Monitoring & Maintaining Endothelial Cell Health

Expert advice on diagnostic tests and their clinical relevance.

By Marianne Price, PhD, and Francis W. Price, Jr., MD

Endothelial cell density and morphology is a key indicator when evaluating and maintaining corneal health. In this article, we discuss the structure and function of the corneal endothelium, the clinical situations in which endothelial cell assessments are most useful, the types of devices available for measuring endothelial cell density and morphology, how to verify and interpret the test results, and issues pertaining to reimbursement.

Endothelial Cell Structure and Function

The human corneal endothelium plays a critical role in maintaining corneal thickness and transparency, yet it has minimal regenerative capacity. The endothelium serves as a barrier to prevent excessive passage of fluid from the aqueous humor into the corneal stroma. In addition, the endothelial cells actively pump fluid out of the cornea to maintain the optimal spacing between collagen fibrils and thereby minimize light scattering.

In young, healthy patients, corneal endothelium appears as an array of hexagonal cells of similar size (Figure 1). The endothelial cell density decreases slowly as we age, and it can decline much more rapidly in certain disease states, in corneal transplants and after trauma, including surgical trauma. As endothelial cells die off, the remaining cells expand laterally and migrate, as needed, to cover the inner corneal surface; as they do so, they become less regular in size and shape.

Figure 1. Specular microscopy image of normal corneal endothelium, which appears as a fairly regular array of hexagonal cells, similar in size.

Infants can have an endothelial cell density in excess of 4000 cells/mm². Amazingly, the cell density can drop down to about 400 to 500 cells/mm² without any change in either corneal thickness or transparency. However, with further attrition the endothelium eventually becomes unable to adequately maintain its barrier and pump functions, and the cornea begins to swell and become cloudy.

The Many Clinical Uses of Endothelial Cell Measurements

Monitoring endothelial cell density and morphology can help you assess the risk of corneal decompensation and alert you to the need for patient counseling and preventative measures.

■ Use with cataract surgery. For cataract surgeons, anterior chamber intraocular lenses used to pose the biggest threat to endothelial cell health. With modern cataract surgical techniques and intraocular lenses, we no longer see the acute or chronic endothelial damage that we used to see 20 or 30 years ago. Nevertheless, after a previous injury, previous intraocular surgery or anything you think might have damaged the corneal endothelium, it's a good idea to do a cell count to assess the risk of corneal decompensation with planned intraocular surgery.
If the endothelial cell density is relatively low, you can inform the patient of the risk of corneal decompensation before performing the surgery. For instance, we have previously reported the relative risk of corneal decompensation with IOL exchanges.¹ Also, a dispersive rather than cohesive viscoelastic can be used to help protect the cornea, because it coats the corneal endothelium better during surgery. These are important considerations with IOL exchange or any secondary intraocular surgery, including cataract removal.

■ Use with refractive surgery. Endothelial cell density measurements also play an essential role in refractive surgery. If a patient has a low endothelial cell density, they should not get a phakic IOL. In those patients who are good phakic IOL candidates, it is important to follow the endothelial cell counts over time to make sure there is no chronic cell loss from inflammation or from lens touch if the phakic IOL was placed in the anterior chamber. In cases of decreasing endothelial cell counts after phakic IOL placement, one needs to determine the underlying problem and treat it. For instance, if there is chronic inflammation, topical steroids may be helpful. In some rare cases, lenses may need to be removed or repositioned.
■ Monitoring cornea transplant health. Regular cell density measurements are also important for monitoring corneal grafts. A sudden drop in cell density may indicate the presence of inflammation or subclinical immunologic rejection/apoptosis that needs to be treated with corticosteroids. Postoperative cell counts may also give us a predictive factor in determining the life expectancy of a corneal graft. The Cornea Donor Study showed that penetrating grafts with low cell density at six months were significantly more likely to fail within five years than grafts with a higher cell density.²
■ Importance after glaucoma surgery. Among the eyes that need to be monitored most closely are those that have undergone previous glaucoma surgery. We haven't yet quantified the corneal endothelial cell damage associated with different types of filtering surgery, but we have noticed that eyes with previous glaucoma surgery are increasingly becoming an indication for corneal transplant surgery due to corneal decompensation.³

Everybody understands how chronic damage can occur from long tubes placed in the eye with shunts: the tubes can touch the cornea and cause endothelial damage due to proximity or eye rubbing. The endothelium can also suffer acute damage from a flat chamber at the time of glaucoma surgery. Less appreciated is that the blood/aqueous barrier appears to be disrupted with filtration surgery, and this may cause long-term damage from apoptosis of endothelial cells. We have recently shown that the blood/aqueous barrier appears to be broken in eyes with shunts, resulting in abnormally elevated levels of serum proteins in the aqueous humor.⁴

Imaging Devices

The corneal endothelium can be imaged with specular or confocal microscopes. Specular microscopes image light reflected from an optical interface. A small amount (0.02%) of incident light is reflected from the interface between the corneal endothelium and the aqueous humor because of the difference in the index of refraction, and this can provide excellent images of the endothelial cells.

Confocal microscopes are designed to focus and detect light rays so that images can be obtained from different depths while minimizing the amount of out-of-focus signal from above and below the focal plane. These devices can clearly image the tear film, corneal epithelium, nerve plexus, stromal keratocytes and endothelium, and can also show the presence of invading organisms, such as Acanthamoeba or filamentous fungi. Therefore, confocal microscopes are generally more versatile but significantly more expensive than specular microscopes.

Most devices include a fixation light to facilitate imaging the central corneal endothelium. Thus, the measurements are usually taken in the visually significant area of the cornea but may not be representative of peripheral areas, particularly if the eye has experienced surgery or trauma.

Contact vs. Non-contact Imaging

Contact and non-contact versions are available for both specular and confocal microscopes. Patients usually prefer non-contact imaging. With this technique, the patient should be instructed to blink to wet the cornea and then hold still during image capture to achieve optimal image sharpness.

The non-contact devices employ auto-tracking and focusing technology to image the endothelium with minimal intervention by the operator. However, if the cornea is thickened, the automated capture mode may fail to capture an image and the operator may need to use manual focusing.

Obtaining clear endothelial images can be especially difficult with thick or hazy corneas. A confocal microscope used in the contact mode generally provides the clearest endothelial images in challenging cases.

Image Analysis

The microscopes designed for clinical use include software for semi- or fully-automated analysis of endothelial cell density and morphology (Figure 2A). Some devices allow manual identification of the cells (Figure 2B) and/or adjustments to the automated detection of the cells.

Figure 2. Automated and manual endothelial cell identification on the same image: A. Identification of the endothelial cells by automated image analysis software.

B. Manual endothelial cell identification. The centers of contiguous cells and the border around the group were manually marked to define the area of analysis (variable frame analysis).

Following either automated or manual cell detection, the software automatically calculates the cell density as well as morphometric indices, including polymegathism (the variation in the cell area as determined by the coefficient of variation) and pleomorphism (the variation in cell shape as represented by the percentage of hexagonal cells).

The accuracy of the image analysis depends upon the image quality, the instrument calibration, how well the sampled area represents the entire population of endothelial cells in that patient, as well as the technician's understanding of the analysis method and skill in performing it.⁵

Counting a larger number of cells and imaging several distinct areas improves the accuracy of the analysis. Particularly when using the “centers method” (Konan Medical), the technician should accurately mark the centers of at least 100 cells (Figure 3), because all peripheral cells along the edge of the marked group are discarded from the analysis. It is also important for the operator to exclude unanalyzable areas from the analysis frame (Figure 4). The most common errors with manual cell identification relate to improper application of the analysis method, such as double marking cells or not marking cells.

Figure 3. Manual endothelial cell identification. In the “centers analysis” method, the technician marks the centers of contiguous cells but not the border around them. With this method, at least 100 cells should be marked, because the information from the peripheral cells is only used to determine the boundary and not included in the cell density and morphology analysis. In this image only 47 cells were identified, which is not enough for accurate measurements.

Figure 4. Example of including unanalyzable areas in the image frame, which can produce erroneous results.

The most common error with automated analysis is failing to recognize when the software has traced the cell borders inaccurately. Automated software can have difficulty with correctly identifying the cells when the image quality is poor or the endothelial cell density and morphology deviate significantly from normal. In fact, it is not uncommon for automated software to erroneously give a reading within the normal range, even in corneas that have dense guttae and no recognizable endothelial cells (Figure 5). Therefore, it is important to use an imaging device that allows the operator to make corrections to the automated cell identification step, such as redrawing cell borders.

Figure 5. Example of automated cell count software erroneously producing a cell density reading of 2398 cells/mm² in a Fuchs' dystrophy patient with dense guttae and no recognizable endothelial cells in the central cornea.

It is essential for the physician, as well as the technician, to critically analyze how well the software matched the cell appearance. In a recent study, we found that some automated software programs were much better than others at matching manual cell identification by an experienced operator.⁶

Reimbursement

Endothelial cell photography is a covered procedure (CPT 92286) when reasonable and necessary for patients who meet one or more of the following criteria:

• Are about to undergo secondary IOL implantation.
• Have had previous intraocular surgery and require cataract surgery.
• Are about to undergo a surgical procedure associated with a higher risk to the corneal endothelium; i.e., phacoemulsification, or refractive surgery.
• Have slit-lamp evidence of endothelial dystrophy or corneal edema.
• Have evidence of posterior polymorphous dystrophy of the cornea (371.58) or irido-corneal-endothelium syndrome.
• Are about to be fitted with extended-wear contact lenses after intraocular surgery.

However, when a presurgical exam for cataract surgery is performed and the only visual problem is cataracts, endothelial cell photography is covered as part of the pre-surgical evaluation, rather than in addition to it. Currently, Medicare reimburses about $100 for endothelial cell photography.

Interpreting and Using the Results

Most patients with a normal cornea, including the elderly, have an endothelial cell density of at least 2000 cells/mm² and less than 300 cells/mm² difference between fellow eyes. However, a minimum cell density of 1000 to 1200 cells/mm² should allow a patient to safely undergo most anterior segment surgery, assuming that cell loss typically falls within the range of 0-30% for any given intraocular procedure. From a morphometric standpoint, a cornea with a coefficient of variation over 0.40 or 50% hexagonal cells should be considered abnormal and at increased risk for postoperative edema.⁵

Assessment of endothelial cell density and morphology allows surgeons to take appropriate measures to protect the corneal endothelium, when necessary, and also facilitates counseling patients about the risk of postoperative corneal edema.

Finally, if a patient does experience endothelial decompensation, the best option is targeted replacement with endothelial keratoplasty, a surgical procedure that is safer and provides faster and more predictable visual recovery, with fewer activity restrictions than a full thickness graft.⁷ OM

References

1. Price FW, Wellemeyer ML. Transscleral fixation of posterior chamber intraocular lenses. J Cataract Refract Surg 1995; 20:567-573.
2. Lass JH, Sugar A, Benetz BA, et al. Endothelial cell density to predict endothelial graft failure after penetrating keratoplasty. Arch Ophthalmol 2010; 128:63-9.
3. Price MO, Fairchild KM, Price DA, Price FW. Descemet's stripping endothelial keratoplasty five-year graft survival and endothelial cell loss. Ophthalmology 2011; 118:725-9.
4. Anshu A, Price MO, Richardson MR, Segu ZM, Lai X, Yoder MC, Price FW. Alterations in the aqueous humor proteome in patients with a glaucoma shunt device. Molecular Vision 2011; 17:1891-1900.
5. Benetz BA, Yee R, Bidros M, Lass J. Specular Microscopy. In: Krachmer JH, Mannis MJ, Holland EJ, eds. Cornea: fundamentals, diagnosis and management. Mosby Elsevier 2011.
6. Price MO, Fairchild KM, Price FW. Comparison of manual and automated endothelial cell density analysis in normal eyes and DSEK eyes. In submission.
7. Price MO, Price FW. Endothelial keratoplasty — a review. Clin Experiment Opthalmol 2010; 38:128–140.

	Marianne Price, PhD, is executive director of the Cornea Research Foundation of America, where she directs research and education programs, oversees finances and supervises daily activities. She can be reached at marianneprice@cornea.org.
	Francis Price, MD, founded the institution in 1988 and is president of the board. He has served as a principal investigator or medical advisor on over 100 clinical studies. He is also medical director of the Price Vision Group. Both organizations are based in Indianapolis.