydration, the ratio of the weight of corneal water to the dry weight of the cornea, is calculated:

H = (wet weight - dry weight) / dry weight.

Data Analysis

Corneal thickness measurements are normalized to a starting thickness of zero by subtracting the initial thickness from each measurement. Data from corneas swollen from the anterior surface are adjusted for endothelial function by subtracting the simultaneous normalized thickness of the paired control cornea. For corneas swollen from the posterior surface, the data is not adjusted for endothelial function since the endothelium has been removed. Data within each experimental group of corneas are pooled. Additionally, data is pooled from both rabbit control groups and from both human control groups. A linear regression model is computed for each data set using SAS/PC (SAS Institute, Ine, Cary, NC). The change in corneal thickness is modeled as a function of the time from the start of the experiment. Since the data was normalized, the y-intercept of the model should be zero. The regression slope is the rate of corneal swelling (µm/min). Finally, the anterior and posterior swelling rates are compared statistically in each species. The hydration data is also pooled as described above. Average hydrations are compared using the appropriate two sample t-test.

RESULTS

This study used six pairs of human eyes and six pairs of rabbit eyes to investigate stromal swelling induced at each isolated corneal surface. One cornea of each pair was used as a control. One of the rabbit control corneas in a posterior swelling experiment was damaged during mounting and was excluded from the study. Figure 1 shows the swelling curves for the corneas studied.

Figure 1: (A-F) Corneal thickness as a function of the time from the start of each perfusion. The graphs show the swelling curve for each control and experimental cornea studied. The graphs are for rabbit (A-C) and human (D-F) data. The scale of each graph is the same.

Figure 1: (A-F) Corneal thickness as a function of the time from the start of each perfusion. The graphs show the swelling curve for each control and experimental cornea studied. The graphs are for rabbit (A-C) and human (D-F) data. The scale of each graph is the same.

Table

Table 1Models of Rabbit Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt + [intercept])*Table 2Models of Human Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt 4- [intercept])*Figure 2: (A,B) Regression models of corneal thickness as a function of the time from the start of each perfusion. The graphs for the (A) rabbit and (B) human experiments are shown. (C = control group, A = anterior swelling group, P = posterior swelling group, P(ad|) = posterior swelling group (adjusted model)).

Table 1

Models of Rabbit Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt + [intercept])*

Table 2

Models of Human Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt 4- [intercept])*

Figure 2: (A,B) Regression models of corneal thickness as a function of the time from the start of each perfusion. The graphs for the (A) rabbit and (B) human experiments are shown. (C = control group, A = anterior swelling group, P = posterior swelling group, P(ad|) = posterior swelling group (adjusted model)).

Tables 1 and 2 list the coefficients of the models developed to describe the swelling of rabbit and human stroma, respectively. In the posterior swelling model for rabbit, a disproportionate influence of the slower swelling corneas upon the model was suggested by the large v-intercept that was significantly different from zero p ~ .0162). To adjust for this, a second model of posterior swelling of the rabbit stroma was computed using data from the first 45 minutes of these experiments (Table 1). After this adjustment, the y-intercept was no longer significantly different from zero (p = .3191). No other models had a γ-intercept that was significantly different from zero (a = .05). Figure 2 is a graphical summary of these results.

The difference between the anterior and posterior swelling rate was significant in both the rabbit (p < .0001) and the human (p < .0001) models. In these experiments, the rabbit stroma swelled 3.65 × faster when the swelling was induced from the posterior surface than when induced from the anterior surface (5.51 × using the adjusted model of posterior swelling). In human stroma, swelling induced from the posterior surface was 13.1 × faster than from the anterior.

Corneal hydration measurements after experimental swelling are summarized in Table 3. Corneas in the posterior swelling group became more hydrated than control corneas, while those in the anterior swelling group became less hydrated. This was true for the average hydrations and for each corneal pair. These findings were statistically significant in the human samples, but only marginally significant in the rabbit samples.

Table

Table 3Corneal Hydration (mg water/mg dry cornea) Following Experimental Swelling (Mean ± Standard Deviation)

Table 3

Corneal Hydration (mg water/mg dry cornea) Following Experimental Swelling (Mean ± Standard Deviation)

DISCUSSION

Other studies have examined various aspects of stromal swelling when the swelling is induced through both corneal surfaces and the cut edge of the corneal button.12-16"19 This experiment sought to isolate the effect of swelling through a single corneal surface denuded of its limiting membrane. The cornea was maintained at a physiological pressure to simulate the mechanical effects of the intraocular pressure on water distribution in the stroma. To induce swelling, deionized water was used as a hypotonic solution. Similarly, the corneal epithelium and endothelium were denuded rather than chemically or metabolically inactivated so that the physical presence of inactive cells would not impede swelling. These experiments showed that stromal swelling occurs more rapidly through the posterior corneal surface than through the anterior surface. Given the many differences between the anterior and posterior corneal stroma, it is not surprising to find an additional difference.

During swelling through the posterior surface, the corneal hydration increased. This is consistent with the theory that swelling results when fluid invades the cornea. However, during swelling through the anterior surface, corneal hydration decreased. This appears paradoxical but may be due to a greater loss of bound water with glycosaminoglycans of the anterior stroma. Alternatively, changes in the thickness of the cornea may result from changes in the ionic milieu of the stroma that alter the molecular conformations of the proteoglycans or glycosaminoglycans in the ground substance. This finding must be interpreted cautiously in view of the small sample size.

These findings have a few limitations due to the experimental design. Since the rate of stromal swelling decreases with time,18 slower swelling rates are ' averaged with the initial maximal swelling rate resulting in an estimation of the time averaged swelling rate, not the maximal swelling rate. Also, in the posterior swelling experiments, the cornea begins to swell during mounting and before an initial thickness measurement can be obtained. This results in an underestimation of the average swelling rate for the posterior swelling experiments. Adjusting for the presence of the endothelium on the control corneas and the corneas swollen through the anterior surface is difficult. Although a functioning endothelium will remove water at a faster rate from the stroma of a swollen cornea than from a normal cornea, the model used for analysis is too conservative in accounting for this difference. While this results in an underestimation of the swelling rate through the anterior stroma, this bias is small and would not result in a meaningful change in the relative swelling rate (because the rate of posterior swelling is large compared to the maximum rate of thinning that can reasonably be attributed to the endothelium). Also, the increases and decreases in hydration with swelling must be correlated with the stromal ultrastructure to explain this difference between the anterior and posterior stroma.

These findings may be applicable to refractive surgery with respect to laser ablative procedures and the chronic management of patients who undergo these procedures. The very slow swelling rate of the cornea through the denuded anterior surface suggests that, even under the worst conditions, stromal swelling would not occur rapidly enough to interfere with current ablative procedures. This agrees with clinical experience. Nevertheless, the potential exists for refractive surgical procedures to alter the overall properties of the cornea by removing part of the anterior stroma. This may leave the remaining cornea with characteristics similar to those of the posterior stroma, leading to a greater likelihood of corneal edema following future ophthalmic procedures. More experience with refractive * surgical procedures will determine if this is a real or theoretical concern.

In summary, the corneal stroma swells more rapidly through the denuded posterior surface than through the denuded anterior surface. Swelling through the posterior surface increases corneal hy- k dration in contrast to the anomalous decrease in hydration during swelling through the anterior surface.

REFERENCES

1. Kanai A, Kaufman HE. Electron microscopic studies of swollen corneal stroma. Ann Ophthalmol. 1973;5:178-190.

2. Goldman J, Kuwabara T. Histopathology of corneal edema. Int Ophthalmol Clin. 1968;8:561-579.

3. Menasche M, Dagonet F, De Waegeneer MJ, Pouliquen Y. Induced corneal swelling: an experimental model for investigation of collagen and interfibrillar substance relationships. Cornea. 1985-86;4: 149-156.

4. Hart RW, Farrell RA. Light scattering in the cornea. J Opt SocAmfAJ. 1969;59:766-774.

5. Goldman JN, Benedek GB, Dohlman CH, Kravitt B. Structural alterations affecting transparency in swollen human corneas. Invest Ophthalmol. 1968;7:501-519.

6. Rangas TA, Edelhauser HF, Twining SS, O'Brien WJ. Loss of stromal glycosaminoglycans during corneal edema. Invest Ophthalmol Vis Sci. 1990;31:1994-2002.

7. Turss R, Friend J, Reim M, Dohlman C. Glucose concentration and hydration of the corneal stroma. Ophthalmic Res. 1971;2:253-260.

8. Castoro J, Bettelheim A, Bettelheim F. Water grathents across bovine cornea. Invest Ophthalmol Vis Sci. 1988;29:963-968.

9. Bettelheim F, Plessy B. The hydration of proteoglycans of bovine cornea. Biochim Biophys Acta. 1975;381:203-214.

10. Komai Y, Ushiki T. The three-dimensional organization of collagen fibrils in the human cornea and sclera. Invest Ophthalmol Vis Sci. 1991;32:2244-2258.

11. Kikkawa Y, Hirayama K. Uneven swelling of the corneal stroma. Invest Ophthalmol. 1970;9:735-741.

12. Lee D, Wilson G. Non-uniform swelling properties of the corneal stroma. Curr Eye Res. 1981;1:457-461.

13. Dikstein S, Maurice D. The metabolic basis to the fluid pump in the cornea. J Physiol (Lond). 1972;221:29-41.

14. Edelhauser H, Gonnering R, Van Horn D. Intraocular irrigating solutions. A comparative study of BSS Plus and lactated Ringer's solution. Arch Ophthalmol. 1978;96:516-520.

15. McCarey B, Edelhauser H, Van Horn D. Functional and structural changes in the corneal endothelium during in vitro perfusion. Invest Ophthalmol. 1973;12:410-417.

16. Hedbys B, Mishima S. The thickness-hydration relationship of the cornea. Exp Eye Res. 1966;5:221-228.

17. Hedbys B, Dohlman C. A new method for the determination of the swelling pressure of the corneal stroma in vitro. Exp Eye Res. 1963;2:122-129.

18. Friedman MH, Green K. Swelling rate of corneal stroma. Exp Eye Res. 1971;12:239-250.

19. Hedbys B, Mishima S. Flow of water in the corneal stroma. Exp Eye Res. 1962;1:262-275.

Table 1

Models of Rabbit Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt + [intercept])*

Table 2

Models of Human Corneal Stromal Thickness, T, as a Function of Time, t (T = [slopelt 4- [intercept])*

Figure 2: (A,B) Regression models of corneal thickness as a function of the time from the start of each perfusion. The graphs for the (A) rabbit and (B) human experiments are shown. (C = control group, A = anterior swelling group, P = posterior swelling group, P(ad|) = posterior swelling group (adjusted model)).

Table 3

Corneal Hydration (mg water/mg dry cornea) Following Experimental Swelling (Mean ± Standard Deviation)

10.3928/1081-597X-19920501-09

Sign up to receive

Journal E-contents