ScT

TMa

b

c

d

e

a

ARRAA

KAPLC

1

lpsthac

S8

0d

Atherosclerosis 210 (2010) 542–547

Contents lists available at ScienceDirect

Atherosclerosis

journa l homepage: www.e lsev ier .com/ locate /a therosc leros is

erum low-density lipoprotein cholesterol level is strong risk factor for acquiredolor vision impairment in young to middle-aged Japanese men:he Okubo Color Study Report 2

akuhei Shoji a,e,∗, Yutaka Sakuraib, Hiroki Satoc, Etsuo Chiharad,asahiro Ishidae, Kazuyuki Omaea

Department of Preventive Medicine and Public Health, School of Medicine, Keio University, Tokyo, JapanDepartment of Preventive Medicine and Public Health, National Defense Medical College, Tokorozawa, Saitama, JapanDepartment of Medical Informatics, National Defense Medical College, Tokorozawa, Saitama, JapanSensho-kai Eye Institute, Uji, Kyoto, JapanDepartment of Ophthalmology, National Defense Medical College, Tokorozawa, Saitama, Japan

r t i c l e i n f o

rticle history:eceived 26 July 2009eceived in revised form 11 October 2009ccepted 13 November 2009vailable online 1 December 2009

eywords:cquired color vision impairmentrevalenceow-density lipoprotein cholesterolardiovascular risk factors

a b s t r a c t

Objective: To investigate associations between blood low-density lipoprotein cholesterol (LDL-C) levelsand the prevalence of acquired color vision impairment (ACVI) in middle-aged Japanese men.Methods: Participants in this cross-sectional study underwent color vision testing, ophthalmic exami-nation, a standardized interview and examination of venous blood samples. Ishihara plates, a Lanthony15-hue desaturated panel, and Standard pseudoisochromatic Plates part 2 were used to examine colorvision ability. The Farnsworth–Munsell 100-hue test was performed to define ACVI. Smoking status andalcohol intake were recorded during the interview. We performed logistic regression analysis adjustedfor age, LDL-C level, systemic hypertension, diabetes, cataract, glaucoma, overweight, smoking status,and alcohol intake. Adjusted odds ratios for four LDL-C levels were calculated.Results: A total of 1042 men were enrolled, 872 participants were eligible for the study, and 31 subjectswere diagnosed with ACVI. As compared to the lowest LDL-C category level (<100 mg/dl), the crude OR of

ACVI was 3.85 (95% confidence interval [CI], 1.24–11.00) for the 2nd highest category (130–159 mg/dl),and 4.84 (95% CI, 1.42–16.43) for the highest level (≥160 mg/dl). The multiple-adjusted ORs were 2.91(95% CI, 0.87–9.70) for the 2nd highest category and 3.81 (95% CI, 1.03–14.05) for the highest level. Testsfor trend were significant (P < 0.05) in both analyses.Conclusions: These findings suggested that the prevalence of ACVI is higher among middle-aged Japanesemen with elevated LDL-C levels. These changes might be related to deteriorated neurologic function

aboli

associated with lipid met

. Introduction

Population studies have reported consistently that increasedow-density lipoprotein cholesterol (LDL-C) is a strong and inde-endent risk factor for cardiovascular diseases [1]. Evidence alsouggests a relation between lipids and vascular changes involving

he brain in dementia [2]. Recent experimental and clinical reportsave discussed in depth the link between late-onset neurodegener-tive diseases and the asymptomatic atherosclerosis, or the bloodholesterol or glycemia levels that occur during mid-life [2–6].

∗ Corresponding author at: Department of Preventive Medicine and Public Health,chool of Medicine, Keio University, 35, Shinanomachi, Shinjyuku-ku, Tokyo 160-582, Japan. Tel.: +81 3 5363 3758; fax: +81 3 3359 3686.

E-mail address: t-shoji@yg7.so-net.ne.jp (T. Shoji).

021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.atherosclerosis.2009.11.039

te abnormalities.© 2009 Elsevier Ireland Ltd. All rights reserved.

The deterioration of the color vision discrimination ability,which is referred to as acquired color vision impairment (ACVI),is generally believed to be an early indicator of various neu-rotoxic and neurodegenerative diseases [7,8]. However, becauseACVI is a secondary feature of a variety of pathological states,there are few studies that have reported on the impact of car-diovascular risk factors or life style factors on ACVI [7–15].In addition, even though these risk factors may influence thecolor vision function via several different mechanisms, at thepresent time, these relationships have yet to be fully under-stood.

The Okubo Color Study has focused attention on the impact ofcardiovascular risk factors on color vision function in male subjectsin the Japanese Self Defense Force (JSDF). The primary aim of thecurrent study was to investigate these relationships with cross-sectional multivariable analysis.

lerosi

2

2

vvbevDdBvws2scctcbutTcp1aSftbfdcAis

2

2

wm

dm

cSha

atrtw

mt

T. Shoji et al. / Atherosc

. Subjects and methods

.1. Study population and design

The Okubo Color Study evaluated the relationship between colorision function, especially ACVI, and ophthalmic disease, cardio-ascular risk factors, and lifestyle factors. The study was reviewedy the National Defense Medical College (NDMC) Ethics Board tonsure that it met the ethical guidelines, with all subjects pro-iding written informed consent in accordance with the Helsinkieclaration prior to their participation in the study. The detailedesign of the current study has been previously described [16].riefly, this study was a large-sample prevalence survey of colorision impairment among male officials aged 20–60 years whoere on active duty in the JSDF, Okubo Garrison, Kyoto, Japan. All

ubjects underwent an annual health examination between April005 and March 2006 at the Okubo Garrison. The authors usedtructured questionnaires, personal physical records, along withonducting detailed interviews in order to collect information onurrent and past diseases, medication use, and lifestyle characteris-ics in all of the subjects. Subsequently, all participants underwentolor vision testing, measurement of visual acuity (VA), slit-lampiomicroscopic examination, measurement of intraocular pressuresing non-contact applanation tonometry, and fundus examina-ion. Venous blood samples were also collected from all subjects.he VA was measured using a Landolt ring chart (with refractiveorrection, if any) at a distance of 5 m. Ishihara pseudoisochromaticlates (Kanehara Shuppan Co. Ltd., Tokyo, Japan), the Lanthony5-hue desaturated panel (D-15 DS) plates (Luneau, Paris, France),nd the Standard pseudoisochromatic plates part 2 (SPP2; Igaku-hoin, Tokyo, Japan) were used to perform color testing. Colorundus photographs were obtained using a fundus camera sys-em (Canon, CR-45NM, Tokyo, Japan) with an angle of 30◦. Venouslood was drawn after an overnight fast in order to measure theasting plasma glucose, total cholesterol (TC), triglyceride (TG), low-ensity lipoprotein cholesterol (LDL-C), high-density lipoproteinholesterol (HDL-C), uric acid (UA), and hemoglobin A1c (HbA1c).ll examiners and photograph readers were masked to all other

nformation on the subject including vision, referring diagnosis,ystemic status and life style characteristics.

.2. Measurements and definitions

.2.1. Examination color testingThe Ishihara pseudoisochromatic plates 1–21 (38-plate edition)

ere used to identify red-green color sense anomaly. Subjects withore than four errors were classified as having failed the test [7].The D-15 DS was used to supplement standard D-15 testing to

iagnose milder congenital anomalies or ACVI [8]. Subjects withore than one major crossing were classified as having failed [7,10].The SPP2 test was designed especially to detect ACVI and also

an be used to identify a blue-yellow color sense abnormality.ubjects with more than one blue-yellow error were classified asaving failed the test. Missing the 2 on Plate 3 was not consideredfailure [7,10].

The Farnsworth–Munsell (F–M) 100-hue test was performedccording to the standard technique [7]. Color discrimination wasested monocularly at a distance of 30 cm with the subjective nearefraction placed in a trial frame. Subjects with scores worse than

he 95th percentile described by Verriest et al. [17] were diagnosedith ACVI.

Subjects were inspected binocularly with the Ishihara plates andonocularly with D-15 DS, SPP2, and the F–M 100-hue test, with

he right eye tested before the left eye.

s 210 (2010) 542–547 543

2.2.2. Cardiovascular risk factor definitionIn accordance with the guidelines, the LDL-C level was cate-

gorized as less than 100 mg/dl, equal to or exceeding 100 mg/dlbut less than 130 mg/dl, equal to or exceeding 130 mg/dl butless than 160 mg/dl, or equal to or exceeding 160 mg/dl [1]. Sys-temic hypertension was defined as a systolic/diastolic pressure of140/90 mmHg or more at the time of examination or a history of adiagnosis of hypertension and current use of medication for hyper-tension [1]. Diabetes was defined as a previous history of diabetestreated with either insulin, oral hypoglycemic agents, or diet. Newlydiagnosed diabetes was defined as no previous medical history ofdiabetes in the presence of an elevated fasting plasma glucose levelof 126 mg/dl or more at the time of the examination [18].

2.2.3. Smoking and drinking statusSubjects were categorized as those who had never smoked and

current and past smokers. Current and past smokers were definedas current or past habitual smokers who had smoked for at least 1year. Subjects who had been smoking for less than 1 year were notconsidered to be smokers. Cigarette smokers were classified intotwo groups, namely, light smokers with 19 or fewer median pack-years of smoking and heavy smokers with 19 or more pack-yearsof smoking.

Classification of alcohol consumption was based on the fre-quency of alcohol consumption (daily, weekly, monthly, or lessoften).

2.2.4. OverweightThe body mass index (BMI), a standard measure defined as the

weight in kilograms divided by the square of the height in meters,was categorized as less than 25 or 25 or higher, with the cutoffpoint chosen as a criterion for obesity because it corresponds tothe standard indicator for overweight in an Asian population [19].

2.2.5. Cataract gradingOphthalmologists used a slit-lamp examination to grade the

cataract level, which was based on Lens Opacity Classification Sys-tem III (LOCS III) [20] standard color photographs and the WilmerClassification [21]. We defined the presence of a nuclear cataract asat least one eye with a nuclear opacity of grade 3.0 or higher on theLOCS III. Cortical cataract and posterior subcapsular cataract wereconsidered present with at least a grade 2 and grade 1 or higher onthe Wilmer Classification, respectively.

2.2.6. Glaucoma definitionSubjects were recruited for visual field evaluation if the intraoc-

ular pressure was more than 20 mmHg, or if the color fundusphotography showed abnormality findings, which included glau-comatous optic disc appearance and nerve fiber layer defect. Visualfield testing was performed with static perimetric analysis using aHumphrey Field Analyzer, 30-2 SITA Standard Program (HumphreyInstruments, Inc., San Leandro, CA, USA) or the Octopus 301 full-threshold G1 program ver. 2.04, Interzeag, Schlieren, Switzerland).The final diagnosis of glaucoma was based on optic disc appear-ance, intraocular pressure, and the results of visual field testingand the clinical records, which included physical records and ques-tionnaires.

2.2.7. Age-related macular degeneration (AMD) gradingEach color fundus photograph was graded based on a compari-

son with standard photographs according to the Rotterdam Study[22] and classified as normal, early age-related maculopathy (stages2 or 3), or AMD.

544 T. Shoji et al. / Atherosclerosis 210 (2010) 542–547

kubo

2

tVhnrwcsa(wcD(tatft

Fig. 1. Schematic diagram of the O

.3. Inclusion and exclusion criteria

The study criteria are summarized in Fig. 1. The inclusion cri-eria included a best-corrected VA of at least 0.7 on the decimalA chart (73 subjects were excluded). Subjects were excluded whoad undergone cataract or retinal surgery and any corneal or reti-al disease, including AMD, central serous chorioretinitis, diabeticetinopathy, and any retinal or vitreous hemorrhage (6 subjectsere excluded). Considering the possibility of the side effects on

olor vision, subjects currently taking or those who had taken aystemic medication such as digitalis glycoside, antiepileptic drugs,ntimalaria drugs, and antituberculosis drugs [9,10] were excluded4 subjects were excluded). Moreover, 75 subjects were excludedho failed testing with the Ishihara plates, had a pattern typi-

al of congenital deutan or protan defect on testing with SPP2 or-15 DS, or had a history of congenital color vision impairment

CCVI) in their medical history records, in which the results of color

esting had been recorded on enlistment. In questionable cases,nomaloscopy was used. Subjects who failed testing with eitherhe SPP-2 or D-15 DS were classified as ACVI suspects and per-ormed the 100-hue test bilaterally. Twelve subjects who declinedhe definition test also were excluded. Subjects exceeding the age-

Color Study and diagnosis of ACVI.

specific limit score of Verriest et al. [17] were diagnosed as havingACVI.

2.4. Statistical analysis

Data were expressed as the mean ± the standard deviation (SD)or median with the interquartile range as appropriate. Variableswere compared with the unpaired t-test for normal distributedvariables or the Wilcoxon rank sum test for non-normal distributedvariables between ACVI and normal subjects. Multiple logisticregression analysis was performed to examine the relationshipbetween LDL cholesterol and ACVI. The model was adjusted withrisk factors including age, smoking, drinking, and overweight, anddiagnosed hypertension, diabetes, cataract, and glaucoma. Oddsratios (ORs) with 95% confidence intervals (95% CI) are presented.Individuals, rather than eyes, were the unit of analysis because the

right and left eyes were not examined independently. A participantwas considered to have CCVI or ACVI when it occurred in either eye.A P-value less than 0.05 indicated a statistically significant differ-ence. All statistical analyses were performed using SAS version 9.1software (SAS Institute, Inc., Cary, NC, U.S.A.).

T. Shoji et al. / Atherosclerosis 210 (2010) 542–547 545

Table 1Baseline characteristics of 872 eligible subjects.

Acquired color vision impairment P-value

Yes (N = 31) No (N = 841)

Age (yr) 40.7 ± 8.2 37.8 ± 8.6 0.063Systolic blood pressure (mmHg) 129 (115, 140) 120 (111, 131) 0.024Diastolic blood pressure (mmHg) 80 (72, 93) 75 (66, 83) 0.014LDL cholesterol (mg/dl) 141.8 ± 34.8 120.6 ± 32.6 <0.001HDL cholesterol (mg/dl) 65.2 ± 18.1 63.7 ± 14.8 0.572Triglycerides (mg/dl) 101.7 ± 60.3 104.5 ± 71.8 0.832Fasting blood glucose (mg/dl) 94 (89, 100) 90 (85, 96) 0.020Height (cm) 169.5 ± 7.1 170.3 ± 5.8 0.489Weight (kg) 70.8 ± 9.7 68.5 ± 9.0 0.172BMI (kg/m2) 24.6 ± 2.7 23.6 ± 2.7 0.051

P ta arew um teL and A

3

3

aattwhC

TR

Ao

lus–minus values are means ± SD. Alternatively, in case of skewed distributions, daith the unpaired t-test for normal distributed variables or by the Wilcoxon rank s

DL, low-density lipoprotein; HDL, high-density lipoprotein; BMI, body mass index

. Results

.1. Baseline characteristics of the study participants

The study included 1042 out of a total of 1152 subjects, forn overall response rate of 90.5%. Of the 1042 enrolled subjects,total of 872 (83.7%) fulfilled the inclusion criteria (Fig. 1). Of

hese subjects, 31 were diagnosed with ACVI. Table 1 presentshe characteristics according to whether or not the participantsere diagnosed with ACVI. Subjects with ACVI were more likely toave higher systolic blood pressure, diastolic blood pressure, LDL-, and fasting plasma glucose. Marginal differences were observed

able 2isk factors for acquired color vision impairment.

Risk factors ACVI (N = 31) No ACVI (N = 841) Crude

LDL cholesterol<100 4 237 1.00100–129 6 306 1.16 (130–159 13 200 3.85 (≥160 8 98 4.84 (

P for Trend

Age, years<40 9 403 1.00≥40 22 438 0.45 (

HypertensionNo 22 740 1.00Yes 9 101 3.00 (

DiabetesNo 27 826 1.00Yes 4 15 8.16 (

CataractNo 30 831 1.00Yes 1 10 2.77 (

GlaucomaNo 29 826 1.00Yes 2 15 3.80 (

OverweightNo 17 606 1.00Yes 14 235 2.12 (

SmokingNever 16 319 1.00Light 6 263 0.46 (Heavy 9 259 0.69 (

DrinkingMonthly or less 9 264 1.00Weekly 12 341 1.03 (Daily 10 236 1.24 (

bbreviations: ACVI, acquired color vision impairment and LDL, low-density lipoproteiverweight (defined by body mass index), diagnosed hypertension, diabetes, cataract, an

medians (25th percentile, 75th percentile). Baseline characteristics were comparedst for non-normal distributed variables between ACVI and no ACVI. Abbreviations:CVI, acquired color vision impairment.

for the age and BMI. There was no significant difference betweenthe groups for HDL-C, triglycerides, height and weight.

3.2. Relation between parameters of cardiovascular risk factorand ACVI prevalence

Table 2 shows the distributions of risk factors in the groups

with and without ACVI and the rude and adjusted ORs with 95%CI for subjects with ACVI compared with no ACVI. Compared withthe reference LDL-C category (less than 100 mg/dl), the crude andmultiple-adjusted OR for the highest LDL-C level (≥160 mg/dl)was 4.84 (95% CI, 1.42–16.43), and 3.81 (95% CI, 1.03–14.05),

ORs (95%CI) P-value Adusted ORs (95%CI) P-value

1.000.32–4.16) 0.080 0.93 (0.25–3.49) 0.0751.24–11.00) 0.042 2.91 (0.87–9.70) 0.1061.42–16.43) 0.014 3.81 (1.03–14.05) 0.032

0.001 0.007

1.000.20–0.98) 0.044 0.68 (0.27–1.71) 0.411

1.001.34–6.69) 0.007 1.84 (0.76–4.48) 0.180

1.002.54–26.22) <0.001 7.76 (1.98–30.39) 0.003

1.000.34–22.34) 0.339 1.07 (0.09–12.97) 0.959

1.000.83–17.39) 0.086 2.99 (0.55–16.28) 0.205

1.001.03–4.38) 0.041 1.40 (0.63–3.11) 0.415

1.000.18–1.18) 0.193 0.64 (0.24–1.76) 0.9600.30–1.60) 0.949 0.44 (0.18–1.09) 0.200

1.000.43–2.49) 0.838 1.31 (0.50–3.41) 0.8320.50–3.11) 0.608 1.45 (0.53–3.92) 0.571

n (mg/dl). The multiple-adjusted ORs were adjusted for age, smoking, drinking,d glaucoma. overweight; Yes; (BMI>25.0) and no (BMI ≤ 25.0).

5 clerosi

rripwTCrd

4

cosFcrgt

tsaptfjsda

rhimaaitcAnrTaorefiri[

thtmnrCsaAc

46 T. Shoji et al. / Atheros

espectively. Tests for trend were significant in multiple-adjustedegression analysis (P = 0.007), which indicated a significant pos-tive correlation between the elevated LDL-C levels and therevalence of ACVI. The crude OR was also significant in subjectsith diagnosed hypertension, diabetes, and those who were obese.

he adjusted OR in subjects with diagnosed diabetes was 7.76 (95%I, 1.98–30.39), which was significant (P = 0.003). Multiple logisticegression analysis did not identify correlations between ACVI andiagnosed hypertension, obesity, cataract, or glaucoma.

. Discussion

Results of the current study demonstrate that there is a signifi-ant relationship between the blood LDL-C level and the prevalencef ACVI. In addition, the multiple logistic regression analysis resultsuggest that the blood LDL-C is an independent risk factor for ACVI.urthermore, the multiple logistic regression analyses also indi-ated there was a significant trend for an increased risk of ACVIisk associated with higher LDL-C blood levels. This finding sug-ests that there might be a dose–response relationship betweenhe increased risk of ACVI and increased LDL-C level.

To the best of our knowledge, few reports documented the rela-ionship between ACVI and blood LDL-C levels. Although one othertudy [23] has suggested there might be a relationship betweenn increased risk of ACVI and increased hypercholesterolemia, thisrevious study has a poorer reliability than our current study dueo the small sample size and the use of an unauthorized methodor evaluating the color vision. In our study, we found 31 ACVI sub-ects out of the 872 enrolled participants. In contrast to the previoustudy, we evaluated the ACVI by using the gold standards for ACVIiagnosis [7], the F–M 100-hue test and the criteria of Verriest etl. [17].

The etiology of ACVI is multifactorial, which includes pre-eceptoral, receptoral, and post-receptoral mechanisms [7]. ACVIas been demonstrated to occur in a variety of patient populations

n association with ocular diseases (senile cataract, age-relatedacular degeneration [AMD], cone degenerations, optic neuritis,

nd glaucoma), use of certain medications (antiepileptic drugs,nti-malarial drugs, and antituberculosis drugs), and various chem-cal substances (solvent mixtures, styrene, perchloroethylene,oluene, carbon disulphide, and ethanol) [7–10]. While it is diffi-ult to elucidate the effects of these factors on the pathology ofCVI [7], strict differential diagnosis criteria between ACVI ando ACVI and the multivariable analysis adjusted for the reportedisk factors [11–15] supports the verity of our current results.he lack of association between ACVI and ocular diseases suchs cataract and glaucoma is most likely because of the exclusionf the severe cases due to the deterioration in visual acuity. Theesults were unchanged if all cataract and glaucoma patients werexcluded from the analysis (table not shown). We were able to con-rm that diabetes without significant retinopathy is also a strongisk factor for ACVI, which is in agreement with previous find-ngs reported in the Early Treatment Diabetic Retinopathy Study11].

The current results would indicate that increased LDL-C is ableo cause morphologic and functional changes in the retina and theigher visual pathways due to atherosclerotic or other effects onhe ocular capillaries or ocular neurons [7,8]. Although the detailed

echanism remains unclear, there are several possible mecha-isms that may account for our findings. Recently, Alcala et al.eported finding a significant positive correlation between the LDL-

level and the deterioration in the visual field [24]. These results

uggest that a rise in cholesterol plasma concentration might neg-tively modify the function of the neuron entrusted with vision.nimal studies have shown that lipid metabolism dysfunctionan induce ultrastructural changes in the retinal cells, increase

s 210 (2010) 542–547

the intraneural immunoreactivity of Amyloid beta-peptide (Ab)[6], and significantly reduce cell numbers in the retinal layers[25]. Moreover, electrophysiologic studies have shown there is animpairment of the visual pathways in hypercholesterolemic mice[25,26]. These studies suggest that blood cholesterol might play animportant role in the deterioration of the retinal or higher visualfunction, which are consistent with current results.

There is growing evidence that cholesterol is linked to the devel-opment of neurodegenerative disease, in particular, Alzheimer’sdisease (AD) [2–6]. It has also been proposed that vascular factorsmight have a toxic effect on the microvasculature of susceptiblebrain regions [2]. Evidence has been reported that suggests thereis axonal degeneration in the optic nerve in AD, with a markeddecreased in the number of ganglion cells [27]. Thus, an alterna-tive hypothesis is that ACVI due to hyperlipemia may be linkedto early subclinical neurological symptoms in the late-onset neu-rodegenerative diseases. Electrophysiologic studies support thenotion that the visual pathways in patients with AD are impairedas compared to those found in patients without AD [28]. Patho-logical and intravascular ultrasound studies of young adults havedemonstrated that histological evidence of neurofibrillary tangleformation can be seen starting from 40 to 50 years prior to theonset of dementia [1]. These pathological findings are not contraryto our results, as the deterioration of color discrimination ability haslong been recognized as an impairment that can precede the onsetof clinically detectable changes [6–9]. Nevertheless, these relation-ships still have to be considered to be hypothetical [28] due to thelong latency that is associated with their incidence.

There are several limitations that should be kept in mind withregard to our current results. First, even though our findings sug-gest that there is a relationship between the LDL-C levels andthe prevalence of ACVI, the etiology and the detailed mechanismshave yet to be clearly elucidated. A second limitation is related todesign of the current study. Cross-sectional studies are not ableto determine the long-term changes in the LDL-C levels and theassociated time points for the incidence of ACVI. Therefore, inorder to assess the causal relationship between the LDL-C levelsand the incidence of ACVI, a longitudinal approach needs to beemployed. A final limitation is that since the study population con-sisted entirely of middle-aged healthy Japanese males, this mightrestrict the external validity of the current study. Woo and Lee havereported that differences in macular pigmentation between Cau-casians and Asians can cause substantial differences in the resultsof the F–M 100-hue test [29]. However, to our knowledge, few stud-ies have investigated the prevalence of ACVI in a large population ofmiddle-aged healthy Japanese males. The prevalence rate that wasdetermined for the 40-year age span of our subjects was approx-imately 4.5%, which appears to be similar to the results reportedby Verriest et al. [17]. Thus, the high participation rate (90.3%), lowamounts of missing data, use of eligibility criteria, and an approveddiagnostic method suggest our study is internally valid. To furtherconfirm our results, a prospective cohort study that employs a moregeneralized population will need to be undertaken.

In conclusion, the findings of this cross-sectional study indicatethat the blood LDL-C levels have a positive relationship with theprevalence of ACVI, which is one of the indicators of neurologicalsymptoms. This relationship might possibly be explained by thedeterioration of the neurological function that is associated withthe abnormality of the lipid metabolites. Our current data is alsouseful and important as it provides information that can be used tofurther investigate many of these associations.

Conflict of interest

The authors have no proprietary interest in any aspect of thisstudy.

lerosi

A

tpKpC

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[27] Hinton DR, Sadun AA, Blanks JC, et al. Optic-nerve degeneration in Alzheimer’sdisease. N Engl J Med 1986;315:485–7.

T. Shoji et al. / Atherosc

cknowledgments

The authors wish to thank The Okubo Color Study team, in par-icular, K. Tanaka for subject recruitment; A. Katagiri for help inreparing the study; K. Sintomi, Y. Usuda, T. Kanazaki, N. Nakai, Y.omatsu, and T. Bessho for clinical input; and the volunteers forarticipating in the study. In addition, we wish to thank Ms. Lyndaharters for her linguistic and editorial support.

eferences

[1] Third Report of the National Cholesterol Education Panel (NCEP) Expert Panelon Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults(Adult Treatment Panel or ATP III). NIH Publication No. 01-3305, May 2001.

[2] Casserly I, Topol E. Convergence of atherosclerosis and Alzheimer’s disease:inflammation, cholesterol, and misfolded proteins. Lancet 2004;363:1139–46.Review.

[3] Hofman A, Ott A, Breteler MM, et al. Atherosclerosis, apolipoprotein E, andprevalence of dementia and Alzheimer’s disease in the Rotterdam Study. Lancet1997;349:151–4.

[4] Luchsinger JA, Reitz C, Honig LS, et al. Aggregation of vascular risk factors andrisk of incident Alzheimer disease. Neurology 2005;65:545–51.

[5] Wolozin B. Cholesterol and the biology of Alzheimer’s disease. Neuron2004;41:7–10. Review.

[6] Trivino A, Ramírez AI, Salazar JJ, et al. A cholesterol-enriched diet inducesultrastructural changes in retinal and macroglial rabbit cells. Exp Eye Res2006;83:357–66.

[7] Melamud A, Hagstrom S, Traboulsi E. Color vision testing. Ophthalmic Genet2004;25:159–87. Review.

[8] Gobba F. Related color vision: a sensitive indicator of exposure to neurotoxins.Neurotoxicology 2000;21:857–62. Review.

[9] Gobba F, Cavalleri A. Color vision impairment in workers exposed to neurotoxicchemicals. Neurotoxicology 2003;24:693–702. Review.

10] Vu BL, Easterbrook M, Hovis JK. Detection of color vision defects in chloroquineretinopathy. Ophthalmology 1999;106:1799–803.

11] Fong DS, Barton FB, Bresnick GH. Impaired color vision associated with diabeticretinopathy: Early Treatment Diabetic Retinopathy Study Report no. 15. Am JOphthalmol 1999;128:612–7.

12] Pacheco-Cutillas M, Edgar DF, Sahraie A. Acquired colour vision defectsin glaucoma-their detection and clinical significance. Br J Ophthalmol1999;83:1396–402. Review.

[

[

s 210 (2010) 542–547 547

13] Bimler D, Kirkland J. Multidimensional scaling of D15 caps: color-vision defectsamong tobacco smokers? Vis Neurosci 2004;21:445–8.

14] Hiratsuka Y, Li G. Alcohol and eye diseases: a review of epidemiologic studies.J Stud Alcohol 2001;62:397–402. Review.

15] Cheung N, Wong TY. Obesity and eye diseases. Surv Ophthalmol2007;52:180–95. Review.

16] Shoji T, Sakurai Y, Chihara E, et al. Reference intervals and discrimination valuesof the Lanthony desaturated D-15 panel test in young to middle-aged Japanesearmy officials: the Okubo Color Study Report 1. Eye 2009;23:1329–35.

17] Verriest G, Van Laethem J, Uvijls A. A new assessment of the normalranges of the Farnsworth-Munsell 100-hue test scores. Am J Ophthalmol1982;93:635–42.

18] American Diabetes Association. Standards of medical care in diabetes. Diab Care2006;(Suppl. 1):S4–42.

19] James PT, Leach R, Kalamara E, et al. The worldwide obesity epidemic. Obes Res2001;(Suppl. 4):228S–33S. Review.

20] Chylack Jr LT, Wolfe JK, Singer DM, et al. The Lens Opacities ClassificationSystem. III. The longitudinal study of cataract study group. Arch Ophthalmol1993;111:831–6.

21] Taylor HR, West SK. A simple system for the clinical grading of lens opacities.Lens Res 1988;5:175–81.

22] van Leeuwen R, Klaver CC, Vingerling JR, et al. The risk and natural course ofage-related maculopathy: follow-up at 6 1/2 years in the Rotterdam study. ArchOphthalmol 2003;121:519–26.

23] Alcalá A, Morell M, Rius F. Chromatic computerized analysis is an early predic-tor of cardiovascular risk associated to hypercholesterolemia. Rev Esp Cardiol2001;54:1417–25.

24] Alcalá A, Jansen S, Farkouh ME, et al. Hypercholesterolemia is associated withvisual field alterations detectable with computerized perimetry. Atherosclero-sis 2007;195:e167–71.

25] Ong JM, Zorapapel NC, Rich KA, et al. Effects of cholesterol and apolipoproteinE on retinal abnormalities in ApoE-deficient mice. Invest Ophthalmol Vis Sci2001;42:1891–900.

26] Bretillon L, Acar N, Seeliger MW, et al. ApoB100, LDLR−/− mice exhibit reducedelectroretinographic response and cholesteryl esters deposits in the retina.Invest Ophthalmol Vis Sci 2008;49:1307–14.

28] Katz B, Rimmer S. Ophthalmologic manifestations of Alzheimer’s disease. SurvOphthalmol 1989;34:31–43. Review.

29] Woo GC. Are ethnic differences in the F-M 100 scores related to macular pig-mentation? Clin Exp Optom 2002;85:372–7.