Following ethics committee approval, all plain ankle radiographs in patients from birth to 14 years of age requested by the emergency department at our institution over a five year period from July 2003 to June 2008 were retrospectively evaluated. Patients with plain radiographs showing traumatic, infective or neoplastic bony abnormalities or not displaying a full lateral view of the calcaneus were excluded from the study, leaving 763 patients with plain lateral radiographs of the ankle and calcaneus that had been reported by a radiologist as normal as our study group. Bohler's angle and the crucial angle of Gissane were measured for every patient in the study group using standard digital radiology software. Each measurement was immediately repeated three times by the principal author, and the mean value for each patient was recorded.
As a comparison group, a consecutive series of 100 plain ankle radiographs displaying a full lateral view of the calcaneus in 100 patients aged 30 years to 70 years of age requested by the emergency department at the adult hospital associated with our institution, that had been reported by a radiologist as normal, were retrospectively evaluated. Bohler's angle and the crucial angle of Gissane were measured for every patient in this group using standard digital radiology software. Each measurement was immediately repeated three times by the principal author, and the mean value for each patient was recorded.
For every patient in the study group and in the comparison group, all three repeated measurements were within a two degree range, indicating acceptable intraobserver measure reliability.
The mean Bohler's angle and the mean crucial angle of Gissane in the paediatric study group were compared with the mean values seen in the adult group. Further comparisons were undertaken by dividing the paediatric study group into five smaller groups according to patient age (0-2 years, 3-5 years, 6-8 years, 9-11 years, and 12-14 years of age) and comparing mean values between each group, and between each group and the adult values. Statistical analysis included one-way analysis of variance testing for comparisons between all groups, t-testing with Tukey adjustments for quantification of differences between groups, and a two sample t-test for comparison between overall paediatric and adult values. It should be noted that the crucial angle of Gissane result in the 0-2 year age group did not have equal variance to results seen in the other age groups and therefore multiple two-sample t-tests with Bonferroni adjustments were undertaken when analysing this result. Statistical significance for the crucial angle of Gissane 0-2 year age group comparisons was therefore defined as a p-value of less than 0.01, while for all other comparisons it was defined as a p-value of less than 0.05. For the statistically significant results, 95% confidence intervals were calculated for the difference between the two means.