ORIGINAL RESEARCH ARTICLE

Multifocal electroretinogram responses in Nepalese diabeticpatients without retinopathy

Prakash Adhikari • Subash Marasini •

Raman Prasad Shah • Sagun Narayan Joshi •

Jeevan Kumar Shrestha

Received: 28 August 2013 / Accepted: 27 May 2014

� Springer-Verlag Berlin Heidelberg 2014

Abstract

Purpose To determine neuroretinal function with

multifocal electroretinogram (mfERG) in diabetic

subjects without retinopathy.

Methods Multifocal electroretinogram (mfERG)

was performed in 18 eyes of 18 diabetic subjects

without retinopathy and 17 eyes of 17 age and gender-

matched healthy control participants. Among 18

diabetic subjects, two had type 1 and 16 had type 2

diabetes. MfERG responses were averaged by the

retinal areas of six concentric rings and four quadrants,

and 103 retinal locations; N1–P1 amplitude and P1-

implicit time were analysed.

Results Average mfERG N1–P1 amplitude (in

nv/deg2) of 103 retinal locations was 56.3 ± 17.2

(mean ± SD) in type 1 diabetic subjects, 47.2 ± 9.3 in

type 2 diabetic subjects and 71.5 ± 12.7 in controls.

Average P1-implicit time (in ms) was 43.0 ± 1.3 in

type 1 diabetic subjects, 43.9 ± 2.3 in type 2 diabetic

subjects and 41.9 ± 2.1 in controls. There was signif-

icant reduction in average N1–P1 amplitude and delay

in P1-implicit time in type 2 diabetic subjects in

comparison to controls. mfERG amplitude did not

show any significant correlation with diabetes duration

and blood sugar level. However, implicit time showed

a positive correlation with diabetes duration in type 2

diabetic subjects with diabetes duration C5 years.

Conclusions This is the first study in a Nepalese

population with diabetes using multifocal electroretinog-

raphy. We present novel findings that mfERG N1–P1

amplitude is markedly reduced along with delay in P1-

implicit time in type 2 diabetic subjects without retinop-

athy. These findings indicate that there might be signif-

icant dysfunction of inner retina before the development

of diabetic retinopathy in the study population, which

have higher prevalence of diabetes than the global

estimate and uncontrolled blood sugar level.

Keywords Multifocal electroretinogram �Diabetes �Diabetic retinopathy � N1–P1 amplitude �P1-implicit time

Introduction

The 2012 global estimate is that there are approxi-

mately 93 million people living with diabetic retinop-

athy (DR) and among them, 28 million are with vision-

P. Adhikari (&)

Visual Science and Medical Retina Laboratories, Institute

of Health and Biomedical Innovation, School of

Optometry and Vision Science, Queensland University of

Technology, 60 Musk Avenue, Kelvin Grove, Brisbane,

QLD 4059, Australia

e-mail: p.adhikari@qut.edu.au

S. Marasini

Ajou University Neuroscience Postgraduate Program,

Suwon, South Korea

R. P. Shah � S. N. Joshi � J. K. Shrestha

Eye Department, Institute of Medicine, Tribhuvan

University, Kathmandu, Nepal

123

Doc Ophthalmol

DOI 10.1007/s10633-014-9447-9

threatening DR [1]. Despite much research, there is

still no effective treatment for saving vision in the late

stages of DR [2]. The vision of many people with

diabetes could be preserved, at least for a longer period,

if the disease can be caught at early stage and treatment

like better blood glucose control can be initiated.

The visual function in diabetic patients may be

disturbed even without any visible structural damage

to the retina. Electrophysiological tests demonstrate

abnormal results in diabetic patients without any

ophthalmoscopic changes [3–7]. Many studies indi-

cate that neural changes take place in the retina of a

diabetic patient before vascular changes are apparent

and these worsen as DR progresses [3–6]. Studies have

shown that multifocal electroretinogram (mfERG)

implicit time and amplitude are highly sensitive

methods of assessment of local retinal function in

diabetes; implicit time being early indicator of DR

before amplitude is affected [7]. However, few studies

on pattern ERG in diabetic patients report abnormal

ERG responses in the presence of diabetic retinopathy

only [8, 9] and normal or even supernormal amplitude

in diabetic patients with nonproliferative DR [10].

Harrison et al. [11] recently determined that mfERG

implicit time is a good predictor of diabetic retinop-

athy in retinas with no retinopathy. The potential of

mfERG amplitude and implicit time to predict the

retinal areas of future retinopathy provides clinicians

an important clinical tool to screen, follow-up, and

even consider early prophylactic treatment of DR [12]

including protein tyrosine kinase pathway inhibitor

[13], aspirin [14] and better metabolic control [15].

We were interested to measure neuroretinal function in

Nepalese diabetic patients because a very high prevalence

(14.6 % [16], 10.6 % [17]) of diabetes has been reported in

this population in comparison to the global estimate

(6.4 %) [18] owing to the lack of public awareness and

poor medical services. This is the first study in diabetic

population in Nepal using mfERG. It aims to compare

mfERG responses (N1–P1 amplitude and P1-implicit time)

in a diabetic population without DR to healthy control

participants without diabetes and to correlate mfERG

responses with diabetes duration and blood sugar level.

Subjects and methods

The study was conducted over the duration of one year

at a tertiary eye centre in Nepal, and purposive

sampling method was applied to select subjects. The

research adhered to the tenets of the Declaration of

Helsinki and ethical clearance was obtained from the

Research Ethics Committee of Tribhuvan University,

Nepal. Informed consent was obtained from the

subjects after explanation of the nature and possible

consequences of the study.

Subjects with known history of diabetes (type 1 and

type 2) for at least 2 years, but without any DR

according to ETDRS classification, were included in

the study. Information about type and duration of

diabetes was obtained by review of medical records of

the subjects. Fasting blood sugar (FBS) was measured

within 10 days of mfERG testing. Subjects with best

corrected visual acuity (BCVA) worse than 6/9, with

visible media opacity, with history of other ocular

disease or surgery, with high degree of refractive error

(spherical equivalent[6.00 DS of myopia and [5.00

DS of hyperopia) were excluded. Similar criteria were

applied to select age and gender-matched controls

without diabetes. All participants underwent complete

ophthalmic evaluation including visual acuity testing

with Bailey-Lovie Log MAR chart, anterior segment

examination by slit lamp biomicroscopy, refraction,

IOP measurement and dilated fundus examination

before mfERG testing. Bailey-Lovie chart was used in

place of more commonly used Snellen’s chart for more

precise visual acuity measurement. The dilated fundus

examination including the macula examination was

performed with 20 D and 90 D Volk lens by two retina

specialists (JKS and SNJ) who were masked to

mfERG findings.

Multifocal electroretinogram

Recording of the fast flicker mfERG was performed

according to the International Society for Clinical

Electrophysiology of Vision (ISCEV) guidelines

using RETISCAN SYSTEM (Roland Consult Elec-

trophysiological Diagnostic System, Brandenburg,

Germany). Pupils were fully dilated with 1 % tropic-

amide. One gold-cup electrode was used as ground

(attached to the forehead) and two electrodes as

reference (attached to the temple) after cleaning the

skin with abrasive gel. After anaesthetizing the cornea

with topical 4 % xylocaine, recording was performed

with both eyes open by using contact lens electrodes

(Jet electrode; LKC Technologies, Gaithersburg,

Doc Ophthalmol

123

MD), but for the purpose of analysis, the data of the

right eye only were considered. The stimuli were

displayed on a 2100 CRT monitor with the frame

frequency of 75 Hz with RETISCAN software version

07/01 using a black-and-white pattern of 103 hexa-

gons. Each hexagon was temporally modulated

between black (\2 cd/m2) and white (120 cd/m2)

with a contrast of 98 %, according to a pseudorandom

binary m-sequence. The viewing distance was 28 cm

which allowed a viewing angle of 30�. Presbyopic

participants used appropriate correction if necessary

for clear visualization of the fixation target. Occa-

sional artefacts, such as blinks during the recording,

were eliminated by the RETISCAN software and that

part of the sequence was immediately repeated to

record a new response free of artefacts. A standard

m-sequence was recorded by dividing into eight short

segments periods. MfERG N1–P1 amplitude was

measured from the trough of the first negative wave

(N1) to the peak of the first positive wave (P1), and P1-

implicit time was defined as the time at which the peak

of P1 occurred [19].

Analysis of responses was done by regional averages

derived from six concentric rings and four quadrants,

and average response from all 103 hexagonal locations.

The mfERG components were analysed among the

studied cohorts along ring 1 (central hexagon, 5�), ring 2

(5�–10�), ring 3 (10�–15�), ring 4 (15�–20�), ring 5 (20�–

25�) and ring 6 (25�–30�); and superonasal (SN),

inferonasal (IN), superotemporal (ST), and inferotem-

poral (IT) quadrants of retina (Fig. 1).

Statistical analysis

SigmaPlot software package was used for data

analysis. Data were described as mean ± SD and

95 % confidence interval; p \ 0.05 was considered

statistically significant. Mean was used to describe

data as the data distribution was symmetrical. One-

way repeated measures ANOVA (parametric test) was

applied to compute the differences in the mfERG

responses between the diabetic and control subjects as

we assumed that our data were normally distributed.

Post-hoc analysis was done with Holm-Sidak method

for pairwise multiple comparisons of ERG responses

in different rings and quadrants between patients and

controls. The correlation of mfERG amplitude and

implicit time with diabetes duration and blood sugar

level was calculated with Pearson’s Correlation.

Multiple linear regression analysis was used to

determine if age, diabetes duration, and FBS would

predict average N1–P1 amplitude and P1-implicit time

in type 2 diabetes.

Results

Multifocal electroretinogram (mfERG) data of 18 right

eyes of 18 diabetic subjects without retinopathy and 17

right eyes of 17 age and gender-matched healthy

control participants were considered for analysis. Two

diabetic subjects had type 1 diabetes, and the remain-

ing 16 had type 2 diabetes. The mean age of diabetic

Fig. 1 Multifocal electroretinogram (mfERG) plots showing different retinal rings (A) and quadrants (B). The rings are described in

methods. ST superotemporal, IT inferotemporal, IN inferonasal and SN superonasal. The quadrants are demarcated by solid bold lines

Doc Ophthalmol

123

subjects was 49.1 ± 12.6 years and that of normal

controls was 50.3 ± 14.0 years. Out of 16 type 2

diabetic subjects, 50 % (8) were male and 50 % (8)

were female and out of 17 controls, 58.8 % (10) were

male and 41.2 % (7) were female. The age (p = 0.145)

and gender (p = 0.432) were not statistically different

among study groups. In diabetic subjects, the mean

duration of diabetes was 6.3 ± 4.0 years and the mean

fasting blood glucose level was 154.4 ± 31.3 mg/dL

(1 mg/dL = 0.0555 mmol/L).

Multifocal ERG findings

Type 2 diabetic subjects showed reduced average

mfERG N1–P1 amplitude and delayed P1-implicit time

in comparison to controls. The results are discussed in

detail below. The statistics were calculated on type 2

diabetes, and control group only as the sample size

(n = 2) for type 1 diabetes group was very small.

MfERG amplitude in 103 retinal locations for a type 2

diabetic subject and a control subject is shown in Fig. 2.

N1–P1 amplitude

Average mfERG N1–P1 amplitude of all (103) hexagonal

locations was 56.3 ± 17.2 and 47.2 ± 9.3 nv/deg2 in type

1 and type 2 diabetic subjects; and 71.5 ± 12.7 nv/deg2

in controls (Table 1). Average mfERG N1–P1 amplitude

of 103 retinal locations was significantly reduced in type 2

diabetic subjects compared to controls (p\0.001). We

found that N1–P1 amplitude in various rings and quadrants

of retina was significantly different between study groups

(p\0.001). Post hoc analysis revealed that the reduction

in amplitude was statistically significant in rings 1, 2, 3 and

in IT quadrant in type 2 diabetic subjects.

P1-implicit time

Average mfERG P1-implicit time of 103 retinal

locations was 43.0 ± 1.3 and 43.9 ± 2.3 ms in type

1 and type 2 diabetic subjects; and 41.9 ± 2.1 ms in

controls (Table 2). Average P1-implicit time was

significantly delayed in type 2 diabetic subjects

compared to controls (p \ 0.010). Post-hoc analysis

on comparing P1-implicit time in different rings and

quadrants of retina showed that rings 1 and 6; and SN

and IT quadrants only showed statistically significant

delay in P1-implicit time in type 2 diabetic subjects

compared to controls.

Association of N1–P1 amplitude and P1-implicit

time with diabetes duration and blood sugar level

Pearson’s correlation was calculated to find the

association of mfERG amplitude and implicit time

with diabetes duration and blood sugar level in type 2

diabetes group (Table 3).

Fig. 2 Multifocal electroretinogram (mfERG) responses from 103 retinal locations of a diabetic patient (A) and a control subject (B).

Note that amplitudes are greater in the control than in the diabetic subject in most of the locations

Doc Ophthalmol

123

The subjects were divided into two subgroups; first

with diabetes duration \5 years and second with

diabetes duration C5 years (Figs. 3, 4). Average N1–

P1 amplitude of 103 retinal locations showed a trend

of decrease with increasing diabetes duration (r =

-0.587, p = 0.075) (Fig. 3B) in the subjects with

diabetes duration C5 years, but the association was

not statistically significant. N1–P1 amplitude in retinal

rings and quadrants showed similar, but not significant

association with diabetes duration. The subjects with

diabetes duration C5 years showed a significant

positive correlation (r = 0.856, p = 0.002) between

P1-implicit time and diabetes duration (Fig. 3D). P1-

implicit time in retinal rings and quadrants showed a

significant positive correlation with diabetes duration

in ring 3, 4, 5 and 6; and IN quadrant (Table 3). Both

average N1–P1 amplitude and P1-implicit time did not

show any correlation with fasting blood sugar (Fig. 4).

Multiple linear regression showed average N1–P1

amplitude can be predicted by the equations, ‘‘Ampli-

tude = 15.99 ? 0.55 (Age) - 4.21 (Diabetes Dura-

tion) ? 0.10 (FBS)’’ for subjects with diabetes

duration\5 years [R2 = 0.63, F(3,2) = 1.14, p =

0.498] and ‘‘Amplitude = 101.68 - 0.44 (Age) - 1.23

(Diabetes Duration) - 0.12 (FBS)’’ for subjects with

diabetes duration C5 years [R2 = 0.68, F(3,6) = 4.33,

Table 1 Mean N1–P1 amplitudes (nV/deg2) for 6 rings and 4 quadrants of retina

Area of retina Type 1 diabetic subjects Type 2 diabetic subjects Controls p value�

Ring 1 67.7 ± 19.3 87.7 ± 22.7 145.4 ± 47.9 \0.001

Ring 2 67.9 ± 28.9 62.5 ± 17.3 92.3 ± 11.2 \0.001

Ring 3 61.6 ± 33.0 46.9 ± 12.3 67.3 ± 12.3 0.013

Ring 4 56.9 ± 10.7 35.2 ± 7.9 49.8 ± 6.3 0.609

Ring 5 44.9 ± 6.0 26.7 ± 7.1 40.3 ± 7.8 0.835

Ring 6 39.1 ± 4.8 24.1 ± 7.1 33.7 ± 5.7 1.000

SN quadrant 46.8 ± 7.4 29.8 ± 6.9 47.9 ± 5.1 0.121

IN quadrant 45.1 ± 4.9 29.2 ± 8.4 47.3 ± 5.2 0.055

ST quadrant 49.0 ± 14.5 30.9 ± 9.0 47.3 ± 5.4 0.253

IT quadrant 47.8 ± 18.3 28.4 ± 8.9 47.6 ± 5.3 0.036

Average of 103 locations 56.3 ± 17.2 47.2 ± 9.3 71.5 ± 12.7 \0.001

� Between type 2 diabetic subjects and controls

Table 2 Mean P1-implicit times (ms) for 6 rings and four

quadrants of retina

Area of

retina

Type 1

diabetic

subjects

Type 2

diabetic

subjects

Controls p

value�

Ring 1 43.7 ± 3.5 45.7 ± 2.0 43.0 ± 2.5 0.001

Ring 2 42.7 ± 2.1 45.0 ± 2.2 42.0 ± 2.3 0.058

Ring 3 42.7 ± 2.1 44.2 ± 2.1 41.5 ± 1.9 0.986

Ring 4 42.2 ± 1.3 44.9 ± 1.8 41.8 ± 1.8 1.000

Ring 5 43.1 ± 0.0 44.6 ± 1.7 41.5 ± 2.5 0.453

Ring 6 43.6 ± 0.7 45.2 ± 1.4 41.3 ± 2.6 0.001

SN quadrant 43.1 ± 0.0 44.6 ± 1.6 41.5 ± 2.7 0.013

IN quadrant 43.2 ± 1.3 44.7 ± 1.4 41.5 ± 3.2 0.633

ST quadrant 43.6 ± 0.7 44.8 ± 2.0 41.4 ± 2.0 0.248

IT quadrant 43.6 ± 0.7 44.7 ± 2.0 41.4 ± 3.2 0.002

Average

of 103

locations

43.0 ± 1.3 43.9 ± 2.3 41.9 ± 2.1 0.010

� Between type 2 diabetic subjects and controls

Table 3 Correlation of P1-implicit time with diabetes dura-

tion in type 2 diabetic subjects

Areas of retina Correlation (r) p value

Ring 1 0.210 0.435

Ring 2 0.113 0.676

Ring 3 0.538 0.032

Ring 4 0.396 \0.001

Ring 5 0.417 \0.001

Ring 6 0.355 \0.001

SN quadrant 0.318 0.230

IN quadrant 0.507 0.045

ST quadrant 0.402 0.123

IT quadrant 0.422 0.104

Average of 103 locations 0.408 0.117

Doc Ophthalmol

123

p = 0.060]. Average P1-implicit time can be predicted

by the equations, ‘‘Implicit Time = 43.88 ? 0.002

(Age) ? 1.43 (Diabetes Duration) - 0.02 (FBS)’’ for

subjects with diabetes duration\5 years [R2 = 0.47,

F(3,2) = 0.609, p = 0.670] and ‘‘Implicit Time =

43.70 ? 0.02 (Age) ? 0.44 (Diabetes Duration) - 0.02

(FBS)’’ for subjects with diabetes duration C5 years

[R2 = 0.91, F(3,6) = 19.58, p = 0.002]. The results of

regression indicated that the effect of the variables was

significant only on implicit time for subjects with

diabetes duration C5 years and among the variables,

diabetes duration and FBS had significant partial

effects on the full model.

Discussion

People with diabetes are at increased risk of eye

complications, and most people with diabetes will get

some form of retinopathy. Considering the huge

strides made in the treatment of diabetic retinopathy,

the earlier the retinopathy is diagnosed, the less sight

threatening damage to the eye may occur.

Multifocal electroretinogram (mfERG) has been

shown to be capable of detecting the neural changes

that occur in retina of diabetic patients before any

vascular changes are visible. The results of the present

study demonstrate on average reduced fast flicker

mfERG N1–P1 amplitude and delayed P1-implicit time

in type 2 diabetic patients without diabetic retinopathy

in comparison to normal control participants.

The results of the present study are consistent with

studies by Fortune et al. [7], Han et al. [20] and Bearse

et al. [12], and this study also presents novel findings

regarding reduced mfERG N1–P1 amplitudes in

diabetes. In the studies by Fortune et al. [7], Han

et al. [20], and Bearse et al. [12], it was shown that

implicit time is more affected in diabetes patients and

is the better predictor of future retinopathy than the

amplitude. Han et al. showed that up to 50 % of

implicit times were delayed significantly in mfERG of

diabetic subjects without retinopathy, and Schneck

et al. [21] also found that 17 % of implicit times were

abnormal in retinal areas that did not have retinopathy.

In contrary to these studies, we found that both mfERG

amplitude and implicit time are affected in type 2

Fig. 3 Correlation of average N1–P1 amplitude and average

P1-implicit time with diabetes duration in type 2 diabetic

subjects. Panel A and B show the correlation of amplitude with

diabetes duration in subjects with diabetes duration \5 years

(triangles) and C5 years (squares). Panel C and D show the

correlation of implicit time with diabetes duration in subjects

with diabetes duration \5 years (triangles) and C5 years

(squares)

Fig. 4 Correlation of average N1–P1 amplitude and average

P1-implicit time with fasting blood sugar (FBS) in type 2

diabetic subjects. Panel A and B show the correlation of

amplitude with FBS in subjects with diabetes duration \5 years

(triangles) and C5 years (squares). Panel C and D show the

correlation of implicit time with FBS in subjects with diabetes

duration \5 years (triangles) and C5 years (squares)

Doc Ophthalmol

123

diabetic patients. Though it has been mentioned in the

literature that high interindividual variability of

mfERG N1–P1 amplitude (standard deviation, SD

ranging from 57 to 98 nV/deg2 in normal eyes) makes

this parameter a poorer predictor of diabetic retinop-

athy than P1-implicit time [7], our study population

had quite less interindividual variability of amplitudes

in diabetic (SD, 9.3 nv/deg2) as well as control

subjects (SD, 12.7 nv/deg2) in comparison to previous

studies.

The explanation given for minimally reduced

amplitude, but significantly delayed implicit time in

the previous studies is that the function of the

photoreceptors and the inner retina including bipolar

cells, which are the primary generators of N1–P1

amplitudes, may be impaired in diabetic patients

without retinopathy, but it is not completely abolished

[22, 23]. Reduced amplitude in our study population

indicate that this population, that has greater preva-

lence of diabetes and uncontrolled blood sugar level,

might have marked dysfunction of the photoreceptors

and bipolar cells before the onset of any clinical

retinopathy. It was interesting to observe that N1–P1

amplitude was reduced in ring 1 in both diabetic

groups indicating even the macula is affected before

any clinical retinopathy occurs. It can be inferred that

though visual acuity was normal in our diabetic

subjects, the quality of vision might have been

affected and the measurement of contrast sensitivity

and colour vision can be important to know macular

function in diabetic subjects. mfERG implicit time has

been reported to increase in type 1 diabetes without

clinical retinopathy, amplitude being unaffected [24,

25]. We found that both amplitude and implicit time

are normal in type 1 diabetes with no clinical

retinopathy. Moreover, no notable difference in

mfERG responses was observed between two diabetic

groups. These results are not conclusive because of

very small sample size in type 1 diabetic group.

There are variable results on association of diabetes

duration and blood sugar level with ERG response.

Shimada et al. [26] who used global flash stimulus

rather than multifocal demonstrated a marginal pro-

pensity for ERG amplitudes to be associated with

longer histories of diabetes. However, Kim et al. [27]

showed that the duration of diabetes and glycemic

control status did not correlate with the local mfERG

responses. In our study, neither average N1–P1

amplitude nor average P1-implicit time showed any

significant correlation with fasting blood sugar level.

Average N1–P1 amplitude was not correlated with

diabetes duration, but average P1-implicit time

showed positive correlation with diabetes duration in

type 2 diabetic subjects with diabetes duration

C5 years, suggesting that P1-implicit time is delayed

with increasing diabetes duration, 5 years after a

person acquires diabetes. Moreover, P1-implicit time

was positively correlated with diabetes duration in

three rings and one quadrant of retina.

The study had few limitations. Visual acuity was

measured with Bailey-Lovie Log MAR chart, but not

with ETDRS chart due to unavailability of the chart in

the research site. However, ETDRS chart is known to

give better visual acuity score than Bailey-Lovie Log

MAR chart [28] and thus our inclusion criteria were

not affected due to the choice of the chart as the visual

acuity was better than 6/9 in all subjects. HbA1c could

have been a more accurate measure than FBS to assess

overall diabetes control of diabetic subjects over a

long duration, but we measured FBS as we wanted to

know the blood sugar level close to ERG testing time.

This is probably the reason for the lack of correlation

between P1-implicit time and blood sugar level. The

information of diabetes duration might not have been

fully reliable as it was obtained from the medical

records of patients and this could have affected our

results to some extent. Though adult male type 2

diabetic subjects without retinopathy show more

abnormal mfERG responses than their female coun-

terparts [29], we could not do the separate analysis for

males and females in our study due to small sample

size.

Conclusion

We demonstrate reduced mfERG N1–P1 amplitude

and delayed P1-implicit time in Nepalese diabetic

patients without retinopathy. Diabetes duration and

fasting blood glucose have significant influence on

implicit time, but not on amplitude. Our findings may

be population specific, and further longitudinal studies

are needed to determine whether those patients

develop more severe forms of diabetic retinopathy.

Acknowledgments The authors would like to thank Beatrix

Feigl for comments on a manuscript draft.

Doc Ophthalmol

123

References

1. Yau JW, Rogers SL, Kawasaki R, Lamoureux EL, Kowalski

JW, Bek T, Chen S-J, Dekker JM, Fletcher A, Grauslund J

(2012) Global prevalence and major risk factors of diabetic

retinopathy. Diabetes Care 35:556–564

2. Ferris F (1996) Early photocoagulation in subjects with

either type I or type II diabetes. Trans Am Ophthalmol Soc

94:505

3. Ghirlanda G, Di Leo MA, Caputo S, Cercone S, Greco AV

(1997) From functional to microvascular abnormalities in

early diabetic retinopathy. Diabetes Metab Rev 13:15–35

4. Shirao Y, Kawasaki K (1998) Electrical responses from

diabetic retina. Prog Retin Eye Res 17:59

5. Tzekov R, Arden G (1999) The electroretinogram in dia-

betic retinopathy. Surv Ophthalmol 44:53–60

6. Barber AJ (2003) A new view of diabetic retinopathy: a

neurodegenerative disease of the eye. Prog Neuropsycho-

pharmal 27:283–290

7. Fortune B, Schneck ME, Adams AJ (1999) Multifocal

electroretinogram delays reveal local retinal dysfunction in

early diabetic retinopathy. Invest Ophthalmol Vis Sci

40:2638–2651

8. Arden GB, Hamilton AM, Wilson-Holt J, Ryan S, Yudkin

JS, Kurtz A (1986) Pattern electroretinogram become

abnormal when background diabetic retinopathy deterio-

rates to a preproliferative stage: possible use as a screening

test. Br J Ophthalmol 70:330–335

9. Nesher R, Trick GL (1991) The pattern electroretinogram in

retinal and optic nerve disease: a quantitative comparison of

the pattern of visual dysfunction. Doc Ophthalmol

77:225–235

10. Jenkins TC, Cartwright JP (1990) The electroretinogram in

minimal diabetic retinopathy. Br J Ophthalmol 74:681–684

11. Harrison WW, Bearse MA, Ng JS, Jewell NP, Barez S,

Burger D, Schneck ME, Adams AJ (2011) Multifocal

electroretinograms predict onset of diabetic retinopathy in

adult subjects with diabetes. Invest Ophthalmol Vis Sci

52:772–777

12. Bearse MA Jr, Adams AJ, Han Y, Schneck ME, Ng J,

Bronson-Castain K, Barez S (2006) A multifocal electro-

retinogram model predicting the development of diabetic

retinopathy. Prog Retin Eye Res 25:425

13. De Juan Jr E (July 6, 1999) Use of a protein tyrosine kinase

pathway inhibitor in the treatment of diabetic retinopathy.

U.S. Patent 5,919,813

14. Colwell JA (1997) Aspirin therapy in diabetes. Diabetes

Care 20:1767–1771

15. Engerman R, Bloodworth J, Nelson S (1977) Relationship

of microvascular disease in diabetes to metabolic control.

Diabetes 26:760–769

16. Singh D, Bhattarai M (2003) High prevalence of diabetes

and impaired fasting glycaemia in urban Nepal. Diabet Med

20:170–171

17. Shrestha U, Singh D, Bhattarai M (2006) The prevalence of

hypertension and diabetes defined by fasting and 2-h plasma

glucose criteria in urban Nepal. Diabet Med 23:1130–1135

18. Shaw J, Sicree R, Zimmet P (2010) Global estimates of the

prevalence of diabetes for 2010 and 2030. Diabetes Res Clin

Pr 87:4–14

19. Hood DC, Bach M, Brigell M, Keating D, Kondo M, Lyons

JS, Marmor MF, McCulloch DL, Palmowski-Wolfe AM

(2012) ISCEV standard for clinical multifocal electroreti-

nography (mfERG) (2011 edition). Doc Ophthalmol

124:1–13

20. Han Y, Bearse M, Schneck M, Barez S, Jacobsen C, Adams

A (2004) Towards optimal filtering of ‘‘standard’’ multifo-

cal electroretinogram (mfERG) recordings: findings in

normal and diabetic subjects. Br J Ophthalmol 88:543–550

21. Schneck M, Bearse M, Han Y, Barez S, Jacobsen C, Adams

A (2004) Comparison of mfERG waveform components

and implicit time measurement techniques for detecting

functional change in early diabetic eye disease. Doc Oph-

thalmol 108:223–230

22. Hood DC, Frishman LJ, Saszik S, Viswanathan S (2002)

Retinal origins of the primate multifocal ERG: implications

for the human response. Invest Ophthalmol Vis Sci

43:1673–1685

23. Hare WA, Ton H (2002) Effects of APB, PDA, and TTX on

ERG responses recorded using both multifocal and con-

ventional methods in monkey. Doc Ophthalmol

105:189–222

24. Laron M, Bearse MA, Bronson-Castain K, Jonasdottir S,

King-Hooper B, Barez S, Schneck ME, Adams AJ (2012)

Association between local neuroretinal function and control

of adolescent type 1 diabetes. Invest Ophthalmol Vis Sci

53:7071–7076

25. Kurtenbach A, Langrova H, Zrenner E (2000) Multifocal

oscillatory potentials in type 1 diabetes without retinopathy.

Invest Ophthalmol Vis Sci 41:3234–3241

26. Shimada Y, Li Y, Bearse MA, Sutter EE, Fung W (2001)

Assessment of early retinal changes in diabetes using a new

multifocal ERG protocol. Br J Ophthalmol 85:414–419

27. Kim SJ, Song SJ, Yu HG (2007) Multifocal electroretino-

gram responses of the clinically normal retinal areas in

diabetes. Ophthalmol Res 39:282–288

28. Kaiser PK (2009) Prospective evaluation of visual acuity

assessment: a comparison of Snellen versus ETDRS charts

in clinical practice (An AOS Thesis). Trans Am Ophthalmol

Soc 107:311

29. Ozawa GY, Bearse MA, Bronson-Castain KW, Harrison

WW, Schneck ME, Barez S, Adams AJ (2012) Neurode-

generative differences in the retinas of male and female

patients with type 2 diabetes. Invest Ophthalmol Vis Sci

53:3040–3046

Doc Ophthalmol

123