Previous histomorphometric studies on Caucasian globes have revealed that in eyes with end-stage glaucoma compared with normal eyes, the lamina cribrosa was significantly thinner, that the part of the outer lamina cribrosa surface directly exposed to the pia mater and indirectly exposed to the cerebrospinal fluid space was significantly wider and that the shortest distance between the intraocular space and the cerebrospinal fluid space was significantly shorter.
9 These studies also showed that in eyes with a markedly elongated axial length, the lamina cribrosa was significantly thinner than in eyes with a normal axial length, thus decreasing the distance between the intraocular space and the cerebrospinal fluid space.
11 In the peripapillary region, a loss of retinal pigment epithelium cells and photoreceptors was described.
14
The present study is the first to characterize laminar and peripapillary scleral architecture in normal eyes, glaucomatous eyes with a normal axial length, and glaucomatous Chinese eyes with a markedly elongated axial length. Its principal findings are as follows. First, laminar thickness: The lamina cribrosa was significantly (P < 0.001) thicker in the normal group than in the glaucomatous group, in which it was significantly (P < 0.001) thicker than in the glaucomatous elongated axial length group. The lamina cribrosa thickness decreased significantly (P < 0.001) with increasing axial length and presence of glaucoma (P < 0.001). Correspondingly in a multivariate analysis, a thin lamina cribrosa thickness was significantly associated with a long axial length (P < 0.001) and with the presence of glaucoma (P < 0.001) as an additional and independent factor. In agreement with the reduced thickness of the lamina cribrosa, the distance between the intraocular space and the cerebrospinal fluid pressure space was significantly (P < 0.001) shorter in the two glaucomatous groups than in the normal group. Second, parallel to the thinning of the lamina cribrosa in the glaucomatous elongated axial length group, the peripapillary scleral thickness decreased significantly with increasing axial length, independent of the presence or absence of glaucoma. Correspondingly, the peripapillary scleral thickness did not vary significantly between the glaucomatous group and the normal group. The part of the sclera passing into the optic nerve meninges (approximately 50%) did not vary significantly in its percentage of the whole sclera thickness between all three study groups. Third, the shortest distance between the intraocular space and the cerebrospinal fluid pressure space was significantly (P < 0.001) longer in the normal group than in the glaucomatous group and than in the glaucomatous elongated axial length group. In all eyes included in the study, the region with the shortest distance between the intraocular pressure space and the cerebrospinal fluid pressure space was located in the periphery of the optic nerve head.
The finding of a thinner lamina cribrosa in the glaucomatous eyes versus the normal eyes in Chinese confirms the finding of a similar study of Caucasians.
9 In that study, one concluded that the thin lamina cribrosa in the eyes with end-stage glaucoma may be one of the reasons why in clinical studies the glaucoma susceptibility increased with advanced stage of the disease.
15 16 17 18 The pathophysiologic basis for the hypothesis that a thinning of the lamina cribrosa may increase glaucoma susceptibility is that the lamina cribrosa forms the border between the intraocular space with a higher pressure and the retrobulbar space with the cerebrospinal fluid with a lower pressure.
19 20 Since the lamina cribrosa is not indefinitely thin, the translaminar pressure difference leads to a translaminar pressure gradient, the steepness of which depends on the pressure difference and the lamina cribrosa thickness. The thin lamina cribrosa in the eyes of the glaucoma group versus the eyes of the normal group may therefore lead to a steepening of the translaminar pressure gradient which, among other factors, such as a change in the biomechanical tissue properties of the lamina cribrosa, may be responsible for the increased glaucoma susceptibility in eyes with advanced glaucoma. In a similar manner, we concluded that the thinning of the lamina cribrosa in glaucomatous eyes with a markedly elongated axial length may be one of the reasons that a markedly elongated length (and thus high myopia) is a risk factor for glaucomatous optic neuropathy.
11 21 22 23 24 25 26 Fitting the hypothesis of the importance of the translaminar pressure difference and gradient, in a recent clinical study, Berdahl et al.
27 suggested that patients with chronic open-angle glaucoma have a lower cerebrospinal fluid pressure than do normal subjects.
With respect to the peripapillary scleral thickness, the present study suggested that it decreases significantly with increasing axial length, independent of the presence of glaucoma. Since the peripapillary sclera may have a major impact on the biomechanics of the lamina cribrosa, as has already been pointed out by Downs et al.,
1 28 Bellezza et al.,
29 Burgoyne and Morrison,
30 Yang et al.,
2 3 and Sigal et al.,
4 the finding may suggest that along with a thinning of the lamina cribrosa, a thinning of the peripapillary sclera may play an additional role in the pathogenesis of increased glaucoma susceptibility in eyes with a markedly elongated axial length. The thickness of the peripapillary sclera did not vary significantly between the non–axially elongated glaucoma group and the normal group, suggesting that the glaucomatous process may not lead to a marked thinning of the peripapillary sclera
(Fig. 6) . Correspondingly, the multivariate analysis including all eyes of the study showed that the presence of glaucoma was not significantly associated with the thickness of the peripapillary sclera. It may suggest that the increase in glaucoma susceptibility with the stage of glaucoma in non–axially elongated eyes may not be mainly due to changes in the thickness of the peripapillary sclera. It was remarkable that approximately 50% of the posterior sclera contributed to the sclera close to the optic nerve scleral canal, whereas approximately 50% merged with the optic nerve meninges. It has remained unclear so far which of these parts or regions of the sclera are to what extent biomechanically important for stress and strain in the lamina cribrosa. Besides glaucoma, the progressive peripapillary scleral thinning with increasing axial length may have clinical and pathogenic importance for posterior staphylomata. The existence of scleral staphylomata which are a hallmark of extremely myopic eyes with extremely elongated axial length fits well and almost confirms the data of the present study in that eyes with a markedly elongated axial length have a thinner peripapillary sclera than non–axially elongated eyes. Since posterior staphylomata are clinically not found in non–axially elongated eyes with advanced glaucoma, they do also fit with the results of our study that the thinning of the peripapillary sclera was associated with markedly elevated axial length and that it is not associated with glaucoma.
The lengths of the inner and outer surface lines in the lamina cribrosa were significantly longer in the glaucoma group than in the normal group
(Table 2) . Since the length of the Bruch’s membrane opening did not vary significantly between both study groups, the parameters inner lamina cribrosa surface line length divided by the length of Bruch’s membrane opening and outer lamina cribrosa surface line length divided by the length of Bruch’s membrane opening were significantly larger in the glaucomatous group than in the normal group. The result suggests that the glaucoma related cupping in non–axially elongated eyes may be associated with an elongation or stretching of the lamina cribrosa leading to longer surface lines. The finding that the lamina cribrosa surface lines were longest in the glaucomatous elongated axial length group may be due to the additional enlargement of the optic disc related to the marked elongation of the globe.
31 Correspondingly, the length of Bruch’s membrane opening was significantly longer in the glaucomatous elongated axial length group than in any of the two other study groups
(Table 2) .
There was generally a notable interindividual variability in the measurements of the lamina cribrosa and the thickness of the peripapillary sclera
(Figs. 4 6) . This interindividual variability was partially due to limitations of the study techniques such as oblique histologic sectioning and variations in the locations from which the sections were taken. The interindividual variability in the measurements led to a marked overlapping between the study groups (
Table 2 ,
Figs. 4 6 ). That the differences between the study groups were often higher than the standard deviations of the means, however, explains the statistical significance of the differences between the study groups for most of the parameters measured. The interindividual variations in the measurements can also be one of the reasons for the interindividually variable susceptibility for glaucomatous optic nerve damage.
The following limitations should be considered in interpreting our results: (1) The normal eyes were not normal eyes, since they contained malignant choroidal melanoma; however, because the tumor cells neither invaded the optic nerve head nor the trabecular meshwork, it is unlikely that the presence of the malignant melanoma influenced the results of the study. (2) In our two glaucomatous groups, only eyes with end-stage glaucomatous damage were studied by what may be the most advantageous method to find differences between normal eyes and eyes with glaucoma. It may be considerably more difficult to find thinning of the lamina cribrosa early in a neuropathy. In fact, in two recent studies, the lamina cribrosa, although clearly posteriorly deformed, was actually thicker in perfusion-fixed monkey eyes with very early experimental glaucoma compared with the contralateral normal eyes.
28 29 Conclusions drawn from the present study may therefore only be that thinning of the lamina cribrosa, perhaps in a sense of compression or collapse, is present in a relatively late stage of glaucomatous optic neuropathies. (3) The ages of our groups were different, with the normal group being significantly younger than the two glaucomatous groups. The differences in the outcome parameters between the study groups were highly significant, however, so that again it may be unlikely that the age difference between the study groups had a marked effect on the study results. In addition, the glaucoma group and the glaucomatous elongated axial length group did not vary in age. Furthermore, the lamina cribrosa thickness was not correlated with age in a multivariate analysis. (4) We measured only a single section per eye, and the section was not consistent among and between groups. Serial sections of the globes were not available so that it was not possible to determine whether the histologic section was located in the very center of the optic disc or ran paracentrally. Since there were no marked differences in the thinning of the lamina cribrosa between the central region versus the peripheral region of the optic disc, the inaccuracy in the exact localization of the histologic section may not have markedly influenced the determinations of the lamina cribrosa thickness and the conclusions of the study. In addition, this limitation in the design of the study again held true for all three study groups so that this systemic error may not have markedly influenced the results of the comparison of the study groups. In a similar manner, the histologic sections were not of a consistent orientation. The glaucomatous eyes were sectioned in a vertical direction, while the eyes of the nonglaucomatous control were sectioned according to the location of the tumor. Considering that the lamina is remarkably variable in its three-dimensional geometry, as shown by Bellezza et al.
29 and Burgoyne and Morrison,
30 the lamina cribrosa may be thinnest in different regions of the optic nerve head. Since the differences in the measurements between the three study groups were highly significant for all locations measured, it may be unlikely that this flaw in the study design markedly influenced the conclusions of the investigation. (5) Unlike in previous studies, we failed to determine a difference in peripapillary scleral thickness in the glaucoma group and may have missed regional differences. (6) The eyes examined in our study were fixed at an intraocular pressure of about 0 mm Hg. Previous studies on nonhuman primates by Burgoyne and Morrison have shown differences in the laminar and peripapillary scleral morphology between eyes fixed at physiologic pressures and at 0 mm Hg. This difference may have introduced an artifact into our study. As long as there were no marked differences in material properties, however, it may not have introduced a profound systemic bias and may not have influenced the results and conclusions of the study. (7) Because of swelling of the tissue after enucleation and due to the histologic preparation of the slides, the measurements given in this study will not represent dimensions as determined intravitally. It was, however, not our purpose to evaluate the measurements of the lamina cribrosa and its surrounding tissues in real dimensions, but to compare the measurements of the lamina cribrosa obtained in markedly axially elongated eyes with measurements taken in non–axially elongated eyes, with or without glaucoma. The systemic error that was introduced by the histologic preparation of the slides will have affected the specimen of all three study groups in a similar manner, since the preparation did not vary between the study groups. It may thus not have markedly affected the conclusions of the study.
Another aspect to be taken into account is that the histomorphometric measurements obtained in the present study on Chinese eyes were generally markedly smaller than those obtained in the previous study of Caucasian eyes.
9 10 11 The reasons for the discrepancies between the studies may be differences in the histologic preparation of the slides including fixation (4% formaldehyde in the present study versus a solution of 4% formaldehyde and 1% glutaraldehyde in the studies of Caucasian eyes) and staining as well as differences in the determination of the anterior surface and posterior border of the lamina cribrosa by the examiners. The differences between the studies may not indicate marked real differences between the two ethnic groups. Since it was not the purpose of the study to compare the lamina cribrosa measurements between Caucasian and Chinese eyes but to evaluate intraethnic differences between groups of Chinese eyes, a potential systemic error in the outlining of the lamina cribrosa may therefore not have markedly influenced the results and conclusions of the study.
It should be noted that our data on the relative thickness of the lamina cribrosa and peripapillary sclera in the glaucomatous group and the glaucomatous elongated axial length group may in fact be profoundly influenced by the age of the patients in each group, since we do not have data on axial length and refractive error before the onset of the ocular problems that led to their enucleation. In fact, our group of patients with a markedly elongated axial length may not have had high myopia before the onset of intraocular pressure elevation. The subset of patients who were myopic at the time of enucleation may have become so secondarily, as a response to chronic high intraocular pressure elevation. It means that it is possible that eyes may axially elongate under the influence of high intraocular pressure depending on scleral material properties. This elongation may occur also after the age of 2 to 4 years. One may assume that if the intraocular pressure is high for a long enough period (i.e., several years), the eyes of relatively young patients, such as those included in our study, may become somewhat elongated. There were no data on the magnitude of the exposure to high intraocular pressure, and there were no hard data on the age at onset. In contrast, the younger age of the patients with the markedly elongated (highly myopic) eyes in our study suggested an elongation of the globes at the onset of high intraocular pressure at a young age. This supposition may hold true particularly since the age at the known onset of the elevation of intraocular pressure was significantly lower in the glaucomatous elongated axial length group than in the non–axially elongated glaucoma group
(Table 1) . The finding that the corneal diameters did not vary between the highly elongated eyes and the nonelongated eyes does not absolutely contradict that the elongation in the highly myopic group was due to the high intraocular pressure. As long as the material properties of the cornea and sclera have not adequately been examined yet, one cannot exclude the possibility that the high intraocular pressure led to a distension of the posterior sclera in the glaucomatous elongated axial length group whereas the cornea kept a normal size due to a difference in the material properties between the posterior sclera and the cornea. In addition, the ocular trauma that led to the high intraocular pressure was usually associated with a corneal scar that may have additionally stiffened the cornea.
Correspondingly, a large body of literature documents profound enlargement of the globe accompanied by thinning of the sclera in a variety of childhood glaucomas.
32 33 34 35 36 37 38 39 40 41 42 43 44 The fact that the mean age of the glaucomatous elongated axial length group was 27.9 years compared with those of the normal and the glaucomatous groups suggests the possibility that a profound enlargement of the globe and thinning of the peripapillary sclera and lamina were part of the response to the chronic, high intraocular pressure elevation in all eyes, and that it reached the extreme in patients who were most susceptible (i.e., had the most compliant sclera and were exposed to the highest levels of intraocular pressure). In this scenario, young patients, with more compliant sclera, thinned most (becoming the highly myopic group on the basis of axial elongation of the globe) and older patients in which the sclera was presumably stiffer and therefore more resistant to deformation, axial elongation was less likely to occur, as was detectable thinning of the peripapillary sclera. Because of its retrospective nature of the study, these two interpretations of our data cannot be discounted.
In view of all the limitations mentioned, the present study as well as the previous histomorphometric investigations using the same methods have clearly to be considered preliminary pilot studies,
9 10 11 until prospectively planned investigations of serial sections have been performed. In these future studies, the techniques used in the present investigation may no longer be adequate but may require fundamental changes.
In conclusion, the present study suggests that the lamina cribrosa and peripapillary sclera are thinnest in eyes with the longest axial length in Chinese eyes with secondary intraocular pressure elevation. The lamina cribrosa was thinnest and associated with a statistically significant thinner peripapillary sclera in those eyes that demonstrated postmortem axial lengths of greater than 27.5 mm. The lamina cribrosa was thinner, but the peripapillary sclera was not thinner in those glaucomatous eyes with an axial length of less than 27.5 mm Hg. Because these data are retrospective, they may not directly pertain to eyes that are myopic in the setting of normal levels of intraocular pressure or eyes that are myopic before the onset of an intraocular pressure elevation.
The authors thank Sun Xian Li for guidance in evaluating the measurements.