Beruflich Dokumente
Kultur Dokumente
2
3 Title: Comparison of Foveal, Macular and Peripapillary Intraretinal Thicknesses
4 between Autism Spectrum Disorder and Neurotypical Subjects
5
6 Running head: Thickness of intraretinal layers in autism
7
8 Authors (10):
9 José Javier García-Medina1-4, María García-Piñero5, Mónica del-Río-Vellosillo6,
10 Jesarán Fares-Valdivia1, Ana Belén Ragel-Hernández7, Salvador Martínez-
11 Saura8, María Dolores Cárcel-López7, Vicente Zanon-Moreno3,4,9, María
12 Dolores Pinazo-Duran3,4,10, María Paz Villegas-Pérez1,2,4
13 Institutions:
14 1. General University Hospital Reina Sofia, Murcia, Spain.
15 2. Department of Ophthalmology, Optometry, Otolaryngology and
16 Pathology. School of Medicine, University of Murcia, Spain.
17 3. Ophthalmic Research Unit “Santiago Grisolia”, Valencia, Spain.
18 4. Spanish Net of Ophthalmic Pathology OFTARED, Institute of Health
19 Carlos III, Madrid, Spain.
20 5. Faculty of Biology, University of Murcia, Spain
21 6. University Hospital Virgen de la Arrixaca, Murcia, Spain.
22 7. Integral Formation Center “Gabriel Perez Carcel”, Murcia, Spain.
23 8. Asociación para la atención de personas con trastornos generalizados
24 del desarrollo de la región de Murcia (ASTRADE), Murcia, Spain.
25 9. Department of Preventive Medicine and Public Health, School of
26 Medicine, University of Valencia, Valencia, Spain
27 10. Department of Ophthalmology, School of Medicine, University of
28 Valencia, Spain
29
30 Corresponding author:
31 Jose Javier Garcia-Medina, MD, PhD.
32 Department of Ophthalmology, General University Hospital Reina Sofia,
33 Avenida Intendente Jorge Palacios, 1, 30003 Murcia, Spain
34 Phone: +34 968 35 90 00; Fax: +34 968 35 98 19
35 Email addresses: josegarciam@yahoo.com, jj.garciamedina@um.es
36
37
1
1 Abstract
6 functioning ASD and 27 age- and gender-matched NT subjects) had two SD-
7 OCT scans: macular fast volume and peripapillary retinal nerve fiber layer
8 (pRNFL). Macula was automatically segmented. The mean foveal and macular
10 considered. Data from the right and left eyes were averaged for each
11 participant. The results were compared between the ASD and NT groups.
14 Results: ASD subjects showed greater foveal thickness at total retina, total
15 inner retina, inner plexiform and inner nuclear layers, and greater macular
16 thickness at total retina and total inner retina. Inferior, nasal inferior and
17 temporal inferior sectors of pRNFL were also thicker in the ASD participants
18 than in the controls (p<0.05, unpaired t-test). Significant correlations were found
19 between some K-BIT results and temporal inferior and inferior pRNFL
23 subjects when compared to NT controls. This fact should be taken into account
2
1 when interpreting SD-OCT examinations in ASD individuals. Plus, some pRNFL
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
3
1 INTRODUCTION
7 States has been identified with ASD.4 The etiology of ASD still remains
8 unknown. However, it has been found that gray and white matter in the brains of
14 physiology and histopathology. Similarly to the brain, the retina has the
15 equivalent of white matter (plexiform and nerve fiber layers) and gray matter
19 permitted the identification in vivo of individual retina layers and the estimation
25 best of our knowledge, only one study has studied ASD subjects by SD-OCT.14
4
1 The authors compared the peripapillary retinal nerve fiber layer (pRNFL)
2 between the ASD and neurotypical (NT) groups, and found a reduced pRNFL in
3 the former group. Nevertheless, as far as we know, the thickness of the retinal
4 layers in the macula has not previously been investigated in ASD. Therefore,
7 thickness of the retinal layers with the cognitive abilities and head
9 METHODS
10 Recruitment
14 Murcia, Spain. All the examinations were performed between September 2015
15 and March 2016 in the General University Hospital Reina Sofia, Murcia, Spain.
16 The study included two groups of patients: ASD and NT subjects. Inclusion
17 criteria for ASD subjects were: (1) Caucasian race; (2) aged under 21; (3)
18 diagnosis of ASD with the Diagnostic and Statistical Manual of Mental Disorders
20 error of less than 6 spherical diopters and 2.5 cylinder diopters; (5) best
22 ocular diseases (including ocular surgery); (7) no systemic disease that may
24 and correctly segmented SD-OCT scans without artifacts (see below) so only
5
1 Control NT subjects were selected age- and gender-matched with ASD
2 subjects. The inclusion criteria for the NT group were the same that to the
3 criteria used for the ASD group except for criterion number 3. An additional
4 inclusion criterion for the NT subjects was to have no family relationship with
5 ASD subjects. The control group was recruited from healthy volunteers who
9 under 18. Older subjects signed the consent themselves or, alternatively,
11 accordance with the tenets of the Declaration of Helsinki and had been
13
14 Ophthalmic examinations
17 BCVA, air-puff tonometry and funduscopy. The data of these examinations were
23 each subject: a macular fast volume scan and a standard pRNFL thickness
24 scan.
6
1 A macular fast volume scan comprised 25 horizontal scans with an area
2 of 666 mm2 centered at the fovea. In this examination, mean retinal thicknesses
3 were estimated in microns in the sectors that correspond to the Early Treatment
4 Diabetic Retinopathy Study (ETDRS) grid (Figure 1A). Macular layers were
8 total retina, nerve fiber layer (mRNFL), ganglion cell layer (GCL), inner
9 plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL),
10 outer nuclear layer (ONL) and photoreceptors. The sums of all the total inner
12 layers were also calculated for this study. OPL and ONL layers were not
14 Henle´s fiber orientation.16 Thus the thicknesses of these two layers were
16 resulting sectors (C0, S1, S2, N1, N2, T1, T2, I1, I2) of the ETDRS grid were
17 averaged and this value was considered to be the macular thickness for each
18 segmentation. The inner 1-mm diameter circle (C0) thickness was considered
22 Fovea-to-disc adjustment (FoDI system) was used in all eyes. With this strategy
23 the reference line of the papillomacular bundle was delineated (Figure 2A) if
24 the patient had foveal fixation. This line is considered to lie in the middle of the
25 temporal quadrant and, with this reference, all the other sectors were
7
1 determined by the device. If the subject was unable to maintain foveal fixation
3 changed.17 In this study the foveal position of all the examinations was
5 (JJGM).
7 sectors: temporal, temporal superior (TS), temporal inferior (TI), nasal, nasal
8 superior (NS) and nasal inferior (NI) (Figure 2B). The superior and inferior
9 quadrant values were also calculated by averaging the TS-NS and TI-NI
10 sectors, respectively.
11 Poor quality images (signal strength under 20) were not included. All the
12 scans were checked by the same experienced operator (JJGM) to exclude any
14 the pRNFL circle, or any other artifacts. Scans were repeated in such cases.
15
17 All the ASD subjects underwent the Kaufman brief intelligence test (K-
18 BIT) by the same experienced psychologist (ABRH). This was a quick test that
21 subtest. The nonverbal subtest consisted of items that contain pictures and
22 abstract symbols, rather than words, and measured the ability to solve problems
24 determined knowledge of words and their meanings. The K-BIT test provided
25 three intelligence quotient (IQ) scores: nonverbal, verbal and composite IQ.
8
1 These three scores had a mean of 100 points and a standard deviation of 15 in
3 calculated as the nonverbal score minus the verbal score, was considered for
4 the analysis in this study because it has been suggested that the nonverbal IQ
5 is generally higher than the verbal IQ in ASD, although this fact has not been
6 universally observed.19,20
11 flexible tape around the widest circumference of the head (above the eyebrow
12 to the most prominent part of the occipital bone). Three measurements were
13 taken and the largest value was considered for each patient.
14
20
21 Statistical analysis
24 were analyzed with the SPSS software (version 22.0; SPSS Inc, Chicago, IL,
25 USA) and R statistical software (version 3.3.1, The R Foundation for Statistical
9
1 Computing, Vienna, Austria). The values of the right and the left eyes of each
4 histograms and using Shapiro-Wilks test. An unpaired Student’s t-test was used
6 groups. Benjamini-Hochberg method was used to adjust the results for multiple
7 comparisons.22
9 results and head circumference, but only if they accomplished two criteria: (1) a
10 statistically significant difference for the Student’s t-test between the ASD and
11 NT groups; (2) this difference is greater than 5 microns which is the threshold of
13 thickness were also obtained to compare our results with the results by Emberti
19
20 RESULTS
21 Thirty-two ASD subjects were initially selected for the study, but five were
22 excluded due to poor collaboration for SD-OCT. Finally, 27 ASD subjects and
23 then 27 age- and gender-matched NT controls were included. Each group was
24 made up of 23 men and 4 women, whose mean age was 13.70+3.03 (range
10
1 from 7 to 20 years). Thus, 54 eyes from each group (108 eyes in all) were
3 The comparisons showed greater foveal thickness of the total retina, total
4 inner retina, IPL and INL in the ASD group (Table 1). In contrast, macular
5 thicknesses only statistically differed between both groups at the total retina and
6 total inner retina layers (Table 2). When comparing the pRNFL results, the TI,
8 than in the NT group (Table 3). The general trends toward thicker
9 segmentations in the foveal, macular and pRNFL thicknesses in the ASD group
11 The K-BIT scores in the ASD group were 85.88+21.04 (range from 46 to
12 123) for the nonverbal IQ, 79.92+20.92 (range from 40 to 112) for the verbal IQ,
14 (range from -30 to 35) for the NVIQ-VIQ D. The correlations of the selected
15 thickness parameters and the K-BIT obtained in the ASD group demonstrated a
16 moderate positive association between the TI sector of pRNFL and three IQs:
17 nonverbal, verbal and composite IQ (Table 4). The inferior pRNFL also showed
18 a positive association with the composite IQ, which came close to statistical
19 significance for the nonverbal and verbal IQs (Table 4). Scatterplots of the
24 (Table 4).
25
11
1 DISCUSSION
7 The pRNFL, the paradigm of white matter in the retina, was significantly thicker
8 in the inferior sectors in the ASD group. However, the layers that represent gray
9 matter did not seemed to present this disparity (except for the foveal INL). The
10 behavior of the inner retinal layers (altered in ASD) also differed from the outer
13 who performed the only research that has compared pRNFL between ASD and
14 NTs published to date.14 Briefly, these authors concluded that young ASD
15 adults (n=24) presented lower nasal pRNFL thickness values than the matched
16 NT subjects. They also divided the ASD subjects into two subcategories: high-
17 functioning autism –HFA- (n=11) and Asperger´s Syndrome -AS- (n=13). They
18 found thinner global, nasal and inferior pRNFL in the HFA group and thinner
19 nasal pRNFL in the AS compared to the controls. Our study considered all the
20 ASD patients since DSM V merged all the subcategories into one: ASD. No
22 The differences between the results of Emberti Gialloreti14 et al. and ours
24 used the Spectralis OCT software, version 5.1, while we used version 6.0. Thus
25 their version did not include fovea-to-disc adjustment (the FoDI system is
12
1 available from software version 5.3). This is relevant because this correction
3 patient head tilt so that scans can be accurately compared point-by-point with
10 of some intraretinal layers. This result was consistent with the results of other
12 structures of the visual pathway such as the thalamus23,24 and visual cortex.25,26
13 More recently, larger relative gray matter volume in the visual network in ASD
20 reduced in this disorder13. Considering that ASD and schizophrenia share some
24
13
1 One of these factors could be that ASD subjects may present an altered
2 programmed cell death (in other words, altered physiological apoptosis) of the
3 brain and the retina during development, which could lead to neuroanatomic
6 NT controls between the ages of 2-4 years that persists until the age of 5-6
7 years, after which no significant difference is detected.6 We did not find any
9 thicknesses in ASD subjects. To the best of our knowledge, this is the first study
10 dealing with these associations in ASD. It should be taken into account that the
11 ages of the patients in our sample ranged from 7 to 20 years and that this study
16 between ASD individuals and NT controls may be related to the atypical visual
18 performance during static visual tasks34-36 but ASD individuals showed inferior
20 recently that the visual field was narrower in ASD patients than in NT controls38
22 are needed to explore whether these functional differences are related to the
23 structural differences.
24 The differences found in this work also suggest that the normative
14
1 subjects. Thus, the interpretation of SD-OCT results in ASD individuals should
3 disease (such as glaucoma, papilledema, optic disk drusen, etc...) in order not
5 for inferior pRFNL sectors because the mean thickenings in ASD subjects at
6 these locations were 11.29 µm, 11.37 µm and 12.30 µm for temporal inferior,
9 applied in this study. This is a controversial issue40 because the risk of false
11 adjustment22, thickness of total retina and total inner layer at the fovea, and
13 1-3).
18 this association in the present study. Instead, we found positive correlations, but
19 only moderate ones, between inferior pRNFL thicknesses and IQs. These
20 findings suggest that the relationship of pRNFL and cognitive abilities in ASD
22 Our study has some limitations. First, the number of individuals included
24 size. Second, our ASD group was made up of young, high-functioning ASD
15
1 low-functioning ASD individuals. Third, our study is cross-sectional and
3 and NT groups will be stable or change over time. Longitudinal studies are
7 measured by SD-OCT. This fact should be taken into account when interpreting
9 between pRNFL thicknesses and cognitive scores were found. Considering that
11 suggest that this in vivo diagnosis technique may be promising to study different
13
14 ACKNOWLEDGMENTS
15 Funding: None.
16
18
19 Contribution of authors:
20 Designed the experiment: all authors; Conducted the experiment: all authors;
23
16
1 support in this study, and Jose Manuel Tamarit (from Heidelberg Engineering,
10
11
12
13
14
15
16
17
18
19
20
21
17
1 REFERENCES
2 1. Wing L. The autistic spectrum. Lancet 1997; 350: 1761-1766.
6 3. Murphy CM, Wilson CE, Robertson DM, et al. Autism spectrum disorder in
14 disorder: brain structure and function across the lifespan. Lancet Neurol
15 2015;14:1121-1134.
22 8. Yanoff M, Sassani JW. Neural (Sensory) Retina. In: Yanoff M, Sassani JW.
18
1 10. Garcia-Martin E, Larrosa JM, Polo V, et al. Distribution of retinal layer
4 11. Garcia-Martin E, Bambo MP, Marques ML, et al. Ganglion cell layer
10 584.
20 2013:1-991.
21 16. Lujan BJ, Roorda A, Croskrey JA, et al. Directional Optical Coherence
22 Tomography Provides Accurate Outer Nuclear Layer and Henle Fiber Layer
19
1 fiber layer thickness using the new FoDi software in spectralis optical
3 18. Kaufman AS, Kaufman NL. K-BIT. Kaufman brief intelligence test. Spanish
5 100
8 2011;129:59-70.
16 Methodol 1995;57:289-300.
17 23. Ecker C, Suckling J, Deoni SC, et al. Brain anatomy and its relationship to
20 24. Lin HY, Ni HC, Lai MC, Tseng WY, Gau SS. Regional brain volume
21 differences between males with and without autism spectrum disorder are
20
1 revealed by cortical thickness analysis and voxel-based morphometry. Hum
8 28. Bigler ED, Abildskov TJ, Petrie JA, et al. Volumetric and voxel-based
16 31. Valenciano AI, Boya P, de la Rosa EJ. Early neural cell death: numbers and
17 cues from the developing neuroretina. Int J Dev Biol 2009;53: 1515-1528.
19 Neurosci 2014;36:13-18
23 48:497-507.
24 35. Simmons DR, Robertson AE, McKay LS, Toal E, McAleer P, Pollick FE.
21
1 36. Samson F, Mottron L, Soulières I, Zeffiro TA. Enhanced visual functioning in
5 38. Song Y, Hakoda Y, Sanefuji W, Cheng C. Can They See It? The Functional
8 39. Constable PA, Gaigg SB, Bowler DM, Jägle H, Thompson DA. Full-field
10 2016;132:83-99.
12 Epidemiology 1990;1:43-46.
13 41. van Koolwijk LME, Despriet DDG, Van Duijn CM, et al. Association of
16 42. Khawaja AP, Chan MP, Yip JL, et al. Retinal Nerve Fiber Layer Measures
19
20
21
22
23
24
25
22
1 Legends
2 Figures
4 sectors (C0, S1, S2, N1, N2, T1, T2, I1 and I2) were averaged to obtain macular
5 thickness values. Thickness in the central circle (C0) was considered the foveal
7 macula.
11 around the optic disc, divided in sectors (see below). Fovea-to-disc line is
13 from the device. The sectors are: temporal (T), temporal superior (TS), temporal
14 inferior (TI), nasal (N), nasal superior (NS) and nasal inferior (NI). Global value
15 (G) represent the mean thickness of all the sectors. The numbers indicate the
17
19 pRNFL thicknesses and K-BIT results in ASD group. (A) Temporal Inferior
21 (B) TI peripapillary RNFL thickness versus verbal IQ. (C) TI peripapillary RNFL
22 thickness versus composite IQ. (D) Inferior peripapillary RNFL thickness versus
23 composite IQ. Thickness is indicated in microns. Cubic regression best fit the
25
23
1 Tables
11
15 0.01 level.
16
17
18
19
20
24
Foveal measurements
Segmented ASD subjects NT subjects Difference of Independent
layer (mean + (mean + means T-test
standard standard (ASD-NT) and P-value
deviation) deviation) % of change
Total retina 279.16 + 20.57 268 + 17.84 11.16 (4.16%) <0.01 **