| Objective:To compare the accuracy of intraocular lens(IOL)power calculation formulae Holladay1,HofferQ,SRK/T,Haigis and Barrett Universal II(hereinafter referred to as Barrett)in children with cataract.Methods:Retrospective case study.Children with cataract surgery of phacoemulsification combined with IOL implantation were involved in the study from January 2011 to January 2018.The predicted refraction of each formula was calculated by preoperative biometric datas(eye axis AL、corneal curvature K、anterior chamber depth ACD)when Implant the same power IOL,and the actual refraction was obtained by the optometry results of 1 to 3 months after surgery.The prediction error(PE)=the actual refraction-the predicted refraction.The absolute prediction error(APE)is the absolute value of the prediction error.Master measurement group and A-scan ultrasound measurement group were analyzed respectively.The patients are grouped according to AL or K,then analyze whether there is any statistical difference between the prediction errors of each formula and 0,and whether the absolute prediction errors of different calculation formulae in different groups are statistically different.The percentage of the prediction error of each formula within ±0.5D,±1D,±2D was analyzed.The effect of age of operation,axis length(AL),keratomery(K),and measuring instrument on the accuracy of different formulae were analyzed by multivariate regression analysis.Results:A total of 45 eyes in the A-scan ultrasound group with an average surgical age of6.30±2.99 years(range,2-14 years),When AL≤22mm,the Barrett formula prediction error(PE)is significantly less than 0(Mean=-0.24,Median=-0.27,P=0.014),and when AL>22mm,the HofferQ formula prediction error(PE)is significantly larger than 0(Mean =0.31,Median=0.33,P=0.039);For the comparison of absolute prediction errors,the Barrett formula APE is significantly smaller than the Holladay1,HofferQ,and SRK/T formulae when K≤43.5D,(mean=0.29,median=0.17).A total of 26 eyes in the Master group with an average surgical age of 7.19±2.86 years(range 4-13 years),In all groups,the Barrett formula prediction error was significantly less than 0(P=0.031,P=0.008,P=0.023,P=0.019);When AL≤22mm or AL>22mm or K>43.5D,the Haigis formula prediction error is also significantly less than 0(P=0.022,P=0.015,P=0.045);For the comparison of absolute prediction errors,There was no significant difference in APE between different formulae in different AL or K groups.In the A-scan ultrasound group,comparing the proportion of prediction errors within ±0.5D of different formulas,the Barrett formula is significantly different from the Holladay1,HofferQ,and SRK/T formulas,and the Barrett formula accounts for a maximum of 68.2%.In the Master group,The proportion of prediction error within ±0.5D、±1D,or ±2D are statistically different from each other.Multivariate regression analysis was performed on the absolute prediction errors of different formulas.It was found that the absolute prediction error of Holladay1 formula was affected by the age of surgery,axial length,corneal curvature and measuring instruments(P=0.006,P=0.008,P=0.008,P=0.021),The absolute prediction error of HofferQ、Haigis formula is affected by the axial direction(P=0.012,P=0.002),and the SRK/T formula is affected by the axial and corneal curvature(P=0.024,P=0.006).The absolute prediction error of Barrett formula is affected by the age of surgery.(P=0.029).Conclusion:For the calculation of IOL power in children with cataract,the postoperative refractive status of HofferQ formula tends to hyperopia.Barrett and Haigis formulae tend to myopia.The Barrett formula is more accurate in children with cataract when using A-scan ultrasound measurement,For the calculation of IOL power in children with cataract,the influencing factors of different formulas are different. |