Continuing Education
Pachymetry - Part 2
The possibility of correcting tonometry measures for corneal thickness
|
Sponsored by |
Professor Michael Doughty and Donald Montgomery consider the application of pachymetry to routine measures of IOP in screening for and management of patients with open-angle glaucoma, and discuss some of the current ideas on further applications of corneal thickness measures (C1806)
THIS MODULE HAS NOW CLOSED.
|
|
THIS ARTICLE IS BEST READ ON A PDF.
APPLICATION OF META-ANALYSIS FOR NORMAL CORNEAL THICKNESS AND RELEVANCE TO TONOMETRY READINGS
In the previous article (OPTICIAN, August 5, p27), the basis of a meta-analysis for defining normal corneal thickness was given.1 A distribution of reported values of central corneal thickness (CCT) was obtained (Figure 1).
From a practical perspective, based on this meta-analysis, there are several options. In what now might be regarded as a too extreme application (but the one intended by the first author), if an individual's CCT was less than 0.473mm, they would have a 'thin' cornea.
|
This module has |
Conversely, if they had a CCT of above 0.597mm, they would have a 'thick' cornea (see arrow at upper end of histogram in Figure 1).
As noted previously,1,2 if it is considered that ultrasound pachymetry is currently the most likely method used to assess corneal thickness, then the same data generally applies, since a wide range of average CCT values have also been reported for ultrasound pachymetry.
The overall mean value and upper limit of the 95 per cent confidence interval (CI) is only slightly different from that obtained using all pachymetry methods, namely 0.544mm (rather than 0.535mm) and 0.611mm (rather than 0.597mm).2
As discussed previously,1 several hand-held ultrasound pachymeters are available and an example of the display on one of them is shown in Figure 2. The corneal thickness reads 664µm, or 0.664mm.
Recent reports of CCT in patients seen at glaucoma clinics have reported average values, by ultrasound pachymetry, of between 0.545mm and 0.569mm.3-7
These results indicate that CCT may be different in glaucoma patients, but also further illustrate the point made previously,1 that different clinics will arrive at different average CCT values.
Regardless of the actual average value in a study, an individual with a CCT value of 0.664mm should be given special consideration.
| FIGURE 1. Distribution of central corneal thickness values in the worldwide population as developed with a metaanalysis approach including all types of pachymetry (Modified from Doughty and Zaman, Surv Ophthalmol, 2000; 44: 367-408). The upper limit of ‘normal’ thickness is shown by the arrow |
From the perspective of patients with glaucoma (or suspected glaucoma), the application of the CCT data to the patient management should be simple.
If their corneas were indeed thicker (for example, 0.664mm), then it should be considered that their IOPs were likely overestimated by either applanation contact or non-contact tonometry.8,9
So, it might well be argued, either that their target pressures under medical therapy need not be so aggressively sought after. Stated another way, perhaps, for some ocular hypertension individuals, they really do not need medical therapy (at least for the time being), but still need to be monitored on a regular basis.
| FIGURE 2. The digital display of a handheld ultrasound pachymeter |
The 'true' IOP of these patients could have been rather higher. If the thin cornea had been identified in screening, then a case can be made that such patients should have received closer attention, and even have been placed on more aggressive medical therapy with the target pressures being set very much on the low side (that is, closer to 10mmHg, rather than perhaps 13mmHg).
The point here that must be stressed is that the application of the meta-analysis for corneal thickness follows the same principle as that used for the interpretation of tonometry data.
It appears that a slightly more conservative approach is also being considered.10 By this, rather than the 95 per cent CI being used, patients could be considered as having a thicker cornea if their value was more than 1 SD above the population mean (in other words a CCT of greater than or equal to 0.567mm) or that they would have a thinner cornea if their CCT value was less than 1 SD below the mean (that is, greater than or equal to 0.505mm).
This conclusion has been reached using the data from the meta-analysis.2 This rather narrow range of CCT, within the expected distribution of values, is shown in Figure 3.
| FIGURE 3. An alternative for defining normal corneal thickness in which values above or below one standard deviation around the population mean value (±1SD) are considered abnormal |
For example, if it were accepted that the population mean IOP ±1SD was 16±3mm, it is not generally held that patients have ocular hypertension if their IOPs are 19mmHg.
A third approach to the application of CCT measures has come into play in very recent times, namely the use of the mean or median value for CCT as the basis for deciding if an individual's corneas are too thin or too thick.
For example, if the population mean CCT was taken to be 0.535mm, then an individual with a CCT of 0.565mm would be regarded as having a thick cornea and an individual with a corneal thickness of 0.505mm would be categorised as having a thin cornea. Such categorisations, or even more complex ones, have been applied in a range of recent studies (see later).
Again, however, the issue that has to be considered is whether such categorisation of patients by their CCT value is consistent with tonometry. For example, if the median value for the measured IOP was taken as 16mmHg, a patient is not automatically considered to be a glaucoma suspect or as having OHT if their IOPs were consistently recorded as being 17 or 18mmHg.
A last issue here, that has yet to be reasonably addressed, relates to what instruments should be used. In line with modern day jargon, it is likely that many will consider Goldmann tonometry as the 'gold standard' in glaucoma clinics.
While ultrasound methods may well be considered as the equivalent gold standard for pachymetry and is obviously easily used before or after Goldmann tonometry (under topical anaesthesia), it remains to be established whether other pachymetry methods could also be acceptable in a glaucoma clinic, (such as an Orbscan).1
THE SECOND STEP
An objective analysis of what the possible impact of corneal thickness might be on clinical measures of IOP as part of the meta-analysis of Doughty & Zaman
This is a different issue to the one outlined above, and brings into play a consideration of whether one can routinely correct for the errors, or even whether there should be a correction (as opposed to simply taking the corneal thickness effect into account).
| FIGURE 4. Assessment of the impact of corneal thickness on the measurement of intraocular pressure in normal human eyes. Each data point represents an average value from a separate published study. The line is that for an unweighted linear regression analysis. (Modified from Doughty and Zaman, Surv Ophthalmol, 2000; 44:367-408) |
Just as reports from different clinics could yield different average CCT values so too, for those that were undertaking studies on glaucoma patients, the average IOP values also differed.
Some of the studies have noted average IOP values that were higher and it could also be noted that, for their particular group of patients, the CCT values were on the higher side.
The opposite was also noted (namely, lower IOP values associated with lower CCT values), but this was not always the case and the proportional difference between the average IOP and CCT values was definitely not the same.
The logic of the meta-analysis approach was that if there was a clinically-relevant effect of CCT on the IOP readings, then this would be manifest by pooling all the data from different studies together, so generating an IOP-CCT inter-relationship.
In all, some 70 studies were found with both sets of data.2 From the outset, it was clear that there was considerable heterogeneity, or lack of agreement between studies.
However, interesting trends were noted, and it was also found that some of the heterogeneity could essentially be compensated for by putting the outcomes of the different studies into different groups.
The two groups of interest here are those designated as having 'normal' eyes and those with chronic disease (including various forms of open-angle glaucoma).
The analyses indicated that the IOP-CCT inter-relationship was different between the two groups.
| FIGURE 5. Assessment of the impact of corneal thickness on the measurement of intraocular pressure in human eyes designated as having some form of chronic disease, mainly some form of glaucoma. Each data point represents an average value from a separate published study. The line is that for an unweighted linear regression analysis. (Modified from Doughty and Zaman, Surv Ophthalmol, 2000; 44:367-408) |
The effect is statistically significant, but the slope is slight. A linear regression analysis (shown by the line) indicates that a 10 per cent difference in the average CCT between studies (groups of individuals) could result in a 1.1mmHg difference in the average IOP value.
Stated a slightly different way, a 0.053mm difference in CCT would be expected to produce a 1.1mm shift in the measured IOP value.
Other ways of stating this are that a 1.1mmHg 'error' in tonometry outcome could arise from a 0.053mm difference in CCT, or that a 2mmHg error in tonometry could arise from a 0.100mm difference in CCT. This is for normal individuals, and presumably this would include many patients presenting for routine eye examinations.
However, for patients with long-standing ocular disease such as open-angle glaucoma, the predicted IOP-CCT inter-relationship was slightly different (Figure 5).2
Here, the relationship was again statistically significant, but the slope was steeper. For a 10 per cent difference in CCT, the relationship predicts that a 2.5 ± 1.1mmHg difference, on average, would occur.
Stated another way, a 2.5mmHg error in tonometry could arise for a 0.053mm difference in CCT, or that a 5mmHg error in tonometry could arise for a 0.100mm difference in CCT.
Going beyond the meta-analysis of Doughty & Zaman, 2000 - can a correction factor for tonometry be applied in routine clinical practice?
The overall outcome of the above meta-analyses have been confirmed in quite a number of subsequent individual studies. Collectively, these indicate that - in a routine clinical setting - it should be accepted that the measurement of IOP can be influenced by a person's corneal thickness.
So, can this effect be compensated for by applying some form of correction factor to the data, or even to the tonometer output?
Such an exercise is currently problematic. Unquestionably, a case can be made that CCT needs to be considered on a case-by-case basis.
If the measurements of a patient's IOPs consistently yielded moderately high values (such as 25mmHg) but there are no accompanying glaucomatous signs or vision changes, then the subsequent finding that the CCTs are greater than 'normal' values (for example, 0.664mm), may reassure the ophthalmologist (or optometrist in shared care) that true ocular hypertension may well not be present, and that further follow-up need only be on an infrequent basis.
Questions that need to be considered are whether any medication strategy, or changes in medication strategy, are needed.
| FIGURE 6. The digital display of one hand-held ultrasound pachymeter to illustrate how the user is provided with a corneal thickness adjusted IOP, from 17mmHg to 12mmHg, based on the just-measured central corneal thickness reading of 0.664mm (Courtesy of Keeler) |
On the other hand, on an individual basis, if a patient presented with a glaucomatous disc and visual field changes, yet their IOP readings have been consistently within a normal range, the finding that they have 'thin' corneas (for example, 0.505mm) could well prompt the ophthalmologist to instigate a relatively aggressive medical treatment in order to pursue a particularly low target IOP for that patient.
Overall, however, beyond individual cases, the application of a global correction factor to all tonometry measures is problematic.
This is simply because the IOP-CCT relationship does not appear to be the same for different groups of individuals and this must surely be taken into account.
Questions need to be asked as to whether one is screening for glaucoma (based on IOP measures) or starting or maintaining a medical treatment of open-angle glaucoma.
Let us consider the two scenarios. In routine practice, it is likely that tonometry is performed as a general indicator of ocular health. The majority of patients will not have ocular disease but the practitioner is concerned that their tonometry readings are not corrected for the possible corneal thickness-related artefact.
While the IOP-CCT inter-relationship may well vary between study groups (namely, from 0.2 to 2.9mmHg for a 10 per cent difference in CCT), the overall effect in many 'normal' individuals is expected to be small.8,9
For young and middle-aged adults without a prior history of ocular disease, the average predicted effect from the meta-analysis is just 1.1 ± 0.6mmHg for a 10 per cent difference in CCT (Figure 4). This will likely fall well within the actual range of reproducibility of the tonometry measures themselves.8,9
One has to question, therefore, whether a correction factor can or should be logically applied in routine screening.
|
FIGURE 7. European Glaucoma Society Terminology and Guidelines for Glaucoma. An essential read! |
This would again be a case-by-case selective review of patients. Similarly, if the results of the meta-analysis and subsequent studies are to be accepted, then the slope of the IOP-CCT relationship is also greater in individuals with a history of ocular disease, including open-angle glaucoma (Figure 5).
The reason for this apparent difference in the IOP-CCT relationship between 'normal' and 'diseased' eyes is really unknown, but it may be simply because POAG patients are generally older. In other words, ageing, per se, changes the IOP-CCT relationship.
It has recently been suggested that this is an expected effect that results from an increase in corneal stiffness (and thus resistance to applanation) in older patients.11
At least in glaucoma patients, therefore, the correction factor that could logically be applied is much greater than for 'normal' individuals. It is around 2.5mmHg for a 10 per cent difference in CCT, or about 2.5mmHg for a 0.053mm difference in CCT, or 5mmHg for a 0.100mm difference in CCT.2
A correction factor of 2.5mmHg for a about a 0.05mm difference in CCT has been applied in some recent studies,4,5,7 and it is this correction factor that has been applied in at least one commercial pachymeter (Figure 6).
A software chip in this particular ultrasound pachymeter allows the user to consider the CCT value (in this case 0.664mm), the patient's measured IOP (in this case 17mmHg) and then to calculate an adjusted IOP value.
This 'T' value is shown to be 12mmHg, that is since the patient had a CCT that was just over 0.100mm higher than the predicted population mean, their measured IOP of 17mmHg could be corrected downwards by 5mm.
While this is an example of where pachymetry correction of tonometry is recommended, it is just as important to note the other end of the spectrum. If a patient's cornea was only 0.445mm, then their measured IOP might be 12mmHg, but their adjusted IOP would be corrected upwards to 17mmHg.
Such patients would still have close-to-average IOPs, but if the cornea were even thinner, then their adjusted pressure might fall into the ocular hypertension range (in other words >21mmHg), although this seems unlikely.7
Notwithstanding, if this were the case, then it might be argued that the so-called detection of glaucoma (as based on tonometry) might be delayed if a correction factor were not applied. In either scenario, this does not mean that the correction factor for the IOP has to be applied, but it could be considered logical to apply a correction factor of this magnitude on a routine basis in reviewing all patients under medical treatment for glaucoma, especially those with OHT or NTG.
Using this type of correction (5mmHg for a 0.100mm difference in CCT), that has been referred to as a linear correction scale, one recent study reported that 55.9 per cent of their glaucoma patients had a 'measurement-significant adjustment' in their IOP measures.4
Other studies have used a slightly greater correction factor of 7mmHg for a 0.100mm difference in CCT (derived from a study on cannulated eyes during surgery),12 and noted that 52 per cent of the glaucoma (OHT) patients had adjusted IOPs that were then less than 21mmHg.13
Overall, however, more data is needed on the IOP-CCT relationship for patients with different forms of open-angle glaucoma to assess whether a single correction factor could be developed and applied (and the authors are currently planning further studies, as well as further meta-analyses).
CORNEAL THICKNESS AS A PREDICTOR OF GLAUCOMATOUS DAMAGE TO THE EYE
As outlined above, there is good evidence that the outcome of routine tonometry measures can be affected by the central corneal thickness.
More recently, however, there has been a new twist to the story, with a claim being made of 'central corneal thickness as a risk factor for advanced glaucoma damage'.3
This presentation of the issue seems to have spread like wild fire in tinder dry forests, despite it being a rather misologous claim and one based on a rather different statistical approach to that applied for both consideration of normal CCT and IOP and also in the IOP-CCT analyses.
Notwithstanding, this risk factor perspective deserves some coverage, if for no other reason than to explain it to the high street optometrist.
Recent studies, just as many had done in the previous 10 to 20 years, have sought to identify those patients who might be at risk for progression of their glaucoma.3,4,6,7,13-17
Stated another way, there will be many individuals who will be diagnosed with 'glaucoma'. However, some of these patients show essentially very little change with each passing year, such that they might well be considered as having a stable or unchanging glaucoma or be 'non-progressors'.
However, as first really highlighted with the OHT issue in the late 1980s, were there also subgroups of patients (for example, those with IOPs in the mid-20s) who were more at risk to develop clinically-significant visual field loss? It was once logically argued that those with higher IOPs would be more likely to show visual field progression, as compared to those with a lower IOP, and this conclusion is still valid.
For example, in one recent study POAG patients who had pressures ³21mmHg, were found to have a 1.67 times higher risk of progressive field loss.15 Another way of examining this is to assess the risk (in other words, the hazard ratio) in relation to the degree of ocular hypertension.
One of the recent OHT studies reported that the hazard ratio for further field loss was 1.11 (rather than ²1) for every extra 1mmHg (presumably above 21mmHg).14
Assuming a linear relationship, a patient with chronic pressures of 24mmHg, would thus have a four times higher risk of developing glaucoma. Now, however, we also have the CCT issue to carefully consider.
The just mentioned study also noted that, in OHT patients, for every 0.04mm decrease in CCT, the risk of development of POAG increased 1.88 times, that is OHT patients with 'thinner' corneas appeared to more at risk to develop glaucoma.14
The banner headlined study3 also examined possible association between the extent of visual field defects and the patient's CCT in POAG patients.
The average CCT, by ultrasound, was 0.545mm, but these authors noted that 'an increase in CCT was associated with an improved mean deviation of visual field', and that 'for an increase of 10µm of CCT, the mean deviation of visual field improved by 0.34dB', and concluded that 'patients with POAG who had a thinner CCT tended to have more severe glaucomatous damage'.
However, their statistical analyses also showed that, for the group of patients studied, 'for an increase of 1mmHg of IOP, the mean deviation improved by 0.21dB'.
Therefore, despite the fact that the reported associations were considered to be statistically significant, one now has to carefully consider whether 'a thicker CCT may be protective against glaucomatous damage' or whether 'thinner' CCTs were a real risk factor for progression of field loss.
In other recent studies on POAG or OHT patients, the average CCTs in those showing field loss progression were reported to be between 0.529 and 0.566mm.6,10,14,16 Some of these are not obviously 'thin' values, but it is the relative effect that needs to be considered.
In one study, patients who did show visual field deterioration had lesser average CCT values compared to those without further field loss, namely 0.553mm against 0.574mm.14 These authors also pooled all their data and then 'divided the entire sample into three approximately equal-sized groups' whose corneal thickness values were <0.555mm, greater than 0.555mm or greater than 0.588mm.
On this grouping, they then estimated that those patients in the 'thin' cornea group (<0.555mm) had a 3.4 times higher risk of progression, compared to those with 'thicker' corneas.14
However, in another recent study,6 the three-way equal grouping strategy was again applied to a heterogeneous group of glaucoma patients and the conclusion drawn that there was no detectable difference in progression when comparing those with thin (< 0.554mm), medium (0.555 to 0.586mm) and thick (> 0.586mm) corneas. It is important to note that these thinner and thicker categories bear no relationship to those defined in the meta-analysis, neither is there any robust statistical reason for the sub-grouping.
Notwithstanding, this issue of whether CCT is a predictor of visual field loss has been considered in several other studies. Another recent study on POAG patients noted that those who showed visual field progression had an average CCT that was lower (at 0.529mm) compared to 'nonprogressors' where the average CCT was 0.547mm.10
More recent studies reported that patients who 'had early and previously undetected glaucoma with visual field defects' and who had a 'low' CCT (less than 0.548mm) value, had a 1.25 times higher risk of field loss progression as compared to those with a 'high' CCT ( greater than or equal to 0.548mm).15
Recent studies on NTG patients used a CCT split at 0.520mm to assess whether patients could be predicted to show visual field deterioration or not.7 The same type of analysis was applied in another study, this time on those diagnosed with glaucoma based on disc changes,16 but with the median split at 0.545mm.
For a related study on the retinal nerve fibre layer thickness in POAG patients, it was noted that those with 'thinner' corneas who had lower retinal nerve fibre thickness values and also higher indicators of subtle damage, as compared to those with 'thicker' corneas.17
However, the 'median' split for CCT was now 0.575mm. In recent studies of disease stability and cost assessments in POAG, the median CCT value derived from ultrasound pachymetry was used, with the conclusion that those glaucoma patients with 'thin' corneas (<0.544mm) showed a different pattern of disease progression to those with 'thick' corneas (namely >0.544mm).18
The point that needs to be realised is that the CCT values in glaucoma patients, and their possible relationship to glaucomatous damage or progression of the same, are not absolute.
The actual values of CCT for which an increased or reduced risk are not the same across the different studies.
CONCLUSIONS AND A LOOK TO THE FUTURE
At the present time, there is more than enough evidence that central corneal thickness (CCT) does need to be considered as part of the interpretation of tonometry data in glaucoma patients.
However, there are still no satisfactory, or statistically robust, ways of generally applying some form of correction factor to all tonometry readings as based on CCT.
One option has been to apply the correction factor for eyes with chronic disease (for example, a 2.5mmHg correction for a 0.05mm difference in CCT).4,7,13 However, others have felt that one could also apply the correction factor derived for normal eyes (namely, just 1mmHg for a 0.05mm difference in CCT),11 or that derived from cannulation studies (namely, approximately 3.5mm for a 0.05mm difference in CCT).13
It has to be noted that different pachymeters can produce somewhat different results, and we still do not know if eyes with stable glaucoma have a different IOP-CCT relationship to those considered to be normal. Patients need to be considered on a case-by-case basis.
As to whether a thin CCT, per se, might be a useful indicator that eyes are more likely to progress to develop glaucoma or progress at a faster rate, no conclusions can yet be drawn.
None of the analyses so far presented use a logical or statistically robust analytical approach, and none have (apparently) used a cut off based on extremes of CCT (that is <0.473mm or >0.595mm) which is the basis of the classification of normal versus abnormal CCT values.1,2
The cut off criteria used to categorise at-risk patients shifts according to the calculations applied by different groups.
While this is obviously different from the time-honored 21mm cut off for IOP (as based on clinical tonometry without correction), it is perhaps a reasonable recognition of current perspectives on IOP and its role in glaucomatous field loss.
However, as also recently concluded,6,10 there remains more than a distinct possibility that patients with thinner corneas may both show unexpected optic disc changes and field loss, as well as then deteriorating faster once they are being monitored, because of a difference in initial patient management strategies.
It is not obvious, or proven, that patients with thinner corneas are at higher risk of glaucoma, but rather that they could have been essentially misdiagnosed at a stage when their glaucoma was first developing.
With IOPs within normal limits and generally unremarkable ophthalmoscopy findings, they were classified as normal and so did not receive any treatment. These are the NTG patients. In contrast, those who presented with higher than normal IOPs (that is, >21mmHg) were likely 'immediately' scheduled for further assessments and perhaps even placed on medical treatment. These are the OHT patients.
So, do we need to worry about thin or thick corneas? This is a fundamental and important question but one for which there is no obvious or easy answer at this time.
The European Glaucoma Society Guidelines (Figure 7), state that 'CCT measurements are unavoidable for the correct measurement of OHT', but that 'CCT measurements have limited value for POAG assessment...'.
The recommendation to include pachymetry as part of assessments of OHT patients is being applied in some optometry-led shared care glaucoma schemes in the UK.19
However, it will be a long time before this recommendation can be usefully applied to general optometric practice in the UK, unless a general decision was made that any referral for 'glaucoma' should be accompanied by not only IOP and visual fields data, but also pachymetry data.
At the present time, in UK optometry, we cannot state with confidence that even the first two assessments would be routinely undertaken as part of an eye test.
References
1 Doughty MJ, Jonuscheit S. OPTICIAN, 2005; 230 (no. 6013), 2732.
2 Doughty MJ, Zaman ML. Surv Ophthalmol, 2005 44: 367-408.
3 Herndon L, et al. Arch Ophthalmol, 2004; 122: 17-21.
4 Shih CY et al. Arch Ophthalmol, 2004; 122, 1270-1275.
5 Brandt JD, et al. Am J Ophthlamol, 2004; 138: 717-722.
6 Jonas JB, et al. Invest Ophthalmol Vis Sci, 2005; 46, 1269-1274.
7 Doyle A, et al. Acta Opthalmol Scand, 2005; 83: 191-195.
8 Aakre BM, Doughty MJ et al. Ophthal Physiol Opt, 2003; 23: 271-283.
9 Doughty MJ, et al. Ophthal Physiol Opt, 2002; 22: 491-504.
10 Kim JW, Chen PP. Ophthalmology, 2004; 111: 2126-2132.
11 Tonnu PA, et al. Br J Ophthalmol, 2005; 89:851-854.
12 Ehlers N, et al. Acta Ophthalmol, 1975; 53: 34-43.
13 Brandt JD, et al. Ophthalmology, 2001; 108: 1779-1788.
14 Gordon MO, et al. Arch Ophthalmol, 2002; 120: 714-720.
15. Leske MC, et al. Arch Ophthalmol, 2003; 121: 48-56.
16. Medeiros FA, et al. Am J Ophthalmol, 2003; 136: 805-813.
17 Henderson PA. Ophthalmology, 2005; 112: 251-256.
18. Walker JH, et al. Curr Med Res Opinion, 2005; 21: 489-494.
19. Jones A. OPTICIAN, 2005; 229 (no. 6007): 25-25.
- Michael Doughty is a professor at Glasgow-Caledonian University. Donald Montgomery is an ophthalmologist, specialising in glaucoma, at Glasgow Royal Infirmary
| SUMMARY |
|
|
|
 | Providing exclusive eye care news, information and educational needs every week, including a FREE CET programme. Subscribe to Optician Print Edition. |
