Foot & Ankle International - September 2020 - 1146

1146

Foot & Ankle International 41(9)

Table 1. Observer Consistency.
Intrarater Reliability (99% CI)

Interrater Reliability

Observer 1
0.936
(0.884-0.965)

Observer 2

Observer 3

All Observers

0.993
(0.986-0.996)

0.986
(0.974-0.992)

0.980
(0.974-0.985)

Table 2. Center-Center Comparisons at 10, 20, and 30 mm.
Axes Compared

Level(s)

Mean Difference
(Degrees)

99% CI

Range

10 mm

0.37

0.31, 0.44

0-1.1

<.001

20 mm

0.36

0.32, 0.48

0-1.1

<.001

30 mm

0.42

0.34, 0.48

0-1

<.001

Center-center vs
center-center

10 mm vs 20 mm

3.06

2.66, 3.46

0.06-5.9

<.001

20 mm vs 30 mm

2.68

2.41, 2.96

1.1-6.3

<.001

Centroid vs centroid

10 mm vs 20 mm

3.22

2.83, 3.62

0.08-6.1

<.001

20 mm vs 30 mm

2.53

2.24, 2.82

0.4-5.6

<.001

Center-center vs
centroid

reliability). For this, each observer was assigned a unique
randomized 25% sample of subjects. Inter- and intraobserver consistency >0.8 is accepted as excellent.
A descriptive univariate analysis was performed and is
reported as means, ranges, and 99% confidence intervals.
An independent samples, 2-tailed t-test was used to compare mean angular differences between axis pairs using
pooled values from the 3 observers. A Bonferroni correction
was implemented to divide the cumulative α of 0.05 among
the 7 comparisons, resulting in statistical significance being
set for P ≤ .007.
A priori sample size calculation for the t test was
informed by the results of pilot work. The pilot indicated
that a minimum of 30 subjects were needed to provide 85%
power to detect a mean difference of 1 degree between a
given axis pair, assuming standard deviation 1 degree for a
2-tailed t test with α = .007.

Results
Interobserver and intraobserver consistencies were found to
be excellent to near perfect (range, 0.94-0.99; Table 1) indicating that the axis measurement was nearly identical when
measured by different observers and when measured at different time points by the same observer.
The alignments of the center-center and centroid axes
differed by an average of only 0.4 degrees with consistently
narrow 99% confidence intervals and ranges at each of the
levels studied. However, these differences were statistically
significant (P < .001; Table 2).
The center-center and centroid axes both externally
rotated a mean of approximately 3 degrees for each increase

in height level. These differences were significant (both P
< .001). Compared to the axis alignments across methods
(center-center vs centroid), there was more variability in
axis alignments across levels as the ranges indicated up to
6.1 degrees of external rotation for 1-level change (Table 2).

Discussion
This CT-based simulation of the clinical center-center technique demonstrated that the center-center and centroid syndesmosis fixation axes were aligned within the zone of tibia
commonly used for syndesmosis fixation. This axis alignment was found to be highly consistent within and between
individuals. There was a mean difference of 0.4 degrees
(mean range, 0.36-0.42 degrees). The strength of this finding was evidenced by very narrow 99% confidence intervals, narrow overall difference ranges, and excellent
observer consistencies (Table 2).
Although the small magnitudes of difference found
between the center-center and centroid axes were statistically
significant, these differences are likely to be clinically insignificant based on the magnitudes of axis deviations studied
previously. Phisitkul et al demonstrated that iatrogenic syndesmosis displacement was minimized when the axis of a
reduction forceps was aligned between the lateral apex of the
fibula and a point on the medial distal tibia metaphysis located
halfway between the anterior and posterior cortices.13 Based
on the images provided by the authors, this axis was within a
few degrees of the centroid and center-center axes. Forceps
axis deviations approximately 15 to 20 degrees from this axis
were found to cause increasing magnitudes of fibula displacement with progressive destabilization of the syndesmosis.

Foot & Ankle International - September 2020

Table of Contents for the Digital Edition of Foot & Ankle International - September 2020