- Email: [email protected]
Distribution of Anterior and Posterior Corneal Astigmatism in Eyes With Keratoconus
Accepted Manuscript Distribution of the Anterior and Posterior Corneal Astigmatism in Eyes with Keratoconus Mohammad Naderan, MD, Mohammad Taher Rajabi, MD, Parviz Zarrinbakhsh, MD PII:
S0002-9394(16)30155-6
DOI:
10.1016/j.ajo.2016.03.051
Reference:
AJOPHT 9705
To appear in:
American Journal of Ophthalmology
Received Date: 25 December 2015 Revised Date:
29 March 2016
Accepted Date: 31 March 2016
Please cite this article as: Naderan M, Rajabi MT, Zarrinbakhsh P, Distribution of the Anterior and Posterior Corneal Astigmatism in Eyes with Keratoconus, American Journal of Ophthalmology (2016), doi: 10.1016/j.ajo.2016.03.051. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Distribution of the Anterior and Posterior Corneal Astigmatism in the Eyes with Keratoconus Abstract
RI PT
Purpose: To investigate the magnitude, with-the-rule (WTR) or against-the-rule (ATR) orientation, and vector components (Jackson astigmatic vectors [J0 and J45] and blurring strength) of the anterior and posterior corneal astigmatism (ACA and PCA) in patients with keratoconus (KC) in a retrospective study, and trying to find suitable cutoff points for ACA and PCA in an attempt to discriminate KC from normal corneas. Design: Retrospective age- and sex-matched case-control study.
SC
Methods: Using the Pentacam images, the aforementioned parameters were compared between 1273 patients with KC and 1035 normal participants.
TE D
M AN U
Results: The mean magnitude of the ACA and PCA was 4.47±2.14 and 0.90±0.43 diopter (D), respectively. The dominant astigmatism orientation of the ACA was ATR in KC patients and WTR in normal participants (p<0.001), while for the PCA it was WTR in KC patients and ATR in normal participants (p<0.001). There was a significant agreement between the axis orientations of ACA and PCA in KC patients (ĸ=0.077, p<0.001), but not in normal group (p=0.626). ACA and PCA magnitude, M, J0, J45, and blur significantly increased by increasing KC severity. There was a trend for increasing anterior ATR and posterior WTR, and decreasing oblique astigmatism on both corneal surfaces by increasing the KC severity according to the AmslerKrumeich classification. A cutoff value of 1.8 D for ACA had 90.2% sensitivity and specificity, and that of 0.4 D for PCA had 89.5% sensitivity and 85.0% specificity for discriminating KC from normal corneas.
EP
Conclusion: Our findings can help clinicians in the diagnosis of KC and lens manufactures in designing suitable contact or intraocular lenses.
AC C
Key Words: Anterior cornea; Astigmatism; Keratoconus; Posterior cornea; Scheimpflug imaging
1
ACCEPTED MANUSCRIPT Distribution of the Anterior and Posterior Corneal Astigmatism in Eyes with Keratoconus Mohammad Naderan MD1,2, Mohammad Taher Rajabi MD1, Parviz Zarrinbakhsh MD3 1
2
RI PT
Eye Research Center, Farabi Eye Hospital, Tehran University of Medical Sciences, Tehran, Iran Ophthalmic Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran 3
Zarrinbakhsh Eye Clinic, Tehran, Iran
SC
Corresponding Author: Mohammad Naderan M.D.
Email:
[email protected]
M AN U
Eye Research Center, Farabi Eye Hospital, Tehran University of Medical Sciences, Tehran, Iran
Address: Eye Research Center, Farabi Eye Hospital, Qazvin Square, Tehran, Iran Zip code: 1336616351, Tel: +98-21-55421006, Fax: +98-21-55416134,
AC C
EP
TE D
Short Title: Corneal Astigmatism in Keratoconus
2
ACCEPTED MANUSCRIPT Introduction
RI PT
Keratoconus (KC) is a progressive, usually bilateral ectatic corneal disorder, characterized by corneal thinning and protrusion.1, 2 KC starts at puberty and progresses to the third or fourth decade of life, causing myopia and astigmatism which results in severe vision distortion and sometimes even blindness.1 Astigmatism is a refractive error that is mostly caused by toricity of the anterior corneal surface leading to visually significant optical aberration. Both the anterior and posterior corneal surfaces contribute to the total corneal astigmatism. Recently, the direct and quantitative measurement of the posterior corneal measurements in a clinical setting has been possible with new imaging technologies such as slitscanning, Scheimpflug, or optical coherence devices.3, 4
M AN U
SC
Assessment of the corneal astigmatism plays an important role in vision correction procedures such as rigid gas-permeable lens prescription or intraocular lens (IOL) implantation in KC patients. Ho et al. 3 reported that neglecting the posterior corneal astigmatism may results in significant deviation in the estimation of the corneal astigmatism. There are several studies evaluated anterior corneal astigmatism in KC patients;5 however, a few studies evaluated the magnitude and orientation of the posterior corneal astigmatism in patients with KC.6 Moreover, several studies tried to differentiate between KC and normal corneas by means of corneal astigmatism and suggested various cutoff points with different sensitivity and specificity.7-12 But the results are inconclusive.
Method
EP
TE D
In the current study, we aimed to investigate the magnitude, orientation, and vector components of the anterior and posterior corneal astigmatism in patients with KC in comparison with normal corneas and according to different KC severity stages. Furthermore, we tried to find a suitable cutoff point for the anterior and posterior corneal astigmatism which could discriminate between KC and normal corneas with the highest sensitivity, specificity, and accuracy.
AC C
A retrospective age- and sex-matched case-control study was conducted in Zarrinbakhsh Eye Clinic, Tehran, Iran, reviewing the documents of the patients diagnosed with KC who attended the clinic from 2010 to 2015. This study was in accordance with the tenets of the Declaration of Helsinki and the Institutional Review Board (IRB) and ethics committee of our clinic approved the study. Since our study was a retrospective chart review, the IRB of our clinic waived the requirement of obtaining informed consent from the participants. The diagnosis of KC was based on the clinical characteristic signs, such as Fleischer ring, Vogt’s striae, stromal scar, using slit-lamp examinations and also corneal topography evaluation using Pentacam (OCULUS Optikgerate GmbH, Wetzlar, Germany). For the control group, a number of age- and sex-matched subjects who primarily attended the clinic for laser in-situ keratomileusis and did not fulfill the diagnostic criteria of the KC or were not KC suspect based on slit-lamp examination and Pentacam imaging, were consecutively selected and included in the
3
ACCEPTED MANUSCRIPT study. Based on the patients' documents, those with a history of ocular trauma or surgery and any corneal or other ophthalmic disorders were excluded from the study.
RI PT
The Pentacam device uses a 475 nm monochromatic blue light-emitting diode (LED) with a 180-degree rotating Scheimpflug camera. The camera rotates around the optical axes of the eye and within 2 seconds captures a total of 25 images and produces a 3-dimensional model of the anterior segment of the eye. The instrument is capable of automatically analyzing the anterior segment, the anterior chamber and the lens, and performing the anterior and posterior topography of the cornea and pachymetry measurements.
M AN U
SC
Before performing the Pentacam imaging, patients were asked to stop using contact lenses for at least 2 weeks. The imaging was performed as follows. The patients were asked to place their chin on the chin rest of the device, press their forehead to the forehead strap, stare at a central target or fixation light and when a perfect alignment between the patient's eye and visual axis was obtained, they were asked to blink and then the imaging was performed. All measurements were based on the data from the annular ring that was 3 mm in diameter around the corneal apex. In our clinic, we routinely assess the Scheimpflug imaging parameters on the 3 mm corneal diameter. This would make our results comparable with the literature.13 The Pentacam images were reviewed, and those with a good quality were included in the study.
TE D
In patients with bilateral KC and in the normal group as well, only one eye of each participant was randomly selected. The randomization was performed using a series of random numbers generated by an independent statistician who were masked to the purpose of the study and diagnosis of the participants, using a computerized randomization program. The following parameters were recorded: anterior and posterior corneal astigmatism (ACA and PCA), astigmatism axis, central corneal thickness (CCT), and anterior and posterior mean keratometry (K). The severity of KC was classified according to the Amsler-Krumeich classification.5, 14
AC C
EP
Anterior corneal astigmatism was classified as with-the-rule (WTR) when the steepest meridian of the corneal surface was between 60-120 degrees and as against-the-rule (ATR) when the steepest meridian was between 0-30 or 150-180 degrees. Since the dioptric power of the posterior corneal surface is negative, posterior corneal astigmatism was classified as WTR when the steepest meridian was between 0-30 or 150-180 degrees and as ATR when the steepest meridian was between 60-120 degrees. The remaining values were classified as oblique astigmatism. Power vector method was employed to quantify the relationship between astigmatism measurements.15 Conventional script notations of manifest refractions (sphere [S], cylinder [C], and axis [α]) were applied to calculate power vector coordinates. The method uses three fundamental vectors, including M = S + C/2, J0 = (-C/2) cos2α, and J45 = (-C/2) sin2α, where S is the sphere power, C is the cylinder power, α is the cylinder axis, and J is the Jackson astigmatic vector. M is the spherical lens equal to the spherical equivalent of the given refractive error. J0 value is the cylinder power set at 90-degree and 180-degree meridians and J45 value refers
4
ACCEPTED MANUSCRIPT to a cross-cylinder set at 45 and 135 degrees.16 The overall blurring strength is calculated through the following formula: B = (M2 + J02 + J452)1/2.
RI PT
Receiver operating characteristic (ROC) curves were produced in order to determine the diagnostic significance of various astigmatism measurements. The area under ROC curves (AUROC) was calculated to describe the predictive accuracy of the different measurements and the optimized cutoff points which could best distinguish KC from normal eyes. An AUROC between 0.90 and 1.0 represents excellent discrimination, between 0.80 and 0.90 good, between 0.70 and 0.80 fair, between 0.60 and 0.70 poor, between 0.50 and 0.60 very poor, and <0.50 represents insufficient measures.
M AN U
SC
The best cutoff point was determined where the tests' sensitivity and specificity were maximized. Sensitivity, specificity, and positive and negative predictive values (PPV and NPV) were calculated for the measurements with AUROC of ≥ 0.900 to assess the validity of cutoff points for predicting KC. The sensitivity (false negative) is defined as the ability of a test to correctly identify the patients with disease. The specificity (false positive) is defined as the ability of a test to correctly identify the patients without disease. PPV is the probability that patients with a positive screening test truly have the disease. NPV is the probability that patients with a negative screening test truly do not have the disease. Statistical Analysis
AC C
EP
TE D
Data analysis was performed using IBM SPSS Statistics (Version 22; IBM Inc., Armonk, NY, US). The normality of the data was assessed using the Kolmogorov-Smirnov test, which revealed the data of anterior and posterior corneal astigmatism were not normally distributed. The results were demonstrated by mean ± standard deviation (SD). The Chi-square test was utilized to analyze categorical variables. The Mann-Whitney U test or Kruskal-Wallis test was used for the analysis of the continuous variables which were not normally distributed. The Spearman correlation test was used to evaluate the correlation between different variables. To evaluate the agreement between anterior and posterior astigmatism orientations, the Cohen kappa coefficient was applied. P value <0.05 was considered statistically significant.
Results
This study included 1273 eyes of 1273 patients with keratoconus and 1035 eyes of 1035 control participants. The clinical characteristics and astigmatism measurements of the KC patients and the control group are demonstrated in Table 1. The mean age of the KC patients was 25 ± 7 years and that of the control group was 25 ± 5. No significant difference was observed between the mean age and the sex of the KC and those of the control group (p>0.05). In the patients with KC, the mean magnitude of ACA and PCA was 4.47 ± 2.14 and 0.90 ± 0.43 diopter (D), respectively, which were significantly higher than that of the normal group (p<0.001). The dominant astigmatism orientation of the anterior corneal surface was ATR in patients with KC and WTR in normal participants (p<0.001). The dominant
5
ACCEPTED MANUSCRIPT astigmatism orientation of the posterior corneal surface was WTR in patients with KC and ATR in normal participants (p<0.001). In addition, spherical equivalent (M), J0 and J45 were significantly higher in the patients with KC (p<0.001). The frequency distribution (percentages) of the anterior and posterior corneal astigmatism in the KC patients is shown in Figure 1.
M AN U
SC
RI PT
The mean magnitude of the anterior and posterior corneal astigmatism according to the sex, age groups, and astigmatism orientation are shown in Table 2. In KC patients, no significant differences were found between males and females regarding ACA and PCA. In the normal group, males had significantly higher PCA (p<0.001). In the KC patients PCA gradually decreased by increasing in age but the difference was not statistically significant (p=0.240). On the contrary, by increasing in age PCA significantly increased in the normal participants (p=0.003). In the anterior corneal surface, the mean magnitude of ATR astigmatism was higher than WTR and oblique orientations in the KC patients (p<0.001). In contrast, WTR orientation of the anterior corneal surface had a higher magnitude in the normal group (p<0.001). Furthermore, in the posterior corneal surface, the mean magnitude of WTR astigmatism was significantly greater in the KC patients, and the mean magnitude of ATR astigmatism was significantly greater in the normal group (p<0.001). The distribution of the anterior and posterior corneal astigmatism of the KC patients in each age group is demonstrated in Figure 2. Most eyes in all of the age groups showed ATR astigmatism in the anterior corneal surface, whereas WTR was the most prevalent astigmatism orientation of the posterior corneal surface.
AC C
EP
TE D
The axis orientation of anterior and posterior corneal surfaces according to each other is presented in Table 3. ACA and PCA had the same axis orientation in 324 eyes (25.4%), in which 20 eyes (1.6%) had WTR, 17 eyes (1.3%) had ATR, and 287 eyes (22.5%) had oblique orientation in both anterior and posterior surfaces. Moreover, there was a significant agreement between the axis orientations of ACA and PCA in the KC patients (ĸ=0.077, p<0.001). On the other hand, no significant agreement was found between the axis orientations of the ACA and PCA in the normal group (ĸ=0.003, p=0.626). Figure 3 shows the correlation scatter-gram between the magnitudes of the anterior and posterior corneal astigmatism in the patients with KC. The Spearman correlation analysis revealed that ACA had a significant positive correlation with PCA in the patients with KC (r= 0.785, p<0.001) and the normal participants (r=0.486, p<0.001). In general, based on the Amsler-Krumeich classifications 427 eyes (33.5%) had stage 1, 502 eyes (39.4%) had stage 2, 181 eyes (14.2%) had stage 3, and 163 eyes (12.8%) had stage 4 of the KC severity. The magnitude, axis orientation, and vector component of the corneal astigmatism according to the disease severity are demonstrated in Table 4. ACA showed a statistically significant increase from stage 1 to 3, revealing that higher disease severity was associated with larger anterior and posterior corneal astigmatism. There was a trend for increasing in the prevalence of ATR astigmatism of the anterior corneal surface by an increase in the disease severity (p<0.001). On the contrary, by increasing the disease severity, WTR astigmatism of the posterior corneal surface increased (p<0.001). In both of the anterior and posterior corneal astigmatism, the oblique orientation gradually decreased. Moreover, M, blur, J0, and J45 components of the anterior and posterior corneal astigmatism were significantly increased by increasing in KC severity.
6
ACCEPTED MANUSCRIPT
RI PT
ROC analysis was performed including all astigmatism variables (Figure 4). The results revealed that ACA and PCA were associated with the largest AUROC and highest diagnostic ability (Table 5). An ACA cutoff value of 1.8 D had 90.2% sensitivity and specificity and a PCA cutoff value of 0.4 D had 89.5% sensitivity and 85.0% specificity in discriminating KC from normal eyes. The calculated sensitivity, specificity, PPV, and NPV based on these cut-offs are demonstrated in Table 6.
Discussion
M AN U
SC
In the past, the direct measurement of the posterior corneal surface was not possible. Posterior corneal parameters were calculated based on the anterior measurements and through a keratometric index (1.3375 in most cases) to compensate for the posterior corneal surface. Recently with the advancement of the newer instruments such as Scheimpflug imaging the evaluation of the posterior corneal surface is achievable. It is reported that in the eyes with higher ACA and total corneal astigmatism, the influence of the PCA increases. Therefore, it is important to evaluate the PCA in the keratoconic eyes, which mostly have a higher magnitude of corneal astigmatism than normal eyes.7
TE D
In our study, the mean magnitudes of the anterior and posterior corneal astigmatism were 4.49 ± 2.16 and 0.90 ± 0.43 D, respectively, which were in line with the findings of Kamiya et al. 6 who evaluated anterior and posterior corneal astigmatism in 137 patients with KC by the means of Pentacam HR and reported that the mean magnitudes of anterior and posterior corneal astigmatism were 3.93 ± 2.74 and 0.93 ± 0.64 D, respectively. Orucoglu et al. 7 evaluated 656 eyes of 338 KC patients by Pentacam and reported lower mean magnitudes of ACA and PCA in comparison to our study. The mean magnitudes of ACA and PCA were 3.05 ± 1.97 and 0.71 ± 0.44 D in their patients, respectively.7
AC C
EP
We found that ATR astigmatism was more prevalent in the anterior corneal surface (57.4%) and WTR astigmatism was more frequent in the posterior corneal surface (63.2%) in the KC patients. In contrast, Kamiya et al. 6 found that WTR astigmatism was more prevalent in the anterior corneal surface (65.7%) and oblique astigmatism was more prevalent in the posterior corneal surface (78.8%). This significant difference in our study with Kamiya et al. 6 could be attributed to the different ethnicity of the study populations. Although we could not exactly explain the reason behind this difference, it is possible that Iranian descent which is mostly Caucasian has different ACA and PCA axis orientations on Pentacam compared to the study by Kamiya et al. 6 whose study population was exclusively Japanese, being an Asian ethnicity. ACA and PCA had the same axis orientations in 25.4% of the eyes, which was higher than those in the research by Kamiya et al. 6 who found similar anterior and posterior orientations in 16.0% of their patients. Moreover, we found a statistically significant agreement between the axis orientations of ACA and PCA (Cohen kappa coefficient), which was again in contrast with Kamiya et al. 6 study who did not find a significant agreement. Additionally, there was a significant correlation between magnitudes of the ACA and PCA which was in accordance with Kamiya et al. 6 study.
7
ACCEPTED MANUSCRIPT
RI PT
Except for stage 4 of the KC severity, we found a gradually significant increase in the magnitude of ACA and PCA. In contrast, Kamiya et al. 6 did not find any significant increase in astigmatism measurements with the progressive stages of the KC. Although similar to them, more severe stages of KC (stages 3 and 4) had larger values of ACA and PCA than less severe stages (stages 1 and 2). Moreover, there was a trend for increasing in the prevalence of the anterior ATR astigmatism and posterior WTR astigmatism, and decreasing oblique astigmatism of both corneal surfaces by an increase in the disease severity. Similar to our findings, Kamiya et al. 6 reported an increase in the prevalence of anterior corneal surface ATR orientation and posterior corneal surface WTR orientation, but the oblique orientation in both surfaces increased along with the severity.
EP
TE D
M AN U
SC
There are several studies tried to distinguish between KC and normal corneas by means of topographic and tomographic parameters which have reported various cutoff points with different sensitivity and specificity.7-12 In the current study, the cutoff value of 1.8 D for ACA had 90.2% sensitivity and specificity, and 95.8% accuracy, and the cutoff value of 0.4 D for PCA had 89.5% sensitivity, 85.5% specificity, and 92.7% accuracy in differentiating KC from normal corneas. Orucoglu et al. 7 stated that the cutoff value of 1.65 D for ACA had 73.3% sensitivity, 81.6% specificity, and 81.5% accuracy, and the cutoff value of 0.45 D for PCA had 70.8% sensitivity, 82.8% specificity, and 80.2% accuracy in discriminating KC from normal corneas. Pinero et al. 12 reported that a cutoff point of 0.50 D for PCA would detect KC with 77.8% sensitivity, 87.1% specificity, and 86.3% accuracy. Fontes et al. 10 suggested that a cutoff point of 2.2 D for corneal astigmatism had 70.1% sensitivity, 89.5% specificity, and 80.3% accuracy. In another study, Toprak et al. 9 proposed a cutoff value of 2.4 D for central astigmatism and reported 74.3% sensitivity, 78.6% specificity, and 81.8% accuracy in discriminating KC from normal corneas. Similar to our findings, Henriquez et al. 8 reported excellent predictive accuracy of ACA (91%) in distinguishing KC from normal corneas, however, they did not suggest cutoff points or analyze sensitivity and specificity. In contrast, Lim et al. 11 did not find out that central astigmatism could significantly discriminate KC from normal corneas (p=0.174). Compared to the previous studies, it seems that the suggested cutoff values for ACA and PCA in the current study would better discriminate KC from normal corneas with the highest accuracy, sensitivity, and specificity.
AC C
Our study has some limitations. The retrospective design, younger age of the patients and higher prevalence of male patients may prone our results to bias. Moreover, we were unable to measure total corneal astigmatism due to the Pentacam limitations. Furthermore, our study measurements were based at 3 mm annular ring and we did not measure the other zones such as 4.5 mm; so it is not clear whether the measurements at other diameters show similar outcomes and trends. Finally, our measurements were performed only with Pentacam, and our results should be confirmed by other corneal imaging methods such as a dual Scheimpflug analyzer or optical coherence tomography and tomography. In conclusion, we found that most keratoconic eyes had ATR orientation of the anterior corneal surface and WTR orientation of the posterior corneal surface. Anterior and posterior corneal astigmatism was correlated with each other. Astigmatism magnitude was significantly associated with KC severity. There was a trend for increasing in the prevalence of anterior ATR astigmatism and posterior
8
ACCEPTED MANUSCRIPT
Acknowledgment
RI PT
WTR astigmatism, and decreasing oblique astigmatism of both corneal surfaces by an increase in the disease severity. ACA and PCA had excellent ability in discriminating KC from normal corneas. A cutoff value of 1.8 D for ACA and 0.4 D for PCA would discriminate between KC and normal corneas with the highest sensitivity, specificity and accuracy. These findings may help ophthalmic physicians and lens manufactures with prescription and designing suitable contact or intraocular lenses.
Funding/Support: No funds, grants or other support were received. Financial Disclosures: No financial disclosures.
AC C
EP
TE D
M AN U
SC
Other Acknowledgments: There were no conflicts of interest.
9
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
1. Romero-Jimenez M, Santodomingo-Rubido J, Wolffsohn JS. Keratoconus: a review. Cont Lens Anterior Eye 2010;33(4):157-166; quiz 205. 2. Naderan M, Shoar S, Rezagholizadeh F, Zolfaghari M, Naderan M. Characteristics and associations of keratoconus patients. Cont Lens Anterior Eye 2015;38(3):199-205. 3. Ho JD, Tsai CY, Liou SW. Accuracy of corneal astigmatism estimation by neglecting the posterior corneal surface measurement. Am J Ophthalmol 2009;147(5):788-795. 4. Koch DD, Ali SF, Weikert MP, Shirayama M, Jenkins R, Wang L. Contribution of posterior corneal astigmatism to total corneal astigmatism. J Cataract Refract Surg 2012;38(12):2080-2087. 5. Naderan M, Shoar S, Kamaleddin MA, Rajabi MT, Naderan M, Khodadadi M. Keratoconus Clinical Findings According to Different Classifications. Cornea 2015;34(9):1005-1011. 6. Kamiya K, Shimizu K, Igarashi A, Miyake T. Assessment of Anterior, Posterior, and Total Central Corneal Astigmatism in Eyes With Keratoconus. Am J Ophthalmol 2015;160(5):851-857. 7. Orucoglu F, Toker E. Comparative analysis of anterior segment parameters in normal and keratoconus eyes generated by scheimpflug tomography. J Ophthalmol 2015;2015:925414. 8. Henriquez MA, Izquierdo L, Jr., Belin MW. Intereye Asymmetry in Eyes With Keratoconus and High Ammetropia: Scheimpflug Imaging Analysis. Cornea 2015;34(Suppl 10):S57-S60. 9. Toprak I, Yaylali V, Yildirim C. A combination of topographic and pachymetric parameters in keratoconus diagnosis. Cont Lens Anterior Eye 2015;38(5):357-362. 10. Fontes BM, Ambrosio Junior R, Jardim D, Velarde GC, Nose W. Ability of corneal biomechanical metrics and anterior segment data in the differentiation of keratoconus and healthy corneas. Arq Bras Oftalmol 2010;73(4):333-337. 11. Lim HB, Tan GS, Lim L, Htoon HM. Comparison of keratometric and pachymetric parameters with Scheimpflug imaging in normal and keratoconic Asian eyes. Clin Ophthalmol 2014;8:2215-2220. 12. Pinero DP, Perez-Cambrodi RJ, Soto-Negro R, Ruiz-Fortes P, Artola A. Clinical utility of ocular residual astigmatism and topographic disparity vector indexes in subclinical and clinical keratoconus. Graefes Arch Clin Exp Ophthalmol 2015;253(12):2229-2237. 13. Naderan M, Shoar S, Naderan M, Kamaleddin MA, Rajabi MT. Comparison of corneal measurements in keratoconic eyes using rotating Scheimpflug camera and scanning-slit topography. Int J Ophthalmol 2015;8(2):275-280. 14. Krumeich JH, Daniel J, Knulle A. Live-epikeratophakia for keratoconus. J Cataract Refract Surg 1998;24(4):456-463. 15. Thibos LN, Wheeler W, Horner D. Power vectors: an application of Fourier analysis to the description and statistical analysis of refractive error. Optom Vis Sci 1997;74(6):367-375. 16. Liu YC, Chou P, Wojciechowski R, et al. Power vector analysis of refractive, corneal, and internal astigmatism in an elderly Chinese population: the Shihpai Eye Study. Invest Ophthalmol Vis Sci 2011;52(13):9651-9657.
10
ACCEPTED MANUSCRIPT Figure Legends
RI PT
Figure1. Histogram of the frequency distribution (percentages) of the magnitude of the anterior (left) and posterior (right) corneal astigmatism in the patients with keratoconus
Figure2. Distribution of the anterior (left) and posterior (right) corneal astigmatism in each age group in the patients with keratoconus
M AN U
SC
Figure3. Correlation scatter-gram between the magnitudes of the anterior and posterior corneal astigmatism (p<0.001, y=0.2+0.16*x, r2=0.608) in the patients with keratoconus
AC C
EP
TE D
Figure4. Receiver operating characteristic curves analysis. ACA: anterior corneal astigmatism, J0 and J45, power vector components of manifest cylinder, PCA: posterior corneal astigmatism
1
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Table1. The baseline characteristics and clinical findings of the KC patients and the control participants Feature Keratoconus Control P value (N=1273) (N=1035) Age (year) 25 ± 7 25 ± 5 0.138* Sex (male) 739 (58.1%) 580 (56.0%) 0.176† Mean K (D) 50.2 ± 4.6 43.1 ± 1.15 <0.001* CCT (µm) 450 ± 38 547 ± 32 <0.001* TCT (µm) 431 ± 40 531 ± 33 <0.001* Astigmatism (D) ACA 4.49 ± 2.16 0.93 ± 0.74 <0.001* PCA 0.90 ± 0.43 0.26 ± 0.14 <0.001* ACA Orientation <0.001† WTR 92 (7.2%) 740 (71.5%) ATR 731 (57.4%) 65 (6.3%) Oblique 450 (35.3%) 230 (22.2%) PCA Orientation <0.001† WTR 805 (63.2%) 10 (1.0%) ATR 94 (7.4%) 925 (89.4%) Oblique 374 (29.4%) 100 (9.7%) M -5.56 ± 4.2 -2.69 ± 1.8 <0.001* Blur 6.24 ± 4.00 2.80 ± 1.81 <0.001* * Mann-Whitney U test. † Chi-square test. ACA: anterior corneal astigmatism, ATR: against-the-rule, Blur: overall blurring strength of the manifest spherocylindrical error, CCT: central corneal thickness, D: diopter, K: keratometry, M: spherical equivalent, PCA: posterior corneal astigmatism, TCT: thinnest corneal thickness, WTR: with-the-rule. Bold values are statistically significant.
ACCEPTED MANUSCRIPT
1
Parameters
ACA (D) Keratoconus Normal
P value
RI PT
Table2. Descriptive measurements of the anterior and posterior corneal astigmatism according to the sex, age groups, and astigmatism orientation of the keratoconus patients and normal subjects PCA (D) Keratoconus Normal
P value
AC C
EP
TE D
M AN U
SC
Sex Males 4.53 ± 2.29 1.00 ± 0.86 <0.001* 0.89 ± 0.45 0.29 ± 0.15 <0.001* Females 4.44 ± 1.97 0.84 ± 0.55 <0.001* 0.91 ± 0.40 0.23 ± 0.11 <0.001* P value 0.933* 0.328* ——— 0.129* <0.001* ——— Age ≤20 4.96 ± 2.41 0.95 ± 0.68 <0.001* 0.95 ± 0.44 0.24 ± 0.11 <0.001* 21-25 4.52 ± 2.09 0.92 ± 0.73 <0.001* 0.91 ± 0.43 0.25 ± 0.13 <0.001* 26-30 4.07 ± 2.01 0.90 ± 0.64 <0.001* 0.88 ± 0.42 0.28 ± 0.15 <0.001* 31-35 4.46 ± 2.37 1.00 ± 1.05 <0.001* 0.88 ± 0.46 0.27 ± 0.15 <0.001* ≥36 4.44 ± 1.95 0.98 ± 0.83 <0.001* 0.82 ± 0.44 0.29 ± 0.15 <0.001* P value 0.002† 0.946† ——— 0.240† 0.003† ——— Orientation WTR 3.50 ± 2.02 1.07 ± 0.80 <0.001* 0.95 ± 0.45 0.10 ± 0.04 <0.001* ATR 4.83 ± 2.27 0.25 ± 0.20 <0.001* 0.63 ± 0.38 0.27 ± 0.14 <0.001* Oblique 4.15 ± 1.86 0.69 ± 0.46 <0.001* 0.85 ± 0.37 0.22 ± 0.10 <0.001* P value <0.001† <0.001† ——— <0.001† <0.001† ——— Vector Component J0 1.05 ± 1.35 -0.25 ± 0.35 <0.001* 1.21 ± 1.33 -0.34 ± 0.45 <0.001* J45 0.02 ± 1.76 -0.11 ± 0.48 0.017* -0.11 ± 1.67 0.19 ± 0.27 <0.001* Data is presented as mean ± SD. * Mann-Whitney U test, † Kruskal-Wallis test. ACA: anterior corneal astigmatism, ATR: against-the-rule, D: diopter, J0 and J45: power vector components of manifest cylinder, PCA: posterior corneal astigmatism, WTR: with-the-rule. Bold values are statistically significant.
ACCEPTED MANUSCRIPT
Table3. Axis orientations of the anterior and posterior corneal astigmatism in keratoconus patients PCA Orientation ATR
AC C
EP
TE D
M AN U
SC
RI PT
Total WTR Oblique ACA Orientation WTR 20 {1.6%} (21.7%) [2.5%] 51 {4.0%} (55.4%) [54.3%] 21 {1.6%} (22.8%) [5.6%] 92 {7.2%} (100%) [7.2%] ATR 648 {50.9%} (88.6%) [80.5%] 17 {1.3%} (2.3%) [18.1%] 66 {5.2%} (9.0%) [17.6%] 731 {57.4%} (100%) [57.4%] Oblique 137 {10.8%} (30.4%) [17.0%] 26 {2.0%} (5.8%) [27.7%] 287 {22.5%} (63.8%) [76.7%] 450 {35.3%} (100%) [35.3%] Total 805 {63.2%} (63.2%) [100%] 94 {7.4%} (7.4%) [100%] 374 {29.4%} (29.4%) [100%] 1273 {100%} (100%) [100%] ACA: anterior corneal astigmatism, ATR: against-the-rule, PCA: posterior corneal astigmatism, WTR: with-the-rule. {% within the whole samples}, (% within each ACA orientation group), [% within each PCA orientation group]
ACCEPTED MANUSCRIPT
Table4. Corneal astigmatism according to the keratoconus severity Parameter
Keratoconus Severity Stage 2 Stage 3 -5.33 ± 3.26 -7.69 ± 4.48 5.95 ± 3.07 8.32 ± 4.25
AC C
EP
TE D
M AN U
SC
RI PT
P value Stage 1 Stage 4 M -3.40 ± 2.75 -9.91 ± 5.71 <0.001* Blur 4.13 ± 2.53 10.70 ± 5.05 <0.001* Anterior corneal astigmatism (ACA) Magnitude (D) 4.16 ± 1.80 4.20 ± 1.84 5.49 ± 2.70 5.17 ± 2.73 <0.001* Orientation <0.001† WTR 43 (46.7%) [10.1%] 33 (35.9%) [6.6%] 4 (4.3%) [2.2%] 12 (13.0%) [7.4%] ATR 218 (29.8%) [51.1%] 283 (38.7%) [56.4%] 116 (15.9%) [64.1%] 114 (15.6%) [69.9%] Oblique 166 (36.9%) [38.9%] 186 (41.3%) [37.1%] 61 (13.6%) [33.7%] 37 (8.2%) [22.7%] J0 0.72 ± 1.16 1.01 ± 1.25 1.46 ± 1.36 1.63 ± 1.75 <0.001* J45 0.01 1.56 0.18 ± 1.78 -0.24 ± 2.00 -0.18 ± 1.89 0.039* Posterior corneal astigmatism (PCA) PCA magnitude (D) 0.85 ± 0.38 0.86 ± 0.38 1.09 ± 0.52 0.96 ± 0.54 <0.001* PCA Orientation <0.001† WTR 224 (27.8%) [52.5%] 330 (41.0%) [65.7%] 121 (15.0%) [66.9%] 130 (16.1%) [79.8%] ATR 29 (30.9%) [6.8%] 41 (43.6%) [8.2%] 12 (12.8%) [6.6%] 12 (12.8%) [7.4%] Oblique 174 (46.5%) [40.7%] 131 (35.0%) [26.1%] 48 (12.8%) [26.5%] 21 (5.6%) [12.9%] J0 0.83 ± 1.08 1.21 ± 1.24 1.48 ± 1.50 1.92 ± 1.69 <0.001* J45 -0.04 ± 1.57 0.05 ± 1.68 -0.41 ± 1.85 -0.48 ± 1.60 0.002* * Kruskal-Wallis test. † Chi-square test. ATR: against-the-rule, Blur: overall blurring strength of the manifest spherocylindrical error, D: diopter, J0 and J45: power vector components of manifest cylinder, M: spherical equivalent, WTR: with-the-rule. (% within each astigmatism orientation group), [% within each severity stage group] Bold values are statistically significant.
ACCEPTED MANUSCRIPT Table5. Receiver operating curve (ROC) analysis AUROC
Asymptotic 95% CI P value Lower bound Upper bound ACA Magnitude 0.958 0.949 0.966 <0.001 PCA Magnitude 0.927 0.914 0.939 <0.001 J0 (PCA) 0.899 0.883 0.915 <0.001 J0 (ACA) 0.863 0.845 0.881 <0.001 Blur 0.798 0.779 0.816 <0.001 J45 (ACA) 0.530 0.503 0.558 0.018 J45 (PCA) 0.442 0.415 0.470 <0.001 M 0.276 0.254 0.298 <0.001 ACA: anterior corneal astigmatism, AUROC: area under receiver operating curve, CI: confidence interval, Blur: overall blurring strength of the manifest spherocylindrical error, J0 and J45: power vector components of manifest cylinder, M: spherical equivalent, PCA: posterior corneal astigmatism. Bold values are statistically significant.
AC C
EP
TE D
M AN U
SC
RI PT
Variables
ACCEPTED MANUSCRIPT Table6. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for predicting keratoconus Sensitivity Specificity PPV NPV (95% CI) (95% CI) (95% CI) (95% CI) ACA ≥ 1.8 D 90.2% 90.2% 92.0% 88.2% (88.4-91.8) (88.3-92) (90.2-93.4) (86.1-90.0) PCA ≥ 0.4 D 89.5% 85.0% 88.0% 86.9% (87.7-91.2) (82.7-87.1) (86.1-89.7) (84.6-88.9) ACA: anterior corneal astigmatism, CI: confidence interval, D: diopter, PCA: posterior corneal astigmatism.
AC C
EP
TE D
M AN U
SC
RI PT
Variables
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0002-9394(16)30155-6
DOI:
10.1016/j.ajo.2016.03.051
Reference:
AJOPHT 9705
To appear in:
American Journal of Ophthalmology
Received Date: 25 December 2015 Revised Date:
29 March 2016
Accepted Date: 31 March 2016
Please cite this article as: Naderan M, Rajabi MT, Zarrinbakhsh P, Distribution of the Anterior and Posterior Corneal Astigmatism in Eyes with Keratoconus, American Journal of Ophthalmology (2016), doi: 10.1016/j.ajo.2016.03.051. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Distribution of the Anterior and Posterior Corneal Astigmatism in the Eyes with Keratoconus Abstract
RI PT
Purpose: To investigate the magnitude, with-the-rule (WTR) or against-the-rule (ATR) orientation, and vector components (Jackson astigmatic vectors [J0 and J45] and blurring strength) of the anterior and posterior corneal astigmatism (ACA and PCA) in patients with keratoconus (KC) in a retrospective study, and trying to find suitable cutoff points for ACA and PCA in an attempt to discriminate KC from normal corneas. Design: Retrospective age- and sex-matched case-control study.
SC
Methods: Using the Pentacam images, the aforementioned parameters were compared between 1273 patients with KC and 1035 normal participants.
TE D
M AN U
Results: The mean magnitude of the ACA and PCA was 4.47±2.14 and 0.90±0.43 diopter (D), respectively. The dominant astigmatism orientation of the ACA was ATR in KC patients and WTR in normal participants (p<0.001), while for the PCA it was WTR in KC patients and ATR in normal participants (p<0.001). There was a significant agreement between the axis orientations of ACA and PCA in KC patients (ĸ=0.077, p<0.001), but not in normal group (p=0.626). ACA and PCA magnitude, M, J0, J45, and blur significantly increased by increasing KC severity. There was a trend for increasing anterior ATR and posterior WTR, and decreasing oblique astigmatism on both corneal surfaces by increasing the KC severity according to the AmslerKrumeich classification. A cutoff value of 1.8 D for ACA had 90.2% sensitivity and specificity, and that of 0.4 D for PCA had 89.5% sensitivity and 85.0% specificity for discriminating KC from normal corneas.
EP
Conclusion: Our findings can help clinicians in the diagnosis of KC and lens manufactures in designing suitable contact or intraocular lenses.
AC C
Key Words: Anterior cornea; Astigmatism; Keratoconus; Posterior cornea; Scheimpflug imaging
1
ACCEPTED MANUSCRIPT Distribution of the Anterior and Posterior Corneal Astigmatism in Eyes with Keratoconus Mohammad Naderan MD1,2, Mohammad Taher Rajabi MD1, Parviz Zarrinbakhsh MD3 1
2
RI PT
Eye Research Center, Farabi Eye Hospital, Tehran University of Medical Sciences, Tehran, Iran Ophthalmic Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran 3
Zarrinbakhsh Eye Clinic, Tehran, Iran
SC
Corresponding Author: Mohammad Naderan M.D.
Email:
[email protected]
M AN U
Eye Research Center, Farabi Eye Hospital, Tehran University of Medical Sciences, Tehran, Iran
Address: Eye Research Center, Farabi Eye Hospital, Qazvin Square, Tehran, Iran Zip code: 1336616351, Tel: +98-21-55421006, Fax: +98-21-55416134,
AC C
EP
TE D
Short Title: Corneal Astigmatism in Keratoconus
2
ACCEPTED MANUSCRIPT Introduction
RI PT
Keratoconus (KC) is a progressive, usually bilateral ectatic corneal disorder, characterized by corneal thinning and protrusion.1, 2 KC starts at puberty and progresses to the third or fourth decade of life, causing myopia and astigmatism which results in severe vision distortion and sometimes even blindness.1 Astigmatism is a refractive error that is mostly caused by toricity of the anterior corneal surface leading to visually significant optical aberration. Both the anterior and posterior corneal surfaces contribute to the total corneal astigmatism. Recently, the direct and quantitative measurement of the posterior corneal measurements in a clinical setting has been possible with new imaging technologies such as slitscanning, Scheimpflug, or optical coherence devices.3, 4
M AN U
SC
Assessment of the corneal astigmatism plays an important role in vision correction procedures such as rigid gas-permeable lens prescription or intraocular lens (IOL) implantation in KC patients. Ho et al. 3 reported that neglecting the posterior corneal astigmatism may results in significant deviation in the estimation of the corneal astigmatism. There are several studies evaluated anterior corneal astigmatism in KC patients;5 however, a few studies evaluated the magnitude and orientation of the posterior corneal astigmatism in patients with KC.6 Moreover, several studies tried to differentiate between KC and normal corneas by means of corneal astigmatism and suggested various cutoff points with different sensitivity and specificity.7-12 But the results are inconclusive.
Method
EP
TE D
In the current study, we aimed to investigate the magnitude, orientation, and vector components of the anterior and posterior corneal astigmatism in patients with KC in comparison with normal corneas and according to different KC severity stages. Furthermore, we tried to find a suitable cutoff point for the anterior and posterior corneal astigmatism which could discriminate between KC and normal corneas with the highest sensitivity, specificity, and accuracy.
AC C
A retrospective age- and sex-matched case-control study was conducted in Zarrinbakhsh Eye Clinic, Tehran, Iran, reviewing the documents of the patients diagnosed with KC who attended the clinic from 2010 to 2015. This study was in accordance with the tenets of the Declaration of Helsinki and the Institutional Review Board (IRB) and ethics committee of our clinic approved the study. Since our study was a retrospective chart review, the IRB of our clinic waived the requirement of obtaining informed consent from the participants. The diagnosis of KC was based on the clinical characteristic signs, such as Fleischer ring, Vogt’s striae, stromal scar, using slit-lamp examinations and also corneal topography evaluation using Pentacam (OCULUS Optikgerate GmbH, Wetzlar, Germany). For the control group, a number of age- and sex-matched subjects who primarily attended the clinic for laser in-situ keratomileusis and did not fulfill the diagnostic criteria of the KC or were not KC suspect based on slit-lamp examination and Pentacam imaging, were consecutively selected and included in the
3
ACCEPTED MANUSCRIPT study. Based on the patients' documents, those with a history of ocular trauma or surgery and any corneal or other ophthalmic disorders were excluded from the study.
RI PT
The Pentacam device uses a 475 nm monochromatic blue light-emitting diode (LED) with a 180-degree rotating Scheimpflug camera. The camera rotates around the optical axes of the eye and within 2 seconds captures a total of 25 images and produces a 3-dimensional model of the anterior segment of the eye. The instrument is capable of automatically analyzing the anterior segment, the anterior chamber and the lens, and performing the anterior and posterior topography of the cornea and pachymetry measurements.
M AN U
SC
Before performing the Pentacam imaging, patients were asked to stop using contact lenses for at least 2 weeks. The imaging was performed as follows. The patients were asked to place their chin on the chin rest of the device, press their forehead to the forehead strap, stare at a central target or fixation light and when a perfect alignment between the patient's eye and visual axis was obtained, they were asked to blink and then the imaging was performed. All measurements were based on the data from the annular ring that was 3 mm in diameter around the corneal apex. In our clinic, we routinely assess the Scheimpflug imaging parameters on the 3 mm corneal diameter. This would make our results comparable with the literature.13 The Pentacam images were reviewed, and those with a good quality were included in the study.
TE D
In patients with bilateral KC and in the normal group as well, only one eye of each participant was randomly selected. The randomization was performed using a series of random numbers generated by an independent statistician who were masked to the purpose of the study and diagnosis of the participants, using a computerized randomization program. The following parameters were recorded: anterior and posterior corneal astigmatism (ACA and PCA), astigmatism axis, central corneal thickness (CCT), and anterior and posterior mean keratometry (K). The severity of KC was classified according to the Amsler-Krumeich classification.5, 14
AC C
EP
Anterior corneal astigmatism was classified as with-the-rule (WTR) when the steepest meridian of the corneal surface was between 60-120 degrees and as against-the-rule (ATR) when the steepest meridian was between 0-30 or 150-180 degrees. Since the dioptric power of the posterior corneal surface is negative, posterior corneal astigmatism was classified as WTR when the steepest meridian was between 0-30 or 150-180 degrees and as ATR when the steepest meridian was between 60-120 degrees. The remaining values were classified as oblique astigmatism. Power vector method was employed to quantify the relationship between astigmatism measurements.15 Conventional script notations of manifest refractions (sphere [S], cylinder [C], and axis [α]) were applied to calculate power vector coordinates. The method uses three fundamental vectors, including M = S + C/2, J0 = (-C/2) cos2α, and J45 = (-C/2) sin2α, where S is the sphere power, C is the cylinder power, α is the cylinder axis, and J is the Jackson astigmatic vector. M is the spherical lens equal to the spherical equivalent of the given refractive error. J0 value is the cylinder power set at 90-degree and 180-degree meridians and J45 value refers
4
ACCEPTED MANUSCRIPT to a cross-cylinder set at 45 and 135 degrees.16 The overall blurring strength is calculated through the following formula: B = (M2 + J02 + J452)1/2.
RI PT
Receiver operating characteristic (ROC) curves were produced in order to determine the diagnostic significance of various astigmatism measurements. The area under ROC curves (AUROC) was calculated to describe the predictive accuracy of the different measurements and the optimized cutoff points which could best distinguish KC from normal eyes. An AUROC between 0.90 and 1.0 represents excellent discrimination, between 0.80 and 0.90 good, between 0.70 and 0.80 fair, between 0.60 and 0.70 poor, between 0.50 and 0.60 very poor, and <0.50 represents insufficient measures.
M AN U
SC
The best cutoff point was determined where the tests' sensitivity and specificity were maximized. Sensitivity, specificity, and positive and negative predictive values (PPV and NPV) were calculated for the measurements with AUROC of ≥ 0.900 to assess the validity of cutoff points for predicting KC. The sensitivity (false negative) is defined as the ability of a test to correctly identify the patients with disease. The specificity (false positive) is defined as the ability of a test to correctly identify the patients without disease. PPV is the probability that patients with a positive screening test truly have the disease. NPV is the probability that patients with a negative screening test truly do not have the disease. Statistical Analysis
AC C
EP
TE D
Data analysis was performed using IBM SPSS Statistics (Version 22; IBM Inc., Armonk, NY, US). The normality of the data was assessed using the Kolmogorov-Smirnov test, which revealed the data of anterior and posterior corneal astigmatism were not normally distributed. The results were demonstrated by mean ± standard deviation (SD). The Chi-square test was utilized to analyze categorical variables. The Mann-Whitney U test or Kruskal-Wallis test was used for the analysis of the continuous variables which were not normally distributed. The Spearman correlation test was used to evaluate the correlation between different variables. To evaluate the agreement between anterior and posterior astigmatism orientations, the Cohen kappa coefficient was applied. P value <0.05 was considered statistically significant.
Results
This study included 1273 eyes of 1273 patients with keratoconus and 1035 eyes of 1035 control participants. The clinical characteristics and astigmatism measurements of the KC patients and the control group are demonstrated in Table 1. The mean age of the KC patients was 25 ± 7 years and that of the control group was 25 ± 5. No significant difference was observed between the mean age and the sex of the KC and those of the control group (p>0.05). In the patients with KC, the mean magnitude of ACA and PCA was 4.47 ± 2.14 and 0.90 ± 0.43 diopter (D), respectively, which were significantly higher than that of the normal group (p<0.001). The dominant astigmatism orientation of the anterior corneal surface was ATR in patients with KC and WTR in normal participants (p<0.001). The dominant
5
ACCEPTED MANUSCRIPT astigmatism orientation of the posterior corneal surface was WTR in patients with KC and ATR in normal participants (p<0.001). In addition, spherical equivalent (M), J0 and J45 were significantly higher in the patients with KC (p<0.001). The frequency distribution (percentages) of the anterior and posterior corneal astigmatism in the KC patients is shown in Figure 1.
M AN U
SC
RI PT
The mean magnitude of the anterior and posterior corneal astigmatism according to the sex, age groups, and astigmatism orientation are shown in Table 2. In KC patients, no significant differences were found between males and females regarding ACA and PCA. In the normal group, males had significantly higher PCA (p<0.001). In the KC patients PCA gradually decreased by increasing in age but the difference was not statistically significant (p=0.240). On the contrary, by increasing in age PCA significantly increased in the normal participants (p=0.003). In the anterior corneal surface, the mean magnitude of ATR astigmatism was higher than WTR and oblique orientations in the KC patients (p<0.001). In contrast, WTR orientation of the anterior corneal surface had a higher magnitude in the normal group (p<0.001). Furthermore, in the posterior corneal surface, the mean magnitude of WTR astigmatism was significantly greater in the KC patients, and the mean magnitude of ATR astigmatism was significantly greater in the normal group (p<0.001). The distribution of the anterior and posterior corneal astigmatism of the KC patients in each age group is demonstrated in Figure 2. Most eyes in all of the age groups showed ATR astigmatism in the anterior corneal surface, whereas WTR was the most prevalent astigmatism orientation of the posterior corneal surface.
AC C
EP
TE D
The axis orientation of anterior and posterior corneal surfaces according to each other is presented in Table 3. ACA and PCA had the same axis orientation in 324 eyes (25.4%), in which 20 eyes (1.6%) had WTR, 17 eyes (1.3%) had ATR, and 287 eyes (22.5%) had oblique orientation in both anterior and posterior surfaces. Moreover, there was a significant agreement between the axis orientations of ACA and PCA in the KC patients (ĸ=0.077, p<0.001). On the other hand, no significant agreement was found between the axis orientations of the ACA and PCA in the normal group (ĸ=0.003, p=0.626). Figure 3 shows the correlation scatter-gram between the magnitudes of the anterior and posterior corneal astigmatism in the patients with KC. The Spearman correlation analysis revealed that ACA had a significant positive correlation with PCA in the patients with KC (r= 0.785, p<0.001) and the normal participants (r=0.486, p<0.001). In general, based on the Amsler-Krumeich classifications 427 eyes (33.5%) had stage 1, 502 eyes (39.4%) had stage 2, 181 eyes (14.2%) had stage 3, and 163 eyes (12.8%) had stage 4 of the KC severity. The magnitude, axis orientation, and vector component of the corneal astigmatism according to the disease severity are demonstrated in Table 4. ACA showed a statistically significant increase from stage 1 to 3, revealing that higher disease severity was associated with larger anterior and posterior corneal astigmatism. There was a trend for increasing in the prevalence of ATR astigmatism of the anterior corneal surface by an increase in the disease severity (p<0.001). On the contrary, by increasing the disease severity, WTR astigmatism of the posterior corneal surface increased (p<0.001). In both of the anterior and posterior corneal astigmatism, the oblique orientation gradually decreased. Moreover, M, blur, J0, and J45 components of the anterior and posterior corneal astigmatism were significantly increased by increasing in KC severity.
6
ACCEPTED MANUSCRIPT
RI PT
ROC analysis was performed including all astigmatism variables (Figure 4). The results revealed that ACA and PCA were associated with the largest AUROC and highest diagnostic ability (Table 5). An ACA cutoff value of 1.8 D had 90.2% sensitivity and specificity and a PCA cutoff value of 0.4 D had 89.5% sensitivity and 85.0% specificity in discriminating KC from normal eyes. The calculated sensitivity, specificity, PPV, and NPV based on these cut-offs are demonstrated in Table 6.
Discussion
M AN U
SC
In the past, the direct measurement of the posterior corneal surface was not possible. Posterior corneal parameters were calculated based on the anterior measurements and through a keratometric index (1.3375 in most cases) to compensate for the posterior corneal surface. Recently with the advancement of the newer instruments such as Scheimpflug imaging the evaluation of the posterior corneal surface is achievable. It is reported that in the eyes with higher ACA and total corneal astigmatism, the influence of the PCA increases. Therefore, it is important to evaluate the PCA in the keratoconic eyes, which mostly have a higher magnitude of corneal astigmatism than normal eyes.7
TE D
In our study, the mean magnitudes of the anterior and posterior corneal astigmatism were 4.49 ± 2.16 and 0.90 ± 0.43 D, respectively, which were in line with the findings of Kamiya et al. 6 who evaluated anterior and posterior corneal astigmatism in 137 patients with KC by the means of Pentacam HR and reported that the mean magnitudes of anterior and posterior corneal astigmatism were 3.93 ± 2.74 and 0.93 ± 0.64 D, respectively. Orucoglu et al. 7 evaluated 656 eyes of 338 KC patients by Pentacam and reported lower mean magnitudes of ACA and PCA in comparison to our study. The mean magnitudes of ACA and PCA were 3.05 ± 1.97 and 0.71 ± 0.44 D in their patients, respectively.7
AC C
EP
We found that ATR astigmatism was more prevalent in the anterior corneal surface (57.4%) and WTR astigmatism was more frequent in the posterior corneal surface (63.2%) in the KC patients. In contrast, Kamiya et al. 6 found that WTR astigmatism was more prevalent in the anterior corneal surface (65.7%) and oblique astigmatism was more prevalent in the posterior corneal surface (78.8%). This significant difference in our study with Kamiya et al. 6 could be attributed to the different ethnicity of the study populations. Although we could not exactly explain the reason behind this difference, it is possible that Iranian descent which is mostly Caucasian has different ACA and PCA axis orientations on Pentacam compared to the study by Kamiya et al. 6 whose study population was exclusively Japanese, being an Asian ethnicity. ACA and PCA had the same axis orientations in 25.4% of the eyes, which was higher than those in the research by Kamiya et al. 6 who found similar anterior and posterior orientations in 16.0% of their patients. Moreover, we found a statistically significant agreement between the axis orientations of ACA and PCA (Cohen kappa coefficient), which was again in contrast with Kamiya et al. 6 study who did not find a significant agreement. Additionally, there was a significant correlation between magnitudes of the ACA and PCA which was in accordance with Kamiya et al. 6 study.
7
ACCEPTED MANUSCRIPT
RI PT
Except for stage 4 of the KC severity, we found a gradually significant increase in the magnitude of ACA and PCA. In contrast, Kamiya et al. 6 did not find any significant increase in astigmatism measurements with the progressive stages of the KC. Although similar to them, more severe stages of KC (stages 3 and 4) had larger values of ACA and PCA than less severe stages (stages 1 and 2). Moreover, there was a trend for increasing in the prevalence of the anterior ATR astigmatism and posterior WTR astigmatism, and decreasing oblique astigmatism of both corneal surfaces by an increase in the disease severity. Similar to our findings, Kamiya et al. 6 reported an increase in the prevalence of anterior corneal surface ATR orientation and posterior corneal surface WTR orientation, but the oblique orientation in both surfaces increased along with the severity.
EP
TE D
M AN U
SC
There are several studies tried to distinguish between KC and normal corneas by means of topographic and tomographic parameters which have reported various cutoff points with different sensitivity and specificity.7-12 In the current study, the cutoff value of 1.8 D for ACA had 90.2% sensitivity and specificity, and 95.8% accuracy, and the cutoff value of 0.4 D for PCA had 89.5% sensitivity, 85.5% specificity, and 92.7% accuracy in differentiating KC from normal corneas. Orucoglu et al. 7 stated that the cutoff value of 1.65 D for ACA had 73.3% sensitivity, 81.6% specificity, and 81.5% accuracy, and the cutoff value of 0.45 D for PCA had 70.8% sensitivity, 82.8% specificity, and 80.2% accuracy in discriminating KC from normal corneas. Pinero et al. 12 reported that a cutoff point of 0.50 D for PCA would detect KC with 77.8% sensitivity, 87.1% specificity, and 86.3% accuracy. Fontes et al. 10 suggested that a cutoff point of 2.2 D for corneal astigmatism had 70.1% sensitivity, 89.5% specificity, and 80.3% accuracy. In another study, Toprak et al. 9 proposed a cutoff value of 2.4 D for central astigmatism and reported 74.3% sensitivity, 78.6% specificity, and 81.8% accuracy in discriminating KC from normal corneas. Similar to our findings, Henriquez et al. 8 reported excellent predictive accuracy of ACA (91%) in distinguishing KC from normal corneas, however, they did not suggest cutoff points or analyze sensitivity and specificity. In contrast, Lim et al. 11 did not find out that central astigmatism could significantly discriminate KC from normal corneas (p=0.174). Compared to the previous studies, it seems that the suggested cutoff values for ACA and PCA in the current study would better discriminate KC from normal corneas with the highest accuracy, sensitivity, and specificity.
AC C
Our study has some limitations. The retrospective design, younger age of the patients and higher prevalence of male patients may prone our results to bias. Moreover, we were unable to measure total corneal astigmatism due to the Pentacam limitations. Furthermore, our study measurements were based at 3 mm annular ring and we did not measure the other zones such as 4.5 mm; so it is not clear whether the measurements at other diameters show similar outcomes and trends. Finally, our measurements were performed only with Pentacam, and our results should be confirmed by other corneal imaging methods such as a dual Scheimpflug analyzer or optical coherence tomography and tomography. In conclusion, we found that most keratoconic eyes had ATR orientation of the anterior corneal surface and WTR orientation of the posterior corneal surface. Anterior and posterior corneal astigmatism was correlated with each other. Astigmatism magnitude was significantly associated with KC severity. There was a trend for increasing in the prevalence of anterior ATR astigmatism and posterior
8
ACCEPTED MANUSCRIPT
Acknowledgment
RI PT
WTR astigmatism, and decreasing oblique astigmatism of both corneal surfaces by an increase in the disease severity. ACA and PCA had excellent ability in discriminating KC from normal corneas. A cutoff value of 1.8 D for ACA and 0.4 D for PCA would discriminate between KC and normal corneas with the highest sensitivity, specificity and accuracy. These findings may help ophthalmic physicians and lens manufactures with prescription and designing suitable contact or intraocular lenses.
Funding/Support: No funds, grants or other support were received. Financial Disclosures: No financial disclosures.
AC C
EP
TE D
M AN U
SC
Other Acknowledgments: There were no conflicts of interest.
9
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
1. Romero-Jimenez M, Santodomingo-Rubido J, Wolffsohn JS. Keratoconus: a review. Cont Lens Anterior Eye 2010;33(4):157-166; quiz 205. 2. Naderan M, Shoar S, Rezagholizadeh F, Zolfaghari M, Naderan M. Characteristics and associations of keratoconus patients. Cont Lens Anterior Eye 2015;38(3):199-205. 3. Ho JD, Tsai CY, Liou SW. Accuracy of corneal astigmatism estimation by neglecting the posterior corneal surface measurement. Am J Ophthalmol 2009;147(5):788-795. 4. Koch DD, Ali SF, Weikert MP, Shirayama M, Jenkins R, Wang L. Contribution of posterior corneal astigmatism to total corneal astigmatism. J Cataract Refract Surg 2012;38(12):2080-2087. 5. Naderan M, Shoar S, Kamaleddin MA, Rajabi MT, Naderan M, Khodadadi M. Keratoconus Clinical Findings According to Different Classifications. Cornea 2015;34(9):1005-1011. 6. Kamiya K, Shimizu K, Igarashi A, Miyake T. Assessment of Anterior, Posterior, and Total Central Corneal Astigmatism in Eyes With Keratoconus. Am J Ophthalmol 2015;160(5):851-857. 7. Orucoglu F, Toker E. Comparative analysis of anterior segment parameters in normal and keratoconus eyes generated by scheimpflug tomography. J Ophthalmol 2015;2015:925414. 8. Henriquez MA, Izquierdo L, Jr., Belin MW. Intereye Asymmetry in Eyes With Keratoconus and High Ammetropia: Scheimpflug Imaging Analysis. Cornea 2015;34(Suppl 10):S57-S60. 9. Toprak I, Yaylali V, Yildirim C. A combination of topographic and pachymetric parameters in keratoconus diagnosis. Cont Lens Anterior Eye 2015;38(5):357-362. 10. Fontes BM, Ambrosio Junior R, Jardim D, Velarde GC, Nose W. Ability of corneal biomechanical metrics and anterior segment data in the differentiation of keratoconus and healthy corneas. Arq Bras Oftalmol 2010;73(4):333-337. 11. Lim HB, Tan GS, Lim L, Htoon HM. Comparison of keratometric and pachymetric parameters with Scheimpflug imaging in normal and keratoconic Asian eyes. Clin Ophthalmol 2014;8:2215-2220. 12. Pinero DP, Perez-Cambrodi RJ, Soto-Negro R, Ruiz-Fortes P, Artola A. Clinical utility of ocular residual astigmatism and topographic disparity vector indexes in subclinical and clinical keratoconus. Graefes Arch Clin Exp Ophthalmol 2015;253(12):2229-2237. 13. Naderan M, Shoar S, Naderan M, Kamaleddin MA, Rajabi MT. Comparison of corneal measurements in keratoconic eyes using rotating Scheimpflug camera and scanning-slit topography. Int J Ophthalmol 2015;8(2):275-280. 14. Krumeich JH, Daniel J, Knulle A. Live-epikeratophakia for keratoconus. J Cataract Refract Surg 1998;24(4):456-463. 15. Thibos LN, Wheeler W, Horner D. Power vectors: an application of Fourier analysis to the description and statistical analysis of refractive error. Optom Vis Sci 1997;74(6):367-375. 16. Liu YC, Chou P, Wojciechowski R, et al. Power vector analysis of refractive, corneal, and internal astigmatism in an elderly Chinese population: the Shihpai Eye Study. Invest Ophthalmol Vis Sci 2011;52(13):9651-9657.
10
ACCEPTED MANUSCRIPT Figure Legends
RI PT
Figure1. Histogram of the frequency distribution (percentages) of the magnitude of the anterior (left) and posterior (right) corneal astigmatism in the patients with keratoconus
Figure2. Distribution of the anterior (left) and posterior (right) corneal astigmatism in each age group in the patients with keratoconus
M AN U
SC
Figure3. Correlation scatter-gram between the magnitudes of the anterior and posterior corneal astigmatism (p<0.001, y=0.2+0.16*x, r2=0.608) in the patients with keratoconus
AC C
EP
TE D
Figure4. Receiver operating characteristic curves analysis. ACA: anterior corneal astigmatism, J0 and J45, power vector components of manifest cylinder, PCA: posterior corneal astigmatism
1
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Table1. The baseline characteristics and clinical findings of the KC patients and the control participants Feature Keratoconus Control P value (N=1273) (N=1035) Age (year) 25 ± 7 25 ± 5 0.138* Sex (male) 739 (58.1%) 580 (56.0%) 0.176† Mean K (D) 50.2 ± 4.6 43.1 ± 1.15 <0.001* CCT (µm) 450 ± 38 547 ± 32 <0.001* TCT (µm) 431 ± 40 531 ± 33 <0.001* Astigmatism (D) ACA 4.49 ± 2.16 0.93 ± 0.74 <0.001* PCA 0.90 ± 0.43 0.26 ± 0.14 <0.001* ACA Orientation <0.001† WTR 92 (7.2%) 740 (71.5%) ATR 731 (57.4%) 65 (6.3%) Oblique 450 (35.3%) 230 (22.2%) PCA Orientation <0.001† WTR 805 (63.2%) 10 (1.0%) ATR 94 (7.4%) 925 (89.4%) Oblique 374 (29.4%) 100 (9.7%) M -5.56 ± 4.2 -2.69 ± 1.8 <0.001* Blur 6.24 ± 4.00 2.80 ± 1.81 <0.001* * Mann-Whitney U test. † Chi-square test. ACA: anterior corneal astigmatism, ATR: against-the-rule, Blur: overall blurring strength of the manifest spherocylindrical error, CCT: central corneal thickness, D: diopter, K: keratometry, M: spherical equivalent, PCA: posterior corneal astigmatism, TCT: thinnest corneal thickness, WTR: with-the-rule. Bold values are statistically significant.
ACCEPTED MANUSCRIPT
1
Parameters
ACA (D) Keratoconus Normal
P value
RI PT
Table2. Descriptive measurements of the anterior and posterior corneal astigmatism according to the sex, age groups, and astigmatism orientation of the keratoconus patients and normal subjects PCA (D) Keratoconus Normal
P value
AC C
EP
TE D
M AN U
SC
Sex Males 4.53 ± 2.29 1.00 ± 0.86 <0.001* 0.89 ± 0.45 0.29 ± 0.15 <0.001* Females 4.44 ± 1.97 0.84 ± 0.55 <0.001* 0.91 ± 0.40 0.23 ± 0.11 <0.001* P value 0.933* 0.328* ——— 0.129* <0.001* ——— Age ≤20 4.96 ± 2.41 0.95 ± 0.68 <0.001* 0.95 ± 0.44 0.24 ± 0.11 <0.001* 21-25 4.52 ± 2.09 0.92 ± 0.73 <0.001* 0.91 ± 0.43 0.25 ± 0.13 <0.001* 26-30 4.07 ± 2.01 0.90 ± 0.64 <0.001* 0.88 ± 0.42 0.28 ± 0.15 <0.001* 31-35 4.46 ± 2.37 1.00 ± 1.05 <0.001* 0.88 ± 0.46 0.27 ± 0.15 <0.001* ≥36 4.44 ± 1.95 0.98 ± 0.83 <0.001* 0.82 ± 0.44 0.29 ± 0.15 <0.001* P value 0.002† 0.946† ——— 0.240† 0.003† ——— Orientation WTR 3.50 ± 2.02 1.07 ± 0.80 <0.001* 0.95 ± 0.45 0.10 ± 0.04 <0.001* ATR 4.83 ± 2.27 0.25 ± 0.20 <0.001* 0.63 ± 0.38 0.27 ± 0.14 <0.001* Oblique 4.15 ± 1.86 0.69 ± 0.46 <0.001* 0.85 ± 0.37 0.22 ± 0.10 <0.001* P value <0.001† <0.001† ——— <0.001† <0.001† ——— Vector Component J0 1.05 ± 1.35 -0.25 ± 0.35 <0.001* 1.21 ± 1.33 -0.34 ± 0.45 <0.001* J45 0.02 ± 1.76 -0.11 ± 0.48 0.017* -0.11 ± 1.67 0.19 ± 0.27 <0.001* Data is presented as mean ± SD. * Mann-Whitney U test, † Kruskal-Wallis test. ACA: anterior corneal astigmatism, ATR: against-the-rule, D: diopter, J0 and J45: power vector components of manifest cylinder, PCA: posterior corneal astigmatism, WTR: with-the-rule. Bold values are statistically significant.
ACCEPTED MANUSCRIPT
Table3. Axis orientations of the anterior and posterior corneal astigmatism in keratoconus patients PCA Orientation ATR
AC C
EP
TE D
M AN U
SC
RI PT
Total WTR Oblique ACA Orientation WTR 20 {1.6%} (21.7%) [2.5%] 51 {4.0%} (55.4%) [54.3%] 21 {1.6%} (22.8%) [5.6%] 92 {7.2%} (100%) [7.2%] ATR 648 {50.9%} (88.6%) [80.5%] 17 {1.3%} (2.3%) [18.1%] 66 {5.2%} (9.0%) [17.6%] 731 {57.4%} (100%) [57.4%] Oblique 137 {10.8%} (30.4%) [17.0%] 26 {2.0%} (5.8%) [27.7%] 287 {22.5%} (63.8%) [76.7%] 450 {35.3%} (100%) [35.3%] Total 805 {63.2%} (63.2%) [100%] 94 {7.4%} (7.4%) [100%] 374 {29.4%} (29.4%) [100%] 1273 {100%} (100%) [100%] ACA: anterior corneal astigmatism, ATR: against-the-rule, PCA: posterior corneal astigmatism, WTR: with-the-rule. {% within the whole samples}, (% within each ACA orientation group), [% within each PCA orientation group]
ACCEPTED MANUSCRIPT
Table4. Corneal astigmatism according to the keratoconus severity Parameter
Keratoconus Severity Stage 2 Stage 3 -5.33 ± 3.26 -7.69 ± 4.48 5.95 ± 3.07 8.32 ± 4.25
AC C
EP
TE D
M AN U
SC
RI PT
P value Stage 1 Stage 4 M -3.40 ± 2.75 -9.91 ± 5.71 <0.001* Blur 4.13 ± 2.53 10.70 ± 5.05 <0.001* Anterior corneal astigmatism (ACA) Magnitude (D) 4.16 ± 1.80 4.20 ± 1.84 5.49 ± 2.70 5.17 ± 2.73 <0.001* Orientation <0.001† WTR 43 (46.7%) [10.1%] 33 (35.9%) [6.6%] 4 (4.3%) [2.2%] 12 (13.0%) [7.4%] ATR 218 (29.8%) [51.1%] 283 (38.7%) [56.4%] 116 (15.9%) [64.1%] 114 (15.6%) [69.9%] Oblique 166 (36.9%) [38.9%] 186 (41.3%) [37.1%] 61 (13.6%) [33.7%] 37 (8.2%) [22.7%] J0 0.72 ± 1.16 1.01 ± 1.25 1.46 ± 1.36 1.63 ± 1.75 <0.001* J45 0.01 1.56 0.18 ± 1.78 -0.24 ± 2.00 -0.18 ± 1.89 0.039* Posterior corneal astigmatism (PCA) PCA magnitude (D) 0.85 ± 0.38 0.86 ± 0.38 1.09 ± 0.52 0.96 ± 0.54 <0.001* PCA Orientation <0.001† WTR 224 (27.8%) [52.5%] 330 (41.0%) [65.7%] 121 (15.0%) [66.9%] 130 (16.1%) [79.8%] ATR 29 (30.9%) [6.8%] 41 (43.6%) [8.2%] 12 (12.8%) [6.6%] 12 (12.8%) [7.4%] Oblique 174 (46.5%) [40.7%] 131 (35.0%) [26.1%] 48 (12.8%) [26.5%] 21 (5.6%) [12.9%] J0 0.83 ± 1.08 1.21 ± 1.24 1.48 ± 1.50 1.92 ± 1.69 <0.001* J45 -0.04 ± 1.57 0.05 ± 1.68 -0.41 ± 1.85 -0.48 ± 1.60 0.002* * Kruskal-Wallis test. † Chi-square test. ATR: against-the-rule, Blur: overall blurring strength of the manifest spherocylindrical error, D: diopter, J0 and J45: power vector components of manifest cylinder, M: spherical equivalent, WTR: with-the-rule. (% within each astigmatism orientation group), [% within each severity stage group] Bold values are statistically significant.
ACCEPTED MANUSCRIPT Table5. Receiver operating curve (ROC) analysis AUROC
Asymptotic 95% CI P value Lower bound Upper bound ACA Magnitude 0.958 0.949 0.966 <0.001 PCA Magnitude 0.927 0.914 0.939 <0.001 J0 (PCA) 0.899 0.883 0.915 <0.001 J0 (ACA) 0.863 0.845 0.881 <0.001 Blur 0.798 0.779 0.816 <0.001 J45 (ACA) 0.530 0.503 0.558 0.018 J45 (PCA) 0.442 0.415 0.470 <0.001 M 0.276 0.254 0.298 <0.001 ACA: anterior corneal astigmatism, AUROC: area under receiver operating curve, CI: confidence interval, Blur: overall blurring strength of the manifest spherocylindrical error, J0 and J45: power vector components of manifest cylinder, M: spherical equivalent, PCA: posterior corneal astigmatism. Bold values are statistically significant.
AC C
EP
TE D
M AN U
SC
RI PT
Variables
ACCEPTED MANUSCRIPT Table6. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for predicting keratoconus Sensitivity Specificity PPV NPV (95% CI) (95% CI) (95% CI) (95% CI) ACA ≥ 1.8 D 90.2% 90.2% 92.0% 88.2% (88.4-91.8) (88.3-92) (90.2-93.4) (86.1-90.0) PCA ≥ 0.4 D 89.5% 85.0% 88.0% 86.9% (87.7-91.2) (82.7-87.1) (86.1-89.7) (84.6-88.9) ACA: anterior corneal astigmatism, CI: confidence interval, D: diopter, PCA: posterior corneal astigmatism.
AC C
EP
TE D
M AN U
SC
RI PT
Variables
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT