Transcript VERTICAL POWER

Repeatability Analysis of Reliability Differences
Between Vertical and Horizontal Corneal Power
Assessments Obtained with Scheimpflug Corneal
Topography
Maria Ximena Núñez
Claudia Blanco Marín
Vision Sana Research Group
Cornea and Refractive Surgery Unit
Clinica de Oftalmologia de Cali
Cali-Colombia
The authors have no financial interest in the subject matter of this poster.
Measurement
Quality
The quality of a measurement depends on :
Validity
Reliability
Measurement must be as close as possible
to the true value - Gold standard
Proximity between various measurements of
the same phenomenon
Purpose
Determine through repeatability
analysis if the corneal power
provided by the scheimplufg camera
via triangle calculation (Pentacam®)
has reliability differences between
the vertical and horizontal axis
assessments.
Power in Vertical Axis
Power in Horizontal Axis
Methods
Subjects
• We included 61 normal eyes from 32 patients
• The average age was 35 years (range, 20 to 45 years).
• These were candidates for refractive surgery and had
normal corneas
• Patients who were previously using contact lenses were
instructed not to, 15 days prior to the topographic
examinations
Methods
The vertical and horizontal corneal powers were calculated
by simulated keratometry
DEFINITION :VERTICAL POWER
The vertical powers included were :
steep keratometry readings in
with-the-rule corneal astigmatism (WRCA)
And flat keratometry readings in
against-the-rule corneal astigmatism (ARCA).
DEFINITION :HORIZONTAL POWER
The horizontal powers included were:
Flat keratometry readings in WRCA
And steep keratometry readings in
ARCA.
Methods :Repeatability Analysis
1.
Two consecutive measurements from vertical and horizontal corneal powers per
eye from each patient were taken by the same operator between 8:00 am and
5:00 pm
2.
The arithmetic mean (AM) of differences, standard deviation (SD) of differences
and 95 % limits of agreement (LOA) of the mean differences (difference mean ±
1,96 x standard deviation) were calculated following the parameters of the British
Standard Institution by the Bland-Altman concordance analysis
3.
The concordance between the measurements was calculated using the coefficient
of repeatability (CR) of Bland Altman obtained from repeated measurements
taken under identical conditions. The CR was defined as 1.96 times the standard
deviations of the differences between the two measurements.
4.
The agreement, precision and correction factor of the measurement trend was
assessed using the concordance correlation coefficient (CCC) of Lin
Methods
• The correlation coefficient of Pearson was used to
assess the linear association between measurements.
• Normal data was defined by means of the kolmogorovSmirnov test.
• The analysis was done using the Medcalc software,
version 11.2.1.0
Results
BLAND AND ALTMAN PLOT
HORIZONTAL POWER
1,0
Differences:
Sample size
=
Arithmetic mean =
95% CI
=
Standard deviation =
Lower limit
=
95% CI
=
Upper limit
=
95% CI
=
61
-0,04754
-0,1074 to 0,01227
0,2335
-0,5053
-0,6080 to -0,4025
0,4102
0,3074 to 0,5129
HP_1 - HP_2
Method A : Horizontal Power Measure 1
Method B : Horizontal Power Measure 2
0,5
+1.96 SD
0,41
0,0
Mean
-0,05
-0,5
-1.96 SD
-0,51
-1,0
-1,5
38
40
42
44
46
48
AVERAGE of Horizontal Power_1 and Horizontal Power_2
Coefficient of repeatability 0,45
Results
BLAND AND ALTMAN PLOT
VERTICAL POWER
4
Method A : VERTICAL POWER MEASURE 1
Method B : VERTICAL POWER MEASURE 2
3
2
+1.96 SD
1,0
1
Mean
0
0,0
-1.96 SD
-1,0
-1
-2
40
41
42
43
44
45
46
AVERAGE of VERTICAL POWER _1 and VERTICAL POWER_2
47
Differences:
Sample size
=
Arithmetic mean =
95% CI
=
Standard deviation =
Lower limit
=
95% CI
=
Upper limit
=
95% CI
=
61
0,006557
-0,1255 to 0,1386
0,5157
-1,0042
-1,2312 to -0,7773
1,0173
0,7904 to 1,2443
Coefficient of repeatability 0,99
Coefficient of Repeatability– Concordance Correlation Coefficient (CCC) –
Precision - Accuracy
Index
Corneal
Power
Vertical Power
Pentacam
Horizontal Power
Pentacam
0,99
0,45
0,92
0,98
Precision
0,92
0,98
Accuracy
0,99
0,99
CR
CCC =1
Discussion
The results in horizontal corneal power were good.
The AM of the differences was similar in vertical and horizontal power with 0.00 and 0.04 respectively.
The SD of the differences was more precise in horizontal (0.23) than in vertical power
(0.51).
The LOA was wider in vertical power (-1.0 – 1.0) than in horizontal power (-0.50 –
0.41).
The coefficient of repeatability was less in horizontal power (0,45) than
in vertical power (0,99)
The precision of horizontal power (0,98) in CCC was better than the precision of
vertical power (0,92)
Conclusions
Differences of up to ± 1 diopter in the vertical corneal power between the first
and second measurements can be expected in 95% of the eyes.
This finding can be due to normal eye motion, mainly horizontal, which is
enough to degrade the elevation precision and affect the precision of the
curvature calculation.
Finally, despite the results of the CR, AM and SD of the differences, the
Pentacam reliability in the calculation of the vertical corneal power is affected
as a result of the wide LOA. This has important implications for clinical
practice such as the calculation of intraocular lenses.