Purpose To develop algorithms for preoperative estimation of the true postoperative intraocular lens (IOL) position to be used for IOL power calculation.

Setting Moorfields Eye Hospital NHS Foundation Trust, London, UK.

Methods Fifty patients were implanted randomly with a 3-piece IOL model in one eye and a 1-piece model in the other eye. Preoperatively, the IOLMaster was used to determine axial length, anterior chamber depth and mean corneal radius. Lens thickness and corneal width were measured with the ACMaster. Postoperative IOL position was measured with the ACMaster. Partial least squares (PLS) regression analysis of IOL position in terms of preoperative parameters was performed with a commercially available software package.

Results The PLS regression analysis showed that age, refraction, corneal width, lens thickness and corneal radius are not significant predictors of postoperative position of the anterior IOL surface, while axial length and in particular anterior chamber depth are. Regression relationships in terms of the above-mentioned predictors were determined for the two models implanted. Surprisingly, it turned out that the position of the posterior IOL surface could be described by a single regression relationship valid for both models. The residual SD for prediction of IOL position was about 0.17 mm for all relationships.

Conclusions Accurate relationships to determine the true postoperative IOL position were obtained. In addition to axial length and corneal radius, which are required for the IOL power calculation as such, they require measurement of preoperative anterior chamber depth only.