gms | German Medical Science

Measuring accuracy has been improved, and formulae have been developed in "generations" for IOL power selection, but nevertheless, the di erences between calculated and manifest refraction could not be reduced below recogition threshold yet.

Before a further improvement of prediction accuracy is possible, the single error contributions must be made more transparent. There are four independent error sources not influencing each other: measuringerrors of axial length and corneal radius, estimation errors of postopertive IOL position, manufacturing errors of the IOL and calculation errors.

Measuring errors of the axial length will be unavoidable also in future due to missing information of the material properties of the crystalline lens which is altered from cataract. In ultrasound measurements, the sound velocity is uncertain, in laser interferometry it is the index of refraction. Corneal radii can be measured by di erent methods with suffcient accuracy in normal eyes. For odd eyes, e.g. after corneal refractive surgery, this is possible only if the radii and the eccentricity are extracted from topography by a complicated three-dimensional model of the whole cornea. The accuracy of the estimation of postoperative IOL position will always be limited by individual di erences in the situation of the capsular bag. However, at least in mean a high accuracy between prediction and and postoperative measurement could be established. According to DIN and ISO standards, IOL manufacturing errors are higher than normally assumed and are caused by measurement and quality management problems which are often overlooked from ophthalmologists. Currently, they are the second largest error source after axial length measurement errors. Calculation errors can be completely avoided if analytical formulae are replaced by numerical ray-tracing methods. Gaussian optics and the analytical formulae based theron is a poor approximation of the optical properties of the human eye. It has caused a lot of misleading approaches and misunderstandings in past.