Measurement Accuracy When Using Spot Vision Screener With or Without Cycloplegia in Young Adults

Introduction

Amblyopia is a unilateral or bilateral reduction in best-corrected visual acuity that is not directly attributed to a structural abnormality of the eye or posterior visual pathways.¹ Unrecognized or untreated amblyopia in the early years of life can often lead to unsuccessful treatment and underdeveloped vision in the affected eye.^1,2 The main risk factors for amblyopia include strabismus and refractive error. The American Association for Pediatric Ophthalmology and Strabismus (AAPOS) has established a cutoff value for “amblyopia risk factor” (ARF) based on strabismus and refractive error.³ Although refractive error is the most common cause of amblyopia,⁴ no gold standard cutoff value for refractive error has been established. In 2022, AAPOS revised the new cutoff values for ARFs for refractive error.⁵

A device that is specifically designed to screen for amblyopia due to refractive error is the Spot Vision Screener (SVS; Welch Allyn, NY, USA), which measures refractive error using a photorefraction method.⁶ SVS is frequently used for screening because it can measure both eyes at the same time and has an extremely high success rate, even in children.^7,8 However, the characteristics of refraction obtained by SVS and the accuracy of screening have not been fully elucidated yet.

Screening for amblyopia of children with SVS has a sensitivity of 87%–95% and specificity of 71%–85% using the 2013 AAPOS criteria.^6,9–11 and a sensitivity of 88% and specificity of 78% using the revised 2022 AAPOS criteria.¹² However, when refractive error is measured using SVS under cycloplegic and noncycloplegic conditions, sensitivity and negative predictive value are higher in noncycloplegic conditions, and specificity and positive predictive value are higher in cycloplegic conditions.¹³

The high success rate of SVS measurements (even in children) is a major advantage for a refraction test,^7,8 and if the accuracy of SVS with cycloplegia is high, it could potentially be used as a diagnostic instrument as well. Yagasaki et al¹⁴ compared cycloplegic refraction measured by SVS and that measured by autorefractometer in children and reported that there was a proportional error in which the hyperopic power of the SVS decreased as hyperopia increased. However, because the measurement accuracy of SVS for cycloplegic refraction in patients with myopia is unknown, this study aimed to investigate the accuracy of SVS measurements with cycloplegia in young adults. The subjective refraction was measured and used as a reference standard, after which it was compared to that of the SVS and auto refractometers in young adults, where there may be numerous people with myopia.

Materials and Methods

Participants

The subjects included 49 university student volunteers who had no eye diseases. In this study, 45 right eyes of the subjects (35 women and 10 men) were analyzed. Four subjects whose refraction measured by SVS were outside the measurement range were excluded. The mean age of the study participants was 19.2 years (standard deviation, 0.3).

Methods

Objective refractions were initially measured by SVS, table-mounted autorefractometer (RT7000; TOMEY, Aichi, Japan), and handheld autorefractometer (Retinomax Screeen; Righton, Tokyo, Japan) at noncycloplegic conditions. Subjective refraction was obtained by a visual acuity test performed by certified orthoptists using a cross-cylinder.¹⁵ Next, cycloplegia was induced by adding 1% cyclopentolate drops at 10-min intervals (two drops in total, at 0 and 10 min); 60 min after the first eye drop, all objective refractive measurements were repeated.

For both noncycloplegic and cycloplegic SVS, one measurement was obtained, and for RT7000 and Retinomax Screeen, the average of five consecutive measurements was obtained. All measurements were obtained on the same day.

Analysis

For the obtained refraction, power vector analysis¹⁶ was used to calculate the spherical equivalent (SE), J0 astigmatic components (J0), and J45 astigmatic components (J45).

One-way repeated-measures analysis of variance (ANOVA) and Bonferroni correction for multiple comparisons were used to compare noncycloplegic SE, J0, and J45 obtained by subjective and objective refraction tests using three devices. Similarly, repeated-measures ANOVA was used to compare cycloplegic SE, J0, and J45 obtained by objective refraction tests using three devices and noncycloplegic SE, J0, and J45. In this study, we considered the subjective refraction as the subject’s true value and examined the difference between the subjective SE and objective SE using Bland–Altman analysis. We calculated the systematic error (fixed and proportional errors) regarding the difference between the subjective and objective SEs. To examine fixed errors, one-sample t-test was performed for differences in SE. To examine the proportionality error, a significance test of the regression coefficients was performed. Two-sided tests were used for statistical significance, and the significance level was set at a p-value of <0.05. Statistical analysis was performed using EZR (Easy R) version 1.41, which is an open-source software based on R and R commands.¹⁷

Ethics Approval and Consent to Participate

This study was conducted under the approval of the Niigata University of Health and Welfare committee (19153-230905). The experiment was conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all subjects.

Results

In the noncycloplegic condition, the subjective SE and objective SE measured by SVS showed significantly less myopic value than the objective SE measured by RT7000 and Retinomax Screeen (p < 0.001) (Figure 1A). The J0 results in the noncycloplegic condition showed no significant difference (p = 0.182) between the measurement methods (Figure 1B). For J45, there was a significant difference between the results of the SVS and Retinomax Screeen measured in the noncycloplegic condition (p = 0.015), but the difference was 0.1 D (Figure 1C).

Figure 1 Comparison of objective refractions and subjective refraction measured under noncycloplegic condition. (A) Mean spherical equivalent (± standard deviation) measured by subjective method, SVS, RT7000, and Retinomax Screeen were −2.56 (3.00), −2.62 (2.38), −3.05 (2.84), and −3.26 (2.97), respectively. (B) J0 astigmatic component measured by subjective method, SVS, RT7000, and Retinomax Screeen were 0.08 (0.34), 0.03 (0.30), 0.12 (0.42), and 0.08 (0.42), respectively (C) J45 astigmatic component measured by subjective method, SVS, RT7000, and Retinomax Screeen were −0.02 (0.14), 0.03 (0.20), −0.04 (0.15), and −0.07 (0.18), respectively. *p < 0.05, **p < 0.01, ***p < 0.001.

In the cycloplegic condition, the objective SE measured by SVS showed significantly less myopic value than SEs measured by other methods (p < 0.001) (Figure 2A). The J0 results in the cycloplegic condition showed no significant difference (p = 0.122) between the measurement methods (Figure 2B). The J45 results in the cycloplegic condition showed no significant difference (p = 0.167) between the measurement methods (Figure 2C).

Figure 2 Comparison of objective refractions measured under cycloplegic condition and subjective refraction measured under noncycloplegic condition. Black symbols indicate results of non-cycloplegic condition and white symbols indicate results of cycloplegic condition. (A) Mean spherical equivalent (± standard deviation) measured by subjective method, SVS, RT7000, and Retinomax Screeen were −2.56 (3.00), −2.07 (2.66), −2.62 (2.98), and −2.66 (3.02), respectively. (B) J0 astigmatic component measured by subjective method, SVS, RT7000, and Retinomax Screeen were 0.08 (0.34), 0.05 (0.30), 0.13 (0.40), and 0.13 (0.41), respectively (C) J45 astigmatic component measured by subjective method, SVS, RT7000, and Retinomax Screeen were −0.02 (0.14), −0.02 (0.18), −0.04 (0.18), and −0.07 (0.19), respectively. ***p < 0.001.

The SE differences between the subjective refraction and the three noncycloplegic objective refractions are shown in Bland–Altman plots (Figure 3). The noncycloplegic SEs measured by RT7000 and Retinomax Screeen had fixed errors that showed more myopic value, and SVS and RT7000 had proportional errors (Table 1). Similarly, the Bland–Altman plots showed the SE difference between the three objective refractions measured with the cycloplegic and subjective refraction (Figure 4). In the cycloplegic condition, SVS showed a fixed error wherein the measured value was more hyperopic and proportional error (Table 1).

Figure 3 Difference in spherical equivalent between subjective refraction and three types of objective refraction under noncycloplegic condition visualized in Bland–Altman plots. (A) Bland–Altman plots of subjective refraction and objective refraction measured by Spot Vision Screener; Limits of agreement (LOA) was 0.05 ± 1.95. (B) Bland–Altman plots of subjective refraction and objective refraction measured by RT7000; LOA was 0.49 ± 0.97. (C) Bland–Altman plots of subjective refraction and objective refraction measured by Retinomax Screeen; LOA was 0.69 ± 1.39.

Figure 4 Difference in spherical equivalent between subjective refraction and three types of objective refraction under cycloplegic condition visualized in Bland–Altman plots. (A) Bland–Altman plots of subjective refraction and objective refraction measured by Spot Vision Screener; Limits of agreement (LOA) was −0.49 ± 1.25. (B) Bland–Altman plots of subjective refraction and objective refraction measured by RT7000; LOA was 0.06 ± 0.83. (C) Bland–Altman plots of subjective refraction and objective refraction measured by Retinomax Screeen; LOA was 0.10 ± 0.64.

Table 1 Fixed and Proportional Errors for Objective Refractions When Compared to Subjective Refraction

Discussion

Retinoscopy is the most accurate objective refraction test when performed by an experienced clinician.¹⁸ However, because retinoscopy’s precision relies on the skills of the clinician, various refractive tests are performed under cycloplegic conditions. With cycloplegia during measurements, an autorefractometer provides accurate refraction, which is unknown for SVS under cycloplegic condition. In this study, we used the SVS before and after cycloplegic conditions and found that the refraction measured by cycloplegic SVS was different from those using other methods.

In noncycloplegic measurements, the SE measured by SVS was closer to subjective test than to other objective tests. Jorge et al¹⁸ reported that noncycloplegic SE measured by table-mounted autorefractometer, retinoscopy, and subjective refraction were −0.74, −0.32, and −0.29 D, respectively. Teberik et al¹⁹ revealed that noncycloplegic SE measured by Retinomax Screeen is more myopic than those measured by other photorefractometers. From these facts, it can be said that the refraction measured by the autorefractometer is myopic. Conversely, it has been reported that the refraction measured by SVS is more hyperopic than those measured by other autorefractometers.²⁰ The same trend was observed in this study. However, there was a significant proportional error in the SE measured by SVS under noncycloplegic condition, indicating the higher the myopia, the lower the SVS value. SVS is a specialized screening device. The measurement accuracy of the SVS published by the manufacturer is ±0.50 D for −3.50 D to +3.50 D, and ±1.00 D for the range beyond this.²¹ Thus, the error is unavoidable due to the characteristics of the device wherein the measurement accuracy becomes unstable at strong refractive values.

Under cycloplegic condition, there was no significant difference between subjective refraction and objective refraction measured by RT-7000 and Retinomax Screeen, nor was there any proportional error or fixed error. Payerols et al²² reported mean (standard deviation) SEs of +1.06 (2.29) and +1.06 (1.57) for cycloplegic table-mounted autorefractometer and Retinomax Screeen, respectively, in children. The same was true in this study, with cycloplegic SEs measured by RT7000 and Retinomax Screeen having high measurement accuracy. Conversely, cycloplegic SE measured by SVS was significantly hyperopic compared to that by other measurement methods, and SVS had fixed and proportional errors. Yagasaki et al¹⁴ compared the SVS and autorefractometer under cycloplegic condition in subjects with hyperopia and reported that there was a proportional error, with higher hyperopia resulting in lower refractive value measured by SVS. In our study, which focused primarily on subjects with myopia, there was a proportional error in which the higher the degree of myopia, the lower the refractive values measured by SVS. The SVS is an off-axis²³ photorefraction device. Therefore, it is assumed that the refraction is calculated based on a built-in fixed value of pupil diameter. Since SVS is a device specialized for screening, it can be expected that the refraction is calculated based on the pupil under noncycloplegic condition (ie, under natural pupil). Therefore, for SVS, the accuracy of the measurement may be worse with cycloplegic condition.

A limitation of this study is that the average age of the subjects was 19.2 years. Generally, the SVS is used for children, but this study aimed to examine the measurement accuracy in patients with myopia, so it was necessary to increase the age of the subjects. In addition, because the experiment was conducted at a single institution, there is a possibility that bias may have been introduced.

Conclusion

The SVS is characterized by its ability to measure both eyes simultaneously in a 1-s fixation period and by its extremely high measurement possibility. Therefore, SVS is beneficial in screening and in children, for whom cooperation from the patient is difficult to obtain. However, cycloplegic SVS measurement showed poor measurement accuracy of refraction. In addition, the measurement under cycloplegic condition has fixed and proportional errors. Therefore, in the measurement under cycloplegic conditions, it was considered preferable to use an autorefractometer rather than a photorefractometer such as SVS.