Currently, cataract surgery can be regarded not only as a vision-improving procedure but also as a refractive surgery that helps decrease patients’ need for eyeglasses either for far or near vision. This is because of its high accuracy for postoperative refractive status and because of the introduction of a modern small-wound clear cornea technique with foldable intraocular lens (IOL) implantation, which has largely decreased surgically induced refractive errors.1,2 Therefore, IOL power calculation has become a crucial factor for determining postoperative refractive status and visual acuity. Predicting IOL power from various regression IOL formulas involves measurements of corneal curvatures and the biometry of the eyeball, which are more accurate in modern technology than they were before. For example, partial coherence interferometry (PCI) devices such as the Zeiss IOLMaster (Carl Zeiss Meditec, Jena, Germany) have been used to improve the accuracy of measuring axial length,3–6 and it has been found that the IOLMaster is better for axial length measurement than ultrasound in predicting IOL power.7–11 As to the keratometry offered by IOLMaster, it measures keratometric values based on the anterior corneal curvature, as conventional automated keratometry does, with high repeatability.12,13 Previous studies have proven that combining PCI with the keratometry of the IOLMaster yielded consistent refractive outcomes postoperatively.14 However, some reports suggested that the results of keratometric measurement of the IOLMaster, in addition to biometric measurement, also needed to be optimized to improve refractive outcomes.15 Furthermore, because a certain percentage of patients failed to respond to axial length measurement by the IOLMaster, we further explored whether the keratometry of the IOLMaster was a better choice than conventional automated keratometers for IOL power calculation in these patients. According to our knowledge, no one has ever compared the accuracy of conventional automated keratometers with that of the keratometry of the IOLMaster based on the same biometry, especially in patients in whom IOLMaster cannot work on the axial length measurement. In this study, we compared the accuracy of the keratometry performed by the IOLMaster with that of the conventional automated keratometry in IOL power calculations.
PATIENTS AND METHODS
Patients receiving phacoemulsification with Accurus (Alcon, Fort Worth, Tex) and IOL implantation from June 2006 to December 2009 with a 2.75-mm clear cornea incision without suturing, at the superotemporal site for right eyes and superonasal site for left eyes, were retrospectively recruited in this study. All operations were performed by one surgeon (Y.-T.H.) at the Department of Ophthalmology, Buddhist Tzu Chi General Hospital, Taipei Branch, Taiwan. Exclusion criteria included previous ocular surgery, severe corneal scarring, IOL implantation outside the capsular bag, anterior capsule tearing or posterior capsule rupture, vitreous loss during the operation, and corneal wound suture. The same type of IOL (Acrysof SA60AT; Alcon) was used for all patients. This research followed the tenets of the Declaration of Helsinki, and institutional review board approval was obtained from the Institutional Review Board of Buddhist Tzu Chi General Hospital, Taipei Branch.
Before surgery, all patients received slit-lamp examination, biometric measurement using both the Zeiss IOLMaster (software version 4) and Alcon OcuScan RxP (Alcon), a contact acoustic biometer, and keratometric measurement using both the Zeiss IOLMaster and Topcon KR-8800 (Topcon, Tokyo, Japan). The severity of cataract was classified according to the Lens Opacities Classification System III. For patients whose axial length measurements could not be obtained with the IOLMaster, IOL power calculations were obtained by measurement with (1) the OcuScan RxP and Topcon KR-8800 (O + T) and (2) the OcuScan RxP and keratometry of the IOLMaster (O + I). For those who were responsive to axial length measurement with the IOLMaster, IOL power calculations were further obtained by measurement with (3) the IOLMaster only for both keratometry and biometry (I + I). The IOL power calculation was performed using the software programs installed in the IOLMaster and OcuScan RxP using the SRK/T formula. During stage 1 (June 2006 to May 2007), the lens constant was obtained from the Carl Zeiss Meditec homepage (http://www.zeiss.de/iol_master) for the IOLMaster (A = 118.7) and was used according to the manufacturer’s suggestion for the OcuScan RxP (A = 118.4). The refraction (spherical equivalent [SE]) was measured with the Topcon KR-8800 1 month after operation, and the errors of predicted refraction (observed minus predicted refraction) for different measurements were calculated. These data were collected for personalized optimization of predicted refraction as follows:
Equation (Uncited)Image Tools
This formula was used for IOL power calculation in stage 2 (June 2007 to December 2009). During stage 2, the SE was measured 1 month after the operation. The errors (observed minus optimized predicted refraction) and absolute errors (absolute values of errors) of predicted refraction, as well as percentages of patients with absolute error of predicted refraction within 0.5, 1, and 2D, were calculated.
The F tests were used for comparison of SDs of mean errors (MEs) of predicted refraction. Paired t tests were used for comparing mean absolute errors (MAEs) of predicted refraction between two different measurement combinations for the same patients. Bowker tests of symmetry were applied to compare the proportions of patients with an absolute error of predicted refraction within 0.5, 1, and 2D between two different device combinations. Kruskal-Wallis tests were applied for comparison of MAEs among more than two groups. Linear mixed models were used to analyze the effects of different devices and other associating factors, including age, status of diabetes mellitus (DM), and severity of cataract on the MAEs of predicted refraction. A random effect was used to capture the correlations between the two eyes of the same patient. Age was treated as a continuous variable, whereas status of DM, severity of nuclear sclerosis (≥NO4/NC4 vs ≤NO3/NC3), cortical opacity (≥C3 vs ≤C2), and posterior subcapsular opacity (≥P3 vs ≤P2) were treated as binary variables. A value of P < 0.05 was considered statistically significant. SAS 9.1 (SAS Institute Inc., Cary, NC) was used for all statistical analyses.
A total of 320 eyes of 249 patients were recruited consecutively in this study. The mean age of the patients was 71.9 ± 10.9 years, and 58.1% of the patients were female. The mean axial length measured by the IOLMaster was 23.79 mm (21.34 to 30.74 mm); the mean axial length measured by the OcuScan RxP was 23.44 mm (21.07 to 30.30 mm).
Of the 88 eyes recruited in stage 1 of the study, the IOLMaster failed to provide a biometric measurement in 26 eyes. The percentages of absolute error within 0.5, 1, and 2D for the three different measurements were shown in Table 1. The MEs of predicted refraction for I + I, O + I, and O + T were 0.18D, −0.06D, and −0.07D, respectively. Therefore, the optimized predicted refractions were the additions of 0.18D, −0.06D, and −0.07D into I + I, O + I, and O + T groups in stage 2, respectively.
Of the 232 eyes recruited in stage 2 of the study, the IOLMaster failed in the axial length measurement in 91 eyes. The MEs of the optimized predicted refraction for I + I, O + I, and O + T were 0.02D, 0.03D, and 0.05D, respectively. For the 141 eyes completing all three measurement combinations, the SD of ME of the optimized predicted refraction was the smallest in I + I (0.48D), followed by O + I (0.58D), and then O + T (0.64D) (Table 2). The MAE of optimized predicted refraction was also the smallest in I + I (0.38 ± 0.28D), followed by O + I (0.49 ± 0.34D), and then O + T (0.54 ± 0.37D) (Table 3). The percentages of absolute error within 0.5, 1, and 2D, respectively, improved to 68.1, 97.8, and 100% in I + I; 57.4, 92.9, and 100% in O + I; and 55.3, 87.9, and 100% in O + T (Fig. 1). For the 91 eyes where the IOLMaster could not determine the axial length measurement, the MAE by O + I (0.57 ± 0.52D) was still superior to that by O + T (0.62 ± 0.56D) (P = 0.03).
Factors Associated with Predictability of Postoperative Refraction
Failure of Axial Length Measurement by IOLMaster
For those eyes where the axial length measurement could not be determined by the IOLMaster, the MAEs of optimized predicted refraction for O + I and O + T were not significantly different from those responsive to the IOLMaster (P = 0.11 for O + I and 0.15 for O + T, respectively).
Different Axial Length and Keratometry Groups
Patients were divided into three groups according to axial lengths: shorter than 22 mm, 22 to 26 mm, and longer than 26 mm. They were then divided into three groups according to keratometry reading: smaller than 42D, 42 to 44D, and larger than 44D. No significant differences of the MAEs of optimized predicted refraction among different axial length groups or keratometry groups were noted, although there was a trend that the MAE was smaller in the group with keratometric readings less than 42D (Tables 4, 5).
Other Associating Factors
The regression coefficients of factors including age, status of DM, and severity of cataract for MAEs using linear mixed models are shown in Table 6. None of these factors were significantly associated with the predictability of refractive outcomes.
Factors Associated with the Failure of Axial Length Measurement Using the IOLMaster
Of the 320 eyes recruited in this study, we failed to measure the axial lengths of 117 eyes using the IOLMaster (36.6%). Severe nuclear sclerosis (P = 0.0005) and severe posterior subcapsular opacity (P < 0.0001) were associated with failed axial length measurement by the IOLMaster, whereas severe cortical opacity was less associated with such failure (P = 0.09).
In this study, we found that the keratometry of the IOLMaster was more accurate than conventional automated keratometry in IOL power calculation. We think that the reason for this is that the radius for light projection on a corneal surface is 2.3 mm for the IOLMaster, which is different from the 3 mm radius of most conventional automated keratometers.13 The corneal curvature measured at a radius of 2.3 mm is closer to the central corneal curvature than that measured at a radius of 3 mm. In addition, the IOLMaster automatically repeats the process until three measurements agree within 0.25D in each meridian. This may be the reason why the IOLMaster is more precise than conventional automated keratometry in calculating IOL power. Furthermore, we also demonstrated that PCI is better than contact acoustic biometry in axial length measurement for IOL power calculation, with less variation in errors of predicted refraction, fewer absolute errors, and higher precision rates. Although the magnitude of difference in MAE was small (0.54D in O + T vs 0.38D in I + I), the proportion of patients with absolute errors of predicted refraction less than 0.5 or 1D was significantly less with measurement of I + I than with measurement of O + T. These results were consistent with those of previous reports.7–11
In our study, we also analyzed the possible associating factors of accuracy of IOL power calculation in terms of its keratometry, as well as its biometry. Previous studies have shown various results for the effects of axial length, keratometry, and severity of cataract on the precision of IOL power calculation.16–19 We found that all of the factors, including age, axial length, keratometry, severity of cataract, and status of DM exerted no different effects on IOL power predictability, no matter whether IOLMaster or acoustic biometry was used. Importantly, our results further demonstrated that the predictability for patients responsive to the IOLMaster was not significantly different from those irresponsive to IOLMaster when the axial length was measured by acoustic biometry. This result indicates that the severity of cataract was not a main factor for the precision of axial length measurement either by an acoustic biometry or by PCI.
Regarding the optimization according to the surgeon’s personal experiences, we measured the errors between observed and expected postoperative refraction in stage 1 and then adjusted the predicted refraction in stage 2 according to the 1-year personally collected data. The MAE of predicted refraction after optimization was 0.38D with I + I after personal optimization, and the percentages of absolute errors within 1D increased from 95.7% in stage 1 to 97.8% in stage 2; this is comparable to or even superior to the results of previous studies.7,9,10,20 As to measurement with O + I and O + T, the percentages of eyes within 1D also increased, respectively, from 83.0 and 80.9% in stage 1 to 92.9 and 87.9% in stage 2. According to these results, we showed that there was an improvement after optimization. Several studies have demonstrated that personalized optimization of IOL constants can reduce the unexpected errors of postoperative refraction.21–24 Although the manufacturer of the IOLMaster has been updating optimized lens constants for the machine, we still suggest that each surgeon should optimize the lens constants personally because the contributing factors of IOL power predictability vary among surgeons and even different patients.
The major problem of the IOLMaster is its failure to determine an axial length measurement in 8 to 18% of cases.20,25,26 Furthermore, in our study, the failure rate was 36.6%, which is much higher than those reported in previous studies. The major causes of this failure may be the lens itself, with severe nuclear sclerosis and posterior subcapsular cataract. Other causes of failure of the IOLMaster include patients having previous ocular surgery, vitreous hemorrhage, or severe corneal scars; such patients were excluded from our study. In these cases, measurement can only be taken through acoustic biometry. Usually, surgeons use the lens constants offered by IOL manufacturers for IOL power calculation. In these cases, we still applied the personalized optimization for IOL power calculation with an acoustic biometry, and we also showed good predictability of postoperative refraction. This proves again that each surgeon has his/her own normogram. Therefore, we suggest that the lens constants used for IOL power calculation be adjusted according to each surgeon’s own experiences whether using PCI or acoustic biometry.
In conclusion, based on our study results, the Zeiss IOLMaster is more accurate than both acoustic biometry and conventional automated keratometry in IOL power calculation. Acoustic biometry, however, is still indispensable for eyes irresponsive to the biometry of the IOLMaster, and the keratometry of the IOLMaster still offers better outcomes than those of conventional automated keratometry in these cases. Personalized optimization improved the results of predicting postoperative refraction both for the IOLMaster and conventional devices.
Department of Ophthalmology National Taiwan University
Hospital No. 7 Chung Shan South Rd
The authors have neither financial or proprietary interests to disclose nor any public or private financial support.
Received May 2, 2012; accepted August 10, 2012.
1. Ermis SS, Inan UU, Ozturk F. Surgically induced astigmatism after superotemporal and superonasal clear corneal incisions in phacoemulsification. J Cataract Refract Surg 2004; 30: 1316–9.
2. Giansanti F, Rapizzi E, Virgili G, Mencucci R, Bini A, Vannozzi L, Menchini U. Clear corneal incision of 2.75 mm for cataract surgery induces little change of astigmatism in eyes with low preoperative corneal cylinder. Eur J Ophthalmol 2006; 16: 385–93.
3. Drexler W, Findl O, Menapace R, Rainer G, Vass C, Hitzenberger CK, Fercher AF. Partial coherence interferometry: a novel approach to biometry in cataract surgery. Am J Ophthalmol 1998; 126: 524–34.
4. Vogel A, Dick HB, Krummenauer F. Reproducibility of optical biometry using partial coherence interferometry: intraobserver and interobserver reliability. J Cataract Refract Surg 2001; 27: 1961–8.
5. Kiss B, Findl O, Menapace R, Wirtitsch M, Drexler W, Hitzenberger CK, Fercher AF. Biometry of cataractous eyes using partial coherence interferometry: clinical feasibility study of a commercial prototype I. J Cataract Refract Surg 2002; 28: 224–9.
6. Kielhorn I, Rajan MS, Tesha PM, Subryan VR, Bell JA. Clinical assessment of the Zeiss IOLMaster. J Cataract Refract Surg 2003; 29: 518–22.
7. Findl O, Drexler W, Menapace R, Heinzl H, Hitzenberger CK, Fercher AF. Improved prediction of intraocular lens power using partial coherence interferometry. J Cataract Refract Surg 2001; 27: 861–7.
8. Rainer G, Petternel V, Findl O, Schmetterer L, Skorpik C, Luksch A, Drexler W. Comparison of ultrasound pachymetry and partial coherence interferometry in the measurement of central corneal thickness. J Cataract Refract Surg 2002; 28: 2142–5.
9. Connors R 3rd, Boseman P 3rd, Olson RJ. Accuracy and reproducibility of biometry using partial coherence interferometry. J Cataract Refract Surg 2002; 28: 235–8.
10. Eleftheriadis H. IOLMaster biometry: refractive results of 100 consecutive cases. Br J Ophthalmol 2003; 87: 960–3.
11. Rose LT, Moshegov CN. Comparison of the Zeiss IOLMaster and applanation A-scan ultrasound: biometry for intraocular lens calculation. Clin Exp Ophthalmol 2003; 31: 121–4.
12. Heatley CJ, Whitefield LA, Hugkulstone CE. Effect of pupil dilation on the accuracy of the IOLMaster. J Cataract Refract Surg 2002; 28: 1993–6.
13. Santodomingo-Rubido J, Mallen EA, Gilmartin B, Wolffsohn JS. A new noncontact optical device for ocular biometry. Br J Ophthalmol 2002; 86: 458–62.
14. Olsen T. Improved accuracy of intraocular lens power calculation with the Zeiss IOLMaster. Acta Ophthalmol Scand 2007; 85: 84–7.
15. Fam HB, Lim KL. Improving refractive outcomes at extreme axial lengths with the IOLMaster: the optical axial length and keratometric transformation. Br J Ophthalmol 2009; 93: 678–83.
16. Ueda T, Ikeda H, Ota T, Matsuura T, Hara Y. Relationship between postoperative refractive outcomes and cataract density: multiple regression analysis. J Cataract Refract Surg 2010; 36: 806–9.
17. Olsen T. Sources of error in intraocular lens power calculation. J Cataract Refract Surg 1992; 18: 125–9.
18. Prinz A, Neumayer T, Buehl W, Kiss B, Sacu S, Drexler W, Findl O. Influence of severity of nuclear cataract on optical biometry. J Cataract Refract Surg 2006; 32: 1161–5.
19. Kim SM, Choi J, Choi S. Refractive predictability of partial coherence interferometry and factors that can affect it. Korean J Ophthalmol 2009; 23: 6–12.
20. Rajan MS, Keilhorn I, Bell JA. Partial coherence laser interferometry vs conventional ultrasound biometry in intraocular lens power calculations. Eye (Lond) 2002; 16: 552–6.
21. Aristodemou P, Knox Cartwright NE, Sparrow JM, Johnston RL. Intraocular lens formula constant optimization and partial coherence interferometry biometry: refractive outcomes in 8108 eyes after cataract surgery. J Cataract Refract Surg 2011; 37: 50–62.
22. Madge SN, Khong CH, Lamont M, Bansal A, Antcliff RJ. Optimization of biometry for intraocular lens implantation using the Zeiss IOLMaster. Acta Ophthalmol Scand 2005; 83: 436–8.
23. Gale RP, Saldana M, Johnston RL, Zuberbuhler B, McKibbin M. Benchmark standards for refractive outcomes after NHS cataract surgery. Eye (Lond) 2009; 23: 149–52.
24. Petermeier K, Gekeler F, Messias A, Spitzer MS, Haigis W, Szurman P. Intraocular lens power calculation and optimized constants for highly myopic eyes. J Cataract Refract Surg 2009; 35: 1575–81.
25. Tehrani M, Krummenauer F, Blom E, Dick HB. Evaluation of the practicality of optical biometry and applanation ultrasound in 253 eyes. J Cataract Refract Surg 2003; 29: 741–6.
26. Nemeth J, Fekete O, Pesztenlehrer N. Optical and ultrasound measurement of axial length and anterior chamber depth for intraocular lens power calculation. J Cataract Refract Surg 2003; 29: 85–8.