Peripheral Ocular Aberrations in Mild and Moderate Keratoconus

Abstract

Purpose.: To investigate the influence of keratoconus on peripheral ocular aberrations.

Methods.: Aberrations in seven mild and five moderate keratoconics were determined over a 42° horizontal × 32° vertical visual field, with a modified aberrometer. Control data were obtained from an emmetropic group.

Results.: Most aberrations in the keratoconics showed field dependence, predominately along the vertical meridian. Mean spherical equivalent M, oblique astigmatism J ₄₅, and regular astigmatism J ₁₈₀ refraction components and total root mean square aberrations (excluding defocus) had high magnitudes in the inferior visual field. The rates of change of aberrations were higher in the moderate than in the mild keratoconics. Coma was the dominant peripheral higher-order aberration in both the emmetropes and the keratoconics; for the latter, it had high magnitudes in the center and periphery of the visual field.

Conclusions.: Greater rates of change in aberrations across the visual field occurred in the keratoconus groups than in the emmetropic control group. The moderate keratoconics had more rapid changes in, and higher magnitudes of, aberrations across the visual field than did the mild keratoconics. The dominant higher-order aberration for the keratoconics across the visual field was vertical coma.

Keratoconus is a progressive and usually bilateral condition that affects the cornea. ¹ Noninflammatory progressive central thinning causes the cornea to assume a conical shape with significant reductions in visual performance. ^1,2 The “irregular” astigmatism associated with keratoconus is difficult to correct with spectacles alone. Once visual acuity becomes unsatisfactory with spectacles, rigid gas-permeable contact lenses are typically prescribed that better neutralize the corneal aberrations by providing a spherical refractive surface. ³  

The concepts of wavefront-guided corneal refractive surgery ⁴ and aberration-correcting contact lenses ^5,6 have sparked interest in higher-order aberration studies. Keratoconus patients are one population that has received attention, with the literature correlating their higher-order aberrations with visual performance. ^7,8 Some papers have addressed the diagnosis of keratoconus before clinical signs and symptoms develop, by examining corneal ^2,9 or ocular higher-order aberrations. ¹⁰ Thus, studies furthering our understanding of aberrations are fundamental to the future clinical management of keratoconic and other abnormal corneas. 

Studies have shown significantly greater magnitudes of on-axis (along the line of sight), higher-order, and total eye aberrations—in particular, comalike aberrations—to be present in keratoconus patients compared with normal subjects. ^{9 –13} As an example, Pantanelli et al. ¹³ used a large-dynamic range Hartmann-Shack wavefront sensor to characterize the on-axis aberrations in 32 eyes with keratoconus. Analysis over a 6-mm pupil showed vertical coma to be the dominant higher-order aberration, followed by trefoil, and then by spherical aberration. Every aberration coefficient up to the fifth order was between two and seven times greater in the keratoconus population than in an emmetropia control group. 

Most research into ocular aberrations, both in normal eyes and in eyes with pathologic conditions such as keratoconus, has been concentrated on axial aberrations. So far, studies of peripheral higher-order aberrations have been conducted by only a few groups. ^{14 –18} Peripheral vision is used for tasks such as detection, ^{19 –21} peripheral motion perception, ²² mobility and postural balance, ^23,24 and driving, ²⁵ which, compared with visual acuity, have less demand for image quality. Peripheral vision is limited, not only by image quality, but also by the low-resolution capacity of the eccentric retina. There are several reasons for investigating peripheral aberrations. Best correcting eccentric fixation after central field loss, ²⁶ the possible adverse effects on peripheral tasks after LASIK, ²⁷ and the idea that peripheral defocus is a cause of myopia progression have all generated interest in peripheral refraction. ^28,29 From the few studies measuring peripheral higher-order aberrations it is well established that second-order aberrations dominate in peripheral vision, ^14,26 but higher-order aberrations, in particular coma, can also be substantial. ^14,27 Mathur et al. ^{30 –32} investigated the variations in aberration coefficients across the visual field of young emmetropic subjects. Although many terms varied across the visual field, only a selection showed obvious trends. Increases in the second-order astigmatic coefficients C ₂ ⁻² and C ₂ ² from the center to the periphery along 45° to 225° and 0° to 180° meridians, respectively, were noted, as were decreases in these terms along the meridians perpendicular to the just-mentioned meridians. The vertical coma coefficient C ₃ ⁻¹ increased linearly from the superior to the inferior field, whereas the horizontal coma coefficient C ₃ ¹ increased linearly from the nasal to the temporal field. The rates of change in C ₃ ⁻¹ across the visual field increased with myopia. ³⁰ The interventions of LASIK and orthokeratology changed the sign of the rate of change of coma across the field. ^33,34  

We extended previous work by investigating peripheral aberrations in keratoconic eyes. We hypothesized that there is a greater rate of change in higher-order aberrations across the visual field within a keratoconic population than in an emmetropic one, and that the use of peripheral measures may amplify the differences in ocular aberrations between keratoconic and normal subjects. 

Methods

This research was approved by the Queensland University of Technology's Human Research Ethics Committee and conformed to the tenets of the Declaration of Helsinki. Information regarding the study was given to the subjects and written consent obtained before testing. 

Twelve subjects with keratoconus were recruited from the University's optometry clinic, research department databases, and contact lens specialist private practices. Other ocular disease and severe keratoconus complications, including corneal scarring and acute hydrops, were exclusion criteria. Rigid gas-permeable contact lens wearers were also excluded. If a patient wore soft contact lenses, a period of 1 day and 1 night without lens wear was enforced before testing, to ensure that the lenses' effect on the corneas did not influence the results. 

Keratoconus was confirmed by a scissoring reflex on retinoscopy, central or paracentral steepening on computerized topography with a videokeratoscope (model E300; Medmont International Pty. Limited, Sydney, Australia), and at least one of central or paracentral corneal thinning, Vogt's striae, or a Fleischer's ring. All subjects exhibited bilateral signs of keratoconus, with various degrees of between-eye symmetry. The keratoconus program of the Pentacam (Oculus Inc., Wetzlar, Germany) offered further confirmation of the condition and was adopted to classify our subjects into seven mild (three men, four women) and five moderate (three men, two women) cases. The Pentacam software provides a keratoconus severity classification (from pre-keratoconus and mild keratoconus, KK-1, to advanced keratoconus, KK-4) adapted from the Amsler grading system and based on eight indices derived from the anterior surface topography and corneal thickness progression. ³⁵ We used this classification to grade the severity of keratoconus for each subject and defined mild keratoconus as being less than KK-2, and moderate keratoconus being KK-2 to -3. Most previous studies categorized keratoconus patients on their keratometry values. ^36,37 Given the Pentacam's recent popularity and demonstrated utility for detecting keratoconus, ³⁸ we felt the use of its indices were justified, as it provided an objective assessment of the severity of keratoconus. 

The mean ages were 28 ± 5 years in the mild keratoconus group (range, 21–34) and 30 ± 7 years in the moderate keratoconus group (range, 22–37). Ocular aberration data from the keratoconus populations were compared with a control population, described in detail elsewhere, ^30,32 consisting of 10 young adult emmetropic subjects with a mean age of 25 ± 3 years (age range, 20–30). Mean steep simulated keratometry (as measured by the Medmont corneal topographer) was 44.2 ± 1.5 D (range, 41.5–47.6) in the emmetropes, 46.8 ± 1.1 D (range, 45.1–48.3) in the mild keratoconics, and 46.8 ± 2.2 D (range, 44.8–50.0) in the moderate keratoconics. 

Each subject underwent an ophthalmic examination that included unaided vision, best corrected visual acuity using both high- and low-contrast Bailey-Lovie charts, retinoscopy, and subjective refraction. The spherical equivalent for the mild and moderate keratoconics was 0.00 ± 0.46 and +1.30 ± 1.44 D, respectively. The best spectacle-corrected high (HCVA)- and low (LCVA)-contrast visual acuities of the keratoconics are described in Table 1. In the emmetropes, only HCVA was measured, which was ≤0.00 logMAR in all subjects. The cone location relative to the pupil center was determined with the videokeratoscope (E300; Medmont International, Pty., Ltd.). The ruler function inbuilt in the software was used to measure the vertical and horizontal displacements from the pupil center to the steepest anterior corneal curvature location on the tangential power map (Table 1). 

Table 1.

 

Visual Acuities and Cone Locations in the Two Study Groups

Group	HCVA (logMAR)	LCVA (logMAR)	Cone Location
			x (mm)	y (mm)
Mild keratoconus	0.02 ± 0.10	0.45 ± 0.17	0.5 ± 0.2	−1.6 ± 0.6
Moderate keratoconus	0.07 ± 0.10	0.48 ± 0.08	0.3 ± 0.3	−2.0 ± 0.2

 Best spectacle-corrected HCVA and LCVA contrast visual acuity and cone locations along the horizontal (x) and vertical (y) corneal meridians in the keratoconus groups. Positive x and y values represent nasal and superior cornea, respectively.

Each subject also had corneal thickness and topography measured with a rotating Scheimpflug camera (Pentacam HR; Oculus). The instrument has been found to provide highly repeatable measures of corneal thickness in patients with keratoconus. ^36,39 A Hartmann-Shack aberrometer (Complete Ophthalmic Analysis System–High Definition; COAS-HD; AMO WaveFront Sciences LLC, Albuquerque, NM) was used to determine peripheral aberrations of each subject's right eye by using a procedure that has been described in detail previously. ³¹ A 100 × 75-cm rear projection screen was placed at a distance of 1.2 m, onto which the fixation targets were projected and viewed via a glass slide beam splitter. Targets were arranged in a 6-row × 7-column matrix to give a visual field of 42° × 32°. The center of the fixation target array was aligned with the aberrometer's internal fixation target. Two images only for each fixation target were taken to reduce the subject's total testing time to less than 2 hours. The dynamic range of the Hartmann-Shack sensor limits the measurement of highly aberrated corneas, ¹³ and therefore more severe cases were not investigated. The quality of the images captured by the modified aberrometer was generally high. Only 10 of 916 images were unsuitable for analysis, which meant that only one image, rather than two images, was available for a particular subject and visual field position. For some images, the software was unable to determine the centroids of some of the points. In such cases, these were estimated by a manual procedure performed by one of the authors. The number of points estimated manually was generally small (≤30 points) relative to approximately 766 points across a 5-mm pupil. With the use of an algorithm (MatLab; The MathWorks, Inc., Natick, MA), the image magnification and contrast were increased, and the observer used a cursor to estimate the centroid. Further analysis was performed with custom software that stretched elliptical pupils to circular pupils and converted from the instrument's 840-nm wavelength to 555 nm. ^40,41 Aberration coefficients up to the sixth order were estimated from the wavefront for a pupil diameter of 5 mm for all subjects. Aberration coefficients for the two images at each visual field position were averaged. 

Because of the large intersubject variations within the keratoconus groups, average group values for each aberration coefficient were used. We determined refractive components' mean spherical equivalent M, oblique astigmatism J ₄₅, and with/against the rule astigmatism J ₁₈₀ refraction components, based on second- to sixth-order aberration coefficients. ^40,41 Total root mean squared aberrations excluding defocus (TotalRMS), and higher-order root mean squared aberrations (HORMS) across the visual field were also determined. 

Contour plots were generated to represent refractive components, aberration coefficients, HORMS, and TotalRMS across the visual field. Further analysis was performed along the vertical visual field meridian by using quadratic fits to the data. Quadratic fits were chosen because they most closely represented the change in most of the aberration coefficients. 

The refractive components, second-, third-, and fourth-order aberration coefficients, HORMS, and TotalRMS were further analyzed with repeated-measures analysis of variance (ANOVA) with field position (38 positions) as the within-subject factor and group as the between-subjects factor. Bonferroni post hoc analysis was also performed for the three groups, and ANOVA was performed (SPSS; SPSS Inc., Chicago, IL). 

Results

The average cone location relative to the pupil center with corneal topography was 1.60 ± 0.6 mm inferiorly and 0.45 ± 0.2 mm nasally in the mild keratoconics and 1.97 ± 0.2 mm inferiorly and 0.34 ± 0.3 mm nasally in the moderate keratoconics. The pupils during corneal topography were typically smaller than 5 mm, and in the larger pupil during aberrometry measurements, a mean absolute pupil shift of approximately 0.21 mm would be likely, but with mean changes in the horizontal and vertical locations of less than 0.03 mm, ⁴² the estimated mean cone position relative to the larger pupil would not be expected to change substantially. 

Refraction components and aberration coefficients across the visual field were displayed as two-dimensional (2-D) contour maps, with a common scale for a refraction component/aberration across all three groups. Negative (−) coordinates represent temporal and inferior visual fields. Figure 1 shows the refraction components across the visual field, Figure 2 shows higher-order wavefront error maps across the eccentric pupil for each location in the visual field, Figure 3 shows the third-order coefficients and spherical aberration coefficient across the visual field, and Figure 4 shows HORMS and TotalRMS across the visual field. 

Figure 1.

 

Mean refraction components across the visual field in diopters (D; right axis) in (a) emmetropes, (b) mild keratoconics (Mild K'conus), and (c) moderate keratoconics (Mod K'conus). (A) Spherical equivalent M; (B) with/against-the-rule astigmatism, J ₁₈₀, and (C) oblique astigmatism, J ₄₅. Pupil size, 5 mm. For any refraction component, the scale is the same in all three groups.

Figure 2.

 

Wavefront error maps across the elliptical pupil at 38 visual field locations for (a) emmetropes, (b) mild keratoconics (Mild K'conus), and (c) moderate keratoconics (Mod K'conus). Pupil size, 5 mm. The scale is the same for all three groups.

Figure 3.

 

Mean aberration components across the visual field in micrometers (μm; right axis) for (a) emmetropes, (b) mild keratoconics (Mild K'conus), and (c) moderate keratoconics (Mod K'conus). (A) Vertical coma coefficient C ₃ ⁻¹; (B) horizontal coma coefficient C ₃ ¹; (C) vertical trefoil coefficient C ₃ ⁻³; (D) horizontal trefoil coefficient C ₃ ³; spherical aberration coefficient C ₄ ⁰. Pupil size, 5 mm. For any coefficient, the scale is the same in all three groups.

Figure 4.

 

Combined aberrations across the visual field in micrometers (μm; right axis) in the (a) emmetropes, (b) mild keratoconics (Mild K'conus), and (c) moderate keratoconics (Mod K'conus). (A) HORMS; (B) TotalRMS. Pupil size, 5 mm. For any combined aberrations, the scale is the same in the three groups.

The rates of change in the refraction components and higher-order coefficients across the visual field were greater in the keratoconus groups than in the emmetropia control group for all aberration terms considered. The moderate keratoconics exhibited greater rates of change than did the mild keratoconics for all aberrations. 

For spherical equivalent M (Fig. 1A), the emmetropes had little variation across the field, whereas the mild keratoconics had high negative (myopic) values in the inferior field, which became less negative in the superior field. The moderate keratoconics had a similar pattern, but the magnitude in the inferior field and rate of change were higher. 

Astigmatism J ₁₈₀ (Fig. 1B) in the emmetropes decreased quadratically from the center to the periphery along the 0° to 180° meridian and increased quadratically along the 90° to 270° meridian. The keratoconus groups had high negative J ₁₈₀ in the superior field that became less negative and eventually positive in the inferior field. The change was quadratic and greater in the inferior field than in the superior field. The moderate keratoconics had greater rate of change than did the mild keratoconics. 

Oblique astigmatism J ₄₅ (Fig. 1C) in the emmetropes decreased from the central to the peripheral field along the 45° to 225° meridian and increased along the 135° to 315° meridian. In the mild and moderate keratoconics, it decreased linearly from the inferior nasal field to the superior temporal field along the 315° to 135° meridian. The moderate keratoconics exhibited a greater rate of change than did the mild keratoconics. 

Figure 2 shows the mean higher-order wavefront error maps across the pupil at each of the 38 visual field positions. In the emmetropes, coma dominated the peripheral visual field. Both keratoconus groups had larger aberrations than did the emmetropes at any visual field position, with the moderate keratoconics having the highest aberrations. Aberrations were again dominated by the comas in the keratoconics, but with some influence of spherical aberration in the inferior visual fields, as indicated by the increase in symmetry of the plots and with some influence of trefoil in the inferior–nasal fields, as indicated by the three-lobed nature of the plots. The mild keratoconics had their most aberrant wavefronts in the superior–temporal field, and the moderate keratoconics had their most aberrant wavefronts along the horizontal meridian. In the emmetropes, the axis of the combined horizontal and vertical comas approximately matched the visual field meridian. The axis of coma in the keratoconics was mainly vertical across the visual field, indicating the dominance of vertical coma, although with some rotation from this in the superior visual field, indicating the influence of horizontal coma. 

The vertical coma coefficient C ₃ ⁻¹ (Fig. 3A) in the emmetropes increased linearly from the superior to the inferior field. In the mild keratoconics, it increased quadratically from the superior to the inferior field. The moderate keratoconics had a high negative C ₃ ⁻¹ in the center of the visual field, which became less negative both superiorly and inferiorly. The horizontal coma coefficient C ₃ ¹ (Fig. 3B) increased linearly from the nasal to the temporal fields in all three groups, with the variation in the moderate keratoconics rotated toward the 135° to 315° meridian. The rates of change were similar in the mild (−0.018 μm/deg) and moderate (−0.016 μm/deg) keratoconics, and these were approximately 2.5 times greater than those in the emmetropes (−0.007 μm/deg). 

The vertical trefoil C ₃ ⁻³ and horizontal trefoil C ₃ ³ coefficients (Figs. 3C, 3D) varied very little across the visual field in the emmetropes. In the keratoconics, however, there was considerable variation in both coefficients, with high negative and high positive values evident in the peripheral fields, particularly in the moderate keratoconus group. 

The spherical aberration coefficient C ₄ ⁰ (Fig. 3E) in the emmetropes varied little across the visual field. In the keratoconics, it increased from the inferior to the superior field. Rates of change were greater in the moderate than in the mild keratoconics, with peak positive values in the moderate keratoconics in the midsuperior field. 

HORMS (Fig. 4A) in all groups showed quadratic variation along the vertical meridian. The magnitudes and rates of change were more pronounced in the moderate keratoconics than in the other groups. As an estimation of variability in aberrations along the vertical field meridian, the standard deviations of HORMS and TotalRMS were calculated for each subject. The average standard deviations of HORMS in the mild and moderate keratoconics (0.20 and 0.30 μm, respectively) were substantially larger than in the emmetropes (0.05 μm). The similarities between the magnitudes and variations of HORMS and vertical coma (Fig. 3A) make it clear that vertical coma was the dominant higher-order aberration across the field. 

TotalRMS (Fig. 4B) in the emmetropes increased quadratically from the center to the periphery. The keratoconics showed increase in TotalRMS from the superior to the inferior of the visual field. The vertical rates of change were similar in the mild and moderate keratoconics (−0.041 μm/deg and −0.052 μm/deg, respectively), but the magnitudes were greater in the latter at any visual field position. The average standard deviations in the mild and moderate keratoconics along the vertical meridian (0.72 and 0.60 μm, respectively) were substantially larger than that in the emmetropes (0.11 μm), indicating the high variability between the individuals within each keratoconus group. 

Table 2 shows the ANOVA results. Repeated-measures ANOVA confirmed the differences in refractive components, aberration coefficients, and HORMS and between the three groups. Refractive group had a significant effect on all refractive components and most aberration coefficients, except C ₂ ⁰, C ₂ ², C ₃ ¹, C ₄ ⁻², C ₄ ⁻⁴ and C ₄ ⁰. As expected, field position had a significant effect on all refractive components and aberration coefficients. There were significant group-field interactions for all the terms, showing that group significantly affected the pattern of aberrations across the field. Bonferroni post hoc analysis showed that most of the terms were significantly different between the emmetropes and keratoconics. M, J ₁₈₀, C ₃ ⁻¹, C ₄ ², C ₄ ⁴, and HORMS differed significantly between the mild and moderate keratoconics. 

Table 2.

 

Probabilities of the Repeated-Measures ANOVA for Refraction Components and Aberration Coefficients

Refraction Component/Aberration Coefficient	Group	Field Position	Group–Field Position Interaction*	Bonferroni Post Hoc Analysis of Groups
				Emm vs. Mild K'conus	Emm vs. Mod K'conus	Mild K'conus vs. Mod K'conus
J 45	0.00*	0.00*	0.00*	0.04*	0.00*	0.28
M	0.01*	0.00*	0.00*	1.00	0.03*	0.02*
J 180	0.04*	0.00*	0.00*	0.16	1.00	0.05*
C 2 −2	0.00*	0.00*	0.00*	0.05*	0.00*	0.08
C 2 0	0.40	0.00*	0.00*
C 2 2	0.08	0.00*	0.00*
C 3 −3	0.00*	0.00*	0.00*	0.00*	0.00*	0.77
C 3 −1	0.00*	0.00*	0.00*	0.00*	0.00*	0.00*
C 3 1	0.10	0.00*	0.00*
C 3 3	0.01*	0.00*	0.00*	0.03*	0.01*	1.00
C 4 −4	0.06	0.00*	0.00*
C 4 −2	0.06	0.00*	0.00*
C 4 0	0.08	0.00*	0.00*
C 4 2	0.00*	0.00*	0.00*	1.00	0.00*	0.00*
C 4 4	0.00*	0.00*	0.00*	0.05*	0.00*	0.00*
HORMS	0.00*	0.00*	0.00*	0.00*	0.00*	0.00*
TotalRMS	0.02*	0.00*	0.00*	0.07	0.03*	1.00

 Refraction components, Zernike aberration coefficients and root-mean-square aberrations with within-subjects factor of field position and between-subjects variable of group. The refraction component M and the defocus coefficient C ₂ ⁰ are relative to their central field values for each subject. Emm, emmetropia; Mild K'conus, mild keratoconus; Mod K'conus, moderate keratoconus.

*  Significant effects (P ≤ 0.05). P < 0.005 is given as 0.00.

Aberrations are affected by cone location. ⁴³ In our keratoconus groups, the cone locations were usually more inferior than nasal, which resulted in higher aberrations for the inferior than for the superior visual fields. We compared the anterior surface power derived from the Medmont corneal topography data along the vertical corneal meridian (Fig. 5, passing through the pupil center) with the ocular refraction and aberrations along the vertical visual field meridian (Fig. 6). The corneal power changed little in the emmetropes, but showed considerable changes in the keratoconus groups (Fig. 5). In the inferior field, the corneal power was, as expected, highest in the moderate keratoconus group. In the superior field, corneal power in the moderate keratoconics was considerably lower than in the other groups. 

Figure 5.

 

Axial anterior corneal power along the vertical meridian (passing through the pupil center, as estimated by corneal topography), relative to the distance from the pupil center in the emmetropes, mild keratoconics, and moderate keratoconics. Error bars, SD; I, inferior cornea; S, superior cornea.

Figure 6.

 

Refraction and aberrations along the vertical visual field meridian. (a) Oblique astigmatism J ₄₅; (b) spherical equivalent M; (c) with/against-the-rule astigmatism J ₁₈₀; (d) vertical trefoil coefficient C ₃ ⁻³; (e) vertical coma coefficient C ₃ ⁻¹; (f) spherical aberration coefficient C ₄ ⁰; (g) HORMS; and (h) TotalRMS. Lines: the quadratic fits to the data; error bars: standard deviations. Different refractions and aberrations have different scales.

Generally the refractive components and aberrations changed quadratically along the vertical meridian in the keratoconus groups. Astigmatic components J ₄₅ and J ₁₈₀ became more positive (Figs. 6a, 6c), and the spherical equivalent M became more negative (Fig. 6b) in the inferior field compared with the respective results in the emmetropic group. The rates of change in the refractive components were highest in the moderate keratoconics followed by those in the mild keratoconics. Vertical trefoil coefficient C ₃ ⁻³ (Fig. 6d) changed little in the emmetropes and mild keratoconics, but was more positive overall across the field in the mild keratoconics than in the emmetropes. It increased linearly from the superior to the inferior field in the moderate keratoconics. Vertical coma coefficient C ₃ ⁻¹ (Fig. 6e) showed the most prominent differences between the three groups. It changed linearly vertically in the emmetropes and quadratically in the mild and moderate keratoconics. The moderate keratoconics also had a high negative vertical coma coefficient C ₃ ⁻¹ across the vertical field meridian. Spherical aberration coefficient C ₄ ⁰ (Fig. 6f) changed little in the emmetropes, but changed linearly in the mild keratoconics and quadratically in the moderate keratoconics. 

HORMS (Fig. 6g) and TotalRMS (Fig. 6h) were highest across the field in the moderate keratoconics followed by those in the mild keratoconics. HORMS decreased as distance from the center of the field increased in the moderate keratoconics, increased toward the superior field in the mild keratoconics, and was dominated by changes in vertical coma C ₃ ⁻¹ in the two groups. TotalRMS increased toward the inferior field because of the increase in astigmatism in the inferior field. The mild and moderate keratoconus groups exhibited five and nine times greater axial TotalRMS, respectively, than the emmetropic subjects. In the inferior field, TotalRMS increased to 6 (mild keratoconus) and 11 (moderate keratoconus) times greater in the keratoconic than in the control subjects. In the keratoconic subjects, a significant correlation was found between the quadratic component of the rate of change in aberrations along the vertical meridian and the corneal power at the cone apex for both vertical coma coefficient (r ² = 0.58, P < 0.001) and HORMS (r ² = 0.41, P = 0.03), indicative of a greater change in aberrations along the vertical meridian for greater corneal powers (Fig. 7). 

Figure 7.

 

Second-order fitting coefficients along the vertical visual field, as a function of axial corneal power (in diopters) at the cone apex in the mild and moderate keratoconics. (a) Vertical coma coefficient C ₃ ⁻¹; (b) HORMS.

Discussion

This study is the first conducted to investigate and quantify peripheral ocular aberrations in keratoconic subjects. The results support our hypothesis of a greater rate of change in aberrations across the visual field in keratoconic than in normal eyes. The magnitudes of aberrations and rates of their change across the visual field were greater for a moderate than for a mild keratoconus group, and with considerable variation within the keratoconus groups. Second-order terms were the dominant aberrations across the visual field for both emmetropes and keratoconics. Rapid changes in M and J ₁₈₀ were noted in the inferior field in keratoconics, with as much as −8.00 D of spherical error occurring in the moderate keratoconus group. For 5-mm pupils, horizontal and vertical comas were the dominant higher-order aberrations for emmetropes across the visual field, but vertical coma alone was the dominant higher-order aberration for keratoconics across most of the visual field. The latter is consistent with on-axis studies of keratoconus. ^10,13  

The characteristics of the peripheral aberrations in our keratoconic population appeared to be related largely to the location and magnitude of the cone. Horizontal coma decreased from the temporal to the nasal field in all three groups. In the keratoconus groups the horizontal coma change appeared to be rotated toward the 135° to 315° meridian. The rotation is explained by the cones being typically located in the inferior–nasal quadrant. Furthermore, the cornea power at the cone apex was associated with the change in certain aberrations (i.e., vertical coma and HORMS) along the vertical field. 

With the advent of laser refractive surgery, there has been considerable interest in identifying those patients with early or subclinical forms of keratoconus, as the presence of keratoconus, even in a subclinical form, is considered a major risk factor in the development of iatrogenic keratoectasia after refractive surgery. ⁴⁴ Although videokeratoscopy is still the most commonly used clinical tool in the diagnosis of keratoconus, several other novel methods may be useful in detecting early keratoconus, including those based on measures of corneal thickness, ³⁸ posterior corneal topography, ³⁹ and on-axis corneal ^2,9 and ocular ^1,45 aberrations. The aberrations are, of course, an indirect measure of the corneal topography. This study shows that substantial differences exist in the magnitude and rate of change of peripheral ocular aberrations between emmetropes and subjects with mild keratoconus. The results suggest that metrics based on peripheral ocular higher-order aberrations or the rate of change in these aberrations across the field would be of use in identifying patients with early or subclinical forms of keratoconus; a complicating factor is that myopes tend to have greater rates of change of coma across the visual field than do emmetropes. ³⁰ Further research is needed in a larger sample of subjects to determine the optimum metrics and the clinical utility and sensitivity of peripheral aberration measurements in the diagnosis and screening of early keratoconus. The lengthy time taken for data collection and analysis with currently available clinical instruments probably precludes these measurements from being viable clinically at this stage. However, with further advancements in aberrometer designs, it is likely to be possible to measure peripheral aberrations in a few minutes. 

It has been hypothesized that peripheral ocular aberrations may be an important factor that influences the growth of the eye and development of myopia, ⁴⁶ and there has been substantial increased interest in peripheral aberrations and refractive error in recent years. ^30,47 Imposed defocus on the peripheral retina has been found to lead to myopic eye growth in monkeys. ^48,49 Furthermore, defocus imposed on local retinal regions in chicks and primates has been found to lead to compensatory eye growth, localized to those regions of the retina. ^{50 –54} Recent work investigating peripheral aberrations in human myopes and emmetropes has demonstrated some difference in peripheral aberrations associated with refractive error (e.g., a greater rate of change in vertical coma across the field in myopes). ³⁰ Although it is unknown at this time whether the increased peripheral aberrations measured in myopic eyes contribute to the development of myopia or are simply a consequence of the ocular biometric changes associated with myopia, the substantial peripheral aberrations evident in keratoconus (particularly vertical coma) may have implications for eye growth control in these subjects. If defocus of the peripheral retina is essential in the control of eye growth in human subjects and if the onset of keratoconus is at a time when the visual system is capable of undergoing axial eye growth in response to peripheral defocus, then the substantial asymmetry in defocus that we observed across the vertical visual field in our keratoconic subjects could lead to compensatory asymmetric growth of the peripheral regions of the eye in these subjects. Given that the typical age of onset of keratoconus is often in the teenage years ^1,55 and that studies of refractive error development have demonstrated the human eye is capable of undergoing axial eye growth at this age, ^{56 –59} future studies using magnetic resonance imaging or peripheral ocular length measures (e.g., partial coherence interferometry) in keratoconic subjects may help to determine whether asymmetric axial eye growth occurs in response to the substantially asymmetric retinal defocus in these subjects. 

In conclusion, we have described in detail the magnitude and pattern of peripheral second- and higher-order ocular aberrations in populations of mild and moderately advanced keratoconic subjects and have compared these aberrations to those from an emmetropic control group. Consistent with previous studies of on-axis aberrations in keratoconus and with the measured corneal topographical changes, our keratoconic subjects exhibited higher magnitudes and rates of change in peripheral ocular aberrations compared with the emmetropic controls, with comatic terms being the dominant higher-order peripheral aberrations. 

Footnotes

 Supported by Australian Research Council Discovery Grant DP0558209.

Footnotes

 Disclosure: D.A. Atchison, None; A. Mathur, None; S.A. Read, None; M.I. Walker, None; A.R. Newman, None; P.P. Tanos, None; R.T. McLennan, None; A.H. Tran, None

The authors thank optometrists Geoff Conwell and Kate Johnson for help with the recruitment of the participants.