الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 94 |  |  |
| | | from | 94 | | |
Abstract
Preoperative IOL power calculation is a must to avoid undesirable
postoperative refractive errors. With the newer generation IOL power
calculation formulas, there is general agreement that the pseudophakic
refraction in eyes of average axial length can be predicted with reasonable
accuracy. However prediction errors still occur specially in unusually
long and short eyes, and it is in such eyes that the calculated IOL power
may differ significantly from one formula to another.
The aim of this work is to evaluate the accuracy ofBU formula in
comparison to other commonly used second and third generation
theoretical and regression derived formulas, namely; Hoffer, Binkhorstll,
SRKII, Holladay and SRK/T formulas.
In this study 35 eyes of 30 patients (9 males and 21 females) were
subjected to planned extracapsular cataract extraction with primary PCIOL
implantation.
All patients had full ophthalmic examination and biometry
(keratometric reading and ultrasonic axial length measurement). Errors in
axial length measuements were avoided as much as possible by using Ascan
probe mounted on the slit-lamp, avoiding any compression on the
cornea, proper allignment and taking only the readings fulfilling the
criteria of good A-scan.
Postoperatively, the spherical equivalent (power of the sphere
added algebrically to t he power of the cylinder) was measured 2
months after sugery. The errors in power prediction within 0.5D, ID, 2D
and greater than 2D were calculated for each eye by each of the formulas
·67·,
------- -
used in the study. Statistical evaluation included the standard error of
estimate in diopters for each of the studied formulas.
Six eyes (2 average, I long and 3 short eyes) had k-readings
beyond the limits ofBU formula (Hamed et al., 1996). Thus the values of
M-factor (the short and long eye correction factor related to keratometric
reading) were not precisely specified and we assumed them in the same
way as (Hamed et al., 1996).
Using retrograde study and regression analysis, the M-factor was
readjusted for those 6 eyes and the error in power prediction as well as
SEE were again calculated.
In eyes of average axial length the BU formula after its
readjustment showed the best results (100% of eyes were predicted
accurately within 20) compared to SRKII formula (94.44%), Holladay
and SRK/T formulas (88.89%).
In moderately long eyes it performs well and its prediction
accuracy (100% of eyes were predicted accurately within 20) is similar to
that of SRKII and SRKIT formulas (100%) and better than that of
Holladay formula (85.71 %). However, the number of eyes was limited in
this group so it needs further evaluation on wide range of eyes of long
axial length.
In unusually short eyes, the prediction accuracy ofBU formula
after its readjustment (70% of eyes were predicted accurately within 20)
is somewhat less than that of SRKII, SRK/T and Holladay formulas
(90%), but still is better than that of Binkhorstll (60%) and Hoffer
formulas (40%).
- 68-
The BlJ formula is a pronusrng formula tor PC-lOL power
calculation. As in other regression derived formulas, application of the
formula on large number of eyes with wide range of axial length together
with retrograde study and regression analysis as well as revision of its
correction factors further improves its accuracy and makes it superior to
other available formulas namely; SRKlI, SRK/T and Holladay formulas.
Errors in taking preoperative measurements, specially the axial
length, have been shown to be the most significant source of error when
using any of the modem formulas for calculating appropriate implant
power. The variability of the A-scan readings is inversely proportional to
the accuracy of the implant power calculated.
Finally, we suggested another way of writing the BU formula,
inorder to make it easier in calculating the IOL power, as presented in
(table, 18).