Visual, refractive and patient-reported outcomes after two trifocal intraocular lens designs: a 12-month comparative cohort study

Trifocal intraocular lenses differ in how their optical architecture apportions incident light between foci, but comparative evidence combining acuity, defocus testing, contrast sensitivity and patient-reported outcomes in large routine-care cohorts is limited. We compared 12-month visual, refractive and patient-reported outcomes after implantation of two trifocal intraocular lens designs. This retrospective comparative cohort screened 432 patients and analysed one right eye from each of 340 patients (170 per group) implanted with a sinusoidal pupil-adaptive or a conventional diffractive trifocal lens. Allocation was non-randomised and individual selection reasons were unavailable. Primary outcomes, selected a priori before access to the study records, were 12-month uncorrected intermediate acuity at 65 cm and uncorrected near acuity at 40 cm. Assessment covered distance acuity, refractive accuracy, defocus curves, contrast sensitivity, dysphotopsia, spectacle independence and the National Eye Institute Visual Function Questionnaire-25. Models adjusted for baseline and operative-period differences, with propensity-score sensitivity analyses. At 12 months the sinusoidal design showed better intermediate acuity (0.09 (0.05) vs. 0.12 (0.05) logMAR; adjusted difference − 0.025, 95% CI − 0.037 to − 0.013) and near acuity (0.10 (0.06) vs. 0.16 (0.05) logMAR; −0.066, 95% CI − 0.080 to − 0.053; both P < 0.001). Propensity-weighted and doubly robust estimates were almost identical (near-acuity E-value 5.4). Refractive predictability was equivalent (83.5% vs. 84.7% within ± 0.50 D; P = 0.88); all eyes were within ± 1.00 D. In exploratory unadjusted analyses, defocus area-under-the-curve, contrast sensitivity at most tested spatial frequencies, halo, glare and questionnaire outcomes also favoured the sinusoidal design; distance-acuity differences were clinically negligible. Both designs achieved excellent distance vision and refractive predictability. The sinusoidal design was associated with better intermediate and near function across several independent measures, and the estimates were robust to propensity-score weighting. Because allocation was not randomised and the groups differed at baseline, residual confounding cannot be excluded and prospective confirmation is required.