EyeWorld Asia-Pacific March 2017 Issue

March 2017 EWAP FEATURE 11 continued on page 15 biggest mistake you can make,” Dr. Hoffer said, adding that if you don’t have an optical biometer and must use ultrasound for this measurement, do an immersion rather than an applanation. With several optical biometers on the market, do any stand out? Dr. Hoffer said he has compared a variety of machines, including the LENSTAR (Haag-Streit, Köniz, Switzerland), IOLMaster 500 and 700 (Carl Zeiss Meditec, Jena, Germany), Aladdin (Topcon Medical Systems, Oakland, New Jersey), and AL-Scan (Nidek, Fremont, California). “They’re all good. You can’t go wrong with any of these optical biometers, and I’m probably one of the few people who has done hands-on comparisons and published the results,” Dr. Hoffer said. Dr. Holladay agreed that newer optical biometers are equally accurate and noted that they have measures like standard deviations and signal-to-noise ratios to confirm the validity of the measurement. In addition to this advice, Dr. Holladay said cataract surgeons should not use any drops or measure IOP before performing keratometry. “It changes the Ks and leads to variability in measurements,” he said. Dr. Hoffer personally advised against using what he called the outdated and error-prone regression formulas like the SRK 1 or SRK 2. Between these two surgeons, the jury is still out on some of the favored formulas. In short eyes (those less than 22 mm), Dr. Hoffer said he prefers his own formula, the Hoffer Q. In eyes 24.5 to 26 mm, Dr. Hoffer said he thinks the Holladay 1 is best, while in eyes greater than 26 mm, he said the SRK/T seems to work well. For eyes between 22 and 24.5 mm, Dr. Hoffer thinks the Holladay 1 and Hoffer Q are on par with each other. The Haigis formula, Dr. Hoffer said, includes triple personalization of three different lens constants (a0, a1, and a2). Dr. Hoffer said this optimization with data from at least 100 eyes should be done by Wolfgang Haigis, PhD, Wurzburg, Germany, through his website and not through others who purport to do it on the internet. “Then you’re dealing with an extremely accurate formula in almost all ranges—short eyes, medium eyes, and long eyes,” he said, noting that it replaces the K with the preop ACD to predict ELP. Dr. Hoffer acknowledged other formulas—the Olsen, Holladay 2, which he said he wouldn’t recommend, and the Barrett Universal II. He also mentioned the new Hoffer H-5, which takes into account the effect of gender and race on IOL power calculations. Dr. Hoffer said a soon-to-be- ublished study, which included 10,000 eyes—2,770 of which were within a quarter diopter of prediction looked at the median absolute error (MedAE) of the Hoffer Q, Holladay 1, SRK/T, and the Hoffer H-5 and found the latter had the lowest MedAE, being 30–40% more accurate. Dr. Holladay said he would recommend recent formulas like the Holladay 2, Olsen 2, and Barrett Universal II. Older formulas, he said, are less accurate. He also recommended taking between five and seven measurements of the eye to predict the effective lens position. Dr. Hoffer noted that the Olsen 2 and Barrett are relatively recent, but the Holladay 2 is now celebrating its 20th birthday. A study of more than 3,000 eyes published in the Journal of Cataract & Refractive Surgery in 2016 compared seven IOL power formulas. 2 IOLMaster 500 and optimized lens constants were used, and other measures were taken to reduce variability. Of the formulas tested—the Barrett Universal II, Haigis, Hoffer Q, Holladay 1, Holladay 2, SRK/T, and T2—the Barrett Universal II was the most accurate over the whole axial length range. Of the 156 eyes in the study that had an axial length of 22 mm or less, the Holladay 1 formula had the lowest mean absolute prediction error (MAE), followed by the SRK/T, T2, Holladay 2, and Barrett Universal II. The study authors reported the Haigis and Hoffer Q had the highest MAEs in this group, but noted that the differences between the formulas were not statistically significant. Of the 2,638 eyes between 22 mm and 24.5 mm in the study, the Barrett Universal II was found to be the most accurate with a statistically significant lower MAE compared to the other formulas. The Holladay 1 was found to be the next most accurate formula, but its difference was not statistically significant compared to the T2 or SRK/T formulas. The study authors wrote that the difference was statistically significant when compared to the Holladay 2, Haigis, and Hoffer Q. The Barrett Universal II also had the lowest MAE in the medium-long eye group in this study (372 eyes), followed by the T2 and Holladay 1. The Barrett Universal II, SRK/T, T2, Haigis, Holladay 2, Holladay 1, and Hoffer Q, respectively, were found the most accurate in predicting IOL power for the 77 eyes that were 26 mm or greater. “[…] to our knowledge, [this] is the only study to assess 7 of the most commonly used IOL power formulas across the entire [axial length] range using a large enough patient cohort to detect differences between formulas,” Kane et al. wrote of the study’s significance. “[…] our study of 2,341 patients provides strong evidence that the Barrett Universal II formula is able to more accurately predict the postoperative refraction for eyes with a medium, medium-long, or long [axial length] compared with the 6 other IOL power formulas,” Kane et al. concluded. “Although none of the formulas performed poorly, the marked difference between the Barrett Universal II formula and the other IOL power formulas for eyes with an [axial length] longer than 22.0 mm should prompt serious consideration of its use in these eyes by cataract surgeons worldwide.” Commenting on this study, Dr. Hoffer said, “Their method of comparing MAEs is inappropriate since absolute errors are not a normal Gaussian distribution. The authors should have compared the MedAE or performed