Abstract
purpose. The pattern of neuroretinal rim loss and increase in the area of parapapillary atrophy in glaucoma depend on the localization of the central retinal vessel trunk in the lamina cribrosa. The purpose of the present study was to determine whether, in a similar way, the pattern of rim loss and progression of parapapillary atrophy are influenced by the presence and position of cilioretinal arteries.
methods. Color stereo optic disc photographs (15°) for morphometric evaluation of the optic nerve head were used to compare the appearance of the optic disc in 41 patients exhibiting unilateral or bilateral cilioretinal arteries in the temporal horizontal disc region with the optic disc morphology of 127 patients without cilioretinal arteries. The areas of the neuroretinal rim and alpha and beta zones of parapapillary atrophy were measured in the total disc and in four disc sectors.
results. Eyes with and eyes without cilioretinal arteries did not differ significantly in the areas of neuroretinal rim and alpha and beta zones of parapapillary atrophy, when measured in the whole optic disc and in the four disc sectors separately; in ratios of the temporal horizontal area to total area of rim and parapapillary atrophy; and in the ratio of temporal horizontal rim area-to-nasal rim area, neither in an interindividual comparison nor in an intraindividual intereye comparison.
conclusions. In contrast to the position of the central retinal vessel trunk, presence and position of cilioretinal arteries do not markedly influence the pattern of neuroretinal rim loss and progression of parapapillary atrophy in glaucoma.
Recent studies have suggested that the pattern of loss of neuroretinal rim and progression of parapapillary atrophy in eyes with chronic open-angle glaucoma is influenced by the location of the exit of the central retinal vessel trunk on the lamina cribrosa.
1 2 The longer the distance to the vessel trunk, the more marked were the decrease in neuroretinal rim, the increase in the beta zone of parapapillary atrophy, and the perimetric loss in the corresponding region of the visual field. Because cilioretinal arteries are anatomically independent of the central retinal vessel trunk but still contribute to the anatomy and physiology of the optic nerve head, we performed the present study to determine whether a cilioretinal artery, similar to the central retinal vessel trunk, influences the pattern of rim loss and the progression of parapapillary atrophy in chronic open-angle glaucoma.
The study included 168 patients (91 women, 77 men) with primary or secondary open-angle glaucoma with a mean age of 55.8 ± 11.4 years and a mean refractive error of −1.15 ± 2.72 D. Eyes with a myopic refractive error exceeding −8 D were excluded due to a different optic disc morphology.
3 The patients were participants in an ongoing glaucoma study (Erlangen Glaucoma Registry). They were selected for inclusion in the present study if they exhibited one or more cilioretinal arteries at the temporal disc border or if they did not show any cilioretinal artery. For an interindividual comparison only one eye per patient was included in the study. For an intraindividual comparison, both eyes of the patient were included if one eye displayed a cilioretinal artery and the fellow eye did not. The methods applied in the study adhered to the tenets of the Declaration of Helsinki for the use of human subjects in biomedical research. Informed consent was obtained from each subject before enrollment. Institutional Review Board and Ethics Committee approvals had been obtained at the baseline of the ongoing glaucoma study (Erlangen Glaucoma Registry).
Criteria for the diagnosis of open-angle glaucoma were an open anterior chamber angle, and glaucomatous visual field defects or glaucomatous abnormalities of the optic nerve head. Glaucomatous visual field defects were defined as a field (Octopus G1; Octopus automated perimeter; Interzeag, Schlieren Switzerland) with (1) at least three adjacent test points having a deviation of equal to or greater than 5 dB and with one test point with a deviation greater than 10 dB, (2) at least two adjacent test points with a deviation equal to or greater than 10 dB, (3) at least three adjacent test points with a deviation equal to or greater than 5 dB abutting the nasal horizontal meridian, and (4) elevated global visual field indices. The number of false-positive or false-negative answers had to be equal to or less than 15% of the total. Glaucomatous abnormalities of the optic nerve head included localized and/or diffuse loss of retinal nerve fiber layer; an abnormally small size and an unusual shape of the neuroretinal rim—for example, with the presence of rim notches; and an optic cup too large in relation to the area of the optic disc.
3 Parapapillary atrophy was not determined for diagnosis.
The total glaucoma group was composed of 94 patients with primary open-angle glaucoma with elevated intraocular pressure without a detectable reason for it; 25 patients with secondary open-angle glaucoma with elevation of intraocular pressure due to pseudoexfoliation of the lens (pseudoexfoliative glaucoma) or primary melanin dispersion syndrome (pigmentary glaucoma); and 49 patients with normal-pressure glaucoma. For the diagnosis of normal-pressure glaucoma, the maximal intraocular pressure readings had to be equal to or less than 21 mm Hg in at least two 24-hour pressure profiles containing measurements at 5 PM, 9 PM, midnight, 7 AM, and noon. Besides glaucoma, other reasons for optic nerve damage had been ruled out by general ophthalmic examination, and in case of doubt, by additional neurologic and neuroradiologic examinations including a magnetic resonance imaging scan.
For all eyes, 15° color stereo optic disc transparencies had been taken with a telecentric fundus camera (Zeiss, Oberkochen, Germany). The disc slides were projected in a scale of 1 to 15. The outlines of the optic cup, optic disc, peripapillary scleral ring, and parapapillary atrophy zones alpha and beta were plotted and morphometrically analyzed. To obtain data in absolute size units (i.e., millimeter or square millimeter), the ocular and photographic magnification was corrected using the Littmann method.
4 The optic cup was defined on the basis of contour and not of pallor. The border of the optic disc was identical with the inner side of the peripapillary scleral ring. Parapapillary atrophy was defined as a central beta zone, with visible sclera and visible large choroidal vessels, differentiated from a peripheral alpha zone with irregular pigmentation. The method has already been described in detail.
5 To obtain measurements in different regions of the optic nerve head, the intrapapillary and parapapillary regions were divided into four sectors. The temporal superior sector and the temporal inferior disc sector were right-angled and were tilted 15° temporal to the vertical optic disc axis. The two other sectors covered the remaining area. The temporal horizontal sector was 60° wide, and the nasal disc sector covered 120°. The areas of parapapillary atrophy and the neuroretinal rim were measured as a whole and separately in each of the four disc sectors
(Tables 1 2) .
For all patients included in the study, the optic disc photographs were searched for the presence of cilioretinal arteries, which were defined as arteries arising at or near the edge of the disc, bending with a characteristic hook before crossing the disc margin, and showing no connection with the central retinal artery or one of its branches. Based on the presence or absence of cilioretinal arteries, the total study group of 168 patients was divided into a group with cilioretinal arteries (
n = 41) and a group without cilioretinal arteries (
n = 127). The two study groups did not differ significantly in age, gender, refractive error, area of the optic disc, mean perimetric mean defect, and type of glaucoma
(Table 1) . In 17 (41.5%) of the 41 patients with cilioretinal arteries, the cilioretinal arteries occurred bilaterally. In 24 (58.5%) patients, the cilioretinal arteries were found unilaterally.
For an interindividual comparison, only one eye per patient was included in the statistical analysis. In the patients with unilateral cilioretinal arteries, the eye containing the cilioretinal artery was included. In patients with bilateral cilioretinal arteries and in the patients without cilioretinal arteries, one eye per patient was chosen at random for inclusion in the statistical analysis. Thus, 41 eyes with cilioretinal arteries were included in the interindividual comparison. The cilioretinal arteries were located in the temporal upper sector in 6 (14.6%) eyes, in the temporal horizontal sector in 27 (65.8%) eyes, and in the temporal lower sector in 8 (19.5%) eyes. One eye displayed two cilioretinal arteries with one artery located in the temporal horizontal sector and one artery located in the temporal lower optic disc sector. For an intraindividual comparison, both eyes were included of patients who showed a cilioretinal artery in one eye.
The presence of cilioretinal arteries was tested on its influence on sectorial data of the morphology of the optic disc by evaluating differences between the two study groups (Mann-Whitney test, χ
2 test, respectively,
Table 1 ) or between the respective fellow eyes (Wilcoxon signed rank test or McNemar test, respectively,
Table 2 ). Because multiple tests were performed, level of significance was set to 0.01 (two-sided). Statistical analyses were performed on computer (SPSSWIN, ver. 10.1; SPSS, Chicago, IL).
Taking into account a sample size of 41 eyes with cilioretinal arteries and 127 eyes without cilioretinal arteries, a significance level of 0.01 (two-sided), and a power of 80%, differences of 0.42 SD (ratio temporal horizontal rim area to nasal rim area) to differences of 0.75 SD (total area of parapapillary atrophy zone beta) between eyes with cilioretinal arteries and eyes without cilioretinal arteries can be detected. Referring to a sample size of 27 eyes with cilioretinal arteries at the temporal horizontal disc sector and 127 eyes without cilioretinal arteries, to a significance level of 0.01 (two-sided), and to a power of 80%, differences of 0.65 SD (total area of parapapillary atrophy zone alpha) to differences of 0.96 SD (total area of parapapillary atrophy zone beta) between eyes with and without cilioretinal arteries can be detected.
The study group with cilioretinal arteries and the study group without cilioretinal arteries did not differ significantly in the total neuroretinal rim area, the neuroretinal rim area measured separately in the four disc sectors, the ratio of the temporal horizontal rim area to total rim area, and the ratio of the temporal horizontal rim area to the nasal rim area
(Table 1) . Correspondingly, both study groups did not vary significantly in frequency and, if present, total area and area in the four disc sectors of the beta or alpha zone of parapapillary atrophy, in the ratio of temporal horizontal beta zone area to total beta zone area, and in the ratio of temporal horizontal alpha zone to total alpha zone area
(Table 1) .
Similar results for all comparisons were obtained, if only eyes with cilioretinal arteries in the temporal horizontal disc sector (n = 27) were compared with eyes without a cilioretinal artery (for all comparisons: P > 0.15)
Taking into account only patients with unilateral cilioretinal arteries and performing an intraindividual intereye comparison, the eyes with cilioretinal arteries and the fellow eyes without cilioretinal arteries did not differ significantly in the total area of the neuroretinal rim, area of the neuroretinal rim in the four disc sectors, the ratio of temporal horizontal rim area to total rim area, and the ratio of temporal rim area to nasal rim area
(Table 2) . In a parallel manner, the eyes with cilioretinal arteries and the contralateral eyes without cilioretinal arteries did not vary significantly in the occurrence and area of beta zone of parapapillary atrophy, measured as a whole and separately in the four sectors, or in the ratio of the area of beta zone in the temporal horizontal sector to the total area of beta zone
(Table 2) .
In the interindividual and the intraindividual comparisons, mean visual field loss was slightly, but not significantly, larger in the eyes with a cilioretinal artery than in the eyes without a cilioretinal artery
(Tables 1 2) . These perimetric measurements were obtained at the baseline of the study when some of the participants were relatively inexperienced in perimetry. In an additional analysis, comparing visual field results of follow-up visits of participants who had undergone a minimum of three computerized perimetric examinations, the visual field parameters mean defect, corrected loss variance and mean defect of the central 10° no longer showed significant differences between eyes with cilioretinal arteries (
n = 20) and eyes without cilioretinal arteries (
n = 70) (
P > 0.70; Mann-Whitney test). Correspondingly, in the intraindividual intereye comparison of patients who had undergone at least three computerized visual field examinations (
n = 12), the perimetric parameters did not show significant differences between eyes with cilioretinal arteries and their respective fellow eyes without cilioretinal arteries (
P > 0.75; Wilcoxon signed rank test).
Cilioretinal arteries, which are present in 6% to 32% of normal eyes, generally arise from the short posterior ciliary artery circulation, or, rarely, directly from the choroidal circulation.
6 7 8 9 10 11 12 13 14 They are true end-arteries supplying a distinct region of the retina. Infrequently, the branches of cilioretinal arteries also supply the prelaminar region.
6 7 8 9 10 11 12 13 14
The role of cilioretinal arteries in glaucoma has been studied with equivocal results so far. Shihab et al.
15 studied 20 consecutive patients with primary open-angle glaucoma and unilateral cilioretinal arteries. The eyes with cilioretinal arteries had a larger average cup-to-disc ratio than fellow eyes without cilioretinal arteries and 15 eyes with cilioretinal arteries showed greater visual field damage than fellow eyes without cilioretinal arteries. In 19 subjects with nonglaucomatous eyes with unilateral cilioretinal vessels, they found no significant difference between the cup-to-disc ratios. They proposed that cilioretinal arteries shunt blood from the posterior ciliary artery circulation away from the optic nerve, which worsens an already compromised situation in glaucoma.
In an interindividual comparison, Lee and Schwartz
16 compared 13 eyes with open-angle glaucoma that exhibited one or two cilioretinal arteries with 21 eyes with open-angle glaucoma that had no cilioretinal arteries.
16 The eyes with cilioretinal arteries significantly more often retained central visual field and a visual acuity of 20/40 or better than eyes without the cilioretinal artery. The authors discussed an improved circulation to the temporal rim of the optic disc which may aid in retaining central field and visual acuity in advanced open-angle glaucoma.
Lindenmuth et al.
17 examined 21 patients with open-angle glaucoma and unilateral cilioretinal artery. They evaluated cup-to-disc ratios derived from a masked evaluation of disc photographs and Goldmann kinetic visual fields. No significant optic nerve or visual field differences were detected in eyes with cilioretinal arteries compared with fellow eyes without cilioretinal artery. Similarly, Mikelberg et al.
18 failed to demonstrate differences in total neuroretinal rim area and global indices of automated visual fields between both eyes of 33 patients with normal-tension glaucoma with unilateral cilioretinal arteries.
18 In an examination of 156 glaucomatous eyes, Chihara and Honda
19 examined the effect of various parameters on the nerve fiber layer defects measured by black-and-white fundus photographs taken under red-free illumination. The presence of cilioretinal vessels in 35 eyes had no influence on the local preservation of the retinal nerve fiber layer in defective areas.
In an both interindividual and intraindividual comparisons, the present study did not demonstrate an effect of cilioretinal arteries on morphologic glaucomatous damage at the optic disc. Global as well as segmental morphometric data for neuroretinal rim and parapapillary chorioretinal atrophy did not differ significantly between eyes with open-angle glaucoma with cilioretinal arteries and those with open-angle glaucoma without cilioretinal arteries. Especially in the temporal disc sector, no significant differences were found in absolute or relative measurements of the areas of the neuroretinal rim and parapapillary chorioretinal atrophy. Of note, the visual field data obtained at baseline of the study pointed toward a more marked loss of visual field in eyes with cilioretinal arteries than in eyes without cilioretinal arteries
(Tables 1 2) . In a further statistical analysis including perimetric follow-up examinations with a higher reliability; however, the differences in perimetric parameters between eyes with cilioretinal arteries and eyes without were markedly outside the range of statistical significance (
P > 0.70). These data support with the finding that the morphometric measurements did not vary significantly between the eyes with cilioretinal arteries and the eyes without cilioretinal arteries.
The rationale for the study was the recent observation that the vicinity of the trunk exit of the central retinal vessels on the lamina cribrosa had a protective effect on neighboring segments of the optic disc.
1 2 The loss of neuroretinal rim was more marked, and the extension of parapapillary choroidal atrophy was more pronounced in the disc quadrant opposite the quadrant containing the vessel trunk than in the disc quadrant containing the central retinal vessel trunk. Pathogenically, it was suggested that the retinal vessel trunk could act as a stabilizing element against glaucomatous changes in the lamina cribrosa. It could render more difficult a mechanical distortion and backward bowing of the lamina cribrosa around the vessel trunk, for example, in elevated intraocular pressure in glaucoma. This hypothesis is supported by photographs of a W-shaped lamina cribrosa in glaucomatous eyes.
20 As an alternative to this mechanical theory, it was speculated that in the close proximity of the retinal vessel trunk, the vascular supply to the adjacent tissue may be better than in a region more distant to the central retinal vessel trunk. A vitally important participation of branches of the central retinal vessels in the nourishment of the optic nerve fibers in the lamina cribrosa, however, has not yet been demonstrated.
In contrast to the central retinal vessel trunk, cilioretinal arteries do not run through the lamina cribrosa of the optic nerve head. They are thought to originate in the short posterior ciliary arteries in the choroid close to the border of the optic disc, and to pierce through the peripapillary scleral ring of Elschnig separating the choroid from the prelaminar part of the optic nerve head.
6 7 8 9 10 11 12 13 14 If the theory of a stabilizing effect of the central retinal vessel trunk against a deformation of the lamina cribrosa due to an increased translaminar cribrosa pressure gradient holds true, cilioretinal arteries cannot give similar support to the lamina cribrosa. That may be the reason that the present study, as most of the studies cited above, and in contrast to investigations examining the influence of the central retinal vessel trunk, did not demonstrate an influence of the presence and location of cilioretinal arteries on the local development of rim loss or local extension of parapapillary chorioretinal atrophy in chronic open-angle glaucoma, regardless of the blood that the cilioretinal arteries may supply to the prelaminar optic nerve.
Several factors may limit the present study. Simple ophthalmoscopy may underestimate the number of detected cilioretinal arteries compared with fluorescein angiography.
11 It was not the purpose of the present study, however, to establish a prevalence of cilioretinal arteries, but to evaluate a possible relationship between the presence and location of cilioretinal arteries and the local susceptibility for glaucomatous optic nerve damage. Because optic disc photographs, on which the presence of a cilioretinal artery could not clearly be detected or ruled out, were generally excluded from the study, the fact that ophthalmoscopic images instead of fluorescein angiographic images were taken may not have markedly influenced the result of the study. Furthermore, the early filling as the key feature of cilioretinal arteries in fluorescein angiograms can reverse with increasing age, leading to an underestimation of cilioretinal arteries on fluorescein angiograms.
21 The present study was not masked in the sense that the cilioretinal arteries were covered when the morphometry of the optic nerve head was performed. However, because the morphometric evaluation of the optic disc photographs was performed before the possible role that the cilioretinal arteries play in local susceptibility for glaucomatous damage in the optic nerve head was decided, the assessment of the neuroretinal rim and parapapillary atrophy may be considered to have been performed independent of the knowledge of whether there were cilioretinal arteries.
In conclusion, in contrast to the position of the central retinal vessel trunk, the presence and position of cilioretinal arteries may not markedly influence the pattern of neuroretinal rim loss and the extension of parapapillary atrophy in glaucoma. This finding may be of interest in the discussion of the pathogenesis of glaucomatous optic nerve damage. It may also have some importance in the diagnosis of glaucomatous optic nerve head changes, because the location of the central retinal vessel trunk in the lamina cribrosa, and not the presence and location of cilioretinal arteries at the border of the optic disc, may indicate the location in which to look for loss of the neuroretinal rim and progression of parapapillary atrophy in patients with chronic open-angle glaucoma.