Orthokeratology (OK) is a clinical contact lens–based technique that is used to intentionally manipulate corneal shape and reduce refractive error with specially designed rigid contact lenses. Overnight wear of OK lenses is currently the modality of choice for most clinicians who use OK for the temporary correction of low to moderate myopia. The clinical efficacy of overnight OK lens wear has now been proven in many studies.1–7 The main variables of interest in studies of overnight OK lens wear are the amount of corrected refractive error, the achieved unaided visual acuity (VA), the change in the shape of the anterior cornea, and the change in corneal thickness. However, the effect of OK on the shape of the posterior cornea is still controversial.
There are two main hypotheses concerning the effects of OK on posterior corneal shape. The first hypothesis is that the OK refractive effect is achieved primarily through remodeling of the anterior corneal layers. Alharbi and Swarbrick3 achieved a good correlation between measured refractive change and predicted changes in refractive error based on corneal thickness changes using the formula of Munnerlyn et al.8 after 3 months of overnight OK. Because the formula of Munnerlyn et al. is based on anterior corneal parameters and assumes no change in posterior corneal shape, they concluded that the refractive effect of OK can be explained by anterior corneal changes alone and specifically by topographical changes in corneal epithelial thickness. According to this analysis, there was no need to postulate corneal bending or any change in the shape of the posterior cornea to explain the refractive effects of OK. Other researchers have since reported no long-term changes in posterior corneal shape during OK using the Pentacam.9,10 On the other hand, using Purkinje images, Owens et al.11 reported that the posterior corneal curvature flattened during the first week of OK lens wear before returning to baseline curvature. This finding implied that there was transient overall bending of the cornea, suggesting stromal involvement in the response.
Understanding any change in the shape of the posterior cornea during myopic overnight OK will provide a more complete picture of overall corneal changes during myopic OK and, in particular, the contributions of epithelium and stroma to the corneal shape change induced by overnight OK.
The aim of this study was to evaluate changes over time in the shape of the posterior cornea at the end of the day during short-term wear of myopic OK in comparison with a rigid gas-permeable (GP) lens–wearing control group based on measurements of anterior corneal topography and corneal thickness across the horizontal corneal meridian.
Subjects in the OK group were asked to wear dispensed OK lenses for a minimum 6 hours per night during a 2-week period. The data for the OK group were collected approximately 8 to 10 hours after eye opening at baseline and 8 to 10 hours after OK lenses were removed on days 1, 4, 7, and 14. Subjects in the conventional GP group (control group) were asked to wear conventional GP lenses overnight for a minimum of 6 hours for one night. The data for the GP group were collected approximately 8 to 10 hours after eye opening at baseline and 8 to 10 hours after GP lenses were removed on day 1.
Eighteen subjects (11 women and 7 men) aged 25.1 ± 3.7 years (mean ± SD; range, 19 to 32 years) agreed to participate in this study, wearing OK lenses overnight once the risks and benefits of OK contact lens wear had been explained. All subjects were low to moderate myopes and had with-the-rule refractive and corneal astigmatism of 1.50 diopters (D) or less. Baseline spherical equivalent refractive error was −2.64 ± 0.99 D.
Ten subjects (8 women and 2 men) aged 24.2 ± 3.9 years (mean ± SD; range, 19 to 32 years) agreed to participate in this research as control subjects wearing conventional GP contact lenses in their right eye (RE) only. All subjects had with-the-rule refractive and corneal astigmatism of 1.50 D or less. Myopic refractive error was not a necessary inclusion criterion because lenses were only to be worn for one night. Baseline spherical equivalent refractive error in this group was −2.38 ± 2.13 D.
This study was conducted in compliance with the Declaration of Helsinki, and informed written consent was gained from subjects before their participation.
BE lenses (Capricornia Contact Lens Pty Ltd, Queensland, Australia) were used in this study. BE lenses are a five-zone reverse geometry design lens, which means that the lenses have a secondary curve steeper than the back optic zone radius. The back optic zone diameter varies between 6.00 and 7.00 mm, which allows a wider treatment zone. BE lenses are made of Boston XO material (hexafocon A), with a nominal total diameter of 11.0 mm and center thickness of 0.22 mm, giving a nominal oxygen transmissibility (Dk/t) of 45.5 × 10−9 (cm/s) (mLO2/mL.mmHg).
Conventional GP Lenses
J-Contour conventional GP lenses (Capricornia Contact Lens Pty Ltd) were fitted on the REs of the control group. J-Contour lenses have an aspheric design and are fitted 0.50 D flatter than the flattest K to achieve an alignment fit. A trial set of 11 aspheric design lenses made of Boston XO material (hexafocon A) was used. The nominal center thickness of the J-Contour lenses is 0.20 mm, giving a nominal Dk/t of 50 × 10−9 (cm/s) (mLO2/mL.mmHg).
Measurements and Calculations
Unaided VA and Refraction
Standard subjective refraction without cycloplegia was conducted at every visit on all subjects. Unaided VA was measured using a LogMAR (logarithm of minimum angle of resolution) chart at 6 m. All VA recordings were conducted in the same room at an illuminance of 456 lux and a chart luminance of 120 cd/m2. Spherical equivalent refraction (sphere + [cylinder/2]) was used for analysis.
Anterior Corneal Topography
Corneal topography was measured with a Medmont E-300 corneal topographer (Medmont International Pty Ltd, Victoria, Australia). The corneal topographic raw data, which included the radial distance data (in millimeters) and corneal height (in millimeters) across the horizontal meridian, were exported from the Medmont E-300 corneal topographer. An ellipse curve for the anterior cornea across the horizontal meridian was calculated from these data using an in-house Interactive Data Language (IDL) computer program (student version 5.0; Research Systems Inc, Boulder CO).
This method for calculating the anterior ellipse curve for the cornea has previously been validated using test surfaces and on human eyes.12 The mean absolute accuracy (±SD) of calculated anterior apical radius and asphericity (Q) of 10 polymethyl methacrylate lenses were 0.009 ± 0.011 mm (95% confidence interval [CI], −0.013 to 0.031 mm) and 0.01 ± 0.01 (95% CI, −0.01 to 0.04), respectively. The mean absolute repeatability (±SD) of the calculated anterior apical radius and Q of five human eyes were 0.02 ± 0.01 mm (95% CI, 0.00 to 0.04 mm) and 0.04 ± 0.03 (95% CI, −0.02 to 0.10), respectively. The repeatability of our technique on human eyes is consistent with the repeatability of apical radius and eccentricity data taken directly from the Medmont topographer as reported by Cho et al.13
Total corneal thickness was measured using the Holden-Payor optical pachometer across the horizontal meridian. The total corneal thickness was measured for the RE only at predetermined corneal locations from the periphery of the nasal cornea to the periphery of the temporal cornea by directing subject fixation to a series of light-emitting diodes (LEDs) as described in Table 1. The equation of Brennan et al.14 was used to estimate the locations on the anterior corneal surface of the thickness measurements for each LED. For this estimation, the apical radius of curvature and Q were assumed to be 7.8 mm and −0.25, respectively.
Five repeated measurements of corneal thickness were taken for each measurement point, and the mean of three measurements was recorded after excluding the maximum and minimum values.
Posterior Corneal Topography
The coordinates of the posterior cornea for the horizontal meridian can be calculated based on the anterior corneal ellipse curve, the positions of corneal optical pachometry measurements, the direction of thickness measurements that are orthogonal to the tangent at the points of measurement on the anterior corneal surface,14 and the real corneal thickness at each position of optical pachometry measurement. Using the posterior corneal coordinates, an ellipse curve for the posterior cornea across the horizontal meridian was calculated using the IDL computer program. This method is a similar method to that proposed previously by Rivett and Ho.15 However, in this new approach, the local radius of curvature at each measurement point and the exact solution for the relationship between real and apparent corneal thickness, as suggested by Brennan et al.,14 were used to more accurately calculate real thickness.
This method for calculating the posterior ellipse curve for the cornea has previously been validated using test lenses and on human eyes.12 The mean absolute accuracy (±SD) of calculated posterior apical radius and Q of 10 polymethyl methacrylate lenses were 0.053 ± 0.044 mm (95% CI, −0.033 to 0.139) and 0.10 ± 0.10 (95% CI −0.10 to 0.31), respectively. The mean absolute repeatability (±SD) of the calculated posterior apical radius and Q of five human eyes were 0.07 ± 0.06 mm (95% CI, −0.05 to 0.19) and 0.09 ± 0.07 (95% CI, −0.05 to 0.23), respectively. The repeatability of posterior corneal measurements using our technique is comparable to that found with the Orbscan16 and of a similar order of magnitude but slightly poorer than repeatability of posterior surface measurements obtained using the Pentacam as reported in previous studies.17,18
The SPSS statistical package (version 14.0) was used to analyze the data collected. Repeated-measures analysis of variance (RM ANOVA) and post hoc pairwise comparisons were used to compare corneal thickness between OK and GP groups at baseline.
Repeated-measures ANOVA was used to examine overall changes from baseline in refractive error, VA, anterior corneal topography, corneal thickness, and posterior corneal topography during the study period in the OK and GP groups. Changes from baseline in these variables were also examined using post hoc paired Student t tests with Bonferroni correction.
A critical value of p = 0.05 was chosen to denote statistical significance for all analyses.
The results reported in this article are for REs only. All data are presented as mean ± SD unless otherwise indicated.
Fig. 1 shows no statistically significant change in VA in the control group after one night of GP lens wear (p > 0.999, paired t test) and a statistically significant improvement in unaided VA in the OK lens–wearing group during 14 days of overnight OK (p < 0.001, RM ANOVA). The improvement in unaided VA in the OK lens–wearing group was statistically significant by day 1 (p < 0.001, paired t test).
Fig. 2 demonstrates no statistically significant change in refractive error in the control group after one night of GP lens wear (p > 0.999, paired t test) and a statistically significant reduction in myopia in the OK lens–wearing group during the 14 days (p < 0.001, RM ANOVA). The reduction in myopia in the OK lens–wearing group was statistically significant by day 1 (p < 0.001, paired t test).
Anterior Corneal Shape
Anterior Corneal Apical Radius
As shown in Fig. 3, there was no statistically significant change in anterior corneal apical radius in the control group after one night of GP lens wear (p = 0.472, paired t test). However, there was a statistically significant flattening of the anterior corneal apical radius in the OK lens–wearing group during the 14 days of overnight OK (p < 0.001, RM ANOVA). The flattening of the anterior corneal apical radius in the OK lens–wearing group was statistically significant by day 1 (p < 0.001, paired t test).
Anterior Corneal Q
Fig. 4 presents the anterior corneal Q over an approximate 9-mm chord (OK group, 9.21 ± 0.28 mm; GP group, 9.18 ± 0.31 mm) during the study.
As shown in Fig. 4, there were no statistically significant changes in anterior corneal Q in the control group after one night of GP lens wear (p > 0.999, paired t test). However, there was a statistically significant increase toward oblate in anterior corneal Q in the OK lens–wearing group during the 14 days of overnight OK (p < 0.001, RM ANOVA). The changes in anterior corneal Q in the OK lens–wearing group were statistically significant by day 4 (p < 0.001, paired t test).
As shown in Table 2, using RM ANOVA, with location as the within-subject factor and group as the between-group factor, there was a significant effect for location (F10,17 = 106.796, p < 0.001). There was no interaction between location and group (F10,17 = 2.228, p = 0.070). There was no statistically significant difference in corneal thicknesses between OK and GP groups (p = 0.190, pairwise comparisons) at baseline.
There were no statistically significant changes in corneal thickness at any corneal location compared with those at baseline (all p > 0.741, paired t tests). The changes in total corneal thickness across the horizontal meridian ranged from −2.1 ± 3.9 μm to +1.9 ± 6.4 μm.
OK Lens–Wearing Group
Fig. 5 shows the changes in total corneal thickness across the horizontal meridian in OK lens–wearing eyes during the 14-day lens-wearing period. A statistically significant reduction in corneal thickness was found at central locations (1.1 N, 0.0, and 1.1 T; all p < 0.001, RM ANOVA). The average thinning of the central cornea on days 1, 4, 7, and 14 was −3.3 ± 8.2 μm, −7.9 ± 7.4 μm, −10.4 ± 6.8 μm, and −12.8 ± 6.2 μm, respectively. Corneal thinning at 1.1 N, 0.0, and 1.1 T was statistically significant (p = 0.027, p = 0.001, and p < 0.001, respectively, paired t tests) by days 4, 4, and 7, respectively.
There were no statistically significant changes in thickness at the mid-periphery of the cornea (3.2 N, 2.1 N, and 3.2 T) during the 14 days of overnight OK (all p > 0.603, RM ANOVA). However, there was a reduction in thickness at 2.1 T (p < 0.001, RM ANOVA), which was statistically significant on day 14 only (p = 0.019, paired t test). The average changes in corneal thickness at the mid-periphery on days 1, 4, 7, and 14 were 0.0 ± 7.9 μm, −1.8 ± 9.2 μm, 0.1 ± 8.7 μm, and −0.7 ± 9.6 μm, respectively.
At the peripheral cornea, there were no statistically significant changes in thickness at 4.7 N, 3.7 N, and 4.6 T during the 14 days of overnight OK (all p > 0.202, RM ANOVA). However, there was an increase in thickness at 3.7 T (p < 0.001, RM ANOVA), which reached statistical significance only on day 7 (p = 0.027, paired t test). The changes in total thickness of the peripheral cornea on days 1, 4, 7, and 14 averaged −3.1 ± 9.8 μm, −3.0 ± 14.1 μm, −0.9 ± 11.9 μm, and −2.0 ± 13.9 μm, respectively.
These data were also analyzed using a fourth-order polynomial fit. There was a statistically significant fourth-order polynomial shape change from baseline in total corneal thickness during the 14 days of overnight OK (p < 0.001, RM ANOVA). The changes from baseline were statistically significant by day 7 (p = 0.002, paired t test).
Posterior Corneal Shape
Posterior Corneal Apical Radius
Fig. 6 presents the posterior corneal apical radius of curvature for control and OK lens–wearing subjects. There were no changes in posterior corneal apical radius in the control group after one night of GP lens wear (p > 0.999, paired t test). Changes in posterior corneal apical radius in the OK lens–wearing group during 14 days of overnight OK also failed to reach statistical significance (p = 0.091, RM ANOVA).
Posterior Corneal Q
Fig. 7 presents the posterior corneal Q over an approximate 8.5-mm chord (OK group, 8.41 ± 0.31 mm; GP group, 8.56 ± 0.31 mm) during the study.
After one night of GP lens wear, there were no statistically significant changes in posterior corneal Q in the control group (p > 0.999, paired t tests). However, there was a statistically significant increase toward oblate in posterior corneal Q in the OK lens–wearing group during the 14 days of overnight OK (p < 0.001, RM ANOVA). The changes in posterior corneal Q were statistically significant on days 4 (p = 0.007, paired t test) and 7 (p = 0.002, paired t test).
The results reported in this article show a significant reduction in myopia from −2.64 ± 0.99 D to −0.39 ± 0.49 D during 14 days of overnight OK lens wear. These results also demonstrated that most of the changes occur within the first 7 nights of lens wear. This fast reduction in myopia during the first 7 days of OK lens wear has been reported in many previous studies.1–5 For example, Soni et al.5 reported that full corneal and visual changes were achieved and were stable for all waking hours after 1 week of OK.
The change in refractive error was also associated with an improvement in unaided VA. In this study, the improvement in unaided VA was statistically significant by day 1 for the evening measurements (p < 0.001, paired t test). After 14 days of overnight OK, unaided logMAR VA averaged 0.11 ± 0.15 (Snellen acuity of approximately 6/7.5 or 20/25) 8 to 10 hours after OK lenses were removed. In the study by Soni et al.,5 mean unaided VA improved significantly from a mean of 0.49 log MAR units (20/60) at baseline to −0.02 log MAR units (20/20) after just one night of overnight OK lens wear. Similar rapid improvements in unaided VA with OK have been reported in many previous studies.2–5,19–21
The amount of flattening of anterior corneal apical radius on days 1, 4, 7, and 14 of overnight OK lens wear were 0.09, 0.22, 0.27, and 0.29 mm, respectively. Thus, 78% of the change in anterior corneal apical radius during the 14 days of overnight OK lens wear was achieved in the first 4 days. The magnitudes of the increase toward oblate in anterior corneal Q for days 1, 4, 7, and 14 were 0.11, 0.31, 0.37, and 0.37, respectively. Thus, 83% of the change in anterior corneal Q was also achieved in the first 4 days of overnight OK lens wear. These rapid changes in anterior corneal apical radius and Q have been reported in other studies2–7,19–22 and are associated with the rapid reduction in myopia and improvement in unaided VA and subsequent stabilization of the OK effect.
In this study, when the changes in the total corneal thickness of 1.1 N, 0.0, and 1.1 T were averaged, we found −12.8 ± 6.2 μm central corneal thinning during 14 days of overnight OK lens wear. Swarbrick et al.23 were the first to report central corneal thinning with OK lenses worn in the open eye. The results of Alharbi and Swarbrick3 confirmed that central corneal thinning (−19.0 ± 2.6 μm) occurred after 3 months of OK lens wear. In addition, using optical pachometry, Alharbi and Swarbrick3 found that the central corneal thinning that occurred during overnight OK was epithelial in origin. Nichols et al.2 also reported −12.0 ± 11.0 μm central corneal thinning using the Orbscan after 60 days of OK lens wear, where a reduction in myopia of 1.83 ± 1.23 D was found. This central corneal thinning has also been confirmed by Soni et al.5 using confocal microscopy and Haque et al.24 using optical coherence tomography.
In this study, when the changes in the total corneal thickness of 3.2 N, 2.1 N, 2.1 T, and 3.2 T were averaged, there was −0.7 ± 9.6 μm mid-peripheral corneal thinning during 14 days of overnight OK lens wear. This change did not reach statistical significance. This absence of a change in mid-peripheral corneal thickness was also reported by Nichols et al.,2 who used the Orbscan to measure corneal thickness. However, Alharbi and Swarbrick3 found significant mid-peripheral thickening (3.50 mm nasal and 3.25 mm temporal) after 4 days of OK lens wear and demonstrated that this was primarily stromal in origin.
The differences in the changes in central and mid-peripheral corneal thickness between studies that used optical pachometry (Alharbi and Swarbrick3 and this study) may be caused by differences in the magnitude of myopic reduction achieved. Alharbi and Swarbrick3 reported a reduction in myopia of −2.63 ± 0.57 D after 90 days of OK lens wear, whereas the reduction in myopia in this study was −2.25 ± 0.72 D after 14 days of overnight OK lens wear. However, a more likely explanation for the differences in corneal thickness changes between these studies may be that the calculation methods that were applied to the optical pachometry results in the two studies were not the same. In the study of Alharbi and Swarbrick,3 a standard value for the local radius of curvature (7.8 mm) was assumed and an approximate solution for the relationship between real and apparent thicknesses was used. However, in this study, the calculation of the real corneal thickness took into consideration the local radius of curvature at the measurement point and the exact solution for real thickness from apparent thickness. The method used in the current study is likely to be a more valid method than that used by Alharbi and Swarbrick3 because local radii of curvature are variable across the anterior cornea, between subjects, and before and after OK treatment.
In this study, there was greater corneal thinning at 2.1 T compared with that at 2.1 N and greater corneal thickening at 3.7 T compared with that at 3.7 N. This phenomenon could be explained by OK lens temporal decentration, which was confirmed by inspection of subtractive tangential maps from the Medmont corneal topographer. In addition, 2.1 T is relatively closer to the lens center than 2.1 N and thus would be located in the relatively thinner central zone of the cornea. Similarly, because 3.7 T is closer to the middle of the reverse curve zone compared with 3.7 N, this would place 3.7 T in the area of the cornea where there was localized relative corneal thickening.
The results of this study show that there were no statistically significant changes in posterior corneal apical radius of curvature during 14 days of overnight OK. This confirms earlier modeling of corneal effects of OK3 based on the formula of Munnerlyn et al.8 and longer term results using the Pentacam.9,10
The results of this study identified a statistically significant increase in posterior corneal Q, tending toward a more oblate shape on days 4 and 7. This unexpected transient shape change in the posterior cornea could result from cumulative minor errors associated with the calculation of posterior corneal shape that, in turn, was based on calculations of the anterior corneal ellipse curve and measurements of topographic corneal thickness. Alternatively, there may have been some subtle peripheral corneal thickness change that was not able to be detected by our methods. This slight, but interesting, transient change in posterior corneal shape warrants further investigation. By day 14, this change was no longer statistically significant. Further research over a longer period is needed to determine if this apparent change in Q continued to recover to baseline levels.
In conclusion, the results of this study confirm previous studies that have reported rapid onset of refractive and anterior topographic effects of overnight OK. This study also confirms previous reports of central corneal thinning, but we were unable to detect any mid-peripheral corneal thickening as previously reported by Alharbi and Swarbrick.3 Most importantly, this study demonstrates that overnight OK lens wear does not cause any change in the central posterior corneal radius of curvature at least in the first 2 weeks of lens wear. The modest and transient change in posterior corneal Q within the first week of lens wear toward a more oblate shape implicates changes in the peripheral cornea as contributing to the corneal response to overnight OK lens wear. Overall, the results of this study support the current hypothesis that the OK refractive effect is achieved primarily through remodeling of the anterior corneal layers, without overall corneal bending.
Jeong Ho Yoon
School of Optometry
University of Choonhae Health Science
Ungchon-myeon Ulju-gun, Ulsan
Republic of Korea
The authors thank Dr Chitralekha Avudainayagam and Dr Kodikullam Avudainayagam for their excellent advice in developing the IDL computer program. The authors thank the Korean College of Behavioural Optometry for financial support (J.H.Y.), Capricornia Contact Lens Pty Ltd for donation of study lenses, and Bausch & Lomb (Australia) for supply of contact lens solutions. The authors also thank Dr Pauline Kang for graphics support.
This research was supported in part by the Australian Research Council Linkage Project Scheme with support from industry partners Polymer Technology Corporation, a Bausch & Lomb Company, BE Enterprises Pty Ltd, and Capricornia Contact Lens Pty Ltd.
The methods and results described in this article have been previously reported at the Seventh Congress of the Orthokeratology Society of Oceania (2008), Gold Coast, Australia, and the 2008 Annual Meeting of the Association for Research in Vision and Ophthalmology, Ft Lauderdale, Fla.
Received July 2, 2012; accepted November 13, 2012.