To explore the accuracy of four new intraocular lens (IOL) calculation formulas, including BU Ⅱ, Kane, EVO 2.0, and LSF, in predicting refractive status after phacoemulsification cataract extraction combined with posterior chamber IOL implantation, vitrectomy, and posterior capsule resection triple surgery.

A total of 104 ophthalmic patients (104 eyes) who visited the Beijing Tongren Eye Center affiliated with Capital Medical University from September 2021 to October 2022 were collected. Among them, there were 26 males (26 eyes) and 74 females (74 eyes) with an average age of (68.4±6.3) years (ranging from 56 to 87 years old). According to the surgical methods, they were divided into a control group and a triple group. The patients of control group underwent phacoemulsification cataract extraction combined with posterior chamber IOL implantation, while those of the triple group underwent phacoemulsification cataract extraction combined with posterior chamber IOL implantation, vitrectomy, and posterior capsule resection triple surgery. The diopter of IOL for all patients using the BU Ⅱ formula and prediction results of Kane, EVO 2.0 and LSF were calculated, respectively. The patient′s age and gender was recorded, and the best corrected visual acuity (BCVA), axial length, corneal curvature, anterior chamber depth, lens thickness, horizontal corneal diameter, and refractive error of implanted IOL before an dafter surgery for 1 month were measured, and the refractive error (PE) of the two groups of patients at 1 month after surgery and the standard deviation (SD), mean absolute error (MAE), and median absolute error (MedAE) of PE and the percentage of PE within the range of ±0.25 D, ±0.50 D, ±0.75 D, and ±1.00 D, formula performance index (FPI) for BU Ⅱ, Kane, EVO 2.0, and LSF formulas were calculated. Age, axial length, corneal curvature, anterior chamber depth, lens thickness, horizontal corneal diameter, and refractive index of implanted IOL all followed a normal distribution, were expressed as ±s and compared by independent sample t-test for inter group. BCVA was described using median and interquartile spacing, and compared by Mann′s-Whitney U test. Gender and eye type were described using percentages, and compared by chi square tests. ME values and zero values were compared by a single sample t test. The MAE and MedAE of different formulas were compared by Friedman test. The percentages of PE within different ranges was compared by Cochran Q test.

The PE values for four formulas, BU Ⅱ, LSF, Kane, and EVO 2.0 in the control group, were (0.032±0.405)D, (0.056±0.389)D, (0.186±0.392)D, and (0.197±0.395)D, respectively. In the triple group, those were (-0.084±0.656)D, (-0.041±0.652)D, (0.047±0.637)D, and (0.050±0.666)D, respectively. In the control group, there was a statistically significant difference in farsightedness system error between Kane and EVO2.0 formulas (t=3.414, 3.599; P<0.05). The systematic error of BU Ⅱ and LSF formulas in the control group was not statistically significant (t=0.570, 1.029; P>0.05). In the triple group, the systematic errors of four formulas were not statistically significant (t=-0.920, 0.530, 0.545, -0.450; P>0.05). The PE of four formulas showed a tendency towards myopia in the triple group compared to the control group, but there was no statistically significant difference (t=1.641, 1.526, 2.315, 1.624; P>0.05). The FPI performance of four formulas was higher in the control group than that of the triple group. The FPI of the BU Ⅱ formula for the control group and triple group were 0.648 and 0.393, respectively, ranking first. The FPI of the LSF formula were 0.526 and 0.381, respectively, ranking second. The FPI of Kane′s formula were 0.479 and 0.351, respectively, ranking third. The FPI of EVO 2.0 formula were 0.463 and 0.346, respectively, ranking fourth.

After surgeryt for one month, BU Ⅱ, LSF, Kane, and EVO 2.0 formulas applied in the triple group do not show hyperopia or myopia drift. Among them, the BU Ⅱ formula has the best performance in predicting accuracy.