INTRODUCTION: The aim of the study was to evaluate clinical and biometric factors leading to a prediction error related to lens position in pars plana vitrectomy. METHODS: This study was conducted as a consecutive retrospective case series at the Department of Ophthalmology, Montpellier University Hospital. All medical files and PCI biometrical reports from a single surgeon were reviewed from 2017 to 2019. Patients who had phacoemulsification with the ASPHINA 509 MP® intraocular lens were selected and stratified into 3 groups: phacoemulsification alone (group 1), phacoemulsification and vitrectomy with gas tamponade (group 2), and phacoemulsification and vitrectomy without tamponade (group 3). Clinical factors and biometry factors from initial and final biometry were collected. Refractive error, actual lens position, C constant, axial length delta, and pre-operative and post-operative anterior and posterior segment variation parameters were calculated. RESULTS: A total of 140 eyes were analyzed, 90 in group 1, and 25 in group 2 and 3. The mean prediction error was 0.10 ± 0.55 D (group 1)
-0.36 ± 0.74 D (group 2)
and -0.12 ± 0.54 D (group 3) with p <
0.05 for group 1 vs. group 2. The mean actual lens position was 5.25 ± 0.29 mm
5.66 ± 0.60 mm
and 5.50 ± 0.43 mm for the 3 groups, respectively (p <
0.001). Axial length delta was -0.10 ± 0.13 mm in group 1, -0.062 ± 0.20 mm in group 2, and -0.022 ± 0.17 mm in group 3 (p = 0.015). Multilinear regression analysis found a significant and independent influence of vitrectomy and gas tamponade on prediction error. CONCLUSION: Myopic shift in the case of vitrectomy is multifactorial, effective lens position is modified by vitrectomy and vitreous refractive index is changing. The integration of these data in formulas may improve refractive outcome after cataract and vitrectomy surgery.