Introduction
Accommodation is a dynamic process of retinal defocus
when the eye focuses at different distances (1). Loss of
accommodation due to presbyopia or cataract surgery is one
of the most common complaints in ophthalmic patients.
Although many basic and clinical researches were conducted
(2,3), the mechanism of accommodation and presbyopia are
not fully understood, and restoring accommodation remains
a challenging task.
The lenticular changes mediated by ciliary muscle
and zonule is commonly regarded as a key factor for
accommodation and presbyopia (4-6). Loss of deformity
of the lens with age is hypothesized to be the primary
etiology of presbyopia (4). Thus a detailed knowledge of
the structural changes of the crystalline lens and optical
properties during accommodation is of great importance
in understanding the mechanism of accommodation
and exploring new techniques for restoring dynamic
accommodation. Additionally, exact data of dynamic
changes amplitude in anterior chamber depth (ACD) and
anterior lens position can help to assess the security of the
implantable Collamer lens (ICL) which may cause cataract
formation due to contact between the ICL and crystalline
lens (7). Previous measurements of crystalline lens in vivo
were mainly performed with A-scan (8), IOLmaster (9),
Purkinje imaging (10), Scheimpflug camera (11), slit-lamp
photography (12), ultrasound biomicroscopy (UBM) (13),
or magnetic resonance imaging (MRI) (14). These
measurements of the crystalline lens, however, were indirect
measures with or without complexity of obtaining cross-sectional images, with limited visibility of the posterior lens
surface, and debatable reproducibility and accuracy (12).
Anterior segment optical coherence tomography (AS-OCT) is a non-invasive method that offers high-resolution
imaging of the anterior segment structures with high-speed,
3D imaging capability and providing coaxial accommodative
targets which enables in vivo and real-time assessment of
accommodation, and thus has gained wide used in clinical
application to assess parameters of the anterior chamber
angle, as well as accommodation due to its advantage of
dynamic measurement (2,15,16). However, challenge exist
to assess the entire crystalline lens in a single frame by
commercial AS-OCT due to the limited scan depth and the
sensitivity decay with depth. Currently, available methods
to image the entire anterior segment with AS-OCT
were obtained by complex phase-splitting techniques for
conjugate removal (17), dual channel dual focus AS-OCT
devices (18) or by overlapping images acquired at different
depths (16), the reliability of which needs to be further
assessed. Furthermore, the complexity of the techniques
limited the possibility to generalize it in larger population-based studies.
In recent years, with the improved scan rate, scan depth,
and scan density of the second generation swept source
AS-OCT (SS-OCT) (CASIA2, TOMEY, Japan), it is now
possible to obtain imaging from the cornea to the posterior
lens surface in one cross-sectional image. The reliability
and repeatability of CASIA2 in measuring anterior segment
structures was high (19). Current studies of accommodation
using CASIA2 are scarce (20-24). Therefore, we aimed to
characterize changes in lens parameters and other anterior
segment parameters including ACD, anterior segment
length (ASL) and its associations using CASIA2 during
-3 diopter (D) accommodative stress. We present the
following article in accordance with the STROBE reporting
checklist
Methods
Participants
One hundred-fifty-nine participants were consecutively
recruited to participant in the current study from Aug 22
to Sep 10, 2019 at Zhongshan Ophthalmic Center (ZOC),
southern China. Exclusion criterion were best corrected
Visual Acuity (BCVA) of less than 20/20; previous history
of ocular disease or surgeries; history of wearing contact
lenses; anterior segment abnormality; poor quality SS-OCT
images; or inability to cooperate during examination. The
study was conducted in accordance with the Declaration of
Helsinki (as revised in 2013). Ethical approval was obtained
from the Zhongshan University Ethics Review Board (No.
2019KYPJ033). Written informed consent was obtained
from all participants.
Ocular examinations
All participants underwent uncorrected visual acuity
(UCVA) and BCVA examination using an Early Treatment
Diabetic Retinopathy Study (ETDRS) visual chart at a
distance of 4 m. Non-cycloplegic refractions were measured
by an autorefractor (KR8800; Topcon, Tokyo, Japan). The
refraction data was further converted to spherical equivalent
(SE) which was defined as spherical power plus half of the cylindrical power. Axial length (AL) was measured by an
optical biometer (IOLmaster700, Carl Zeiss Meditec, Jena,
Germany). A slit lamp examination was adapted to detect
abnormalities of the anterior and posterior segments. Each
examination was performed by the same experienced nurse
by a standard protocol to ensure reliability.
SS-OCT examinations
The SS-OCT device (CASIA2, Tomey Corporation,
Nagoya, Japan) with a laser wavelength of 1,310 nm was
used to obtain images of the anterior segment before and
during accommodation. All measurements were conducted
under a standard procedure in a darkened room with
the same device settings by the same research nurse. SS-OCT scans of the lens analysis model were obtained at the
accommodative resting state and at an accommodative stress
on -3 D after refractive compensation by a build-in system
of the CASIA2. Before scanning, refractive error of the
tested eye was corrected by the built-in system. The anterior
segment images of the tested eye in unaccommodated
state was captured with the participant focusing on the coaxial internal fixation target image set for distance in the
CASIA2 device for the first 5 seconds. Afterward, -3.0 D
accommodative stress by the built-in system of the CASIA2
was added to stimulate the physiologic accommodation, and
images were captured when the participant could clearly
look forward at the internal fixation target for 5 seconds.
The eye was centered by the active eye tracker system in
CASIA2. Corneal topographic axis, defined as a reference
line connecting the fixation point in the topographer to
the vertex normal on the cornea, was used as positional
reference for comparative images obtained from the
unaccommodated and accommodated states (25,26). To
avoid eye movement, rotation, convergence or other
movements during scanning, participants were guided
to keep their jaw and forehead on the fixed trestle, while
staring at the co-axial internal fixation target during
scanning. Then the device automatically analyzed and
displayed values of ACD, anterior and posterior lens
curvature, lens thickness (LT) and lens diameter at
3-dimensional space. In order to ensure the reliability of the
automatic results generated by the CASIA2, the “semi-auto”
setting was used to check the accuracy of the automatic
outline and re-modify the outline in scans with poorly
automatically detected outlines. Three-dimensional data of
the anterior segment parameters were analyzed. ACD was
defined as the distance between the corneal epithelium and
the anterior surface of the lens. ASL was defined as ACD
plus LT. Lens central point (LCP) was defined as ACD
plus half of the LT. The corneal model was used to obtain
measurements of corneal power.
Accommodative response
The accommodative response was defined as changes in
refractive power of the eye during accommodation based on
a schematic eye model in paraxial approximation (Eq. [2]).
Where P, PC and PL refer to the power of the equivalent
total eye, cornea and lens, respectively. Values of PC were
obtained from the CASIA2, nl is the refractive index of the
lens which is adjusted for age (27), nh=1.336 is the refractive
index of the aqueous humor, Ra and Rp represent the radius
of anterior and posterior lens curvatures.
Statistical
analysis
All statistical analyses were performed with Stata (ver. 12.0;
Stata Corp, College Station, TX). All data were presented
for right eyes. The Skewness-Kurtosis test and histogram
were used to test the normality of variables. Continuous
variables were presented as the mean ± standard deviation
(SD) or median [interquartile range (IQR)]. Analysis of
variance (ANOVA) for normal distributions, or the Kruskal
Wallis test for very skewed distributions were used to
compare anterior segment parameters between age groups.
Changes in anterior segment parameters, including anterior
lens curvature, posterior lens curvature, LT, lens diameter,
ACD, LCP and ASL, were calculated by subtracting
anterior segment parameters at the accommodative
resting state from anterior segment parameters at ?3
D accommodative stress. Paired t-test was used to test
for changes in anterior segment parameters during ?3
D accommodative stress. Univariable and multivariable
regression models were used to determine associated
factors with changes in anterior lens curvature, posterior
lens curvature and accommodative response of ?3 D
accommodative stimulus. A P value of <0.05 was defined to
indicate statistical significance.
Results
Of the 159 recruited participants, 19 were excluded due to
having BCVA of less than 20/20 (9 participants), history of
wearing contact lenses (3 participants), anterior segment
abnormality (2 participants) and poor-quality SS-OCT
images (5 participants), leaving 140 eligible participants
for the final analysis (Figure 1). The distribution of basic
characters of the included participants is presented in Table 1.
In general, images of 140 participants with a median (IQR)
age of 30 (23 to 44) years (age range, 10–59 years) and a
male proportion of 40.71% were obtained under both the
static state and at an accommodative amplitude of ?3 D.
The median (IQR) SE was ?1.13 (?4.13, 0) D and the mean
AL was 24.28±1.52 mm.
Table 2 presents result of changes in anterior segment
biometry during ?3 D accommodative stimulus by age
groups. Compared to non-accommodative status, ACD
(2.952±0.402 vs. 2.904±0.382 mm, P<0.001), anterior
(10.771±1.801 vs. 10.086±1.571 mm, P<0.001) and posterior
lens curvature (5.894±0.435 vs. 5.767±0.420 mm, P<0.001),
lens diameter (9.829±0.338 vs. 9.695±0.358 mm, P<0.001)
and LCP (4.925±0.274 vs. 4.900±0.259 mm, P=0.010)
tended to decreased and LT thickened (9.829±0.338 vs.
9.695±0.358 mm, P<0.001), while ASL (6.903±0.279 vs.
6.898±0.268 mm, P=0.568) did not change significantly
during accommodation. The changing amplitude of anterior
lens curvature, posterior lens curvature, and LT tended to
be smaller with increasing age (all P<0.05).
Table 3 shows the associated factors with changes
in anterior and posterior lens curvature during ?3 D
accommodative stress. In the univariable model, participants
who were younger (β=0.033, 95% CI: 0.025 to 0.041,
P<0.001), female gender (β=0.411, 95% CI: 0.191 to 0.630,
P<0.001), had myopic SE (β=0.083, 95% CI: 0.045 to
0.120, P<0.001), longer AL (β=?0.180, 95% CI: ?0.258 to
?0.101, P<0.001), deeper ACD (β=?0.778, 95% CI: ?1.031
to ?0.525, P<0.001), larger anterior (β=?0.189, 95% CI:
?0.243 to ?0.134, P<0.001) and posterior lens curvature
(β=?0.399, 95% CI: ?0.650 to ?0.148, P=0.002), thinner
LT (β=0.847, 95% CI: 0.619 to 1.074, P<0.001) and more
backward of LCP (β=?0.655, 95% CI: ?1.053 to ?0.258,
P=0.001) were more likely to have larger steeping of the
anterior lens curvature after ?3 D accommodative stimulus.
Younger age (β=0.029, 95% CI: 0.020 to 0.038, P<0.001)
and larger anterior (β=?0.071, 95% CI: ?0.138 to ?0.003,
P=0.040) lens curvature were still associated with larger
accommodation-induced steeping of the anterior lens
curvature after adjusting for age, gender and SE. Younger
age (β=0.004, 95% CI: 0.001 to 0.007, P=0.014) and larger
posterior (β=?0.102, 95% CI: ?0.184 to ?0.021, P=0.014)
lens curvature were associated with larger accommodationinduced steepening of posterior lens curvature after
adjusting for confounders. There was a positive correlation
between changes in anterior lens curvature and changes
in posterior lens curvature (r=0.518, P<0.001) and lens
diameter (r=0.743, P<0.001), ACD (r=0.847, P<0.001), and
LCP (r=0.601, P<0.001), while changes in anterior lens
curvature was negatively correlated with changes in LT
(r=?0.881, P<0.001) and ASL (r=?0.387, P<0.001) (Figure 2).
A similar correlation was also observed for changes in
posterior lens curvature (Figure 3).
Changes in optical power of the whole eye and the
isolated lens are calculated using Eqs. [1,2]. The optical
power of the whole eye at 0 D and ?3 D accommodative
stress was 62.49±2.28 and 63.27±2.29 D, respectively
(P<0.001). Accordingly, the lens power was 24.03±1.89 D
at 0 D and which increased to 24.90±2.05 D during ?3 D of
accommodative stress (P<0.05). Younger age was associated
with greater accommodative response (β=?0.027, 95% CI:
?0.038 to ?0.016, P<0.001) in the multivariable regression
model (Table 4).
Discussion
The current study found that the anterior segment
biometry including ACD, anterior lens curvature, posterior
lens curvature, lens diameter, LCP and LT changed
significantly during ?3 D accommodative stress, while ASL
did not change during accommodation. Participants with
older age and smaller anterior lens curvature tended to have
smaller decreasing amplitude of the anterior lens curvature
during accommodation. The accommodative response
of ?3 D stimulus decreased with age. Previous studies on
accommodation have been based on small study samples or
based on traditional or compound imaging which cannot
directly observe changes of the entire anterior segment, thus
decreased reliability (13,14,16,18). To further understand
the mechanism of accommodation and on-set of presbyopia
and its related factors, studies of larger sample sizes with
scans in which images are entirely visible are needed. To the
best of our knowledge, this is so far the study containing
the largest study sample by the CASIA2, the scanning depth
of which enables us to obtain a single anterior segment
image from anterior cornea to posterior lens surface and
automatically generate data of anterior segment parameters
with build-in software.
Our findings that lens curvature steepened followed
by LT thickened and ACD shallowed are in accordance
with the widely accepted Schachar mechanism (5,28-30).
Accommodation-related decrease of anterior lens curvature,
increase of LT and shallow of ACD have been well verified
by numerous studies using Scheimpflug, AS-OCT, and
other methods (2,14). However, lack of available methods
for imaging the posterior lens curvature, most of which
are currently obtained indirectly or through complex
machine or overlapping images, limited our knowledge
of accommodation-related changes in the posterior lens
curvature. Some studies found a statistically significant
reduction in posterior lens surface curvatures during
accommodation using MRI (14) or SD-OCT (31), but other
studies have indicated no changes (32,33). The possible
reason for the controversy may be due to the low resolution
of images and low reliability of the synthesized images.
The enhanced penetration and speed of the CASIA2
enables the visibility of the entire anterior segment in one
cross-sectional image in real-time and automatic analysis,
which is a useful method in evaluating dynamic changes
in lens parameters during accommodation. Our study
indicated that posterior lens curvature steepened during
?3 D accommodation. Current studies on accommodation
using CASIA2 are scarce. One study of 30 participants aged
20 years and older from Japan presented preliminary
values of anterior lens curvature, posterior lens curvature
and LT before and after ?3 D accommodative stress.
However, it is difficult to determine whether changes in
lens parameters during accommodation happened because
this paper did not make statistical test for these changes (23).
The same researchers did a more detailed analysis of
accommodation-related changes in lens parameters in
another study, with the results indicating that anterior
and posterior lens curvatures steepened and LT thickened
during accommodation in 96 included eyes, but the exact
values of lens biometry were not presented (20). Thus, we
cannot make direct comparison between our study and the
aforementioned studies (20,23).
Studies on changes of the position of the posterior lens
pole (ASL) during accommodation is controversial, with
some studies having indicated no significant changes (16,34),
but others found a significant backward-displacement of
the posterior lens position (3,31). Variation of different
measurements, or imaging length of technique may partly explain these disparities. Our finding that ASL did not change
with accommodation supports the Schachar theory that the
lens position remains stable during accommodation (5,28).
Our and previous studies (14,35,36) indicate that
the accommodation induces changes in anterior lens
curvature is larger than the posterior lens curvature
during ?3 D accommodative stress. Additionally, the
accommodation-induced changes in the amplitude and
slope of the anterior lens curvature are different from
the posterior lens curvature. One 3D MRI study showed
that changes in anterior lens curvature were larger than
those of posterior lens curvature, and changes in slope of
the anterior and posterior lens curvature were not linear
during accommodation (14). Gambra et al. (35) found that
there was a decrease of 0.73 mm/D for the anterior lens
curvature and 0.20 mm/D for the posterior lens curvature
with increasing accommodative demand from 0 to 6 D by
1-D step. Results from an AS-OCT study indicated that
changes in lens biometry were limited to the anterior lens
curvature if accommodative stress was less than 1.5 D,
while the posterior lens curvature was additionally altered
in the process of accommodation only above the 1.5 D
accommodative stress level (36). The reason for the
differences in response of anterior and posterior lens
curvatures to accommodation may be due to differences in
insertion position and direction of the zonulas fiber or the
relative position of the ciliary muscle to the lens (3).
The human lens, which provides one-third of the
ocular optical power, is responsible for refocusing on
targets at different distances through alteration of its
shape to decrease or increase ocular optical power
during accommodation. Previous studies showed that the
deformability of the crystalline lens decreased with age,
but the contractility of the muscle did not decline with age
(2,27). One study using a combination of two SD-OCT
systems to acquire synchronized images of the ciliary body
and lens showed that accommodation-induced changes in
LT was larger in younger subject, and younger subject had
faster responses in the crystalline lens after contraction
of the ciliary muscle (37). Richdale et al. (38) found that
accommodative response increased linearly with increasing
accommodation demand among subjects younger than
40 years but increased little among subjects aged 40 years
and older; the lost rate of maximum accommodative response
was 0.2 D/age. The current study found that the amplitude
of reshaping of anterior and posterior lens curvatures during
?3 D accommodative stimulus was significantly decreased
with age, as well as the accommodative response which
represents the real accommodative power under ?3 D
accommodative stimulus (0.027 D per year).
The mean value of ?3 D accommodation-induced
changes in ocular refractive power of the whole eye in our
study was 0.79 D (range from ?0.10 to 3.04 D), which is
smaller than found in other studies. One OCT study of
4 subjects indicated that the changes of refractive power
was about 3.8 D during 6 D accommodative demands (35).
Another OCT study of 19 eyes from 13 subjects with a
mean age of 28.6±4.4 years showed that the changes of
refractive power was 4.14 D during 0–6 D accommodative
demands in 1.5 D steps (31). Richdale et al. (38) found that
the response amplitudes of 6 D accommodation were 0.03
to 5.15 D among 26 emmetropic adult aged 30 to 50 years
using an auto-refractor. The relative wide age range in the
current study may partly explain the difference. However,
differences in measurement technique, study design, and
race prohibited a direct comparison of our study with
previous studies (31,35,38).
Strength of our study included the relatively large
study sample size, standardize study protocol, and the
new SS-OCT device which enable visibility of the entire
lens figure and automatic analysis. Several limitations of
the current study should be mentioned. First, analysis
of some accommodation-related mechanisms including
changes to ciliary muscle and zonular fibers were not
available in the current study. Second, understanding
accommodation-induced changes in lens surface area and
lens volume, followed by refractive index (39-42) and
crystalline lens density may help to provide a more indepth understanding the mechanism of accommodation.
However, currently available SS-OCT device cannot
measure lens surface areas and lens volumes. Third,
characteristics of lens biometry changes in other
accommodative states than presently examined (>?3 D
or maximum accommodative response) should be further
assessed. Fourth, the fact that accommodative response was
estimated by equation instead of auto-refractor which is
an objective measurement of the accommodative response
may have skewed our data. Compare to the equation of
optical power, an indirect measurement, evaluating the
accommodative response directly using an auto-refractor or
other objective measurements may increase the reliability
of the measurements. Fifth, the dynamic behavior of the
lens during accommodation may potentially influence
the accuracy of the measurements which were obtained
only once at each status. Finally, it is difficulty to evaluate
changes in equatorial diameter because the edge of the lens
cannot be visualized due to the presence of the iris by using
AS-OCT (15,43). The lens diameter in the current study
was obtained by fitting the anterior and posterior circular
curves which may not exactly represent the real shape of the
lens during accommodation
In conclusion, the current study found that the lens
is elastically deformed during accommodation which
include steepening anterior and posterior lens curvature,
decreased lens diameter and increased LT, followed by
decreasing ACD and forward-displacement of LCP in
almost all age subgroups. We found no significant changes
in ASL in accommodation. Our findings support Schachar’s
accommodative hypotheses that the lens is stable during
accommodation. The accommodative response of ?3 D
stimulus is age-dependent. These data may help us
further understand the age-dependent mechanisms of
accommodation.
Acknowledgments
Funding: This study was supported by the National
Natural Science Foundation of China (Nos. 81770905 and
81873675) and the Construction Project of High-Level
Hospitals in Guangdong Province (No. 303020102).
