^{1}

^{2}

^{3}

^{1}

^{2}

^{1}

^{2}

^{1}

^{2}

^{1}

^{2}

^{1}

^{2}

^{1}

^{2}

^{3}

Primary IOL implantation has become a common and well-accepted method for optical correction in infants and young children [

At present, most of the previous studies have a large age range and small sample size, and more importantly, very few studies have investigated specific age stages. This study aims to assess and compare the predictability of refractive outcomes using SRK/T, Hoffer Q, Holladay 1, Haigis, and Holladay 2 formulas in children of different ages who accepted cataract extraction and primary intracapsular IOL implantation.

All patients underwent pediatric cataract surgery with IOL implantation in the capsular bag by an experienced surgeon (Y. E. Z.) over a 5-year period (September 2013–December 2018) at the Eye Hospital of Wenzhou Medical University (Hangzhou branch), Hangzhou, China. The mean age of the patients was 41.43 months (ranging from 6 months to 84 months). This study was in accordance with the Declaration of Helsinki, and it was approved by the Office of Research Ethics, Eye Hospital of Wenzhou Medical University. Written informed consent was obtained from their parents.

Eyes with traumatic cataracts, corneal disorders such as leukoplakia, retinal diseases, IOL implantation in the ciliary sulcus or secondary implantation, severe complications, and patients who did not finish follow-up on time were excluded from the study.

For poorly cooperative patients, measurements were performed under sedation. AL and ACD were measured via an A-scan (Axis nano, Quantel Medical, France).

The IOL power was calculated by different formulas (with reference to

Target diopter reference law.

Age (years) | Target diopter (D) |
---|---|

<1 | +6.00 |

1.5 | +4.00 |

2 | +3.00 |

3 | +2.50 |

4 | +2.00 |

5 | +1.50 |

6 | +1.00 |

7 | 0.00 |

All patients underwent surgery under general anesthesia. A 2.2 mm superior scleral tunnel incision was created followed by two 1.0 mm side incisions. Lensectomy and limited anterior vitrectomy were applied for all children; the diameter of anterior capsule was 4.5–5.0 mm, and an IOL (SA60AT/ZCB00) was implanted into the capsular bag. Finally, the main incision was closed with a 10-0 nylon suture for older children with side port incisions sealed by stromal hydration, while for younger (younger than 4-5 years old) children, all the incisions were sutured, and the side port sutures were removed 2–4 weeks after operation according to the corneal healing and the tightness of the suture.

Patients were required to attend regular clinical follow-ups, and the refraction was examined at 1 month after surgery by retinoscopy. Spherical equivalent (spherical equivalent = spherical power + cylinder power/2) was converted to calculate the prediction error (PE) as follows: PE = predicted refraction–actual refraction. And absolute PE (APE) was the absolute value of the PE. All the patients’ age, sex, horizontal corneal diameter, AL, K, lens thickness (LT), IOL type and power, ACD, PE, and APE were included for accuracy analysis. The APE percentages of patients within 0.5 diopter (D), 1.0 D, and 2.0 D were calculated to determine which formula could generate the smallest error.

75 patients were divided into three groups according to their age at the time of surgery: group 1 (age: ≤2 years), group 2 (>2 years and ＜5 years), and group 3 (>5 years). Data were analyzed with the statistical software SPSS 21. The Kolmogorov–Smirnov test was used to evaluate normality. The ANOVA test was used to assess intragroup difference about PE and APE results among five formulas. The chi-square test was used to analysis the different percentage of APE within 0.5 D and 1 D among five formulas. Multiple linear regression was used to evaluate the impact caused by the factors that were analyzed.

A total of 102 unilateral and bilateral eyes (75 patients; 50 males and 25 females) were enrolled in this study. The characteristics of the patients are shown in Table

Patients demographic characteristics (

Age ≤ 2 years ( | >2 years and ＜5 years ( | Age ≥ 5 years ( | |

Average age in month (range) | 17.63 (6–24) | 40.33 (24–54) | 71.59 (60–84) |

No. of eyes | 35 | 38 | 29 |

Average AL in mm (range) | 20.62 (18.14–23.81) | 21.96 (19.44–25.85) | 23.00 (20.62–25.63) |

Average K in diopter (range) | 44.55 (40–51.25) | 43.96 (41.23–49.81) | 44.26 (41.98–49.25) |

Average ACD in mm (range) | 2.97 (2.03–4.5) | 3.05 (2.26–3.73) | 3.42 (2.20–4.83) |

Average WTW in mm (range) | 10.26 (9.10–11.50) | 10.85 (9.8–12.40) | 11.52 (10.50–12.50) |

Average IOL power in diopter (range) | 23.64 (17–30) | 22.87 (12–28) | 20.84 (7–28) |

In all three groups, the Haigis formula got the smallest PE among all formulas, although there was no difference overall in the ANOVA test (Figure

PE results of five formulas in the three study groups.

Refractive outcomes of all the pediatric patients using five intraocular lens power calculation formula.

Age ≤ 2 years | >2 years and <5 years | Age ≥ 5 years | ||||
---|---|---|---|---|---|---|

PE ± SD | APE ± SD | PE ± SD | APE ± SD | PE ± SD | APE ± SD | |

SRK/T | 0.67 ± 1.57 | 1.13 ± 1.27 | 0.78 ± 0.97 | 1.018 ± 0.70 | 0.23 ± 1.04 | 0.80 ± 0.70 |

Hoffer Q | 0.32 ± 1.70 | 1.29 ± 1.14 | 0.77 ± 0.97 | 1.03 ± 0.69 | 0.32 ± 0.93 | 0.77 ± 0.60 |

Holladay 1 | 0.70 ± 1.66 | 1.27 ± 1.26 | 1.01 ± 1.56 | 1.51 ± 1.07 | 0.34 ± 0.93 | 0.77 ± 0.61 |

Haigis | −0.11 ± 1.95 | 1.46 ± 1.27 | 0.25 ± 1.04 | 0.87 ± 0.61 | −0.04 ± 0.82 | 0.63 ± 0.50 |

Holladay 2 | 0.80 ± 2.07 | 1.58 ± 1.54 | 1.02 ± 1.84 | 1.71 ± 1.20 | 0.44 ± 0.92 | 0.78 ± 0.65 |

0.196^{a} | 0.630^{a} | 0.084^{a} | <0.001^{a} | 0.360^{a} | 0.854^{a} | |

<0.5 | <1 | <0.5 | <1 | <0.5 | <1 | |

SRK/T | 40 | 54.29 | 26.32 | 50 | 41.38 | 72.4 |

Hoffer Q | 22.86 | 51.43 | 23.68 | 57.89 | 37.93 | 65.52 |

Holladay 1 | 28.57 | 60 | 18.42 | 42.11 | 37.93 | 68.97 |

Haigis | 28.57 | 45.71 | 34.21 | 60.53 | 44.83 | 79.31 |

Holladay 2 | 25.71 | 51.43 | 18.42 | 36.84 | 58.62 | 75.86 |

0.578^{b} | 0.824^{b} | 0.469^{b} | 0.185^{b} | 0.483^{b} | 0.786^{b} |

SD, standard deviation; PE, prediction error; and APE, absolute postoperative refraction. ^{a}ANOVA test. ^{b}Chi-square test.

With regard to APE, there were no statistical differences overall among the formulas in the ANOVA test except group 2. According to further analysis of multiple comparisons, in group 1, the SRK/T formula got a little smaller APE and the Holladay 2 formula a little larger, while in group 3, the Haigis formula a little smaller and the SRK/T formula a little larger, without significant difference. In group 2, the Haigis formula had the lowest APE (0.87 ± 0.61 D), while the Holladay 2 formula had the largest (1.71 ± 1.20 D,

APE results of five formulas in the three study groups.

The percentage of eyes achieved the targeted absolute errors by the five formulas in the three groups is shown in Figure

Percentage of APE of five formulas in the three study groups.

Therefore, the SRK/T and the Holladay 1 formulas were relatively accurate in patients younger than 2 years old, while the Haigis formula was better in patients older than 2.

Multiple linear regression was used to evaluate the impact caused by the factors (AL,

The eyes of younger patients have a shorter AL, a steeper cornea with a higher

In the present study, in group 1, the mean AL was 20.62 mm, and the Haigis formula had the smallest PE, while the SRK/T formula got the best accuracy within 0.5 D in 40.00% and the Holladay 1 formula within 1 D in 60.00%. Kekunnaya et al. [

In group 2, the Haigis formula got the smallest PE and APE and achieved the best accuracy within 0.5 D in 34.21% and 1.0 D in 60.53%. Vasavada et al. [

In group 3, the Haigis formula got the lowest PE and APE in average, while the Holladay 2 formula achieved the best prediction within 0.5D in 58.62% and the Haigis formula within 1.00D in 79.31%. Eibschitz et al. [

For patients younger than 2, more mistakes may occur in examinations due to the poor cooperation comparing other groups. The accuracy of the Haigis formula seemed not very satisfying, and the SRK/T and Holladay I formulas got better results. While for patients older than 2, the Haigis formula performed better. The possible reason is that the Haigis formula can predict better the effective position of IOL, although it may need further study, and the a1 constant of the Haigis formula is tied to the measured ACD, which can be an issue in children of different eye dimensions [

Multiple linear regression analysis demonstrated that age was a major factor affecting APE. Li et al. [

The present study has some limitations. First, we checked the eye parameters and retinoscopy under sedation for poor cooperative children, and measurement errors might occur. Second, A-scan and IOL MASTER 500 were both applied to examinations, and errors may be produced by different apparatus.

Despite the limitations, the present study had some advantages in dividing the children into three different age groups and analyzing the accuracy of five formulas. In conclusion, our results show that the SRK/T and Holladay 1 formulas were relatively accurate in patients younger than 2 years old, while the Haigis formula performed better in patients older than 2.

The data used to support the findings of this study are available from the corresponding author upon request.

The authors declare that they have no conflicts of interest.

This article was supported by the National Natural Science Foundation of China (Grant No. 81870680), the Innovation Discipline of Zhejiang Province (lens disease in children) (2016cxxk1), the Zhejiang Provincial Key Research and Development Program (2018C03012), and the Natural Science Foundation of Zhejiang Province (LY18H120008).