**3. Impact of CCT on tonometry**

**2. Corneal anatomy and histology**

228 Glaucoma - Basic and Clinical Aspects

transparency, surface smoothness, contour, and refractive index.

and is adjacent to the epithelial basement membrane.

nents of the extracellular matrix (ECM).

thesize both collagen and proteoglycans.

The cornea, the primary refractive ocular structure that contributes to focusing the external images on the retina, measures 11 to 12 mm horizontally, 10 to 11 mm vertically, and is about 0.5 mm thick centrally. The corneal thickness increases gradually toward the periph‐ ery to about 0.7 mm. Corneal nutrition depends on both glucose diffusing from the aqueous humor and oxygen supplied from the air through the tear film and in the peripheral cornea from the limbal blood vessels [11]. The cornea accounts for more than two thirds of the total ocular refractive power. Any slight change in the corneal contour can cause a substantial change in the ocular refractive power. The corneal optical properties are determined by its

The cornea is comprised of five layers: the epithelium, Bowman's layer, stroma, Descemet's membrane, and endothelium. The epithelium, the most anterior layer, is comprised of nonkeratinizing stratified squamous epithelial cells. The epithelium and tear film form an opti‐ cally smooth surface. The Bowman's layer is the most anterior part of the corneal stroma,

The structural and optical features depend mainly on the structure and composition of the corneal stroma, which represents up to 90% of the corneal thickness. Corneal transparency basically depends on the regular spacial distribution of the stromal cells and the stromal la‐ mellae, and also on the water content of the stroma, that must be kept at a constant level of about 78%. The keratocytes are highly scattered and do not affect transparency. The lattice structure of the corneal collagen fibers, within a distance of 0.5 microns of the visible wave‐ length, is responsible for corneal transparency. Any decrease (dehydration) or increase (ede‐ ma) in this distance results in a loss of transparency. Fibrillar collagen types I and V, which are intertwined with type VI collagen filaments (collagen types III, XII, and XIV have also been found in the stroma) and corneal proteoglycans (mainly decorin associated with der‐ matan sulfate and lumican associated with keratan sulfate), are the fundamental compo‐

Negatively charged stromal glycosaminoglycans tend to repel each other, producing the corneal swelling pressure (SP) (of about 50 mmHg in the excised cornea), and can absorb and retain large amounts of water. The keratocytes lie between the corneal lamellae and syn‐

The diameter of each collagen fiber and the distance between the collagen fibers are homo‐ geneous and measure less than half of the wavelength of the visible light (400-700 nm). This anatomic distribution of fibers is responsible for the fact that the incident light rays scattered by each collagen fiber are cancelled by the interference of other scattered rays, which allows

Descemet's membrane, the basement membrane of the endothelium, is highly elastic and

The endothelium, the innermost corneal layer, is a monolayer of hexagonally shaped endo‐ thelial cells arranged in a mosaic pattern. The integrity of this layer and the correct function

the incident light to pass through the cornea without optical disruption.

can withstand high pressure. When injured, it can regenerate.

The Ocular Hypertension Treatment Study (OHTS) [13] was a multicenter, randomized, prospective clinical trial of the efficacy of topical ocular hypotensive medications in delaying or preventing glaucoma onset in patients with ocular hypertension (OHT). Based on the OHTS, the CCT measured by pachymetry (Figure 1) has become important in glaucoma, and the study showed that the CCT is a significant predictor of the patients with OHT who are at higher risk of developing glaucoma, with a hazard ratio of 1.82 for each 40-µm thin‐ ning of the CCT.

**Figure 1.** Ultrasound Pachymeter DGH 500 (Pachette™)

Eyes with a CCT of 555 µm or less had a three-fold greater risk of developing glaucoma compared with eyes that had a CCT exceeding 588 µm. In the multivariate model of baseline characteristics predictive of conversion oh OHT to glaucoma, the CCT had the greatest im‐ pact on the risk. These findings were confirmed in the European Glaucoma Prevention Study [14].

The CCT can be easily and accurately measured, it remains quite constant over a patient's lifetime, and, thus, just one CCT measurement is adequate in most patients. It is not clear why the CCT is such a strong predictor of the development of primary open-angle glauco‐ ma (POAG) in OHT patients. In a multivariable model including age, baseline GAT IOP, op‐ tic disc topography (cup to disc [c/d] ratio), and visual field (pattern standard deviation [PSD]), although the CCT and IOP have independent effects on the risk of developing POAG, the two factors interact. Nevertheless, because GAT measurements depend on the CCT, it was impossible in the original model to completely disassociate the effects of both. These findings prove that CCT is an independent risk factor for glaucoma development. The CCT artifacts the GAT, so the IOP may be overestimated or underestimated in thick or thin corneas, respectively.

sult could hardly indicate a major true independent contribution of CCT as a prognostic

Cornea and Glaucoma

231

http://dx.doi.org/10.5772/53017

The fundamental concept supporting this correction formula is that as corneas become thin‐ ner, the GAT measurements become too low. If the CCT is an average value, the GAT value is essentially correct, and if the cornea is thicker than average, the GAT overvalues the true manometric IOP. Although the Ehlers formula was based on manometric data, the weakness of the formula arises from the small number of subjects studied and the high degree of vari‐ ability among the subjects. Ehler's data showed a tendency for the Goldmann IOP to in‐ crease with increasing CCTs; however, a close look at that data indicates that many subjects clearly defy that trend, i.e., the Goldmann values were too low in some subjects with thick corneas and too high in some with thin corneas. The correlations between the IOP and CCT with Ehler's data and data from similar studies are too low to allow definitive clinical deci‐ sion-making based on these formulas. However, adjusting the IOP using CCT-based formu‐ las has resulted in poorer agreement with Pascal dynamic contour tonometry, a slit-lamp mounted tonometer for measuring IOP wich seems to be independent of the corneal proper‐ ties (Figure 3), compared with unadjusted GAT IOP values [20]. It suggested that, although the CCT may be useful in population analyses, CCT-based correction formulas should not

CCT correction formulas for GAT measurements are probably of little value in clinical prac‐ tice [21]. It might be advantageous to incorporate the risk information from validated pre‐ dictive models of glaucoma development or progression [19,21], so clinicians have to account for baseline older age, higher IOP, larger vertical c/d ratio, thinner CCT, and greater PSD in the visual field. The hypothesis that CCT is a true independent risk factor for glauco‐

In addition, the CCT is becoming more important clinically because of the large number of pa‐ tients who undergo laser in situ keratomileusis (LASIK), which causes high IOP elevations in‐ traoperatively [22] and a permanent corneal thinning, that, therefore, affects the IOP evaluation.

ma is currently not validated and requires further investigation.

factor of glaucoma development.

be applied to individuals.

**Figure 3.** Dinamic contour tonometer.

In 1975, Ehlers cannulated 29 eyes undergoing cataract surgery and found differences be‐ tween the cannulated IOP and GAT IOP that were related to the CCT [15]; the GAT IOP was most accurate when the CCT was 520 µm. These results indicated that the CCT varies among individuals, and that this variations significantly affect the GAT IOP (Figure 2); therefore, deviations from the 520-µm reference value produced under- and overestimates of 7 mmHg for every 100 µm of deviation.

**Figure 2.** Goldmann applanation tonometer on a slit-lamp.

Investigators have attempted to design nomograms or correction formulas to account for the effect of CCT on GAT-IOP measurement [15-18], but none has been satisfactory.

The use of the available formulas to obtain a CCT-corrected GAT IOP does not improve the accuracy of the models to predict the risk of glaucoma development [19]. The predictive abilities were similar between the original OHTS model that included CCT, and other mod‐ els that did not include the CCT but only the CCT-corrected IOP. This may mean that the CCT is relatively unimportant in the final predictive ability of the multivariable model as long as the CCT-corrected IOP is included. For example, a model including the IOP values corrected by the Ehlers formula [15] (a commonly used CCT correction formula that exclud‐ ed the CCT) had a predictive ability almost identical to the original OHTS model. Such a re‐ sult could hardly indicate a major true independent contribution of CCT as a prognostic factor of glaucoma development.

The fundamental concept supporting this correction formula is that as corneas become thin‐ ner, the GAT measurements become too low. If the CCT is an average value, the GAT value is essentially correct, and if the cornea is thicker than average, the GAT overvalues the true manometric IOP. Although the Ehlers formula was based on manometric data, the weakness of the formula arises from the small number of subjects studied and the high degree of vari‐ ability among the subjects. Ehler's data showed a tendency for the Goldmann IOP to in‐ crease with increasing CCTs; however, a close look at that data indicates that many subjects clearly defy that trend, i.e., the Goldmann values were too low in some subjects with thick corneas and too high in some with thin corneas. The correlations between the IOP and CCT with Ehler's data and data from similar studies are too low to allow definitive clinical deci‐ sion-making based on these formulas. However, adjusting the IOP using CCT-based formu‐ las has resulted in poorer agreement with Pascal dynamic contour tonometry, a slit-lamp mounted tonometer for measuring IOP wich seems to be independent of the corneal proper‐ ties (Figure 3), compared with unadjusted GAT IOP values [20]. It suggested that, although the CCT may be useful in population analyses, CCT-based correction formulas should not be applied to individuals.

**Figure 3.** Dinamic contour tonometer.

The CCT can be easily and accurately measured, it remains quite constant over a patient's lifetime, and, thus, just one CCT measurement is adequate in most patients. It is not clear why the CCT is such a strong predictor of the development of primary open-angle glauco‐ ma (POAG) in OHT patients. In a multivariable model including age, baseline GAT IOP, op‐ tic disc topography (cup to disc [c/d] ratio), and visual field (pattern standard deviation [PSD]), although the CCT and IOP have independent effects on the risk of developing POAG, the two factors interact. Nevertheless, because GAT measurements depend on the CCT, it was impossible in the original model to completely disassociate the effects of both. These findings prove that CCT is an independent risk factor for glaucoma development. The CCT artifacts the GAT, so the IOP may be overestimated or underestimated in thick or thin

In 1975, Ehlers cannulated 29 eyes undergoing cataract surgery and found differences be‐ tween the cannulated IOP and GAT IOP that were related to the CCT [15]; the GAT IOP was most accurate when the CCT was 520 µm. These results indicated that the CCT varies among individuals, and that this variations significantly affect the GAT IOP (Figure 2); therefore, deviations from the 520-µm reference value produced under- and overestimates

Investigators have attempted to design nomograms or correction formulas to account for the

The use of the available formulas to obtain a CCT-corrected GAT IOP does not improve the accuracy of the models to predict the risk of glaucoma development [19]. The predictive abilities were similar between the original OHTS model that included CCT, and other mod‐ els that did not include the CCT but only the CCT-corrected IOP. This may mean that the CCT is relatively unimportant in the final predictive ability of the multivariable model as long as the CCT-corrected IOP is included. For example, a model including the IOP values corrected by the Ehlers formula [15] (a commonly used CCT correction formula that exclud‐ ed the CCT) had a predictive ability almost identical to the original OHTS model. Such a re‐

effect of CCT on GAT-IOP measurement [15-18], but none has been satisfactory.

corneas, respectively.

230 Glaucoma - Basic and Clinical Aspects

of 7 mmHg for every 100 µm of deviation.

**Figure 2.** Goldmann applanation tonometer on a slit-lamp.

CCT correction formulas for GAT measurements are probably of little value in clinical prac‐ tice [21]. It might be advantageous to incorporate the risk information from validated pre‐ dictive models of glaucoma development or progression [19,21], so clinicians have to account for baseline older age, higher IOP, larger vertical c/d ratio, thinner CCT, and greater PSD in the visual field. The hypothesis that CCT is a true independent risk factor for glauco‐ ma is currently not validated and requires further investigation.

In addition, the CCT is becoming more important clinically because of the large number of pa‐ tients who undergo laser in situ keratomileusis (LASIK), which causes high IOP elevations in‐ traoperatively [22] and a permanent corneal thinning, that, therefore, affects the IOP evaluation. Because the IOP is an important risk factor for glaucoma, accurate measurement is important, and it can be achieved by intraocular manometry; however, this is an invasive method that obvi‐ ously cannot be used in a clinical setting,

have been a less powerful predictor. Some investigators interpreted the OHTS results to in‐ dicate that the CCT is an independent risk factor for glaucoma development. Because the GAT measurements ultimately depend on the CCT, Medeiros and Weinreb [26] stated that it is impossible, based on the original model, to disassociate the effects of both. Some groups have evaluated [19,27,28] whether the OHTS prediction model could be improved using CCT-corrected IOP using previously published formulas (Table 1), evaluated using the c sta‐ tistics (a measure of concordance), and calibration chi-squares. The c statistic is the fraction of patients with an outcome among pairs of patients, in which one has the outcome and one does not; the patient with the higher predictive value is classified as the one with the out‐ come. The c statistic varies between 0.5 when a model provides no information and 1.0 in sensible models. The CCT also remained a significant predictor of glaucoma development in

Cornea and Glaucoma

233

http://dx.doi.org/10.5772/53017

a multivariable model that included the CCT-corrected IOP.

**CCT in microns IOP correction in mm Hg**

**Table 1.** Correction values for IOPs based on CCT [ 8,17].

The only way to fully evaluate the possible independent role of CCT as a prognostic factor for glaucoma development is to include in the predictive model the IOP measurements obtained by a CCT-independent tonometer. The Pascal dynamic contour tonometer (DCT), is a slit-lampmounted, nonapplanation, digital contact tonometer that provides continuous tonometry re‐ cordings that measure the IOP and the ocular pulse amplitude, which is the difference between the minimal and maximal values of the pulsatile IOP wave contour, and does not require cor‐ neal applanation and the DCT IOP measurements seem to agree closely with manometric meas‐ urements [23]. Therefore, including DCT measurements with the CCT in a predictive model for glaucoma might better assess the true independent value of CCT compared with use of only the CCT-corrected GAT values. This has been investigated in patients undergoing phacoemulsifi‐ cation, that had the anterior chamber cannulated in a closed system and the IOP was set to 15, 20, and 35 mmHg by a water column. The IOP measurements then were taken by DCT. The results showed that the DCT agree well with the intracameral IOP. Interestingly, the CCT had a low but significant effect on the DCT measurements [23].

The DCT measurement principle is based on contour matching, which assumes that if the eye were enclosed by a contoured, tight-fitting shell, the forces generated by IOP would act on the shell wall. Replacing part of the shell wall with a pressure sensor would enable measurement of these forces and therefore the IOP. The DCT has a central gauge surrounded by a contoured plastic tip that is in contact with the cornea and creates a tight-fitting shell. The DCT compen‐ sates for all forces exerted on the cornea and an electronic sensor measures IOP independent of the corneal properties.
