Similarity in corneal biomechanics between 2 major glaucoma types: More alike than expected?

Corneal biomechanical properties measured with corneal visualisation Scheimpflug technology (Corvis ST; Oculus Optikgeräte GmbH) differ between primary open-angle glaucoma (POAG) and primary angle-closure glaucoma (PACG). However, Corvis ST (CST) parameters are influenced by age, sex, axial length (AL), intraocular pressure (IOP) and central corneal thickness (CCT)—factors that also differ between glaucoma subtypes—raising the question of whether previously reported differences reflect true pathophysiological variation or baseline confounding.¹

A retrospective cross-sectional study published in Scientific Reports by Nakaniida and colleagues at Hiroshima University and collaborating institutions in Japan addresses this question using propensity score matching and linear mixed-effects model (LMM) analysis adjusted for key covariates, finding no statistically significant differences in CST parameters between POAG and PACG eyes after correction for multiple comparisons.¹

Background and rationale

In a previous comparison of POAG and PACG eyes by the same group, significant differences were identified in five CST parameters. However, the PACG group included a higher proportion of women and had a significantly shorter AL, and all five parameters were significantly correlated with AL. This raised uncertainty as to whether the observed differences reflected glaucoma subtype-specific biomechanics or were attributable to baseline imbalances. The present study was designed to clarify this by applying propensity score matching based on age, sex, AL, IOP and CCT, followed by covariate-adjusted LMM analyses and correction for multiple comparisons using the Benjamini–Hochberg procedure.¹

Study design and methods

The study included 48 eyes with POAG and 48 eyes with PACG, matched 1:1 from 144 potentially eligible POAG eyes and 92 potentially eligible PACG eyes. Data were collected at Hiroshima University Hospital, Saneikai Tsukazaki Hospital and Seirei Hamamatsu General Hospital between August 2019 and December 2021. All eyes were phakic; eyes with a history of cataract surgery or pseudophakia were excluded. Diagnoses were made according to the Japan Glaucoma Society Guidelines for Glaucoma (fifth edition).¹

CST examinations were performed three times per eye on the same day using software v1.6r2031, with the mean of three quality-checked readings used for each parameter. Twelve CST parameters were evaluated: A1 time, A1 velocity, A1 length, A1 deformation amplitude, A2 time, A2 velocity, A2 length, A2 deformation amplitude, highest concavity (HC) deformation amplitude, peak distance, radius and whole-eye movement. AL was measured using the IOLMaster 700 (Carl Zeiss Meditec), IOP by Goldmann applanation tonometry and CCT by the CST device.¹

Propensity scores were estimated using logistic regression including age, sex, IOP, AL and CCT. One-to-one nearest-neighbour matching without replacement was applied on the logit of the propensity score, with a caliper width of 0.2 standard deviations. Twelve CST parameters were then compared using LMMs adjusted for age, sex, AL, IOP and CCT, with subject ID as a random intercept to account for inter-eye correlation. P-values for the group effect were adjusted using the Benjamini–Hochberg false discovery rate (FDR) procedure.¹

Results

After propensity score matching, standardised mean differences (SMDs) were substantially reduced for age, sex, IOP and CCT, all falling below 0.10. Residual imbalance remained for AL (SMD 0.19) and propensity score distance (SMD 0.19). No significant differences were observed between matched groups in age, sex, AL, IOP or CCT (all Q ≥ 0.507).¹

Comparing the 12 CST parameters, PACG eyes showed longer A2 time (P = .013, Q = 0.099), greater A2 deformation amplitude (P = .025, Q = 0.099) and greater whole-eye movement (P = .019, Q = 0.099) compared with POAG eyes on unadjusted testing. However, none of these differences remained statistically significant after correction for multiple comparisons using the Benjamini–Hochberg method. No significant differences were observed in A1 time, A1 velocity, A1 length, A1 deformation amplitude, A2 velocity, A2 length, HC deformation amplitude, peak distance or radius (all Q ≥ 0.143).¹

Discussion

The authors note that the three parameters showing uncorrected trends—A2 time, A2 deformation amplitude and whole-eye movement—relate to the corneal recovery phase and whole-globe responses to the air-puff stimulus. Longer A2 time may reflect delayed corneal recovery; greater A2 deformation amplitude may indicate a larger deformation response during the second applanation phase; and whole-eye movement may reflect not only corneal properties but also ocular rigidity, scleral biomechanics and orbital tissue compliance. According to Nakaniida et al, these uncorrected trends toward greater indentation-related responses in PACG eyes should be regarded as exploratory and hypothesis-generating, given that they did not survive FDR correction and that PACG-related anatomical variables such as anterior chamber depth and lens vault were not available for inclusion in the models.¹

The authors also note that corneal hysteresis (CH), measured using the Ocular Response Analyzer (Reichert Ophthalmic Instruments), has generally shown no significant differences between POAG and PACG in prior studies, suggesting that subtype-specific biomechanical variations may not be fully captured by CH alone.¹

Limitations identified by the authors include the relatively small sample size and consequent limited statistical power after covariate adjustment and multiple-comparison correction; retrospective design with potential selection bias, as PACG eyes were referred for laser or surgical treatment; absence of refractive error data; unavailability of standardised glaucoma severity indices at the time of CST measurement; absence of PACG-related anatomical parameters; and lack of data on topical antiglaucoma medications, which may influence corneal biomechanical properties. The authors call for future prospective studies incorporating standardised visual field, OCT, anterior segment imaging and medication data.¹

Nakaniida et al conclude that, after accounting for age, sex, AL, IOP and CCT through propensity score matching and covariate adjustment, no statistically significant differences in CST parameters were demonstrated between POAG and PACG eyes.¹

Reference

1. Nakaniida Y, Tokumo K, Nakakura S, Asaoka R, Kiuchi Y, Sakaguchi H. Corvis ST parameters in primary open angle and angle closure glaucoma after propensity score matching. Sci Rep. Published online June 11, 2026. doi:10.1038/s41598-026-57462-w