Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=icey20

Current Eye Research

ISSN: 0271-3683 (Print) 1460-2202 (Online) Journal homepage: https://www.tandfonline.com/loi/icey20

Predictive Role of Paracentral Corneal ToricityUsing Elevation Data for Treatment ZoneDecentration During Orthokeratology

Zhouyue Li, Dongmei Cui, Wen Long, Yin Hu, Liying He & Xiao Yang

To cite this article: Zhouyue Li, Dongmei Cui, Wen Long, Yin Hu, Liying He & Xiao Yang(2018) Predictive Role of Paracentral Corneal Toricity Using Elevation Data for TreatmentZone Decentration During Orthokeratology, Current Eye Research, 43:9, 1083-1089, DOI:10.1080/02713683.2018.1481516

To link to this article: https://doi.org/10.1080/02713683.2018.1481516

Accepted author version posted online: 26May 2018.Published online: 17 Jul 2018.

Submit your article to this journal

Article views: 84

View Crossmark data

Predictive Role of Paracentral Corneal Toricity Using Elevation Data for TreatmentZone Decentration During OrthokeratologyZhouyue Li, Dongmei Cui, Wen Long, Yin Hu, Liying He, and Xiao Yang

State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China

ABSTRACTPurpose: To investigate the influence of paracentral corneal toricity using elevation data on thetreatment zone decentration of spherical and toric orthokeratology (Ortho-k) lens.Methods: Corneal elevation difference (CED) was defined as the difference of corneal elevation betweenthe two principle meridians at 8-mm chord, representing the paracentral corneal toricity. Seventy-fivesubjects included in this prospective study were divided into a low CED (LCED) group (LCED<30μm,n = 25) and a high CED (HCED) group (HCED≥30μm, n = 50). All subjects in the LCED group and 25subjects in the HCED group (HCED I) were fitted with spherical Ortho-k; the other 25 subjects in theHCED group (HCED II) were fitted with toric Ortho-k. Corneal topography data from the right eyes wereobtained at baseline and after 1 month of lens wear. The amount and direction of treatment zonedecentration among the three groups were compared, and their relationships with corneal shapeparameters, including central and paracentral corneal toricity, corneal asymmetry, flat-k and eccentricity,and lens diameter were analyzed using univariable and multivariate linear regression models.Results: The magnitude of treatment zone decentration was the greatest in the HCED I group ((LCED vs.HCED I vs. HCED II: 0.47 ± 0.15mm vs. 0.73 ± 0.15mm vs. 0.47 ± 0.19mm, respectively; ANOVA, p < 0.01).Among participants fitted with spherical Ortho-k, the magnitude of treatment zone decentration wassignificantly correlated to paracentral CED after adjusting for the other corneal parameters and lensdiameter (standard β = 0.599, p < 0.01). No significant correlation between these parameters was foundamong those fitted with toric Ortho-k.Conclusions: Eyes with greater paracentral CED tend to have increased decentration of spherical Ortho-k lens, whereas toric Ortho-k appears to reduce the amount of lens decentration in eyes with CED at 8-mm chord above 30 μm.

ARTICLE HISTORYReceived 10 January 2018Revised 27 March 2018Accepted 22 May 2018

KEYWORDSOrthokeratology;decentration; paracentralcorneal toricity; cornealelevation difference;treatment zone

Introduction

Orthokeratology (Ortho-k), with a reverse-geometry design,has been shown effective in improving unaided visual acuityduring daytime after removal of the lens.1,2 Recent studies havedemonstrated the efficacy of Ortho-k in slowing the axialelongation in myopic children by approximately 50% comparedwith single-vision spectacles or soft contact lens.3–5 However,despite successful lens fitting, decentration of the Ortho-k lenstreatment zone is common,6–8 potentially causing visual dis-turbance due to the induction of astigmatism,9 higher-orderaberrations and reduction in contrast sensitivity.10,11

The exact mechanism of treatment zone decentration is stillunclear, but appears to be multifactorial. Greater baseline cor-neal toricity was recently reported to cause a higher amount oftreatment zone decentration after single overnight use of sphe-rical Ortho-k lens in eyes with minimal (≤1.50 diopters cylin-der [DC]) and moderate (1.50–3.50 DC) corneal toricities.6

Lens diameter appears to be another influencing factor.7,8

Thus, a smaller lens diameter is related to greater amounts oftreatment zone decentration. Data from paracentral cornealregion, where the first alignment curve of an Ortho-k lensmostly likely falls (between the chords of 7 and 9 mmm),12

may also play a critical role in treatment zone decentration.The relationship between lens decentration and corneal eleva-tion asymmetry at 8-mm chord was confirmed by a recentstudy showing that both direction and magnitude of lensdecentration were influenced by paracentral cornealasymmetry.13 Compared to corneal curvature data,14 paracen-tral corneal elevation difference (CED) between the principlemeridians reflects paracentral corneal toricity in a morestraightforward manner. The relation between central cornealtoricity and treatment zone decentration suggests that differ-ence between two principle meridians may be an importantpredictor for treatment zone decentration. However, to ourknowledge, no previous study has investigated the relationshipbetween paracentral corneal toricity using elevation data andtreatment zone decentration. Furthermore, it is important toidentify the main contributor to treatment zone decentrationbecause that can help us provide better Ortho-k lens fitting inclinical practice. Therefore, the aim of this study was to inves-tigate the influence of paracentral corneal toricity using eleva-tion data at 8-mm chord on the treatment zone decentration ofspherical and toric Ortho-k lens. In addition, multiple linearregression models were used to evaluate the relative

CONTACT Xiao Yang Yangx_zoc@163.com Zhongshan Ophthalmic Center, 54 S. Xianlie Road, Guangzhou 510060, ChinaColor versions of one or more of the figures in the article can be found online at www.tandfonline.com/icey.

CURRENT EYE RESEARCH2018, VOL. 43, NO. 9, 1083–1089https://doi.org/10.1080/02713683.2018.1481516

© 2018 Taylor & Francis Group, LLC

contribution of potential factors in determining the treatmentzone decentration of spherical and toric Ortho-k lens.

Subjects and methods

Participants

This study was conducted at Zhongshan Ophthalmic Center(Guangzhou, China) between December 2016 and March2017. The study adhered to the tenets of the Declaration ofHelsinki and was approved by the ethical committee ofZhongshan Ophthalmic Center, Sun Yat-sen University.Written consent was obtained from guardians of all childrenbefore enrollment.

The inclusion criteria included age between 8 and 18 years,mean sphere between −1.00 and −4.00 D, refractive astigma-tism no greater than −1.50 D and visual acuity correctable to20/20 or better. The exclusion criteria included ocular surfacediseases, fundus diseases, or a history of Ortho-k treatment.Seventy-five right eyes of 75 subjects were enrolled. Twenty-five subjects fitted with spherical Ortho-k, of which the CEDat 8-mm chord was less than 30 μm, were categorized as thelow CED (LCED) group. Fifty subjects with CED at 8-mmchord more than 30 μm were categorized as the high CED(HCED) group. Twenty-five subjects in the HCED groupwere fitted with spherical Ortho-k lens (served as the HCEDI group); the other 25 subjects were fitted with toric Ortho-klens (served as the HCED II group).

Lens fitting

Lenses used in this study were the spherical and toric four-zone reverse-geometry gas-permeable rigid contact lenses(Emerald series; Euclid, Herndon, VA). For both lens types,the back optic zone diameter was 6.2 mm, the width of thereverse curve was 0.5 mm and the width of the peripheralcurve was 0.5 mm. The total lens diameter for a typical triallens was 10.6 mm. All these lenses were made by BOSTONEQUALENS II (oprifocona) and had a nominal Dk of127 × 10–11 (cm2/s) (ml O2/ml_mmHg) (ISO/Fatt) accord-ing to the fitting guidelines supplied by the manufacturer.

For spherical lens fitting, subtraction of the spherical refrac-tive error and a Jessen factor of 0.75 from the baseline cornealflat-k were used to determine the back optical zone radius. Thebaseline corneal flat-k and eccentricity over the 8-mm chord ofthe corneal topography were used to determine the alignmentcurve radius of the first lens. Lens diameter (about 90% HVID)was determined according to the horizontal visible iris diameter(HVID) data. The maximum lens diameter that could beordered was 11.4 mm. For subjects in the HCED II group,toric lens design was chosen according to the CED at 8-mmchord, i.e., 1.00 D toricity for 30 μm difference and adding 0.25D for every 7 ~ 8 μm of increasing CED.

Lens-fitting evaluation was performed using fluorescein. Thealignment curve was determined until a classic bull’s eye wasshown, with a central touch surrounded by a narrow and deepannulus of tears trapped in the reverse curve area. Ideal lensfitting was defined as Ortho-k lens with good centration (lensedge should not go beyond the limbus) and appropriate

movement (approximately 1 mm on a blink). If the cornealresponse after the first overnight was poor, showing, for exam-ple, displacement, smiley and frowny face topographic pattern,a new lens with adjusted parameters would be used. Subjectswho could not show satisfactory fits despite repeated modifica-tions (three pairs of lenses) were excluded from the study. Inthe current study, only two subjects in the HCED I grouprequired a second pair of lenses due to poor lens decentrationat the overnight visit. All the participants were instructed toreturn for a follow-up visit one day, one week and one monthafter the primary lens fitting. Ortho-k lens fitting, visual acuity,ocular health and corneal topography (Medmont E-300,Australia) were evaluated at each visit. After lens dispensing,all subjects were requested to wear their Ortho-K lenses everynight for at least seven consecutive hours.

Determination of CED and corneal asymmetry vector

Corneal topography was conducted using the Placido ring-based Medmont E300 (Australia, Medmont Company) beforeand after lens delivery at each visit. The baseline CED at 8-mm chord was obtained from at least four topography read-ings. First, the mean of average height from the flat and steepmeridians at 8-mm chord were recorded. In this manner, theflat and steep meridians determined for CED calculation werethe same as those from K readings. CED values were thendetermined by subtracting the average height at the flat mer-idian from the average height at the steep meridian. Theasymmetry between different corneal quadrants was calcu-lated according to the method reported by Chen et al.13 Theelevation difference of the two principle meridians of cornealastigmatism at 8-mm chord was recorded. Then, vector ana-lyses based on these differences were conducted to determinethe magnitude and direction of the corneal asymmetry vector.The angle acquired by counterclockwise rotating around thekeratometery center from 3 o’clock of the cornea to the vectorwas defined as the direction of the corneal asymmetry vector.

Determination of the decentration of the treatment zone

The magnitude and direction of treatment zone decentrationwere determined using Image-Pro Plus (IPP) software (pro-duced by Media Cybernetics Corporation, USA). Our previousstudy15 has demonstrated that both the magnitude and directionof treatment zone decentration could be determined using IPPsoftware with good repeatability and reproducibility. After onemonth of Ortho-k lens treatment, a difference map was obtainedby subtracting the pre-orthokeratology (pre-Ortho-k) tangentialcurvature map from the post-orthokeratology (post-Ortho-k)tangential curvature map. Figure 1 illustrates how the decentra-tion of the treatment zone was determined. The step diopter ofthe tangential subtractive map was set to 0.01 D in customsettings, and the whole image was captured with a format settingat a maximum resolution of 1366*768 pixels. Various pointssurrounding the border of the treatment zone area on whichthe powers are all zero were indicated manually and then con-nected to be a closed zone using IPP software. The distance fromthe center of the depicted circle (defined as the center of thetreatment zone) to the corneal vertex normal was defined as

1084 Z. LI ET AL.

treatment zone decentration. The angle (0 ~ 359°) acquired bycounterclockwise rotating around the vertex normal from 3o’clock was defined as the angle of treatment zone decentration.In the current study, one experienced observer who was maskedto the study groups assessed treatment zone centration.

Data analysis

Only data from the right eye were used for statistical calcula-tion. Statistical analysis was performed using SPSS Version 16.0(SPSS 16.0, Inc., Chicago, IL). p < 0.05 at two tails was con-sidered to be statistically significant. The baseline corneal shapeand lens parameters in this study included baseline cornealasymmetry vector and CED at 8-mm chord, corneal flat-k,toricity based on the simulated keratometry or Sim K calculatedapproximately at a 3-mm chord diameter, eccentricity and lensdiameters. Data were first tested for normality using a SampleK-S test. Chi-squared test or one-way ANOVA was used to testthe difference in demographic data, baseline corneal shape andlens parameters among three groups where appropriate. Pairedt-test was used to analyze the change in corneal flat-k, toricityand eccentricity after Ortho-k. Differences in the magnitudeand direction of treatment zone decentration, change of cornealflat-k, toricity and eccentricity after Ortho-k among the threegroups were also analyzed using one-way ANOVA, withBonferroni post hoc testing on indication and correction formultiple comparisons.

The association between corneal shape and lens parametersand treatment zone parameters was tested using univariablelinear regression analyses. High correlations among explana-tory variables are likely to cause multicollinearity, which arti-ficially inflates the variance of the estimated regressioncoefficients and thus makes the estimators less trustworthyfor prediction. The multiple linear regression models includedonly independent factors based on the result of univariablelinear regression analyses, which were performed to test theassociation among all the baseline corneal shape and lens

variables. In this study, CED was significantly correlated tocorneal toricity in both spherical Ortho-k (LCED and HCED Igroups combined) (adjusted R2 = 0.525, ANOVA, F = 52.972,p < 0.01) and toric Ortho-k (HCED II group) groups(adjusted R2 = 0.134, ANOVA, F = 4.707, p = 0.041). Twomultiple linear regression models were established for parti-cipants fitted with spherical (LCED and HCED I groupscombined) and toric Ortho-k. Model 1 explored the associa-tion of treatment zone decentration with the magnitude ofbaseline CED at 8-mm chord, adjusting for the corneal asym-metry vector at 8-mm chord, flat-k, eccentricity and lensdiameter against the magnitude of treatment zone decentra-tion. Model 2 explored the association of treatment zonedecentration with the magnitude of baseline corneal toricity,adjusting for the corneal asymmetry vector at 8-mm chord,flat-k, eccentricity and lens diameter against the magnitude oftreatment zone decentration.

Results

All the 75 subjects completed this one-month study. Baselineinformation with respect to the demographic data, cornealshape and lens parameters are summarized in Table 1. Atbaseline, there was no statistically significant difference ofgender (Chi-squared test, p = 0.913), age, refractive sphere,flat-k, corneal eccentricity along the flattest meridian, amountand angle of corneal asymmetry vector at 8-mm chord, HVIDand lens diameter in the three groups (ANOVA, all p > 0.05).There was no significant difference in baseline refractivecylinder (post hoc multiple comparisons, p = 0.846), cornealtoricity (post hoc multiple comparisons, p = 0.162) and CEDat 8-mm chord (post hoc multiple comparisons, p = 0.303)between the HCED I and HCED II groups. Baseline refractivecylinder, corneal toricity and CED at 8-mm chord in theLCED group were significantly smaller than those in theHCED I and HCED II groups (post hoc multiple compari-sons, all p < 0.01).

Figure 1. Determination of the decentration of treatment zone using Image-Pro Plus software. Various points surrounding the border of the treatment zonearea on which the powers are all zero are indicated manually. Then, all the points were connected to form a yellow closed zone, which represents the border of thetreatment zone. The red cross symbol is the center of treatment zone. Distance of the solid yellow line represents the amount of decentration, and the angle αrepresents the direction of decentration.

CURRENT EYE RESEARCH 1085

After one month of Ortho-k treatment, both the flat-k andcorneal eccentricity along the flattest meridian significantlydecreased from baseline in the three groups (paired t-test, all p <0.01). The magnitude of change in the flat-k (ANOVA, p = 0.170)and corneal eccentricity along the flattest meridian (ANOVA,p = 0.505) were similar among the three groups (Table 2).Corneal toricity did not significantly change from baseline in theLCED group (paired t-test, p = 0.204), but significantly decreasedfrom baseline in the HCED I and HCED II groups (paired t-test,both p < 0.01) (Table 2). The magnitude of corneal toricity changein the HCED I group was significantly smaller than that in theHCED II group (post hoc multiple comparisons, p = 0.045).

As shown in Figure 2A, there was no significant difference inthe magnitude of treatment zone decentration between theLCED and HCED II groups (post hoc multiple comparisons,0.47 ± 0.15 mm vs. 0.47 ± 0.18 mm, respectively; p = 0.965). Themagnitude of treatment zone decentration in the HCED I group(0.73 ± 0.15 mm) was significantly greater than that in the othertwo groups (ANOVA, p < 0.01). As shown in Figure 2B, themean angle of treatment zone decentration was similar amongthe three groups (LCED vs. HCED I vs. HCED II:206.01 ± 52.91° vs. 210.11 ± 45.50° vs. 202.02 ± 58.55°, respec-tively; ANOVA, p = 0.863). For overall displacement, inferotem-poral decentration was the most common (LCED vs. HCED I vs.

Table 1. Comparison of demographic data, baseline corneal shape parameters and orthokeratology lens diameter among the three groups (mean ± SD).

Spherical Toric

Parameters LCED, n = 25 HCED I, n = 25 HCED II, n = 25 p-value

Gender (F/M) 12/13 11/14 13/12 0.913Age (years) 11.68 ± 2.01 12.72 ± 2.01 11.88 ± 1.94 0.154Sphere (D) −2.67 ± 0.80 −2.94 ± 0.91 −2.57 ± 0.82 0.278Cylinder (D) −0.16 ± 0.29 −0.53 ± 0.38 −0.55 ± 0.41 <0.001Flat-K (D) 42.90 ± 1.38 42.52 ± 1.18 42.87 ± 1.25 0.504Corneal toricity (D) 0.71 ± 0.36 1.38 ± 0.43 1.54 ± 0.36 <0.001Corneal eccentricity 0.66 ± 0.07 0.66 ± 0.08 0.64 ± 0.07 0.612CED at 8-mm chord (μm) 15.92 ± 7.00 37.96 ± 6.18 40.07 ± 10.36 <0.001Corneal asymmetry vector (μm) 29.17 ± 12.69 31.05 ± 11.52 29.02 ± 12.70 0.507Angle of asymmetry vector (°) 192.22 ± 43.50 193.68 ± 54.15 194.21 ± 49.49 0.989HVID (mm) 11.61 ± 0.25 11.56 ± 0.36 11.54 ± 0.33 0.728Lens diameter (mm) 10.49 ± 0.22 10.47 ± 0.28 10.42 ± 0.28 0.667

Corneal eccentricity: corneal eccentricity along the flattest meridian; CED at 8-mm chord: corneal elevation difference at 8-mm chord; HVID: horizontal visible irisdiameter. Difference in gender was tested using the Chi-squared test, and difference in the other variables was tested using ANOVA analysis with post hoc multiplecomparisons. p < 0.05 at two tails was considered to be statistically significant. Statistically significant numbers are in bold face.

Table 2. Change of corneal flat-k, toricity and corneal eccentricity after one month of Ortho-k lens wear among the three groups (mean ± SD).

Spherical Toric p-value

Parameters LCED, n = 25 HCED I, n = 25 HCED II, n = 25 (ANOVA)

△Flat-K (D) −2.10 ± 0.52* −2.31 ± 0.64* −1.96 ± 0.74* 0.170

△Corneal toricity (D) −0.10 ± 0.40 −0.32 ± 0.46* −0.57 ± 0.41* 0.001

△Corneal eccentricity −0.35 ± 0.11* −0.33 ± 0.15* −0.30 ± 0.16* 0.505

△Corneal eccentricity: change of corneal eccentricity along the flattest meridian. *: paired t-test p < 0.01.

Figure 2. Distribution of treatment zone decentration. A: Comparison of the magnitude of treatment zone decentration among three groups. B: Overview oftreatment zone lens decentration. 0–360 degrees represent the meridian degree set on the corneal topography map. Dashed and solid circles outline the distance of0.5 mm and 1 mm to the corneal vertex, respectively. **: p-value was less than 0.01.

1086 Z. LI ET AL.

HCED II: 68% vs. 76% vs. 56%, respectively), whereas super-ionasal decentration was the rarest (LCED vs. HCED I vs. HCEDII: 4% vs. 4% vs. 8%, respectively) for the three groups(Figure 2B). The angle of treatment zone decentration and thebaseline angle of the corneal asymmetry vector were significantlycorrelated in the spherical group (LCED and HCED I groupscombined) (adjusted R2 = 0.074, ANOVA, F = 4.889, p= 0.032),but not in the toric group (HCED II) (adjusted R2 = 0.018,ANOVA, F = 1.439, p = 0.243).

Results of univariable and multiple linear regression analysesevaluating the relationship between the amount of treatmentzone decentration and corneal shape and lens parameters inthe spherical Ortho-k group (LCED and HCED I groups com-bined) are summarized in Table 3. In the spherical Ortho-kgroup, the magnitude of lens decentration was significantlyassociated with baseline CED at 8-mm chord, corneal toricityand corneal asymmetry vector at 8-mm chord according to theresults of univariable analyses (all p < 0.01). In multiple linearregression model 1, both the CED (standard β (95% CI): 0.737(0.563, 0.910), p < 0.01) and corneal asymmetry vector at 8 mm(standard β (95% CI): 0.245 (0.071, 0.418), p = 0.007) signifi-cantly contributed to the magnitude of treatment zone decentra-tion (adjusted R2 = 0.645, ANOVA, F = 45.458, p < 0.01),whereas the other factors, including corneal flat-k, eccentricityand lens diameter, did not influence the treatment zone decen-tration (all p > 0.05). In multiple linear regression model 2, bothof the corneal toricity (standard β (95% CI): 0.599 (0.387, 0.811),p < 0.01) and corneal asymmetry vector at 8 mm (standard β(95% CI): 0.298 (0.086, 0.510), p = 0.007) significantly contrib-uted to the magnitude of treatment zone decentration (adjustedR2 = 0.463, ANOVA, F = 22.105, p < 0.01), whereas the otherfactors, including corneal flat-k, eccentricity and lens diameter,did not influence treatment zone decentration (all p > 0.05). Inthe toric Ortho-k group (HCED II group), both univariable andmultiple linear regression analyses showed no significant asso-ciation between the magnitude of treatment zone decentrationand any corneal shape and lens parameters (all p > 0.05).

Discussion

Despite the efficacy of Ortho-k lens in reducing refractive error,visual disturbance due to treatment zone decentration9–11

remains a problem in clinical practice. Corneal toricity,6 cor-neal asymmetry13 and eccentricity16 are some of the factors thatmay influence lens centration. To our knowledge, this is thefirst study that investigates the potential effect of paracentralcorneal toricity using elevation data on treatment zone

decentration with different spherical Ortho-k lens designs. Inaddition, the effect of toric Ortho-k lens on treatment zonedecentration in eyes with greater paracentral corneal toricitywas also investigated.

It is reasonable to assume that good alignment of the lensback-surface with the corneal surface will enhance lens-fittingstability. Maseedupally et al.14 found that the difference ofcorneal curvature between temporal versus nasal sectors orinferior versus superior sectors in the central 5-mm zone wassignificantly smaller than that in the paracentral corneal zonebetween the chords of 5 and 8 mm. In addition, the alignmentcurve, which supports most of the weight of the Ortho-k lens,plays an important role in the stabilization of an Ortho-k lenson the cornea. Thus, data from the peripheral corneal regionbetween the chords of 7 and 9 mm, where the first alignmentcurve of the Ortho-k lens most likely falls on,12 are critical forlens centration and needs to be examined carefully. Instead ofusing the curvature data, we used a simplified method basedon paracentral corneal toricity elevation.14 The magnitude oftreatment zone decentration in eyes with CED at 8-mm chordabove 30 μm was greater than that in eyes with minimal CEDat 8-mm chord with spherical Ortho-k. This finding is furthersupported by the positive relationship when all the data fromeyes fitted with spherical Ortho-k were combined betweenbaseline CED at 8-mm chord and the magnitude of treatmentzone decentration after adjustment for baseline corneal asym-metry, flat-k, eccentricity and lens diameter.

Most myopic children are also astigmatic.17,18

Conventional fitting of spherical Ortho-k lens is mainlybased on the matching of lens sag height with the cornealsag height along the flattest corneal meridian. When theamount of corneal toricity increases, this method of selectinglens parameters is thought to influence lens-fitting stabilitydue to the unequal alignment of the Ortho-k lens along theprinciple meridians. Swarbrick et al.6 found that eyes with ahigher amount of central corneal toricity display greateramounts of treatment zone decentration during sphericalOrtho-k. Consistent with Swarbrick’s result, we found a sig-nificant association between central corneal toricity and treat-ment zone decentration. However, paracentral corneal toricityusing elevation data may be a better predictor for decentrationthan central corneal toricity, as the standard coefficients ofunivariable and multiple linear regression analyses for para-central CED were greater than those for central corneal tori-city. The relative contribution of central and paracentralcorneal toricity to spherical lens decentration could be furtherinvestigated by their significant correlation. However, as also

Table 3. Results of univariable and multiple linear regression analyses evaluating the relationship between treatment zone decentration with corneal shape and lensparameters in the spherical Ortho-k group.

Treatment zone decentration

Univariable Multiple*

standard β (95% CI) p standard β (95% CI) p

Model 1CED at 8-mm chord 0.775 (0.592, 0.958) <0.001 0.737 (0.563, 0.910) <0.001Corneal asymmetry vector 0.360 (0.089, 0.631) 0.010 0.245 (0.071, 0.418) 0.007

Model 2Corneal toricity 0.630 (0.405, 0.855) <0.001 0.599 (0.387, 0.811) <0.001Corneal asymmetry vector 0.360 (0.089, 0.631) 0.010 0.298 (0.086, 0.510) 0.007

*Adjusted for corneal flat-k, eccentricity and lens diameter using multiple linear regression analyses.

CURRENT EYE RESEARCH 1087

found by Maseedupally et al., these parameters did not corre-spond one to one.14 Thus, it appears that the curvature in thecentral corneal zone is significantly different from that in theparacentral region. During Ortho-k lens fitting, it is thereforeimportant to assess paracentral CED in eyes with high centralcorneal toricity or in cases where bad lens centration occursdespite low central corneal toricity.

Another important finding in the current study was thatwhen fitted with toric Ortho-k, the magnitude of treatmentzone decentration was significantly smaller than when fittedwith spherical Ortho-k in eyes with similar great CED at 8-mm chord (above 30 μm). This indicates that the toric fittingtechnique improves lens-fitting stability in eyes with greaterparacentral CED. Interestingly, the association between treat-ment zone decentration and other potential factors showedthat all factors, including CED, corneal asymmetry and centralcorneal toricity, were independent of lens decentration intoric Ortho-k subjects. The absence of any associationbetween the amount of treatment zone decentration and themagnitude of corneal shape parameters in spherical Ortho-ksubjects versus toric Ortho-k subjects suggests that the toricfitting technique may help improve lens centration by weak-ening the influence of these corneal shape parameters, espe-cially for CED at 8 mm more than 30 μm. It should be noticedhere that the sample size in the toric Ortho-k group wasrelatively small, and further studies with larger sample sizeare therefore warranted.

In a previous study,7 we found that the amount of lensdecentration with a lens diameter of 11.0 mm was signifi-cantly smaller than that with lens diameters between 10.2 and10.6 mm (0.41 mm vs. 0.77 mm, respectively). Hiraoka et al.8

found a relatively higher amount of lens decentration of0.85 mm when all the eyes were fitted with the same lensdiameter of 10.0 mm. This suggests that a larger lens diametermay help limit lens decentration. However, uniform lens sizeis not feasible in clinical practice because the corneal sizevaries with different subjects. Lens diameter in the currentstudy was determined according to the HVID result. In con-trast, both our current data and Chen et al.’s data16 showedthat the HVID and lens diameter were independent of lensdecentration during spherical Ortho-k when lens diameterwas tailored to the subject’s corneal size.

Regarding the direction of decentration for overall displa-cement, inferotemporal decentration was the most common,whereas superionasal decentration was the rarest for all thethree groups. This is consistent with results from previousstudies.6–8,16 Baseline corneal asymmetry vector has been sug-gested as a possible explanation.16 In the current study, themean directions of the corneal asymmetry vector among thethree groups were all inferotemporal, and both the amountand direction significantly contributed to the magnitude andangle of treatment zone decentration.

Eyelid force is one potential factor influencing the align-ment between the lens and the corneal surface that was notinvestigated in the current study. During Ortho-k lens fitting,many factors should be taken into account, including cornealparameters, lens design and eyelid forces. Further studies areneeded to investigate how a combination of these factors mayinfluence Ortho-k lens fitting.

In conclusion, the amount of paracentral corneal tori-city using elevation data at 8-mm chord appears to be abetter predictor of lens decentration than central cornealtoricity and corneal asymmetry. When the CED at 8-mmchord was above 30 μm, a toric Ortho-k fitting appears toimprove lens centration and should therefore berecommended.

Disclosure Statement

All the authors declare that they have no competing interests.

Funding

This work was supported by the National Natural Science Foundation ofChina, China [81200716].

References

1. El HS, Leach NE, Miller W, Prager TC, Marsack J, Parker K,Minavi A, Gaume A. Empirical advanced orthokeratologythrough corneal topography: the University of Houston clinicalstudy. Eye Contact Lens. 2007;33(5):224–35. doi:10.1097/ICL.0b013e318065b0dd.

2. Chan B, Cho P, Cheung SW. Orthokeratology practice in childrenin a university clinic in Hong Kong. Clin Exp Optom. 2008;91(5):453–60. doi:10.1111/j.1444-0938.2008.00259.x.

3. Li SM, Kang MT, Wu SS, Liu LR, Li H, Chen Z, Wang N. Efficacy,safety and acceptability of orthokeratology on slowing axial elon-gation in myopic children by meta-analysis. Curr Eye Res. 2016;41(5):600–08. doi:10.3109/02713683.2015.1050743.

4. Sun Y, Xu F, Zhang T, Liu M, Wang D, Chen Y, Liu Q.Orthokeratology to control myopia progression: a meta-ana-lysis. PLoS One. 2015;10(4):e0124535. doi:10.1371/journal.pone.0124535.

5. Si JK, Tang K, Bi HS, Guo DD, Guo JG, Wang XR.Orthokeratology for myopia control: a meta-analysis. Optom VisSci. 2015;92(3):252–57. doi:10.1097/OPX.0000000000000505.

6. Maseedupally VK, Gifford P, Lum E, Naidu R, Sidawi D, Wang B,Swarbrick HA. Treatment zone decentration during orthokeratol-ogy on eyes with corneal Toricity. Optom Vis Sci. 2016;93(9):1101–11. doi:10.1097/OPX.0000000000000896.

7. Yang X, Zhong X, Gong X, Zeng J. Topographical evaluation ofthe decentration of orthokeratology lenses. Yan Ke Xue Bao.2005;21:132–5, 195.

8. Hiraoka T, Mihashi T, Okamoto C, Okamoto F, Hirohara Y,Oshika T. Influence of induced decentered orthokeratology lenson ocular higher-order wavefront aberrations and contrast sensi-tivity function. J Cataract Refract Surg. 2009;35(11):1918–26.doi:10.1016/j.jcrs.2009.06.018.

9. Hiraoka T, Furuya A, Matsumoto Y, Okamoto F, Sakata N,Hiratsuka K, Kakita T, Oshika T. Quantitative evaluation ofregular and irregular corneal astigmatism in patients having over-night orthokeratology. J Cataract Refract Surg. 2004;30(7):1425–29. doi:10.1016/j.jcrs.2004.02.049.

10. Hiraoka T, Okamoto C, Ishii Y, Kakita T, Oshika T. Contrastsensitivity function and ocular higher-order aberrations followingovernight orthokeratology. Invest Ophthalmol Vis Sci. 2007;48(2):550–56. doi:10.1167/iovs.06-0914.

11. Hiraoka T, Okamoto C, Ishii Y, Kakita T, Okamoto F, Oshika T.Time course of changes in ocular higher-order aberrations andcontrast sensitivity after overnight orthokeratology. InvestOphthalmol Vis Sci. 2008;49(10):4314–20. doi:10.1167/iovs.07-1586.

12. Tahhan N, Du Toit R, Papas E, Chung H, La Hood D, Holden AB.Comparison of reverse-geometry lens designs for overnight

1088 Z. LI ET AL.

orthokeratology. Optom Vis Sci. 2003;80(12):796–804.doi:10.1097/00006324-200312000-00009.

13. Chen Z, Xue F, Zhou J, Qu X, Zhou X. Prediction of orthoker-atology lens decentration with corneal elevation. Optom Vis Sci.2017;94(9):903–07. doi:10.1097/OPX.0000000000001109.

14. Maseedupally V, Gifford P, Lum E, Swarbrick H. Central and para-central corneal curvature changes during orthokeratology. OptomVis Sci. 2013;90(11):1249–58. doi:10.1097/OPX.0000000000000039.

15. Mei Y, Tang Z, Li Z, Yang X. Repeatability and reproducibility ofquantitative corneal shape analysis after orthokeratology treat-ment using Image-Pro Plus software. J Ophthalmol.2016;2016:1732476. doi:10.1155/2016/1732476.

16. Li J, Yang C, Xie W, Zhang G, Li X, Wang S, Yang X, Zeng J.Predictive role of corneal Q-value differences between nasal-tem-poral and superior-inferior quadrants in orthokeratology lensdecentration. Medicine (Baltimore). 2017;96(2):e5837. doi:10.1097/MD.0000000000005837.

17. Fan DS, Rao SK, Cheung EY, Islam M, Chew S, Lam DS.Astigmatism in Chinese preschool children: prevalence, change,and effect on refractive development. Br J Ophthalmol. 2004;88(7):938–41. doi:10.1136/bjo.2003.030338.

18. Kleinstein RN, Jones LA, Hullett S, Kwon S, Lee RJ, Friedman NE,Manny RE, Mutti DO, Julie AY, Zadnik K. Refractive error andethnicity in children. Arch Ophthalmol. 2003;121(8):1141–47.doi:10.1001/archopht.121.8.1141.

CURRENT EYE RESEARCH 1089