Reconsidering the Gold Standard of Tonometry
A glaucoma specialist argues for a new method of applanation tonometry.
IOP is one of the key metrics used to monitor the health of an eye, especially an eye with glaucoma. Tonometry is the act of estimating the IOP. It is not a measurement of eye pressure, only an estimate. An exact measurement of IOP requires live manometry. At present, no clinical tonometer is capable of directly measuring IOP, and very few have been studied against live manometry. The concept of estimation is critical, because every tonometer has limitations based on the mechanism of the instrument and assumptions of the method. This article will detail the mechanisms of the most prominent tonometers and highlight the strengths and limitations of each.
UNDERSTANDING IOP METRICS
Discussing tonometers requires understanding a couple of important metrics. The first is accuracy, which refers to the ability of an instrument to produce the true value. The determination of accuracy requires an absolute standard. As noted earlier, the only way to determine the accuracy of a tonometer is to compare it to simultaneous live manometry. It is not possible to use one tonometer to calibrate another. Covarying and confounding errors of each tonometer cannot be overcome with any technique other than simultaneous manometry. Unfortunately, few studies employ this method, which is why accuracy is often the least-researched aspect of a tonometer.
Precision, the second metric, is the ability of a tonometer to read the same result with repeated testing at exactly or about the same time on the same eye. Precision is very easy to study and is often the most touted feature of an instrument. An instrument’s precision, however, has no influence on its accuracy. These features are completely independent. Too many studies mistake precision as a measure of a tonometer’s accuracy. This practice is completely incorrect and misleading.
THE CURRENT GOLD STANDARD
Imagine a tonometer that is heavily influenced by the operator, so much so that a masked reader is required when IOP is a main outcome measure in a study. One person physically applies the tonometer, while another reads the result. This same tonometer is also markedly influenced by corneal thickness and many other factors— known and unknown. Of course, I am referring to the Goldmann tonometer, which has been and often still is considered the gold standard of tonometry, yet this instrument’s inaccuracy and imprecision are well documented. Why is it still considered a gold standard?
I believe the history of tonometers helps to explain the devotion to the Goldmann tonometer. The concept of a tonometer was born in Germany in the mid-1800s. Von Graefe and Donders invented crude devices, but more important, they lit the spark among scientists for a working tonometer. Hundreds of variations were developed over the next decades. Eventually, Schiotz came up with an indentation tonometer in 1904 that appeared to work (it did not and does not). The shortcomings of indentation filled volumes of journals. Some 50 years later, Goldmann presented his version of the applanation tonometer. Applanation was a dramatic improvement over indentation, and I believe this leap of technology is what solidified its position as the ultimate tonometer. The FDA’s requiring its use for IOP studies and the historical consensus of opinion bolster this belief.
Consensus is a wonderful thing for people driving on a road. Otherwise, chaos and carnage would result. It is also great for a group of diners when deciding on a place to eat, but consensus in science is horrible and exactly the opposite of the scientific spirit of inquiry. No group, no matter how large, knows the truth. Therefore, any group declaration of consensus can only lead to a stifling of exploration, stagnation of knowledge, and ultimately, a delay in scientific progress. With this in mind, I will now discuss the fundamentals of tonometry without regard to historical views.
GOLDMANN APPLANATION TONOMETRY
Applanation is the mechanical flattening of the cornea. There are two varieties: constant area with variable force (eg, Goldmann applanation tonometry) and constant force with a variable area (eg, Maklakov applanation tonometry). The mechanism of applanation is classic Newtonian force versus countervail. The tip presses against the cornea to a defined endpoint, where the calibration curve is then used to estimate the IOP.
Goldmann applanation tonometry, the force is changed until the defined area of flattening is achieved. This is simple in theory but awkward in practice. Difficulties in achieving a proper endpoint are numerous and well documented, and they explain why this instrument is so dependent on the operator. Many common issues render this tonometer unusable. The fixed-area endpoint does help to neutralize the nonlinear elements of corneal flattening to a large degree. However, it still results in a pressure curve that is much too flat with respect to manometry such that the higher the actual pressure is, the less change is noted by the tonometer. This is a dangerous situation in that the higher the IOP, the greater the underestimation and likelihood that the physician will not take appropriate action. The main advantages of the Goldmann tonometer are its simple mechanics and relatively low cost.
MAKLAKOV APPLANATION TONOMETRY
Maklakov applanation tonometry uses a constant force and then determines the amount of area flattened. In theory, this style of applanation should be less accurate than Goldmann style because corneal flattening is not linear, but this has not been well studied because the instrument has mostly been abandoned. A newer device, the Pascal Dynamic Contour Tonometer (Ziemer Ophthalmic Systems AG, Port, Switzerland), uses a high-sensitivity pressure transducer that measures the countervail of the cornea when 1 g of pressure is applied. This transducer is able to measure what I call a virtual area because of its sensitivity. It is not really measuring the area of flattening but the countervail produced by the area flattened. I use the word flattened loosely, because the tip is curved and does not actually make the cornea flat, only flatter. The corneal force is directly correlated with the amount of corneal flattening by the constantly applied force of the tonometer’s tip and therefore qualifies as a Maklakov applanation tonometry device. The entire system is automated so that the operator is only required to properly align the tip on the cornea.
The manufacturer of the Pascal Dynamic Contour Tonometer disputes my description of its instrument and insists that it measures IOP by the “Pascal principle,” which has little in common with Pascal’s principle in physics. The company describes the term to mean that the pressure is equal on each side of a thin membrane, which is a version of the Imbert-Fick principle. The manufacturer claims that the tip relaxes the cornea and allows it to act like a flaccid membrane. This is complete nonsense, because the cornea is not and cannot become a thin, flaccid membrane. However, the tip can stabilize surrounding corneal tissue and allow a more consistent measurement in the center, as discussed later in this article.
The Pascal Dynamic Contour Tonometer is definitely not a Goldmann-type instrument, but it is an applanation instrument of the Maklakov type. It is a better tonometer than the Goldmann device, because the former removes the operator bias completely. The microapplanation permitted by the very sensitive pressure sensor also greatly reduces corneal influences. The instrument must be mounted on a slit lamp, it cannot function on eyes with nystagmus or significant corneal abnormalities, and it is less studied than the Goldmann. Overall, I rank the Pascal Dynamic Contour Tonometer well above and greatly prefer it to the Goldmann tonometer.
The Tono-Pen (Reichert Ophthalmic Instruments, Inc., Depew, NY) and others are descendants of the Mackay- Marg tonometer. The Mackay-Marg tonometer was an outstanding instrument that used dynamic applanation to determine the IOP. It belongs in the Goldmann family, because it measured a corneal countervail at a fixed flattened area. The tip had a flat circular area with a central plunger. As the tip was applied to the cornea, the user monitored the position of the plunger with respect to the flat disc head. If the tip was extended, then the pressure applied was too low. If the tip was pushed beyond the plane of the head, then the pressure was too high. As the tip was applied to the eye, the plunger would move from too low to too high and back to too low. These transitional points, one going up and the other down, produced small inflections in the pressure-time curve and looked like a human silhouette, so they were often called shoulders. The mean pressure of the two shoulders was the IOP reading.
To fully appreciate the mechanism of this instrument, I have to discuss thick material physics. Most everyone learns of the Imbert-Fick principle when the topic of applanation arises, and the originators of the Mackay- Marg tonometer used Imbert-Fick as the basis of their instrument, too. This principle relies on a number of concepts that do not exist. The main one is the infinitely thin sphere of the cornea, similar to the flaccid membrane noted earlier. The cornea is certainly not infinitely thin. Nor is it a sphere.
Tonometry is probably better thought of as a subspecialty of rheology—the science and study of plastic and non-Newtonian fluids. The cornea has viscoelastic and static structural properties. Surrounding corneal tissue also influences any spot measurement. A rough estimate of this influence is about 1.5X corneal thickness. A 1-mm-diameter contact point on a 500-μm thick cornea is therefore actually influenced by 0.75 mm of additional tissue (0.5 X 1.5) all the way around. This means that the area of cornea influencing the contact point is really 2.5 mm in diameter. The 1-mm-diameter contact is only 0.78 mm2 of corneal surface area, but there is almost a 5-mm2 area of cornea influencing that center.
The Mackay-Marg tonometer had a 5-mm-diameter disc with a 1-mm-diameter central plunger. The genius of this design was that the tip stabilized the surrounding tissue so that the central plunger functioned under more uniform conditions. Reviewing the original literature reveals that the authors were likely unaware of this issue and had a fortuitous design. Regardless of their intent, no tonometer has been found more accurate or precise than the Mackay- Marg, but it never surpassed the Goldmann tonometer in popularity and was ultimately abandoned. Modern descendants such as the Tono-Pen do not follow the same mechanics as the original and have not demonstrated equivalent accuracy. Older versions of the Tono-Pen had sampling rates too slow to accurately determine the inflection points. Sampling rates have reportedly improved with newer versions, but data remain unpublished.
Model 30 Pneumatonometer
The third member of the Goldmann applanation tonometry family is the Model 30 pneumatonometer (Reichert Ophthalmic Instruments, Inc.). This instrument has a 5-mm tip like the original Mackay-Marg, but all similarities end there. The Model 30’s tip is a silicone membrane cap attached by a semiflexible tube to a metal tube that floats on an air bearing. Air is pumped through the tube to the membrane at a constant flow. The air escapes via the membrane, and the resistance is generated by the applied force at the cornea, causing the membrane to block outflow through the tube. This is also a dynamic form of tonometry. The membrane seals the gas’ exit until the gas’ force exceeds the membrane’s force. Gas then escapes, which decreases the internal pressure and causes the membrane to seal again. The membrane frequency is approximately 200 Hz. The software samples the line pressure at 40 Hz, providing a real time pressure curve accurate enough to determine ocular pulse amplitudes. The semiflexible tube allows for a gimbal-like effect that helps correct for abnormal applanation angles. The air bearing and continuous tonometry make this the only tonometer able to measure IOP in the presence of nystagmus and tremor. The 5-mm membrane tip stabilizes the applanated area, similar to a Mackay-Marg tonometer, because the actual gas outlet is in the central 1 mm. The software determines the endpoint, allowing true independence from the operator. The combination of a stabilized cornea, continuous dynamic measurement, an air bearing, and a small handpiece makes the Model 30 extremely robust, accurate, and precise.
Numerous studies have demonstrated that this instrument functions correctly in settings that cause other tonometers to fail such as after LASIK or penetrating keratoplasty; in the presence of corneal scarring and astigmatism, nystagmus, poor epithelium, bandage contact lens use, and a keratoprosthesis; and in severely obese patients and young children. The Model 30 can be used while the patient is seated or supine (subtract 1.5 mm Hg while the patient is used supine), and it can even be used on sclera for a rough estimate.
The main disadvantages of the Model 30 are its cost and the training required to properly install the membrane tip; most poor results with this instrument can be traced to incorrect installation of the tip. The effect of corneal thickness on the Model 30 remains unknown. LASIK studies have shown only a 1-mm Hg difference before and after LASIK, regardless of ablation amount, making this the preferred tonometer to use in eyes that have undergone LASIK. However, manometric studies assessing the effects of natural corneal thickness are absent.
For 56 years, we have relied on the Goldmann tonometer. The principle of applanation remains sound, but it is long past time for us to replace the original instrument with improved ones like the Pascal Dynamic Contour Tonometer and the Model 30 pneumatonometer. Eventually, a new method of tonometry will succeed applanation, but this will only happen when we finally abandon the misguided consensus that the Goldmann is the gold standard.
Dan L. Eisenberg, MD, is a glaucoma specialist at the Shepherd Eye Center in Las Vegas. He acknowledged no financial interest in the products or companies mentioned herein. Dr. Eisenberg may be reached at firstname.lastname@example.org.