4.2.1 Heights
Figure 4 - 1 displays the trajectories of height (cm) for each refractive error group and their fitted LOWESS smoothing curves. Figure 4 - 2 shows their best- fitting FP models of heights. All these plots show similar trends, that is, the increases in height remained constant with age for all the refractive error groups.
55 4.2.2 Spherical equivalent of refractive error
Figure 4 - 3 displays the LOWESS curves and trajectories of SE (D) for each refractive error group. Comparing the best-fitting FP curves in Figure 4 - 4, the plots revealed that the change in SE was faster in the younger children with myopia than those with emmetropia and hyperopia, and slowed down after 11 years of age for all refractive error groups. The SE of the children with persistent hyperopia and emmetropising hyperopia changed faster before their age of 8 years but thereafter the changes were minimal (Figure 4 - 4).
4.2.3 Axial length
Figure 4 - 5 displays the trajectories and LOWESS curves for 10 randomly selected children from each refractive error group. For each group, the spread in AL among the children at the younger ages were similar to the spread when they were 13 years old. These plots also show that there was substantial within-subject variability in AL. This was discerned from the somewhat jagged appearance of the line segment that joined the repeated measures of AL on any subject. In addition, there was also very substantial between-subject variability. That is, some of the children have consistently longer AL at all ages, while others have consistently shorter AL. The LOWESS smoothing curve for each refractive error group fitted using the data of all children showed in Figure 4 - 5 highlighted that the AL was elongating steadily over time for all children but at different rates. The elongation for the children with newly developed myopia and persistent myopia were more rapid than their peers at younger ages. These curves indicate a second-order FP model might be appropriate.
The best-fitting FP models for each refractive error group are given in Figure 4 - 6. The AL of all children was elongated when they grew from 6 to 13 years old but the elongation slowed with age. The elongation tapered off after 10 years of age for
56 all children. Throughout the period, the longest AL was observed in children with persistent myopia, followed by newly developed myopia, persistent emmetropia, emmetropising hyperopia and finally persistent hyperopia (Figure 4 - 6). The elongation rate was similar for children with persistent hyperopia and emmetropising hyperopia compared to those with persistent emmetropia. Children with myopia have faster elongation than those with persistent emmetropia, especially at age less than 10 years old. The best-fitting FP models of AL are given in Table 4 - 3.
Pairwise comparisons between growth patterns of AL for children with persistent emmetropia and other refractive error groups suggest that the elongation of children with newly developed myopia (p < 0.001) and persistent myopia (p < 0.001) were significantly faster than for those with persistent emmetropia, but the elongation of children with persistent hyperopia (p = 0.099) and hyperopia who emmetropise (p
= 0.196) were not significantly different from children who were emmetropic, after adjusting for gender, ethnicity, father’s education level and height (Figure 4 - 7).
4.2.4 Vitreous chamber depth
Figure 4 - 8 presents the trajectories and LOWESS curves of VCD for 10 randomly selected children of each refractive error group. The VCD were growing steadily over time at different rates where the rate for children with persistent hyperopia, emmetropising hyperopia and persistent emmetropia were similar. Figure 4 - 9 shows that the children with persistent myopia had the deepest vitreous chamber between 6 to 13 years of age, followed by newly developed myopia, persistent emmetropia, emmetropising hyperopia and children with persistent hyperopia. For all children, the vitreous chamber was deepened as the child grew older but the deepening tapered off after 10 years of age. A similar rate of deepening in vitreous
57 chamber was observed among children who were emmetropic, hyperopic and emmetropising hyperopic. Children who were myopic and children who had newly developed myopia appeared to have faster deepening in VCD compared to those who were emmetropic, especially at younger ages. The best-fitting FP models of VCD are given in Table 4 - 3.
When compared to persistent emmetropia, children with myopia (newly developed myopia: p < 0.001; persistent myopia: p < 0.001) but not those with hyperopia (persistent hyperopia: p = 0.091; hyperopia who emmetropise: p = 0.059), had significantly faster deepening of the VCD, after adjusting for gender, ethnicity, father’s education level and height (Figure 4 - 10).
4.2.5 Anterior chamber depth
The trajectories and LOWESS curves for 10 randomly selected children per group are displayed in Figure 4 - 11. The ACD curves for children with persistent emmetropia, persistent myopia and newly developed myopia followed an inverted U- shape; while those with persistent hyperopia and emmetropising hyperopia had a constant ACD has demonstrated by a flat line (Figure 4 - 12). Children with persistent hyperopia had the shallowest ACD, followed by children with emmetropising hyperopia, throughout the time they were observed. The ACD of children with persistent emmetropia was slightly deeper than for children with newly developed myopia at younger ages, but the ACD became shallower after 8.5 years of age. Children with persistent myopia had the deepest ACD between 6 to 13 years of age. The peak of the ACD curve occurred at around 9 years in children with persistent emmetropia and 10 years in those with persistent myopia and newly developed myopia. The ACD of children with persistent myopia deepened at a faster
58 rate than those with persistent emmetropia at younger ages, but their rates were similar after the age of 10 years. Table 4 - 3 provides the best-fitting FP models of ACD.
In Figure 4 - 13, the deepening in ACD for children who had newly developed myopia (p < 0.001) was significantly faster at younger ages, but the rate of reduction was slower at older ages compared to those who were emmetropic. This occurred after adjusting for gender, ethnicity, father’s education level and height. However, no significant differences in the growth rates of ACD were observed from the comparisons between children with persistent emmetropia and other groups (persistent hyperopia: p = 0.421; hyperopia who emmetropise: p = 0.769; persistent myopia: p = 0.657).
4.2.6 Lens thickness
Figure 4 - 14 displays the trajectories and LOWESS curves of LT for 10 randomly selected children per group. In contrast to ACD, the growth of LT followed a U-shape pattern for all children, but those who were persistently hyperopic had a
‘flat line’ (Figure 4 - 15). A trough in the LT measurements were observed at around 9 years of age for children with persistent emmetropia and emmetropising hyperopia, while for those with persistent myopia and newly developed myopia, this was observed at around 10 years of age. Children with persistent emmetropia have thinner lenses than those with persistent hyperopia and emmetropising hyperopia. When compared to those with newly developed myopia, the lenses of children who were emmetropic were thicker before 10 years of age, but thinner after that. The thinnest lens was observed in children with persistent myopia. The best-fitting FP models of LT are given in Table 4 - 3.
59 Similar lens thinning rates were observed for children with persistent emmetropia and emmetropising hyperopia at younger ages (Figure 4 - 16). However, the rate of lens thickening was faster for children with emmetropising hyperopia after the trough than for those with persistent emmetropia. The rate of lens thinning of children with newly developed myopia was faster compared to those who were emmetropic in the initial years, and the subsequent thickening of the lens in later years was slower in children who developed myopia. After adjusting for gender, ethnicity, father’s education level and height, the growth rate of LT in children with persistent emmetropia differed from those with newly developed myopia (p < 0.001; persistent hyperopia: p = 0.853; hyperopia who emmetropise: p = 0.549; persistent myopia: p = 0.166).
4.2.7 Corneal radius of curvature
Between the age of 6 and 13 years old, there were minimal changes in CR for children who were myopic while a flat line was observed for children who had persistent hyperopia and persistent emmetropia (Figure 4 - 17). Children with persistent hyperopia had the largest CR, followed by children with persistent emmetropia, newly developed myopia, emmetropising hyperopia and those with persistent myopia. The best-fitting FP models of CR are given in Table 4 - 3 and Figure 4 - 18. No difference was found in the growth rate of CR between the groups after the adjusting for gender, ethnicity, father’s education level and height of the children (Figure 4 - 19).