Hand-drawn picture in Prof. Ernst Fuchs’ original paper showing epithelial bullae on the right side (Fuchs )
Cornea guttata (Fig. 3.2) is thus a typical finding in FECD , but not each cornea guttata will progress to FECD . In its strict sense, cornea guttata represents an isolated clinical sign of the endothelial basement membrane (also called Descemet’s membrane) without affecting the microscopic tangible architecture of corneal endothelium and corneal stroma . FECD in contrast is a bilateral, often asymmetric dystrophy of the corneal endothelium characterized by morphological changes in corneal endothelial mosaic leading to corneal edema and secondary stromal and epithelial changes .
Cornea guttata: excrescences of Descemet’s membrane as seen in scanning electron microscopy
3.2 Clinical Phenotypes of FECD
There are two variants of FECD, early-onset FECD and late-onset FECD . Most cases begin in the fourth decade of life or later, but the early variant that counts for 1 % of all FECD patients may start in the first decade . The early form can be attributed to a specific genetic defect, i.e., alpha 2 collagen VIII (COL8A2) on the 1p34.3-p32 gene locus . The guttae may develop already in the first decade of life. They are typically small and monomorphic and tend to spread almost from limbus to limbus. This indicates a several-fold more rapid deposition of DM extracellular matrix, supported by the fact that Descemet’s membrane in these patients is thicker (up to 38 μm) compared to normal controls where secretion of extracellular matrix by corneal endothelial cells extends tonu lo lypsis throughout the whole life . The inheritance of early-onset FECD is autosomal dominant [9, 13].
The more typical late-onset form of FECD progresses through four clinically defined stages . Here, larger and different-sized guttae start to develop in the central area of the cornea. Late-onset FECD has female predominance and the disease can need decades to develop . Patients complain about intermittent reduced vision from epithelial/stromal edema. Typically, visual acuity is worse in the morning due to increased stromal/epithelial edema after an overnight period of eye closure . However, best-corrected visual acuity typically increases during the day during the first years and even decades after the onset of the disease.
Guttae are best diagnosed at the slit lamp after dilation of the pupil in retroillumination of focusing on Descemet’s level (Fig. 3.3) and/or using a broad slit and directing the investigator’s gaze next to the slit (Fig. 3.4).
Right eye of a 67-year-old male with late-onset cornea guttata +++. Guttae are best diagnosed at the slit lamp after dilation of the pupil in retroillumination (a) of focusing directly on Descemet’s membrane (b)
In advanced stages, guttae may also be well visualized using a broad slit beam and directing the investigator’s gaze next to the slit (arrows). Please note typical melanin deposits in the endothelial layer, in the area of the broad beam
Guttae alone may not cause corneal edema and mild corneal edema may not result in decreased vision . On the other side, corneal guttae without edema may cause the quality of vision to deteriorate. Intraocular forward light scatter caused by corneal guttae alone may result in visual disturbances . In a study to differentiate cornea guttata and FECD, Jackson et al. discovered pigment deposits on the posterior endothelial surface in all but one examined eye .
Recently, Amin SR et al. stressed that anterior corneal cellular and structural abnormalities begin early in the course of FECD, before the onset of clinically evident edema. The chronicity of these changes can explain their incomplete resolution after endothelial keratoplasty .
Increasing epithelial and stromal edema in more severe stages of the disease is associated with decreased vision (Fig. 3.5). Pain, photophobia, and epiphora are common in advanced stages due to epithelial erosions resulting from burst epithelial bullae occurrence . In addition to pain, these patients are at risk of infectious keratitis having lost their protecting epithelial barrier.
Severe epithelial and stromal edema in FECD associated with decreased vision and multiple epithelial bullae (arrows) (“bullous keratopathy”)
Typically FECD does not recur on the graft after lamellar or penetrating keratoplasty. Thus, guttae on the graft are typically transplanted, because they were not detected in the eye bank during processing of the donor tissue .
3.3 Interpretation of Morphological Examinations
Because FECD is a clinical entity for more than 100 years, several technical devices have been established to aid the slit-lamp-based diagnosis. These may be helpful in determining the clinical stage and, therefore, determining the appropriate time for surgery. Whether a cornea guttata with a best-corrected visual acuity of 20/25 without any endothelial decompensation should be treated surgically with Descemet’s membrane endothelial keratoplasty (DMEK) remains at least questionable! Due to their noninvasive nature, all examinations may be performed at all stages of the disease. These include:
Noncontact specular microscopy
Anterior segment OCT
In vivo confocal microscopy
3.3.1 Noncontact Specular Microscopy
Corneal endothelium is – in combination with the excrescences from Descemet’s membrane – the key histological layer for diagnosis of FECD. In the morphological picture, two main entities may be estimated:
Pattern of corneal guttae
Morphology of endothelial cell mosaic
Typically modern (semi)automated specular microscopes display (1) number of analyzed cells; (2) density of cells per 1 mm2; (3) average size of analyzed cells; (4) standard deviation of analyzed cell size; (5) coefficient of variance of the analyzed cells, depending on the relation of the mean cell size and the standard deviation; (6) size of the largest analyzed cell; and (7) size of the smallest analyzed cell (Fig. 3.6).
An 80-year-old male with asymmetrical expression of FECD. (a) OD: specular microscopic image of the endothelium with few guttae of different sizes, normal tomographic central thickness of 542 μm (Pentacam, c), and VA of 20/20. (b) OS: no specular microscopic image possibly due to major edema, increased tomographic central thickness of 686 μm (Pentacam, c), and VA of 20/100
Interestingly, cornea guttata can be seen on the slit lamp before it can be demonstrated in the specular microscopy . The extent of cornea guttata seems to follow a clinically identifiable pattern course. The endothelial mosaic has firstly a strawberry pattern  which is the first sign for cornea guttata. This strawberry pattern can present small hyperreflective dots suggestive of protrusions of Descemet’s membrane in the center of hyporeflective areas . As cornea guttata progresses, these hyporeflective areas get larger resulting in a craterlike appearance . According to Zoega, the specular image can be divided in five grades. In the fifth grade, guttae occupy more than 50 % of the image .
Healthy corneal endothelial cells appear as a regular array of mainly hexagonal cells which exhibit bright cell bodies and dark cell borders . The appearance of guttae in specular microscopy and confocal microscopy is similar . Endothelial cell morphology is one key for the diagnosis of FECD: the main changes include polymegalism (variation in cell size) and pleomorphism (variation in cell shape) . Patients with cornea guttata have reduced proportion of hexagonal cells, but increased proportion of tetragonal, pentagonal, heptagonal, and octagonal cells . This is contradicted by Zoega et al. who did not detect a difference in hexagonality between corneas with and without guttae during a 7-year follow-up . This is in accordance with a study by Kitagawa et al. who also did not detect changes in hexagonality, pachymetry, or visual function in a group of Japanese patients with primary cornea guttata . Therefore, it is insecure whether polymegalism and pleomorphism of the endothelium in the presence of cornea guttata are already diagnostically decisive in order to talk about FECD. Nevertheless, it is a common sense today that in FECD, polymegalism and pleomorphism do increase . The authors attribute this to oxidative stress in general or from hypoxia.
Several attempts have been made to standardize specular microscopy. They focus on the fact that endothelial cell density per square millimeter is notoriously overestimated being very much depending on the individual investigator. It is difficult to find the same area once measured again for a sequential measurement, leading to a considerable variation of images taken from the center of the cornea with guttae . Neither manual nor automated endothelial cell count will give valid results in case of unevenly distributed guttae (Fig. 3.7). There is a possibility to create an effective cell density, correlating the fraction of clear endothelium and the fraction of the image area free of guttae. But this method is based on several assumptions. One assumption is that the scanned area is representative of viable cell density . This depends on the severity of FECD. It could be demonstrated that most of the changes in FECD occur in the central 0.6 mm of the cornea . Corneal edema can lead to light scattering, thus reducing image quality and making central analysis more difficult. In the corneal periphery at 3.7 mm, there is more damage to cells inferotemporally, but no positive correlation with pachymetry could be established so far .
Automated endothelial cell in specular microscopy count will not give valid results in case of unevenly distributed guttae. Especially in large guttae and large endothelial cells, the automatic algorithms result in a too high number of endothelial cells, because too many cell borders are typically falsely introduced, even within the area of a large gutta (arrows)
The pachymetry may range up to 1100 μm in the late stage of FECD compared to normal thickness of 530 μm [21, 29]. However, the change of central corneal thickness during clinical follow-up visits is more important than the overall value . It is of importance that the thickness measured also depends on the system used. Given the difficulty in defining the exact center of the cornea, separate readings should be obtained if the central corneal thickness is manually measured using ultrasound. Today, tomography (e.g., using the Pentacam) allows better comparison of central corneal thickness values during individual follow-up. A central corneal thickness greater than 640 μm usually indicates corneal edema  (Fig. 3.6). Patients with central corneal pachymetry higher than 650 μm have greater than 85 % probability of epithelial edema occurrence . Four millimeters from the center, the cornea is still statistically thicker than normal, but this seems to be clinically much less important. It has been suggested to use a central-to-peripheral thickness ratio (CPTR), i.e., measuring the center of the cornea and at 4 mm to detect individual progression of the disease .
3.3.3 Anterior Segment OCT
A theoretical advantage of anterior segment OCT is the possibility to measure corneal thickness layer by layer and not as overall thickness, therefore being able to better differentiate between stromal and epithelial edema. This method holds a second advantage in FECD patients: Descemet’s membrane is visualized as a thickened band of two opaque lines. The anterior line is smooth and represents the stromal face of Descemet’s membrane, while the posterior line is irregular and wavy with local thickenings strongly corresponding to the histological image . This might allow to determine the progression of the disease. Unfortunately, this clear separation is restricted to ultrahigh-resolution anterior segment OCTs. In contrast, commercially available anterior segment OCTs have limited ability even to depict Descemet’s membrane at all.
3.3.4 In Vivo Confocal Microscopy (IVCM)
IVCM is able to demonstrate microstructural changes that correlate well with histological studies. IVCM showed marked diminishment of total nerve number and number of nerve branches in the subbasal corneal nerve plexus with increasing stages of FECD compared to a control group . Corneal sensation is mildly reduced in FECD patients at stage IV when tested with a Cochet-Bonnet esthesiometer . It has to be kept in mind, however, that there is also a decrease of nerve density in the subbasal plexus with age . But nerve alterations in the subbasal plexus can already be demonstrated in early stages of FECD . They seem at first not to have an impact on corneal sensitivity. But in a study on 42 eyes after Descemet’s stripping endothelial keratoplasty, corneal sensitivity remained lower than in controls for 36 months after surgery . Pleomorphism and polymegalism can be demonstrated in IVCM and noncontact specular microscopy [36–38]. Guttae can be demonstrated in IVCM as hyporeflective round structures with ill-defined borders. Sometimes, there is a sharp hyperreflective spot in the center of a gutta [36, 37]. In addition, IVCM can help to assess endothelial cell count and morphology in analogy to specular microscopy
3.4 Clinical Stages and Disease Progression
Four clinical stages have been described by Adamis et al. . Cornea guttata starts centrally and spreads peripherally, but the vision is not affected (stage I). Some patients demonstrate cornea guttata but never progress to later stages, and others develop stromal edema due to endothelial decompensation with concomitant mild loss of vision (stage II). Corneal endothelium has a beaten metallike appearance with or without pigment dusting. Corneal guttae in adult-onset FECD are larger than those seen in early-onset FECD. Descemet’s membrane is thickened. Stromal edema may progress further toward the epithelium causing intra- and interepithelial edema (epithelial bullae) (so-called bullous keratopathy…) with loss of vision (stage III). Patients with central corneal pachymetry higher than 650 μm have a greater than 85 % probability of epithelial edema occurrence . Subepithelial fibrous tissue (“pannus”), stromal scarring, and peripheral superficial vascularization may occur in long-standing cases from chronic edema, and the pain subsides (stage IV) (Fig. 3.8) .
(a, b) Severe bullous keratopathy with subepithelial fibrosis (“pannus”) and peripheral superficial vascularization (b) in long-standing FECD (stage IV)
This grading system still strongly relies on subjective parameters which are exclusively taken from the slit lamp. Stages I and II can be difficult to differentiate. Therefore, it is no surprise that in a large study with 100 FECD patients, only moderate agreement between two corneal specialists for this clinical grading was found .
The two main functions of the endothelium, barrier and water pump, follow different courses of decline. The barrier function depends on the ability of endothelial cells to fully cover the stromal surface and to maintain cell-to-cell tight junctions. Both appear to be intact until end-stage disease. However, the pump function is gradually compromised, as evidenced by reductions in Na+/K+ATPase expression with subsequent stromal edema [30, 39]. The stromal changes are likely to occur secondary to corneal edema rather than due to FECD per se.
Another possibility of staging was presented by Krachmer et al. who focused on the number and confluence of corneal guttae and developed a grading scale from 1 to 6  currently used by the Fuchs endothelial dystrophy genetics multicenter study group . It has been reported that guttae are larger in the center of the cornea compared to paracentral areas .
Significant differences in central corneal thickness are also detectable at early grades of FECD; especially 1–2 mm of confluent guttae may begin to develop central corneal thickening  – even resulting in localized (para)central bullous keratopathy (Fig. 3.9). In a large multicenter study, female sex and smoking were associated significantly with higher stages of cornea guttata, but not with increased central corneal thickness . It has been suggested to use central corneal thickness to determine the grade of disease progression, but this alone might not be unequivocal because of well-known interindividual variety of central corneal thickness .
A 73-year-old male with good BCVA of 20/30 (!) despite (para)centrally localized focal bullous keratopathy in FECD. (a) Overview and (b) slit beam evaluation
Concerning contact lens wearing, it was shown that morphology of the endothelial cells is changing toward increasing proportions of polymegalism and pleomorphism during increasing corneal thickness . This phenomenon points toward the conclusion that a deviation of the endothelial morphology from a homogeneous hexagonal shape goes along with a deterioration of the endothelial pumping function.
From our point of view, progression should be defined by looking at the following criteria:
Best-corrected visual acuity (BCVA)
Difference of BCVA between morning (typically worse due to closed eyelids overnight) and afternoon
Subjective glare (0, +, ++, +++) especially at night
Corneal thickness profile (tomography, e.g., Pentacam or anterior segment OCT)
Endothelial cell morphology (pleomorphism, polymegalism) (specular microscopy or confocal microscopy)
BUT NOT “endothelial cell count,” because this very much depends on the investigator and is notoriously overestimating the amount of the endothelium, strongly depending on where the “region of interest (ROI)” is placed 
These criteria are also important for judging whether cataract surgery is still a reasonable option in FECD. Some authors state that below a corneal thickness of 640 μm, modern phacoemulsification would be feasible [43, 44], although Afshari NA et al. found that pachymetry-determined corneal thickness was a poor predictor of visual acuity until extreme levels of corneal edema were reached. At a corneal thickness of approximately 775 μm and greater, mean BCVA was 20/100, compared with a BCVA of 20/60 below that thickness level .
3.5 Differential Diagnosis
Differential diagnosis of endothelial dystrophies is typically an easy task. It includes:
Posterior polymorphous corneal dystrophy (PPCD)
Congenital hereditary endothelial dystrophy (CHED)
X-linked endothelial corneal dystrophy (XECD)
3.5.1 Posterior Polymorphous Corneal Dystrophy (PPCD)
In posterior polymorphous dystrophy, the endothelial cells undergo a transformation to epithelial-like cells  (Fig. 3.10a, b). Scattered endothelial vesicles can be shown with retroillumination . Of note is that the vesicles in retroillumination can be visualized in IVCM as large oval structures that contain abnormal endothelial cells  which allow a clear distinction from FECD corneas. A railroad-shaped pattern of the corneal endothelium can be seen at the slit lamp and is almost pathognomonic  (Fig. 3.10c). Eyes with posterior polymorphous corneal dystrophy (PPCD) often need no therapy for decades. However, the intraocular pressure has to be monitored on a regular basis.
In posterior polymorphous corneal dystrophy (PPCD), the endothelial cells undergo a transformation to epithelial-like cells (a overview. b slit beam). Often a railroad-shaped pattern (arrows) of the corneal endothelium can be seen at the slit lamp which is almost pathognomonic (c)
3.5.2 Congenital Hereditary Endothelial Dystrophy (CHED)
Congenital corneal clouding in the form of a milky ground glass appearance is the landmark of CHED, usually already present at birth  (Fig. 3.11). Specular microscopy in patients only mildly affected has not yet been published. But specular microscopy in parents of CHED patients may show cornea guttata .
Congenital hereditary endothelial dystrophy (CHED). (a) Overview, and (b) slit beam
3.5.3 X-linked Endothelial Corneal Dystrophy (XECD)
According to the heredity, only males are affected . The so-called endothelial moon craters are the hallmark of this disease (Fig. 3.12a) that can also be demonstrated in female carriers. These alterations compromise the whole corneal endothelium. In later life, a band keratopathy (Fig. 3.12b) may develop. Congenital corneal clouding can occur. Until now, no data on noncontact specular microscopy or corneal thickness are available in the literature because of the rarity of this disease.
X-linked endothelial corneal dystrophy (XECD). (a) “Moon craters.” (b) Band-shaped keratopathy (arrows) in later stages
Differential diagnosis of other alterations of the corneal endothelium include:
Melanin dispersion syndrome
Pseudophakic bullous keratopathy
3.5.4 PEX Keratopathy
Whereas fibroblastic metaplasia and the production of abnormal and excess extracellular material by corneal endothelium are believed to be a nonspecific common response to stress, injury, or disease, the production of PEX fibers is a specific and pathognomonic feature of the PEX syndrome. Because, however, this pathognomonic feature is not evident throughout all five stages  of the disease, the diagnosis of PEX keratopathy may rely on other clinical and histopathologic criteria, such as diffuse edema, irregular thickening of Descemet’s membrane, and pronounced melanin phagocytosis in the presence of an ocular PEX syndrome [49–52].
Circumscribed excrescences of the posterior Descemet’s membrane (“guttae”) are the clinical and histopathologic hallmark of FECD. The endothelium is attenuated, but mostly intact and shows moderate fibroblastic transformation and moderate amounts of phagocytosed melanin granules. FECD shows obvious similarities and differences with PEX endotheliopathy: the diffuse decompensation pattern on PEX eyes can be easily clinically distinguished from the corneal edema in classical FECD, which usually starts centrally and spreads peripherally, and also from bullous keratopathy (e.g., after anterior chamber lens implantation), which usually starts in the limbal region and spreads centrally (Fig. 3.13). Further clinical differences between Fuchs and PEX endotheliopathies include the presence of typical guttata in FECD which is usually lacking in PEX eyes (Figs. 3.14 and 3.15) and the extent of endothelial loss, which is often more pronounced in PEX eyes (Table 3.1). Besides more severe polymegalism and pleomorphism, PEX eyes may show hyperreflective retrocorneal whitish deposits at slit-lamp examination (Fig. 3.16).