Purpose
To assess the intrasession and intersession precision of higher-order aberrations (HOAs) measured using a commercial Hartmann-Shack wavefront sensor (Zywave; Bausch & Lomb) in refractive surgery candidates.
Design
Prospective, experimental study of a device.
Methods
To analyze intrasession repeatability, 1 experienced examiner measured 30 healthy eyes 5 times successively. To study intersession reproducibility, the same clinician obtained measurements from another 30 eyes in 2 consecutive sessions at the same time of day 1 week apart.
Results
For intrasession repeatability, excellent intraclass correlation coefficients (ICCs) were obtained for total ocular aberrations, total HOAs, and second-order terms (ICC, > 0.94). The ICCs for third-order terms also were high (ICCs, > 0.87); however, fourth-order ICCs varied from 0.71 to 0.90 (Z 4 0 = 0.90); and fifth-order ICCs were less than 0.85. For intersession reproducibility, only total ocular aberrations, total ocular HOAs, second-order terms, Z 4 0 , Z 3 1 , and Z 3− 3 had ICCs of 0.90 or more. Bland-Altman analysis showed that the limits of agreement (were clinically too wide for most higher-order Zernike terms, especially for the third-order terms (> 0.21 μm).
Conclusions
Total ocular aberrations, total HOAs, and second-order terms can be measured reliably by Zywave aberrometry without anatomic recognition. Third-order terms and Z 4 0 are repeatable, but not as reproducible between visits. Fourth-order terms, except for Z 4 0 , and fifth-order terms are not sufficiently reliable for clinical decision making or treatment. Because the variability of Zywave can be a major limitation of a truly successful wavefront-guided excimer laser procedure, surgeons should consider treating HOA magnitudes that are more than the intrasession repeatability values (2.77 × S w ) as those presented in this study.
Clinical wavefront sensors evaluate optical quality, and their main area of application is refractive surgery. There are several types of aberrometers depending on the principle used: Tscherning, automated retinoscopy, laser ray tracing, and Hartmann-Shack. Zywave (Bausch & Lomb, Rochester, New York, USA) is a Hartmann-Shack aberrometer that, when coupled to a corneal topographer (Orbscan II, Bausch & Lomb), generates an individual excimer laser ablation profile.
Excimer laser keratorefractive surgery is the most popular procedure worldwide for correcting refractive errors and is in continuous technologic evolution. In the last decade, the ability to quantify higher-order aberrations (HOAs) allowed refractive surgeons to design a refined laser treatment. Wavefront-guided excimer laser procedures have been introduced to reduce postoperative reports of glare, halos, and decreased night vision after conventional laser surgery. However, retreatments still are needed after wavefront-guided and conventional excimer laser surgery to satisfy patient expectations.
The success of the wavefront-guided excimer laser ablation relies on several intraoperative factors, such as laser beam homogeneity, cyclotorsional misalignment or decentration during excimer laser ablation, and the corneal biomechanical and wound-healing responses. However, the reliability of the acquisition process of the wavefront sensor during the measurement of the optical aberrations is one of the most important variables, because the wavefront-guided ablation profile depends on its measurement data.
The Zywave has been used to assess the optical quality after blinking, the presence of lenticular opacities, contact lens wear, pseudophakic intraocular lens (IOL) implantation, and the diurnal variations in HOAs, and its safety and efficacy already have been evaluated in patients who underwent wavefront-guided LASIK. A number of investigators have assessed the Zywave test–retest variability in healthy subjects. However, there were no consistent definitions of both repeatability and reproducibility within the same study, and after a thorough review of the literature, no study has evaluated the instrument’s ability to measure HOAs in such a wide range of ametropias like the current Zywave measurement variability analysis or was sufficiently detailed to allow differentiation of each Zernike term. Therefore, the purpose of the current study was to quantify the reliability of the Zywave aberrometer using definitions of reproducibility and repeatability based on those recommended by Bland and Altman. Our goal was to point out that targeting a HOA treatment at less than the repeatability of the wavefront sensor used can result in a major miscalculation that limits the outcome of the wavefront-guided excimer laser treatment, as previous authors have already pointed out (Stahl ED. Comparing multiple devices. Paper presented at the Fourth Annual Wavefront Congress. February 14-16, 2003; San Francisco, California).
Methods
Patients were recruited over a 3-month period. Exclusion criteria included the presence of suspected keratectasia, progressive myopia, or astigmatism; active ocular disease; any other ocular diagnosis that may alter the optical quality; and contact lens wear. All eyes underwent a complete ophthalmic examination before the Zywave measurement.
Intrasession Repeatability and Intersession Reproducibility
Under repeatability conditions, the same operator (A.L.M.) obtained independent test results using the same method on the same subject and the same equipment with the shortest time possible between successive sets of readings. We investigated the repeatability by performing 5 examinations of 30 eyes after ensuring proper focusing and alignment.
Under the reproducibility conditions, sets of readings were obtained using the same method. To investigate the intersession reproducibility, the same experienced operator performed 2 examinations of 30 eyes 1 week apart. Examinations were carried out from 11:00 am to 1:00 pm to minimize the effect of diurnal variations in HOAs.
To calculate the intrasession repeatability, the within-subject standard deviation (S w ) of 5 consecutive measurements was calculated by obtaining the square root of the value, referred to as the residual mean square in 1-way analysis of variance. The precision, from a statistical standpoint, is the difference between a subject’s measurement and the true value (average value that would be obtained over many measurements) for 95% of observations and was defined as ± 1.96 × S w . Repeatability, or the value below which the difference between 2 measurements will lie with a probability of 0.95, also was analyzed and was defined as 2.77 × S w . We also calculated the intrasession reliability of the measurement method with the intraclass correlation coefficient.
To calculate the intersession reproducibility, graphs of the differences from the means were plotted to ascertain that there was no relationship between the differences and the range of measurement and that the differences between measurements were approximately normally distributed. The 95% limits of agreement then were defined as the mean difference in measurements performed during 2 different sessions (standard deviation, ± 1.96), with lower values indicating higher reproducibility. We also calculated the intersession S w with an analysis of variance. The precision of the measurements in different sessions also was calculated and was defined as 1.96 times the intersession S w . As it happens, 2.77 times this S w is equal to the reproducibility of the measurement in different sessions. The paired t test also was used to establish whether there was a significant systematic bias between measurements. The intersession reliability of the measurement method was calculated with the intraclass correlation coefficient.
Zywave Measurements
The Zywave wavefront system is based on the Hartmann-Shack principle. The Zywave aberrometer uses a 600-μm pitch lenslet array with a focal length of 40 mm. The instrument uses a near-infrared 785-nm laser beam and measures up to 76 locations, depending on the pupil diameter. The wavefront is measured using a lenslet array that is conjugate with the pupil plane. The lenslet array focuses the wavefront on a charge-coupled device (CCD) detector, and then the data are computed. The details of the Zernike coefficients up to the fifth order may be obtained. Zywave measures and compensates for the refractive errors before taking the HOA measurement and tries to avoid patient accommodation by fogging the eye before the wavefront is acquired. Each run consists of 5 measurements; the system computes the average of the 3 best runs after discarding the 2 with higher deviations from the mean. Zywave measurements were obtained without pharmacologic mydriasis and dark adaptation; however, we always obtained measurements from pupils that were 6.5 mm or larger. Patients were instructed to blink between measurements, and acquisition was obtained after a blink to ensure higher-quality results by limiting tear film disruption. One experienced technician (A.L.-M.) obtained the measurements to avoid interobserver variability in the results. Zywave calibration was performed before a measurement session following the manufacturer’s guidelines.
The Zywave software computes ocular wavefront aberration data in standardized Zernike terms up to the fifth order. This information was exported for 6-mm pupil diameters. Zywave shows the aberration data considering that positive propagation is toward the eye. Therefore, exported data are normalized Zernike polynomial coefficients, but do not strictly conform to those of the Optical Society of America because the coefficient sign is the opposite value. In the current study, the sign of the coefficients was not reversed, as other authors have done previously, to avoid clinician misunderstanding when comparing our results and the data commonly displayed by the Zywave in their eye clinics and because assessing the measurement variability, which was the primary aim of the current study, does not depend on the direction of light propagation. All Zywave data provided throughout the study were always for a 6-mm pupil size.
Statistical Analyses
Data from the prospectively completed forms were entered into a database, and statistical calculations were performed using SPSS software version 15.0 for Windows (SPSS, Inc, Cary, North Carolina, USA). The means and standard deviations were calculated for normally distributed data. For all statistical tests, a 2-tailed P < .05 was considered significant.
Results
Intrasession Repeatability
The repeatability was studied in 30 eyes of 30 patients (15 men, 15 women; average age, 32.20 ± 9.34 years; range, 20 to 55 years). The mean ametropia was −4.26 ± 1.94 diopters (D; range, −7.50 to 1.50 D); the mean astigmatism was −0.93 ± 0.83 D (range, −3.00 to −0.00 D). Table 1 shows the overall average ocular aberration values corresponding to the repeated measures, the intrasession S w , the precision, the repeatability, and the intraclass correlation coefficient.
| Aberration (μm) | Overall Mean (Range) | S w | Precision (1.96 × S w ) | Repeatability (2.77 × S w ) | ICC (95% CI) |
|---|---|---|---|---|---|
| TOA RMS | 6.60 (0.47 to 13.76) | 0.131 | 0.256 | 0.362 | 0.99 (0.99 to 0.99) |
| HOA RMS | 0.44 (0.14 to 0.81) | 0.039 | 0.076 | 0.109 | 0.96 (0.93 to 0.98) |
| HOA RMS (Z 4 0 ) | 0.37 (0.12 to 0.73) | 0.039 | 0.076 | 0.109 | 0.94 (0.90 to 0.97) |
| Z 2 0 | −6.405 (−13.190 to −0.029) | 0.130 | 0.25 | 0.360 | 0.99 (0.99 to 0.99) |
| Z 2 −2 | 0.079 (−1.309 to 3.284) | 0.069 | 0.135 | 0.191 | 0.99 (0.98 to 0.99) |
| Z 2 2 | 0.582 (−1.351 to 3.703) | 0.089 | 0.174 | 0.246 | 0.99 (0.99 to 0.99) |
| Z 3 −1 | −0.016 (−0.625 to 0.354) | 0.116 | 0.227 | 0.321 | 0.82 (0.72 to 0.89) |
| Z 3 1 | 0.008 (−0.297 to 0.408) | 0.035 | 0.068 | 0.097 | 0.96 (0.93 to 0.98) |
| Z 3 −3 | 0.073 (−0.143 to 0.363) | 0.054 | 0.105 | 0.150 | 0.87 (0.80 to 0.93) |
| Z 3 3 | −0.009 (−0.293 to 0.289) | 0.042 | 0.082 | 0.117 | 0.90 (0.84 to 0.94) |
| Z 4 0 | −0.161 (−0.696 to 0.124) | 0.067 | 0.131 | 0.186 | 0.90 (0.84 to 0.94) |
| Z 4 −2 | 0.010 (−0.102 to 0.127) | 0.019 | 0.037 | 0.053 | 0.87 (0.80 to 0.93) |
| Z 4 2 | 0.028 (−0.130 to 0.168) | 0.022 | 0.043 | 0.060 | 0.88 (0.82 to 0.94) |
| Z 4 −4 | −0.001 (−0.066 to 0.145) | 0.022 | 0.043 | 0.060 | 0.83 (0.73 to 0.90) |
| Z 4 4 | −0.019 (−0.105 to 0.093) | 0.029 | 0.056 | 0.080 | 0.71 (0.58–0.83) |
| Z 5 −1 | −0.004 (−0.244 to 0.096) | 0.021 | 0.041 | 0.059 | 0.85 (0.76 to 0.91) |
| Z 5 1 | −0.001 (−0.062 to 0.053) | 0.016 | 0.031 | 0.044 | 0.59 (0.43 to 0.74) |
| Z 5 −3 | 0.007 (−0.061 to 0.169) | 0.020 | 0.039 | 0.055 | 0.69 (0.55 to 0.82) |
| Z 5 3 | 0.006 (−0.039 to 0.041) | 0.016 | 0.031 | 0.045 | 0.54 (0.38 to 0.71) |
| Z 5 −5 | 0.007 (−0.096 to 0.091) | 0.020 | 0.039 | 0.056 | 0.64 (0.49 to 0.78) |
| Z 5 5 | 0.002 (−0.051 to 0.090) | 0.021 | 0.041 | 0.059 | 0.45 (0.28 to 0.64) |
Intersession Reproducibility
Intersession reproducibility was studied in 30 eyes of 30 patients (16 men, 14 women; average age, 27.90 ± 3.63 years; range, 21 to 35 years). The mean ametropia was −0.12 ± 2.06 D (range, −4.50 to 6.00 D); the mean astigmatism was −0.75 ± 0.63 D (range, −2.25 to −0.00 D). Table 2 shows the overall average ocular aberration values corresponding to the 2 examinations, the intersession S w , the precision, the reproducibility, and the intraclass correlation coefficient corresponding to each variable assessed. Table 3 shows the 95% limits of agreement corresponding to the intersession variability ( Supplemental Figures 1, 2, and 3 for Bland-Altman plots, available at AJO.com ). There were significant differences between measurements only in total ocular aberrations and defocus during sessions 1 and 2 ( Supplemental Table , available at AJO.com ).
| Aberration (μm) | Overall Mean (Range) | S w | Precision (1.96 × S w ) | Reproducibility (2.77 × S w ) | ICC (95% CI) |
|---|---|---|---|---|---|
| TOA RMS | 2.07 (0.43 to 6.69) | 0.165 | 0.323 | 0.458 | 0.99 (0.98 to 0.99) |
| HOA RMS | 0.34 (0.14 to 0.80) | 0.042 | 0.082 | 0.117 | 0.91 (0.82 to 0.95) |
| HOA RMS (Z 4 0 ) | 0.30 (0.14 to 0.80) | 0.040 | 0.078 | 0.125 | 0.88 (0.76 to 0.94) |
| Z 2 0 | −0.915 (−6.430 to 6.602) | 0.209 | 0.409 | 0.580 | 0.99 (0.98 to 0.99) |
| Z 2 −2 | 0.024 (−0.698 to 1.880) | 0.076 | 0.148 | 0.210 | 0.97 (0.94 to 0.98) |
| Z 2 2 | 0.247 (−0.848 to 2.034) | 0.120 | 0.235 | 0.332 | 0.97 (0.95 to 0.98) |
| Z 3 −1 | −0.078 (−0.422 to 0.240) | 0.050 | 0.098 | 0.138 | 0.87 (0.76 to 0.94) |
| Z 3 1 | −0.009 (−0.383 to 0.524) | 0.038 | 0.074 | 0.105 | 0.95 (0.89 to 0.97) |
| Z 3 −3 | 0.105 (−0.102 to 0.297) | 0.040 | 0.078 | 0.112 | 0.90 (0.81 to 0.95) |
| Z 3 3 | 0.008 (−0.268 to 0.188) | 0.050 | 0.098 | 0.145 | 0.79 (0.61 to 0.89) |
| Z 4 0 | −0.111 (−0.417 to 0.153) | 0.040 | 0.078 | 0.116 | 0.90 (0.81 to 0.95) |
| Z 4 −2 | −0.005 (−0.074 to 0.087) | 0.024 | 0.047 | 0.067 | 0.59 (0.30 to 0.78) |
| Z 4 2 | −0.011 (−0.204 to 0.096) | 0.029 | 0.056 | 0.080 | 0.83 (0.67 to 0.91) |
| Z 4 −4 | −0.004 (−0.126 to 0.058) | 0.028 | 0.054 | 0.077 | 0.66 (0.41 to 0.82) |
| Z 4 4 | −0.009 (−0.100 to 0.314) | 0.030 | 0.058 | 0.083 | 0.83 (0.67 to 0.91) |
| Z 5 −1 | −0.002 (−0.134 to 0.093) | 0.018 | 0.035 | 0.050 | 0.79 (0.60 to 0.89) |
| Z 5 1 | 0.002 (−0.056 to 0.047) | 0.017 | 0.033 | 0.046 | 0.58 (0.28 to 0.77) |
| Z 5 −3 | 0.002 (−0.061 to 0.073) | 0.014 | 0.027 | 0.040 | 0.77 (0.58 to 0.88) |
| Z 5 3 | 0.001 (−0.067 to 0.069) | 0.018 | 0.035 | 0.050 | 0.40 (0.06 to 0.66) |
| Z 5 −5 | to 0.005 (−0.120 to 0.041) | 0.019 | 0.037 | 0.054 | 0.71 (0.48 to 0.85) |
| Z 5 5 | 0.005 (−0.066 to 0.063) | 0.017 | 0.033 | 0.048 | 0.58 (0.28 to 0.77) |
| Aberration (μm) | Upper LoA (95% CI) | Lower LoA (95% CI) | Width of (95%) LoA |
|---|---|---|---|
| TOA RMS | 0.327 (0.187 to 0.467) | −0.522 (−0.661 to −0.382) | 0.849 |
| HOA RMS | 0.124 (0.085 to 0.163) | −0.114 (−0.153 to −0.075) | 0.238 |
| HOA RMS (Z 4 0 ) | 0.128 (0086 to 0.169) | −0.126 (−0.167 to −0.084) | 0.254 |
| Z 2 0 | 0.654 (0.493 to 0.815) | −0.323 (−0.484 to −0.162) | 0.977 |
| Z 2 −2 | 0.227 (0.157 to 0.296) | −0.195 (−0.264 to −0.126) | 0.422 |
| Z 2 2 | 0.326 (0.215 to 0.437) | −0.349 (−0.460 to −0.237) | 0.675 |
| Z 3 −1 | 0.149 (0.102 to 0.195) | −0.131 (−0.177 to −0.085) | 0.280 |
| Z 3 1 | 0.112 (0.077 to 0.147) | −0.101 (−0.135 to −0.066) | 0.213 |
| Z 3 −3 | 0.102 (0.065 to 0.139) | −0.122 (−0.159 to −0.085) | 0.224 |
| Z 3 3 | 0.135 (0.087 to 0.183) | −0.156 (−0.204 to −0.108) | 0.291 |
| Z 4 0 | 0.099 (0.061 to 0.137) | −0.129 (−0.167 to −0.092) | 0.228 |
| Z 4 −2 | 0.070 (0.048 to 0.093) | −0.064 (−0.086 to −0.042) | 0.134 |
| Z 4 2 | 0.077 (0.051 to 0.104) | −0.084 (−0.111 to −0.058) | 0.161 |
| Z 4 −4 | 0.073 (0.047 to 0.099) | −0.082 (−0.108 to −0.057) | 0.155 |
| Z 4 4 | NA a | NA a | NA a |
| Z 5 −1 | 0.048 (0.031 to 0.064) | −0.053 (−0.070 to −0.037) | 0.101 |
| Z 5 1 | 0.042 (0.027 to 0.057) | −0.050 (−0.065 to −0.035) | 0.092 |
| Z 5 −3 | 0.044 (0.030 to 0.057) | −0.037 (−0.050 to −0.024) | 0.081 |
| Z 5 3 | 0.049 (0.032 to 0.065) | −0.052 (−0.068 to −0.035) | 0.101 |
| Z 5 −5 | 0.049 (0.031 to 0.066) | −0.059 (−0.077 to −0.041) | 0.108 |
| Z 5 5 | 0.046 (0.030 to 0.062) | −0.050 (−0.066 to −0.035) | 0.096 |
a See Bland M. An Introduction to Medical Statistics. 3rd ed. Oxford, UK: Oxford University Press; 2000.
Results
Intrasession Repeatability
The repeatability was studied in 30 eyes of 30 patients (15 men, 15 women; average age, 32.20 ± 9.34 years; range, 20 to 55 years). The mean ametropia was −4.26 ± 1.94 diopters (D; range, −7.50 to 1.50 D); the mean astigmatism was −0.93 ± 0.83 D (range, −3.00 to −0.00 D). Table 1 shows the overall average ocular aberration values corresponding to the repeated measures, the intrasession S w , the precision, the repeatability, and the intraclass correlation coefficient.
| Aberration (μm) | Overall Mean (Range) | S w | Precision (1.96 × S w ) | Repeatability (2.77 × S w ) | ICC (95% CI) |
|---|---|---|---|---|---|
| TOA RMS | 6.60 (0.47 to 13.76) | 0.131 | 0.256 | 0.362 | 0.99 (0.99 to 0.99) |
| HOA RMS | 0.44 (0.14 to 0.81) | 0.039 | 0.076 | 0.109 | 0.96 (0.93 to 0.98) |
| HOA RMS (Z 4 0 ) | 0.37 (0.12 to 0.73) | 0.039 | 0.076 | 0.109 | 0.94 (0.90 to 0.97) |
| Z 2 0 | −6.405 (−13.190 to −0.029) | 0.130 | 0.25 | 0.360 | 0.99 (0.99 to 0.99) |
| Z 2 −2 | 0.079 (−1.309 to 3.284) | 0.069 | 0.135 | 0.191 | 0.99 (0.98 to 0.99) |
| Z 2 2 | 0.582 (−1.351 to 3.703) | 0.089 | 0.174 | 0.246 | 0.99 (0.99 to 0.99) |
| Z 3 −1 | −0.016 (−0.625 to 0.354) | 0.116 | 0.227 | 0.321 | 0.82 (0.72 to 0.89) |
| Z 3 1 | 0.008 (−0.297 to 0.408) | 0.035 | 0.068 | 0.097 | 0.96 (0.93 to 0.98) |
| Z 3 −3 | 0.073 (−0.143 to 0.363) | 0.054 | 0.105 | 0.150 | 0.87 (0.80 to 0.93) |
| Z 3 3 | −0.009 (−0.293 to 0.289) | 0.042 | 0.082 | 0.117 | 0.90 (0.84 to 0.94) |
| Z 4 0 | −0.161 (−0.696 to 0.124) | 0.067 | 0.131 | 0.186 | 0.90 (0.84 to 0.94) |
| Z 4 −2 | 0.010 (−0.102 to 0.127) | 0.019 | 0.037 | 0.053 | 0.87 (0.80 to 0.93) |
| Z 4 2 | 0.028 (−0.130 to 0.168) | 0.022 | 0.043 | 0.060 | 0.88 (0.82 to 0.94) |
| Z 4 −4 | −0.001 (−0.066 to 0.145) | 0.022 | 0.043 | 0.060 | 0.83 (0.73 to 0.90) |
| Z 4 4 | −0.019 (−0.105 to 0.093) | 0.029 | 0.056 | 0.080 | 0.71 (0.58–0.83) |
| Z 5 −1 | −0.004 (−0.244 to 0.096) | 0.021 | 0.041 | 0.059 | 0.85 (0.76 to 0.91) |
| Z 5 1 | −0.001 (−0.062 to 0.053) | 0.016 | 0.031 | 0.044 | 0.59 (0.43 to 0.74) |
| Z 5 −3 | 0.007 (−0.061 to 0.169) | 0.020 | 0.039 | 0.055 | 0.69 (0.55 to 0.82) |
| Z 5 3 | 0.006 (−0.039 to 0.041) | 0.016 | 0.031 | 0.045 | 0.54 (0.38 to 0.71) |
| Z 5 −5 | 0.007 (−0.096 to 0.091) | 0.020 | 0.039 | 0.056 | 0.64 (0.49 to 0.78) |
| Z 5 5 | 0.002 (−0.051 to 0.090) | 0.021 | 0.041 | 0.059 | 0.45 (0.28 to 0.64) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
