Accurate IOL calculations begin with the identification of the patient's visual goals. No one needs to explain this concept to a LASIK surgeon, but cataract surgeons all too often overlook this basic starting point. With patients' expectations continuing to increase, it is becoming more and more important to follow the lead of our refractive colleagues and take the time to agree upon a refractive goal with each patient prior to cataract surgery. It may come as a surprise that more commonly than expected your patients will give you an answer that is something different than emmetropia. For example, a patient of mine is a well-known chef who told me that his world exists mainly at arm's length. We agreed before surgery that -1.50D would be our refractive target. Conversely, for a pilot, the refractive goal would undoubtedly be plano. So, the first step in the process of IOL power calculations is to have a clear idea as to what your patient may prefer, especially if a specific occupational concern is important. This extra initial step has the potential to convert those patients who clearly know what they want into a small army of ambassadors for your practice.

COMPONENTS OF ACCURATE CALCULATIONS

For normal eyes, using the best aspects of today's technology makes it possible to consistently achieve highly accurate postoperative results. As always, however, the devil remains in the details, meaning that we have to execute all components of the exercise correctly. Patient selection, accurate keratometry, the method of biometry, the IOL power formula selected, and even the surgical technique all play important roles. To concentrate all of our attention on biometry is generally to miss the point. For example, if the keratometry is off by 0.75D, then the final postoperative refraction will be off by that same amount. If the SRK/T formula is used in the setting of high axial hyperopia, there will probably be an undercorrection. If the capsulorhexis is much larger than the optic of the IOL, a myopic shift may occur following contraction of the capsular bag. Finally, knowing when to repeat a measurement that does not fall within an established set of validation criteria is as important as knowing how to do it correctly in the first place. Highly accurate IOL power calculations are the result of a collection of many nuances, all linked together, each of which must be optimized.

AN EXCITING TIME

Refractive surprises have occurred ever since Sir Harold Ridley implanted the first IOL in 1949. With steadily progressing technological advances, the trend has been that the overall accuracy of our refractive outcomes has doubled every 5 to 10 years. With the introduction of the IOLMaster (Carl Zeiss Meditec Inc., Dublin, CA) in North America in 2000, refractive outcomes within 0.25D of the targeted refraction became a reality for the first time. Reaching this milestone will now allow us to set our sights on far more sophisticated endeavors, such as the correction of third- and fourth-order aberrations. We are clearly entering a very exciting time in the history of IOL power calculations.

There are several patient groups for whom it is not always possible to deliver on the promise of a highly accurate refractive result. At present, consistently accurate refractive outcomes for those with prior keratorefractive surgery, keratoconus, extreme axial myopia with posterior staphyloma, nanophthalmic eyes, or eyes with silicone oil remain elusive.

BASIC KERATOMETRY

When was the last time you calibrated all the keratometers in your office? As stated earlier, a keratometric error of 0.75D before surgery becomes a 0.75D error after surgery. If your office uses more than one keratometer or several different methods (eg, simulated keratometry, automated keratometry, and manual keratometry), multiple instruments will introduce another variable into the process. I strongly recommend assigning a single instrument that has been recently calibrated against a set of standard calibration spheres to the task of all pre- and postoperative keratometry.

It is also helpful for each office to establish a set of keratometry validation guidelines. In our office, if the Ks are very flat (less than 40.00D) or very steep (greater than 48.00D), a second person double checks the measurements and signs the chart. If the total keratometric power between eyes is greater than 1.50D, a second staff member repeats the measurements. If the mires are distorted, or the total astigmatism for either eye is greater than 4.00D, we will typically obtain a topographic axial map to screen for keratoconus. Lastly, if there are any difficulties obtaining the measurements that cannot be resolved, we ask the patient to return for repeat keratometry on another day.

BIOMETRY
There are currently four methods available for ophthalmic biometry: applanation; immersion A-scan; immersion A/B-scan; and partial coherence interferometry using the IOLMaster.
Surgeons interested in highly accurate outcomes have mostly abandoned applanation biometry. This technique yields a falsely short axial length due to variable amounts of corneal compression, is highly operator dependent, and often leads to corneal irritation.

Of the ultrasound-based biometry methods, immersion biometry is a much better choice. Even though it has the same 10-MHz resolution as the applanation method, it is much more consistent because there is no corneal compression and the measurement displayed is closer to the true axial length. Contrary to popular belief, the immersion technique is actually quite simple to perform, especially when used in conjunction with the Prager shell. Moreover, because immersion biometry is far more consistent than applanation biometry, it often takes less time.

The main limitation to accuracy with 10-MHz A-scan ultrasound is that it uses a relatively broad, low-resolution sound wave to measure the distance from the corneal vertex to the vitreoretinal interface. Adding to this is the fact that the region surrounding the fovea has a variable retinal thickness with the foveal center being thinner than the area immediately adjacent to it. Typically, both of these areas are included in an A-scan biometry measurement.

The most sophisticated form of ultrasound-based biometry is a combination immersion vector A/B-scan. By this technique, familiar to our retina colleagues, a horizontal immersion B-scan is carried out with a simultaneous vector A-scan that can be manually positioned to measure from the center of the corneal vertex to the location of the fovea.1 The disadvantages of A/B biometry are that the equipment is generally expensive and a high level of operator skill is required.

THE IOLMaster

In my opinion, the IOLMaster represents the single most important advance in IOL power calculations since the introduction of ultrasound biometry 3 decades ago. Interestingly, the technological basis of this instrument is principles laid down during the 19th century by the German-American physicist Albert Michelson.2 More than 100 years after its invention, the Michelson interferometer was introduced to ophthalmology via our colleagues in astronomy and physics. It is very likely that similar technological advancements will come to us from other unrelated disciplines and will have an equally important impact on ophthalmic surgery.

One of several reasons why the IOLMaster has a much higher resolution than ultrasound is that the axial-length measurement is based on a very short 780-nm light wave, rather than a much longer 10-MHz sound wave. By partial coherence interferometry, the instrument measures the distance from the corneal vertex to the retinal pigment epithelium (not affected by variations in retinal thickness) and then subtracts the foveal thickness. This approximation to an axial length by immersion ultrasound is based on a comparison to the exquisitely accurate Grieshaber Biometric System, an ultra high-resolution ultrasound biometer that employs four 40-MHz counters and is capable of an astonishing accuracy of 20µm.3 In essence, the IOLMaster is the equivalent of an upright, noncontact, immersion A-scan but with a fivefold increase in accuracy.4

There are four situations in which the IOLMaster is best suited for accurate biometry: (1) nanophthalmia or extreme axial hyperopia, because small errors in axial length are important; (2) extreme axial myopia, especially in the presence of a peripapillary posterior staphyloma; (3) prior retinal detachment with silicone oil; and (4) pseudophakia and phakic IOLs. Ophthalmologists are now starting to measure eyes that develop cataracts after phakic IOL implantation, and the IOLMaster can measure straight through the phakic IOL on the phakic setting with excellent results.5

When the IOLMaster debuted, it was presented mostly as a point-and-shoot device where the axial-length display with the highest signal-to-noise ratio was considered the best choice. Unfortunately, it is not quite that simple. Using the IOLMaster requires a correct interpretation of the axial-length display, with the signal-to-noise ratio a helpful, but not the most important, item in determining the overall quality of the measurement. Ideally, the axial-length display should have tall and slender primary maxima, with a thin and well-defined termination, much like the familiar silhouette of the Chrysler building in New York City.4 Careful attention to the axial-length display will avoid double peaks and other problems that could lead to potentially inaccurate measurements.

VALIDATION GUIDELINES FOR AXIAL LENGTH

If the preoperative refraction is equal between both eyes, but one eye measures 28mm and the other eye measures 26mm, something is obviously wrong. If a 27-mm axial length is displayed for a patient with a +4.00D refractive error, there is an apparent problem. As originally suggested by Holladay,6 it is important to follow a set of axial-length validation guidelines as the basis of a protocol to double check any measurements that may not seem right. In our office, if the difference between eyes is greater than 0.33mm, a second person independently verifies the results. If the axial length is less than 22mm or greater than 26mm, a second person repeats the measurements. We do likewise if the axial length correlates poorly with the refractive data or if there is any difficulty in obtaining consistent measurements.

IOL FORMULas

In North America, the three commonly used theoretical IOL power calculation formulas (Hoffer Q, Holladay 1, and SRK/T) are derived from the same mathematical backbone. The main difference between each of these third-generation, two-variable formulas is the way in which they calculate the final position of the IOL, commonly known as the effective thin-lens position.7

Limitations of all third-generation theoretic two-variable formulas are that they work best near schematic eye parameters, apply a number of broad assumptions to all eyes, and, apart from the lens constant, predict the final position of the optic of the IOL based solely on central corneal power and axial length. For example, some formulas assume that the anterior and posterior segments of the eye are mostly proportional, or that there is always the same relationship between central corneal power and the effective thin-lens position, which is not always true, especially in axial hyperopia.7,8

By the late 1980s, the Holladay 1 formula was available, which works well for eyes with normal and long axial lengths. This formula was followed in 1990 by the SRK/T formula, which works well for normal to moderately long axial lengths. Several years later, the Hoffer Q formula was added, which works well for eyes with short and normal axial lengths. Regression formulas like Binkhorst II, SRK I, and SRK II soon became of historical interest only. Interestingly, SRK II is still used by many in spite of its obvious limitations.

In 1991, Wolfgang Haigis, the head of the biometry Department of the University of Würzburg Eye Hospital in Germany, published the Haigis formula. Using the same mathematical backbone as other theoretic formulas, the Haigis formula approaches the problem of IOL power accuracy with three constants (a0, a1, and a2) and adds a measured anterior chamber depth for a third required variable. With the a0 constant optimized in a manner similarly to SRK/T, and the a1 and a2 constants based on schematic eye parameters, the formula performs similar to most third-generation two-variable formulas. When all three constants are optimized by regression analysis based on surgeon-specific IOL data, however, the range of the Haigis formula can be extended greatly to cover both high-axial hyperopia and high-axial myopia. The main limitation to using the Haigis formula for all axial lengths is that only Dr. Haigis or I presently carries out the required regression analysis, and a patient database of approximately 200 cases containing a wide range of axial lengths is required.

The Holladay 2 formula, available since 1998, is considered by many to be the most accurate of the theoretic formulas currently offered. The formula is easy to optimize and works well across a wide range of axial lengths. Its main limitation is that it requires the manual input of seven variables and is relatively expensive to purchase. Surgical practices serious about their refractive outcomes will typically use the Holladay 2 formula.

Because most biometry equipment already comes with several theoretic formulas, a simple rule to follow is to use the Holladay 1 formula for normal-to-long eyes and the Hoffer Q formula for normal-to-short eyes. However, it should eventually be the goal of every surgical practice to plan for the use of a more modern formula, such as Haigis or Holladay 2.

What are useful IOL power calculation validation guidelines? Of course, both eyes should be measured at the same time to serve as a basis for comparison. If the IOL power difference between eyes is greater than 1.00D, or there is any question about the accuracy of the axial length or keratometry, double check the results. Also, if the calculated IOL power does not match what you expected to see, such as a +28.00D IOL recommended for an axial myope, repeat the measurements.

SURGICAL TECHNIQUE

For highly accurate refractive outcomes, the capsulorhexis becomes the defining portion of the surgical procedure. Ideally, the capsulorhexis should be round, smaller than the optic, and centered (Figure 1). If carried out correctly, the optic of the lens implant should remain at the plane of the zonules. If the capsulorhexis is much larger than the optic, the forces of capsular bag contraction may shift the optic anteriorly, inducing a myopic shift late in the postoperative course. At the conclusion of the case, the optic of the IOL should be centered directly beneath the capsulorhexis so that the capsular bag can uniformly shrinkwrap around it (Figure 2). This approach is another important step for ensuring consistent outcomes.

SUMMARY

Understand the current limitations of technology; there are some patients whom you cannot promise a highly accurate outcome. Assign a single instrument for the task of keratometry for added consistency. Use the IOLMaster or immersion biometry rather than an applanation technique. Develop a set of validation criteria for each part of the measurement process and have a second person repeat any part that falls outside of these guidelines. Consider switching to one of the newer IOL power calculation formulas for improved accuracy. Optimize your surgical technique by making the capsulorrhexis round, smaller than the optic, and centered. By embracing current technology and paying careful attention to details, highly accurate refractive outcomes are an achievable goal for every ophthalmologist.

Warren E. Hill, MD, FACS, is Medical Director of East Valley Ophthalmology in Mesa, Arizona. He is a consultant for Alcon Laboratories, Inc., and Carl Zeiss Meditec Inc. Dr. Hill may be reached at (480) 981-6111; hill@doctor-hill.com.
1. Hill WE, Byrne SF. Complex axial length measurements & Unusual IOL Power Calculations. In: Focal Points: Clinical Modules for Ophthalmologists. San Francisco: American Academy of Ophthalmology; 2004; 22: 9.
2. Hill WE. The IOLMaster. Techniques in Ophthalmology. 2003;1:1:62.
3. Haigis W, Lege B, Miller N, Schneider B. Comparison of immersion ultrasound biometry and partial coherence interferometry for intraocular lens calculation according to Haigis. Graefes Arch Clin Exp Ophthalmol. 2000;238:765-773.
4. Vogel A, Dick B, Krummenauer F. Reproducibility of optical biometry using partial coherence interferometry. Intraobserver and interobserver reliability. J Cataract Refract Surg. 2001;27:1961-1968.
5. Salz JJ, Neuhann T, Trindade F, et al. Consultation section: cataract surgical problem. J Cataract Refract Surg. 2003;29:1058-1063.
6. Holladay JT, Prager TC, Chandler TY, et al. A three part system for refining intraocular lens power calculations. J Cataract Refract Surg. 1988;14:17-24.
7. Holladay JT. Standardizing constants for ultrasonic biometry, keratometry, and intraocular lens power calculations. J Cataract Refract Surg. 1997;23:1356-1370.
8. Holladay JT, Gills JP, Leidlen J, Cherchio M. Achieving emmetropia in extremely short eyes with two piggyback posterior chamber intraocular lenses. Ophthalmology. 1996;103:1118-1123.