Predicting progression

Emerging evidence suggests that measuring circadian intraocular pressure (IOP) patterns with a contact lens sensor (CLS) could help identify glaucoma patients at higher risk of progression. But more work is needed to make it clinically useful, Arthur J Sit SM, MD told the Glaucoma Subspecialty Day at the 2018 American Academy of Ophthalmology Annual Meeting in Chicago, USA. In a recent multi-centre retrospective study involving 445 patients, a model based on 24-hour CLS IOP patterns correlated more closely with actual progression than did Goldmann tonometry readings obtained in clinic in the same patient population (De Moraes CG et al. JAMA Ophthalmol 2018; 136(7):779085). The CLS model was based on 55 variables in IOP patters as measured by a Triggerfish CLS over a 24-hour period, said Dr Sit, of the Mayo Clinic, Rochester, Minnesota, USA. The resulting CLS model correlated with fast progression rates, defined as visual field mean deviation slopes >1 dB/year, at R2=30.1%. By comparison, Goldman mean IOP measured directly during office hours correlated at R2=27.8%. Other retrospective studies have linked greater variation in CLS-measured circadian IOP with faster progression in patients with closed-angle glaucoma (Tan et al. IOVS 2015) and normal tension glaucoma (Tojo N et al. J Glaucoma 2017; 26(3): 195–200). However, Dr Sit urged caution interpreting CLS-generated IOP data. CLS DOES NOT MEASURE IOP While specific 24-hour CLS patterns are associated with progression, reproducibility of such CLS patterns varies widely, reducing the predictive power of the model overall. One study comparing two sets of 24-hour CLS readings from the same patients on different days found overall reproducibility of R=0.59, ranging from 0.80 to 0.20 for individual patients (Mansour K et al. Arch Ophthalmol 2012 Aug 13: 1-6). Moreover, CLS measurements do not directly correlate to IOP measured by Goldmann applanation tonometry, Dr Sit said. That’s because CLS does not measure IOP, but rather changes in corneal curvature resulting from corneal distension related to IOP changes. Nonetheless, studies show the pattern of IOP and CLS measurement and the timing of peaks correlate strongly (Mansouri K et al. PLoS One 2015;10(5): e0125530). However, the amplitude of the two measures do not (Liu et al. PLoS One 2015 June 15;10(6): e0129529). While this suggests CLS and IOP patterns are not equivalent, CLS may yet prove useful, Dr Sit said. “To really understand where [CLS data] fits into our clinical armamentarium, we are probably going to need prospective data.” Arthur Sit: sit.arthur@mayo.edu