Continuing Education

Defining normal corneal thickness and normal IOP measures

Professor Michael Doughty and Sven Jonuscheit discuss ascertaining whether a cornea is 'thin' or 'thick' as part of routine screening for and management of patients with open-angle glaucoma. CET module C1805

THIS MODULE HAS NOW CLOSED.

View PDF

THIS ARTICLE IS BEST READ ON A PDF.

OVERALL PERSPECTIVE

Figure 1. Slit-lamp biomicroscope with Haag-Streit pachymeter

Research over the last decade has confirmed observations made in the 1970s that corneal thickness can have an impact on the clinical measurement of intraocular pressure (IOP) by either applanation tonometry (AT) or non-contact tonometry (NCT). As a result, it is becoming increasingly commonplace for glaucoma screening to not only involve tonometry, ophthalmoscopy and visual fields, but a measurement of central corneal thickness (CCT) as well. The European Glaucoma Society recently concluded that 'CCT measurements are unavoidable for the correct management of OHT' (ocular hypertension).¹

The main issue nowadays is therefore not whether pachymetry (CCT measurements) should be considered as part of the management of open-angle glaucoma patients, but what to do with these measurements. There are clearly different options depending on one's perspective on what the differences in corneal thickness might mean. For a detailed coverage of many aspects of this topic, the reader is encouraged to consult the review article written by the author.² This article will summarise some of the pertinent findings made in that review and update selective aspects, some of which are based on continuing studies being carried out by the first author's research team at Glasgow-Caledonian University, which includes the second author.


MEASURING CORNEAL THICKNESS
The human cornea is made up of five essential layers, namely the multilayered epithelium, the anterior limiting lamina (Bowman's membrane), the corneal stroma (with its keratocytes), the posterior limiting lamina (Descemet's membrane) and the single layer of corneal endothelium.³ All of these make a contribution to the total corneal thickness. In clinical practice, it is the thickness at the centre of the cornea, close to the optical axis of the eye, that is usually measured. The technique used is some form of pachymetry (also known as pachometry) and a measurement is made of central corneal thickness, now usually referred to as the CCT value. CCT can be measured in several ways.

Optical pachymetry, based on slit-lamp techniques, requires that the observer judge the location of light reflected off each surface of the cornea in contrast to the tear film and aqueous humour. The location of one reflex is ascertained and then the other, and a sliding manual scale on the slit lamp used by the observer to note the difference in location between the two reflexes. It was introduced into contemporary practice in the 1960s as an attachment to a Haag-Streit slit lamp (Figure 1). The sliding scale on the instrument makes it likely that one can only estimate CCT to within 0.01mm. A difficulty with the technique, necessarily carried out at high magnification, is to reproducibly take measures at the central region of the cornea. While later refined by the addition of a set of fixation lights to assist in alignment of the slit lamp beam to the central region of the cornea, it nonetheless is a subjective technique and prone to error from both alignment problems and also as to whether the specular reflex is uniform.

Figure 2. Ultrasound pachymeter probe, being applied to central region of cornea

The 1970s saw the introduction of the ultrasound-based pachymeter in which the location of the two corneal surfaces is automatically detected from the position of reflection of ultrasound beams. The most likely location of the measures are in the centre of the pupillary region, which should be fairly easy to judge by simple visual inspection. The pachymeter probes are hand-held (Figure 2) and numerous portable options are available such that pachymetry can be performed in any type of clinic.

Through the late 1980s and to this day, ultrasound pachymetry is likely the most widely used instrument for measurement of CCT. The technique is objective, although the reproducibility may be affected by either the alignment of the probe and/or the pressure exerted on the corneal surface by the probe (under topical anaesthesia). Depending on the electronic set up and tuning of the instrument, it is likely that the measures can routinely made to a resolution of better than 0.01mm, but that the absolute values of the CCT are likely to be slightly higher (2 to 3 per cent) than with a slit-lamp based method.

In the 1990s, an alternative to slit-lamp based optical pachymetry became widely available with the introduction of the clinical specular microscope. The instruments can be based on a non-contact principle or require that the objective lens be in contact with the corneal surface (under topical anaesthesia). The former method is the one that appears to be most widely used, but all instruments are somewhat bulky and, like a conventional slit lamp, can hardly be regarded as portable.

Figure 3. Specular microscope print out to show endothelial image and corneal thickness value

The specular microscopes were essentially designed to allow the ophthalmologist to assess the appearance of the corneal endothelium both before and after intraocular surgery (eg cataract operations). The various instruments, unlike a slit lamp, provide a very high magnification view of the endothelial surface by specular reflection. However, similar to a slit lamp, the specular microscope operating principle is based on the internal prism-based reflection of light from the tear film and aqueous interfaces but the assessment of the overall difference between the two sets of reflections is usually done automatically, and the instrument provides a print out of the endothelial image and the CCT value. This is in millimetres and designated as 'T' in the example shown

(Figure 3). Likely as a result of the automation, the measures can likely be made to better than 0.01mm and the absolute values of CCT generally appear to be between those measured with a slit lamp and by ultrasound pachymetry.

Since the later 1990s, a number of alternatives to the ultrasound pachymeter have been developed of which the laser scanning slit method (also known by its proprietary name, the Orbscan) is perhaps the best known. The approach used in this technique is non-contact and has similarities to slit-lamp based pachymetry except that the process of identifying the location of reflection of light off the two surfaces of the cornea is repeated many times. To allow this to be done, the subject (patient) needs to maintain a stable position for a few seconds. This is achieved with subject using a chin and forehead rest (similar to a slit lamp) as a set of fine laser beams are automatically moved across the cornea surface and the locations of the reflections automatically recorded by the detector system. As a result, a 3-D map is built up of the locations of the two surfaces and the CCT value is obtained, by subtraction, using a customised computer software programme.

FIgure 4. Orbscan pachymeter map with zones for averaged thickness readings

The instrument provides numerous print-out options, but the most relevant one here is the pachymetry map (Figure 4). In the example shown, which appears to be the most common output used, the averaged thickness value, within a central circle of 2mm diameter, is reported, ie as 561µm in this example. The other circles represent thickness values at paracentral (mid-peripheral) regions. The instrument is bulky, is not portable and at this time is likely found in refractive surgery clinics rather than glaucoma clinics. As with specular microscopy, the automation of the technique means that it should be routinely possible to take measures on a normal cornea to within 0.01mm, but that the absolute measures of CCT are likely to be between 5 and 10 per cent higher than measures obtained with other instruments mentioned above.

In conclusion, the ease of use, and cost considerations, mean that ultrasound pachometry is likely to remain the mainstay of corneal thickness measures for the foreseeable future.

THE CORNEA THICKNESS, ITS RELEVANCE TO GLAUCOMA AND THE NEED FOR PACHYMETRY
From the side (slit lamp) optical section, the cornea not only appears transparent (Figure 5a) but also incredibly thin. In routine external evaluation of the cornea in section, including as part of managing glaucoma patients, it is likely acceptable to simply comment on whether there was any disruption of the uniformity of the transparency, eg by regions of haze or the presence of an infiltrate or ulcer etc. The situation is expected to be rather different for an angle-closure glaucoma (acute closed angle glaucoma) where a hazy appearance of the cornea (especially the epithelium) could well be part of the needed signs for a diagnosis of this condition along with the very high IOP values and patient symptoms (Figure 5b). This assessment does not need a slit-lamp, but its use would likely be useful to see changes in the corneal endothelium associated with the substantially elevated IOP.

However, if a specific consideration of corneal thickness is to be part of the assessment of glaucoma patients, an accurate means of measuring its 'thinness' is needed because even at 40X objective  it is still hard to visually judge what the actual thickness might be. For example, is it

(a)

(b)

Figure 5. External appearance of a normal eye (a) and one with angle-closure glaucoma (b) (Figure b courtesy of Clinical Ophthalomology. A clinical systematic approach (1999) by J Kankski)

0.475mm, 0.525 or 0.575mm? The point here is that an eye care practitioner does not need a pachymeter to tell him or her that a cornea is thickened in angle-closure glaucoma. The combination of the distinct oedematous appearance of the corneal epithelium, the loss of iris detail, and accompanying conjunctival congestion surely cannot be missed (Figure 5b). However, the practitioner does need to be able to record what must be considered as subtle differences in corneal thickness that are not accompanied by any obvious changes in the overall appearance of the external eye or the cornea in optical section.

This ability to detect subtle differences in corneal thickness is especially relevant to screening for various forms of open angle glaucoma and, in very recent times, also to be considered as part of the routine management of some of these patients. In both cases, the need for pachymetry is driven by a quest to provide optimum management for patients. The interest in pachymetry has developed in three phases, namely in the assessment of ocular hypertensive (OHT) patients, normal tension glaucoma (NTG) patients and, most recently, as part of general interest in the outcomes of medical treatment of open-angle glaucoma. These might be considered as separate issues, but are all largely coming together by current perspectives. In simple terms, the OHT patients (sometimes referred to as glaucoma suspects) are those with measured IOPs that are routinely in the mid 20s but who show no other glaucomatous signs in routine assessments (ophthalmoscopy and screening visual fields).

The NTG patients are those who routinely present with measured pressures in the normal range but who have glaucomatous changes. In considering the outcome of medical treatment, the goals have traditionally been to lower and keep the measured IOP values within a so-called normal range (ie 10 to 20mm Hg).1 If the IOPs were 'very high', as opposed to 'slightly elevated', then much more aggressive medical treatment would likely be instigated, with the target pressures being closer to 10 rather than 20mm Hg. Such aggressive treatment might involve two or more sets of eyedrops being recommended, sometimes several times a day. If the pressures were 'normal' or close to normal, the ophthalmologist was faced with the dilemma of whether there was going to be any benefit of subjecting the patients to the daily routine of instilling various anti-glaucoma medications.

An important point here, seemingly all too often overlooked, is that if the patient has 'normal' pressures then it is less likely that they will be subjected to intensive IOP-lowering therapies. As a result, the rate or extent of disease progression may be substantial (especially if visual fields were not performed as part of a routine eye examination) before the glaucomatous damage was recognised, and then decisions made to try to slow the progression of the disease by medical (or surgical) intervention.

CORNEA OEDEMA, THICKNESS, AND TONOMETRY
In the original reports on the Goldmann applanation tonometer,⁴ the authors accepted that the instrument could only be expected to work on 'normal' corneas, but there was no specific correction made for corneal thickness per se. They made an assumption as to what 'normal' corneal thickness was and, according to which accounts are to be believed, this could have been either 0.500 or 0.520mm. This issue has clearly caused considerable confusion and was part of the reason why the first author undertook the meta-analysis (see next section). However, in simple terms, Goldmann and Schmidt noted that if there was substantial compromise of the epithelial surface, perhaps with some obvious epithelial oedema, then adequate alignment of the tonometer mires (with fluorescein) would not be achieved and the IOP readings would be unpredictable and unreliable. Those wishing to defend the routine use of NCT in optometric practice might well argue that its utility is because its effective use does not require the alignment of a fluorescein pattern, although it has to be noted that the European Glaucoma Society does not recommend NCT for evaluating patients with glaucoma.¹

Figure 6. Distribution of measure IOP values, by applanation tonometry, in a population sample. (Modified from Graham and Hallows, (Trans Ophthalmol Soc UK, 1964; 84:597-613)

However, it also has to be noted that further specific work from the same period,5 indicated that this unreliability in applanation tonometry was perhaps predictable in that lower IOP readings would likely be obtained on corneas with epithelial compromise. While the data available was limited, it can be said that these corneas were not 'thin' but rather that it was easier to applanate a 'soggy' (oedematous) epithelium and so get lower IOP readings. The current irony is that it is this issue of 'thin' corneas that underpins the latest banner headlines on the importance of corneal thickness in open-angle glaucoma.

However, it was the opposite effect that started a rather circuitous journey towards the contemporary interest in corneal thickness and routine screening for 'glaucoma'. In the late 1960s and early 1970s, some reports were published from both Scandinavia and Japan that noted that the measured IOP value appeared to depend on CCT (as measured by optical pachymetry). It is very important to note that these were tonometry and pachymetry measures on intact eyes. A series of studies over the following decade clearly indicated that this was not an irrelevant and irreproducible finding, but that there might be a predictable relationship between the central corneal thickness measures and the clinically measured IOP by conventional tonometry. We are now in a position to state with considerable certainty that routine clinical tonometry measures can be affected by the central corneal thickness, as based on clinical pachymetry measures.

The first step - an objective analysis of what 'normal' central corneal thickness should be as part of the meta analysis of Doughty & Zaman

Figure 7. Examples of the distribution of corneal thickness values obtained by pachymetry from two different eye clinics

A series of rather well profiled articles and editorials in the mid-to-late 1990s raised an alarm on the accuracy or reliability (or both) of tonometry in routine practice. It applies just as much to NCT as it does to different forms of applanation tonometry. The issue that was of most substantial importance at that time was the possible misdiagnosis of glaucoma as based on tonometry data. It was concluded that patients with thicker than 'normal' corneas would be more likely to have higher IOP readings, and it really did not seem to matter what the method of tonometry was, ie it was not an artifact that could be unambiguously assigned to a particular instrument or instrument type. Stated another way, if a patient had a so-called 'thick' cornea, they might have IOP readings that were routinely over 21mm Hg but, it was argued, they didn't really have that 'true' IOP value because the tonometer gave a reading that was too high. This particular issue will be looked at shortly, but the first point that needs to be addressed is how might one define 'normal' corneal thickness (and so then decide if a cornea was 'thinner' or 'thicker' than normal). There are clearly different opinions, and an attempt will be made to address all of these. One of them, largely adopted by the author, follows the same principles of deciding whether a patient's IOP is too high or too low, and is based on a statistical evaluation of a 'normal' distribution.

Figure 8. Distribution of corneal thickness values in the worldwide population as developed with the meta analysis approach including all types of pachymetry. (Modified from Doughty and Zaman, Surv Ophthalmol, 2000; 44: 367-408)

So, as just mentioned, one approach to deciding what 'normal' CCT might be is based on the expected distribution of measured values from a sample of patients. This same idea forms the basis of our decisions on what a normal IOP is.

Figure 6 shows an example of the distribution of IOP measures in a population sample. This is one representative example from UK-based studies first reported in 1964 by Graham and Hollows.6 It is one of several major studies of tonometry outcomes that were made many years ago at a time when glaucoma was often considered as synonymous with elevated IOP.

Other notable studies were conducted in Germany, ⁷ and the USA,^8-10 and all produced similar conclusions. Several points can be taken from the graph shown in Figure 6. Firstly, some IOP values are observed a lot more frequently than others and so, not surprisingly, the average IOP value is around 16mm Hg. What is just as important however is that there are very few values below 11mm Hg and only relatively few above 20mm Hg. So, this distribution allows us to draw the general conclusion that normal IOPs should generally be between 11 and 20mm Hg.

Based on this type of distribution of IOP measures from applanation tonometry, some further statistical analysis allowed investigators in this period to come to the decision that an abnormal IOP would be 21mm Hg, and we need perhaps to remind ourselves of just how this came about. From distributions of this type, it is possible to calculate a standard deviation (SD) and, in the example shown, it was 2.86mm Hg, about the mean value of 15.9mm Hg. An almost identical result was obtained from a study in Ohio using Schiotz tonometry,9 where the mean IOP was 15.4mm Hg and the standard deviation was 2.65mm Hg. If one calculates the IOP at two SDs above the arithmetic mean then, as specifically discussed by Johnson,9 one arrives at an IOP value of just below 21mm Hg, ie (15.4 + 5.3mm Hg) = 20.7mm Hg. One can go further and calculate a 95 per cent confidence interval (95 per cent CI) to define the range of 'normal' IOP. This might be done on ±2 SD, or more usually, on ±1.96 SD, so this would be 15.4 ±5.2mm Hg for a 95 per cent CI of between 10.2 and 20.6mm Hg. For the Graham and Hollows study,6 the result would be 10.3 to 21.5mm Hg, and is the definition of so-called cut-off values for 'high' (= 21mm Hg) and 'low' (= 10mm Hg).

One of the key aspects of these sorts of studies is that a special clinical sample was not used, but a very large number of individuals from the population at large. However, it is not always possible to set up such studies, and so a different approach with a slightly different outcome can be used. This is the meta-analysis approach.

One principle behind a meta-analysis is to use as many different sources of information to arrive at a result. There are many subtle variances on this approach, but the one used was openly acknowledged as all but all-inclusive and unweighted.² The idea is as follows. If a clinic routinely measured CCT, then, over a period of time, it would end up having a range of different values on the basis that all individuals are not the same.

The actual values could be somewhat unpredictable on occasions but, over time, a predictable distribution of values could be obtained from a single clinic. So, as with the tonometry measures, the average CCT value might be the same between two clinics but the actual fine detail of these distributions of CCT data could still be slightly different (Figure 7). They could be narrow or slightly broader and, as with the distributions of IOP measures, the sets of CCT measures might not perfectly conform to a normal (Gaussian) distribution. Notwithstanding, with reason, such distributions of CCT data can also be compared by calculation of the standard deviation (SD) and a 95 per cent confidence interval to estimate the expected values for corneal thickness in normal eyes.

However, and this is the important point underpinning the meta-analysis, different clinics might come up with a set of normally distributed data, but their actual average value could be slightly different to other clinics. The reasons for this might be that their instruments were calibrated slightly differently, or that they used slightly different instruments (see earlier). Equally importantly, the different clinics could actually be taking measures from different cross-sectional samples of the population whose CCT values might actually be slightly different. These individual differences might be gender, age, refractive error, ethnic origin or, perhaps most important from the perspective of this article, that some had normal healthy eyes and others had long-standing disease such as open-angle glaucoma. So, the first step of the meta-analysis was to try to find every article wherein a clinic had reported CCT data on normal healthy eyes, ie they were not considered to have glaucoma, they were not contact lens wearers nor had they any systemic diseases (eg diabetes) that might affect the cornea, neither had they had ocular surgery. The project had been under way for some years by the author, but he was spurred into completing the task by the interest from a new graduate student, Zaman (whose full name is actually Mohammad Laiquzzaman). With the added assistance, it took about a year to assemble all the data (Figure 8) in which each of some 300 average CCT values published from different clinics around the world were included. Now, as a first approximation, one has a population distribution that should have, by random chance, included all those measures with slightly different instruments and, in some cases, on slightly different people. From this population-based study, the population mean value for CCT was estimated to be very close to 0.535mm.

In the meta-analysis of pachymetry data, as was done for analyses of tonometry data from population studies, the distribution of all the average values from different clinics was used to calculate a population-based SD (at ±0.031mm) and a 95 per cent CI for the worldwide population for the CCT was estimated to be between 0.473 and 0.597mm. From this, one can logically make a decision as to whether the measured CCT of an individual, at any time and place, is within 'normal' limits. One can also therefore logically define a 'thin' cornea (²0.475mm) and a 'thick' cornea (³0.600mm). It needs perhaps to be stressed that this approach conforms to that used for analysis and interpretation of IOP (tonometry) data. The 95 per cent CI, as a first approximation, assists one in deciding the expected range for 'normal' as based on the SD which is a measure of the shape of the distribution.

At the present time, there is only slow movement towards accepting and applying the outcome of this meta-analysis for normal CCT, but there are several applications. For some of the large scale glaucoma treatment studies, patients have been excluded if their CCT was less than 0.475mm, ie they needed a different treatment strategy. Prior to refractive surgery, this lower limit of CCT has also been considered as a suitable criterion for excluding patients from undergoing Lasik for example, ie their corneas were too thin for safe surgical intervention.

In the next article, an examination will be undertaken of how this knowledge about CCT can now be applied, and specifically to the evaluation of patients with abnormal IOPs. 
 

References
1 European Glaucoma Society terminology and Guidelines for Glaucoma. EGS, Savona, Italy, 11th Edn, 2003 (www.eugs.org).

2 Doughty MJ, Zaman ML. Surv Ophthalmol, 2000; 44: 367-408.

3 Bergmanson JPG, Doughty MJ. Anatomy, morphology and electron microscopy of the cornea and conjunctiva. In - Clinical Contact Lens Practice (Bennett E, Weissman BA, eds). Philadelphia, Lippincott-Williams & Wilkins. 2005; Chapter 2, pp 11-39.

4 Goldman VH, Schmidt T. Ophthalmologica, 1957; 134: 221-242.

5 Ytteberg J, Dohlman CH. Arch Ophthalmol, 1965; 74: 477-484.

6 Graham P, Hollows FC Trans Ophthalmol Soc UK 1964; 84: 597-613.

7 Leydhecker W, Akiyama K, Neumann HG. Klin Monatsbl Augenheilkd, 1958;133: 662-670.

8 Armaly MF. Invest Ophthalmol 1962; 1: 618-628.

9 Johnson LV. Am J Ophthalmol, 1965 59: 680-689.

10 Colton T, Ederer F. Surv Ophthalmol, 1980 25: 123-129.

Michael Doughty is a professor at Glasgow Caledonina University where Sven Jonuscheit, Augenoptikermeister, is an optometrist and PhD student

SUMMARY

Routine tonometry is likely affected by central corneal thickness, as based on clinical pachymetry measures

Patients with thick corneas may also be misdiagnosed as being a 'glaucoma suspect' or having ocular hypertension (OHT) because the routine tonometry overestimates the true IOP value

Patients with thin corneas may be misdiagnosed (as being normal) because the clinically measured IOP values (by Goldmanm or NCT) are actually underestimated. These are patents who may progress to normal tension glaucoma (NTG)

 

Providing exclusive eye care news, information and educational needs every week, including a FREE CET programme. Subscribe to Optician Print Edition.