Eyeworld

EW RESIDENTS 162 April 2017 ORA and the AL-optimized Holladay 1 in axial hyperopia and keratoco- nus. Other patient categories in need of further validation should include eyes of all axial lengths with mature cataracts where pre-operative optical biometry is not feasible and ultra- sound biometry, whether contact or immersion, is the default measure- ment and potentially less accurate or more cumbersome. The ultrasound scan may measure the anatomical axial length and not necessarily the optical axial length especially with poor fixation due to mature cataracts and staphylomas due to myopia. The axial length range in this study did not include pathologic myopia, as the range of included eyes was from 25 to 29.5 mm. Pathologic my- opia can be fraught with even great- er inaccuracy in IOL power predic- tions and increased risk of hyperopic surprise. 2 This category of patients with pathologic myopia may be rare but would be most likely to benefit from more accurate predictions of IOL power, and should be included in future studies. Another study limitation was that this was a retrospective, small study conducted at a single center using data from a single surgeon. However, the data is thought-pro- voking and could serve as a pilot for either an equivalence or non-infe- riority cut-off when comparing any of these newer methods/formulas to the older generation formulas. Thus, future larger studies designed as a randomized control trial would provide further insight into the accuracy of the ORA and Hill-RBF compared to older formulas. EW References 1. Huelle JO, Katz T, Druchkiv V, Pahlitzsch M, Steinberg J, Richard G, Linke SJ. First clinical results on the feasibility, quality and repro- ducibility of aberrometry-based intraopera- tive refraction during cataract surgery. Br J Ophthalmol. 2014 Nov;98(11):1484–91. 2. Kapamajian MA, Miller KM. Efficacy and safety of cataract extraction with negative power intraocular lens implantation. Open Ophthalmol J. 2008 Feb 15;2:15–9. Contact information Challa: pratap.challa@duke.edu office. We cannot ascertain the total number of optometrists involved or how reproducible their results were. Results One of the primary outcomes was a hyperopic outcome, which the authors defined as "any outcome that is hyperopic relative to the SE and is not necessarily a positive SE." While this definition makes sense when trying to assess how close predictions are to an outcome, the goal was emmetropia for all enrolled subjects. Ultimately, patients may be quite happy with a SE that is more hyperopic than the predicted out- come as long as it is closer to plano and not truly hyperopic (i.e., a pos- itive SE). Thus, we believe that true hyperopia (i.e., a positive SE) should be included as another secondary outcome. The results Table 1 did not include any specific demographic information (i.e., age, gender, race) about the study population from which the eyes were enrolled. The subsequent Tables 2–5, demonstrat- ed many statistically significant p-values (p < 0.05). However, the fact that so many statistical tests were run increased the statistical likeli- hood of finding significant p-values due to chance. It would be prefera- ble to limit the number of statistical tests run and to emphasize the key questions of the paper, which was the comparison of the third and fourth generation formulas against either the ORA or the Hill-RBF. The 95% confidence intervals would also be helpful given the sample size. Finally, the tables did not present any point estimates for the statistical tests. For example, in the analysis of the proportion of "0.5 D of pre- dicted, within 1.0 D of predicted, and with hyperopic outcomes," the authors stated that they ran logistic regression. However, odds ratios were not reported and should be included. The authors concluded that the mean and median numerical errors in the total analysis were similar to the final results of a subgroup analysis in which toric lenses were excluded. When the authors eval- uated the prediction methods for the sample without toric IOLs, the number of eyes decreased to 38. The results generally seemed similar, the loss of 13 eyes also impacted the power of the study. We think it would be preferable to use multivar- iate analysis to adjust for the type of IOL used, so that the study results are less underpowered. Discussion The authors concluded that the ORA was "as effective" as the AL-opti- mized Holladay 1. However, the data suggests that the AL-optimized Holladay 1 produced significantly lower MNEs than the ORA. The take- away should be that the Wang-Koch adjustment is still easier, cheaper, more accessible, and more effective as using the ORA. Another highlight of this study is that it is the first to compare the results of the Hill-RBF formula in axial myopes. The merits of intraoperative aberrometry have been previously validated in post-refractive surgery patients. Additionally, the ORA can guide the positioning of the toric IOL implants and, therefore, it is useful for improving the surgeons' confidence with toric lenses. We agree with this study's recommenda- tion to also evaluate the merit of the Review continued from page 161 Intraoperative Aberrometry versus Preoperative Biometry for Intraocular Lens Power Selection in Axial Myopes Intraoperative aberrometry effectively targets refractive outcomes in axial myopes Darren C. Hill, MD, Shruti Sudhakar, BS, Christopher S. Hill, BS, Tonya S. King, PhD, Ingrid U. Scott, MD, MPH, Brett B. Ernst, MD, Seth M. Pantanelli, MD J Cataract Refract Surg. 2017;43(4). Article in press Purpose: To compare accuracy of the new Hill-RBF formula and intraoperative aberrometer (ORA) to other formulas based on preoperative biometry in predicting residual refractive error after cataract surgery in axial myopes. Setting: Suburban private practice Design: Retrospective consecutive case series Methods: Fifty-one eyes with axial length (AL) > 25.0 mm underwent cataract extraction with intraocular lens (IOL) implantation. For each eye, the 1-center Wang-Koch AL-optimized Holladay 1 formula was used to select an IOL targeting emmetropia. Residual refractive error was predicted pre-operatively using SRK/T, Holladay 1, Holladay 2, and Barrett Universal II, Hill-RBF and intraoperatively using the ORA. Postoperative refraction was compared to the preoperative and intraoperative prediction methods. Results: Mean numerical errors (MNE±STDERR) associated with using the SRK/T, Holladay 1, AL-optimized Holladay 1, Holladay 2, Barrett Universal II, Hill-RBF, and ORA were 0.20 ± 0.06, 0.33 ± 0.06, -0.02 ± 0.06, 0.24 ± 0.06, 0.19 ± 0.06, 0.22 ± 0.06, and 0.056 ± 0.06 D, respectively (p < 0.001). The proportion of patients within 0.5 D of predicted was 74.5%, 62.8%, 82.4%, 79.1%, 73.9%, 76.7% and 80.4%, respectively (p = 0.090). Hyperopic outcomes occurred in 70.6%, 76.5%, 49.0%, 74.4%, 76.1%, 74.4%, and 45.1%, respectively (p = 0.007). Conclusions: ORA was superior to all formulas based upon pre- operative biometry and as effective as the AL-optimized Holladay 1 formula when predicting residual refractive error and reducing hyperopic outcomes after cataract surgery in eyes with axial myopia. The Hill-RBF formula's performance was similar to that of the 4th generation formulas.

• Blank Page
• Blank Page (1)