Background: A rise in intraocular pressure (IOP) in upgaze is regarded as a diagnostic sign in Graves' ophthalmopathy (GO). However, the question of erroneous IOP measurement due to applanation carried out on the peripheral cornea has never been addressed.
Methods: In 22 healthy volunteers, as well as in 51 GO patients, applanation tonometry was performed in the primary position of gaze and at 20 degrees of upgaze. In addition, applanation tonometry was repeated using a flexible chin rest to incline the head and produce 20 degrees upgaze. This enabled applanation on the central cornea.
Results: In healthy controls, mean IOP in conventional upgaze showed a significant rise compared to primary position (p < 0.0001). IOP measurements in 20 degrees upgaze/head inclination were significantly lower compared to conventional upgaze tonometry (p < 0.0001) and comparable to mean IOP in primary position (p = 0.7930). Mean IOP in GO patients was also significantly higher in conventional upgaze compared to primary position (p < 0.0001). The upgaze measurements obtained by head inclination were significantly lower than those from conventional upgaze tonometry (p < 0.0001), but showed a statistically significant rise compared to mean IOP in primary position (p < 0.0001). The overlap of IOP readings in upgaze between normal individuals and GO patients was considerable, even in patients with severely impaired ocular motility.
Conclusion: In both normal volunteers and patients suffering from GO, a rise in IOP was observed in conventional upgaze tonometry. However, this increase in IOP was partially due to applanation on the peripheral cornea. Measurements in upgaze by head inclination on the central cornea led to a significant lowering of the gaze-dependent IOP change. The discriminating power of the IOP difference between upgaze and primary position to diagnose GO was found to be limited. The broad overlap of IOP between normal individuals and GO patients as detected by conventionally performed upgaze tonometry leads us to conclude that this sign may not be of relevant differential diagnostic value in patients with a clinically undetermined diagnosis.