The transparent nature of the cornea makes it a difficult structure to examine in vivo. Confocal microscopy has been shown to be a viable clinical research tool for examining cellular changes within the cornea. 1–7 Confocal microscopy has also been widely used in the areas of corneal disease and refractive surgery. 8–20 Its application in studying both qualitative and quantitative changes in corneal cell morphology is proving to be extremely valuable both clinically and diagnostically.
The instrument used in this research was the Tomey Confo- Scan, a slit-scanning confocal microscope with a 40×/0.75 NA immersion objective lens. This provided a magnification of ×680. It was possible to obtain good quality images of Bowman’s layer, the anterior and posterior stroma, Descemet’s layer, and the endothelium. Images of the corneal epithelium could not be obtained in all subjects.
Research using the specular microscope has established that the cellular density of the corneal endothelium decreases with age. It is now possible to examine the keratocyte population of the in vivo corneal stroma utilizing confocal microscopy. Previous studies have relied on invasive techniques. The word “keratocyte” is a Latin term meaning corneal cell (kerato:corneal and cyte:cell). Keratocytes are the principle cells of the corneal stroma and play an active role in maintaining corneal transparency and health. They may also have a role in corneal wound healing. They are responsible for the synthesis of collagen fibrils and the intercellular matrix. 21 Changes to the keratocyte population in the in vivo cornea may have dramatic implications in corneal disease, refractive surgery, and contact lens wear. Moller-Pedersen and Ehlers 22 examined the density and distribution of keratocytes using biochemical measurements of the DNA content within the corneal stroma. It was shown that cellularity increased toward the corneal periphery (60 to 70% higher cell density at the limbus); the central stroma showed 30% lower cellularity in the subendothelial region compared with the subepithelial region; and the four main quadrants of the central cornea had a uniform cell density, whereas in the peripheral areas, a 10% increase in cell density was found superiorly. A close intraindividual correlation was demonstrated. Gender had no influence on the keratocyte density, and no correlation was found between the density of the endothelial cells and keratocytes. Mustonen et al. 23 demonstrated similar results in a study using confocal microscopy. It was shown that central corneal cell densities decreased significantly with age only in the endothelium; no correlation was found with age for the other corneal layers.
The confocal microscope provides us with an ideal opportunity to examine changes to the different cellular populations of the cornea; however, to be able to monitor and document these changes within the abnormal cornea, it is essential that researchers have normative data with which they can compare. The aim of this study was to document cellular changes in the normal human cornea with age.
This study was conducted by three examiners in the form of a single center, masked, randomized trial. Ethical committee approval was obtained before commencing the study, and fully informed consent was obtained from all subjects. One hundred twenty subjects (64 male and 56 female; average age 41 ± 18) were recruited for this study via letter, e-mail or telephone (Fig. 1). Subjects of any race or gender were included to reflect the normal population. The ethnic distribution of subjects was as follows: 110 white, 1 mixed race, 2 Arabic, 1 Bangladeshi, 2 Oriental, 2 Indian, and 2 black Carribean. The following inclusion criteria were used to ensure that there were no factors present that might affect corneal physiology: no ocular or systemic disease, no pregnancy or lactation, no corneal disorder or disease, no history of contact lens wear, no previous injuries or operations to the eye. After arrival for the study visit a full history and symptoms was taken from each subject, visual acuity was measured, and slitlamp biomicroscopy was performed. Before the confocal microscopy, one drop of topical local anesthetic (Benoxinate 0.4%, Chauvin Pharmaceuticals) was applied to the lower fornix of both eyes. A drop of polymer gel (Viscotears, CIBA Vision) was applied to the microscope probe before the examination to optically couple the microscope objective lens to the cornea. Confocal microscopy was then performed on the central corneas of both eyes of each subject. Images were taken of all corneal layers and stored on SVHS videotape. The process of obtaining usable images was more difficult in some subjects. After confocal microscopy, subjects’ eyes were checked on the slitlamp biomicroscope for corneal staining before subjects were discharged. There were no cases that needed further care.
Storing and analysis of the images was carried out by all three examiners in a randomized manner. Where possible, images were saved of the epithelium, the basal membrane, Bowman’s Layer, the anterior and posterior stroma, Descemet’s layer, and the endothelium. Images of the anterior stroma were taken as being immediately posterior to Bowman’s layer, and images of the posterior stroma were taken to be immediately anterior to Descemet’s layer or the endothelium if Descemet’s layer is not visible. Where possible, at least three images of each corneal layer were saved to obtain an average reading. Analysis was performed using the dedicated Tomey base software (ConfoCommander 2.7.1) in a semiautomatic manner.
To ensure that the methodology used in this study was consistent, both intraexaminer and interexaminer variability were assessed. Fifteen subjects were randomly chosen, and images were saved of the anterior stroma, posterior stroma, and endothelium. Three images were saved of each layer. To assess intraexaminer variation, each examiner analyzed all images twice, 1 week apart. To assess interexaminer variation, the data obtained from the first analysis of the intraexaminer variation was taken for each examiner and subsequently analyzed statistically. Statistical analysis was carried out on a PC using the SPSS 10.0 statistical package.
Images of all corneal layers were successfully stored in 120 patients for both right and left eyes (240 eyes). Subsequent analysis was only possible in 164 eyes for the anterior stroma (68% success), 239 eyes for the posterior stroma (99.5% success), and 236 eyes for the endothelium (98% success). It was not possible to analyze the remaining images due to poor image quality.
Intraexaminer variation was assessed using paired sample t-tests. No statistically significant difference was demonstrated, indicating that repeated measurements from individual examiners were consistent. Interexaminer variation was assessed using analysis of variance tests. No statistically significant difference was demonstrated between examiners for anterior keratocyte density (p = 0.529), posterior keratocyte density (p = 0.258), or endothelial cell density (p = 0.475).
As a whole, the clarity of the images decreased slightly throughout life, demonstrating the established reduction in translucency of the cornea with age. Epithelial morphology and the basal membrane appeared to remain constant throughout life. The amount of nerves within Bowman’s membrane seemed to decrease from the sixth decade of life.
The mean ± SD anterior keratocyte density was 1037 ± 169 cells/mm2 (range, 504 to 1536), the average posterior keratocyte density was 571 ± 76 cells/mm2 (range, 360 to 768), and the average endothelial cell density was 3061 ± 382 cells/mm2 (range, 2132 to 4161). No statistically significant difference was detected between right and left eyes for anterior keratocyte density, posterior keratocyte density, or endothelial cell density using paired t-tests (p = 0.11, 0.52, and 0.65, respectively). Cellular densities were also shown to be unaffected by the sex of the subject with p values of 0.46, 0.55, and 0.50 for anterior keratocyte density, posterior keratocyte density, and endothelial cell density, respectively (multivariate analysis of variance).
The anterior keratocyte density, posterior keratocyte density, and endothelial cell density were all correlated with age (p ≤ 0.0001). Because no statistically significant difference was demonstrated between eyes, the mean value of the right and left eye was used for further analysis. Anterior keratocyte density was shown to decrease at a rate of 0.48% per year (Fig. 2).
Posterior keratocyte density was shown to decrease at a rate of 0.22% per year (Fig. 3). Keratocyte nuclei were found to be longer and more spindle shaped in the posterior stroma compared with the anterior stroma (Figs. 4 and 5). A 45% reduction in the keratocyte density in an anterior-posterior direction was demonstrated (p ≤ 0.0001). Folds were noted at the level of the posterior stroma from the sixth decade of life (Fig. 6). Folds were documented in 10, 18, and 29% of the population in the sixth, seventh, and eighth decades, respectively.
Descemet’s layer appeared to become more visible with increasing age. When examining the videotapes of younger subjects, this layer was not visible; the frame adjacent to the endothelium contains keratocyte nuclei and was taken to be the posterior stroma. However when sequential images of older subjects were observed, an acellular layer was seen between the endothelium and the posterior stroma.
The endothelial cell density decreased at a rate of 0.33% per year (Fig. 7). A positive correlation was demonstrated between the coefficient of cell variation and age, demonstrating an increase in polymegathism with age (Fig. 8). Endothelial guttata were shown to be present in 6, 12, and 29% of the population in the sixth, seventh, and eighth decades of life, respectively (Fig. 9).
No difference was found between right and left eyes. The sex of the subjects was also shown to have no influence on cellular densities. This is in accordance with previous published work. 23,24
The decrease in clarity seen in all corneal layers with increasing age corresponds with both the decrease in epithelial luster and the increased stromal relucency that is known to occur in the aging eye. 25 The constancy of epithelial morphology and the basal membrane is expected because it is well known the epithelium regenerates within a period of approximately 24 h. The reduction in the number of nerves at Bowman’s membrane after the sixth decade correlates with research demonstrating reduced corneal sensitivity after the age of 40 years. 26 These findings are clearly demonstrated in a reference grid published previously. 27
In this study, the mean (±SD) anterior keratocyte density was 1037 ± 169 cells/mm2 (mean age, 41 years). It was demonstrated that the anterior keratocyte density decreased at a rate of 0.48% per year. The mean posterior keratocyte density was 571 ± 76 cells/mm2 (mean age, 41 years), decreasing at a rate of 0.22% per year. Moller-Perdersen 24 showed a physiological decline of 0.3% in the keratocyte density and endothelial cell density of human corneas. Various specular microscopic and in vitro studies have demonstrated a gradual reduction in the endothelial cell density throughout life. 28–32 In our study, the mean endothelial cell density was 3061 ± 382 cells/mm2 (mean age, 41 years). We also demonstrated the established decline in the endothelial cell density, showing a drop of 0.33% per year, as well as an increase in polymegathism and pleomorphism. The cellular decline in anterior keratocyte density and posterior keratocyte density reported in our work is similar to that reported by Moller-Pedersen, 24 although different methodologies were used. It has been demonstrated that there is great diversity in the intersubject endothelial cell density 31 and that the rate of decline may be affected by a polymegathous endothelium. 32 This may account for differing results found between studies for the decline in endothelial cell density per year. There may also be an amount of intersubject variability in the keratocyte density that may help to account for the differences between published work in the same area.
It has been demonstrated that automated digital analysis of confocal microscopy images in vivo is repeatable and agrees with keratocyte density estimated from histologic sections. 33,34 This is on analysis of images obtained with the tandem scanning confocal microscope. The images obtained with the nonapplanating, real-time scanning slit confocal microscope are sharp and high contrast. It is been shown that there is no need for digital image processing techniques to enhance the contrast of single video images. 5 The analyses done by our group assessing interexaminer and intraexaminer variation verifies the repeatability of the methodology used. However, digitized analysis packages may produce an even higher level of reproducibility. It is not possible to correlate the results obtained with histological analysis, but future work may involve looking at the same data with a more automated analysis package.
It has been demonstrated that keratocyte apoptosis occurs in response to chronic epithelial injury. 35 Throughout life, there will be instances where epithelial disruptions occur in the normal eye. This study demonstrated a greater decrease per year in cell density at the level of the anterior stroma (0.48%) compared with the posterior stroma (0.22%). It is possible that this occurs due to keratocyte apoptosis in the anterior stroma.
Previous in vitro studies have demonstrated a lower cellular density in the posterior stroma compared with the anterior stroma. Petroll et al. 36 demonstrated a 30% decrease in cell density over the entire anteroposterior stromal thickness using laser scanning confocal microscopy on rabbit corneas. Patel et al. 33 also demonstrated a decrease in the rabbit cornea using a tandem scanning confocal microscope. Moller-Pedersen et al. 22 found the same rate of decrease in humans (30%). Hahnel et al. 37 showed that the cellular density of the stroma decreases progressively from the anterior to posterior stroma by 46.3% using confocal laser scanning fluorescence microscopy on human corneas. Our in vivo study demonstrated a 45.1% reduction in the keratocyte density of the posterior stroma compared with the anterior stroma. This is consistent with Mustonen et al., 23 who showed a statistically significant difference between anterior and posterior keratocyte densities.
Due to the limitations of the Tomey ConfoScan, it was only possible to analyze the anterior and posterior stroma; it was not possible to look at changes in the keratocyte density at any other stromal depth. The Tomey ConfoScan does not allow a confocal microscopy through focusing examination to be carried out. A confocal microscopy through focusing examination allows precise depth location of images throughout the corneal stroma and would therefore allow detailed investigation of cellular changes at all stromal depths. The anterior and posterior stroma were able to be precisely located in this study because they were taken as being the frames directly adjacent to Bowman’s layer and Descemet’s membrane.
The presence of dark bands in the posterior stroma was noted in some older subjects in this study. Dark bands of a similar nature, although more striking in appearance, have been documented in subjects with advanced Fuch’s endothelial dystrophy using the confocal microscope and are known to correspond to folds in the cornea caused by pathological edema. To the best of our knowledge, the frequency of folds appearing in the posterior stroma with age has not been previously documented. This work indicates that posterior folds are found more frequently with increasing age.
The fact that Descemet’s membrane becomes more visible in the later decades of life is thought to demonstrate a thickening of this layer with age. In a study measuring specimens from 12 weeks of gestation through birth and up to 60 years of age, Murphy et al. 38 showed that the thickness of Descemet’s membrane is 3 μm at birth, increasing to 10 to 12 μm at 60 years of age if no disease process intervenes. Due to the fact that the depth of field of the Tomey ConfoScan is 10 μm, Descemet’s membrane becomes more visible in the later decades of life. Endothelial guttata are an established senescent change of the endothelium. This work suggests a frequency with which they may occur.
In this study, it was only possible to analyze 68% of the images obtained from the anterior stroma. Other published work has demonstrated a higher success rate. 23 The main reason for poor image quality was lack of cooperation by the subject during the examination or an inability to hold the eyes open or fixate properly. Most of the subjects used in this research had limited experience with ocular investigations and were found to be slightly more apprehensive than subjects such as contact lens wearers who are familiar with ocular investigation techniques. A lower success rate was found for the anterior stroma compared with the posterior stroma and endothelium because this was the last area to be examined during the confocal microscopy examination. The confocal microscope was initially focused on the endothelium to ensure that it was perpendicular to the cornea; it was then moved forward to obtain images of the anterior stroma. Subjects’ fixation was found to be generally less steady toward the end of the examination, possibly due to tiredness. This may also explain why it was not possible to obtain images of the epithelium in all subjects. The success rate of 99.5 and 98% for the posterior stroma and endothelium, respectively, was similar to that reported by Mustonen et al. 23
This study demonstrates a linear decrease in both anterior and posterior keratocyte density as a function of age. It also documents the appearance of folds in the posterior stroma occurring from the sixth decade of life. These results also verify the well-established decrease in endothelial cell density with age. This data constitutes essential normative data that can be used as a control in further research into abnormal corneal conditions.
Received February 22, 2001; revision received May 25, 2001.
1. Cavanagh HD, Petroll WM, Jester JV. The application of confocal microscopy to the study of living systems. Neurosci Biobehav Rev 1993; 17: 483–98.
2. Bohnke M, Masters BR. Confocal microscopy of the cornea. Prog Retin Eye Res 1999; 18: 553–628.
3. McGhee CNJ, Keller PR. In vivo
confocal microscopy of living tissue: the cornea at a cellular level. Eyenews 1994; 5: 14–20.
4. Beuerman RW, Laird JA, Kaufman SC, Kaufman HE. Quantification of real-time confocal images of the human cornea. J Neurosci Methods 1994; 54: 197–203.
5. Masters BR, Thaer AA. Real time scanning slit confocal microscopy of the in vivo
human cornea. Appl Optics 1994; 33: 695–701.
6. Koester CJ, Auran JD, Rosskothen HD, Florakis GJ, Tackaberry RB. Clinical microscopy of the cornea utilizing optical sectioning and a high-numerical-aperture objective. J Opt Soc Am A 1993; 10: 1670–9.
7. Wiegand W, Thaer AA, Kroll P, Geyer OC, Garcia AJ. Optical sectioning of the cornea with a new confocal in vivo
slit-scanning videomicroscope. Ophthalmology 1995; 102: 568–75.
8. Cavanagh HD, McCulley JP. In vivo
confocal microscopy and Acanthamoeba keratitis. Am J Ophthalmol 1996; 121: 207–8.
9. Florakis GJ, Moazami G, Schubert H, Koester CJ, Auran JD. Scanning slit confocal microscopy of fungal keratitis. Arch Ophthalmol 1997; 115: 1461–3.
10. Somodi S, Hahnel C, Slowik C, Richter A, Weiss DG, Guthoff R. Confocal in vivo
microscopy and confocal laser-scanning fluorescence microscopy in keratoconus. Ger J Ophthalmol 1996; 5: 518–25.
11. Werner LP, Werner L, Dighiero P, Legeais JM, Renard G. Confocal microscopy in Bowman and stromal corneal dystrophies. Ophthalmology 1999; 106: 1697–704.
12. Rosenberg ME, Tervo TM, Petroll WM, Vesaluoma MH. In vivo
confocal microscopy of patients with corneal recurrent erosion syndrome or epithelial basement membrane dystrophy. Ophthalmology 2000; 107: 565–73.
13. Mustonen RK, McDonald MB, Srivannaboon S, Tan AL, Doubrava MW, Kim CK. In vivo
confocal microscopy of Fuchs’ endothelial dystrophy. Cornea 1998; 17: 493–503.
14. Kaufman SC, Beuerman RW, Kaufman HE. Diagnosis of advanced Fuchs’ endothelial dystrophy with the confocal microscope. Am J Ophthalmol 1993; 116: 652–3.
15. Chiou AG, Kaufman SC, Beuerman RW, Ohta T, Soliman H, Kaufman HE. Confocal microscopy in cornea guttata and Fuchs’ endothelial dystrophy. Br J Ophthalmol 1999; 83: 185–9.
16. Prydal JI, Dilly PN. In vivo
confocal microscopy of the cornea and tear film. Scanning 1995; 17: 133–5.
17. Heinz P, Bodanowitz S, Wiegand W, Kroll P. In vivo
observation of corneal nerve regeneration after photorefractive keratectomy with a confocal videomicroscope. Ger J Ophthalmol 1996; 5: 373–7.
18. Kauffmann T, Bodanowitz S, Hesse L, Kroll P. Corneal reinnervation after photorefractive keratectomy and laser in situ
keratomileusis: an in vivo
study with a confocal videomicroscope. Ger J Ophthalmol 1996; 5: 508–12.
19. Linna T, Tervo T. Real-time confocal microscopic observations on human corneal nerves and wound healing after excimer laser photorefractive keratectomy. Curr Eye Res 1997; 16: 640–9.
20. Moller-Pedersen T, Li HF, Petroll WM, Cavanagh HD, Jester JV. Confocal microscopic characterization of wound repair after photorefractive keratectomy. Invest Ophthalmol Vis Sci 1998; 39: 487–501.
21. Muller LJ, Pels L, Vrensen GF. Novel aspects of the ultrastructural organization of human corneal keratocytes. Invest Ophthalmol Vis Sci 1995; 36: 2557–67.
22. Moller-Pedersen T, Ehlers N. A three-dimensional study of the human corneal keratocyte density. Curr Eye Res 1995; 14: 459–64.
23. Mustonen RK, McDonald MB, Srivannaboon S, Tan AL, Doubrava MW, Kim CK. Normal human corneal cell populations evaluated by in vivo
scanning slit confocal microscopy. Cornea 1998; 17: 485–92.
24. Moller-Pedersen T. A comparative study of human corneal keratocyte and endothelial cell density during aging. Cornea 1997; 16: 333–8.
25. Kotulak JC, Brungardt T. Age-related changes in the cornea. J Am Optom Assoc 1980; 51: 761–5.
26. Millodot M, Owens H. The influence of age on the fragility of the cornea. Acta Ophthalmol (Copenh) 1984; 62: 819–24.
27. Efron N, Perez-Gomez I, Mutalib HA, Hollingsworth J. Confocal microscopy of the normal human cornea. Contact Lens Anterior Eye 2001; 24: 16–23.
28. Sherrard ES, Novakovic P, Speedwell L. Age-related changes of the corneal endothelium and stroma as seen in vivo
by specular microscopy. Eye 1987; 1: 197–203.
29. Sturrock GD, Sherrard ES, Rice NS. Specular microscopy of the corneal endothelium. Br J Ophthalmol 1978; 62: 809–14.
30. Bahn CF, Glassman RM, MacCallum DK, Lillie JH, Meyer RF, Robinson BJ, Rich NM. Postnatal development of corneal endothelium. Invest Ophthalmol Vis Sci 1986; 27: 44–51.
31. Hoffer KJ, Kraff MC. Normal endothelial cell count range. Ophthalmology 1980; 87: 861–6.
32. Blatt HL, Rao GN, Aquavella JV. Endothelial cell density in relation to morphology. Invest Ophthalmol Vis Sci 1979; 18: 856–9.
33. Patel SV, McLaren JW, Camp JJ, Nelson LR, Bourne WM. Automated quantification of keratocyte density by using confocal microscopy in vivo
. Invest Ophthalmol Vis Sci 1999; 40: 320–6.
34. Prydal JI, Franc F, Dilly PN, Kerr Muir MG, Corbett MC, Marshall J. Keratocyte density and size in conscious humans by digital image analysis of confocal images. Eye 1998; 12: 337–42.
35. Kim WJ, Helena MC, Mohan RR, Wilson SE. Changes in corneal morphology associated with chronic epithelial injury. Invest Ophthalmol Vis Sci 1999; 40: 35–42.
36. Petroll WM, Boettcher K, Barry P, Cavanagh HD, Jester JV. Quantitative assessment of anteroposterior keratocyte density in the normal rabbit cornea. Cornea 1995; 14: 3–9.
37. Hahnel C, Somodi S, Weiss DG, Guthoff RF. The keratocyte network of human cornea: a three-dimensional study using confocal laser scanning fluorescence microscopy. Cornea 2000; 19: 185–93.
38. Murphy C, Alvarado J, Juster R. Prenatal and postnatal growth of the human Descemet’s membrane. Invest Ophthalmol Vis Sci 1984; 25: 1402–15.
Keywords:© 2001 American Academy of Optometry
confocal microscopy; normal cornea; keratocyte density; endothelial cell density