Related Experiment Video
Updated: Oct 26, 2026
Rotating the Intraocular Lens to Prevent Posterior Capsular Opacification in Cataract Surgeries
Published on: July 7, 2023
Comparison of the Moussa Intraocular Lens Formula to contemporary formulas in a multicenter study
Peter Kiraly1, Hisham Hamze2, Soon Ch'ng2
1Manchester Royal Eye Hospital, Manchester, Oxford Road, M13 9WL, U.K.
Purpose:
To evaluate the refractive predictive accuracy of the Moussa intraocular lens (IOL) formula compared with Kane, EVO, PEARL-DGS, K6, Hoffer-QST, and Barrett Universal II (BUII).
Setting:
Birmingham Midland Eye Centre (Birmingham), Manchester Royal Eye Hospital and Newmedica (Manchester), UK.
Design:
Multicenter, retrospective, consecutive, nonrandomized observational case series (September 2022-June 2025).
Methods:
Included were one eye per patient undergoing uncomplicated phacoemulsification. Preoperative biometry used the IOLMaster 700, Lenstar LS900, or Pentacam AXL Wave. Lens-constants were optimized to zero mean prediction error (PE) for each formula and lens-biometer combination. Primary outcome was mean absolute error (MAE), evaluated using heteroscedastic studentized bootstrap analysis (2000 iterations; Holm correction). Secondary outcomes included median absolute error, root-mean-square error, standard deviation of PE, and proportions of eyes within ±0.25diopter (D) and ±0.50D.
Results:
Included were 2550 patients (2550 eyes; mean age 74.0±10.0 years; 1068 male patients [41.9%]). Moussa produced a lower MAE (0.348 D) than Kane (0.359D), EVO (0.363D), PEARL-DGS (0.367D), K6 (0.367D), Hoffer-QST (0.377D), and BUII (0.379D; all p<.001). In short eyes (≤22.0mm), Moussa (0.412D) demonstrated lower error than K6 (p<.001), BUII (p<.001), Hoffer-QST (p=.008), EVO (p=.008), PEARL-DGS (p=.014), and Kane (p=.039). In long eyes (≥26.0 mm), Moussa (0.342 D) had lower error than Kane, BUII, Hoffer-QST (all p<.001), and K6 (p=.045), but not EVO (p=.068) or PEARL-DGS (p=.202).
Conclusions:
The Moussa formula demonstrated high predictive accuracy across multiple biometry platforms, with statistically lower overall MAE than the comparator formulas. Absolute differences from the closest-performing formulas were modest, and ±0.50D outcomes were similar to Kane and EVO. These findings support Moussa as a clinically relevant contemporary option for IOL power calculation, with encouraging results in short and long eyes requiring further targeted validation.