摘要目的:比较4种不同的人工晶状体度数计算公式在原发性闭角型青光眼合并白内障患者中的准确性。方法:回顾性病例对照研究。纳入山西省眼科医院2022年8月至2022年10月收治的原发性闭角型青光眼合并白内障75例(75眼)。所有患者使用4种不同的人工晶状体计算公式(Hoffer Q、Barrett Universal Ⅱ、Kane及PEARL-DGS)计算所需人工晶状体度数，再行超声乳化白内障吸除人工晶状体植入联合房角镜下前房角分离术。术后3个月进行屈光检查，再行主觉验光。比较4种公式的绝对预测误差(APE)，预测误差(PE)在±0.25 D、±0.50 D、±0.75 D、±1.00 D内所占的百分比及分析术前生物参数和术后眼压与4种公式PE的相关性。结果:Hoffer Q、Barrett Universal Ⅱ、Kane及PEARL-DGS 4种公式的APE分别为0.25(0.10，0.55)、0.27(0.14，0.49)、0.26(0.12，0.39)、0.27(0.09，0.48) D，与0.50 D比较，差异有统计学意义( Z＝4.77、4.72、4.13、4.04，均 P＝0.001)；4种公式PE在±0.25 D ( χ2＝0.25， P＝0.969)、±0.50 D ( χ2＝2.43， P＝0.487)、±0.75 D ( χ2＝3.64， P＝0.303)、±1.00 D ( χ2＝1.61， P＝0.657)、>±1.00 D ( χ2＝1.61， P＝0.657)内的百分比之间的差异均无统计学意义；晶状体厚度与Hoffer Q公式的PE和Barrett Ⅱ公式的PE呈独立负相关( B＝－0.19， t＝－2.50， P＝0.015； B＝－0.19， t＝－2.71， P＝0.008)，前房深度与Kane公式的PE呈独立正相关( B＝0.29， t＝2.23， P＝0.029)，术前术后眼压变化和术前各项生物参数与PEARL-DGS公式的PE无相关关系( F＝1.09， P＝0.381)。 结论:原发性闭角型青光眼合并白内障患者的人工晶状体度数计算公式中Kane及PEARL-DGS的准确性高于计算公式Barrett Universal Ⅱ及Hoffer Q公式，其中PEARL-DGS最优。

abstractsObjective:To compare four different intraocular lens power calculation formulas in primary angle-closure glaucoma patients with cataract.Methods:This was a retrospective case-control study. Seventy-five cases (75 eyes) of primary angle-closure glaucoma with cataract admitted to Shanxi Eye Hospital from Aug. 2022 to Oct. 2022 were included. Four different intraocular lens power calculation formulas (Hoffer Q, Barrett Universal Ⅱ, Kane and PEARL-DGS) were used for all patients to calculate the absolute prediction error (APE), and they underwent phacoemulsification, intraocular lens implantation and goniosynechialysis. Refractive examination was performed on all cases at 3 months after operation, followed with subjective refraction. The absolute prediction error (APE) of 4 formulas were compared, and the percentage of prediction error (PE) in the range of ±0.25 D, ±0.50 D, ±0.75 D and ±1.00 D was measured respectively. The correlation among preoperative biometric parameters, postoperative intraocular pressure, and PE of the four formulas was analyzed.Results:The APE of Hoffer Q, Barrett Universal Ⅱ, Kane, and PEARL-DGS was 0.25(0.10, 0.55), 0.27(0.14, 0.49), 0.26(0.12, 0.39), and 0.27(0.09, 0.48) D, respectively ( Z=4.77, 4.72, 4.13, 4.04; all P=0.001), with significant difference compared with the 0.50 D. There was no significant difference among the percentages of the four formulas within the PE range of ±0.25 D( χ2=0.25, P=0.969), ±0.50 D( χ2=2.43, P=0.487), ±0.75 D( χ2=3.64, P=0.303), ±1.00 D( χ2=1.61, P=0.657), or >±1.00 D( χ2=1.61, P=0.657). Lens thickness was negatively correlated with Hoffer Q PE and Barrett Ⅱ PE independently ( B=-0.19, t=-2.50, P=0.015; B=-0.19, t=-2.71, P=0.008). The depth of anterior chamber was positively correlated with Kane PE independently ( B=0.29, t=2.23, P=0.029). The changes of intraocular pressure before and after surgery and preoperative biometric parameters were not correlated with PEARL-DGS PE ( F=1.09, P=0.381). Conclusion:In primary angle-closure glaucoma combined with cataract surgery, the accuracy of artificial intelligence formulas, i.e. Kane and PEARL-DGS, is higher than that of Barrett universal Ⅱ and Hoffer Q formula, of which PEARL-DGS is the best.