Reduction of Stromal Swelling Pressure after UVA-Riboflavin Cross-Linking

Abstract

Purpose.: We evaluated whether UVA-riboflavin collagen cross-linking (CXL) reduces stromal swelling pressure (SP) in porcine and human corneas ex vivo.

Methods.: Porcine corneas (n = 35) were divided into five groups: Full-thickness buttons, riboflavin-treated buttons, CXL-treated buttons, riboflavin-treated anterior and posterior lamellae, and CXL-treated anterior and posterior lamellae. Riboflavin- or CXL-treated human corneas (n = 6) were cut into anterior and posterior lamellae. The force exerted by the specimens during swelling in saline was recorded and SP was calculated. Dry weights were obtained for solids correction.

Results.: In full-thickness porcine buttons, a significantly reduced SP was observed after CXL (40.07 ± 3.86 mm Hg) compared to riboflavin treatment (68.13 ± 11.39 mm Hg, P = 0.02) at 5% compression. Also, a trend toward reduced SP in the CXL-treated anterior human lamellae (9.9 mm Hg) was found compared to the riboflavin group (25 mm Hg) at 5% compression. In the anterior porcine lamellae a significant SP reduction was observed in the CXL group versus the riboflavin group (P < 0.001, two-way ANOVA). Likewise, in the posterior porcine lamellae, a significant SP reduction was observed in the CXL group versus riboflavin alone (P < 0.05, two-way ANOVA). Posterior human lamellae did not differ in SP when comparing CXL and riboflavin groups.

Conclusions.: Our study demonstrated a significant reduction in anterior stromal SP in porcine and human corneas after CXL. This finding suggested that CXL may reduce corneal SP in vivo, and thereby reduce edema and improve vision. Thus, in the clinical setting, patients suffering from corneal edema may benefit from CXL treatment.

Introduction

The biophysical properties of the cornea depend on precise maintenance of tissue hydration. The corneal stroma has an innate tendency to imbibe fluid and swell to a degree higher than any other connective tissue in the body. This capability and the transparency in vivo are unusual properties for a connective tissue. ^1,2 When the stroma swells it loses its transparency ^3–6 because of increased light scattering. ⁷ The physiologic control of swelling, therefore, is crucial to the maintenance of transparency. ⁸ UVA-riboflavin cross-linking (CXL) is an evolving procedure originally introduced in 1998 to increase stiffness in keratoconic corneas to halt progression of the disease. ⁹ Since its introduction, several clinical applications for CXL have been proposed, including treatment of corneal edema. ^10–13  

There are two ways in which a gel can swell: by volume exclusion ¹⁴ or in aqueous media, by the additional force of net electric charges fixed to the insoluble matrix of the gel (Donnan swelling). ¹⁵ This type of swelling is 600 times faster than swelling by volume exclusion and is the physiologically important type of swelling. ⁵ The corneal stroma consists of several hundred sheets of lamellae containing parallel fibrils. ¹⁶ There seems to be no significant cross-linking between collagen fibrils, which in other tissues act to impede the swelling phenomena. ⁵ The corneal stroma is held in a relatively dehydrated state, according to the pump-leak hypothesis proposed by Maurice. ¹⁷ This negative fluid pressure in the stroma is maintained by active ion transport over the cellular layers. ^17,18 The gel pressure or swelling pressure (SP) of a tissue can be defined as “the mechanical pressure applied to its surfaces which is needed to prevent swelling when the tissue has free access to a physiologic aqueous medium.” ^18,19 Several methods have been assessed over the past 50 years to estimate corneal SP ex vivo, including osmometry ^4,20,21 or by mechanical methods. ^21–23 In vivo, the swelling pressure has been measured by direct stromal cannulation ²² or with hydrogel implants. ²⁴ In our study SP is measured ex vivo while stromal hydration is stabilized mechanically by a fixed distance between a pair of porous compression plates. 

Materials and Methods

Riboflavin Solution

A solution of 0.1% riboflavin-5-phosphate (Unikem, Copenhagen, Denmark) in PBS-based 20% Dextran T-500 (Amersham Biosciences, Uppsala, Sweden) was prepared. Osmolarity was adjusted to 300 mOsm/kg and pH to 7.0, as described previously. ²⁵  

Ambient Light Measurement

The riboflavin-only group and the untreated group were prepared under dim light conditions. To rule out any effects of incident light, ambient light intensity was measured using a photodiode (S120VC; Thorlabs Sweden AB, Göteborg, Sweden) and a power meter (PM100A compact power meter console; Thorlabs Sweden AB). Additionally, light intensity was measured with a 450/40 nm filter to quantify light intensity around riboflavins' greatest upper absorption maximum. ²⁵ Ambient light intensity without any filter was kept below 9 μW and 0.1 μW as measured with a 450/40 nm filter. 

Preparation of Full-Thickness Porcine Corneal Buttons

Fresh porcine eyes were obtained from the local abattoir (Danish Crown, Horsens, Denmark) on the day of the experiments and testing was initiated within two hours postmortem. The eyes were dissected from the pigs immediately after death and before any scalding was performed. One eye from each pig was used. The central corneal thickness (CCT) was measured in the intact eye ex vivo with ultrasound pachymetry (Sonomed Pacscan 300P; Sonomed, Inc., New Hyde Park, NY). After a few seconds of 96% volume alcohol application, the corneal epithelium was removed with a Beaver knife, and the CCT was measured. In a dim environment, two drops of 0.1% riboflavin-5-phosphate solution were applied to the denuded, anterior stromal surface every 5 minutes for 30 minutes. In the riboflavin group (n = 8), the riboflavin application was continued another 30 minutes in the dim environment, to avoid influence of ambient light. In the treatment group (n = 8), the denuded surface was irradiated with an UV-A lamp (UV-X; PESCHKE Meditrade GmbH, Switzerland) at 370 nm with an intensity of 3 mW/cm² for 30 minutes, giving a total dose of 5.4 J/cm². The radiation source was calibrated with the UV-meter (Lutron YK-34UV), supplied by the UV-lamp manufacturer (Lutron Electronic Enterprise Co., Ltd., Taipei, Taiwan). During the 30-minute radiation period, riboflavin application was continued. After cross-linking or riboflavin application alone, 8-mm buttons were trephined and the endothelium was removed with a Beaver knife. The stromal buttons then were weighed using an Ohaus Discovery DV215CD digital scale (Ohaus Europe GmbH, Nänikon, Switzerland). The weighing was repeated after two hours of preswelling in isotonic saline at room temperature and again immediately after swelling pressure measurement after gentle blotting. Finally, the dry weights of the specimens were obtained, directly after lyophilization for >24 hours (VirTis Benchtop K Manifold freeze dryer; SP Industries, Gardiner, NY). 

Further, five corneal buttons were used as controls without riboflavin or UVA treatment. Three porcine corneal buttons with the epithelium and endothelium removed were processed for histologic examination. These were stained with hematoxylin-eosin (HE) and periodic acid-Schiff stain (PAS) to confirm whether the epithelium and endothelium were removed completely. 

Anterior versus Posterior Stromal Lamellae in Porcine Eyes

The CCT was measured in the intact eye using ultrasound pachymetry. After a brief application of 96% volume alcohol, the corneal epithelium and endothelium were removed with a Beaver knife, and the CCT measured by US-pachymetry as above. 

The cornea, including a 4-mm scleral rim, was excised, and the specimen was inverted and mounted in a Barron artificial anterior chamber (Katena Products, Inc., Denville, NJ). The chamber was filled with air to avoid swelling. Pressure was maintained at 70 cm H₂O with a saline infusion set connected to the air-filled artificial anterior chamber. A 300 μm thick (ø = 8 mm), posterior lamella was cut with a ZEISS VisuMax femtosecond laser (FSL; Carl Zeiss Meditech, Jena, Germany). Leaving the lamella in situ, the specimen was reverted and an anterior lamella was cut (300 μm, ø = 8 mm). The laser raster pattern left sufficient bridged tissue across the cutting plane to allow the lamellae to stay attached. With both lamellae still attached, two drops of 0.1% riboflavin-5-phosphate solution were applied to the denuded, anterior stromal surface every 5 minutes for 30 minutes. In the riboflavin group (n = 7), the riboflavin application was continued for another 30 minutes. In the treatment group (n = 7), CXL was performed as in the full-thickness buttons. 

After cross-linking or riboflavin application alone, the lamellae were dissected carefully and weighed. The weighing was repeated after 2 hours of preswelling in isotonic saline and immediately after swelling pressure measurement after gentle blotting. The lamella (anterior or posterior) to be tested last biomechanically was kept in a moist chamber until preswelled and tested (approximately 5 hours from detachment). The procedure was repeated in the riboflavin group (n = 7). Finally, the dry weights of the specimens were determined as above. 

An additional 24-hour lyophilization and reweighing was performed in five corneas to control for residual water content. The research was conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Human Corneal Lamellae

Six human corneas from four different donors were obtained from the Rocky Mountain Lions Eye Bank (Aurora, CO). The tissue was procured and processed in compliance with the Eye Bank Association of America Medical Standards and government regulations, and in compliance with the Declaration of Helsinki. The tissue was transported in Optisol GS (Bausch & Lomb, Inc., Rochester, NY). Lamellae from the riboflavin and CXL groups were prepared as in the porcine groups above. In the FSL-procedure, the human corneas were divided equally into an anterior and posterior part. The cut depth was calculated from the ultrasound pachymetry after epithelial removal. The FSL cut segment diameter was 7.5 mm. Finally, a 6.5-mm button was punch-trephined through both lamellae of the 7.5 mm-button (Fig. 1b). 

Figure 1. 

 

Lamellae cut with the femtosecond laser. (a) Anterior and posterior 300 μm lamellae are cut in porcine corneas. (b) Human corneas are divided equally in two lamellae. Finally, the ø = 6.5 mm button is trephined (dashed vertical lines).

Biotester

A biaxial load cell–based system was customized for SP measurement (Biotester 5000; CellScale Biomaterials Testing, Waterloo, Canada). This design allows simultaneous recording of x- and y-axis displacement, force, and time. However, in our study, the x-axis was kept fixed and only the force in the y-axis (swelling force) was recorded. The semiconductor strain gauge–based load cells were sampled at 100 Hz with hardware averaging over 8 samples to reduce noise. A stepper motor under open loop control drove the actuator. Synchronously, a CCD-camera recorded images at 1280 × 960 pixels for documentation. Customized software showed real-time plots of force, displacement, and time. 

Fixation Unit

The excised corneal button or FSL-cut lamella was fixed vertically by compression between two flat, porous plates (GenPore, Reading, PA). The plates were treated hydrophilically, with ultra high molecular weight polyethylene with 50 μm pores with a density of 40% to 50% void volume (data supplied by manufacturer). These permitted bulk diffusion of water to the corneal surfaces. The plates were changed regularly to avoid restriction of water diffusion by debris blocking the pores. The pores also served as a rough surface to keep the cornea in place. The fixture arms were machined in inert and rigid Delrin polymer. The fixation unit (Fig. 2) was submerged into a heated isotonic saline bath at 37°C for 30 minutes before measurements to ensure full hydration and correct temperature of the plates. The corneal samples were mounted and allowed to imbibe isotonic 9 mg/mL sodium chloride solution (B. Braun Melsungen AG, Melsungen, Germany). The resulting swelling force was measured as a function of time and displacement. 

Figure 2

 

Schematic drawing of the customized BioTester fixation unit (top view).

Software and Data Acquisition

Data were recorded in LabJoy version 9, provided by the BioTester manufacturer (CellScale Biomaterials Testing). The swelling pressure was recorded at 1 Hz at 5 compression steps (+30%, +20%, +10%, +5%, and −5% of initial stromal thickness, respectively). Each compression step was fixed for 30 minutes allowing the swelling force to reach a plateau phase, giving a total recording time of 2.5 hours per specimen. A distance and force calibration procedure was performed before measurements of every corneal button. Likewise, the actuators were reset before each measurement. The movement of the Biotester actuators was calibrated from zero (when the compression plates touch). The distances between markings on these plates and the gap between these were measured optically in Image J v1.45 (Rasband WS; National Institute of Mental Health, Bethesda, MD) to confirm readings in the LabJoy software. 

Calculations were made in Microsoft Excel 2011 (Microsoft Corp., Redmont, WA) and Prism 5 (GraphPad Software, Inc., La Jolla, CA). 

Swelling Pressure

When the corneal button swells in saline, it exerts a force orthogonal to its faces (y-direction). This swelling force divided by the area of the corneal button is calculated as the swelling pressure. The area is assumed constant, as the corneal button swells predominantly in the y-direction ^5,26 and is assumed to have a negligible in-plane swelling ability (the tangential x-plane). ²⁷  

The hydration was calculated as:  

 

where w_wet = wet weight, w_dry = dry weight.  

The change in volume of the corneal button from swelling to a thickness of t is:  

 

where t is the CCT, $t_{wet}$ is the measured central corneal thickness at the weight $w_{wet}$, $A=\pi r^2=area$ (assumed constant).  

The hydration from Equation 1 then can be expressed as:  

 

where ${\rho }_{H_{2}O}=1mg/ml$.  

Statistics

Statistical analysis was performed in Prism GraphPad and Microsoft Excel, and included paired and unpaired Student's t-tests. Olsen et al. found that the SP follows a straight-line dependence on stromal thickness, when plotted in a double logarithmic scale. ²³ This was expressed as SP(T) = aT^b , where T is stromal thickness, and a and b are constants of the cornea. ²³ Our data were fitted to this model using least squares. 

Swelling pressure as a function of relative CCT (compression) was compared between groups with two-way ANOVA. All results are reported as mean ± SD, unless otherwise stated. In the human corneas, statistical analysis was omitted due to the small number of human corneas available. 

Results

Porcine Full-Thickness Buttons

After the 2-hour preswelling period, a significant difference in wet weight was observed in the CXL group compared to the riboflavin-only group (P < 0.01, Table 1). During CXL treatment, the hydration in the treatment group was reduced significantly compared to the group treated with riboflavin alone (2.8 ± 0.8 vs. 4.3 ± 0.7 mg H₂O/mg dry weight, respectively, P < 0.05, Table 1). The CCT reduction after 60 minutes of the CXL treatment or 60 minutes riboflavin alone was identical in the groups (88.0 ± 23.5 vs. 87.9 ± 49.7 μm, respectively, P = 1.0, unpaired t-test Table 1). In the CXL and riboflavin groups, the greatest reduction in thickness was observed within the first 30 minutes of riboflavin application, followed by a smaller further reduction from 30 to 60 minutes of riboflavin application. The CXL-treated porcine full-thickness corneal buttons demonstrated a decreased swelling force for a given stromal thickness compared to the untreated group (Fig. 3). A significant reduction (P < 0.05) in swelling force in the CXL group was observed at all compressions compared to the untreated group (Fig. 3). The CXL-treated buttons compared to untreated buttons showed a significant reduction in swelling force at +30% (P < 0.001), +20% (P = 0.02), and −5% (P = 0.03) compressions. Interestingly, the riboflavin-treated corneas exhibited significantly higher average swelling pressures at all compressions compared to the untreated corneas (Fig. 3). Plotting swelling pressure versus hydration for individual corneal samples showed a similar trend, and revealed the sample variation within the groups, most pronounced in the riboflavin- and CXL-treated groups (Fig. 4). 

Figure 3

 

Porcine full-thickness corneal buttons. Swelling force as a function of relative compression calculated from initial CCT. Mean ± SD. At all compressions a significant reduction in swelling force was observed in the CXL group compared to the riboflavin-only group (P < 0.05, unpaired t-test). Nonlinear regression lines fitted to SP(T) = aT^b .

Figure 4

 

Swelling pressure as a function of tissue hydration in individual porcine full-thickness buttons.

Table 1. 

 

Porcine Full-Thickness Corneas (Mean ± SD)

ø = 8 mm	CCTstart without Epithelium, μm	CCT, 30 Min, μm	CCT, 60 Min, μm	CCTstart – Epithelium – CCT, 60 Min, μm	Wet Wt after Preswell × 10−3, g	Dry Wt × 10−3, g	Hstart, (No Epithelium) mg H2O/ mg Dry Wt	ΔH after 60 Min, mg H2O/ mg Dry Wt	H after Preswell, mg H2O/ mg Dry Wt
Riboflavin only, n = 8	751 ± 68	685 ± 80	663 ± 68	88 ± 24	86.2 ± 10.2*	9.4 ± 0.9	3.9 ± 0.5†	−4.3 ± 0.7*†	8.2 ± 1.0*
CXL treated, n = 8	772 ± 60	696 ± 40	684 ± 27	87 ± 50	68.6 ± 7.1*	9.1 ± 0.5	3.7 ± 0.3†	−2.8 ± 0.8*†	6.5 ± 0.7*
Untreated, n = 5	762 ± 28	–	–	–	91.0 ± 4.6	9.1 ± 0.5	4.2 ± 0.4	–	8.9 ± 0.8

 CCT, 30 Min, central corneal thickness after 30 minutes of riboflavin treatment; CCT, 60 Min, ce7ntral corneal thickness after 60 minutes of riboflavin treatment (riboflavin group), or after 30 minutes of riboflavin treatment + 30 minutes of riboflavin and UVA radiation (CXL group).

*   P < 0.05.

†  Calculated from Equation 3.

‡   P < 0.001.

Porcine FSL-Cut Corneal Lamellae

No significant difference was observed in the dry weight of the 300 μm lamellae between CXL-treated anterior lamellae compared to lamellae in the riboflavin group (P = 0.41, Table 2). 

Table 2. 

 

Porcine FSL-Cut Lamellae

ø = 8 mm Lamella = 300 μm	CCT Intact. Start, μm	Epithelium, μm	Weight after Preswell × 10−3, g	Dry Wt × 10−3, g	ΔH after Preswell, mg H2O/mg Dry Wt
Anterior riboflavin, n = 7	874 ± 21	105 ± 5	32.0 ± 1.9*	3.3 ± 0.1	5.8 ± 1.6
Posterior riboflavin, n = 7			43.3 ± 2.4†	4.2 ± 0.1	5.4 ± 1.5
Anterior CXL, n = 7	865 ± 21	101 ± 6	18.3 ± 0.5*	3.4 ± 0.1	1.5 ± 0.7
Posterior CXL, n = 7			32.7 ± 1.8†	4.3 ± 0.2	2.9 ± 0.8

 All data are reported as mean ± SD. ΔH calculated from hydrations obtained from Equation 3.

*   P < 0.001.

†   P = 0.0038.

After a 2-hour preswelling period in isotonic saline, a significantly increased weight was found in the anterior lamellae in the riboflavin group compared to the corresponding CXL group (P < 0.001, Table 2). The posterior lamellae did not differ significantly in dry weight (P = 0.57, Table 2). The posterior CXL group imbibed significantly less saline compared to the corresponding riboflavin group (P < 0.005, Table 2). In both anterior and posterior CXL groups, a significant difference was observed when compared to the respective riboflavin-treated groups (Fig. 5). 

Figure 5. 

 

Porcine FSL-cut 300 μm lamellae. Relative compression versus swelling force. (Mean, SD). Nonlinear regression lines, SP(T) = aT^b . In the anterior lamellae, a significant difference in swelling force was observed between the CXL and riboflavin groups (two-way ANOVA, P < 0.001). In the posterior lamellae a significant difference in swelling force was likewise observed between the CXL and riboflavin groups (two-way ANOVA, P = 0.005).

Individual plots of swelling force versus hydration in the riboflavin-only and CXL groups show greatly diminished swelling pressures in the anterior CXL-treated samples compared to all other groups (Fig. 6). The posterior samples exhibited a pronounced variation within groups, but appeared to have higher swelling pressures at comparable hydrations in both groups (Fig. 6). 

Figure 6. 

 

Porcine FSL-cut 300 μm individual lamellae. Individual plots of swelling pressure versus hydration in the riboflavin- and CXL-treated groups show reduced swelling pressure in CXL treated anterior samples.

Human FSL-Cut Corneal Lamellae

Mean CCT in the intact donor tissue was 644 μm, range 585 to 696 μm (n = 3), and 681 μm, range 613 to 776 μm (n = 3) in the riboflavin and CXL groups, respectively. 

The dry weights of the human corneal lamellae differed, as the lamellar thicknesses were different (range 260–300 μm). For each cornea, the lamellar thickness was estimated as the total ultrasound pachymetry reading (CCT intact, start − epithelial thickness) divided by 2 (Table 3). 

Table 3. 

 

Human FSL-Cut Lamellae

ø = 6.5 mm	CCT Intact, Start – Epithelium, μm	Wet Weight after Preswell × 10−3, g	Dry Wt × 10−3, g	ΔH after Preswell, mg H2O/mg Dry Wt
Anterior riboflavin, n = 3	573 ± 28	19.4 ± 0.6	2.7 ± 0.1	5.4 ± 0.7
Posterior riboflavin, n = 3		20.0 ± 0.3	2.2 ± 0.2	7.5 ± 1.3
Anterior CXL, n = 3	636 ± 39	13.5 ± 1.4	2.3 ± 0.1	4.1 ± 0.8
Posterior CXL, n = 3		15.8 ± 2.4	2.1 ± 0.1	5.7 ± 1.2

 All data are reported as mean ± SD. Significant differences in dry weights may be attributed to differences in FSL-cut thickness (lamella thickness = [CCT start − epithelium thickness]/2). ΔH = hydration change (using wet weights calculated from Equation 3).

In the human FSL-cut lamellae, a pronounced reduction in swelling force in the anterior lamellae was observed in the CXL-treated lamellae compared to the riboflavin-treated lamellae, as expected. However, the swelling force was greater in the anterior lamellae compared to posterior groups. No difference was observed between the posterior groups (Figs. 7, 8). 

Figure 7. 

 

Human FSL-lamellae. Swelling force versus relative compression in the CXL-treated groups (open dots) and riboflavin-treated groups (solid dots) with exponential nonlinear regressions. Dotted lines: CXL treated. Solid lines: riboflavin only.

Figure 8

 

Human FSL-cut lamellae. Swelling force versus hydration in anterior and posterior riboflavin-only and CXL-treated groups.

Histology

Histologic evaluation confirmed complete removal of epithelium and endothelium in all three porcine corneas. 

Discussion

The presented data demonstrated the effects of CXL on corneal hydration and swelling properties in porcine and human eyes. We hypothesized that CXL reduces corneal swelling pressure, and thereby reduces edema and potentially improves vision. Thus, in the clinical setting, patients suffering from stromal edema may benefit from CXL. 

The corneal stroma has an innate tendency to swell when it has free access to water. Corneal hydration and swelling behavior are important in biomechanical testing, including stress–strain measurements, indentation and inflation experiments. As the clinical cross-linking procedure affects corneal biomechanics, ^28,29 we expected to see an effect on swelling properties. ³⁰ We expected this effect to be most pronounced in the anterior stromal segment, as previous experiments have described enzymatic, ³¹ biomechanical, ²⁸ histologic, ^30,32 and swelling ³⁰ changes to be greatest anteriorly. 

The CCT was measured in the center of the cornea by ultrasound pachymetry. This is an estimate for the global cornea thickness, as thickness varies over the corneal surface and increases towards the periphery. In corneal edema, CCT measurement by ultrasound might overestimate corneal thickness. Silverman et al. conclude that a 20% increase in CCT would result in a 7 μm overestimate in ultrasound pachymetry due to a decrease in sound velocity. ³³ This decrease may be attributed to backscatter, as stromal edema increases backscatter as a result of increasing disorganization of collagen fibrils. 

The CCT in vivo was not measured. The CCT measured in the intact porcine eye within 2 hours postmortem was used as the initial, normohydrated value to minimize sample variation. ³⁴ As all testing protocols in this study depend on this initial value, a dehydrated or overhydrated cornea at the time of CCT measurement will produce different results. The cornea may swell ex vivo or dry during handling and dissection, although dissection and preparation times were short. To correct for this, dry weights were obtained and used in the analysis. An additional freeze-drying period of another 24 hours did not reduce the dry weight significantly (paired t-tests: P = 0.2080, n = 5 in the riboflavin group and P = 0.0705, n = 5 in the CXL group). 

The cornea swells in only one direction (orthogonal to its faces) in free solution. After it contacts the rigid plates of the pressure transducer, however, it spreads slightly laterally. Corneal SP thereby may be overestimated because the force recorded on the transducer is divided by the original area of the tissue, ⁵ that is, a lateral spreading of 0.5 mm would overestimate the SP approximately 20% at 0% compression. 

The placement of the corneal button should not affect swelling behavior and force readings significantly, as the fixation unit is made of a rigid material and the compression plates are kept parallel. The load cells are zeroed before each test and a calibration routine is run in the software before every test to subtract system forces from the measured forces. 

The effect of CXL depends primarily on the concentration of the riboflavin in the stroma, ^25,35–37 available UVA light, and the availability of oxygen. ³⁸ The riboflavin solution used was examined with absorption and emission spectra as described previously, ²⁵ which agree with literature results. 

In human and porcine corneal stroma we found a diminished tendency to swell after the CXL procedure. The CXL-treated tissue imbibed less saline after a soaking period in saline compared to the control groups (untreated or treated with riboflavin solution alone). From previous studies on corneal biomechanics after CXL, the treatment affects primarily the anterior 200 μm, ^25,37,39 which agrees with our findings. The SP is a function of stromal thickness, the ion-strength of the medium, and the temperature. ^18,21 These parameters were kept constant and at physiologic values in the Biotester bath, changing only the stromal thickness. The exponential regression describing the relation between SP and stromal thickness in this study agrees with the findings of Olsen et al. in human corneas. ²³  

In human full-thickness corneas, Olsen et al. found a swelling pressure of 85 mm Hg at a standard stromal thickness of 500 μm. ²³ In our study, the swelling pressure in the full-thickness porcine corneas at 0% compression, at 760 μm is approximately 45 mm Hg. This is lower than the reported human swelling pressure in the literature. ^23,40 The in vivo CCT in porcine corneas was not measured. The correct thickness at 0% compression, therefore, may be significantly lower (and the resulting SP higher), as the corneas may have swelled in the transport medium post mortem. Swelling pressure values in porcine tissue, however, have not been reported previously by other investigators to our knowledge. 

Swelling pressure in isolated corneal layers, to our knowledge, has been described only in a recent study by Petsche et al. ⁴⁰ Compared to their work, the SP in the FSL-cut human corneal anterior and posterior segments at 0% axial strain (0% compression in this study) have similar values of approximately 1.2 to 2.5 kPa (11–19 mm Hg). 

The SP in full-thickness porcine corneas was reduced approximately from 58 mm Hg (riboflavin group) to 34 mm Hg (riboflavin group) at 0% compression (Fig. 3). 

Plots of SP versus tissue hydration for individual full-thickness porcine samples (Fig. 4) showed increased variation in the riboflavin- and CXL-treated groups compared to untreated samples. This may be attributed to differences in tissue response to the riboflavin solution and/or CXL treatment. 

The interesting finding that the SP is increased in porcine riboflavin–only-treated buttons, compared to the untreated corneal buttons (Fig. 3), could be an effect of dextran entering the stroma. When the corneal button is placed in aqueous medium, this would lead to an initial swelling due to the slight increase in osmolarity of the stromal fluid. ⁴¹ It is unlikely that the low ambient light gave rise to any cross-linking in the riboflavin-only group. 

The surprising finding that the posterior human lamellae exerted a lower swelling pressure than the anterior lamellae could be explained by hydration differences in the corneal stroma at the time of laser surgery. Postmortem, the tissue swells somewhat in the transport medium (thickness range was 585–776 μm). However, the displacement of fluid within the cornea postmortem is unknown. The dry-weights of the anterior lamellae were slightly greater than the posterior lamellae, which could explain why the swelling ability was greatest anteriorly. This indicates that the posterior stroma could have contained more water than the anterior stroma and, therefore, the cut lamellae would differ in dry weight, when cut into two lamellae of same thickness. However, dry weight–corrected plots of swelling force versus hydration confirm the greater swelling in anterior lamellae (Fig. 8). The composition of the Optisol GS transport medium, which includes dextran and chondroitin sulfate, might affect hydration dynamics in the corneal tissue and give rise to these unexpected results. The human corneas were stored up to three weeks until used, and the epithelium may have been damaged or missing. During Optisol storage, the presence of an intact epithelial layer is required for maintaining hydration and sodium levels, ⁴² which may explain the observed edema. 

Edema may have led to detection of a lower SP at 0% compression than the actual. The clinically important aim in edematous corneas, however, is to lower the SP with CXL. Thus, the test results reflect the CXL treatment of edematous corneas. 

The mechanism of CXL is not yet fully elucidated, and we speculate that effects of CXL in vivo in part may be due to healing processes or inflammatory response, which might increase or decrease over time. The present ex vivo study presents immediate effects on hydration properties. In future studies, the swelling studies should be repeated, that is, in porcine eyes, after three or six months after CXL treatment to assess treatment durability. 

As pointed out by Doughty et al., significant errors may be associated with sample selection, and handling for this type of research and intrinsic sample variation is very important. ⁴³ Thus, the hydration of samples in this study (as mg H₂O/mg dry mass) was done only to one decimal place, due to variation in sample mass and error. 

Empirically, patients suffering from early-stage Fuchs dystrophy have an effect from CXL treatment on diurnal hydration fluctuations. ¹² Further, in bullous keratopathy, a benefit from CXL has been described. ⁴⁴ The present observations indicated that CXL affects SP and supported the clinical findings and that corneal edema may be managed by CXL clinically. ¹³  

However, as the disease progresses, guttae appear and are not treated in this procedure, and would require further surgery, for example, posterior lamellar keratoplasty. Therefore, treatment may not change the outcome of the progressive disease, but may relieve pain, increase comfort, and improve vision. 

Further studies should be conducted to assess how the clinical protocol could be adapted to optimize treatment in these patients. Deswelling the edematous cornea before CXL, either with glycerol 70% ¹² or glucose 40%, ⁴⁴ has been reported clinically to be effective in edema management. Staged CXL with femtosecond laser facilitated intrastromal riboflavin administration was introduced recently ^45,46 as an alternative method for managing bullous keratopathy. Although not fully studied, the investigators found the method to be a safe and temporizing alternative. Questions as how much stroma must be treated to be effective in edema management and if deswelling before CXL increases the success rate remain. 

Acknowledgments

The authors thank Carl Christian Danielsen and Eva Mikkelsen at the Department of Anatomy at Aarhus University for assistance in freeze-drying and histology of the corneal samples. 

Footnotes

 Supported by The Danish Eye Health Society, Jørgen Bagenkop Nielsens Myopia Foundation, and Maskinfabrikant Jochum Jensen og hustru Mette Marie Jensen f. Poulsens Mindelegat (Fond).

Footnotes

 Disclosure: A.P. Søndergaard, Carl Zeiss Meditec AG (F); A. Ivarsen, None; J. Hjortdal, None