Int Ophthalmol DOI 10.1007/s10792-015-0053-y
ORIGINAL PAPER
Optic nerve head characteristics in eyes with papillomacular bundle defects in glaucoma Aparna Rao • Sujoy Mukherjee Debananda Padhy
•
Received: 15 June 2014 / Accepted: 16 February 2015 Ó Springer Science+Business Media Dordrecht 2015
Abstract To evaluate the pattern of retinal nerve fibre layer defects (RNFLD) with regard to involvement of papillomacular bundle (PMB) in glaucoma. This observational study included patients attending glaucoma imaging services at our centre from 2011 to 2012. All images were exported to Image J software for analysis and rescaled to a unified scale for measurement of degree of RNFLD defined by its angular width, pattern of involvement with regard to involvement or sparing of PMB in particular and horizontal and vertical distance of central vessel trunk (CVT) from the disc margin. Association of clinical data with pattern of defects with regard to PMB involvement was analysed. Sixty-two fundus photographs with discernible nerve fibre layer defects on red free images were selected, including 48 normal tension glaucoma, two primary angle closure glaucoma (PACG) and 12 primary open angle glaucoma (POAG) eyes. Discernible PMB involvement was seen in 35 eyes which included 31 defects in inferior quadrant while CVT exit was placed in the quadrant opposite to the quadrant of RNFLD in that eye. The mean vertical distance from the nearest disc margin
Electronic supplementary material The online version of this article (doi:10.1007/s10792-015-0053-y) contains supplementary material, which is available to authorised users. A. Rao (&) S. Mukherjee D. Padhy LV Prasad Eye Institute, Patia, Bhubaneswar, India e-mail: [email protected]; [email protected]
was greater in eyes without PMB involvement, 0.4 ± 0.02 mm, than eyes with PMB defects, 0.3 ± 0.01 mm, p \ 0.001. On multivariate logistic regression, PMB involvement was significantly associated with decreased linear horizontal of the CVT from the disc margin, p = 0.003. Selective involvement of superior and inferior PMB suggests different retinotopic representation within the optic disc. Exit of the CVT towards the disc margin may be a predisposing factor for RNFLD and involvement of the PMB. Keywords Glaucoma Papillomacular bundle Central vessel trunk Retinotopy
Introduction Structural retinal nerve fibre layer defects (RNFLD) precede functional changes on the visual field and are often related to the retinotopic localisation of fibres within the optic disc [1–3]. While several earlier studies have identified the retinotopic map of the nerve fibres subserving different retinal fields, there is still scope for studies on localisation of fibres within the retina or differential blood supply to these fibres at several levels of their course towards the visual cortex [4–6]. It is further unclear as to why certain eyes are predisposed to greater structural damage in the form of RNFLD as opposed to others with more functional visual field damage.
123
Int Ophthalmol
Though the paracentral area is usually spared until the end stages [2], early paracentral scotomas threatening fixation may develop in some eyes during early stage which are at greater risk for impaired quality of vision [6]. Atypical RNFLD involving the papillomacular bundle (PMB) are more likely in eyes with larger axial length, larger optic discs and in normal tension glaucoma (NTG) [6]. While those that develop early PMB defects may be phenotypically different from eyes without them, current literature is sparse regarding the pattern of involvement of RNFLD and PMB defects in glaucoma as also the possible reasons to early involvement or sparing with similar intraocular pressure (IOP). To address this lacuna, we conducted this study to evaluate the differences in optic disc phenotypes in eyes with different patterns of RNFL defects including PMB.
Methods Participants Medical records of patients attending the glaucoma imaging centre at our tertiary centre from 2011 to 2012 were reviewed retrospectively. Data that were retrieved included the best corrected visual acuity, spherical equivalent, gonioscopy, slit lamp biomicroscopy, Goldmann applanation IOP at the time of imaging, baseline untreated IOP, refractive error, dilated non-stereoscopic optic disc photographs, Humphrey visual fields (Carl Zeiss Meditec, 24-2, 01-2 and macular programme). All procedures adhered to the tenets of the Declaration of Helsinki and the study was approved by the institutional review board.
All available photographs were assessed by a single examiner (APR) to identify presence (or absence of) RNFL defects on red free photographs. Inclusion criteria included fundus images with identifiable RNFD (seen as dark black wedge-shaped or diffuse defects reaching the disc margin) whose width at one disc diameter away from the disc was greater than a major retinal vessel. Those with hazy media, refractive errors C-6 dioptre (D), associated posterior segment lesions like diabetic retinopathy, age related macular degeneration, vascular occlusions, were excluded. Image analysis To avoid magnification errors, only images showing disc, macula and posterior pole were selected for analysis using Image J software while magnified images capturing the disc alone were excluded for Image J analysis. All colour images were exported to Image J software (http://rsbweb.nih.gov/ij/; www.nih.gov, National Institutes of Health, Bethesda, MD).for analysis and rescaled to a unified scale, and measurement of degree of RNFLD measurement was done as explained in previous study with some modifications [3, 4]. Image analysis was done using the methods described elsewhere. Briefly, distances between two points of interest were measured in millimetre scale after rescaling in each of the images. Figure 1 shows the methods of measurement of different parameters used in this study. All included photographs were analysed for the following by two independent examiners (SM and APR) blinded to clinical details and an average of three measurements were used for analysis. Each image was then analysed using the following protocol.
Fundus photography protocol at imaging centre Assessment of pattern of RNFL involvement All images were acquired by one examiner (SM) blinded to the diagnosis or details of the patient using the Visupac version 4.4.4 (FF 450 plus IR Carl Zeiss Ltd USA). Pupils were dilated by one drop of fixed combination drops containing 1 % tropicamide and 2.5 % phenylephrine. Images with motion artefacts, reduced contrast or low quality making it difficult to distinguish disc margins were excluded from the study. Photographs were taken in colour and red free frames.
123
A line from the centre of the optic disc to the foveal centre (the reference line), was drawn on the images. A RNFL defect was considered to involve the PMB when the proximal border of the nearest defect was located within 20° of the reference line while those in the superior or inferior quadrants outside this area were termed arcuate defects. After identifying the RNFLD and the reference line, the number of RNFL defects along with the
Int Ophthalmol
Visual field testing Only reliable fields were included which included fixation losses, false-positive errors and false-negative errors were less than 20 %. A glaucomatous field defect was defined by a glaucoma hemifield test outside normal limits, the presence of at least three non-edge test points in the same hemifield on the pattern deviation probability plot at p less than 5 % with at least one point at p less than 1 % and excluding points directly above or below the blind spot. The visual field indices including mean deviation (MD), pattern standard deviation (PSD) and visual field index (VFI) were noted. Fig. 1 Method of measurement of linear distance of central vessel trunk from nasal disc border or nearest (superior or inferior) disc margin and angular width (inset) after drawing reference circle of 3.4 mm using Image J software after rescaling to mm scale
location of the defects (superotemporal, inferotemporal or PMB superior or inferior) was noted. The pattern of PMB involvement (superior or inferior quadrant or sparing of either area) was noted. The two borders of each RNFL defect were defined by drawing lines from the centre of the optic disc to each of the points at which the borders of the RNFL defect intersected the circle. The angular width of the RNFL defect was measured in degrees measured by the angle measurement tool in the image J software as the angle formed by the two border lines of the RNFL defect. For eyes with multiple defects, the total width of the defects was calculated by summing widths of all the RNFL defects of that eye. The photograph of the disc was divided into four quadrants by two lines passing through the centre of the disc into superotemporal, superonasal, inferotemporal and inferonasal quadrants to evaluate the location of RNFLD and exit of central vessels from the optic disc.
Statistical analysis An average of three readings per image was used for analysis. Agreement between two observers for measurement of angular width of RNFLD and linear horizontal distance of CVT was assessed using intraclass coefficient. Descriptive data are described as means and deviation for normally distributed variables and median and range for non-parametric variables and normality was assessed using Shapiro–Wilk test. Association of pattern of PMB involvement with independent clinical variables including age, IOP at time of imaging and baseline untreated IOP, linear horizontal and vertical distance of CVT, location of RNFLD and visual field indices were analysed using univariate and multivariate logistic regression. Additional analysis of association of extent of RNFL defect with the mentioned clinical variables was evaluated by backward step-wise approach with significance set at 5 %. We also analysed differences between eyes with and without PMB involvement by student t-test for unpaired observation samples for the parametric parameters and Mann–Whitney U test for non-parametric data with an alpha error set at \0.05.
Position of central vessel trunk (CVT)
Results
The exit position of the central vessel trunk (CVT) on the lamina cribrosa surface was noted in relation to the optic disc quadrant. The distance of the CVT was measured horizontally from the nasal disc border and vertically from the nearest (superior or inferior) disc margin by calliper function in the image J software.
We included 62 photographs with evident RNFL defects while 13 with poor quality of images, three with refractive error [6 dioptres and two with associated diabetic retinopathy were excluded. Table 1 gives the demographic and clinical characteristics of all participants included in the study. This
123
Int Ophthalmol Table 1 Subject demographics and clinical variables of participants Age (years)
54 ± 10.3
IOP at time of imaging (mmHg)
18 ± 3.6
MD (dB)
-14 ± 7.9
PSD (dB)
9 ± 4.2
VFI (%)
63 ± 24.1
Diagnosis
NTG:POAG:PACG = 42:12:2
No of RNFL defects per eye
1.1 ± 0.5
VFI visual field index, MD mean deviation, PSD pattern standard deviation, IOP intraocular pressure, NTG normal tension glaucoma, POAG primary open angle glaucoma, PACG primary angle closure glaucoma
included 48 NTG, two primary angle closure glaucoma (PACG) and 12 primary open angle glaucoma (POAG) eyes (Table S1) with a mean spherical equivalent of -2 (±1.3 dioptres). The average MD and VFI was -14 ± 7.9 dB (-31 to -2.1 dB) and 63 ± 12.4 % in all patients included for the study. Intraclass correlation coefficient between the two observers [95 % confidence interval (CI)] for the angular width of RNFL and horizontal distance of CVT was 0.94 (0.81–0.98) and 0.91 (0.86–0.95), respectively. Retinal nerve fibre measurements There was an average of 1.19 ± 0.5 RNFL defects per eye with eight having two RNFL defects each in
superotemporal and inferotemporal quadrant and one eye having superior, inferior arcuate and inferior PMB involvement. PMB involvement was seen in 35 eyes which included 31 defects in inferior quadrant. Table 2 gives the difference in RNFL variables among eyes with and without PMB involvement. The mean angular width of the RNFL defects was 64° ± 32.4° (range from 17° to 121°) with no difference in eyes with or without PMB involvement. The location of the RNFLD was seen in the inferotemporal quadrant (ITQ) in 36 eyes, superotemporal quadrant (STQ) in 13 eyes, isolated inferior PMB in two eyes, and combined defects in nine eyes. The exit of CVT was situated in the superonasal in 26 eyes and inferonasal quadrant in 17 eyes which included 21 eyes and 12 eyes with PMB involvement, respectively. The mean horizontal distance of CVT from nasal disc border was 0.3 ± 0.1 and 0.5 ± 0.03 mm in eyes with and without PMB involvement, respectively, p \ 0.001, Table 3. None of the eyes with horizontal CVT distance [0.6 mm had PMB involvement while 24 eyes with PMB involvement had CVT placed within \0.4 mm from the disc, Fig. 2. The mean vertical distance from the nearest disc margin was greater in eyes without PMB involvement, 0.4 ± 0.02 mm than eyes with PMB defects, 0.3 ± 0.01 mm, p \ 0.001. The CVT was placed vertically in 26 eyes opposite to the quadrant of the RNLFD (namely inferotemporal quadrant in 36 eyes). There was no statistical difference between other clinical variables like age or treated IOP between the
Table 2 Pattern of RNFL defects in those with and without PMB involvement Variables
Without PMB involvement
With PMB involvement
p
Angular width of RNFLD
56 ± 30.9 (43.9–68.9)
71 ± 32.4 (59.8–83.2)
0.07
No of RNFLD
1 ± 0.5 (1.02–1.4)
1 ± 0.5 (0.9–1.3)
0.7^
MD VFI
-10 ± 5.3 (-12.9 to -8.3) 73 ± 17 (65.6–80.4)
-16 ± 8.5 (-19.8 to -13.9) 56 ± 26.2 (47.8–65.8)
0.002
Generalised
0
2
Inferotemporal quadrant
14
22
Combined defects
4
5
Location of RNFLD
Superotemporal quadrant
9
4
Isolated PMB involvement
0
2
RNFLD retinal nerve fibre layer defect, MD mean deviation, VFI visual field index ^
Chi-square test
123
Int Ophthalmol Table 3 Central vessel trunk distance-relation to optic disc in patients with and without PMB involvement Variables
Without PMB involvement
With PMB involvement
p
Horizontal distance (mm)*
0.56 ± 0.03 (0.53–0.56)
0.3 ± 0.1 (0.29–0.36)
\0.00
Vertical distance (mm)
0.4 ± 0.02 (0.3–0.41)
0.3 ± 0.01 (0.1–0.2)
\0.001
CVT \0.4 CVT [0.6
0 3
24 0
\0.001^ 0.02
CVT central vessel trunk ^
Chi-square test
* Linear distance from nasal disc border
Fig. 2 Number of eyes with and without PMB involvement with different linear horizontal distance of central vessel trunk from nasal disc margin
two groups. The visual field indices including MD and VFI differed significantly between the two groups, p = 0.002 and p = 0.01, respectively. On multivariate logistic regression, PMB involvement was significantly associated with decreased horizontal and vertical distance of the CVT from the disc margin, p = 0.003. The angular width of RNFLD involvement in all eyes (n = 62) was significantly associated on multivariate linear regression with horizontal distance from
disc margin, p = 0.002 and PSD, p = 0.04. There was no difference in the age, IOP or other visual field indices in those with \ or [48° (50th percentile) of RNFLD. Figures 3 and 4 give representative examples of those with PMB involvement or sparing and angular width of RNFLD with corresponding visual field involvement, both depicting central and extensive defects with decreased distance of CVT from the horizontal and vertical disc margins.
123
Int Ophthalmol
Fig. 3 a Color fundus photographs of the right eye showing inferior PMB involvement (inset and black solid arrows) with exit of CVT in the superonasal quadrant. b Color fundus image
showing 3 RNFL defects involving the superior and inferior arcuate area as well as inferior papillomacular bundle (solid black arrows)
Discussion
About 90 % of the retinal ganglion cells (RGC’s) in this location are small midget cells projecting to the parvocellular layer [4, 8, 9]. While these fibres are responsible for central vision and higher contrast, such anatomic differences alone fail to explain why superior PMB fibres should be spared preferentially, as seen in this study. Preferential involvement of fibres in the inferior quadrant with sparing of the superior PMB fibres or corresponding inferior central vision possibly may be attributed to possible anatomical differences in vessels at the optic nerve head or the arrangement of PMB fibres at the optic nerve head which can be studied in future studies. Studies exploring binocular visual field with involvement of the central field in an eye may give insight into the actual impact of PMB involvement early in the disease process on the quality of life. Atypical RNFLD involving the PMB are more likely in NTG, eyes with larger optic disc and in high myopes [3, 10, 11]. The retinal circulation perfuses the inner retina (bounded by the nerve fibre layer anteriorly and the inner nuclear layer posteriorly) via the central retinal artery (CRA). The source of the CRA is the ophthalmic artery, which is the major conduit to both the retina and choroid [12]. Cilioretinal vessels radiate temporally from the optic nerve through the PMB, providing an accessory blood supply to the macular inner retina from the ciliary circulation [13]. At the optic disc, the axons are closely packed along with the central vessels and the pattern of distribution of branches to each area from the central vessels is not very clearly delineated. Proximity to the vessel trunk
In this study, PMB involvement was seen in 35 eyes, 31 of which were seen preferentially in the inferior quadrant. The exit of the CVT was seen correspondingly in superonasal quadrant in 21 of 35 eyes, with decreased horizontal and vertical distance of CVT from the disc margin strongly correlating with PMB involvement. Fibres that originate within 0.5 mm of the fovea temporal to the disc are called papillomacular. While topographic differentiation of PMB fibres relative to fibres from other retinal areas is described, this bundle is often referred to as a single bundle at the optic disc with little differentiation of fibres in inferior, superior and central fibres [2–5]. In this study, PMB involvement was seen more commonly in the inferior quadrant (thereby affecting the superior visual field) with sparing of the corresponding inferior visual field in all these eyes. This suggests difference in risk of involvement of the superior and inferior fibres constituting the PMB at the level of the retina. It has been demonstrated that approximately 130 million RGCs connect to about 1.2 million nerve fibres [5]. Visual function of the subcortex is resultant of the precise orderly organisation of ascending axons of retinal ganglion cells as they course towards the optic disc and further to the cortex [4–8]. The structural differences in the PMB area as compared to the rest of the retina has been well delineated [8, 9].One study revealed small diameter bundle of fibres aligned in the inner most row adjacent to the central vessels [9].
123
Int Ophthalmol
Fig. 4 a Color and red free fundus photographs (inset) showing diffuse RNFL loss sparing of the superior PMB layer (black solid arrows), b Visual fields macular programme showing corresponding sparing of the inferior field, c Color and red free (inset) fundus photographs of the right eye showing inferior
diffuse loss involving the inferior PMB (solid black arrows) and superior RNFL defect and sparing of the superior PMB, d Visual fields of the patient showing corresponding defect involving superior half involving fixation and inferior arcuate area on 24-2 and sparing of the inferior central field on 10-2
offers an advantage to these axons in terms of mechanical support as well as possibly more vascular twigs or straighter course of vessels providing nourishment [14, 15]. It is possible that the superior and inferior fibres of the PMB have an independent vascular supply at the retinal level possibly from different twigs from the cilioretinal artery or direct branches of the central artery itself. This would need histological studies in the future to investigate this possibility. The CVT exit was placed in the quadrant opposite to that of the RNFLD in 26 eyes in this study. Previous studies have provided proof that the distance between
the retinal vessels in the lamina cribrosa and the neuroretinal rim may be associated with the pattern of glaucomatous rim loss [14, 15]. Displacement of the CVT and nerve fibres due to compression within the lamina cribrosa in advanced or progressive glaucomatous cupping would lead to the CVT exit moving closer to the disc margins and therefore away from the central macular fibres within the optic disc [13]. The CVT is thought to be protective with regards to providing mechanical support as well vascular supply to areas in close proximity [14, 15]. Our findings support this hypothesis though the relation between the vascular twigs supplying the different fibres of the PMB versus
123
Int Ophthalmol
their pattern of involvement in glaucoma remains to be explored. Park et al. observed that eyes with involvement of superior and inferior parafoveal area may be thought to be phenotypically different from those that do not have involvement even till late stages of glaucoma [16]. This phenotypic difference may be attributed to differences in the topographical location of the fibres within PMB itself or differential involvement of vascular twigs supplying these areas [17]. High myopes have abnormally shaped tilted discs which may alter the structural organisation of the fibres within the disc [3, 6, 7, 18–20]. Kim et al. recently reported that progressive tilting of the optic disc with nasal shift of temporal optic disc margin with corresponding enlargement of peripapillary atrophy occurred in children with myopic shift [20]. Such displacements may also shift the CVT trunk farther away from the macular fibres. We excluded high myopes and eyes with tilted discs. Yet, we found factors independent of axial length or refractive error responsible for early involvement of the central visual field in some eyes. An inferior paracentral scotoma can significantly affect a patient’s quality of life and it is important to detect these abnormalities. Since structural alterations like RNFLD precede functional visual field changes, detection of superior PMB defects could be useful to predict development of inferotemporal paracentral scotomas. In summary, placement of the CVT trunk away from the centrally located PMB fibres may be a phenotypic risk factor for early involvement of these fibres. This would help us predict the future outcomes of eyes at risk in terms of visual function and quality of life. Conflict of interest
None
References 1. Radius RL, Anderson DR (1979) The course of axons through the retina and optic nerve head. Arch Ophthalmol 97:1154–1158 2. Ogden TE (1984) Nerve fiber layer of the primate retina: morphometric analysis. Invest Ophthalmol Vis Sci 25:19–29 3. Kimura Y, Hangai M, Morooka S, Takayama K, Nakano N, Nukada M, Ikeda HO, Akagi T, Yoshimura N (2012) Retinal nerve fiber layer defects in highly myopic eyes with early glaucoma. Invest Ophthalmol Vis Sci 53:6472–6478
123
4. Rao A, Mukherjee S (2014) Anatomical attributes of the optic nerve head in eyes with parafoveal scotoma in normal tension glaucoma. PLoS One. doi:10.1371/journal.pone. 0090554 5. Jonas JB, Mueller-Bergh JA, Schloetzer-Schrehardt UM, Naumann GOH (1990) Histomorphometry of the human optic nerve. Invest Ophthalmol Vis Sci 31:736–744 6. Quigley HA, Addicks EM (1981) Regional differences in the structure of the lamina cribrosa and their relation to glaucomatous optic nerve damage. Arch Ophthalmol 99:137–143 7. Jonas JB, Schmidt AM, Mueller-Bergh JA, SchloetzerSchrehardt UM, Naumann GOH (1992) Human optic nerve fiber count and optic disc size. Invest Ophthalmol Vis Sci 33:2012–2018 8. FitzGibbon T, Taylor SF (1996) Retinotopy of the human retinal nerve fibre layer and optic nerve head. J Comp Neurol 375:238–251 9. Pavlidis M, Stupp T, Hummeke M, Thanos S (2006) Morphometric examination of human and monkey retinal ganglion cells within the papillomacular area. Retina 26:445–453 10. Caprioli J, Sears M, Miller JM (1987) Patterns of early visual field loss in open-angle glaucoma. Am J Ophthalmol 103:512–517 11. Araie M, Sekine M, Suzuki Y, Koseki N (1994) Factors contributing to the progression of visual field damage in eyes with normal-tension glaucoma. Ophthalmology 101:1440–1444 12. Lieberman MF, Maumenee EA, Green RW (1976) Histologic studies of the vasculature of the anterior optic nerve. Am J Ophthalmol 82:405–423 13. Lee SS, Schwartz B (1992) Role of the temporal cilioretinal artery in retaining central visual field in open-angle glaucoma. Ophthalmology 99:696–699 14. Huang H, Jonas JB, Dai Y, Hong J, Wang M, Chen J, Wu J, Sun X (2013) Position of the central retinal vessel trunk and pattern of remaining visual field in advanced glaucoma. Br J Ophthalmol 97:96–100 15. Jonas JB, Ferna´ndez MC (1994) Shape of the neuroretinal rim and position of the central retinal vessels in glaucoma. Br J Ophthalmol 78:99–102 16. Park SC, De Moraes CG, Teng CC, Tello C, Liebmann JM, Ritch R (2011) Initial parafoveal versus peripheral scotomas in glaucoma: risk factors and visual field characteristics. Ophthalmology 118:1782–1789 17. Pavlidis M, Stupp T, Naskar R, Cengiz C, Thanos S (2003) Retinal ganglion cells resistant to advanced glaucoma: a postmortem study of human retinas with the carbocyanine dye DiI. Invest Ophthalmol Vis Sci 44:5196–5205 18. Mitchell P, Hourihan F, Sandbach J, Wang JJ (1999) The relationship between glaucoma and myopia: the blue mountains eye study. Ophthalmology 106:2010–2015 19. Yang H, Downs JC, Girkin C, Sakata L, Bellezza A, Thompson H, Burgoyne CF (2007) 3-D histomorphometry of the normal and early glaucomatous monkey optic nerve head: lamina cribrosa and peripapillary scleral position and thickness. Invest Ophthalmol Vis Sci 48:4597–4607 20. Kim TW, Kim M, Weinreb RN, Woo SJ, Park KH, Hwang JM (2012) Optic disc change with incipient myopia of childhood. Ophthalmology 119:21–26
ORIGINAL PAPER
Optic nerve head characteristics in eyes with papillomacular bundle defects in glaucoma Aparna Rao • Sujoy Mukherjee Debananda Padhy
•
Received: 15 June 2014 / Accepted: 16 February 2015 Ó Springer Science+Business Media Dordrecht 2015
Abstract To evaluate the pattern of retinal nerve fibre layer defects (RNFLD) with regard to involvement of papillomacular bundle (PMB) in glaucoma. This observational study included patients attending glaucoma imaging services at our centre from 2011 to 2012. All images were exported to Image J software for analysis and rescaled to a unified scale for measurement of degree of RNFLD defined by its angular width, pattern of involvement with regard to involvement or sparing of PMB in particular and horizontal and vertical distance of central vessel trunk (CVT) from the disc margin. Association of clinical data with pattern of defects with regard to PMB involvement was analysed. Sixty-two fundus photographs with discernible nerve fibre layer defects on red free images were selected, including 48 normal tension glaucoma, two primary angle closure glaucoma (PACG) and 12 primary open angle glaucoma (POAG) eyes. Discernible PMB involvement was seen in 35 eyes which included 31 defects in inferior quadrant while CVT exit was placed in the quadrant opposite to the quadrant of RNFLD in that eye. The mean vertical distance from the nearest disc margin
Electronic supplementary material The online version of this article (doi:10.1007/s10792-015-0053-y) contains supplementary material, which is available to authorised users. A. Rao (&) S. Mukherjee D. Padhy LV Prasad Eye Institute, Patia, Bhubaneswar, India e-mail: [email protected]; [email protected]
was greater in eyes without PMB involvement, 0.4 ± 0.02 mm, than eyes with PMB defects, 0.3 ± 0.01 mm, p \ 0.001. On multivariate logistic regression, PMB involvement was significantly associated with decreased linear horizontal of the CVT from the disc margin, p = 0.003. Selective involvement of superior and inferior PMB suggests different retinotopic representation within the optic disc. Exit of the CVT towards the disc margin may be a predisposing factor for RNFLD and involvement of the PMB. Keywords Glaucoma Papillomacular bundle Central vessel trunk Retinotopy
Introduction Structural retinal nerve fibre layer defects (RNFLD) precede functional changes on the visual field and are often related to the retinotopic localisation of fibres within the optic disc [1–3]. While several earlier studies have identified the retinotopic map of the nerve fibres subserving different retinal fields, there is still scope for studies on localisation of fibres within the retina or differential blood supply to these fibres at several levels of their course towards the visual cortex [4–6]. It is further unclear as to why certain eyes are predisposed to greater structural damage in the form of RNFLD as opposed to others with more functional visual field damage.
123
Int Ophthalmol
Though the paracentral area is usually spared until the end stages [2], early paracentral scotomas threatening fixation may develop in some eyes during early stage which are at greater risk for impaired quality of vision [6]. Atypical RNFLD involving the papillomacular bundle (PMB) are more likely in eyes with larger axial length, larger optic discs and in normal tension glaucoma (NTG) [6]. While those that develop early PMB defects may be phenotypically different from eyes without them, current literature is sparse regarding the pattern of involvement of RNFLD and PMB defects in glaucoma as also the possible reasons to early involvement or sparing with similar intraocular pressure (IOP). To address this lacuna, we conducted this study to evaluate the differences in optic disc phenotypes in eyes with different patterns of RNFL defects including PMB.
Methods Participants Medical records of patients attending the glaucoma imaging centre at our tertiary centre from 2011 to 2012 were reviewed retrospectively. Data that were retrieved included the best corrected visual acuity, spherical equivalent, gonioscopy, slit lamp biomicroscopy, Goldmann applanation IOP at the time of imaging, baseline untreated IOP, refractive error, dilated non-stereoscopic optic disc photographs, Humphrey visual fields (Carl Zeiss Meditec, 24-2, 01-2 and macular programme). All procedures adhered to the tenets of the Declaration of Helsinki and the study was approved by the institutional review board.
All available photographs were assessed by a single examiner (APR) to identify presence (or absence of) RNFL defects on red free photographs. Inclusion criteria included fundus images with identifiable RNFD (seen as dark black wedge-shaped or diffuse defects reaching the disc margin) whose width at one disc diameter away from the disc was greater than a major retinal vessel. Those with hazy media, refractive errors C-6 dioptre (D), associated posterior segment lesions like diabetic retinopathy, age related macular degeneration, vascular occlusions, were excluded. Image analysis To avoid magnification errors, only images showing disc, macula and posterior pole were selected for analysis using Image J software while magnified images capturing the disc alone were excluded for Image J analysis. All colour images were exported to Image J software (http://rsbweb.nih.gov/ij/; www.nih.gov, National Institutes of Health, Bethesda, MD).for analysis and rescaled to a unified scale, and measurement of degree of RNFLD measurement was done as explained in previous study with some modifications [3, 4]. Image analysis was done using the methods described elsewhere. Briefly, distances between two points of interest were measured in millimetre scale after rescaling in each of the images. Figure 1 shows the methods of measurement of different parameters used in this study. All included photographs were analysed for the following by two independent examiners (SM and APR) blinded to clinical details and an average of three measurements were used for analysis. Each image was then analysed using the following protocol.
Fundus photography protocol at imaging centre Assessment of pattern of RNFL involvement All images were acquired by one examiner (SM) blinded to the diagnosis or details of the patient using the Visupac version 4.4.4 (FF 450 plus IR Carl Zeiss Ltd USA). Pupils were dilated by one drop of fixed combination drops containing 1 % tropicamide and 2.5 % phenylephrine. Images with motion artefacts, reduced contrast or low quality making it difficult to distinguish disc margins were excluded from the study. Photographs were taken in colour and red free frames.
123
A line from the centre of the optic disc to the foveal centre (the reference line), was drawn on the images. A RNFL defect was considered to involve the PMB when the proximal border of the nearest defect was located within 20° of the reference line while those in the superior or inferior quadrants outside this area were termed arcuate defects. After identifying the RNFLD and the reference line, the number of RNFL defects along with the
Int Ophthalmol
Visual field testing Only reliable fields were included which included fixation losses, false-positive errors and false-negative errors were less than 20 %. A glaucomatous field defect was defined by a glaucoma hemifield test outside normal limits, the presence of at least three non-edge test points in the same hemifield on the pattern deviation probability plot at p less than 5 % with at least one point at p less than 1 % and excluding points directly above or below the blind spot. The visual field indices including mean deviation (MD), pattern standard deviation (PSD) and visual field index (VFI) were noted. Fig. 1 Method of measurement of linear distance of central vessel trunk from nasal disc border or nearest (superior or inferior) disc margin and angular width (inset) after drawing reference circle of 3.4 mm using Image J software after rescaling to mm scale
location of the defects (superotemporal, inferotemporal or PMB superior or inferior) was noted. The pattern of PMB involvement (superior or inferior quadrant or sparing of either area) was noted. The two borders of each RNFL defect were defined by drawing lines from the centre of the optic disc to each of the points at which the borders of the RNFL defect intersected the circle. The angular width of the RNFL defect was measured in degrees measured by the angle measurement tool in the image J software as the angle formed by the two border lines of the RNFL defect. For eyes with multiple defects, the total width of the defects was calculated by summing widths of all the RNFL defects of that eye. The photograph of the disc was divided into four quadrants by two lines passing through the centre of the disc into superotemporal, superonasal, inferotemporal and inferonasal quadrants to evaluate the location of RNFLD and exit of central vessels from the optic disc.
Statistical analysis An average of three readings per image was used for analysis. Agreement between two observers for measurement of angular width of RNFLD and linear horizontal distance of CVT was assessed using intraclass coefficient. Descriptive data are described as means and deviation for normally distributed variables and median and range for non-parametric variables and normality was assessed using Shapiro–Wilk test. Association of pattern of PMB involvement with independent clinical variables including age, IOP at time of imaging and baseline untreated IOP, linear horizontal and vertical distance of CVT, location of RNFLD and visual field indices were analysed using univariate and multivariate logistic regression. Additional analysis of association of extent of RNFL defect with the mentioned clinical variables was evaluated by backward step-wise approach with significance set at 5 %. We also analysed differences between eyes with and without PMB involvement by student t-test for unpaired observation samples for the parametric parameters and Mann–Whitney U test for non-parametric data with an alpha error set at \0.05.
Position of central vessel trunk (CVT)
Results
The exit position of the central vessel trunk (CVT) on the lamina cribrosa surface was noted in relation to the optic disc quadrant. The distance of the CVT was measured horizontally from the nasal disc border and vertically from the nearest (superior or inferior) disc margin by calliper function in the image J software.
We included 62 photographs with evident RNFL defects while 13 with poor quality of images, three with refractive error [6 dioptres and two with associated diabetic retinopathy were excluded. Table 1 gives the demographic and clinical characteristics of all participants included in the study. This
123
Int Ophthalmol Table 1 Subject demographics and clinical variables of participants Age (years)
54 ± 10.3
IOP at time of imaging (mmHg)
18 ± 3.6
MD (dB)
-14 ± 7.9
PSD (dB)
9 ± 4.2
VFI (%)
63 ± 24.1
Diagnosis
NTG:POAG:PACG = 42:12:2
No of RNFL defects per eye
1.1 ± 0.5
VFI visual field index, MD mean deviation, PSD pattern standard deviation, IOP intraocular pressure, NTG normal tension glaucoma, POAG primary open angle glaucoma, PACG primary angle closure glaucoma
included 48 NTG, two primary angle closure glaucoma (PACG) and 12 primary open angle glaucoma (POAG) eyes (Table S1) with a mean spherical equivalent of -2 (±1.3 dioptres). The average MD and VFI was -14 ± 7.9 dB (-31 to -2.1 dB) and 63 ± 12.4 % in all patients included for the study. Intraclass correlation coefficient between the two observers [95 % confidence interval (CI)] for the angular width of RNFL and horizontal distance of CVT was 0.94 (0.81–0.98) and 0.91 (0.86–0.95), respectively. Retinal nerve fibre measurements There was an average of 1.19 ± 0.5 RNFL defects per eye with eight having two RNFL defects each in
superotemporal and inferotemporal quadrant and one eye having superior, inferior arcuate and inferior PMB involvement. PMB involvement was seen in 35 eyes which included 31 defects in inferior quadrant. Table 2 gives the difference in RNFL variables among eyes with and without PMB involvement. The mean angular width of the RNFL defects was 64° ± 32.4° (range from 17° to 121°) with no difference in eyes with or without PMB involvement. The location of the RNFLD was seen in the inferotemporal quadrant (ITQ) in 36 eyes, superotemporal quadrant (STQ) in 13 eyes, isolated inferior PMB in two eyes, and combined defects in nine eyes. The exit of CVT was situated in the superonasal in 26 eyes and inferonasal quadrant in 17 eyes which included 21 eyes and 12 eyes with PMB involvement, respectively. The mean horizontal distance of CVT from nasal disc border was 0.3 ± 0.1 and 0.5 ± 0.03 mm in eyes with and without PMB involvement, respectively, p \ 0.001, Table 3. None of the eyes with horizontal CVT distance [0.6 mm had PMB involvement while 24 eyes with PMB involvement had CVT placed within \0.4 mm from the disc, Fig. 2. The mean vertical distance from the nearest disc margin was greater in eyes without PMB involvement, 0.4 ± 0.02 mm than eyes with PMB defects, 0.3 ± 0.01 mm, p \ 0.001. The CVT was placed vertically in 26 eyes opposite to the quadrant of the RNLFD (namely inferotemporal quadrant in 36 eyes). There was no statistical difference between other clinical variables like age or treated IOP between the
Table 2 Pattern of RNFL defects in those with and without PMB involvement Variables
Without PMB involvement
With PMB involvement
p
Angular width of RNFLD
56 ± 30.9 (43.9–68.9)
71 ± 32.4 (59.8–83.2)
0.07
No of RNFLD
1 ± 0.5 (1.02–1.4)
1 ± 0.5 (0.9–1.3)
0.7^
MD VFI
-10 ± 5.3 (-12.9 to -8.3) 73 ± 17 (65.6–80.4)
-16 ± 8.5 (-19.8 to -13.9) 56 ± 26.2 (47.8–65.8)
0.002
Generalised
0
2
Inferotemporal quadrant
14
22
Combined defects
4
5
Location of RNFLD
Superotemporal quadrant
9
4
Isolated PMB involvement
0
2
RNFLD retinal nerve fibre layer defect, MD mean deviation, VFI visual field index ^
Chi-square test
123
Int Ophthalmol Table 3 Central vessel trunk distance-relation to optic disc in patients with and without PMB involvement Variables
Without PMB involvement
With PMB involvement
p
Horizontal distance (mm)*
0.56 ± 0.03 (0.53–0.56)
0.3 ± 0.1 (0.29–0.36)
\0.00
Vertical distance (mm)
0.4 ± 0.02 (0.3–0.41)
0.3 ± 0.01 (0.1–0.2)
\0.001
CVT \0.4 CVT [0.6
0 3
24 0
\0.001^ 0.02
CVT central vessel trunk ^
Chi-square test
* Linear distance from nasal disc border
Fig. 2 Number of eyes with and without PMB involvement with different linear horizontal distance of central vessel trunk from nasal disc margin
two groups. The visual field indices including MD and VFI differed significantly between the two groups, p = 0.002 and p = 0.01, respectively. On multivariate logistic regression, PMB involvement was significantly associated with decreased horizontal and vertical distance of the CVT from the disc margin, p = 0.003. The angular width of RNFLD involvement in all eyes (n = 62) was significantly associated on multivariate linear regression with horizontal distance from
disc margin, p = 0.002 and PSD, p = 0.04. There was no difference in the age, IOP or other visual field indices in those with \ or [48° (50th percentile) of RNFLD. Figures 3 and 4 give representative examples of those with PMB involvement or sparing and angular width of RNFLD with corresponding visual field involvement, both depicting central and extensive defects with decreased distance of CVT from the horizontal and vertical disc margins.
123
Int Ophthalmol
Fig. 3 a Color fundus photographs of the right eye showing inferior PMB involvement (inset and black solid arrows) with exit of CVT in the superonasal quadrant. b Color fundus image
showing 3 RNFL defects involving the superior and inferior arcuate area as well as inferior papillomacular bundle (solid black arrows)
Discussion
About 90 % of the retinal ganglion cells (RGC’s) in this location are small midget cells projecting to the parvocellular layer [4, 8, 9]. While these fibres are responsible for central vision and higher contrast, such anatomic differences alone fail to explain why superior PMB fibres should be spared preferentially, as seen in this study. Preferential involvement of fibres in the inferior quadrant with sparing of the superior PMB fibres or corresponding inferior central vision possibly may be attributed to possible anatomical differences in vessels at the optic nerve head or the arrangement of PMB fibres at the optic nerve head which can be studied in future studies. Studies exploring binocular visual field with involvement of the central field in an eye may give insight into the actual impact of PMB involvement early in the disease process on the quality of life. Atypical RNFLD involving the PMB are more likely in NTG, eyes with larger optic disc and in high myopes [3, 10, 11]. The retinal circulation perfuses the inner retina (bounded by the nerve fibre layer anteriorly and the inner nuclear layer posteriorly) via the central retinal artery (CRA). The source of the CRA is the ophthalmic artery, which is the major conduit to both the retina and choroid [12]. Cilioretinal vessels radiate temporally from the optic nerve through the PMB, providing an accessory blood supply to the macular inner retina from the ciliary circulation [13]. At the optic disc, the axons are closely packed along with the central vessels and the pattern of distribution of branches to each area from the central vessels is not very clearly delineated. Proximity to the vessel trunk
In this study, PMB involvement was seen in 35 eyes, 31 of which were seen preferentially in the inferior quadrant. The exit of the CVT was seen correspondingly in superonasal quadrant in 21 of 35 eyes, with decreased horizontal and vertical distance of CVT from the disc margin strongly correlating with PMB involvement. Fibres that originate within 0.5 mm of the fovea temporal to the disc are called papillomacular. While topographic differentiation of PMB fibres relative to fibres from other retinal areas is described, this bundle is often referred to as a single bundle at the optic disc with little differentiation of fibres in inferior, superior and central fibres [2–5]. In this study, PMB involvement was seen more commonly in the inferior quadrant (thereby affecting the superior visual field) with sparing of the corresponding inferior visual field in all these eyes. This suggests difference in risk of involvement of the superior and inferior fibres constituting the PMB at the level of the retina. It has been demonstrated that approximately 130 million RGCs connect to about 1.2 million nerve fibres [5]. Visual function of the subcortex is resultant of the precise orderly organisation of ascending axons of retinal ganglion cells as they course towards the optic disc and further to the cortex [4–8]. The structural differences in the PMB area as compared to the rest of the retina has been well delineated [8, 9].One study revealed small diameter bundle of fibres aligned in the inner most row adjacent to the central vessels [9].
123
Int Ophthalmol
Fig. 4 a Color and red free fundus photographs (inset) showing diffuse RNFL loss sparing of the superior PMB layer (black solid arrows), b Visual fields macular programme showing corresponding sparing of the inferior field, c Color and red free (inset) fundus photographs of the right eye showing inferior
diffuse loss involving the inferior PMB (solid black arrows) and superior RNFL defect and sparing of the superior PMB, d Visual fields of the patient showing corresponding defect involving superior half involving fixation and inferior arcuate area on 24-2 and sparing of the inferior central field on 10-2
offers an advantage to these axons in terms of mechanical support as well as possibly more vascular twigs or straighter course of vessels providing nourishment [14, 15]. It is possible that the superior and inferior fibres of the PMB have an independent vascular supply at the retinal level possibly from different twigs from the cilioretinal artery or direct branches of the central artery itself. This would need histological studies in the future to investigate this possibility. The CVT exit was placed in the quadrant opposite to that of the RNFLD in 26 eyes in this study. Previous studies have provided proof that the distance between
the retinal vessels in the lamina cribrosa and the neuroretinal rim may be associated with the pattern of glaucomatous rim loss [14, 15]. Displacement of the CVT and nerve fibres due to compression within the lamina cribrosa in advanced or progressive glaucomatous cupping would lead to the CVT exit moving closer to the disc margins and therefore away from the central macular fibres within the optic disc [13]. The CVT is thought to be protective with regards to providing mechanical support as well vascular supply to areas in close proximity [14, 15]. Our findings support this hypothesis though the relation between the vascular twigs supplying the different fibres of the PMB versus
123
Int Ophthalmol
their pattern of involvement in glaucoma remains to be explored. Park et al. observed that eyes with involvement of superior and inferior parafoveal area may be thought to be phenotypically different from those that do not have involvement even till late stages of glaucoma [16]. This phenotypic difference may be attributed to differences in the topographical location of the fibres within PMB itself or differential involvement of vascular twigs supplying these areas [17]. High myopes have abnormally shaped tilted discs which may alter the structural organisation of the fibres within the disc [3, 6, 7, 18–20]. Kim et al. recently reported that progressive tilting of the optic disc with nasal shift of temporal optic disc margin with corresponding enlargement of peripapillary atrophy occurred in children with myopic shift [20]. Such displacements may also shift the CVT trunk farther away from the macular fibres. We excluded high myopes and eyes with tilted discs. Yet, we found factors independent of axial length or refractive error responsible for early involvement of the central visual field in some eyes. An inferior paracentral scotoma can significantly affect a patient’s quality of life and it is important to detect these abnormalities. Since structural alterations like RNFLD precede functional visual field changes, detection of superior PMB defects could be useful to predict development of inferotemporal paracentral scotomas. In summary, placement of the CVT trunk away from the centrally located PMB fibres may be a phenotypic risk factor for early involvement of these fibres. This would help us predict the future outcomes of eyes at risk in terms of visual function and quality of life. Conflict of interest
None
References 1. Radius RL, Anderson DR (1979) The course of axons through the retina and optic nerve head. Arch Ophthalmol 97:1154–1158 2. Ogden TE (1984) Nerve fiber layer of the primate retina: morphometric analysis. Invest Ophthalmol Vis Sci 25:19–29 3. Kimura Y, Hangai M, Morooka S, Takayama K, Nakano N, Nukada M, Ikeda HO, Akagi T, Yoshimura N (2012) Retinal nerve fiber layer defects in highly myopic eyes with early glaucoma. Invest Ophthalmol Vis Sci 53:6472–6478
123
4. Rao A, Mukherjee S (2014) Anatomical attributes of the optic nerve head in eyes with parafoveal scotoma in normal tension glaucoma. PLoS One. doi:10.1371/journal.pone. 0090554 5. Jonas JB, Mueller-Bergh JA, Schloetzer-Schrehardt UM, Naumann GOH (1990) Histomorphometry of the human optic nerve. Invest Ophthalmol Vis Sci 31:736–744 6. Quigley HA, Addicks EM (1981) Regional differences in the structure of the lamina cribrosa and their relation to glaucomatous optic nerve damage. Arch Ophthalmol 99:137–143 7. Jonas JB, Schmidt AM, Mueller-Bergh JA, SchloetzerSchrehardt UM, Naumann GOH (1992) Human optic nerve fiber count and optic disc size. Invest Ophthalmol Vis Sci 33:2012–2018 8. FitzGibbon T, Taylor SF (1996) Retinotopy of the human retinal nerve fibre layer and optic nerve head. J Comp Neurol 375:238–251 9. Pavlidis M, Stupp T, Hummeke M, Thanos S (2006) Morphometric examination of human and monkey retinal ganglion cells within the papillomacular area. Retina 26:445–453 10. Caprioli J, Sears M, Miller JM (1987) Patterns of early visual field loss in open-angle glaucoma. Am J Ophthalmol 103:512–517 11. Araie M, Sekine M, Suzuki Y, Koseki N (1994) Factors contributing to the progression of visual field damage in eyes with normal-tension glaucoma. Ophthalmology 101:1440–1444 12. Lieberman MF, Maumenee EA, Green RW (1976) Histologic studies of the vasculature of the anterior optic nerve. Am J Ophthalmol 82:405–423 13. Lee SS, Schwartz B (1992) Role of the temporal cilioretinal artery in retaining central visual field in open-angle glaucoma. Ophthalmology 99:696–699 14. Huang H, Jonas JB, Dai Y, Hong J, Wang M, Chen J, Wu J, Sun X (2013) Position of the central retinal vessel trunk and pattern of remaining visual field in advanced glaucoma. Br J Ophthalmol 97:96–100 15. Jonas JB, Ferna´ndez MC (1994) Shape of the neuroretinal rim and position of the central retinal vessels in glaucoma. Br J Ophthalmol 78:99–102 16. Park SC, De Moraes CG, Teng CC, Tello C, Liebmann JM, Ritch R (2011) Initial parafoveal versus peripheral scotomas in glaucoma: risk factors and visual field characteristics. Ophthalmology 118:1782–1789 17. Pavlidis M, Stupp T, Naskar R, Cengiz C, Thanos S (2003) Retinal ganglion cells resistant to advanced glaucoma: a postmortem study of human retinas with the carbocyanine dye DiI. Invest Ophthalmol Vis Sci 44:5196–5205 18. Mitchell P, Hourihan F, Sandbach J, Wang JJ (1999) The relationship between glaucoma and myopia: the blue mountains eye study. Ophthalmology 106:2010–2015 19. Yang H, Downs JC, Girkin C, Sakata L, Bellezza A, Thompson H, Burgoyne CF (2007) 3-D histomorphometry of the normal and early glaucomatous monkey optic nerve head: lamina cribrosa and peripapillary scleral position and thickness. Invest Ophthalmol Vis Sci 48:4597–4607 20. Kim TW, Kim M, Weinreb RN, Woo SJ, Park KH, Hwang JM (2012) Optic disc change with incipient myopia of childhood. Ophthalmology 119:21–26
