The aim of this study was to compare the accuracy of six intraocular lens (IOL) power calculation formulas in patients with congenital cataract.

Twenty-seven (38 eyes) subjects with congenital cataract who underwent cataract extraction and primary IOL implantation in Beijing Tongren Eye Center, Beijing Tongren Hospital affiliated with Capital Medical University from January 2014 to December 2018 were retrospectively analyzed. Among them, there were 11 male (17 eyes), and 16 female (21 eyes) with the average age of (98.5±37.4) months (ranged from 43 to 162 months). The prediction error (PE) and absolute prediction error (APE) of 6 formulas (SRK Ⅱ, SRK/T, Holladay 1, Hoffer Q, Barrett, Haigis) were calculated. PE of different formulas were described as ±s, APE as ±s or median (upper and lower quartile). The percentage of APE value less than or equal to 0.50 D, 1.00 D, 2.00 D for per formulas, respectively were calculated. The analysis of APE for 6 formulas was performed by Friedman test; the percentage for 6 formulas performed by Cochran′s Q test.

The mean PE of all formulas was negative, suggesting myopia. The median of APE for SRK Ⅱ, SRK/T, Holladay 1, Hoffer Q, Barrett and Haigis formulas from small to large were 0.71 D, 1.04 D, 1.10 D, 1.19 D, 1.20 D, 1.60 D, respectively; the mean APE from small to large were 1.03 D, 1.03 D, 1.11 D, 1.20 D, 1.32 D, 1.60 D. There were significant difference among 6 IOL calculation formulas (χ²=28.33, P<0.05). After comparing using post hoc multiple comparisons for APE of 6 formulas, SRK/T was better than Barrett and Haigis with significant difference (χ²=1.80, 1.43; P<0.05); SRK Ⅱ better than Barrett and Haigis with significant difference (χ²=1.65, 1.28; P<0.05); Holladay 1 better than Haigis with significant difference (χ²=1.38, P<0.05). The eyes and percentage of APE value≤0.50 D, 1.00 D, 2.00 D were different from formulas to formulas. According to the calculation of APE value≤0.50 D, there were significant difference on the percentage for 6 formulas (χ²=15.08, P<0.05); SRK Ⅱ was the most accurate, and 34.21% of patients′ APE value were less than or equal to 0.50 D, while Haigis was the worst accurate, and 10.53% of patients′ APE value were less than or equal to 0.50 D. After comparing using post hoc multiple comparisons for APE of 6 formulas, SRK Ⅱ was better than Haigis with significant difference (χ²=-0.24, P<0.05). According to the calculation of APE value≤1.00 D, there were significant difference on the percentage for 6 formulas (χ²=16.59, P<0.05). SRK Ⅱ was the most accurate, while Haigis was the worst accurate. After comparing using post hoc multiple comparisons for APE of 6 formulas, SRK Ⅱ was better than Haigis with significant difference (χ²=-0.34, P<0.05). According to the calculation of APE value≤2.00 D, there were significant difference on the percentage for 6 formulas (χ²=18.75, P<0.05). Holladay 1 was the most accurate, while Haigis was the worst accurate. After comparing using post hoc multiple comparisons for APE of 6 formulas, Hollday 1and SRK/T were better than Haigis with significant difference (χ²=-0.29, -0.24; P<0.05).

The accuracy of six formulas is not good in subjects with congenital cataract. It is necessary to develop a new IOL calculation formula for children with congenital cataract.