Skip to main content

  • Predicting postoperative refraction: Which formula is best?

    Cataract/Anterior Segment, Refractive Mgmt/Intervention

    Add to My Bookmarks

    This retrospective case series compared the accuracy of 7 popular IOL formulas for predicting postoperative refraction in eyes of all ocular dimensions

    Study design

    The analysis included 18,501 eyes receiving an Alcon monofocal IOL (SN60WF or SA60AT IOL) over an 18-month period. All eyes had biometry with the Lenstar 900 optical biometer.

    Investigators analyzed the error in predicted spherical equivalent of each formula (Barrett Universal II, Haigis, Hoffer Q, Holladay 1, Holladay 2, Olsen, and SRK/T), including the Wang-Koch adjustment (Holladay 1, Hoffer Q, Haigis, SRK/T) for eyes with axial length longer than 25 mm.

    Outcomes

    For the SN60WF, the Barrett Universal II had the significantly lowest mean absolute error followed by the Olsen, Haigis 1, Holladay 1 Wang-Koch, Holladay 2, Holladay 1, Hoffer Q Wang-Koch, SRK/T, SRK/T Wang-Koch, Hoffer Q, and Haigis Wang-Koch. The results were similar for the SA60AT IOL.

    For long eyes (>25.5mm), Olsen had the lowest mean error followed by Barrett, whereas Haigis Wang-Koch, Hoffer Q, and Holladay 1 were worst. For short eyes (<22.5mm), Barrett and Holladay 1 Wang-Koch were most accurate, and Hoffer Q, Holladay 2 and Hoffer Q Wang-Koch were least accurate.

    Of note, all 7 formulas without the WK adjustment give results that within 0.1 D of predicted spherical equivalent for eyes with axial lengths between 23 and 25 mm. The major reason for the difference between the formulas is their performance outside this range.

    Limitations

    These results may not apply to other IOL models and manufacturers. Due to a retrospectively optimized Haigis a0 constant and a large number of surgeons in the study (127 for SN60WF and 95 for SA60AT), the results may not hold true for individual surgeons. A significant number of patients (6,476) were excluded due to missing postoperative refractive information in the 2-week to 4-month postoperative period, possibly skewing results.

    Clinical significance

    Findings from this large study suggest the Barrett Universal II is the best overall formula, and had a high accuracy for both short and long eyes. The Hoffer Q was least accurate for short eyes. In general, applying the Wang-Koch modification of axial length for long eyes resulted in a shift from hyperopic to myopic outcomes.