摘要目的:研究Kane公式应用于不同眼轴长度的白内障患者人工晶状体屈光度计算的准确性。方法:回顾性病例系列研究。选取开封市眼病医院2022年4至12月行白内障手术158例(158眼)，根据眼轴长度将患者分为3组：A组49例，眼轴≤22.0 mm；B组55例，22.0 mm<眼轴<26.0 mm；C组54例，眼轴≥26.0 mm。术前行IOL Master 700检查，分别用SRK/T、Hoffer Q、Holladay 2、Haigis、BarrettⅡ、Kane公式计算人工晶状体屈光度，术后1个月行主觉验光，比较人工晶状体屈光度预测误差均值(ME)、绝对值误差中位数(MedAE)。结果:A、C组患者每个公式的ME与0比较差异有统计学意义(均 P<0.05)，MedAE差异有统计学意义( χ2＝34.30、12.27； P<0.001， P＝0.031)，B组所有公式ME与0比较差异无统计学意义( P>0.05)。A组和C组ME组内比较显示Kane公式与其他公式比较差异有统计学意义(均 P<0.01); A组49例不同公式ME在±0.50 D范围内的比率分别为59.18%(29/49)、63.26%(31/49)、67.35%(33/49)、57.14%(28/49)、69.38%(29/49)、71.43%(35/49)，差异有统计学意义( χ2＝12.47, P＝0.018)，C组不同公式ME±0.50 D范围内的比率分别为57.41%(31/54)、53.70%(29/54)、68.52%(37/54)、66.67%(36/54)、72.22%(39/54)、75.93%(41/54)，差异有统计学意义( χ2＝17.83， P＝0.001)。但是A组在ME±1.0 D的比率分别为83.67%(41/49)、85.71%(42/49)、87.75%(44/49)、83.67%(41/49)、93.88%(46/49)、95.92%(47/49)差异无统计学意义( χ2＝10.87， P>0.05)，C组则差异有统计学意义( χ2＝12.44， P<0.05)。 结论:Kane在线公式在短眼轴和长眼轴情况下都有着较好的屈光预测准确性，优于其他公式。

abstractsObjective:To evaluate the accuracy of Kane formula in calculating intraocular lens diopter in cataract patients with different axial lengths.Methods:A retrospective case series study was carried out by selecting (158 eyes) of 158 cases of cataract surgery performed from Apr. to Dec. 2022 in Kaifeng Eye Hospital. The patients were divided into 3 groups by axis lengths: group A, 49 cases with axial length ≤ 22.0 mm; group B, 55 cases with 22.0 mm < axial length < 26.0 mm; and group C, 54 cases with axial length ≥ 26.0 mm. Preoperative IOL Master 700 examination was performed, and the refractive powers of intraocular lenses were calculated using SRK/T, Hoffer Q, Holladay 2, Haigis, Barrett II, and Kane formulas. At 1 month after surgery, subjective refraction was performed to compare the mean prediction error (ME), mean absolute error (MAE) and median absolute error (MedAE) of refractive power of intraocular lens.Results:The ME of each formula in groups A and C was significantly different from 0 ( P<0.05), and the MAE had a significant difference ( χ2=34.30, 12.27; P<0.001, P=0.031). There was no significant difference in the ME of all formulas in group B compared with 0 ( P>0.05). Comparison of ME within groups A and C showed that the Kane formula had a significant difference compared with the other formulas (all P<0.01). The ratios of ME±0.50 D for different formulas in group A were 59.18%(29/49), 63.26%(31/49), 67.35%(33/49), 57.14%(28/49), 69.38%(29/49), and 71.43%(35/49), respectively, with statistically significant differences ( χ2=12.47, P=0.018). The ratios of ME ±0.50 D for different formulas in group C were 57.41%(31/54), 53.70%(29/54), 68.52%(37/54), 66.67%(36/54), 72.22%(39/54) and 75.93%(41/54), respectively, with statistically significant differences ( χ2=17.83, P=0.001). But there was no significant difference in the ratio of ME±1.0 D [83.67%(41/49), 85.71%(42/49), 87.75%(44/49), 83.67%(41/49), 93.88%(46/49), 95.92%(47/49)] in group A ( χ2=10.87, P>0.05), while there was a significant difference in group C ( χ2=12.44, P<0.05). Conclusion:The Kane online formula has better refractive prediction accuracy in both short and long axial length, which is superior to other formulas.