This study is intended to assess the predictability of desired refractive outcomes in the immediate postoperative period in pediatric patients with cataract undergoing lens aspiration with primary intraocular lens implantation.

In our study of 24 patients over 3 years period, a preponderance of the children was boys, (54.16%). This is consistent with the 51.5% to 72.0% of congenital and developmental cataract reported in the literature [2, 18, 19].

Our data also showed that the distribution of age at time of surgery were more towards older children. Only 4 eyes in children of less than 3 years old were enrolled in our study. This reflects the parental awareness is poor with regards to the diagnosis, early surgical intervention and the need for earlier visual rehabilitation.

A study by Teresa et al [20] of 138 patients has similar findings. The mean age at time of surgery was 3.1 (3.7) years (range 1 week to 20.5 years). In their paper, they did not precisely state the age distribution of their study patients. In a retrospective review by Daniel et al [17] of 101 eyes over 5-year period from June 1998 to August 2003, the mean age at time of surgery was 5.7 (4.4) years old. Their patients' age ranged from 22 days to 18 years old. Their findings compared favourably to us.

### Prediction error

The mean prediction error of the refractive outcome obtained in our study was 1.03 D (SD, 0.69 D) for SRK II group and 1.14 D (SD, 1.19 D) in Pediatric IOL Calculator group. This showed both groups ended more myopic than anticipated. However, there was no statistically significant different in the mean prediction error in both groups. Even though, the SRK II group had a lower prediction error of 0.11 D compared to the Pediatric IOL Calculator, we could not prove that SRK II is better than Pediatric IOL Calculator or vice versa. Our results showed that for the overall group, the prediction error is satisfactory and is comparable with errors demonstrated in adult group [21].

In our study, we further divided and analyzed the prediction error according to age group, axial lengths and keratometry. We divided the groups according to the age at time of surgery to less than 3 years old, and equal or older than 3 years old; axial length of less than 22 mm, and equal or more than 22 mm; and keratometry of less than 46 D and equal or more than 46 D. This was based on the fact that the surgical predictions of appropriate implant power become increasingly complicated in children under age 3 years, especially those under 1 year of age [17, 22].

Although our sample size was too small to reach significance level, there was a trend towards a smaller prediction error in eyes in children equal or older than 3 years old in both formulas. The trends also observed in eyes with axial length equal or more than 22 mm and keratometry equal or more than 46 D (which demonstrated in Pediatric IOL Calculator group). Majority of our patients were equal or older than 3 years at the time of surgery. Only 13% (4 eyes) were less than 3 years old. As most of our patients were equal or older than 3 years old, the eyes also had relatively normal axial length.

We postulated that if we were to get equal representation of samples in both age groups, we would be able to prove the benefit of the Pediatric IOL Calculator. It is well known that the rapid growth of the eye in children, especially in the first year of life, increased tissue reactivity, decreased scleral rigidity and alteration in growth patterns of pseudophakic eyes are the major issue in pediatric cataract surgery [13, 14, 17, 23]. These can lead to postoperative surprise.

Tromans et al [24] did a fairly similar study like us. They showed a similar trend of prediction errors. Our mean prediction errors for both groups were also comparable to them. They did a study using SRK II and SRK/T to determine the accuracy of intraocular lens power calculation in a group of 52 pseudophakic eyes of 40 infants and children. Their interest was on the prediction error at three months post operatively in different axial lengths and age.

In their study they divided the groups according to axial length, keratometry reading and also age at surgery. For the overall group the mean prediction error was 1.40 D (SD, 1.60 D). The mean prediction errors in eye with axial length less than 20 mm was 2.63 D (SD, 2.65 D), and in eyes 20 mm or more was 1.07 D (SD, 0.98D). The mean prediction errors in eyes in children aged 36 months or more was 1.06 D (SD, 1.02 D), while patients less than 36 months was 2.56 D (SD, 2.50 D). The differences between the prediction errors for both axial length and age were statistically significant (p < 0.05).

### Accuracy of predictability

In term of accuracy of the predictability in our study, both formulas did not show any significant difference. The Pediatric IOL Calculator group accurately predicted the immediate postoperative refraction within ±2 D in 80% of patient; while the SRK II group in 87.5% of patients. A study by McClatchey [13, 14] using this Pediatric IOL Calculator found that this formula accurately predicted the refraction within 3 D in 24% of eyes operated before two years old, and in 77% of eyes operated after this age. The accuracy criterion in our study was very small in which we set the cut point of ± 0.5 D as accurate.

The accuracy of postoperative refraction in this study is comparable with studies done by Daniel et al [17] in which 77% achieved ± 2.0 D. In fact our study showed a better result with 62.5% of eyes achieved ± 1.0 D of prediction error in SRK II group and 46.67% in Pediatric IOL Calculator group.

Although it is difficult to compare individual studies because of the variations between the inclusion criteria, exclusion criteria, certain general conclusion can be drawn. Most of the authors agree that pseudophakic eyes grow normally and a significant shift after intraocular lens implantation is to be expected. The resulting myopic shift would lower the estimated refraction, and this should be borne in mind when comparing estimated and actual refractive outcomes [24].

However, from our study there was no significant difference between SRK II and Pediatric IOL Calculator which showed Pediatric IOL Calculator with its myopic shift element did not outperform the SRK II. Even though McClatchey et al [13, 14] with their Pediatric IOL Calculator tried to consider myopic shift as one of important element in the intraocular lens power calculation, still we were unable to obtain a favourable and significant result of predictability in our study. The relatively small sample size and unequal distribution for the group of less than 3 years old could be the explanation for our results.

More information is needed about the growth pattern of the eyes following cataract removal and intraocular lens implantation. With a better understanding of the factors influencing pediatric eye growth will assist in intraocular lens power calculation and the prediction of refractive changes after intraocular lens implantation. This also will increase the confidence of the surgeon in choosing the optimum power of the intraocular lens and most important a better visual quality for the child.

To the best of our knowledge, we are the first to report the predictability of post operative refraction between SRK II and Pediatric IOL Calculator. Though we are unable to get a significant difference between the two formulas, we believe if we were to have a larger sample with longer study duration, we will have more favourable and significant result.

These findings emphasize the differences between adult and pediatric cataract surgery and lend support to arguments for development of a new intraocular lens power calculation formula that addresses the specific needs of the pediatric population. The question of appropriate intraocular lens power selection will require many more years of follow-up in a large number of infants and children before enough data is accumulated to accurately predict the expected refractive change during the rapid growth period.