TWELKER, J DANIEL OD, PhD, FAAO; HARBISON, SHAUNA C. OD; BAILEY, IAN L. OD, MS, FAAO
Ultraviolet radiation (UVR) can damage human tissues, especially anterior ocular tissues that lack the protective effect of melanin. UVC (190 to 280 nm) has the potential for the greatest damage because of its relatively short wavelength and high energy levels. Fortunately, the majority of UVC does not reach Earth’s surface because it is filtered out by ozone in the atmosphere. UVB (280 to 315 nm) is partially absorbed by the cornea (290 to 295 nm) and the crystalline lens (295 to 320 nm). UVA (315 to 400 nm), as well as part of the longer wavelength UVB, can be transmitted to the retina and choroid, but a large proportion is absorbed by the crystalline lens.1,2
UVR and visible radiation incident on the temporal peripheral cornea can be focused at the nasal crystalline lens and limbus. Kwok and Coroneo3 have estimated that the scattered light incident at the temporal limbus can be concentrated up to 20-fold at the nasal limbus. This observation is interesting given that 90% of pterygia are located nasally,4 and their occurrence has been associated with UVR.5,6 Cortical cataracts have been associated with UVR7,8 and are most commonly located in the inferior nasal quadrant of the crystalline lens. Narayanan et al.9 created a model cornea and anterior chamber that was placed in the orbit of a human skull. Using light-sensitive paper in the anterior chamber, they exposed the model to summer sunlight at various times of the day. The inferonasal section of the anterior chamber received the majority of sunlight exposure.
Cullen et al.10 used a modified slitlamp to measure the degree of peripheral light focusing in 20 subjects. They used a converging lens to modify the slit beam and create a collimated beam. While directing this beam toward the temporal limbus, they measured the range of the incident angles at the temporal limbus that created a concentration or bright patch of light at the nasal limbus. These authors also measured refractive error, central and peripheral corneal radius, palpebral fissure height at the temporal limbus, and anterior chamber depth. They found that the magnitude of the temporal catchment angle was positively correlated with anterior chamber depth. The temporal catchment angle was defined as the range of incident angles at the temporal limbus that created a concentration or bright patch of light at the nasal limbus. Maloof et al.11 also investigated the degree to which sharp focus is made at the nasal limbus. Using computer-assisted ray tracing techniques, they speculated that differences in corneal topography account for the clinical observation of individual variation in the degree of peripheral light focusing.
In this study, a new head-mounted instrument was designed and built to measure the temporal catchment angle. We assessed the reliability of measurements with this instrument and investigated ocular parameters that might be associated with the magnitude of the temporal catchment angle.
The instrument was a head-mounted apparatus that allowed a beam of light to be directed toward the temporal limbus of the right or left eye at a variable angle. Attached to a headband, there was a mounting block to which two symmetrical acrylic plates were attached so they were in a horizontal plane. In both plates, there was a pair of concentric slots that allowed a box carrying the light source to be rotated in an arc about a vertical center of rotation located at the temporal limbus of either the right or left eye. The light source was a Welch Allyn (Orange, CA) transilluminator (Model 41100), which had a halogen light source directed into a fiberoptic guide whose exit aperture had a 3.5-mm diameter (Fig 1). A 30-D lens, 19 mm in diameter, was positioned 33 mm from the end of the fiberoptic. The lens was 10 cm from the axis about which the system rotated. The range of angular rotation was from 75 to 140° relative to straight ahead. Angular positions were marked in 5° steps and further subdivided into 1° increments.
Thirty subjects participated in this study, 23 women and 7 men ranging in age from 21 to 37 years. Subjects were recruited from the student body of the University of California, Berkeley School of Optometry. The protocol and informed consent forms were reviewed and approved by the University of California, Berkeley Committee for the Protection of Human Subjects. Informed consent was obtained from the subjects after explaining the nature and possible consequences of the study. The research followed the tenets of the Declaration of Helsinki.
Two examiners (JDT and SCH) completed measurements of the temporal catchment angle on the right eye only of each subject. JDT completed the measures twice, separated by 10 minutes, to assess intraobserver reliability. SCH completed each measure once, and this allowed assessment of interobserver reliability.
The head-mounted instrument was positioned comfortably on the subject’s head. The instrument was adjusted so that the center of rotation of the light beam was centered on the temporal limbus. The subject was asked to keep his or her head still and fixate a large letter on an acuity chart so that the eyes were in primary gaze. The examiner began with the light source rotated all the way back at 140° from the primary line of sight. Fig. 2A shows the light source incident on the temporal limbus before nasal focus appears. The examiner then rotated the light source, moving it anteriorly until the bright patch of light first appeared at the nasal limbus. The point at which the light just began to focus at the nasal limbus was called the posterior limit of the catchment angle. The examiner continued to move the light source anteriorly. Fig. 2B shows the nasal focus at its peak brightness. Eventually, an angle is reached at which the light patch disappears. Fig. 2C shows the point at which the light was no longer focused at the nasal limbus but was refracted onto the nasal iris, called the anterior limit of the catchment angle.
For each measurement, the examiner made three settings of the anterior and posterior limits of the temporal catchment angle. An unbiased observer read and recorded the instrument readings and did not reveal the measurements to the examiners. After the completion of testing for all the subjects, the data were entered into a spreadsheet software program that calculated the mean value for each anterior and posterior limit measurement for each set of measurements. The magnitude of the temporal catchment angle was obtained by subtracting the anterior limit from the posterior limit.
Measurements were made of selected dimensions and ocular parameters of the eyes. A millimeter ruler was used to measure the height of the palpebral fissure at the temporal limbus. A subjective manifest refraction and the best-corrected visual acuity were measured. Central corneal curvature was measured by keratometry. The measurements of refractive error and the keratometry were analyzed to determine the refractive error and the corneal power along the horizontal meridian, and these values were used in the analyses. The midperipheral corneal power at 4 mm temporal to the corneal apex was measured using a Keratron corneal analyzer. This instrument takes a digital image of the anterior eye surface, including the limbus. All the images appeared to be in good focus. Using an image analysis subroutine, we estimated the corneal diameter for each subject. The anterior chamber depth was estimated using the Smith method,12 in which the slit width of a standard slitlamp is narrowed to 2 mm and oriented in the horizontal orientation. With the illumination source at an angle of 60° temporal to the line of sight, the calibrated slit length dial is adjusted until the two inner edges of the reflections from the cornea and crystalline lens just meet. The measure has been shown to be reliable.13,14
Table 1 presents the mean temporal catchment angle for the 30 subjects as measured by examiner 1 (JDT). For measure 1, it was 17.1° (SD, 6.5°), and for measure 2, it was 16.6° (SD, 6.6°). There was no statistically or clinically significant bias between the first and second measurements (t = −0.92; p = 0.37; two-sided paired t-test). The mean difference between measure 1 and measure 2 was 0.5° (SD, 3.0°), and the intraobserver 95% limits of agreement were ±6.2°. Fig. 3 presents the graph of the difference between the measures vs. their mean.15 Bland and Altman16 recommend these graphs when comparing two measures because the use of simple regression graphs and correlation coefficients can be misleading when comparing two similar measures. The x axis shows the mean of the measure 1 and measure 2. The y axis shows the difference between measure 1 and measure 2. The solid line represents the mean difference between the two measures for all the subjects. There is no statistically significant bias between measure 1 and measure 2. The gray shaded area represents the 95% limits of agreement of measures 1 and 2.
To assess interobserver reliability, measurements between two examiners were compared in Table 1. The mean temporal catchment angle of the 30 subjects measured by examiner 2 was 15.9 (SD, 5.6°). The mean difference between the first measurement of examiner 1 and the measurement of examiner 2 was 1.2° (SD, 4.7°), and the interobserver 95% limits of agreement were ±9.7°. There was no statistically or clinically significant difference between the measurements made by examiner 1 and examiner 2 (t = 1.37; p = 0.18; two-sided paired t-test). Fig. 4 presents the graph of the difference between the measures for examiner 1 and examiner 2 vs. their mean.
Other Ocular Parameters and Their Associations with the Temporal Catchment Angle
The mean height of the palpebral fissure at the temporal limbus was 8.8 mm (SD, 1.5 mm); the mean refractive error in the horizontal meridian was −3.36 D (SD, 2.79 D); and the mean anterior chamber depth was 4.1 mm (SD, 0.4 mm). The mean central corneal power and peripheral corneal power in the horizontal meridian were 42.74 D (SD, 1.49 D) and 39.00 D (SD, 2.71 D), respectively. The mean corneal diameter was 11.07 mm (SD, 0.35 mm). Table 2 shows a correlation coefficient matrix for various ocular parameters.
Simple linear regression analysis was performed using the Stata software program (College Station, TX)17 and a Macintosh G3 personal computer (Apple, Cupertino, CA). The magnitude of the temporal catchment angle was not associated with palpebral fissure height or refractive error (p = 0.43 and 0.53, respectively). The magnitude of the temporal catchment angle was associated with central corneal power, peripheral corneal power, anterior chamber depth, and corneal diameter (central corneal power, p = 0.02; peripheral corneal power, p = 0.01; anterior chamber depth, p = 0.002; and corneal diameter, p = 0.002; Fig. 5). Fig. 5A shows the correlation graph for the temporal catchment angle vs. the corneal diameter. The x axis shows the corneal diameter. The y axis shows the magnitude of the temporal catchment angle, and the solid line represents best-fit line (p = 0.002). Fig. 5B shows the correlation graph for the temporal catchment angle vs. the anterior chamber depth. The solid line represents best-fit line (p = 0.002). Fig. 5C shows the correlation graph for the temporal catchment angle vs. the central corneal power. The solid line represents best-fit line (p = 0.018). Fig. 5D shows the correlation graph for the temporal catchment angle vs. the peripheral corneal power. The solid line represents best-fit line (p = 0.012).
A multivariate linear regression analysis was performed to determine which of the many ocular dimensions best predicts the magnitude of the temporal catchment angle. Central corneal power in the horizontal meridian and corneal diameter, in combination, best predicted the magnitude of the temporal catchment angle [y = 2.61 (central corneal power) + 14.22 (corneal diameter) − 251.93; adjusted R2 = 0.73; p < 0.0001]. This nested model was not significantly different from the full model, which included all the parameters [F(5,22) = 1.31; p > 0.05] but was much better than the best bivariate linear model using only corneal diameter [F(1,27) = 37.93; p < 0.001]. The multivariate regression model using central corneal power and corneal diameter explains 73% of the variance associated with the temporal catchment angle.
The Anterior and Posterior Limits of the Temporal Catchment Angle
The mean anterior limit of the temporal catchment angle was 104.1° (SD, 6.5) with a range of 91 to 120° among the 30 subjects. The mean posterior limit of the temporal catchment angle was 121.0° (SD, 5.5) with a range of 111 to 136°. The intraobserver 95% limits of agreement for the anterior limit were 16.6° and for the posterior limit were 15.7°. The interobserver 95% limits of agreement for the anterior limit were 15.0° and for the posterior limit were 13.6°. Fig. 6 shows the correlation graph for the anterior (solid squares) and posterior (open squares) angle limits vs. the corneal diameter. The x axis shows the corneal diameter, and the y axis shows the anterior and posterior angle limits. There is no correlation between the posterior angle limit and corneal diameter or any other ocular parameter. The solid line shows the best-fit line for the correlation between the anterior angle limit and the corneal diameter (p = 0.0006). The anterior limit was also correlated with anterior chamber depth (p = 0.006).
The mean magnitude of the temporal catchment angle was 17.1° with a range from as small as 3° and as great as 32°. This range and relatively normal distribution of the measurements in the study sample suggest that there is substantial variation among persons in the population. This variation supports the notion that the temporal catchment angle could be a risk factor for ocular disease, pterygia or cortical cataract in particular. If everyone had the same angle, or none at all, then the risk would be a constant in the population.
The light originated from behind and to the side of the study subject’s head and was focused to the nasal limbus. If there is a relationship between the temporal catchment angle and ocular disease, and to date there have been no studies to investigate the matter, then the risk could be coming from behind the patient’s head. A person who averts his or her head and eyes away from bright sunlight could escape direct exposure, only to expose the eye to risk from the side and behind. As suggested by Kwok and Coroneo,3 the risk could be much greater than direct exposure because of the approximately 20 times concentration of the light.
Kwok et al.18 made a physical model eye with a UV sensor at the nasal limbus. They exposed the model eye to a laboratory UV light source and ambient sunlight. They reported that maximal UV intensity at the nasal limbus was obtained from an angle of peak incidence 120° from fixation. In our project, using visible light, the range of nasal focus was from a mean anterior limit of 104.1° to a mean posterior limit of 121.0°. We did not measure the angle of peak incidence; rather, we measured the range of incident angles that created nasal focus. The data from Kwok et al. show that the peak intensity is skewed toward the posterior end of the range, allowing us to estimate the peak intensity at 117° from fixation. This estimate is similar to the 120° obtained by Kwok et al. using UVR. The difference may be because of the decreased angle of refraction by the longer wavelength visible light as compared with UVR.
Using multivariate linear regression analysis, it was found that the model that includes central corneal power and corneal diameter best predicts the magnitude of the temporal catchment angle. The model explains 73% of the variance associated with the temporal catchment angle, which is remarkably efficient. Eyes with steep, large-diameter corneas have a high propensity to refract light to the nasal limbus. The relationship is so robust that one may not need to measure the temporal catchment angle. According to these data, a simple measure of central corneal curvature and corneal diameter could fairly successfully predict whether a person will be at risk of increased focus of light rays to the nasal limbus.
Our study found a positive, linear relationship between the temporal catchment angle and central corneal power, peripheral corneal power, and corneal diameter. It was a surprise to find an association between corneal diameter and temporal catchment angle, not to mention its relative strength (p < 0.002). It is possible that this finding is caused by the strong correlation between corneal diameter and anterior chamber depth (Table 2). Subsequent analysis shows that the anterior chamber depth was positively correlated with corneal diameter (p = 0.001). Eyes with large-diameter corneas appear to have deep anterior chambers.
This study is one of the few to measure the temporal catchment angle in vivo. Fortunately, the measurements are reliable. When two independent measures from one examiner are compared, the intraobserver 95% limits of agreement were ±6.2°. Although one always strives for precision, and there is room for improvement in this measure, this level of reliability is within acceptable limits. The limits of agreement are well within the full range of measures from 3 to 32°. The interobserver 95% limits of agreement, where measures from two different examiners were studied, were wider at ±9.7°. This is to be expected because observers will have different thresholds for judging the appearance and disappearance of focused light at the nasal limbus. To improve repeatability, further training and practice could help to make more uniform measures.
The intraobserver and interobserver 95% limits of agreement for the anterior and posterior limits were about twice as high as for the temporal catchment angle. That is, there was less measurement precision with the measure of the anterior and posterior limits as compared with the solid angle between them. The discrepancy reflects the positioning of the instrument between measurements. The instrument is mounted on the subject’s head. The entire instrument can rotate up to 5° or even 10° degrees between measurement sets. It is difficult for the examiner to notice and correct for this rotation. Therefore, the absolute measures of anterior and posterior measures can vary from one measurement set to the next, whereas the difference (the temporal catchment angle) will remain constant because the anterior and posterior limits are affected equally by the rotation. Fortunately, the measure of interest is not the anterior and posterior limits; it is the temporal catchment angle that remains relatively unaffected by the imprecision of the placement of the instrument. Alignment of the apparatus could be improved if the subject were to view the fixation target through a sighting aperture attached to the plates.
In the interest of making an instrument that was relatively inexpensive and practical to use, many approximations and compromises were made. For example, the instrument uses the visible portion of the electromagnetic spectrum instead of UVR. An UVR source, although it is the region of the spectrum that causes damage to ocular tissues, would be an impractical choice for a light source that is purposefully directed toward the research subject’s eye. First, there is a safety concern for all but very low levels of UVR. Furthermore, it would be difficult for the examiner to directly detect the nasal illumination because UVR is not visible to the human eye. UVR has a shorter wavelength than visible light and will have a greater angle of refraction. As a result, it would be expected that the range of incident angles that results in nasal focus to be shifted posteriorly if using UVR and anteriorly while using visible light.
Cullen et al.10 used a different apparatus to measure the temporal catchment angle, although the fundamental concept was the same. They used a slitlamp biomicroscope that was modified to give a collimated beam rather than a focused visible slit. In developing this project, a similar adaptation was attempted. The advantage of the slitlamp is that the viewing optics are coincident with the slitlamp beam at one point in space, which can be positioned at the temporal limbus. Conversely, properly positioning the patient in the chin rest is difficult because the structure of the normal headrest often blocked the illumination beam. Even with young, agile subjects turning their heads to facilitate access to the temporal limbus, the task was difficult, and it would be almost impossible with a more general clinical population. In the end, the head-mounted instrument was selected because it was portable and could be comfortably fitted to practically all the research subjects. The ease of taking the measurement could be important when using such an instrument in a larger clinical study.
Cullen et al. found a positive linear relationship between anterior chamber depth and the temporal catchment angle, and our study also confirmed this result. Given that anterior chamber depth tends to decrease with age because of the increasing thickness of the crystalline lens, there could be a natural protective effect over time. This effect is analogous to the yellowing of the crystalline lens and its associated protective effect on the retina. The time course and latency of the development of pterygia in relation to exposure to ultraviolet light or another risk factor is not known. Even though risk of ultraviolet light exposure at the nasal limbus may decrease over time because of a shallow anterior chamber, sufficient damage may have already been done to cause a pterygium later in life.
Is an eye with a large temporal catchment angle at higher risk for ocular diseases such as pterygium and cortical cataract? At this time, the answer is not known. Epidemiologic studies have shown an association between UVR and pterygium5,6 and UVR with cortical cataract.7,8 The exact nature of the etiologic mechanism has not yet been clearly demonstrated. Anecdotal evidence comes from veterinary science literature in which Taylor and Hanks19 described peripheral light focusing from sunlight in the eyes of Hereford cows. They suggested that this could play a role in the development of bovine ocular squamous cell carcinoma. The next step in this line of investigation could be to study the association between the temporal catchment angle or its correlates of central corneal curvature and corneal diameter, and the occurrence of anterior segment eye diseases such as pterygium and cortical cataract.
A reliable method of measuring the temporal catchment angle was described using an optical instrument. The temporal catchment angle was positively correlated with central corneal power, peripheral corneal power, anterior chamber, and corneal diameter. Central corneal power and corneal diameter, in combination, best predicted the magnitude of the temporal catchment angle. Further investigation could determine whether a large temporal catchment angle is a risk factor for eye diseases associated with UVR.
Supported by National Eye Institute, National Institutes of Health grant K23-EY00372, an Ezell Fellowship from the American Optometric Foundation, Knights Templar Eye Foundation, and Research to Prevent Blindness.
J. Daniel Twelker
University of Arizona
Department of Ophthalmology
655 N. Alvernon Way, Suite 108
Tucson, AZ 85711
1. Kinsey VE. Spectral transmission of the eye to ultraviolet radiations. Arch Ophthalmol (Chicago) 1948;39:508–13.
2. Pitts DG, Kleinstein RN. Environmental Vision. Boston: Butterworth-Heinemann; 1993.
3. Kwok LS, Coroneo MT. A model for pterygium formation. Cornea 1994;13:219–24.
4. Adamis AP, Starck T, Kenyon KR. The management of pterygium. Ophthalmol Clin North Am 1990;3:611.
5. Taylor HR, West SK, Rosenthal FS, Munoz B, Newland HS, Emmett EA. Corneal changes associated with chronic UV irradiation. Arch Ophthalmol 1989;107:1481–4.
6. Taylor HR. Studies on the tear film in climatic droplet keratopathy and pterygium. Arch Ophthalmol 1980;98:86–8.
7. Klein BE, Klein R, Linton KL. Prevalence of age-related lens opacities in a population. The Beaver Dam Eye Study. Ophthalmology 1992;99:546–52.
8. Schein OD, West S, Munoz B, Vitale S, Maguire M, Taylor HR, Bressler NM. Cortical lenticular opacification: distribution and location in a longitudinal study. Invest Ophthalmol Vis Sci 1994;35:363–6.
9. Narayanan P, Merriam JC, Vazquez ME, Dillon J. Experimental model of light focusing of the peripheral cornea. Invest Ophthalmol Vis Sci 1996;37:37–41.
10. Cullen AP, Olanrewaju MO, Voisin AC. Anterior eye focusing of peripheral ultraviolet and visible radiation albedo. Clin Exp Optom 1997;80:80–6.
11. Maloof AJ, Ho A, Coroneo MT. Influence of corneal shape on limbal light focusing. Invest Ophthalmol Vis Sci 1994;35:2592–8.
12. Smith RJ. A new method of estimating the depth of the anterior chamber. Br J Ophthalmol 1979;63:215–20.
13. Barrett BT, McGraw PV, Murray LA, Murgatroyd P. Anterior chamber depth measurement in clinical practice. Optom Vis Sci 1996;73:482–6.
14. Osuobeni EP, Oduwaiye KA, Ogbuehi KC. Intra-observer repeatability and inter-observer agreement of the Smith method of measuring the anterior chamber depth. Ophthalmic Physiol Opt 2000;20:153–9.
15. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10.
16. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346:1085–7.
17. Stata. Statistics/Data Analysis. College Station, TX; 1999.
18. Kwok LS, Kuznetsov VA, Ho A, Coroneo MT. Prevention of the adverse photic effects of peripheral light-focusing using UV-blocking contact lenses. Invest Ophthalmol Vis Sci 2003;44:1501–7.
19. Taylor RL, Hanks MA. Developmental changes in precursor lesions of bovine ocular carcinoma. Vet Med Small Anim Clin 1972;67:669–71.