Transient Corneal Response After Goldmann Three-Mirror Gonioscopy and Its Limited Effect on Optical Biometry and Calculated Intraocular Lens Power: A Prospective Within-Subject Controlled Observational Study

Background and Objectives: Contact gonioscopy and optical biometry often happen at the same preoperative visit, with no defined interval between them. We asked whether uncomplicated Goldmann three-mirror gonioscopy was followed by short-term changes in optical biometry and calculated intraocular lens (IOL) power and whether any mean change could be positively affirmed as small enough not to matter clinically. Materials and Methods: Sixty-four cataract patients (128 eyes) underwent NIDEK AL-Scan biometry at baseline and 5, 15 and 30 min after unilateral Goldmann three-mirror gonioscopy in this prospective repeated-measures study (2015–2016). The untreated fellow eye served as an internal control. Our primary analysis was a within-patient paired difference-in-differences (DiD) comparison, which modeled the paired-eye structure directly; nested patient/eye mixed-effects models followed as a sensitivity analysis. We recomputed IOL power as a continuous variable for four formulas, SRK/T, Haigis, Hoffer Q and Holladay 1, using the published constants for the implanted lens. Equivalence within ±0.25 D was tested by two one-sided tests (TOSTs), with Holm adjustment across the twelve tests performed. Results: Central corneal thickness (CCT) rose by 3.14 µm at 5 min (95% CI, 1.94 to 4.35; p < 0.001) and by 1.66 µm at 15 min (p = 0.015), returning to baseline by 30 min. That signal sat inside the ±6.2 µm short-term repeatability seen in untreated control eyes. Keratometry moved by no more than 0.11 D, and all four IOL formulas behaved almost identically, the largest mean difference being −0.18 D at 15 min. Equivalence within ±0.25 D was confirmed in unadjusted testing at 5 and 30 min for all four formulas and at 15 min for SRK/T; after Holm adjustment only the four 30 min conclusions and the 5 min conclusion for SRK/T were retained. The selected 0.5 D lens power changed in 33–47% of study eyes and in a comparable 31–52% of untreated control eyes. Conclusions: Procedure-related differences in calculated IOL power were small throughout. The largest was −0.18 D, inside the ±0.25 D margin and well below the 0.5 D interval on which lenses are supplied. Demonstrated equivalence was a narrower claim. After adjustment for multiplicity, it held at 30 min for all four formulas and at 5 min for SRK/T, and at 15 min it was not demonstrated for Haigis, Hoffer Q or Holladay 1. These data therefore do not establish that biometry is unaffected across the whole first half hour, and we make no recommendation about deferring it. The transient CCT signal fell within device repeatability. Step-level changes in selected lens power were as common in eyes that never underwent gonioscopy. All analyses were exploratory and limited to a partial-coherence-class biometer of the study period and to the four formulas tested.