rtbrtbrtbrt

1. ORIGINAL PAPER
Precision of 5 different keratometry devices
Jonas Vejvad Nørskov Laursen • Peter Jeppesen •
Thomas Olsen
Received: 24 January 2015 / Accepted: 23 March 2015
Ó Springer Science+Business Media Dordrecht 2015
Abstract To compare the precision among currently
available keratometry devices. The corneal power was
measured on two separate visits with the Nidek
TonoRef II Autorefractor/Keratometer, the Zeiss
IOLMaster 500, the Haag-Streit Lenstar LS 900, the
Oculus Pentacam, and the Oculus Keratograph 4M.
The precision was evaluated as the mean absolute
intersession difference (MAD) between the corneal
power measurements for each patient. Only the non-
operated eye was included in the study. The Ker-
atograph was found to have the highest MAD
(0.215 D), which was significantly different from the
other devices except for the IOLMaster. Nidek ARK
had the lowest MAD (0.097 D), but this was not
significant compared to Pentacam (0.124 D), Lenstar
(0.132 D), or IOLMaster (0.140 D). Only one out of
29 patients had a precision difference exceeding
0.25 D with the Nidek ARK. Among the devices
studied, the Nidek ARK was found to have the highest
and the Keratograph was found to have to the lowest
precision for the measurement of corneal power.
Keywords Keratometry Corneal power
Precision Measurement reliability
Introduction
The corneal power is an important variable in the
calculation of IOL power. As an error of 1D in the
measurement of corneal power will translate into an
IOL power error of approximately the same magni-
tude, the precision by which this variable is measured
is crucial [1]. This study examines the precision, by
means of intersession difference, of five devices that
through ‘K-reading’ obtain the corneal power.
Methods and materials
Twenty-nine eyes of 29 patients scheduled for routine
cataract surgery were enrolled in this prospective
study. Inclusion criteria were no previous eye-surgery
and compliance during measurement. The demo-
graphics of the patients can be seen in Table 1.
During preoperative examination (first visit), cor-
neal curvature (simK) was measured with all devices.
When the patient returned for the postoperative visits
of the first eye (second visit), the non-operated eye was
measured again. The measurements took place at
Aarhus University Hospital, Aarhus, Denmark, from
June 2013 to October 2013. Average time between
visits was 47 days. The devices studied were Nidek
TonoRef II Autorefractor/Keratometer, Nidek/Japan,
IOLMaster 500 (version 1.1.0), Zeiss/Germany, Len-
star LS 900 (version 7.1.2.0042), Haag-Streit/Switzer-
land, Pentacam (version 1.20b23), Oculus/Germany,
J. V. N. Laursen P. Jeppesen T. Olsen
Department of Ophthalmology, Aarhus University
Hospital, Aarhus C, Denmark
J. V. N. Laursen ()
Skovvejen 87, 8000 Aarhus C, Denmark
e-mail: jonaslaursen@dadlnet.dk
123
Int Ophthalmol
DOI 10.1007/s10792-015-0069-3

2. and Keratograph 4M (version 6.03r17), Oculus/Ger-
many. Nidek ARK, IOLMaster, and Lenstar measure-
ments were obtained by nurses trained in performing
these measurements. Pentacam and Keratograph mea-
surements were obtained by a physician (JVL), trained
to use these apparatus. Measurements were taken in no
particular order. Only Pentacams ‘‘anterior curvature’’
measurement (simK) were used, since this was the
only parameter the other devices were able to asses.
All examinations were done before dilating eye drops
were administered. The study adhered to the Decla-
ration of Helsinki and local conduct.
Statistical analysis
The statistics were calculated in Microsoft Office
Excel 2007. The absolute difference was calculated as
the absolute difference between the spherical equiva-
lent of the first and second measurement for each
patient. The mean value of the absolute differences
(MAD) was then compared to the other devices using
Wilcoxon signed-rank test and analysis of variance
(ANOVA). Sw describes the standard deviation
between first and second measurement for each
patient.
All corneal power readings were expressed as
Diopters which was calculated from the corneal
curvature (simK) using the formula D ¼ 1;33751
R ,
where D is the corneal power and R is the anterior
corneal radius of curvature (in meters).
Results
In Table 2, the mean corneal power from the different
devices is shown. No statistically significant differ-
ence was found.
In Table 3, the mean absolute difference of the
repeated K-readings among the different devices is
shown. The Keratograph was found to have the
significantly highest MAD and number of patients
with an absolute difference above 0.25 D compared to
the other devices, except for the IOLMaster.
Figure 1 illustrates the mean absolute difference of
the various devices.
Discussion
This study found no overall statistical difference
between mean corneal powers from the various
devices. Compared to the existing literature, this was
in full agreement regarding Nidek ARK versus
Pentacam [2, 3], split regarding IOL-master versus
Lenstar [4–6], and general discordance regarding
IOLMaster versus Pentacam [7–11]. Almost all stud-
ies found that IOLMaster measured a slightly higher
corneal power compared to other devices; this was
also the case in this study.
Only a few studies seem to have addressed the
question of precision. Some have reported the auto-
mated placido-based keratometers to be more precise
than Scheimpflug devices, which in turn are more
precise than topography-based devices [7, 12]. In our
study, we did not find a lower precision of Pentacam as
compared to the other devices. Our finding of a MAD
of 0.14D with the IOLMaster is in full agreement with
Table 1 Patient demographics
Gender
Female 15
Male 14
All 29
Age (years ± SD)
Average 73.4 ± 10.0
Oldest 88.5
Youngest 39.5
Visit interval (Days ± SD) 47.4 ± 16.9
Eye
Right 18
Left 11
Table 2 Mean corneal power from different devices
Power ± SD (D)
IOLmaster 43.64 ± 1.37
Pentacam 43.59 ± 1.43
Lenstar 43.58 ± 1.36
Nidek ARK 43.57 ± 1.36
Keratograph 43.54 ± 1.36
Overall 43.58 ± 1.37
ANOVA all NS (p = 0.996)
NS Non significant
Int Ophthalmol
123

3. the findings of Shammas and Chan [13]. Who also
found a mean absolute difference of 0.14D.
Several reports have tried to evaluate the accuracy
of keratometers by comparing the error of the IOL
power calculation, i.e. as the error between the
predicted and the observed refraction using different
sources of keratometry.
Lam et al. [2] found that the IOL power prediction
accuracy using Nidek ARK was significantly higher
compared to the Pentacam. Shammas et al. [3] did not
find any significant difference between the two,
although the mean absolute error was highest in the
latter. Symes et al. [11] and Shirayama et al. [14].
observed no significant difference between Pentacam
and IOLMaster. Savini et al. [15] found a significantly
lower accuracy with the Pentacam compared to a
Tomey Topographer.
Rabsilber et al. [5] observed no difference in
accuracy between IOLMaster and Lenstar.
Comparing the Topcon autokeratometer to the
IOLMaster, Hsieh and Wang [16] found the latter to
have a significantly higher accuracy.
In conclusion, we have found the Keratograph to
have the lowest precision of all devices, except when
compared to the IOLMaster. In our study, the most
precise device for measuring corneal power was found
to be Nidek ARK, followed by Lenstar. Further studies
with a larger sample size seem indicated.
Acknowledgments Jonas V. Laursen received a grant from
Grosserer Chr. Andersens Legat (fund), Copenhagen, Denmark.
No author has a financial or proprietary interest in any material
or method mentioned.
References
1. Olsen T (2007) Calculation of intraocular lens power: a
review. Acta Ophthalmol Scand 85:472–485
Table 3 Precision comparison between devices
MAD (D) Sw (D) CI (D) AD B 0.25 D (%)
Nidek ARK 0.097 0.088 0.173 97à
Pentacam 0.124 0.112 0.219 86à
Lenstar 0.132 0.120 0.235 86à
IOLmaster 0.140 0.132 0.260 79
Keratograph 0.215* 0.189 0.371 62**
MAD mean absolute difference, Sw within subject standard deviation, CI 95 % confidence interval, AD absolute difference
* Significantly higher than at p = 0.05 with both ANOVA and Wilcoxon signed-rank test
** Significantly lower than à at p = 0.05 with Z-test for proportions
Fig. 1 Mean absolute
difference (MAD) for each
device
Int Ophthalmol
123

4. 2. Lam S, Gupta BK, Hahn JM et al (2011) Refractive out-
comes after cataract surgery: Scheimpflug keratometry
versus standard automated keratometry in virgin corneas.
J Cataract Refract Surg 37:1984–1987
3. Shammas HJ, Hoffer KJ, Shammas MC (2009) Scheimpflug
photography keratometry readings for routine intraocular
lens power calculation. J Cataract Refract Surg 35:330–334
4. Buckhurst PJ, Wolffsohn JS, Shah S et al (2009) A new
optical low coherence reflectometry device for ocular bio-
metry in cataract patients. Br J Ophthalmol 93:949–953
5. Rabsilber TM, Jepsen C, Auffarth GU et al (2010) In-
traocular lens power calculation: clinical comparison of 2
optical biometry devices. J Cataract Refract Surg
36:230–234
6. Cruysberg LPJ, Doors M, Verbakel F et al (2010) Eval-
uation of the Lenstar LS 900 non-contact biometer. Br J
Ophthalmol 94:106–110
7. Wang Q, Savini G, Hoffer KJ et al (2012) A comprehensive
assessment of the precision and agreement of anterior cor-
neal power measurements obtained using 8 different de-
vices. PLoS One 7(9):e45607
8. Symes RJ, Ursell PG (2011) Automated keratometry in
routine cataract surgery: comparison of Scheimpflug and
conventional values. J Cataract Refract Surg 37:295–301
9. Elbaz U, Barkana Y, Gerber Y et al (2007) Comparison of
different techniques of anterior chamber depth and kerato-
metric measurements. Am J Ophthalmol 143:48–53
10. Karunaratne N (2013) Comparison of the Pentacam
equivalent keratometry reading and IOL Master ker-
atometry measurement in intraocular lens power calcula-
tions. Clin Exp Ophthalmol 41(9):825–834
11. Symes RJ, Say MJ, Ursell PG (2010) Scheimpflug keratometry
versus conventional automated keratometry in routine cataract
surgery. J Cataract Refract Surg 36:1107–1114
12. Shirayama M, Wang L, Weikert MP et al (2009) Compar-
ison of corneal powers obtained from 4 different devices.
Am J Ophthalmol 148:528–535
13. Shammas HJ, Chan S (2010) Precision of biometry, ker-
atometry, and refractive measurements with a partial co-
herence interferometry–keratometry device. J Cataract
Refract Surg 36:1474–1478
14. Shirayama M, Wang L, Koch DD et al (2010) Comparison
of accuracy of intraocular lens calculations using automated
keratometry, a Placido-based corneal topographer, and a
combined Placido-based and dual Scheimpflug corneal to-
pographer. Cornea 29:1136–1138
15. Savini G, Barboni P, Carbonelli M et al (2009) Accuracy of
Scheimpflug corneal power measurements for intraocular lens
power calculation. J Cataract Refract Surg 35:1193–1197
16. Hsieh YT, Wang IJ (2012) Intraocular lens power measured
by partial coherence interferometry. Optom Vis Sci 89:
1697–1701
Int Ophthalmol
123