Tonometry reliability questioned

Key Points

New Orleans-Rather than rely exclusively on tonometry measurements, clinicians should develop a healthy skepticism of tonometry and extend that skepticism to thick and thin corneas, said James D. Brandt, MD, during glaucoma subspecialty day at the annual meeting of the American Academy of Ophthalmology.

"The problem of measuring pressure is an old one," said Dr. Brandt, professor of ophthalmology and director, glaucoma service, at the University of California, Davis, in Sacramento. "With the recognition that glaucoma was a disease associated with elevated pressure, some of the greatest minds in ophthalmology have tinkered with devices to measure IOP. At the same time, luminaries of ophthalmology voiced a healthy skepticism about these new devices, and it seems that their skepticism was probably well founded.

"Attempts to refine individual IOP estimates will not help in the care of individual patients," Dr. Brandt said. "My recommendation is that you categorize corneas as thin (<510 μm), average (510 to 580 μm), and thick (>580 μm) and integrate that information into your global assessment of risk. Attempts to be more precise than this are simply not supported by the data."

Use CCT routinely

It is widely recognized that pachymetry has become an integral part of the glaucoma exam. This recognition has been driven primarily by the growing recognition that Goldmann applanation tonometry is far less accurate than generally appreciated, Dr. Brandt said. The Ocular Hypertension Treatment Study (OHTS) demonstrated that central corneal thickness (CCT) is a powerful predictive factor for the development of glaucoma, and it is now commonly accepted that CCT should be determined as a routine part of the glaucoma examination.

In the mid-1990s, pachymetry was introduced to the OHTS protocol in response to a small but growing body of literature confirming Goldmann's original conjecture that variations in CCT would be an important confounder for his tonometer. This introduction was particularly relevant among patients with ocular hypertension, Dr. Brandt said. In addition, all of the pachymetric and biometric studies to date had been performed in Caucasian populations in Scandinavia. Researchers wondered whether racial differences in CCT might underlie variations in disease susceptibility. Because 25% of the OHTS cohort was African-American, investigators had an excellent opportunity to explore this question.

"At a more fundamental level, though, our hope was that by correcting IOP data with corneal thickness, we would find a stronger dose-response relationship between pressure and glaucoma risk," Dr. Brandt said. "This has not proven to be the case."

OHTS revealed that CCT was a powerful, independent predictive factor for the development of primary open-angle glaucoma among patients with ocular hypertension. At a given IOP level or cup-to-disk ratio, patients with thinner corneas were at much higher risk of developing glaucoma than those with thicker corneas.

"What is not appreciated by most is that the OHTS statisticians have attempted to correct IOP using the various published algorithms, and corneal thickness does not drop out of the multivariate model," Dr. Brandt said. Although the reasons for this occurrence are complex, the net result is that clinicians need more than two pieces of information (Goldmann tonometry and corneal thickness) to arrive at a more accurate IOP measurement.

According to Dr. Brandt, two terms that often mistakenly are used interchangeably must be understood clearly: accuracy and precision. In the context of tonometry, accuracy represents the degree to which a measurement reflects true IOP, whereas precision represents the repeatability of the measurement.

To adjust or not adjust?

For the practicing ophthalmologist who treats patients with glaucoma, the main concerns revolve around what to do with CCT and tonometry measurements acquired from the patient in the exam chair, and whether to adjust a patient's IOP measurement, Dr. Brandt said.

"I would argue that you should not," he added.

Dr. Brandt posed three arguments for not attempting to correct pressure measurements.