The purpose of our study was to examine the effects of pharmacologic pupil dilation on biometric parameters measured by OLCR biometry, and whether these changes affect the power of the calculated IOL. Specifically, we sought to examine the effect of these changes on IOL power calculated using multivariable formulas (Barrett Universal II, Olsen, Hill-RBF and Haigis) when compared to third-generation formulas such as SRK/T and Holladay I using both lowest myopic residual spheroequivalent residual refraction (LMP) and ideal power for emmetropia (IPE) methods (Method A and Method B, respectively).
We saw no significant changes in keratometry values pre-dilation versus post-dilation, which is consistent with previous biometry studies.5–8,13 The significant increase in ACD post-dilation that we observed is also consistent with previous biometric studies, although the increase in our study is smaller than ACD increases reported in other studies.5–8 Nevertheless, this increase in ACD was still large enough to cause statistically significant increases in the IOL power for formulas that included ACD as a variable.
As mentioned previously, multivariable formulas also rely on other variables, such as LT, to calculate IOL power.3 To our knowledge, our study is the first to find a statistically significant decrease in LT after the administration of tropicamide and phenylephrine. Read et al.16 reported significant LT changes after accommodation, but other studies have failed to find changes in LT after administration of tropicamide 1%.5,17 It is also possible that the changes in LT after pharmacological dilation affected the calculated IOL power in our study.
Two studies reported increases in corneal thickness from administration of tropicamide drops.14,15 Our finding of a statistically significant increase in CCT is consistent with these findings. We also observed a small but statistically significant increase in axial length post dilation; this may be partially explained by the increase in CCT. Notably, while an increase in axial length may theoretically lead to a lower IOL power calculation, the magnitude of AL change observed in our study was not large enough to have any significant impact on IOL power calculations.
Several studies have found no change in IOL power related to pupil dilation when using the SRK/T formula.5–7 The SRK/T formula is based on a linear regression method that uses a theoretical, thin-lens eye model relying primarily on AL and K values.18 Although we had a statistically significant increase in AL, the magnitude of this change was not large enough to cause a significant change in IOL power chosen. We noted similar results when performing calculations using both the SRK/T and the Holladay I formula. Though it may not be optimal clinical practice, the presence or absence of pharmacological dilation at the time of biometry will likely not significantly affect the calculated IOL power when using these third-generation formulas.
The Haigis formula was the first of the multivariable formulas to incorporate ACD, and used ACD measurements as a proxy for effective lens position (ELP). Rodriguez-Raton et al.7 compared pre-dilation and post-dilation biometry results in 107 eyes and observed a significant increase (0.098, p = 0.01) in IOL power when using this formula. The authors attributed this result to the increase in ACD seen in the study; because dilation is associated with a shift of the ELP posteriorly, the formula may estimate for an IOL with greater dioptric power in order to reach the target refraction. Khambhiphant et al.8 similarly found a statistically significant increase in IOL power in 373 eyes calculated by the Haigis formula after pharmacological dilation.
While we decided to analyze the potential impact of pre- versus post-dilation measurements on actual IOL powers (Method A), as well as on theoretical custom IOL dioptric power (Method B), we believe that Method A is more relevant for cataract surgeons as it simulates clinical practice as compared to Method B, which has primarily theoretical implications. In the United States, the majority of lens manufacturers make IOLs in increments of 0.5D.
To the best of our knowledge, this is the first study to investigate and report a difference in IOL power using two different calculation methods when performed in an undilated versus post-dilated state using the Barrett, Olsen, and Hill-RBF formulas. Our data also confirm the findings reported previously regarding IOL calculations performed in an undilated versus post-dilated state when using third-generation formulas and the Haigis formula. Our results suggest that in some patients, post-dilation biometry measurements may lead to a higher selected IOL power (when using non-custom IOLs), increasing the risk of myopic surprise. Based on these findings, our recommendation is that cataract surgeons should consider obtaining IOL measurements in an undilated state, particularly if using non-custom IOLs.
We presume that our results have relevance for clinic workflow. Detection of a visually significant cataract by clinicians after dilation may sometimes lead to patients undergoing post dilation biometry measurements. As a result of our findings, we have instituted a protocol at our institution to ensure biometry measurements are done in an undilated state to optimize refractive accuracy (Figure 2). In some cases, this may require patients to make an additional visit, though we have instructed ancillary staff and trainees to obtain biometry measurements in all phakic patients with symptoms consistent with visually significant cataract prior to patient dilation and examination by the attending surgeon. We offer that these measures may be of interest to cataract surgeons to consider for their own respective pre-operative clinical testing protocols.
One limitation of our study is that we did not include the Holladay 2 formula, which also incorporates ACD and LT into its calculations. Further studies would be needed to determine whether a similar change in predicted IOL power would occur with this formula. Furthermore, post-surgical validation of mean absolute error (MAEs), median absolute errors (MedAEs), and percentage of eyes within ±0.25, ±0.50 and ±1.00 D of predicted refraction would be required for our findings in order to demonstrate whether or not the predicted myopic refractive error(s) would occur post-operatively. Based on our preoperative findings, we could not justify using any post-dilation biometry measurements for IOL implantation and therefore cannot comment on the post-surgical MAE and MedAE at this time. Future studies may wish to incorporate a study design to appropriately assess post-operative refractive results based on pre-dilation and post-dilation biometry measurements.
Our study also did not have the required number of eyes to stratify patients based on axial length values (AL), such as short (AL<22.0mm), average (AL 22.0—24.5mm) and long (AL > 24.5 mm) AL values. Similarly, our study did not have the required number of eyes to stratify patients based on keratometry values. Future studies on this topic may be able to address these limitations.
As the current multivariable formulas are increasing in popularity, and given the increased demands for accuracy with IOL calculations, especially when considering advanced technology lenses, surgeons should give increased consideration to pre-operative biometric testing in order to optimize their surgical results. Our data suggest that post-dilation biometry measurements may favor a higher power IOL in a significant portion of cases, increasing the likelihood of a myopic surprise and a potential increase in MAEs and MedAEs. Therefore, we recommend that cataract surgery patients receive biometry testing in an undilated state, especially if multivariable IOL calculation formulas are being used.