From a large refractive outcomes database the chart below reflects our experience with several popular formulas.

For axial lengths from 22.00 mm to 26.00 mm, and central corneal powers ranging from 40.00 D to 47.00 D, almost any modern IOL power calculation formula will give very good outcomes. However, for eyes outside this range, our results indicate that the Holladay 2, or the Haigis formula (with properly optimized a0, a1 and a2 constants) are better choices.

As presented by Dr. Wolfgang Haigis, at the 2001 San Diego ASCRS meeting, the individual geometry of each intraocular lens plays a role in determining which formula works best. For example, with the Alcon MA-series 3-piece intraocular lenses, we have found that the Holladay 1, Holladay 2 and Haigis formulas give the optimum results. For other types of intraocular lenses, SRK/T, or Hoffer Q, may offer better outcomes, depending on the axial length and the individual geometry of the particular lens used. This small advantage of one formula over another is the result of the power prediction curve of, say the Hoffer Q formula being a better match for the AMO SN40 3-piece lens at higher IOL powers than the Holladay 1 formula.

In the partially optimized form (IOLMaster software), with the a0 constant optimized and the a1 and a2 constants set at default values of 0.4 and 0.1 respectively, the Haigis formula performs well for eyes of normal axial lengths and relatively short axial lengths. When the a0, a1 and a2 Haigis constants have been fully by multi-variable regression analysis, the Haigis formula performs well across a very wide range of axial lengths. Please visit our Download Area for instructions and an Excel spreadsheet that will allow you to carry out this three variable optimization for the Haigis formula.

Worth repeating, our recommendation is that you carefully track your outcomes, using several formulas for each IOL. The accuracy of these formulas is noticeably increased when the A-constant, ACD, or Surgeon Factor are "personalized." For this reason, maintaining, and systematically reviewing, an Outcomes Database should be an essential part of your practice, and well worth the extra effort.

Utilizing a fully optimized version of the Haigis formula, or the Holladay 2 formula, in conjunction with the Zeiss IOL Master, our mean absolute post-operative prediction error runs slightly better than ±0.25 D. Given the fact that IOLs presently come in 0.50 D steps, these results approach the theoretic limit of the exercise.

The chart below represents our experience with the Alcon SA60AT and the Alcon MA60MA intraocular lenses, and is displayed in terms of the best possible mean absolute prediction error. These data represent outcomes for patients with keratometry below 50.0 D and above 40.0 D and excludes cases of keratoconus, or prior keratorefractive surgery.