Using the surgical result in the first eye to calculate intraocular lens power for the second eye

Using the surgical result in the first eye to calculate intraocular lens power for the second eye Thomas Olsen, M.D., Niels LlISgstruP, M.D., Henrik Olesen, M.D., Leif Corydon, M.D.

ABSTRACT This paper examines the possibility of using the surgical result in the first eye when planning the intraocular lens power for the second eye. Two methods were considered: (1) an empirical method by which one regards the second procedure as a repeat of the one in the first eye and calculates the power from the actual refractive error obtained in the first eye and (2) a theoretical method by which one measures the pseudophakic anterior chamber depth (ACO) of the first eye and uses this value to plan for the second eye. Based on the data from 136 second eye procedures using extracapsular cataract extraction, the prediction error of the empirical method ranged from -10.5 to +9.5 diopters. The error of the theoretical method ranged from -2.3 to +2.8 diopters, which was significantly more accurate than the empirical method (P < .001). We conclude that the fellow eye ACO may be used as a guideline for the assumed ACO of the second eye. However, such use of the fellow eye ACO could not be shown to improve power calculation predictions significantly. Key Words: anterior chamber depth, biometry, empirical, formula, intraocular lens power calculation, prediction error, pseudophakia, refraction, theoretical

The sources of error in intraocular lens (lOL) power calculation include measurement errors, errors in calibration, variations in surgical technique, and errors in the formula itself.I-4 The prevalence of errors often leaves the surgeon uncertain about selecting a particular IOL power according to any ofthe available formulas. 5-9 Once the eye has been operated on, it is often easier to determine what the IOL power should have been from the final refraction produced by the operation. Because of the biological symmetry between the two eyes, a second eye operation may be regarded as a repeat operation of the first eye. Hence, it should be possible to use the information presented by the first eye to select the IOL power for the second eye. One piece of information presented by the first eye is the actual refraction obtained in that eye. Knowing that measurement error is a major source ofIOL calculation error, the surgeon might consider the deviation of the refraction from the one expected as evidence of mea-

surement error and choose the second IOL power accordingly. This is a purely empirical method and does not require biometry of the second eye. Another piece of information is the pseudophakic anterior chamber depth (ACD) of the first eye. Recent investigations have found the error associated with the prediction of the postoperative ACD to be responsible for 20% to 40% of the total refractive prediction error.4 Any method that improves the prediction of the pseudophakic ACD is therefore likely to improve the accuracy of IOL calculation. One such method might be to take the actual ACD of the first eye and use it in calculating the assumed ACD of the second eye.

MATERIALS AND METHODS One hundred thirty-six consecutive patients having cataract extraction of the second eye were examined. They comprised 40 men and 96 women, ranging in age

From the Department of Ophthalmology, Vejle Sygehus, Vejle, Denmark (L(Jgstrup, Olesen, Corydon), and the University Eye Clinic, Aarhus Kommunehospital, Aarhus, Denmark (Olsen). Supported in part by the Danish Medical Research Council, grant no 12-8138, and Karen Elise Jensens Fond. Reprint requests to Thomas Olsen, M.D., University Eye Clinic, Aarhus Kommunehospital, DK-8000 Aarhus C, Denmark.

36

J CATARACT REFRACT SURG-VOL 19, JANUARY 1993

from 44 to 96 years, mean 74 years. The surgical technique was phacoemulsification or planned extracapsular cataract extraction (ECCE) with in-the-bag placement of a posterior chamber IOL. Three types of lenses were used: Pharmacia 700B, a planoconvex one-piece poly(methyl methacrylate) (PMMA) lens (61 eyes); 3M style iSle, i71e, or 8151e, meniscus-shaped, all-PMMA twopiece lenses (68 eyes); 3M 17xe or other biconvex allPMMA lenses (7 eyes). The axial length and the anterior chamber depth were measured with a 10 MHz A-scanner with a solid transducer probe (Sonometrics DBR 400 Ocuscan). The assumed velocity of ultrasound was 1,550 m/s for the cataractous eye as a whole, and 1,532 m/s for the anterior chamber depth. The corneal curvature was measured with an autokeratometer (Nidek KM-800) in two principal meridians and averaged. The clinical data appear in Table 1. The postoperative measurements were obtained four to six months after surgery as part of a routine follow-up of the accuracy of IOL power calculation. Only cases with a postoperative visual acuity of20/40 or better were included.

second eye. As a measure of the shift in average power required, we may take the difference between the SRK A-constants of the lenses: PowerShift = A.rk(lens B) - A.rk(lens A)

From these relations, the expected refraction may now be calculated as: Rx = (PO + PowerShift - Piol)/1.25

where Rx = expected refraction of the second eye, PO = IOL power that would have produced emmetropia in the first eye, Piol = IOL power actually implanted in the second eye.

Method 2: The Theoretical Method

If the IOL to be implanted in the second eye is the same type as the IOL in the first eye, the ACD obtained in the first eye may be taken directly as the assumed ACD of the second eye. If the IOL is a different type, one should correct for the observed shift from the mean ACD in the first eye: ACDshift = ACD(first eye) - ACDmean(first lens)

where ACD(first eye) is the observed pseudophakic ACD of the first eye as measured by ultrasound, and Method 1: The Empirical Method ACDmealjirst lens) is the observed ACD constant for From the actual refraction produced by a given IOL, the given lens type (= mean value in a representative it is possible to estimate in retrospect what the power of population). The expected ACD of the second eye the lens should have been if it were to produce emmetropia. A common conversion formula, which has been (ACDexp(second eye» now becomes used in the original Sanders-RetzlafT-KrafT (SRK) ACDexp(second eye) = ACDshift + ACDmean(second lens) formula,lo is where ACDmealsecond lens) indicates the ACD con(I) stant for the IOL to be implanted in the second eye. PO = Piol + Rx * 1.25 The ACDexp(second eye) was entered as the assumed where PO = IOL power that would have produced emACD in CATREFRACT,TM an IBM PC-based IOL calmetropia, Piol = actual IOL power, Rx = spherical culation software system based on physiological optics equivalent of the spectacle correction. If one were to developed by one of the authors. 9 The standard ACD implant a lens of the same type in the second eye, one prediction of this software was based on a multiple remight adjust the power according to equation 1. gression formula incorporating the preoperative ACD If one were to implant a different lens type in the and the axial length. 11 second eye, one would have to correct for possible differences in average effective powers between the lenses: Statistical Evaluation Iflens B on the average required 1 diopter (D) stronger To minimize off-set errors (which may conceal the power than lens A to achieve the same refraction, one true error associated With the different methods) each would have to add 1 D when implanting lens B in the formula was optimized in retrospect before evaluation. This involved determining the A-constant and the mean ACD for each lens style based on the postoperative meaTable 1. Clinical and biometric data of 136 patients having ECCE surements of the second eye. The accuracy was in the second eye; mean ± SD. evaluated as the prediction error, i.e., the difference between the observed and the expected value, the range Axial Postop Corneal Length IOL Power of the error, the absolute error (all errors positive), Refraction Power and the correlation between the observed and the ex(D) (D) (D) (mm) Eye pected value. The F-test was applied to test differences 23.79 20.13 -0.89 42.64 First in variance. (±2.29) (±5.30) (±1.50) (±1.57) 20.14 42.71 23.64 Second -0.95 RESULTS (±2.20) (±5.40) (±1.53) (±1.40) The refractive prediction error associated with the -0.15 +0.01 +0.08 -0.06 Difference respective methods is shown in Table 2. The mean error (±2.45) (±0.43) (±0.85) (±1.65) is noted to be close to zero with all methods, as expected J CATARACT REFRACT SURG-VOL 19, JANUARY 1993

37

Table 2. Accuracy of IOL calculation according to the empirical method using the fellow eye refraction, the theoretical method using the fellow eye ACD, and the standard method, which is a theoretical IOL calculation formula using a regression formula for the prediction of the ACD. The error is stated as the mean difference between the observed and the expected refraction (±SD, standard deviation), the range ofthe error, the mean value ofthe absolute difference, and the correlation between the observed and the expected refraction.

Mean Error

Absolute Correlation Error Coefficient (D) (D) r -2.35 - 2.84 0.75 0.73 1.32

0.40

0.71

0.74

from the optimization which minimizes off-set errors. The accuracy of prediction, i.e., the scatter around the mean value, with the empirical method was poor; errors ranged from -10.50 D to +9.47 D (Figure 1). The two largest errors were found in two long eyes of two patients who were found to have an inter-eye difference in axial lengths of 2.0 mm and 5.4 mm. A more accurate prediction was observed with the theoretical method using the fellow eye ACD. This

NUMBER 60

50

D

Theoretical

_

Empirical

40

This study has shown that it is possible to use the ACD of the previously operated fellow eye in the IOL power calculation for the second eye when the ACD of the first eye is used in an optical IOL power calculation formula. To do this, the IOL calculation formula should include the anatomical (measurable) ACD as a parameter. The original theoretical formulas such as the Binkhorst5 and the Colenbrander formula 6 as well as the formula of the present study meet this requirement. However, several of the newer theoretical formulas such as the SRK/T7 and the Holladay 8 are based on calculated ACDs which need not correspond with the distance from the corneal vertex to the anterior lens surface, as measured by ultrasound. If the measured ACD of the fellow eye is to be used by these formulas, a further mathematical conversion may be necessary. High prediction errors may occur if the surgeon uses the refraction error obtained in the first eye as the only guideline for the IOL power of the second eye. The reason for this seems to be the variation in axial lengths that can occur between the two eyes of the same person, especially in myopes. In the present study this inter-eye difference was not apparent by the inter-eye difference in refraction: In one ofthe two high-error cases an intereye difference in axial length of 2 mm was associated with a difference in refraction of 4 D. In another case an inter-eye difference in axial length of5.4 mm was found, but the patient was wearing only a minor myopic correction of the longer eye, which was considered amblyopic. We therefore conclude that one should perform

Table 3. Accuracy of ACD prediction according to the theoretical method using the fellow eye ACD and the standard method, which uses a regression formula based on the preoperative ACD and the axial length. The error is stated as the difference between the observed and the expected value (±SD), the range of the error, the mean absolute error, and the correlation between the observed and the expected value.

30

20

10

o L - _....Il-_--"..... -10

-5

o

5

10

15

PREDICTION ERROR (D)

Fig. 1. (Olsen) Range of prediction errors with the theoretical and empirical methods. 38

DISCUSSION

Range

Method (D) Theoretical -0.05 (±0.96) Empirical +0.01 -10.52 - 9.47 (±2.09) Standard -0.08 -2.59 - 3.20 (±0.94)

-15

method resulted in a comparatively smaller range of errors (from -2.4 to 2.8 D), which was significantly less than that of the empirical method (P< .001). However, compared to our regression method, the fellow eye ACD did not improve the prediction accuracy significantly (P> .05). This was true for the refractive predictions as well as for the specific accuracy in the ACD prediction (Table 3).

Mean Error (mm) Method Theoretical 0.04 (±0.36) 0.02 Standard (±0.31)

JCATARACT REFRACT SURG-VOL 19, JANUARY 1993

Range (mm) -0.98 - 1.14 -0.64 - 1.01

Absolute Correlation Error Coefficient (mm) r 0.28 0.72 0.25

0.77

the biometry and use the axial length reading rather than rely on the refraction obtained in the fellow eye. We could not improve the ACD prediction by using the fellow eye ACD as we can using our regression method, which is based on a multiple regression analysis on the preoperative ACD and the axial length. However, the regression formula already predicts the pseudophakic ACD with a high accuracY,11 and a minor improvement may only be found in a larger series. If a method of individual ACD prediction is not available, that is, if only a fixed value is used, a safe procedure would be to use the fellow eye ACD as described.

4. 5. 6. 7. 8. 9.

REFERENCES 1. Richards SC, Olson RJ, Richards WL, et al. Clinical evaluation of six intraocular lens calculation formulas. Am Intra-Ocular Implant Soc J 1985; 11: 153-158 2. Holladay JT, Prager TC, Ruiz RS, et al. Improving the predictability of intraocular lens power calculations. Arch Ophthalmol 1986; 104:539-541 3. Singh K, Sommer A, Jensen AD, Payne JW. Intraocular lens power calculations; a practical evaluation in normal

10. 11.

subjects at the Wilmer Institute. Arch Ophthalmol 1987; 105: 1046-1050 Olsen T. Sources of error in intraocular lens power calculation. J Cataract Refract Surg 1992; 18:125-129 Binkhorst RD. Intraocular Lens Calculation Manual. A Guide to the Author's TI 58/59 IOL Power Module, 2nd ed. New York, RD Binkhorst, 1981 Colenbrander Me. Calculation of the power of an iris clip lens for distant vision. Br J Ophthalmoll973; 57:735-740 Retzlaff JA, Sanders DR, KraffMe. Development of the SRK/T intraocular lens implant power calculation formula. J Cataract Refract Surg 1990; 16:333-340 Holladay JT, Musgrove KH, Prager TC, et al. A three-part system for refining intraocular lens power calculations. J Cataract Refract Surg 1988; 14: 17-24 Olsen T, Thim K, Corydon L. Accuracy of the newer generation intraocular lens power calculation formulas in long and short eyes. J Cataract Refract Surg 1991; 17: 187-193 Sanders DR, Retzlaff J, Kraff Me. Comparison of the SRK II'" formula and other second generation formulas. J Cataract Refract Surg 1988; 14:136-141 Olsen T, Olesen H, Thim K, Corydon L. Prediction of postoperative intraocular lens chamber depth. J Cataract Refract Surg 1990; 16:587-590

J CATARACT REFRACT SURG-VOL 19, JANUARY 1993

39