Since its introduction, the toric intraocular lens (IOL) has become a useful tool to correct astigmatism in the field of cataract and refractive surgery.1,2 These lenses improve the uncorrected distance visual acuity (UDVA) outcomes in patients with regular corneal astigmatism. One systematic review of outcomes following toric IOL placement found that approximately 30% of patients require spectacles following toric IOL implantation compared to 53% of patients who received non-toric IOLs, with or without limbal-relaxing incisions.3
Keratometers, corneal topographers, and anterior segment tomographers can provide the corneal measurements necessary to accurately predict the ideal toric IOL power and alignment meridian to correct astigmatism at the time of cataract surgery. For further refinement of the selected toric IOL, surgeons may use open-access, online calculators that take into account variables such as posterior corneal astigmatism and surgically induced astigmatism.1,4 The Barrett Toric Calculator5 is one such tool and is available on the American Society of Cataract and Refractive Surgery web site ( www.ascrs.org/barrett-toric-calculator).
The need for spectacles following toric IOL placement may be due in part to errors in the preoperative astigmatism measurements or differences in the predicted surgically induced corneal astigmatism.1 To postoperatively determine the “ideal” toric IOL alignment meridian, the Berdahl & Hardten Toric Results Analyzer6 (BHTRA), an open-access online tool, was developed ( www.astigmatismfix.com or ascrs.org). This calculator uses the toric IOL position and toricity, as well as the postoperative manifest refraction, to estimate the “ideal” IOL position that would produce the minimum residual astigmatism. The BHTRA thus provides a valuable method for evaluating the ability of keratometers, topographers, and tomographers to estimate the optimal meridian for toric IOL alignment.
The primary aim of this study was to determine which keratometer, topographer, or tomographer among those tested was superior in predicting the ideal toric IOL alignment meridian. The devices tested were the Cassini color light-emitting diode (LED) topographer (i-Optics Corporation, The Hague, The Netherlands), the Galilei G4 Dual-Scheimpflug/Placido Analyzer (Ziemer Ophthalmic Systems AG, Port, Switzerland), and the Lenstar LS 900 optical low-coherence reflectometer/autokeratometer (Haag-Streit AG, Koeniz, Switzerland). A secondary aim of this study was to determine whether the devices are biased in their errors in the clockwise or counterclockwise direction. To the authors' knowledge, no other study has compared these imaging modalities using the BHTRA as the ideal alignment meridian.
Patients and Methods
A retrospective review was performed to identify patients who underwent phacoemulsification cataract surgery with toric IOL implantation from November 2014 to November 2016. Exclusion criteria included previous ocular trauma or surgery, the use of corneal relaxing incisions, cataract incisions greater than 2.5 mm, postoperative corrected distance visual acuity (CDVA) worse than 20/30, no postoperative follow-up examination, active eye pathology that could affect accurate measurements, and/or the inability to obtain accurate measurements. Institutional review board approval was granted by Baylor College of Medicine.
Preoperatively, all patients received visual acuity assessment, manifest refraction, slit-lamp examination, and dilated fundus examination. Corneal measurements were obtained with the color light-emitting diode (LED) topographer, dual Scheimpflug/Placido analyzer, and optical low-coherence reflectometer/autokeratometer (OLCR). Scans were considered poor quality and were excluded if there was “reduced signal” on the dual Scheimpflug/Placido analyzer, quality less than 85% on the color LED topographer, or a flat and/or steep meridian magnitude standard deviation greater than 0.25 diopters (D)7 or astigmatism meridian standard deviation greater than 3.5° on the OLCR.
On the day of surgery, marking of the reference meridian was performed in every patient under topical anesthesia. Patients were seated in an upright position to compensate for cyclotorsion that might occur in the supine position. They were instructed to gaze at a distant target point8 and corneal limbal marks were placed at the 3- and 9-o'clock positions.
All surgeries were performed by one of four surgeons (Douglas D. Koch, MPW, ZA, SSK) at a single academic institution. IOL selection was based on surgeon preference and corneal measurements. Implanted IOLs included the Tecnis toric IOL (Johnson and Johnson, Inc., New Brunswick, NJ) and the AcrySof IQ toric IOL (Alcon Laboratories, Inc., Fort Worth, TX). Intraoperative toric IOL alignment was performed using multiple techniques: manual alignment with pre-placed marks and surface landmarks, the Optiwave Refractive Analysis (ORA) System (Alcon Laboratories, Inc.), the TrueVision 3-D computer-guided system (Truevision Systems, Inc., Santa Barbara, CA), or intrastromal corneal marks made with the Catalys femto-second laser (Johnson and Johnson, Inc.). The manifest refraction and actual toric IOL alignment meridian were measured at the postoperative week 3 visit. The toric IOL alignment meridian was measured after dilation at the slit lamp using an adhesive gauge, as described in a previous study.9
Barrett Toric Calculator
The autokeratometry measurements of the OLCR and the simulated keratometry (SimK) measurements of the Galilei G4 and Cassini do not take into account posterior corneal astigmatism. The Barrett Toric Calculator, which estimates posterior corneal astigmatism, was used to calculate the “Barrett-Corrected axis” for each of these keratometry values. The total corneal power (TCP2) measurement of the Galilei was also included in the comparison. The Galilei TCP2, which is calculated via ray tracing through the anterior and posterior cornea surfaces, takes into account posterior corneal astigmatism and thus was not adjusted using the Barrett Toric Calculator. In addition, the Barrett Toric Calculator allows one to input data from several devices to calculate a combined axis. This “combined Barrett axis” was also calculated using anterior corneal data from all three devices.
Berdahl & Hardten Calculation
Using the freely available BHTRA6 (available at www.astigmatismfix.com or ascrs.org), the ideal alignment meridian for the implanted lens was calculated using the IOL alignment meridian measured at the postoperative week 3 examination.
The primary analysis compared the mean absolute error (MAE) of the alignment meridian for all patients. The absolute error for an individual patient was calculated as the absolute value of the difference between the BHTRA ideal alignment meridian and the meridian predicted by each device, as well as the difference between the BHTRA meridian and the Barrett-Corrected axis. Thus, there were eight comparison groups. Four groups compared the direct readings from the four devices (OLCR auto-keratometry, dual Scheimpflug/Placido analyzer SimK and TCP2, and color LED topographer SimK) with the BHTRA ideal alignment meridian. The remaining four groups compared the Barrett-Corrected readings from the four devices with the BHTRA ideal alignment meridian.
For the secondary analysis, the one-sample Wilcoxon signed-rank test was performed to determine whether the mean relative error for each of the groups was different from zero. Because the relative error for each patient was calculated as the difference between the BHTRA ideal alignment meridian and the meridian predicted by the device, a negative value significantly different from zero would indicate a counterclockwise bias for that device, whereas a positive value would indicate a clockwise bias. In addition, a subanalysis of MAE by device was performed with eyes grouped by the astigmatism location (against-the-rule [ATR] = steep meridian from 0° to 30° or 150° to 180°; with-the-rule [WTR] = steep meridian from 61° to 120°; and oblique = steep axis from 31° to 60° or 121° to 150°).
Statistical analyses were performed using SPSS for Windows software (version 12.0; SPSS, Inc., Chicago, IL), and a P value of less than .05 was considered statistically significant. Patient characteristics were categorized with descriptive statistics. The Friedman test was used to compare the MAE between groups and a Bonferroni correction procedure was used for multiple comparisons. For sample size calculation, we wished to detect a difference of one-half of the standard deviation of differences between two groups in alignment meridian. With a significance level of 5% and a test power of 80%, 32 eyes are required. In this study, we enrolled 56 eyes.
There were 56 eyes of 56 patients who met the inclusion criteria in the study period. The average age was 73.3 ± 9.0 years and there were no statistically significant differences in male:female (28:28) or right:left eye (30:26) distributions. Implanted IOLs included 37 Tecnis Toric ZCT and 19 AcrySof SN6ATT. Descriptive analyses are shown in Table A (available in the online version of this article). The double-angle plot of preoperative astigmatism levels and location as determined by the OLCR is shown in Figure A (available in the online version of this article), with its mean vector (centroid) located at 0.58 ± 1.18 D at 175°.
Patient Characteristics (by Eye)
Double-angle plots showing preoperative corneal astigmatism measured by the Lenstar (Haag-Streit AG, Koeniz, Switzerland). Each ring represents 0.25 diopters (D), and the outer ring represents 1.00 D. red dot = centroid, red line = standard deviation; ATR = against-the-rule astigmatism; WTR = with-the-rule astigmatism
Table 1 shows the MAE for the eight groups. Friedman's test indicated that there were significant differences in MAE among groups (P < .001). The Bonferroni-corrected multiple comparison between pairs only showed statistically significant differences between the Dual Scheimpflug/Placido Analyzer SimK and Combined-Barrett-Corrected Axis (7.6° vs 5.3°, P < .05) and the Dual Scheimpflug/Placido Analyzer SimK and Color LED Topographer SimK (7.6° vs 5.2°, P < .05).
MAE for Each Comparison Group
For ATR eyes (n = 42) in the secondary analysis, there were significant differences in MAE among groups (Friedman's test, P < .001). The Bonferroni-corrected multiple comparisons between pairs only showed statistically significant differences between the Dual Scheimpflug/Placido Analyzer SimK and Combined-Barrett-Corrected Axis (MAE 7.4 vs 4.5) and Dual Scheimpflug/Placido Analyzer SimK and Color LED Topographer SimK (MAE 7.4 vs 4.5). For WTR eyes (n = 9), there were also significant differences in MAE among groups (Friedman's test, P < .026). The Bonferroni-corrected multiple comparisons between pairs only showed statistically significant differences between Dual Scheimpflug/Placido Analyzer SimK and Color LED Topographer SimK (MAE 7.2 vs 5.1). There were no significant differences in MAE between devices among oblique eyes (n = 5). When comparing the mean relative error to zero, only the Dual Scheimpflug/Placido Analyzer SimK group had a mean relative error significantly different from zero (mean relative error= −2.7° ± 9.1°; P = .036).
In addition, post-hoc analyses were performed to better understand characteristics of eyes for which devices had an absolute error greater than 15°. These characteristics are shown in Table 2. Twelve eyes accounted for all instances where the MAE was greater than 15°. Of these 12 eyes, 8 had only one device where the MAE was greater than 15° and 4 eyes had three or more devices with MAE greater than 15°. The number of eyes with absolute error greater than 15° varied between 2 and 5 across the eight groups, and the astigmatism was less than 1.50 D in 92% (22 of 24) of these instances. Table 3 shows the percentage of patients for each device that had a MAE less than 5°, as well as those with a MAE of 5° or greater, subdivided into those who had baseline cylinder magnitude of less than 2.00 D or 2.00 D or greater.
Instances Where the Absolute Error Was > 15°
Device Count by MAE and Preoperative Cylinder Magnitude
Determining the optimal meridian for toric IOL alignment is critical to achieving the best possible visual outcomes for patients receiving these IOLs. Thus, comparing the results obtained for toric IOL implantation using the measurements produced by commonly used, commercially available instruments should be of great interest to cataract surgeons. The current study compared the steep meridian predicted by three different devices using both the initial output and an adjusted meridian calculated by the Barrett formula to the ideal meridian predicted by the BHTRA. To the authors' knowledge, this is the first study to make this comparison.
Although the multiple comparisons procedure found that the Dual Scheimpflug/Placido Analyzer SimK was significantly worse at predicting the ideal meridian than the Combined Barrett-Corrected Axis and the Color LED Topographer SimK, the difference in MAEs was less than 2.5° and likely not clinically significant. Thus, this study found that none of the devices studied were clinically more accurate at predicting the ideal meridian than the others. This was true for the steep meridian initially predicted by the device and the adjusted meridian predicted by the Barrett formula for the OLCR Auto-K, the Dual Scheimpflug/Placido Analyzer SimK, and the Color LED Topographer SimK, which adjusted the steep meridian to account for posterior corneal astigmatism. This is a significant finding because it can often be difficult for surgeons to determine which of the instruments available will predict the optimal meridian for toric IOL placement. The findings of the current study suggest that any of the studied devices may be used.
Findings of the subanalysis may also be useful. ATR eyes accounted for most eyes (42 of 56) and unsurprisingly had results that were identical to the whole group, with significant differences between the Dual Scheimpflug/Placido Analyzer SimK and the Combined Barrett-Corrected Axis and Color LED Topographer SimK. WTR eyes had a similar difference between the Dual Scheimpflug/Placido Analyzer SimK and Color LED Topographer SimK, but not the Combined Barrett-Corrected Axis. Oblique eyes showed no differences among groups. These slight differences between groups may be indicative of a small sample size unable to reach statistical significance, or may represent a real difference in the ability of these devices to predict the ideal alignment meridian in eyes with different types of astigmatism. More studies with larger sample sizes of each eye type are needed to adjudicate this question.
Examining Table 3, five of the eight groups had at least 50% of patients with a toric IOL alignment meridian 5° or greater different from the BHTRA ideal meridian. Additionally, there was no effect of baseline cylinder power on the absolute error magnitude distribution for any of the groups (data not shown). These findings again support the idea that MAE does not vary significantly from group to group and that baseline astigmatism magnitude is likely not a useful measurement for deciding which device would be most accurate in predicting the ideal alignment meridian for an individual patient. In the 12 patients who accounted for all instances where at least one group had an MAE greater than 15°, only one patient had astigmatism greater than 2.00 D and the average amount of astigmatism was 1.26 ± 0.58 D. Because prior research has found that there is greater variability with lower levels of astigmatism,10,11 it is not surprising that higher errors were found in eyes with low levels of astigmatism.
Results from the analysis of the mean relative error indicate that there is not a consistent directionality in the error of meridian prediction by any of the devices. Although the Dual Scheimpflug/Placido Analyzer SimK did show a statistically significant error of 2.7° in the counterclockwise direction, this likely is not a meaningful clinical difference.
Others have also compared the visual outcomes of patients whose toric IOL planning was performed by different devices. Gundersen and Potvin12 reported on 50 eyes where the OLCR was used for preoperative planning and found that visual outcomes were equivalent or superior to manual keratometry for astigmatism planning. A prospective study13 of 25 eyes receiving AcrySof toric IOLs compared the keratometric error (defined as: [true residual astigmatism – anticipated residual astigmatism] / toricity of implanted IOL) for four devices (manual keratometer, autokeratometer, IOLMaster, and Pentacam topographer). Similar to the current study, they found no significant difference in the keratometric error between devices, and all achieved an average keratometric error of less than 1.00 D. In a prospective study with 102 eyes, Potvin et al.14 evaluated clinical outcomes of patients whose toric IOL calculations were based on the Lenstar LS900 OLCR dual zone automated keratometer and found that more than 76% of eyes had 0.50 D or less of astigmatism by 3 months, which was better than what had been reported previously when manual keratometry was used. Another recent study15 compared, among other things, the astigmatism meridian as measured by five devices (VERION, Placido-based corneal topography [OPD-Scan III], LenStar LS900 OLCR, AL-Scan, and auto-kerato-refractometer KR8800). Using the VERION as their reference, they compared the pre-operative corneal astigmatism meridian with each of the other devices. The largest difference in meridian from the VERION was the KR8800 (approximately 3°), but there was no statistically significant difference in the meridian determined by the different devices. However, this differs importantly from the current study where the reference meridian was not one of the devices but an ideal meridian determined with information distinct from any of the devices studied.
Another approach to obtaining the best possible meridian has been to combine measurements from different devices. Browne and Osher16 reported on 87 eyes for which measurements were taken with a manual keratometer and four different automated keratometers (IOLMaster 500, Atlas 995, Lenstar LS 900, and iTrace). Although none of the devices had a significant difference from the combined values for the steep meridian, there were several outliers for each of the devices. Thus, using the approach of combining values from different devices may be a way to increase the precision of alignment meridian determination for individual patients. In our study, the Barrett Toric Calculator was used to both combine the different steep meridians predicted by the devices and account for posterior corneal astigmatism. However, this group was not significantly different in its ability to predict the ideal toric alignment meridian.
There are several strengths of this study. First, all patients received measurements with all three devices, ensuring that any differences found were not due to differences between eyes. A second strength is the relatively large sample size that is comparable to previous studies. Limitations of the study include its retrospective design and the small sample size of WTR eyes. It is also important to note that astigmatism is a two-variable vector with both a direction (meridian) and magnitude (power). The BHTRA and current study focus solely on finding the “ideal” meridian, and thus do not address the problem of choosing the best device for determining the ideal power of a toric IOL for astigmatism correction.
To our knowledge, this study is the first to use the BHTRA in the evaluation of the Lenstar LS 900 optical low-coherence reflectometer, the Galilei G4 Dual Scheimpflug/Placido Analyzer, and the Cassini color LED topographer. It found no difference in the ability of these devices to predict the ideal meridian for toric IOL alignment. The results suggest that cataract surgeons may confidently use any of these devices in preoperative planning for toric IOLs.
- Ventura BV, Wang L, Weikert MP, Robinson SB, Koch DD. Surgical management of astigmatism with toric intraocular lenses. Arg Bras Oftalmol. 2014;77:125–131.
- Miyake T, Kamiya K, Amano R, Iida Y, Tsunehiro S, Shimizu K. Long-term clinical outcomes of toric intraocular lens implantation in cataract cases with preexisting astigmatism. J Cataract Refract Surg. 2014;40:1654–1660. doi:10.1016/j.jcrs.2014.01.044 [CrossRef]
- Kessel L, Andresen J, Tendal B, Erngaard D, Flesner P, Hjortdal J. Toric intraocular lenses in the correction of astigmatism during cataract surgery: a systematic review and meta-analysis. Ophthalmology. 2016;123:275–286. doi:10.1016/j.ophtha.2015.10.002 [CrossRef]
- Visser N, Berendschot TT, Bauer NJ, Nuijts RM. Vector analysis of corneal and refractive astigmatism changes following toric pseudophakic and toric phakic IOL implantation. Invest Ophthalmol Vis Sci. 2012;53:1865–1873. doi:10.1167/iovs.11-8868 [CrossRef]
- American Society of Cataract and Refractive Surgery. Barrett Toric Calculator. http://www.ascrs.org/barrett-toric-calculator. Accessed on July 21, 2017.
- Berdahl J, Hardten D. Berdahl & Hardten Toric IOL Calculator. http://doctor-hill.com/biometry_validation.html. Accessed on July 21, 2017.
- East Valley Opthalmology. Biometry Validation Guidelines. www.doctor-hill.com/biometry_validation. Accessed on January 9, 2018.
- Popp N, Hirnschall N, Maedel S, Findl O. Evaluation of 4 corneal astigmatic marking methods. J Cataract Refract Surg. 2012;38:2094–2099. doi:10.1016/j.jcrs.2012.07.039 [CrossRef]
- Montes de Oca I, Kim EJ, Wang L, et al. Accuracy of toric intraocular lens axis alignment using a 3-dimensional computer-guided visualization system. J Cataract Refract Surg. 2016;42:550–555. doi:10.1016/j.jcrs.2015.12.052 [CrossRef]
- Frings A, Katz T, Richard G, Druchkiv V, Linke SJ. Efficacy and predictability of laser in situ keratomileusis for low astigmatism of 0.75 diopter or less. J Cataract Refract Surg. 2013;39:366–377. doi:10.1016/j.jcrs.2012.09.024 [CrossRef]
- Katz T, Frings A, Linke SJ, Richard G, Druchkiv V, Steinberg J. Laser in situ keratomileusis for astigmatism < 0.75 diopter combined with low myopia: a retrospective data analysis. BMC Ophthalmol. 2014;14:1. doi:10.1186/1471-2415-14-1 [CrossRef]
- Gundersen KG, Potvin R. Prospective study of toric IOL outcomes based on the Lenstar LS 900® dual zone automated keratometer. BMC Ophthalmol. 2012;12:21. doi:10.1186/1471-2415-12-21 [CrossRef]
- Chang M, Kang SY, Kim HM. Which keratometer is most reliable for correcting astigmatism with toric intraocular lenses?Korean J Ophthalmol. 2012;26:10–14. doi:10.3341/kjo.2012.26.1.10 [CrossRef]
- Potvin R, Gundersen KG, Masket S, et al. Prospective multi-center study of toric IOL outcomes when dual zone automated keratometry is used for astigmatism planning. J Refract Surg. 2013;29:804–809. doi:10.3928/1081597X-20131115-03 [CrossRef]
- Lin HY, Chen HY, Fam HB, Chuang YJ, Yeoh R, Lin PJ. Comparison of corneal power obtained from VERION image-guided surgery system and four other devices. Clin Ophthalmol. 2017;11:1291–1299. doi:10.2147/OPTH.S137878 [CrossRef]
- Browne AW, Osher RH. Optimizing precision in toric lens selection by combining keratometry techniques. J Refract Surg. 2014;30:67–72. doi:10.3928/1081597X-20131217-07 [CrossRef]
MAE for Each Comparison Group
|Machine||MAE (Mean ± SD)a||Min/Max Absolute Error for Each Machine|
|Lenstar||5.4 ± 5.1||0/26|
|Lenstar Barrett-Corrected||5.8 ± 4.9||0/22|
|Galilei SimK||7.6 ± 5.7||0/23|
|Galilei SimK Barrett-Corrected||6.1 ± 4.9||0/21|
|Galilei Total K||6.2 ± 4.5||0/19|
|Cassini SimK||5.2 ± 5.6||0/34|
|Cassini SimK Barrett-Corrected||5.8 ± 5.5||0/29|
|Combined Barrett-Corrected||5.3 ± 4.9||0/22|
Instances Where the Absolute Error Was > 15°
|Machine||No.||No. With Astigmatism < 1.50 D|
|Galilei SimK Barrett-Corrected||4||2/4|
|Galilei Total K||2||1/1|
|Cassini SimK Barrett-Corrected||3||3/3|
Device Count by MAE and Preoperative Cylinder Magnitude
|< 5° MAEa||30 (53.6)||27 (48.2)||21 (37.5)||28 (50.0)||21 (37.5)||36 (64.3)||29 (51.8)|
| < 2.0 Db||24 (80.0)||20 (74.1)||15 (71.4)||21 (75.0)||6 (28.6)||27 (75.0)||21 (72.4)|
| ≥ 2.00 Db||6 (20.0)||7 (25.9)||6 (28.6)||7 (25.0)||15 (71.4)||9 (25.0)||8 (27.6)|
|≥ 5° MAEa||26 (46.4)||29 (51.8)||35 (62.5)||28 (50.0)||35 (62.5)||20 (35.7)||27 (48.2)|
| < 2.00 Db||18 (69.2)||22 (75.9)||27 (77.1)||21 (75.0)||27 (77.1)||15 (75.0)||21 (77.8)|
| ≥ 2.00 Db||8 (30.8)||7 (24.1)||8 (22.9)||7 (25.0)||8 (22.9)||5 (25.0)||6 (22.2)|
Patient Characteristics (by Eye)
|Intraocular lens type|
| Tecnis ZCT (total)||37|
| AcrySof SN6 (total)||19|