Our study evaluated the variability of ocular biometry before cataract surgery using the IOLMaster 700 and investigated the effects of DED on ocular biometric measurements. K values and the amount of corneal astigmatism showed a relatively higher variability between two repeated measurements compared with AL and ACD. The axis of corneal astigmatism was also variable. In 25% of cases, IOL power was changed from the initial surgical plan because of variability. Mean K difference ≥ 0.25D was significantly associated with lower preoperative TBUT, higher preoperative CSS, and postoperative refractive surprise (MAE > 0.5D).
Accurate determination of AL and keratometry values is essential for precise IOL power calculations. Deviations in refractive outcomes after cataract surgery primarily result from fluctuations in AL and anterior keratometry (mean K). Several methods, including optical low coherence reflectometry (OLCR), swept-source optical coherence tomography (ssOCT), and Scheimpflug technology, are currently employed for ocular biometric measurements. Despite the utilization of new devices such as the IOLMaster 700 (ssOCT) and the Lenstar LS 900 (Haag-Streit, Koeniz, Switzerland) (OLCR), as well as modern IOL power formulas including BUII, EVO, Hill-Radial Basis Function (Hill-RBF), Kane Formula, and Ladas Super Formula (LSF), the calculation of IOL power is not entirely foolproof 20.
The IOLMaster 700 has demonstrated higher reproducibility and measurement accuracy compared with other devices 4,21,22. This device utilizes ssOCT and offers a measurement based on images, enabling the examiner to observe the entire longitudinal section of the eye. The IOLMaster 700 assesses AL through the mean values obtained from three scans along each of the six meridians 4. In patients with posterior subcapsular cataracts, there are instances where the IOLMaster 500 and other ocular biometric devices have failed to acquire AL measurements 23. The IOLMaster 500 may yield inaccurate AL measurements in cases of significant lens opacities and in situations where the infrared light with a wavelength of 780 nm is reflected off areas other than the fovea in cases of poor vision. In comparison, the IOLMaster 700 successfully completes AL measurements without encountering failures 4. In a study comparing the IOLMaster 500 and the IOLMaster 700, the IOLMaster 500 failed to attain measurements in 12 eyes (9.7%), whereas the IOLMaster 700 experienced failure in only 3 eyes (2.4%) 24. In another study by Hirnschall et al., the IOLMaster 500 exhibited a failure rate of 6.4%, whereas the IOLMaster 700 had a lower failure rate of only 0.5% 25. The 95% limits of agreement (LoA) were found to be -0.01 to + 0.06 mm for AL, -0.44 to + 0.27 D for corneal power, and − 0.18 to + 0.17 mm for ACD 26. The IOLMaster 700 demonstrated repeatability and reproducibility limits of ± 0.014 and ± 0.023 mm for AL, ± 0.26 and ± 0.27 D for corneal power, and ± 0.02 and ± 0.02 mm for ACD 26. A similar study by Srivannaboon et al. which assessed the repeatability and reproducibility of the IOLMaster 700 in comparison to the IOLMaster 500 with 100 eyes of patients with cataract demonstrated that the 95% LoA of AL for repeatability were − 0.02 to + 0.02 mm for the IOLMaster 700 and − 0.03 to + 0.03 mm for the IOLMaster 500 5. However, despite the excellent repeatability and reproducibility observed with the IOLMaster 700, noticeable discrepancies were identified.
Previous studies have reported inaccuracies in AL measurements for patients with DED 27. The repeatability of AL measurement demonstrated a notable adverse correlation with TBUT (r = − 0.199, p = 0.041) 27. However, in our study, the IOLMaster 700 accurately measured AL, and the range of AL difference between the two measurements was almost negligible at 0–0.06mm. In a study assessing the reliability of the IOLMaster 700, the repeatability of ocular biometric measurements, specifically WTW values, was found to be lower in patients with cataract compared with that in healthy children and adults 28. The diminished repeatability of WTW in this study is presumed to be associated with corneal arcus development in patients with cataract, possibly because of their advanced age 29. In addition, the IOLMaster 700 utilizes ssOCT images for ACD and keratometric measurements, reducing measurement errors and failure rates 4. The IOLMaster 700 and IOLMaster 500 both utilize the telecentric keratometry measurement method, projecting light onto the cornea to measure its curvature 30. However, in this study, the differences in K values were larger, with the K1 difference averaging 0.21 ± 0.21D (max range 1.25D), the K2 difference averaging 0.21 ± 0.23D (max range 1.49D), and the mean K difference averaging 0.18 ± 0.17D (max range 0.88D) (Table 3). Unavoidable errors were revealed when the same patient underwent measurement twice by the IOLMaster700, separated by a 2-week interval. Our study demonstrated that despite the use of advanced instruments like the IOLMaster 700, inconsistencies in ocular biometric measurements may arise, particularly in patients with DED and especially in those with decreased TBUT or increased corneal staining score.
Our findings demonstrated that mean K values, in particular, may contribute substantially to errors in IOL power. In addition, we demonstrated the influence of DED on this variability. In the group with mean K ≥ 0.25D, TBUT was significantly shorter at 3.1 ± 1.3 seconds, and SICCA was higher at 1.1 ± 0.6 (Table 4.). This suggests that DED may influence ocular biometric measurements. These findings are consistent with previous studies. Hiraoka et al. assessed the precision of biometric measurements by conducting the procedure twice using the IOLMaster 500 (Carl Zeiss Meditec) within the same day 27. The repeatability of both keratometry and AL measurements exhibited a negative correlation with TBUT 27. A shorter TBUT was associated with decreased repeatability in both keratometry and AL measurements 27. However, no such correlations were identified for abnormal corneal staining 27. Furthermore, Epitropoulos et al. demonstrated that average K variability was significantly high in the hyperosmolar group (n = 50) compared with that in the normal group (n = 25) 13. Teshigawara et al. also demonstrated that a 2% rebamipide ophthalmic suspension (Mucosta ophthalmic suspension UD2%; Otsuka Pharmaceutical Co.) improved the tear film stability and significantly reduced intra-patient discrepancies in astigmatism power and axis measurements obtained through two repeated measurements using the IOLMaster 700, compared with the Mytear artificial tear ophthalmic solution (Mytear®; Senju Pharmaceutical Co., Ltd., Osaka, Japan) 31. Some studies have reported the worsening of ocular surface condition owing to the instability of the tear film 32,33. Shin et al. demonstrated that there was a negative correlation between the average non-invasive keratograph break-up time and the absolute differences observed in Kmax and astigmatism 15. Thus, all these findings indicate that a stable tear film is essential for accurate measurements using the IOLMaster 13. Astigmatic axis change was observed in two ocular biometric measurements with the IOLMaster 700. Accurately measuring the astigmatic axis is crucial for determining the angular orientation of toric IOLs. In this study, 25 eyes (41.7%) exhibited a difference of 10 degrees or more in the astigmatic axis, and among them, eight eyes (13.3%) showed differences of 30 degrees or more. In a study by Matossian et al, 68% of patients showed a change in the astigmatism axis of orientation after treatment with a thermal pulsation system 34. Preoperative management of DED may be essential to reduce the frequency of toric IOL repositioning.
Although there were 5 cases of refractive surprise in both K variability groups, the rate in the group with mean K differences of 0.25D or more was significantly higher at 45.4% (p = 0.037). However, when dividing the groups based on K1 and K2, there was no significant difference in refractive surprise rates (Table 5.). Kim et al. analyzed the 1-month postoperative MAE in two groups: one group without preoperative treatment and another group that received preoperative treatment starting 2 weeks before cataract surgery, including topical 0.5% Lotepro eye drops, topical 0.05% cyclosporine A eye drops, eyelid scrub, and warm compression 35. In the group without preoperative treatment, refractive surprises occurred at a rate of 17.3% (n = 9) using the Sanders Retzlaff Kraff/Theoretical (SRK/T) formula and 15.4% (n = 8) using the BUII formula 35. However, after preoperative treatment with topical steroid eye drops, the refractive surprises significantly decreased to 3.8% (n = 2) using the SRK/T formula and to 1.9% (n = 1) using the BUII formula, indicating a meaningful difference 35. Furthermore, upon comparing the variability in mean K differences between the dry eye and non-dry eye control groups based on two measurements, the dry eye group exhibited significantly higher variation 36. Some studies have reported a higher variability in mean K in hyperosmolar groups compared with that in normal groups, with a greater proportion showing IOL power differences of 0.5D or more 13. Although there are studies suggesting the use of treatments like rebamipide 37 and cyclosporine 0.09% 38 to reduce refractive errors, further research is required. However, it is necessary to recognize the American Society of Cataract and Refractive Surgery clinical committee's recommendation that adequate treatment for ocular surface disease before cataract or refractive surgery is essential as a matter of great importance 39. Further research is warranted, including a comparative analysis and comparison of variability between the group undergoing preoperative DED management and the group without any intervention.
This study had some limitations. First, the sample size was relatively small. Second, this study was a single-center study using a retrospective design. The study evaluated outcomes 1 month after surgery, providing a relatively short follow-up period. However, in this study, we were able to confirm for the first time that, even with the IOLMaster 700 which is recognized as the latest equipment for ocular biometry, notable differences in measurement errors were observed because of DED. These differences were clinically significant enough to impact cataract surgery planning and postoperative refractive outcomes. Further research is necessary to explore the specific parameters of DED that significantly impact variation in biometric measurements. Additionally, the optimal management strategies for preoperative dry eye before measurement of ocular biometric parameters for cataract surgery must be investigated.
In conclusion, variability in mean K and the amount and axis of corneal astigmatism was more frequent compared to variability in AL and ACD when measured twice with the IOLMaster 700 before cataract surgery. Lower TBUT and higher corneal staining scores were significantly associated with increased variability in mean K values. Vigorous management of preoperative DED is crucial to obtain accurate K values in IOL power calculation.