Comparison of Dynamic Retinoscopy and Autorefraction for Measurement of Accommodative Amplitude

Abstract

Significance.

This study promotes the use of dynamic retinoscopy to obtain objective measures of accommodative amplitude (AA) in the clinical setting in lieu of the subjective push-up technique.

Purpose.

This study compared the agreement between open-field autorefraction and a modified dynamic retinoscopy for the objective measurement of AA.

Methods.

AA was measured using two objective techniques for subjects age 5-60 years. Test order was randomized and monocular AA measured as subjects viewed printed letters 0.9mm in height with their dominant eye and distance refraction. For retinoscopy, subjects held a near rod and viewed the target at the nearest (most proximal) point of clear vision. The examiner then performed dynamic retinoscopy along the horizontal meridian and identified the physical location of neutrality of the reflex, which was converted to AA in Diopters (D). Autorefraction was performed obtaining repeated measures of refraction beginning from a target demand of 2.5D and increasing in discrete steps until there was no subsequent increase in accommodative response. Refractions were converted to power in the horizontal meridian and expressed as accommodation in D with the maximum value termed the AA. Distance over-refractions were measured for both techniques to adjust AA for any uncorrected refractive error. Difference versus mean analysis was used to compare agreement between tests.

Results.

The 95% limits of agreement (LOA) between techniques were calculated after removal of two young outliers who responded poorly to one of the techniques. The overall mean difference for 95 subjects was 0.02D ± 0.97D with LOA spanning −1.87D to 1.92D. No significant linear relationship between the magnitude of the AA and the differences between techniques was observed.

Conclusions.

Agreement between dynamic retinoscopy and open-field autorefraction was less than 2D with no systematic bias, suggesting dynamic retinoscopy may be a suitable clinical technique to measure objective AA.

Accommodation is a change in shape of the crystalline lens, which increases the plus power of the lens for near viewing. Total accommodative amplitude, often referred to simply as accommodative amplitude, is the maximum focusing power of the corrected eye measured in diopters. Accommodative amplitude is a standard clinical measurement in routine eye care and is well known to decrease with increasing age until it effectively reduces to zero in the sixth or seventh decade of life.^1,2

The most common clinical method to measure accommodative amplitude is the subjective push-up technique. This technique involves moving a near target toward the distance corrected patient until the patient first reports a sustained blur, at which time the viewing distance expressed in diopters represents the accommodative amplitude. Hofstetter’s published equations estimate a linear decrease in amplitude with age and can be used to interpret clinical findings and determine whether the measured amplitude meets age expectations.³

More recently, however, accommodative amplitude studies utilizing objective open-field autorefraction or optometers have demonstrated that measurements obtained with the subjective push-up test overestimate accommodative amplitude.^4-7 Subjective techniques do not directly measure the refractive power of the eye. Instead these techniques quantify the target position as the accommodative amplitude, a strategy that includes the depth of field (the range in space over which a target will remain clear without a change in accommodation), thus overestimating the true refractive response of the eye. This overestimation of accommodative amplitude with subjective techniques has been demonstrated to be greatest in young children.⁵ Overestimation of amplitude in children may not be exclusively related to the inclusion of the depth of field in the measurement, but also due to children having difficulty comprehending the first blur endpoint.⁸ In one study, subjective push-up was shown to overestimate accommodative amplitude by a factor of two in preschool aged children as compared to objectively measured accommodative amplitude.⁵ Clinically interpreting a subjective push-up measure as true accommodative ability could misinform the clinician’s analysis of how much uncorrected hyperopia a given patient can naturally compensate, and thus adopting a technique that provides true accommodative measures is desirable for clinical care. While open-field autorefraction is utilized successfully in the research setting to provide objective measures of accommodative amplitude, the technique is not likely to be adopted in the clinic for measurement of accommodative amplitude due to the specialized instrumentation required, time to obtain measures (i.e. testing numerous demands to derive a stimulus response function), and mathematical conversions required (i.e. averaging of multiple measures at each demand, converting refraction to accommodation, and applying spectacle lens effectivity formulae when applicable). In addition, efforts to develop formulae to convert subjective push-up measures into objective amplitude estimates lacked precision, and thus a more efficient test with less costly equipment is needed.⁵

A previously published dynamic retinoscopy technique for the measurement of accommodative amplitude has recently been evaluated and shows promise as an alternative to the subjective push-up test in the clinical setting.^{9, 10} The technique first requires the patient to place a near target at the most proximal point of clear vision (subjective pull-away accommodative amplitude test), after which an examiner performs dynamic retinoscopy to determine the accommodative response of the eye to the target. While this technique does include a subjective component to determine target position, it is still described in the literature as an objective test since the refractive power of the eye is measured to determine the accommodative amplitude. Accommodative amplitude measures using dynamic retinoscopy were found to have better within session and inter-session repeatability than subjective testing, and also produced lower estimates of accommodative amplitude since the method does not include the depth of field.⁹ The age-related changes in accommodative amplitude as measured by dynamic retinoscopy also agreed well with previously reported measurements obtained with open-field autorefraction in subjects aged 5 – 60 years, giving further evidence that the dynamic retinoscopy technique may be a suitable objective measure of accommodative amplitude in the clinical setting.¹⁰

While the age-related trends of accommodative amplitude matched well between the dynamic retinoscopy technique and previously published values for open-field autorefraction, the agreement between the two techniques was not evaluated on the same individuals. The purpose of this study is to compare agreement between open-field autorefraction and the modified dynamic retinoscopy technique for objective measurement of accommodative amplitude measured on the same individuals in a sample with a large age range. If good agreement is observed, it will lend further support to adopting dynamic retinoscopy as the standard for assessment of accommodative amplitude in the clinic in lieu of techniques that do not measure the refractive power of the eye (e.g. subjective push-up).

METHODS

This study adhered to the tenets of the Declaration of Helsinki and was approved by the University of Houston Committee for the Protection of Human Subjects. Informed consent was obtained from all adult participants, and parental permission and child assent obtained for all participants less than 18 years of age.

One hundred subjects, aged 5 – 60 years old, were recruited through the University of Houston’s College of Optometry staff, faculty and students, as well as their friends and families. Subject recruitment, not to exceed 100 participants, was targeted by age with the intent to recruit a minimum of 8 subjects into five year age bins (10 bins spanning 5 to 54 years) and an 11^th bin ranging from 55 to 60 years. While it would be unusual to measure accommodative amplitude on older presbyopic patients in a clinical exam, the intent of including these age bins in the study was to use these subjects as a surrogate for a young patient with significant accommodative deficiency to determine whether the agreement between methods is negatively impacted by minimal or absent accommodative amplitude. Subjects were screened with correction (or unaided if they did not wear habitual correction) and excluded from further participation if they had presenting near visual acuity worse than 20/50, strabismus, or residual/uncorrected astigmatism greater than 1.25 DC, as described below. Subjects with acuity meeting the inclusion criteria, but later found to have significant uncorrected refractive error (+1.50D or greater hyperopia; −1.00D or greater myopia) were excluded from analysis. Medical history questionnaires were administered to all participants to document past medical/ocular history conditions, or medication use that could impact accommodation.

Preliminary Testing

Preliminary testing was performed on both eyes of study participants. Monocular distance acuity was measured for each eye with a Snellen acuity chart presented on an M&S SmartSystem® (M&S Technologies, Inc. Niles, IL) in the dark. Monocular near acuity was measured for each eye using a near card held at 40cm from the subject’s eye with full room illumination and a near lamp directed at the card. Following acuity testing, eye dominance was determined via the hole-in-the-card method. Cover test was performed at distance and near, and distance autorefraction performed over subjects’ habitual correction (or unaided if they presented with no correction) with the WAM-5500 Grand Seiko Autorefractor (RyuSyo Industrial Co., Ltd. Hiroshima, Japan). Subjects with strabismus, near acuity worse than 20/50, or greater than 1.25 DC of uncorrected astigmatism in their dominant eye were excluded from further participation.

After preliminary data were acquired and subject eligibility confirmed, accommodative amplitude was measured by two different techniques: one using dynamic retinoscopy and the other using open-field autorefraction. Measurement of accommodative amplitude was performed on the dominant eye previously determined via the hole-in-the-card method and test order was determined via coin toss. Subjects wore their habitual distance correction for both measurement techniques. For subjects wearing progressive addition lenses or a near add, their distance spectacle prescription was determined by automatic lensometry and placed into a trial frame with loose lenses for study measures. For subjects wearing multifocal or monovision contact lenses, single vision distance powered trial contact lenses were provided for study measures.

Dynamic Retinoscopy

For the dynamic retinoscopy technique, distance loose lens over-retinoscopy was first performed in full room illumination along the horizontal meridian of each eye to quantify residual spherical refractive error in the horizontal meridian. Subjects then viewed printed letters on a near target mounted on a custom slide attached to a 50cm near rod with a forehead rest. The near target was printed on white cardstock paper (2.5” square) with a rectangular hole (3/4” × 7/8”) cut through the center to enable the examiner to perform retinoscopy along the visual axis. The target consisted of 6 letters (A H T V Z O) printed in black and measuring 0.9mm in height. In addition to full room illumination, a reading lamp was directed at the target, providing an approximate luminance of 143 cd/m².

The subjects’ non-dominant eye was occluded with a patch and subjects were instructed to hold the near rod with the arm corresponding to their non-dominant eye and move the target out from the most proximal position until the letters appeared “first clear” to their viewing eye. At this point, subjects were instructed to keep the letters clear while the examiner performed retinoscopy along the horizontal meridian.

To perform retinoscopy, the examiner first assessed the retinoscopic reflex at twice the distance of the near target. At this distance, the retinoscopic reflex motion was opposite the retinoscopic beam. The examiner then moved closer to the target, continuously assessing the retinoscopic reflex until neutrality was observed. To ‘bracket’ the neutral point, the examiner moved further toward the target, until the retinoscopic reflex was observed to move in the same direction as the retinoscopic beam, and then moved more distal from the target until neutrality was once again observed. The distance at which neutrality occurred was determined by measuring the length of a string extended from the retinoscope to the near target and summing the distance with the distance of the target to the eye. Both the position that the subject placed the target and the position of neutrality were recorded. This procedure was performed three times per subject.

The position at which neutrality was observed was taken to be the near point of accommodation. The average position of neutrality (obtained from three measures) was converted to diopters and added to the distance retinoscopy finding to account for any residual refractive error in the measurement of total accommodative amplitude. For example, if the average position of neutrality was 20 cm and the distance retinoscopy finding was +0.50 D, the total accommodative amplitude would be +5.50 D.

For older subjects who could not read the letters clearly at any position along the 50cm near rod, a plus lens was used to bring their focal point to a position on the near rod for the examiner to perform the retinoscopy technique. All of these subjects were first presented with a +1.00D lens and, if needed, subsequent increasing powers in +0.50D increments until they were able to achieve clarity of the target at some point along the near rod. In calculating the accommodative amplitude for these subjects, the power of the added lens was subtracted from the mean accommodative amplitude finding.

Autorefraction

The non-dominant eye of each subject was occluded with an eyepatch, while the dominant eye viewed a 0.9mm letter “E” first positioned at 40cm (2.5D accommodative demand). Five repeated measures of sphere, cylinder, and axis were taken in rapid succession on the viewing eye as the subject was instructed to try to keep the letter clear. The target was subsequently brought in toward the subject at distances equal to the following dioptric demands: 3, 4, 5, 6, 7, 7.9, 10.5, 12, 15, 20, 25, and 30D. The large range of demands used for this technique was to ensure that the maximum accommodative amplitude would be identified for all subjects, particularly those of young age. While it was not anticipated that subjects would have accommodative amplitudes nearing a 30D demand, due to the depth of focus of the eye, the demand at which a subject elicits their maximum accommodative response may be quite high, and thus large demands were included given that they have previously been shown to be successful for eliciting maximum accommodative responses across the age range tested.⁵ For demand positions from 2.5 – 7.9D, the target was mounted on the examiner side of the beamsplitter along a near rod. For demand positions from 10.5 – 30D, the target was mounted on the subject’s side of the beamsplitter using custom-built attachments described in a previous publication.⁵ For all stimulus positions, the target was a 0.9mm “E”, however, for lower demands on the examiner side of the beamsplitter, the letter was printed in the center of a 2.5cm square of white paper, whereas when it was placed on the subject’s side of the beamsplitter, it was cut out with minimal border and glued along clear thread suspended in a custom-built open frame to allow simultaneous viewing of the target and measurement of the eye. The target was viewed under full room illumination with an additional bendable book light mounted to the forehead rest alongside the non-dominant eye to maintain adequate illumination for target positions underneath the housing of the autorefractor. The target for the autorefraction technique had an approximate luminance of 178 cd/m² when placed on the doctor’s side of the autorefractor and 142 cd/m² when placed under the hood on the patient’s side of the autorefractor.

As subjects viewed the stimulus at increasing proximity, five repeated measures of refraction (sphere, cylinder, and axis) were captured, printed, and later converted to power in the 180 degree meridian (Equation 1)¹¹ to match the meridian measured with the retinoscopy technique, and then averaged for each dioptric demand tested. Measures were obtained rapidly for each demand, within a matter of seconds, and were visible in real time on the autorefractor screen. If the examiner obtained a measurement with 2 diopters of cylinder or greater, it was considered to be erroneous, as subjects in this study did not have large amounts of uncorrected astigmatism. If this occurred, additional measurements were taken until 5 measurements with less than 2 diopters of cylinder were obtained. No other real-time criteria were applied in assessing the quality of the measurements; however, in reviewing the dataset, the five measurements for a given demand spanned a range of no greater than 1 diopter spherical equivalent an estimated 99% of the time, with few exceptions.

Demands of increasing proximity were tested until subjects failed to demonstrate an increase in accommodation for at least two sequential demands (as determined by refraction measures remaining the same, or becoming more plus), or until the 30D demand was reached. During testing, subjects were also asked to report whether or not each stimulus position appeared clear. The first demand for which the subject reported sustained blur was noted.

$$P_m=\mathrm{S}+\mathrm{C}\ast {\mathit{sin}}^2(\phi )$$ (eq.1)

Equation 1 is the alculation of power in the 180 degree meridian where S represents the sphere component of the subject’s average autorefraction measurement, C represents the cylinder component, and φ (phi) is the distance of the axis from the 180 meridian in degrees.

For the determination of accommodative amplitude, an accommodative stimulus response function was plotted for each subject. To achieve this, both refraction measures and accommodative demands had to be adjusted for lens effectivity for any subjects tested while wearing loose trial lenses or spectacles. The effectivity formulae have been described previously and were calculated using a 13mm vertex distance as an estimate of the vertex distance of spectacles worn.¹² While the actual vertex distance of subject spectacles was not determined, 13mm is a reasonable estimate and any minor deviations from this distance would not be expected to impact the results given that the largest magnitude lens power worn by any subject with spectacles was −5.50D. Effective accommodative response versus demand was then plotted for each subject and the maximum accommodative response identified. This value was then added to the mean power in the 180 degree meridian distance over-refraction measurement previously obtained with the Grand Seiko autorefractor to adjust accommodative amplitudes for any residual refractive error.

Refractive Error Classification

To summarize subject characteristics, refractive error classification was based on lensometry of subjects’ habitual distance spectacles, self-report of contact lens powers, or unaided distance autorefraction with the WAM-5500 Grand Seiko Autorefractor (RyuSyo Industrial Co., Ltd. Hiroshima, Japan). Myopia was defined as greater than −0.50D in the most plus meridian, hyperopia as greater than +1.00D in the most plus meridian, emmetropia as powers between +1.00D and −0.50D in both meridians (with ≤ −0.75DC), and mixed astigmatism as plus in one meridian and minus in the other.

Data Analysis

Agreement between dynamic retinoscopy and autorefraction techniques for the measurement of accommodative amplitude was assessed using Bland-Altman analysis and the calculation of the 95% limits of agreement (±1.96 * standard deviation of the differences).¹³ The stimulus position (i.e. accommodative demand) used to elicit the accommodative amplitude for each technique was compared with a descriptive analysis plotting difference in demand versus difference in measured amplitude to identify trends in differences as a function of demand tested. Only subjects aged 5 to 35 years were included in the analysis comparing demands between the two methods, thereby eliminating subjects whose near point of clear vision for dynamic retinoscopy was beyond the extent of the near rod (i.e. requiring the addition of plus lenses during testing). To evaluate the repeatability of dynamic retinoscopy from the three repeated measures per subject, the within subject standard deviation (Sw) was estimated as square root of the residual variance of an Analysis of Variance model (Stata Corp, College Station, TX) and used to calculate the coefficient of repeatability as 1.96*√2* Sw.

RESULTS

One hundred subjects aged 5 to 60 years old were evaluated for the study with 8 to 10 subjects in each of the pre-defined age bins. One adult was excluded due to uncorrected myopic refractive error (−2.00 D) identified on distance autorefraction. A second adult was excluded due to the presence of 2 D of uncorrected hyperopic refractive error identified with distance autorefraction. Lastly, a 19 year old subject was excluded from data analysis since the subject did not reach a definitive peak accommodative response for the autorefraction method even with the largest accommodative demand presented (8.16D accommodative response for the 30D demand). No subjects were identified with medical conditions or medication use (such as stimulants to manage ADHD, or anti-depressants) suspected to interfere with accommodation.

The mean and standard deviation age of the 97 subjects included in analysis was 32 ± 16 years. There were 69 females and 28 males, with a distribution of 61 right eyes tested and 36 left eyes tested. Forty-five subjects were tested unaided, 21 wore contact lenses (all myopic) and 31 wore spectacles (1 astigmat, 8 hyperopes, 22 myopes). All subjects had 20/30 or better distance visual acuity, with the exception of two subjects in their 50s who had distance visual acuity of 20/40. Subjects 45 years and younger all had near visual acuity of 20/30 or better. Near visual acuity through distance correction for subjects 46 to 60 years old ranged from 20/20 to 20/50. For all subjects combined, mean uncorrected cylinder was −0.48 ± 0.31 DC (range = 0 to −1.25 DC) and mean residual spherical equivalent refractive error was 0.17 ± 0.42 D (range = −0.90 to +1.34 D), as determined by distance corrected over-refraction with the Grand Seiko autorefractor.

Accommodative Amplitude by Subject Age

Figure 1 depicts the accommodative amplitude measured by each technique as a function of subject age. Both techniques demonstrate the expected decrease in amplitude with increasing age, with amplitude approaching zero in the sixth decade of life.

Figure 1.

Accommodative amplitude as a function of age as measured by two different techniques.

Agreement between Accommodative Amplitude Measures

Figure 2 depicts the difference versus mean analysis for the accommodative amplitude data acquired. As described in the methods, reported amplitudes are net findings (i.e. distance over-refraction in the horizontal meridian by each technique is summed with each subject’s measured near accommodative response). No significant linear relationship was observed between the magnitude of the measured response and the difference between techniques (p=0.97). Two subjects (open circles) were visually identified as outliers and thus excluded from calculation of the mean difference and 95% limits of agreement. One of these outliers was an 8 year old subject who demonstrated twice the amplitude for the autorefraction technique as for the dynamic retinoscopy technique, and the other was a 12 year old subject demonstrating the opposite pattern. With the outliers excluded from analysis, the mean difference and standard deviation for the group was 0.02D ± 0.97D with 95% limits of agreement ranging from −1.87D to 1.92D.

Figure 2.

Agreement between autorefraction (AR) and dynamic retinoscopy (DR) techniques for all subjects. Gray circles indicate two outliers who were not included in the calculation of the 95% limits of agreement (dashed lines).

Evaluation of Demand Used for Each Technique

Accommodative demand used to elicit accommodation varied between the dynamic retinoscopy and autorefraction techniques. For dynamic retinoscopy, subjects were instructed to place the target at the most proximal position where the target remained clear. For the autorefraction technique, multiple demands were tested, ranging from 2.5 to 30D, until the maximum accommodative response was elicited. This meant that for the autorefraction technique, a demand more proximal or more distal than the near point of clear vision could have been utilized to elicit the maximum accommodative amplitude. In order to compare demands between techniques, the physical distance of the stimulus used to elicit the accommodative amplitude in centimeters was used to avoid the compression that occurs at close working distances when expressing demand in diopters.

Figure 3 is a descriptive analysis that depicts the difference in accommodative amplitude between the two techniques versus the difference in demand used to elicit the accommodative amplitude in centimeters for the two measurement techniques for 53 subjects aged 5 to 35 years. The number of subjects falling into each of four categories was tallied to determine whether there was a trend toward greater or lesser demands used with the AR technique, as well as a trend toward greater or lesser accommodative amplitude obtained with demands that were more proximal or more distal to the near point of clear vision.

Figure 3.

Difference in accommodative amplitude for autorefraction (AR) versus dynamic retinoscopy (DR) as related to the difference in target position in centimeters (cm) used to elicit the accommodative amplitude for each technique for subjects aged 5 to 35 years.

On average, the demands that elicited the maximum accommodative response for the AR technique were 0.7 cm closer to the subject than those used during the DR technique (mean and standard deviation along the x-axis = −0.7 ± 3.6 cm). Regarding accommodative amplitude measures, the AR technique measured higher accommodative amplitude (mean and standard deviation along the y-axis = 0.15 ± 1.5 D) for this subset of subjects aged 5 to 35 years, similar to the differences between techniques observed for the entire group, as shown in Figure 2. Despite these small differences, the distribution of subjects among the four categories identified in Figure 3 was relatively balanced; indicating that use of a more proximal demand did not always result in an increased accommodative amplitude measure, or vice versa.

Variability of Dynamic Retinoscopy

For the dynamic retinoscopy technique, three repeated measurements were performed for each subject and the average accommodative amplitude analyzed. Figure 4 shows the within subject standard deviation for the three repeated measurements of accommodative amplitude by dynamic retinoscopy as a function of the average accommodative amplitude. Visual inspection reveals a relationship between accommodative amplitude magnitude and the overall amount of measurement variability. Approximately 63% of subjects had variability less than 0.50D and 90% had variability less than 1.00D. Both children and adults were included among those with variability exceeding 1.00D. The within-subject variance of the entire sample was 0.515 D, yielding a coefficient of repeatability of 1.99 D.

Figure 4.

Within subject variability of accommodative amplitude as measured by dynamic retinoscopy

DISCUSSION

This study sought to determine the agreement between accommodative amplitude measured by open-field autorefraction versus a dynamic retinoscopy technique. The 95% limits of agreement between the two techniques ranged from −1.87D to 1.92D after removal of two young outliers and the measurement bias between techniques showed a small, clinically insignificant difference of greater amplitudes measured with autorefraction (0.02 D for the entire group versus 0.15 D for subjects aged 5 to 35 years). There was also no relationship between differences and the magnitude of the accommodative amplitude measured indicating that agreement was not impacted by the presence or absence of robust accommodative ability (e.g. the presbyopic subjects included to simulate accommodative deficiencies). Given the strong relationship between the magnitude of accommodative amplitude and age, this would also suggest that the bias between measurement techniques is unrelated to subject age. However, it is important to note that two young subjects (8 and 12 years) were identified as outliers due to excessively large differences between measurements of accommodative amplitude with the two techniques (Figure 2). One of these subjects had substantially greater amplitude with the autorefraction technique, while the other performed substantially better with the dynamic retinoscopy technique. These findings suggest that there may be children who respond better to the instructions for one test than another, as with any clinical technique. In this sample, there was not a trend of consistent difficulty with one technique over the other.

One potential source of differences between the techniques is that the dynamic retinoscopy technique utilized measurement of accommodation at a single accommodative demand (the near point of clear vision), whereas the autorefraction technique measured accommodation at multiple demands, determining the maximum amplitude from a stimulus response function. Thus one may question whether the near point of clear vision (i.e. the subjective amplitude) truly occurs coincident with the maximum refractive accommodative amplitude of the eye, or if individuals still exert additional accommodative effort once a near stimulus becomes blurred. The analysis depicted in Figure 3 is an attempt to answer this question and indicates that while 15 individual subjects did demonstrate an increased accommodative amplitude once the target was positioned more proximal than the near point of clear vision, an almost equal number of 19 individual subjects demonstrated an increased accommodative amplitude when the target was positioned more distal than the near point of clear vision. The analysis in Figure 3 demonstrates that the near target position at which the maximum accommodative amplitude is obtained differs across subjects; however, without a means to determine the behavior of a single individual a priori, utilizing the near point of clear vision for dynamic retinoscopy seems to provide a good balance of the varied behaviors of the group as a whole. It is possible that testing multiple near target positions with dynamic retinoscopy would improve agreement between the retinoscopy and autorefraction techniques; however, this would come at the cost of clinical time that may not yield a meaningful benefit in accuracy.

Pupil diameter is an additional potential source for differences between autorefraction and dynamic retinoscopy, but was not measured in this study. In the objective measurement of accommodative amplitude; however, pupil diameter likely has less influence on measurements, particularly when using the autorefraction technique. In the autorefraction technique, a stimulus response function is derived, testing numerous demands well beyond the near point of clear vision to obtain a maximum accommodative response. For the dynamic retinoscopy technique performed in this study; however, pupil diameter may have influenced the target demand used given that the subject was instructed to place it at the nearest point of clear vision. Pupil diameter could be one factor that accounts for the finding that maximum accommodative amplitude is found at demands more proximal to the near point of clear vision for some subjects and yet more distal for others (Figure 3).

A previous publication sought to determine whether a predictive equation could be used to convert subjective amplitude findings to objective amplitude for clinical assessment of accommodation. That study found the predicted amplitudes fell within ±1.50 D of the objective amplitude for 80% of subjects.⁵ In the present study, differences between dynamic retinoscopy findings and autorefraction findings fell within the range of ±1.50 for 87% of subjects once the two young outliers were excluded. While this demonstrates some improvement over the previously published predictive equation, ±1.50 D is still a relatively large difference when comparing accommodative amplitudes that do not exceed 10D. However, when considered in the context that the coefficient of repeatability for dynamic retinoscopy in this study was 1.99 D, the limits of agreement between dynamic retinoscopy and autorefraction were comparable, suggesting agreement is not outside that expected from measurement variability. The repeatability of the dynamic retinoscopy technique in the present study was elevated compared to that reported in a previous study (1.99 D vs. 0.80 D);⁹ however, the prior study included subjects between the ages of 18 and 30 years, whereas the present study included children as young as 6 years, likely negatively impacting repeatability due to both subject cooperation and the greater magnitude of accommodative amplitudes observed (Figure 4). Given that objective measurement of accommodative amplitude is an emerging technique that has not yet been adopted clinically, it is difficult to determine a clinically acceptable level of agreement between tests. However, when compared to the subjective push-up test which has been shown to over-estimate true accommodative amplitudes by more than 6D in young children,⁵ agreement between objective tests within 2D appears to be a notable improvement.

The amplitudes measured in this study agree well with the age-related magnitudes observed in other studies utilizing objective measurement techniques.^{5, 6, 10} As previously demonstrated, accommodative amplitudes measured objectively are lower than that assumed by a subjective assessment of near point of clear vision. To continue to shift clinical thinking away from interpretation of a subjective push-up test as the actual refractive change of the eye due to accommodation, objective techniques that can be utilized efficiently in the clinic are needed. Grand-Seiko autorefraction is unlikely to be adopted due to instrumentation and time constraints; however, dynamic retinoscopy performed while patients view a target at the near point of clear vision shows promise as a viable clinical technique. In conducting this study, the time to complete each study measure was not recorded, but could be estimated to be 1 to 3 minutes for dynamic retinoscopy versus 10 to 15 minutes for autorefraction. The time to complete autorefraction is not meant to suggest subjects were continuously fixating and accommodating to the near stimulus throughout that time, but includes the time to reposition the target for multiple stimulus demands and print the accommodative response measures for each demand tested. By this assessment of test time, dynamic retinoscopy is certainly more clinically feasible.

One potential limitation to the generalizability of these results is that the study population included adults (students and faculty) from the University of Houston, College of Optometry who may have had greater attention and cooperation for study measurements than the general patient population due their familiarity with accommodative testing. Despite this limitation, agreement between tests did not vary with accommodative amplitude (Figure 2) which is strongly related to subject age (Figure 1), and thus it is unlikely that inclusion of adults with familiarity of accommodative testing swayed the overall agreement between tests.

A second limitation to the generalizability of these findings is that of examiner experience. The examiner performing retinoscopy in this study was a student who had just completed year one of the four year doctor of optometry program at the University of Houston, College of Optometry. While this investigator had completed training and demonstrated strong proficiency in the performance of retinoscopy prior to the study, it is possible that the repeatability of dynamic retinoscopy findings within subjects, or the agreement of accommodative amplitude measures between tests would be even more robust for a clinician with specific expertise and years of experience in the performance of retinoscopy.

Debate regarding the optimal target size for accommodative studies and whether or not the target should be scaled by viewing distance to subtend a consistent visual angle on the retina is not uncommon. This study utilized a fixed target size of 0.9mm for both dynamic retinoscopy and autorefraction, and thus both techniques would be impacted by changes in visual angle as the target is moved closer during testing. With respect to the autorefraction, varying target size has been shown to have minimal impact on the measured accommodative amplitude due to the methodology of obtaining a stimulus response function, despite whether accommodative responses at specific demands differ with target size.⁵ Given that the dynamic retinoscopy technique does not derive its amplitude measurement from a stimulus response function, it may be more prone to influence by target size; however, the goal of this study was to evaluate a previously published technique potentially applicable for a clinical setting. Thus scaling the target based on testing position for the dynamic retinoscopy technique would not be clinically practical.

Efforts were made to exclude subjects with significant uncorrected refractive error, as identified by distance autorefraction. That said, subjects did not undergo cycloplegia, and thus individuals with latent hyperopia may not have been identified. Even so, subjects were recruited from the University of Houston, College of Optometry staff, faculty, students, and their family and friends and thus the majority, if not all, subjects had received a comprehensive eye examination in the past and are unlikely to have significant amounts of uncorrected latent hyperopia. In addition, this study was designed to evaluate agreement between tests which would likely be impacted similarly by an individual’s uncorrected refractive error.

In conclusion, this study found no systematic bias, or effect of accommodative amplitude magnitude on the differences between measures obtained by dynamic retinoscopy versus autorefraction. While differences reached 2D for some subjects, the majority had differences that did not exceed the variability of dynamic retinoscopy itself and good agreement was observed for both children and adults.