| Objective:With the development of modern ophthalmology microsurgery,cataract surgery has been transformed from blindness prevention surgery to refractive surgery,so it is very important to accurately predict the refractive state of the postoperative eye.The aqueous humor and silicone oil filled eyes often be excluded in traditional studies about the accuracy of IOL power calculation formulas,because of its special refractive state.In recent years,it’s very important to pay attention to special refractive state eyes,because the number of cataract combined surgery has been increasing and patients with history of retinal surgery need basic eyesight for daily life after cataract surgery.The purpose of this study is to compare the accuracy of seven different IOL power calculation formulas(including Barrett Universal Ⅱ,Kane,EVO,Hoffer Q,SRK/T,Haigis,Holladay Ⅱ formula)in different refractive media eyes(vitreous body,silicone oil and aqueous humor),and discuss the applicable scope,finally to guide clinical application which kind of furmula is better.Methods: This study was a retrospective study.A total of 128 eyes of 101 cataract patients were included.These patients are from Ophthalmology center of the second hospital affiliated to Lanzhou University,in June 2020 to January 2021.All the eyes were divided into three groups according to intraocular refractive media,namely:pure cataract surgery group(98 eyes of 73 people)、silicone oil filled eyes cataract surgery group(10 eyes of 10 people)、water filled eyes cataract surgery group(20eyes of 18 people).Used seven kinds of IOL calculation formulas(including Barrett,Universal Ⅱ,Kane,EVO,Hoffer Q,SRK/T,Haigis,Holladay Ⅱ formula)to predict postoperative SE.Compared the SE one month after surgery with the theoretical SE of the above seven formulas,and calculated the average refractive error(ME)、the mean absolute refractive error(MAE)and the median absolute refractive error(Med AE).SPSS 26.0 software was used for statistical analysis.Analysis of variance(ANOVA)or multi-factor analysis of variance(MANOVA)was used for comparison between groups with normal distribution,and further LSD-t test was performed.Non-parametric test was used for comparison between groups with non-normal distribution,Wilcoxon rank sum test was used for comparison between related samples,and Mann-Whitney U test was used for comparison between independent samples.Results: The simple cataract group was divided into three subgroups according to the axial length,K-S test and P-P chart test were performed,and the MAE values of seven formulas of IOL degree in the three subgroups were in line with normal distribution.In the short axial subgroups(AL 22 mm or less),ANOVA showed a significant difference of 0.932(P > 0.05)between groups,further lines of LSD-t test are compared,and any two of the seven formulas showed no significant difference.Hoffer Q formula and Holladay Ⅱ formula had smaller MAE value compared with other formulas.In the medium length axial subgroups(22 mm < AL<26 mm),ANOVA showed a significant difference of 0.979(P > 0.05)between groups,further lines of LSD-t test also shows no significant difference.Holladay Ⅱ formula had smaller MAE value compared with other formulas.In Long axial subgroups(AL≥26mm),ANOVA showed a significant difference of 0.001(P < 0.05)between groups.Significant differences were found in at least one pair of the seven formulas in the long axis subgroup.Further lines of LSD-t test shows: 1.There is no significant difference between any two of the three formulas including Hoffer Q、Haigis and Holladay Ⅱ formula.2.There’s significant difference between the five generation formulas(Barrett Universal Ⅱ,Kane,EVO)and three clinical commonly used formulas(Hoffer Q,Haigis,Holladay).3.There’s no significant difference between any two of the three five generation formulas(Barrett Universal Ⅱ,Kane,EVO).4.There is no significant difference between SRK/T formula and the other six formulas. K-S test and P-P chart test were performed on the water filled eyes group,and the data accorded with normal distribution.The MANOVA analysis of the MAE value of the seven formulas showed that the significant difference of the seven formulas on the MAE value was 1.00(P > 0.05),and the significant difference of the AL on the MAE value was 0.00(P < 0.05).The influence of AL on MAE value was removed,and LSD-t test was carried out for multiple comparisons of the influence of formula factors on MAE value.The results also showed that there was no significant difference between the seven formulas.Considering the smaller value of MAE,therefore Barrett Universal Ⅱ,Kane and EVO these three formula for five generations in water filled eyes group performed better.From water filled eyes group from 12 cases of AL ≥ 26 mm research object and simple cataract group leader eye axis and the group of 20 cases of the object of study for two independent sample t-test,the results found that seven kinds of IOL Power calculation formula of significant difference were greater than 0.05(P > 0.05),so The filling of vitreous cavity with water does not significantly affect the accuracy of postoperative refractive prediction.L-S test and P-P chart test were performed for silicone oil-filled eyes group,and the data did not conform to normal distribution.Each IOL power calculation formula for MAE value between Wilcoxon rank and inspection are compared.There were significant differences between the seven pairs(P < 0.05),it showed that Holladay Ⅱformula had the largest MAE error and Kane formula had the minimum MAE error.the Med AE value also supports this conclusion.From silicone oil filled eyes from 7 cases of AL ≥ 26 mm research object and simple cataract group leader eye axis and the group of 20 cases of the object of study for two independent samples nonparametric test,choose the Mann-Whitney U test,comparison of seven kinds of IOL power calculation formulas on the glass for silicone oil and vitreous body cavity filling refractive media when the prediction accuracy of the presence of significant differences.The results were as follows:The MAE of the seven formulas for predicting IOL degree in the silicone oil-filled eyes undergoing cataract surgery group was significantly higher than that in the simple cataract surgery group,and the errors were statistically significant.Conclusion : In simple cataract group,Hoffer Q and Holladay Ⅱ formula performed better in short axial subgroups.Any two of seven formulas showed no significant difference in medium length axis subgroups.In long axial subgroups,three five generation formulas performed better including Kane formula、EVO formula and Barrett Universal Ⅱ formula.In the water-filled eyes group,the prediction accuracy of the three five generation formulas was better,and vitreous cavity filling with water had no significant influence on the accuracy of postoperative refractive prediction.In the silicone oil-filled eyes group,the errors of all the seven formulas were very large,and the error of Kane’s formula was the smallest.Silicone oil filling in vitreous cavity would have a significant impact on the accuracy of postoperative refractive prediction. |
PDF Full Text Request