For more than 50 years, Goldmann Applanation Tonometry (GAT) has been the gold standard for measuring intraocular pressure (IOP) [1] . However, several biomechanical limitations of this method have been characte- rized. These include the need for mathematical adjustment of the obtained values for central corneal thickness, uncertain accuracy in the presence of corneal scarring and corneal curvature, and the known underestimation of IOP in eyes with corneal edema [2] -[7] . Therefore, we still lack an IOP measurement device that yields repro- ducible and exact IOP values in patients suffering from corneal pathologies like Fuchs’ endothelial dystrophy, or for patients who have undergone penetrating keratoplasty [8] [9] .

Several new devices aiming for correct IOP measurement that is less dependent on the cornea’s biomechani- cal properties have been developed. Two of these devices are the Dynamic Contour Tonometer (PASCAL^®, DCT; Ziemer Ophthalmic Systems AG, Port, Switzerland) and the Tono-Pen XL^® (TP; Reichert, Depew, New York, United States).

The purpose of this study was to investigate the accuracy of GAT and these new devices compared to direct intracameral IOP readings in a prospective in vivo study in patients suffering from corneal pathologies.

8 patients (a total of 8 eyes) scheduled for penetrating keratoplasty due to corneal pathology were included in this study. Four patients were experiencing bullous keratopathy after cataract surgery; one patient had non-her- petic scars; one patient had Fuchs’ endothelial dystrophy; and two patients needed to undergo keratoplasty again due to graft failure or corneal scarring. All eyes included in this group were pseudophakic. In a control group, 10 patients without corneal pathology who were scheduled for cataract surgery were included.

The indications for penetrating keratoplasty and cataract surgery were independent from participation in this study. Exclusion criteria included any history of ocular surgery in the 4 weeks prior to penetrating keratoplasty or cataract surgery, or any active intraocular inflammation. Patients scheduled for penetrating keratoplasty had to be pseudophakic to participate in the study. The presence of any corneal pathology was an exclusion criterion for the control group.

Before surgery, central corneal thickness (CCT) was measured with an ultrasonic pachymeter (Pachette^TM, DGH Technology Inc., Exton, Pennsylvania, United States).

For penetrating keratoplasty, patients were anesthetized with general anesthesia. Cataract surgery was per- formed either in general anesthesia or in retrobulbar anesthesia (injection of 5 ml mepivacaine [Scandicain] 2% 15 minutes prior to surgery; no oculopression was performed).

All IOP measurements were performed before surgery by one of the authors (MN). With the aid of a laser water scale, the tubing system used to manometrically determine the IOP was adjusted to the height of each pa- tient’s anterior chamber before the measurement was taken.

The experimental setting for adjusting the IOP has been described elsewhere in detail [10] . The anterior chamber was cannulated at the temporal corneal limbus with a cannula connected to an adjustable water column and a pressure sensor used as reference (the DCT tip was directly connected with the water column, Figure 1). The IOP was directly calibrated by changing the height of the water column in relation to the anterior chamber (open system) to three different IOP levels (10, 20, and 30 mmHg). After each IOP level was reached, the sys- tem was closed, and three consecutive measurements (DCT, GAT, Tono-Pen, and simultaneously with the ref- erence sensor) were taken (closed system). The maintenance of a tension-free position for the cannula was main- tained with high priority. This tension-free position was reached by using sterile adhesive tape (Steri-Strip^TM Adhesive Skin Closures; 3M Deutschland, Neuss, Germany) and sterile pads to fix the cannula (Figure 2).

Measurement with the DCT (the measurement principle for the DCT has been described in detail elsewhere [11] ) was performed with a modified Perkins tonometer, allowing performance of dynamic contour tonometry measurement on a patients in the supine position, with a DCT tip mounted on the body of a Perkins tonometer

(Figure 2). This instrument was developed by Ziemer Ophthalmic Systems AG (Port, Switzerland). During the measurement, lubricant eye drops (Corneregel^® Fluid Eye drops; Mann Pharma) were used to ensure a normal corneal surface and optimal quality for the DCT measurements. If no audible pulsation was present, the mea- surement of the IOP was aborted and not included in the statistical analysis. The DCT provides an internal qual- ity assessment of the measurement, with Grade 1 representing the best and Grade 5 representing the worst qual- ity. Only measurement quality 1 to 3 was accepted. If it was not possible to gain a measurement with a quality grade of 3 or better in a patient after three attempts, the patient was excluded from the study.

The Tono-Pen^® XL (TP; the way it works has been detailed elsewhere [12] ) performs four independent IOP readings and provides a reliability index for each single measurement. Measurements were only used if the scat- ter index was below 5%. Three of the measurements with an index below 5% were taken, and the average of these measurements was calculated.

Three consecutive measurements with GAT were performed using the Perkins tonometer (HAAG-STREIT AG, Koeniz, Switzerland); the average of these three measurements was calculated.

In the following description, the term DCT measurement refers to the measurement with the handheld DCT; the term GAT refers to the measurement with the Perkins tonometer; the term TP refers to measurements taken with the Tono-Pen^® XL; and the reference point is the DCT tip integrated into the tube system to obtain intra- cameral measurements.

All patients signed informed consent before entering the study. The study was performed in accordance with the Declaration of Helsinki and was approved by the institutional ethics committee of the Medical Department of the University of Freiburg.

Statistical analysis: We calculated the measurement errors from the difference between the intracameral ref- erence pressure and the respective measurement result. We computed average measurement errors for each de- vice and reference pressure separately. We plotted measurement errors against reference pressure for each de- vice, and performed a regression analysis by means of a linear model. Intercept and slope of the regression lines were used to rank the devices for validity. The regression and the dependency of the measurement results on CCT were tested for statistical significance by means of analysis of variance (ANOVA). All computations were performed with the R programme (www.r-project.org).

Control group (Figure 3): In the control group of 10 patients scheduled for cataract surgery, all three devices showed good agreement with the adjusted IOP as measured by the reference sensor. The mean deviation for all adjusted IOP levels was −0.9 mmHg ± 5.5 mmHg (mean ± SD) for DCT; −0.9 mmHg ± 3.4 mmHg for GAT; and −0.9 mmHg ± 2.5 mmHg for TP. Intercept and slope of the regression line for each device can be found in Table 1.

Group with corneal pathologies (Figure 3): When taking into account all examined patients, the TP yielded the most exact IOP values in the group with corneal pathologies; we found a mean deviation of −0.9 mmHg ± 3.2 mmHg. Mean deviation was −2.9 mmHg ± 3.3 mmHg for GAT and −5.0 mmHg ± 7.9 mmHg for DCT. In- tercept and slope of the regression line for each device are displayed in Table 1.

Bullous keratopathy (Figure 4): In this group, all 3 devices underestimated the adjusted IOP. We found a mean deviation of −0.2 mmHg ± 3.5 mmHg for TP, −4.0 mmHg ± 3.2 mmHg for GAT, and −6.5 mmHg ± 5.1 mmHg for DCT, respectively. This underestimation of the adjusted IOP increased with increasing IOP during measurements obtained with GAT and DCT.

Fuchs’ endothelial dystrophy (Figure 4): Both TP and GAT underestimated the IOP (mean deviation −2.3 mmHg ± 0.4 mmHg (TP) and −1.6 mmHg ± 1.7 mmHg (GAT)). DCT showed a greater scattering in compari- son with the other devices (mean deviation +0.1 mmHg ± 3.7 mmHg).

Post-keratoplasty (Figure 4): All three devices underestimated the adjusted IOP with a great degree of scat- tering, especially in the instance of DCT. Mean deviations were −0.5 mmHg ± 3.2 mmHg (TP), −1.9 mmHg ± 4.1 mmHg (GAT), and −0.6 mmHg ± 11.4 mmHg (DCT).

Non-herpetic scars (Figure 4): All three devices underestimated the adjusted IOP, with the DCT showing the greatest deviation and scattering. Mean deviations were −3.6 mmHg ± 1.8 mmHg (TP), −1.9 mmHg ± 2.8 mmHg (GAT), and −13.1 mmHg ± 5.9 mmHg (DCT).

Central corneal thickness (CCT) (Figure 5): Mean CCT was 559 µm (±19.5 µm) in the control group and 616 µm (±28.4 µm) in the group with corneal pathologies. None of the measurement devices showed a statistically relevant dependency on the CCT in the control group. In the group with corneal pathologies, the DCT and the GAT showed a statistically significant negative correlation with the CCT (DCT: t = −2.3, P = 0.03; GAT: t = −2.2, P = 0.04). In this group, only the Tono-Pen showed no statistically relevant dependency on CCT (t = −0.0244, P = 0.99).

Table 1 shows the intercept and slops for each measurement device according to diagnosis. No statistically relevant dependency on astigmatism or axial length was found in any of the devices.

The gold standard for measuring IOP, Goldmann Applanation Tonometry (GAT), is known to depend on CCT and several other corneal properties (edema, tear film, astigmatism) [1] -[4] . A device that can yield exact and reproducible IOP values is still necessary, especially in patients suffering from corneal pathologies such as bul- lous keratopathy, Fuchs’ endothelial dystrophy, or corneal scaring [7] -[9] . Accurate IOP measurement remains difficult after penetrating keratoplasty, but is important nonetheless, as elevated IOP values negatively affect graft survival [13] .

In recent years, several methods to measure IOP independent from corneal properties have been developed. Dynamic Contour Tonometry (PASCAL^®, DCT) is based on the physics of Pascal’s law. This device is said to measure IOP independent from CCT or other biomechanical corneal properties; this has been confirmed in sev- eral clinical trials [10] [14] [15] . After LASIK eye surgery, DCT showed no change in IOP measurement, while GAT revealed lower IOP levels [16] . Nevertheless, there have been recent reports that CCT and corneal edema may influence DCT measurements [17] [18] . Considering this method’s measurement principles, the cornea’s

viscoelastic properties should also be taken into account as another potential influencing factor. After penetrat- ing keratoplasty, Viestenz found that DCT yielded reproducible IOP values that were higher in comparison to those for GAT, and were disturbed by postoperative intrastromal sutures [19] .

The Tono-Pen^® XL (TP) is a handheld tonometer operating on the same principle as the Mackay-Marg tono- meter [12] . The accuracy of the TP is still a matter of discussion. Several authors have reported that its mea- surements are not dependent on CCT, while others did find a positive correlation [20] [21] . When comparing TP to GAT, some authors found the Tono-Pen to underestimate or overestimate the IOP, with a particular trend to- ward overestimation at higher IOP values [20] [22] [23] . In patients who had undergone penetrating keratoplasty, both Shemesh and Rao found the TP to show good agreement with IOP values yielded by GAT [24] [25] .

In or study, we wanted to examine which of the three devices for measuring IOP (GAT, TP, DCT) yielded the most exact IOP values in patients with corneal pathologies. We therefore used a closed system that allowed for adjustment of the IOP of each patient manometrically to 10, 20, and 30 mmHg, respectively, and for control of this adjusted IOP through a reference sensor. At each pressure level, we measured the IOP using each of the three devices.

This study was performed in 8 patients scheduled for penetrating keratoplasty due to bullous keratopathy (n = 4), Fuchs’ endothelial dystrophy (n = 1), non-herpetic scars (n = 1); or after keratoplasty (n = 2), as well as in 10 patients scheduled for cataract surgery who served as the control group.

In the control group, we found all three devices to yield exact IOP values. Each device showed a mean devia- tion from the IOP value of −0.9 mmHg. These results indicate the reliability and accuracy of our measurement configuration. None of the three devices in the control group showed a statistically relevant dependency on CCT. For the DCT, this is in accordance with most studies; the literature has generally reported that DCT is indepen- dent of CCT [10] [26] . Nevertheless, most of the authors report a positive correlation for GAT and TP with CCT [2] [27] . Our 10-patient control group was small; our results therefore allow no obvious conclusion concerning the relationship of GAT and TP and CCT with respect to measuring IOP.

In the group with corneal pathologies, the TP yielded the most accurate IOP measurements for all three de- vices, followed in accuracy by GAT and DCT. For the entire group of patients with corneal pathologies, the IOP measurement results for TP were just as accurate as those for the control group. Several other studies have re- ported similar results, with the TP yielding exact IOP values after penetrating keratoplasty [24] [25] . Neverthe- less, these groups compared TP measurements of IOP to IOP values obtained by GAT, and not to direct intra- cameral readings. The TP not only yielded the lowest mean deviation over all diagnoses (−0.9 mmHg, compared with −2.9 mmHg [GAT] and −5.0 mmHg [DCT]), but also the lowest standard deviation (−3.2 mmHg [TP], −3.3 mmHg [GAT], −8.0 mmHg [DCT]), indicating the reliability of the measurement. The DCT in particular had a very high standard deviation for its various measurements; the results for this device showed a high degree of scattering around the adjusted IOP.

Despite the small number of patients, a statistically significant negative correlation existed between the CCT and the IOP (GAT) and especially the IOP (DCT) measurement results in the group with corneal pathologies for all diagnoses. Most patients in the study group suffered from pathologies that led to edematous swelling of the cornea (Fuchs’ endothelial dystrophy, bullous keratopathy, graft failure). Indeed, mean CCT in the group with corneal pathologies was significantly higher than in the control group (559 µm vs. 616 µm; P < 0.05). Our re- sults are therefore in accordance with Simon and also Oh, who recorded significantly lower IOP (GAT) meas- urement results in edematous corneas [5] [6] . Francis reported the same negative correlation between CCT and IOP (DCT) measurement results on edematous corneas [17] . Tono-Pen measurements showed no statistically relevant correlation between the IOP measurement results and the CCT; this is in accordance with the accurate IOP measurement results delivered by this device in our study, especially in the bullous keratopathy group.

Due to the small number of patients in the respective group suffering from Fuchs’ endothelial dystrophy, non- herpetic scars, bullous keratopathy, and post-keratoplasty, an interpretation of our results with respect to each individual diagnosis is difficult (Figure 4). Nevertheless, especially in the group with bullous keratopathy, TP yielded the most accurate measurement results, whereas GAT and, to an even greater degree, DCT underesti- mated the IOP at all adjusted pressure levels, especially with increasing IOP. In the group of patients with non- herpetic scars or who had undergone keratoplasty, we found that GAT measurement results were surprisingly accurate despite the irregular corneal surface. Tono-Pen also yielded quite exact IOP values, whereas the DCT showed a great variance. All three devices yielded relatively accurate IOP values on Fuchs’ endothelial dystro- phy.

One possible explanation for the accuracy of the IOP measurement results obtained by TP may be the small contact area of this device with the cornea, and the minor distortion of the cornea during measurement. Possible biomechanical properties of the cornea, such as edema or a higher degree of stiffness due to scarring, therefore may have less influence on the measurement procedure. The higher resistance against distortion caused by cor- neal scarring, or the lower resistance due to corneal edema, might explain the worse results obtained by GAT and especially by DCT in our study.

The small number of patients in the study group was a major important limitation of our study. This small number makes statistically significant statements regarding the respective diagnoses difficult. Nevertheless, when taking all diagnoses into account, the results concerning the high accuracy of TP and its independence from CCT measurements as well as the dependance of GAT and DCT from CCT measurements were all statis- tically significant. Regarding DCT, the measurement procedure with the modified Perkins tonometer was prob- lematic, as the device had to be held in a stable position over the supine patient for several seconds. It was therefore sometimes complicated to obtain measurement results of adequate quality. Nevertheless, as only measurement results with a quality grade of 3 or better were provided in the statistical analysis, the difficulties that occurred during the measurement procedure could not have played a role in the interpretation of the poor quality of the measurement results obtained by this device.

In summary, our results suggest that the Tono-Pen XL^® is the most accurate measurement device for the de- termination of IOP in patients with corneal pathologies, especially in patients suffering from corneal edema (bullous keratopathy), probably due to its relatively small contact area with the cornea. These results are in ac- cordance with several other groups reporting on its accuracy compared to GAT after penetrating keratoplasty. To our knowledge, our study is the first to reveal this accuracy when comparing the obtained IOP values to di- rect intracameral readings.

GAT yielded surprisingly exact IOP values in patients with an irregular corneal surface (non-herpetic scars, post-keratoplasty), and underestimated IOP in patients suffering from corneal edema. Apart from Fuchs’ endo- thelial dystrophy, DCT showed great deviations from the adjusted IOP in all patient groups and should not be used to determine IOP in corneas with any pathologies or disorders.