OPTOMETRY ORIGINAL PAPER

Transcription

1 C L I N I C A L A N D E X P E R I M E N T A L OPTOMETRY ORIGINAL PAPER Repeatability and agreement of two A-scan ultrasonic biometers and IOLMaster in non-orthokeratology subjects and post-orthokeratology children Clin Exp Optom 2006; 89: 3: Ben Chan BSc(Hons)Optom Pauline Cho PhD FAAO Sin Wan Cheung MPhil FAAO School of Optometry, The Hong Kong Polytechnic University, Hong Kong SAR, China soben@polyu.edu.hk Submitted: 22 September 2005 Revised: 2 December 2005 Accepted for publication: 8 December 2005 DOI:1111/j x Purpose: Our aim was to determine the repeatability of measurements of axial length (AL) and anterior chamber depth (ACD) made with two ultrasonic biometers and the IOLMaster in a group of non-orthokeratology (ortho-k) adult subjects and to investigate the agreement among instruments in children undergoing ortho-k therapy and in children wearing spectacles. Methods: To determine repeatability, AL and ACD were measured twice in 22 non-orthok young adults using two A-scan ultrasonic biometers (A-5500 and A-2500) and the IOLMaster. To determine agreement, AL and ACD were measured with the same instruments in 30 children undergoing ortho-k therapy and 30 spectacle-wearing children. Results: In the adult subjects, there were no significant differences in ACD and AL measurements obtained from the three instruments (repeated measures ANOVAs, p > 5). There was also no significant between-measurement difference for each instrument. The between-measurement agreement was better for the IOLMaster (95% limits of agreement (LA): -4 and +5 mm for both AL and ACD) than for the two A-scan ultrasonic biometers (95% LA: 2 and +1 mm for AL; 2 and +7 mm for ACD). Among the children, AL measurements with all three instruments were not significantly different from each other for both the children undergoing ortho-k therapy and those wearing spectacles (repeated measures ANOVAs, p > 5). The 95% LA of differences obtained from any two instruments were also comparable for both groups of subjects (within 0 mm and +0 mm). ACD measurements of the children were significantly different among the three instruments (repeated measures ANOVAs, p < 5). No significant differences in ACD measurements were found between A-5500 and A-2500 for both groups of children (paired t tests, p > 17). Conclusions: The repeatability of AL and ACD measurements with the IOLMaster was very good, and was better than with the A-scan ultrasonic biometers. The agreements in AL measurements between A-scan ultrasonic biometers and IOLMaster were comparable in both the ortho-k and the spectacle-wearing subjects, and were comparable to the repeatability of the A-scan ultrasonic biometers. ACD measurements between A-scan ultrasonic biometry and the IOLMaster were not comparable. AL measurements with the IOLMaster can replace the measurements from the two A-scan ultrasonic biometers used, however, the reverse is not true. AL and ACD measurements with all three instruments were unaffected by the flattened cornea following ortho-k lens wear. Key words: axial length, anterior chamber depth, orthokeratology, A-scan ultrasonic biometer, IOLMaster, children 160

2 Axial length (AL) is the measurement of the length of the eyeball along the ocular axis of symmetry, which is the distance from the cornea to the retinal surface. It is generally accepted that AL elongation is the major cause of myopic progression; the longer the AL, the higher the myopia. The significant relationship between AL elongation and increase in myopia is well documented. 1 6 Grosvenor and Scott 5 reported that all myopia is axial in origin and that myopia due to an increase in corneal or lens power without AL elongation is non-existent. Measurement of anterior chamber depth (ACD) not only provides useful information in screening for glaucoma 7 but also provides necessary information for the calculation of intra-ocular lens power in cataract surgery. 8 ACD is unchanged with increases in myopia. 9 A-scan ultrasonic biometry is an objective method used for AL measurement and it gives both AL and ACD values in a single measurement by calculating the return time of ultrasonic echoes from different ocular media. 10 It is an invasive technique that requires the use of topical anaesthetic and contact of the measuring probe with the corneal surface. Most topical anaesthetics will cause an initial stinging or burning sensation and irritation 11 and some patients may have an allergic response to the drug, resulting in acute conjunctival injection. 12 One possible source of error in ultrasonic AL and ACD measurements is the contact of the probe on the corneal surface, which can cause corneal indentation if excessive pressure is exerted on the cornea. Misalignment of the ultrasonic probe is another source of error A non-contact instrument is currently available for both AL and ACD measurements the IOLMaster (Zeiss Humphrey System, CA, USA). This instrument is designed for calculating the power of intraocular lenses. As the technique requires no contact with the corneal surface, topical anaesthesia is unnecessary. A detailed explanation of the principle and design of the IOLMaster has been published elsewhere Using the IOLMaster, AL is measured optically based on the technique known as partial coherence interferometry (PCI) using an infrared beam of wavelength 780 nm. The dual beam of the split coherence length is transmitted and reflected from different ocular media. The time delay taken from the reflection off the cornea and the retinal pigment epithelium is converted to a geometric AL by dividing the optical path length with a mean refractive index of the refractive media. Dual beam PCI technique is unaffected by longitudinal eye movement during AL measurement as the cornea is used as a reference surface. 14,18 Therefore, repeatable and precise results are obtained when compared with A-scan ultrasonic biometry, which uses acoustic waves. 17,21 The IOLMaster measures ACD using a slit-beam photographic technique. A flickering lateral slit beam along the visual axis produces a cross-sectional image of the anterior eye segment. Although measurement of ACD with this technique has been found to have a good agreement with PCI in the phakic eye, a large variation has been reported when compared with ultrasonic biometry. 19 AL and ACD measurements made with A-scan ultrasonic biometers and the IOLMaster in normal adults have been reported to be repeatable and reproducible. 17,22 25 However, there are few articles in the literature comparing the repeatability of both A-scan ultrasonic biometry and the IOLMaster. 17,21 A number of studies have also shown good agreement between the IOLMaster and A-scan ultrasonic biometry in measuring AL Some studies reported deeper ACD values measured with the IOLMaster compared to those with A-scan ultrasonic biometry. 17,19,21,22,26 Lam and colleagues 22 measured ACD using the two techniques in a group of young adults, who would be expected to have active accommodation but interestingly, the results were similar to the recent studies where measurements were performed in young adults 21 and children 17 under cycloplegia. ACD is expected to be affected by the status of the crystalline lens that is, ACD is shallower when the eye accommodates and vice versa. 27 Normally children have active accommodation and hence ACD measured without cycloplegic would be expected to be shallower than those made under cycloplegia. To date, it is unclear if ACD of children measured under non-cycloplegic and cycloplegic conditions using A-scan ultrasonic biometry would differ. As A-scan ultrasonic biometry provides both AL and ACD in a single measurement and the IOLMaster incorporates a measuring mode for ACD, one of the aims of this study was to compare the ACD measurements with the two techniques in children without cycloplegia. Orthokeratology (Ortho-k) gives a temporary reduction of myopia. 28 The treatment involves changing the shape of the cornea using a specially-designed rigid gas permeable contact lens. The central cornea is flattened and thinned, resulting in a reduction of refractive power of the eye Ortho-k has also been shown to be effective in slowing myopic progression. 33 In Hong Kong, most ortho-k patients are children undergoing the treatment for myopic control. 34 The use of the non-contact IOLMaster would be the preferred option for the measurement of AL in young children, if the results obtained are repeatable and comparable to those obtained with ultrasonic biometry. After ortho-k lens wear, a flattened cornea may affect the alignment of the measuring probe in A-scan ultrasonic and the ACD measurements using the IOLMaster. Previous studies on the repeatability and reproducibility of the two techniques were limited to subjects with normal corneas. It is unclear whether different results would be obtained in subjects with corneas that are flattened after ortho-k lens wear. Therefore, it is important to know the relationship between the results obtained from the IOLMaster and A-scan ultrasonic biometry in children with normal corneas or corneas flattened after commencement of ortho-k treatment. To allow us to determine the between-instrument agreements, we also need to know the repeatability (betweenmeasurement agreement) of each instrument. So, additional aims of this study were to determine the repeatability of AL and ACD measurements with the two A-scan ultrasonic biometers and the IOLMaster 161

3 in a group of normal young adults and to evaluate the levels of agreement among these measurements with the three instruments in children undergoing ortho-k (flattened corneas) and children wearing spectacles (normal corneas). SUBJECTS AND METHODS Orthokeratology (n = 30) Spectacle-wearing (n = 30) Age (years) 12.1 ± ± 2.10 Refractive sphere (D) ± ± 1.13 Refractive cylinder (D) 7 ± 8 8 ± 9 Spherical equivalent refraction (D) ± ± 1.18 Table 1. Age and subjective refraction (mean ± SD) of the orthokeratology (pre-therapy) and spectacle-wearing children Repeatability Twenty-two young adult subjects (8 male, 14 female) of mean age of 23.5 ± 2.2 (SD) years were recruited. Subjects had no previous ortho-k experience. In a single visit, two representative measurements were taken with each of the three instruments A-5500 ultrasonic biometer and A-2500 ultrasonic biometer (Sonomed Inc., NY, USA) and the IOLMaster by the same examiner. For the measurements with the A-scan biometers, a resting interval of about one minute was allowed between measurements. For the IOLMaster, the subject was asked to re-position his/her head after each measurement. Agreement among instruments Thirty ortho-k subjects (14 male, 16 female) with a mean age of 12.1 ± 1.8 years and 30 spectacle-wearing subjects (11 male, 19 female) of mean age 1 ± 2.1 years were recruited. The two groups of subjects were matched for refractive error (spherical equivalent). All the ortho-k subjects had been wearing ortho-k lenses for at least six months and the spectacle-wearing subjects had no previous experience in contact lens wear. Table 1 shows the demographics of the subjects in the two study groups. All children were recruited from the Optometry Clinic of The Hong Kong Polytechnic University. AL and ACD were measured with the three instruments one representative measurement per instrument. All subjects had good general and ocular health and were not on medication. The tenets of the Declaration of Helsinki as revised in 2002 were followed and ethics approval was obtained from the Departmental Research Committee of the Department of Optometry and Radiography (now School of Optometry) of The Hong Kong Polytechnic University before the recruitment of subjects. Informed consent was obtained from each subject after the nature of the study had been fully explained and before the commencement of the study. Only the right eye was measured and used for analysis. For each subject, unaided (ortho-k group) or habitual (young adult subjects or spectacle-wearing group) vision and monocular subjective refraction were taken, followed by slitlamp biomicroscopy. All subjects had less than Grade 1 staining in the central cornea and less than Grade 2 in the peripheral cornea (Efron Scale) 35 prior to AL and ACD measurements. AL and ACD were measured with the IOLMaster first, followed by A-scan ultrasonic biometry. For the IOLMaster, AL and ACD measurements were determined according to the manufacturer s instructions. For AL measurement, five consecutive readings were made for each measurement; each reading showing signal-to-noise ratio above 2.0 and differing by less than 0 mm from the others. For ACD measurement, corneal curvature was measured before activating the ACD measuring mode; five ACD readings were obtained for each measurement; any reading which varied by more than 5 mm was deleted and ACD re-measured (as recommended by the manufacturer). The averages of the AL and ACD readings were used for analysis. Prior to determination of the AL and ACD with the A-scan ultrasonic biometers, the cornea was anaesthetised with % benoxinate and the measuring probe was sterilised with 70 per cent alcohol swab. The order of measurements with the two instruments was random. During measurements, all subjects were instructed to fixate at a distance target six metres away and the measuring probe was kept perpendicular to the cornea for each measurement. Five readings with a standard deviation less than 0 mm (as recommended by the manufacturer) were obtained and the average value recorded for analysis. Corneal integrity was re-examined at the end of the study to ensure that there were no clinically significant corneal abrasions after A-scan biometry. No subject exhibited significant corneal staining after the measurements. Treatment of data The distributions of the data were not significantly different from normal (Kolmogorov-Smirnov tests, p > 5), so parametric tests were used for statistical analysis. Repeatability Repeated measures analysis of variance (ANOVA) was used to test the betweenmeasurement differences for all three instruments in normal adults. To assess the repeatability, distribution of the differences in AL and ACD measurements and the relationship between the differences and their means were determined. If the differences were normally distributed and if there was no significant relationship between the differences and the means, a graph of differences versus means was plotted. The 95% limits of agreement (LA) were determined 36 as a measure of repeatability of each instrument. 162

4 Axial length (mm, mean ± SD) Anterior chamber depth (mm, mean ± SD) A-5500 A-2500 IOLMaster A-5500 A-2500 IOLMaster Measurement ± ± ± ± ± ± 3 Measurement ± ± ± ± ± ± 3 Difference -06 ± ± ± ± ± ± 25 95% limits of agreement 2 to +1-9 to +0-4 to +5 7 to +9 2 to +7-4 to +5 Table 2. Repeatability of axial length and anterior chamber depth measurements with the three instruments in young adult subjects (n = 22) A. A-5500 B. A-2500 C. IOL Master Figure 1. Plots of the between-measurement differences versus means for axial length for each instrument in 22 young adult subjects: (A) A-5500, (B) A-2500, (C) IOLMaster. The dotted lines represent the lower and upper limits of agreement (mean difference ± 1.96 standard deviation of the differences). Each solid line represents the mean of the differences. Agreement among instruments ANOVA was used to analyse measurements among instruments for the spectacle-wearing children and children undergoing ortho-k therapy. If a significant difference was found, post-hoc tests with Bonferroni correction were performed to identify which pair(s) were significantly different. Pearson correlations, with Bonferroni correction for multiple correlations where appropriate, were used to study the association between instruments. To evaluate agreement among instruments, the distribution of the differences in AL and ACD and the relationship between the differences and their means were determined. If the differences were normally distributed and if there was no significant relationship between the differences and the means, a graph of differences versus means was plotted. The 95% limits of agreement were determined 36 to assess the extent to which the measurements from any two instruments agreed with each other. RESULTS Repeatability Table 2 shows the results of AL and ACD measurements of the 22 young adults obtained from the three instruments. There were no statistically significant differences among AL and ACD measurements from the three instruments (repeated measures ANOVA (Wilks Lambda): AL: F (5, 17) = 75, p = 59; ACD: F (5, 17) = 35, p = 85) and between measurements for each instrument. Also, no significant correlations between the differences of the two measurements (in AL and ACD) and their means were found for all three instruments (AL: Pearson r: 8 < r < 0, p > 5; ACD: 3 < r < 3, p > 5). For A-scan ultrasonic biometry (A-5500 and A-2500), the 95% LA were 2 to +1 mm for AL measurements, and 2 to +7 mm for ACD measurements (Table 2). For the IOLMaster, the 95% LA were -4 to +5 mm for both AL and ACD measurements. Figures 1 and 2 present the plots of between-measurement differences in AL and ACD against their means for each instrument. Agreement among instruments AL Table 3 presents a summary of AL results from the three instruments in ortho-k and spectacle-wearing children. There were no significant differences in AL measurements among instruments for each group of subjects (repeated measures ANOVA (Wilks Lambda): ortho-k: F (2, 28) = 0.59, p = 0.56; spectacle-wearing: F (2, 58) = 1.55, p = 2). There were no relationships among the differences in AL measured by different instruments and their means for both groups of subjects (ortho-k: Pearson r: 6 < r < -2, p > 5; spectacle-wearing: 8 < r < 3, p > 5) except for a weak negative correlation between IOLMaster and A in the spectacle-wearing group 163

5 A. A-5500 B. A-2500 C. IOL Master Figure 2. Plots of the between-measurement differences versus means for anterior chamber depth for each instrument in 22 adult subjects: (A) A-5500, (B) A-2500, (C) IOLMaster. The dotted lines represent the lower and upper limits of agreement (mean difference ± 1.96 x standard deviation of the differences). Each solid line represents the mean of the differences. (Pearson r = 5, p = 12). Figure 3 shows the plots of between-instrument differences in AL against their means with 95% LA. On average, the agreement between any two instruments was good the mean difference was equal to or smaller than 3 mm. The 95% LA of AL measurements in both groups of subjects are summarised in Table 3. No 95% LA was determined between A-2500 and IOL- Master for the spectacle-wearing children, as there was a significant correlation between their differences and means. Agreement among instruments ACD Table 3 also shows ACD results obtained from the three instruments in ortho-k and spectacle-wearing children. Results among instruments were significantly different for both groups of subjects (repeated measures ANOVA (Wilks Lamba): ortho-k: F (2, 58) = 37.82, p < 01; spectaclewearing: F (2, 28) = 73.06, p < 01). Posthoc tests with Bonferroni corrections indicated that the differences were mainly due to significantly deeper ACD measured with the A-5500 and A-2500 when compared to the IOLMaster by 8 and 5 mm, respectively for the ortho-k group (adjusted p < 17) and by 6 and 3 mm, respectively for the spectaclewearing group (p < 17). There was no significant difference in ACD measured by the A-5500 and A-2500 in both groups of subjects (p > 17). The 95% LA between the two A-scan ultrasonic biometers were 0 to +8 mm for the ortho-k subjects and 2 to +0 mm for the spectaclewearing subjects (Table 3 and Figure 4). There were significant differences in ACD measurements among instruments for all subjects and the measurements among instruments were significantly correlated (Pearson r > 2, p < 01). DISCUSSION The results of current study are consistent with recent studies 17,21 that the repeatability of the IOLMaster is better than A-scan ultrasonic biometry in AL and ACD measurements. Our results show better repeatability than those reported by previous investigators (Table 4). 17,21 The betweenmeasurement difference in AL was less than per cent for either A-scan ultrasonic biometer and less than per cent for the IOLMaster. Therefore, for AL assessment, measurements from all three instruments are repeatable but the performance of the IOLMaster is superior to the A-scan biometers. For ACD measurements the variation can be up to 7.3 per cent for either A-scan ultrasonic biometer but not more than 1.4 per cent for the IOLMaster. Therefore, the repeatability of ACD with the A-scan ultrasonic biometer is unsatisfactory and ACD measured with the IOL- Master is far more reliable. For both AL and ACD measurements, the comparison of repeated measurements between the A-scan ultrasonic biometers revealed larger limits of agreement than comparison of repeated measurements with the IOLMaster. This means that AL and ACD measurements with A-scan ultrasonic biometry have a relatively greater range of variance. The comparatively poor repeatability of the A-scan biometers (compared to IOLMaster) may be due to the eye being in contact with the measuring probe. Even though a local anaesthetic was applied prior to measurement, involuntary eye movement during the measurement will affect the accuracy. It has been reported previously that a five degree off-axis shift in AL measurement will cause a 0 mm variation. 37 The accuracy between two sets of measurement was dependent on the experience of the examiner and whether he/she places the measuring probe on the same location each time. The better repeatability of the A-scan ultrasonic biometer in the current study (compared to results reported previously) may be due to: 1. The experience of our practitioner who has researched A-scan ultrasonic biometry for two years or 2. Different brands of A-scan ultrasonic biometers. Alternatively, the IOLMaster utilises optical coherence in determining AL and a slit-beam in determining ACD and neither 164

6 A-5500 A-2500 IOLMaster p* A-5500 vs. A-2500 A-5500 vs. IOLMaster A-2500 vs. IOLMaster MD ± SD # 95% limits of agreement MD ± SD # 95% limits of agreement MD ± SD # 95% limits of agreement Orthokeratology (n = 30) Axial length (mm) ± ± ± ± 5 0 to +0 2 ± 9 6 to +0 2 ± 8 4 to ± ± ± 1 <17 4 ± 2 0 to +8 8 ± 0 5 ± 4 Anterior chamber depth (mm) Spectacle-wearing (n = 30) Axial length (mm) ± ± ± ± 9 5 to +1 1 ± 7 3 to +5-2 ± ± ± ± 3 <17 4 ± 8 2 to +0 6 ± 7 3 ± 1 Anterior chamber depth (mm) * p: probability values of repeated measures ANOVA for differences in measurements among instruments post-hoc tests showed the differences to be due to deeper measurements with A-scan ultrasonic biometry than the IOLMaster Mean difference ± standard deviation No limits of agreement were determined due to significant differences between measurements or significant correlation between differences and means. # Table 3. Summary of axial length and anterior chamber depth measurements from the A-scan biometers and the IOLMaster for the two groups of children technique involves contact with the eye, resulting in less eye movement. Moreover, the internal fixation target ensures that each measurement is taken close to the visual axis. hence, more accurate and repeatable results can be obtained with this instrument. On average, there was good agreement (mean close to zero) in AL measurements among the three instruments for both ortho-k and spectacle-wearing children. This suggested that AL determined with these instruments may not be affected by the reshaped cornea after ortho-k. Compared to previous results in adults, 17,21 23 our findings showed better agreement between A-scan ultrasonic biometers (either A-5500 or A-2500) and IOLMaster. Moreover, in the present study, the 95% LA between the two A-scan biometers were 0 to +0 mm (ortho-k) and 5 to +1 mm (spectacle-wearing) and these findings were reasonably close to the repeatability results of A-5500 and A-2500 (from 2 to +1 mm). This suggests that, where appropriate or necessary, AL measurements with A-5500 and A-2500 may be used interchangeably. Taking into account the repeatability of AL measurements in the current study, measurements with the IOLMaster may be used to replace those from A-5500 or A- 2500, if necessary, as there were no significant differences in AL measurements between A-scan and IOLMaster, and the 95% LA of between-instrument differences were reasonably comparable to the 95% LA of between-measurement differences with the A-scan biometer. However, the reverse is not true; measurements made by IOLMaster cannot be replaced by measurements from A-scan, as the 95% LA of between-instrument differences were not comparable to the 95% LA of between-measurement differences with the IOLMaster. The IOLMaster may replace A-scan ultrasonic biometry for AL measurement, however, the cost of IOLMaster is far greater than the A-scan ultrasonic biometer and not every optometric practice can afford to install such an instrument. In our study, we found that there is a good agreement in AL measurement between the two 165

7 - - A. A-5500 A-2500 (orthokeratology) C. A-2500 IOL Master (orthokeratology) - Figure 3. Plots of the differences versus means of axial length with any two instruments in 30 orthokeratology children: (A) A-5500 and A-2500, (B) A-5500 and IOLMaster, (C) A-2500 and IOLMaster; in 30 spectacle wearing children: (D) A-5500 and A-2500, (E) A-5500 and IOLMaster. The dotted lines represent the lower and upper limits of agreement (mean difference ± 1.96 standard deviation of the differences). Each solid line represents the mean of the differences. techniques on flattened corneas resulting from ortho-k treatment. Hence, A-scan ultrasonic biometry can still be used for monitoring eye elongation in ortho-k patients. Our results also showed that ACD measurements with the IOLMaster were - E. A-5500 IOL Master (spectacle wearing) - B. A-5500 IOL Master (orthokeratology) D. A-5500 A-2500 (spectacle wearing) significantly less (shallower) than those measured with the A-5500 and A-2500 in both groups of children. ACD determined from these instruments did not appear to be affected by the reshaped corneas after ortho-k. Although the ACD differences measured with IOLMaster and A-scan ultrasonic biometry are in agreement with the findings of Santodomingo-Rubido and colleagues, 23 other investigators have reported significantly deeper ACD with the IOLMaster than with the A-scan ultrasonic biometer. 21,22,26 Some authors claimed that this may due to the silt-beam technique used in the IOLMaster the slit illumination projects from the lateral side of the eye resulting in larger ACD measurements. 19,22 It may also be due to the different axes of measurement in the two techniques, which resulted in measurements made along the visual axis in the IOLMaster but along the optical axis in A-scan ultrasonic biometry. 19,21,38 Lam and colleagues 22 also speculated that the error in determining the corneal curvature by IOLMaster may lead to shallower ACD measurements, as the calculation of the ACD involves the input of the corneal radius. The study by Vogel and colleagues 24 reported that there was no significant difference in corneal curvature measured with the IOLMaster and the conventional keratometer. In the current study, accommodation in young children during measurements with the IOLMaster may contribute to a shallower ACD measurement compared to A-scan ultrasonic biometry. Carkeet and colleagues 17 reported deeper ACD measurements with the IOLMaster than with an ultrasonic biometer in children under cycloplegia. During ACD measurements with the IOLMaster, subjects were asked to fixate on the internal fixation light, whereas in A-scan ultrasonic biometry, they were asked to fixate a target a few metres away. The close proximity of the internal fixation target in the IOLMaster may induce accommodation in our young subjects resulting in a thickening of the crystalline lens, causing the lens to move forward and resulting in a shorter ACD. This is supported by a recent study of Sheng and colleagues, 21 who compared the pre- and post-cycloplegic ACD from IOLMaster and A-scan biometry in a group of young adults. Their results showed that eyes under cycloplegia produced significantly deeper ACD values with the IOLMaster. Our results showed that the ACD measured with A-scan ultrasonic biometry 166

8 Author Repeatability (mm) Agreement (mm) A-scan biometry IOLMaster A-scan biometry vs IOLMaster Axial length Anterior chamber depth Axial length Anterior chamber depth Axial length Current 1 to +1 2 to +7-4 to +5-4 to +5 6 to +0 Carkeet et al 17 5 to to to to to +4 Sheng et al 21 5 to +5 9 to +1-6 to +5-6 to +4-9 to +4 Lam et al 22 5 to +6 Santodomingo- Rubido et al 23 1 to +5 Table 4. Summary of the 95 per cent limits of agreement of current and previous studies A. Orthokeratology B. Spectacle wearing change in the corneal refractive index after ortho-k can influence axial biometry, further studies are necessary to clarify if there is a need to change the refractive index used in calculating biometric measurements following ortho-k Figure 4. Plots of the differences versus means of anterior chamber depths between A-5500 and A-2500 in (A) orthokeratology and (B) spectacle-wearing children. The dotted lines represent the lower and upper limits of agreement (mean difference ± 1.96 standard deviation of the differences). Each solid line represents the mean of the differences. and the IOLMaster were significantly different. Hence, the results from these instruments were not comparable and therefore, we did not determine the limits of agreement between A-scan biometers and the IOLMaster. Our findings indicated that ACD measurements from A and A-2500 were comparable for both the ortho-k and the spectaclewearing group (Table 3). ACD measured with A-5500 and A-2500 may be used interchangeably as the variations in between-instrument measurements were comparable to the between-measurement differences A change of corneal refractive index after corneal reshaping by ortho-k may affect measurements with the A-scan, however, there appear to be no reports on this matter. One study reported that the corneal refractive index changed after photorefractive keratectomy (PRK). 39 Like ortho-k, PRK reduces the degree of myopia by thinning and flattening the central cornea. Thinning of the central epithelium after ortho-k may result in close packing of the corneal tissue and this modification of the organisation of corneal tissue may effect the refractive index. As currently there are no data on how CONCLUSIONS AL and ACD measurements with the IOL- Master are more repeatable than those with the A-scan ultrasonic biometers (A and A-2500). For AL measurements, less than per cent between-measurement variation is expected for any of the three instruments, however, a significant variation is expected for ACD measurements. AL measurements with both the A-scan ultrasonic biometer and the IOL- Master were comparable in both the orthok and the spectacle-wearing subjects, and were comparable to the repeatability of either biometer. ACD measurements with the two techniques are not comparable, probably due to the relatively poor repeatability of ACD using the A-scan biometers. AL or ACD measurements with the A-scan ultrasonic biometers or the IOLMaster appear to be unaffected by the reshaped corneal surface resulting from ortho-k treatment. GRANTS AND FINANCIAL SUPPORT This study was supported by a grant from The Hong Kong Polytechnic University (A356). 167

9 REFERENCES 1. Adams AJ. Axial length elongation, not corneal curvature, as a basis of adult onset myopia. Am J Optom Physiol Opt 1987; 64: Grosvenor T. High axial length/corneal radius ratio as a risk factor in the development of myopia. Am J Optom Physiol Opt 1988; 65: Hosaka A. The growth of the eye and its components. Japanese studies. Acta Ophthalmol Scand Suppl 1988; 185: Lam CS, Goh WS. The incidence of refractive errors among school children in Hong Kong and its relationship with the optical components. Clin Exp Optom 1991; 74: Grosvenor T, Scott R. Three-year changes in refraction and its components in youthonset and early adult-onset myopia. Optom Vis Sci 1993; 70: Lin LL, Shih YF, Lee YC, Hung PT, Hou PK. Changes in ocular refraction and its components among medical students-a 5- year longitudinal study. Optom Vis Sci 1996; 73: Devereux JG, Foster PJ, Baasanhu J, Uranchimeg D, Lee PS, Erdenbeleig T, Machin D, Johnson GJ, Alsbirk PH. Anterior chamber depth measurement as a screening tool for primary angle-closure glaucoma in an East Asian population. Arch Ophthalmol 2000; 118: Haigis W, Lege B, Miller N, Schneider B. Comparison of immersion ultrasound biometry and partial coherence interferometry for intraocular lens calculation according to Haigis. Graefes Arch Clin Exp Ophthalmol 2000; 238: Rabsilber TM, Becker KA, Frisch IB, Auffarth GU. Anterior chamber depth in relation to refractive status measured with the Orbscan II Topography System. J Cataract Refract Surg 2003; 29: Storey JK, Rabie EP. Ultrasound a research tool in the study of accommodation. Ophthalmic Physiol Opt 1983; 3: Perez-Gomez I, Hollingsworth J, Efron N. Effects of benoxinate hydrochloride % on the morphological appearance of the cornea using confocal microscopy. Contact Lens Ant Eye 2004; 27: Norden LC. Adverse reactions to topical ocular anesthetics. J Am Optom Assoc 1976; 47: Findl O, Kriechbaum K, Sacu S, Kiss B, Polak K, Nepp J, Schild G, Rainer G, Maca S, Petternel V, Lackner B, Drexler W. Influence of operator experience on the performance of ultrasound biometry compared to optical biometry before cataract surgery. J Cataract Refract Surg 2003; 29: Hitzenberger CK, Drexler W, Dolezal C, Skorpik F, Juchem M, Fercher AF, Gnad HD. Measurement of the axial length of cataract eyes by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1993; 34: Tehrani M, Krummenauer F, Blom E, Dick HB. Evaluation of the practicality of optical biometry and applanation ultrasound in 253 eyes. J Cataract Refract Surg 2003; 29: Foster PJ, Alsbirk PH, Baasanhu J, Munkhbayar D, Uranchimeg D, Johnson GJ. Anterior chamber depth in Mongolians: variation with age, sex, and method of measurement. Am J Ophthalmol 1997; 124: Carkeet A, Saw SM, Gazzard G, Tang W, Tan DT. Repeatability of IOLMaster biometry in children. Optom Vis Sci 2004; 81: Drexler W, Findl O, Menapace R, Rainer G, Vass C, Hitzenberger CK, Fercher AF. Partial coherence interferometry: a novel approach to biometry in cataract surgery. Am J Ophthalmol 1998; 126: Kriechbaum K, Findl O, Kiss B, Sacu S, Petternel V, Drexler W. Comparison of anterior chamber depth measurement methods in phakic and pseudophakic eyes. J Cataract Refract Surg 2003; 29: Hitzenberger CK. Optical measurement of the axial eye length by laser Doppler interferometry. Invest Ophthalmol Vis Sci 1991; 32: Sheng H, Bottjer CA, Bullimore MA. Ocular component measurement using the Zeiss IOLMaster. Optom Vis Sci 2004; 81: Lam AK, Chan R, Pang PCK. The repeatability and accuracy of axial length and anterior chamber depth measurements from the IOLMaster. Ophthalmic Physiol Opt 2001; 21: Santodomingo-Rubido J, Mallen EA, Gilmartin B, Wolffsohn JS. A new non-contact optical device for ocular biometry. Br J Ophthalmol 2002; 86: Vogel A, Dick HB, Krummenauer F. Reproducibility of optical biometry using partial coherence interferometry: intraobserver and interobserver reliability. J Cataract Refract Surg 2001; 27: Rudnicka AR, Steele CF, Crabb DP, Edgar DF. Repeatability, reproducibility and intersession variability of the Allergan Humphrey ultrasonic biometer. Acta Ophthalmol Scand 1992; 70: Nemeth J, Fekete O, Pesztenlehrer N. Optical and ultrasound measurement of axial length and anterior chamber depth for intraocular lens power calculation. J Cataract Refract Surg 2003; 29: Gao L, Zhuo X, Kwok AK, Yu N, Ma L, Wang J. The change in ocular refractive components after cycloplegia in children. Jpn J Ophthalmol 2002; 46: Nichols JJ, Marsich MM, Nguyen M, Barr JT, Bullimore MA. Overnight orthokeratology. Optom Vis Sci 2000; 77: Coon LJ. Orthokeratology. Part II: Evaluating the Tabb method. J Am Optom Assoc 1984; 55: Swarbrick HA, Wong G, O Leary DJ. Corneal response to orthokeratology. Optom Vis Sci 1998; 75: Sridharan R, Swarbrick H. Corneal response to short-term orthokeratology lens wear. Optom Vis Sci 2003; 80: Alharbi A, Swarbrick HA. The effects of overnight orthokeratology lens wear on corneal thickness. Invest Ophthalmol Vis Sci 2003; 44: Cho P, Cheung SW, Edwards M. Longitudinal orthokeratology research in children (LORIC) in Hong Kong: A pilot study on refractive changes and myopic control. Curr Eye Res 2005; 30: Cho P, Cheung SW, Edwards MH. Practice of orthokeratology by a group of contact lens practitioners in Hong Kong-Part 1. General overview. Clin Exp Optom 2002; 85: Efron N. Clinical application of grading scales for contact lens complication. Optician 1997; 213(5604): Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1(8476): Longstaff S. Factors affecting intraocular lens power calculation. Trans Ophthalmol Soc U K 1986; 105: Tehrani M, Krummenauer F, Kumar R, Dick HB. Comparison of biometric measurements using partial coherence interferometry and applanation ultrasound. J Cataract Refract Surg 2003; 29: Ferrara G, Cennamo G, Marotta G, Loffredo E. New formula to calculate corneal power after refractive surgery. J Refract Surg 2004; 20: Corresponding author: Ben Chan School of Optometry The Hong Kong Polytechnic University Hong Kong, SAR CHINA soben@polyu.edu.hk 168