вход по аккаунту


Early factors associated with axonal loss after optic neuritis.

код для вставкиСкачать
Early Factors Associated with Axonal Loss
after Optic Neuritis
Andrew P. D. Henderson, FRACP,1 Daniel R. Altmann, DPhil,1,2 S. Anand Trip, PhD,1
Katherine A. Miszkiel, FRCR,3 Patricio G. Schlottmann, MD,4 Steve J. Jones, PhD,5
David F. Garway-Heath, FRCOphth,4 Gordon T. Plant, MD,6
and David H. Miller, FMedSci1
Objective: Acute optic neuritis due to an inflammatory demyelinating lesion of the optic nerve is often seen in
association with multiple sclerosis. Although functional recovery usually follows the acute episode of visual loss,
persistent visual deficits are common and are probably due to axonal loss. The mechanisms of axonal loss and early
features that predict it are not well defined. We investigated clinical, electrophysiological, and imaging measures at
presentation and after 3 months as potential markers of axonal loss following optic neuritis.
Methods: We followed 21 patients after their first attack of acute unilateral optic neuritis for up to 18 months.
Axonal loss was inferred from optical coherence tomography measures of retinal nerve fiber layer (RNFL) thickness at
least 6 months following the episode. Visual function, visual evoked potential, and optic nerve magnetic resonance
imaging measures obtained during the acute episode and 3 months later were investigated for their association with
later axonal loss.
Results: After multivariate analysis, prolonged visual evoked potential latency and impaired color vision, at baseline
and after 3 months, were significantly and independently associated with RNFL thinning. Low-contrast acuity
measures exhibited significant univariate associations with RNFL thinning.
Interpretation: The association of RNFL loss with a prolonged visual evoked potential (VEP) latency suggests that
acute and persistent demyelination is associated with increased vulnerability of axons. VEP latency and visual
function tests that capture optic nerve function, such as color and contrast, may help identify subjects with a higher
risk for axonal loss who are thus more suitable for experimental neuroprotection trials.
ANN NEUROL 2011;70:955–963
cute optic neuritis due to an inflammatory demyelinating lesion of the optic nerve is often seen in association with multiple sclerosis (MS). Although functional recovery usually follows the acute episode of visual loss,
persistent visual deficits are common. In vivo measurements
of thinning of the retinal nerve fiber layer (RNFL) suggest
that the extent of axonal loss is associated with the degree
of persistent visual dysfunction following optic neuritis.1
The factors that contribute to axonal loss following optic
neuritis are of significant interest not only because of the
important contribution that axonal loss makes to visual deficits following optic neuritis per se but also, through the
effects of relapses of MS in general, to disability in MS.
The mechanisms of axonal loss following a relapse and
early features that predict it are not well defined.
Axonal loss following optic neuritis has been
inferred in vivo in several ways. Defects in the RNFL have
been observed on fundoscopy following optic neuritis,2,3
and atrophy of the optic nerve has been measured using
magnetic resonance imaging (MRI).4,5 More directly, thinning of the RNFL has been documented with optical coherence tomography (OCT).1,6 Quantitative OCT measurements of the RNFL thickness have the advantage over
optic nerve cross-sectional area measures that they are not
subject to confounding by loss of myelin. Despite the generally good visual prognosis following optic neuritis, a
decrease in the thickness of the RNFL of 15 to 20% on
average compared to the unaffected fellow eye is seen following a single episode of unilateral optic neuritis.6–8
View this article online at DOI: 10.1002/ana.22554
Received Mar 16, 2011, and in revised form Jun 1, 2011. Accepted for publication Jul 27, 2011.
Address correspondence to Dr Henderson, NMR Research Unit, University College London, Institute of Neurology, Queen Square, London WC1N 3BG,
United Kingdom. E-mail:
From the 1NMR Research Unit, University College London, Institute of Neurology; 2Medical Statistics Department, London School of Hygiene and Tropical
Medicine; 3Department of Neuroradiology, National Hospital for Neurology and Neurosurgery; 4Glaucoma Research Unit, Moorfields Eye Hospital; and
Departments of 5Neurophysiology and 6Neurology, National Hospital for Neurology and Neurosurgery, London, United Kingdom.
C 2011 American Neurological Association
of Neurology
The availability of noninvasive OCT-measured
RNFL thickness to measure axonal loss provides a unique
potential for in vivo investigation of factors associated
with axonal loss resulting from the sentinel optic nerve
lesion in optic neuritis. It is also possible to acquire
detailed quantitative measures of visual impairment, to
study nerve conduction through the lesion using visual
evoked potentials (VEPs), and to investigate structural
abnormalities within the symptomatic lesion using optic
nerve MRI techniques. The identification of features during the earlier stages of acute optic neuritis that are associated with the extent of subsequent axonal loss would aid
subject selection for trials of neuroprotection and may give
insights into the mechanisms of axonal loss and potential
strategies for its prevention. We therefore investigated the
associations of early visual, VEP, and optic nerve MRI
measures with final axonal loss (measured using OCT) in
a cohort of patients who were followed serially from presentation with acute unilateral optic neuritis for up to 18
months. The serial OCT-measured RNFL findings of the
cohort have been previously reported.8
Patients and Methods
Clinical Methods
Patients were recruited from the Medical Eye Clinic at Moorfields Eye Hospital. All patients presented with their first ever
episode of unilateral optic neuritis, and no patient experienced
a recurrence during the study period. Patients were studied as
soon as possible after onset of symptoms (median days elapsed,
17; range, 10–32 days), and again after 3 months (median, 95
days; range, 83–136 days), 6, 12, and 18 months. No patient
had any other neurological or ophthalmological disease at presentation. At each study point, all subjects had the following
Clinical Visual Testing
As previously described,8 visual acuity was measured with Early
Treatment of Diabetes Retinopathy Study acuity charts, and
expressed as the 4m logarithm of the minimum angle of resolution (LogMAR). Sloan 25%, 5%, and 1.25% contrast charts
were used to measure low-contrast visual acuity, with acuity
expressed as the LogMAR equivalent. If patients were unable to
perceive any letters on either the logMAR or Sloan charts at
4m, the chart was moved to a distance of 1m, the test was
repeated, and an adjustment for the change in distance was
made according to the method proposed in the paper by Ferris
and colleagues.9 Color vision was measured using the Farnsworth-Munsell 100-Hue test,10,11 and the result was expressed
as the square root of the total error score, as this is normally
distributed in normal subjects.12 Visual sensitivity in the central
30 of vision was measured using a Humphrey visual field analyzer (Carl Zeiss Meditec, Dublin, CA), with correction for refractive errors using wide-angle lenses where needed.
RNFL Imaging
As previously described,8 circumpapillary images of the RNFL
were acquired using a Stratus OCT device (Carl Zeiss Meditec,
Dublin, CA), using the Fast RNFL scan. The images were segmented automatically to give the RNFL thickness.
All MRI sequences were performed on a Signa 1.5T imager
(General Electric, Milwaukee, WI), and analyses were performed using methods described below.
OPTIC NERVE MEAN AREA. Optic nerve mean area was
measured using the method of Hickman and colleagues. The
patients’ optic nerves were imaged with the coronal-oblique
short-echo fast fluid-attenuated inversion recovery sequence
described by Hickman and colleagues4,13: repetition time (TR)
¼ 2,740 milliseconds, echo time (TE) ¼ 16 milliseconds,
inversion time ¼ 1,072 milliseconds, number of excitations ¼
6, echo train length ¼ 6, matrix size ¼ 512 384, field of
view ¼ 24 18cm, in-plane resolution ¼ 0.47 0.47mm,
interleaved contiguous slices ¼ 16 3 mm, acquisition time ¼
13.5 minutes. Mean cross-sectional area was calculated by a
blinded observer using the methods described previously.4
RATIO. Whole optic nerve magnetization transfer ratio
(MTR), and lesion and nonlesion MTR were measured using
the methods described by Hickman and colleagues.14 A
3-dimensional sequence was acquired with coronal 60 1.5mm thick slices, in-plane resolution ¼ 0.75 0.75mm, TR
¼ 23.1 milliseconds, TE ¼ 5.6 milliseconds, excitations ¼ 2,
flip angle ¼ 12 , matrix size ¼ 256 192, field of view ¼
19 14.25cm, acquisition time ¼ 18 minutes, both with and
without a prepulse to saturate the broad resonance of immobile
macromolecular protons (offset frequency ¼ 2kHz, equivalent
on-resonance flip angle ¼ 500 ). Segmentation of the optic
nerves and calculation of MTR were performed by a blinded
observer as described by Hickman and colleagues.14
Sequences as previously described were used to detect the T24
and contrast-enhanced15 optic nerve lesion and measure its
extent and location. The lesions characteristics were reported by
a neuroradiologist who was not aware of the side of the lesion
or other clinical features. The sequence details were:
• A fat-saturated dual echo fast spin echo proton density/T2-
weighted sequence: TR ¼ 2,300 milliseconds, TEeffective ¼
58/145 milliseconds, echo-train length ¼ 8, excitations ¼ 2,
field of view ¼ 24 18cm, matrix size ¼ 512 384, inplane resolution ¼ 0.5 0.5mm, interleaved contiguous coronal oblique slices ¼ 16 3mm, acquisition time ¼ 11
• Gadolinium-enhanced (0.3mmol/kg dimeglumine gadopentate) fat-saturated T1-weighted spin echo sequence: TR ¼
600 milliseconds, TE ¼ 20 milliseconds, excitation ¼ 1,
matrix ¼ 256 3 192, field of view ¼ 24 3 18cm,
in-plane resolution ¼ 0.94 3 0.94mm, interleaved
Volume 70, No. 6
Henderson et al: Axonal Loss after Optic Neuritis
contiguous coronal oblique slices ¼ 16 3 3mm, acquisition time ¼ 3 minutes.15
The length of the lesion (in millimeters) was determined
as the number of slices involved multiplied by 3. The anatomical position of involvement (whether intraorbital, intracanalicular, intracranial, chiasmatic, or a combination of these) was
noted, and this was used to generate binary variables for statistical analysis. The globe–lesion distance (in millimeters) was calculated by multiplying the number of slices between the globe
and the anterior edge of the lesion by 3.
VEP latency and amplitude were recorded using the method
described by Brusa and colleagues.16 Briefly, monocular stimuli
using reversal of achromatic checks with a contrast of 93% and
an individual check angle of 40 at the eye were presented in a
darkened room. The whole field stimulus measured 28 20 ,
and the central field stimulus measured 4 . Recordings were
made using a symmetric lateral chain of 5 electrodes, 5cm
above the inion, with a referential electrode at Fz (10–20 electrode system). The electrode 5cm above the inion was used for
calculating P100 amplitude and latency.
The main analyses aimed to identify baseline (and secondarily
3-month) associations with eventual RNFL loss. As there was
negligible additional loss of RNFL thickness after 6 months,8
we used a minimum of 6 months as an estimate of final RNFL
loss. RNFL loss was estimated by subtracting the last recorded
RNFL thickness in the affected eye (in all subjects at least 6
months from onset of symptoms) from the baseline fellow eye
RNFL thickness. Multiple regressions were carried out with this
RNFL loss outcome on the potential baseline associations in 3
stages, to minimize the number of variables in any single model.
First, potential baseline associations were entered singly. Second,
significant baseline associations from this first stage were entered
together in 3 separate models, for MRI, VEP, and visual measure
associations, respectively. Third, any remaining significant associations from these 3 models were entered together into 1 model,
with nonsignificant variables then removed.
These stages were repeated for 3-month associations, and
remaining significant 3-month associations were entered into
the baseline model to assess the relative performance of baseline
and 3-month associations.
Baseline fellow eye RNFL thickness was included as a
covariate in all these regression models (including first-stage single potential associations) for 2 reasons:
• Regressions of RNFL loss as defined above on any associations contain this additional covariate implicitly but in constrained form, because such models are mathematically
equivalent to regressing the last recorded RNFL thickness in
the affected eye on these same variables with baseline fellow
eye RNFL thickness as an additional covariate but with its
coefficient constrained to take the value 1; this is generally
December 2011
an implausible assumption, and including the baseline fellow
covariate explicitly releases this constraint.
• Entering this covariate also tends to reduce between-subject
residual variability and thus increase precision of all regression coefficients. P values for univariate associations reported
in Table 1 were therefore from models also including this
All statistical analysis was conducted in Stata 9.2 (StataCorp, College Station TX).
We have previously reported the relationship between
RNFL loss in the first 3 months and later RNFL loss and the
lack of relationship between early RNFL swelling and later
RNFL loss.8 The early RNFL measures are not included in our
present analysis, which aimed to identify whether there were
clinical (visual), electrophysiological (VEP), or optic nerve
structural (MRI) measures that were associated with axonal loss.
Twenty-one patients were recruited (mean 6 standard
deviation [SD] age, 31 6 6 years; 15 females). Seventeen
patients had their final examination at 18 months, 3 at 12
months, and 1 at 6 months. Baseline visual function and
visual evoked potential data are presented in Table 2. The
median (range) LogMAR acuity at onset was 0.12 (0.1
to 1.7), and the median LogMAR acuity at the final examination was 0.02 (0.10 to 0.22). During the study, 4
patients experienced a further clinical episode of demyelination (all in the spinal cord), and fulfilled the revised
McDonald criteria for the diagnosis of MS.17 Eight further
patients developed new MRI lesions, fulfilling the revised
McDonald criteria for the diagnosis of MS. Of the remainder, 8 had MRI lesions consistent with inflammatory demyelination, but did not fulfill the McDonald criteria during the study. One patient had no brain or spine lesions.
Serial RNFL measures of this cohort have been reported
previously.8 The presently reported cohort does not include
2 patients (reported in the previous paper) who did not
have RNFL thickness measured at 6 months or later.
RNFL Thickness
As reported previously,8 affected RNFL thickness
declined from initial swelling at baseline to reduced
thickness at the final measure. At the final study point,
the mean 6 SD RNFL in the affected eye was 83.4 6
16.9lm, which was reduced (p < 0.001, paired t test)
compared with the baseline fellow eye value (102.7 6
9.6lm). Swelling of the RNFL was evident (p ¼ 0.006,
paired t test) at the first study point (mean 6 SD
affected eye RNFL ¼ 132.8 6 51.0lm vs unaffected eye
RNFL ¼ 102.7 69.8lm). The degree of this swelling
was not related to later RNFL loss.8
of Neurology
TABLE 1: Univariate Regression Characteristics of Candidate Predictors of Final Retinal Nerve Fiber Layer Loss
Baseline Measures
3-Month Measures
T2 lesion length
Gd-positive lesion length
Globe–lesion distance (T2 weighted)
Globe–lesion distance (Gd enhanced)
Optic nerve mean area
Presence of intracanalicular lesion
Presence of intracranial lesion
Presence of chiasmal lesion
Lesion MTR
Whole nerve MTR
Nonlesional MTR
MRI measures
VEP measures
Whole field P100 latency
Whole field P100 amplitude
Central field VEP latency
Central field VEP amplitude
LogMAR acuity
Sloan 25% contrast acuity
Sloan 5% contrast acuity
Sloan 1.25% contrast acuity
Visual function measures
Central 30 visual field mean deviation
FM-100 Hue test
Result shown includes 3 patients with no recordable VEP (ie, zero amplitude) at baseline. If these patients are excluded, n ¼ 18
and p ¼ 0.053.
Square root of total error score.
Gd ¼ gadolinium; LogMAR ¼ logarithm of the minimum angle of resolution; MRI ¼ magnetic resonance imaging; MTR ¼
magnetization transfer ratio; VEP ¼ visual evoked potential.
Predictors of RNFL Loss
The univariate analysis of potential associations with final
RNFL loss are shown in Table 1.
VISUAL ASSESSMENTS. At baseline, FM-100 Hue
color vision score, visual field mean sensitivity, full contrast LogMAR visual acuity (as previously reported8),
Sloan 25% contrast acuity, and Sloan 5% contrast visual
acuity had significant univariate associations with final
RNFL thickness. At 3 months, FM-100 Hue color vision
score, visual field mean sensitivity, Sloan 25% contrast,
Sloan 5% contrast, and Sloan 1.25% contrast had significant univariate associations with final RNFL thickness.
In all cases, poorer early visual function was associated
with greater final RNFL loss.
Multivariate regression of the significant baseline
visual function associations identified only color vision as
having an independent association with RNFL loss.
OPTIC NERVE MRI MEASURES. Only the length of
the optic nerve lesion, as determined by the gadoliniumenhanced images obtained at baseline, was positively
Volume 70, No. 6
Henderson et al: Axonal Loss after Optic Neuritis
TABLE 2: Baseline Visual Function and Visual Evoked Potential Measures of the Patient Cohort
Affected Eye
Fellow Eye
Mean (SD)
0.57 (0.75)
0.07 (0.06)
Median [range]
0.12 [0.1 to 1.7]
0.08 [0.18 to 0.40]
Mean (SD)
0.64 (0.73)
0.00 (0.06)
Median [range]
0.20 [0.04 to 1.7]
0.00 [0.10 to 0.20]
Mean (SD)
0.92 (0.66)
0.22 (0.10)
Median [range]
0.61 [0.16 to 1.7]
0.20 [0.04 to 0.42]
1.31 (0.44)
0.70 (0.32)
1.60 [0.52 to 1.7]
0.62 [0.42 to 1.7]
Mean (SD)
19.7 (11.2)
12.3 (15.8)
Median [range]
13.6 [7.8 to 36.6]
8.3 [4.5 to 17.9]
Mean (SD)
13.8 (12.1)
3.2 (3.0)
Median [range]
7.4 [33.5 to 0.0]
2.4 [14.3 to 0.0]
Mean (SD)
115.6 (15.0)
95.3 (4.4)
Median [range]
117.7 [92.8 to 140.0]
96.2 [88.0 to 101.6]
Mean (SD)
6.5 (5.5)
11.2 (6.5)
Median [range]
5.3 [0.0 to 20.4]
8.8 [4.0 to 28.8]
LogMAR visual acuity
Sloan 25% contrast visual acuity
Sloan 5% contrast visual acuity
Sloan 1.25% contrast visual acuity
Mean (SD)
Median [range]
Farnsworth-Munsell 100 Hue test
Visual field mean deviation, dB
Visual evoked potential latency, ms
Visual evoked potential amplitude, mV
Square root of total error score.
LogMAR ¼ logarithm of the minimum angle of resolution; SD ¼ standard deviation.
associated with final RNFL loss (p ¼ 0.019) (see Table
1). A lower baseline lesion MTR was borderline significantly (p ¼ 0.059) associated with greater RNFL loss.
VEPS. The latency of the P100 component of the
whole field VEP was significantly associated with final
RNFL loss when measured at baseline (p ¼ 0.007) and
at 3 months (p ¼ 0.007). Decreased amplitude of the
P100 component of the whole field VEP was associated
with increased final RNFL loss when measured at baseline (p ¼ 0.012), and was on the borderline of significance at 3 months (p ¼ 0.050). Increased latency of the
central field VEP at 3 months was associated with final
RNFL loss (p ¼ 0.001). Decreased amplitude of the central field VEP was associated with greater RNFL loss
December 2011
when measured at baseline (p ¼ 0.003) and at 3 months
(p ¼ 0.030).
In the multivariate analysis of VEP variables, both
baseline and 3-month whole field VEP latencies were independently associated with final RNFL loss, whereas
whole field VEP amplitude was not at either time point,
nor was either amplitude or latency of the central field
VEP. VEP latency was significantly (p ¼ 0.009) longer at
3 months if final RNFL loss was >20% of the baseline
fellow RNFL thickness (Table 3).
To determine whether shortening of whole field
VEP latency was associated with a reduction in final
RNFL loss, we regressed RNFL loss against the change
in VEP latency in the first 3 months, and found a significant relationship, with a reduction in VEP latency of 1
of Neurology
TABLE 3: Relationship of VEP Amplitude and Latency to RNFL Loss >20% of Fellow Eye Value
Mean (SD) VEP Latency, ms
Mean (SD) VEP Amplitude, Mv
3 Months
3 Months
>20% RNFL loss
123.0 (17.2), n ¼ 6
131.4 (19.8), n ¼ 7
3.5 (4.6), n ¼ 6
6.0 (3.5), n ¼ 7
<20% RNFL loss
111.9 (13.1), n ¼ 12,
p ¼ 0.146a
107.2 (13.5), n ¼ 11,
p ¼ 0.009a
8.8 (5.2), n ¼ 12,
p ¼ 0.027a
10.0 (6.1), n ¼ 11,
p ¼ 0.143a
Comparison made by t test.RNFL ¼ retinal nerve fiber layer; SD ¼ standard deviation; VEP ¼ visual evoked potential.
millisecond over the first 3 months being associated with
a 0.37lm decrease in final RNFL loss (p ¼ 0.001; 95%
confidence interval [CI], 0.16–0.59; R2 ¼ 0.14).
Overall Model for Final RNFL Value
The only overall independent baseline associations with
final RNFL loss were color vision and whole field VEP
latency. In a model with these 2 variables and baseline
fellow eye RNFL as covariate, a 1U increase in the
square root of the total error score for color vision predicted an RNFL loss greater by 0.81lm (95% CI, 0.40–
1.21; p ¼ 0.001), and a 1-millisecond increase in whole
field VEP latency was associated with an RNFL loss
greater by 0.50lm (95% CI, 0.21–0.80; p ¼ 0.003),
with 77% of variance in the loss explained by the model.
These variables also were independently associated with
final RNFL when measured at 3 months (p < 0.001 and
0.003 for color vision and VEP latency, respectively),
with 83% of variance explained. However, the increase in
variance explained by adding the 3-month variables to
the baseline model was not a significant improvement
(p ¼ 0.22) over the baseline model.
Although our cohort was recruited in an unselected manner with regard to their degree of visual dysfunction, it
appears to be milder than other large cohorts in some
but not all respects; for example, baseline visual field
mean deviation was milder (13.4dB vs 23.0dB compared with the Optic Neuritis Treatment Trial cohort18),
but mean VEP interocular latency difference was similar
(20 milliseconds vs 23 milliseconds) to the cohort studied by Brusa et al.16 The longer time between symptom
onset and baseline investigation in our study compared
to the Optic Neuritis Treatment Trial (mean 18 6 6
days vs mean 5 6 1.6 days) might contribute to the
milder initial visual dysfunction observed. It may also
have enabled more recordable VEP and color vision
responses at the first study time point. Overall, we feel
that the clinical features and course of our cohort falls
within the usual spectrum of unilateral optic neuritis due
to inflammatory demyelination.
Although multiple measures had a univariate association with final RNFL loss, in this cohort only color vision
and the latency of the VEP—at baseline and after 3
months—were independently associated with final RNFL
loss. The results of a multivariate model suggest that color
vision and VEP P100 latency are associated with final
RNFL loss independently of each other, suggesting that
they may reflect different prognostic and/or pathogenic
elements of the inflammatory lesion in optic neuritis.
A previously reported pathological study by Evangelou and colleagues of the anterior visual pathway of
patients with MS19 found that the parvocellular layers of
the lateral geniculate nucleus (which are responsible for
color vision) and the corresponding axons in the optic
nerve were preferentially affected over the magnocellular
layers (which are responsible for perception of luminance) and their corresponding optic nerve axons.20 The
authors suggested that neurons with small axons (from
parvocellular layers) are selectively vulnerable to damage
after inflammatory demyelination. Our results, demonstrating a strong relationship between color vision (which
is transmitted through ganglion cells with small-diameter
axons), and final axonal loss in the RNFL are consistent
with this idea. However, there is no direct evidence that
functional deficits in small fibers at onset have any relationship with the fibers that actually degenerate, and
some authors have proposed that color vision deficit is
due to loss of the fibers in the optic nerve that correspond to the fovea and immediate surrounding area.20
Notwithstanding, our results suggest that color vision
may be more specific than low-contrast acuity or visual
field sensitivity regarding the potential to anticipate eventual retinal axonal loss.
Although other measures of visual function were
not significant in the multivariate analysis that included
color vision, they were significant univariate predictors of
RNFL loss. The absence of an independent relationship
between the other measures of visual function and final
Volume 70, No. 6
Henderson et al: Axonal Loss after Optic Neuritis
RNFL loss is probably explained by the stronger relationship of a related visual function measure—color vision—
with final axonal loss. In general, our findings suggest a
link between the extent of visual impairment during the
acute episode and later axonal loss. This association is despite a relatively weak relationship between acute visual
impairment and eventual visual outcome in optic neuritis.21–23 A possible explanation for this may be that vision
improves with central adaptation despite axonal loss.24
In the univariate analysis, the significant associations of low-contrast visual acuity measures at baseline
and 3 months with final RNFL loss are particularly noteworthy; low-contrast visual acuity may be a practical
alternative to the FM-100 Hue test in clinical trial settings, as the time required is significantly less than for the
FM-100 Hue test. The 1.25% Sloan chart has floor effects
that may have reflected the lower univariate significance
level compared to the 5% and 25% Sloan measures at
baseline. The 2.5% contrast chart was not tested in this
study but has been used as an endpoint in other studies.
The VEP in optic neuritis often demonstrates one of
the classical pathophysiological features of demyelination,
that of prolongation of conduction time, with relatively
preserved conducted waveforms.25–27 The prolonged
latency is thought to reflect slowing of conduction velocity
through the demyelinated segment of optic nerve,28,29
although it may also reflect the effects of inflammatory
mediators such as nitric oxide.30 The remainder of the nerve
should conduct normally, and the extent of the demyelinating lesion may be inferred by the magnitude of the conduction delay. We found that the extent of whole field VEP
P100 wave delay observed at baseline and at 3 months was
independently associated with final loss of axons in the retina, and by inference, in the optic nerve.
The association between baseline VEP latency and
eventual RNFL loss could be explained by a greater propensity for acute inflammation-mediated axonal damage,
including transection, when the axon is more extensively
demyelinated and thus exposed to acute inflammatory
mediators.31 Demyelination has been shown to make
axons vulnerable to a number of toxic compounds,
including nitric oxide,32 which may induce mitochondrial dysfunction33 and by so doing, render axons vulnerable to an influx of calcium following reversal of the
sodium–calcium exchanger.34
The association of prolongation of VEP latency at
3 months with greater RNFL loss suggests that demyelinated axons remain vulnerable after the initial stages after
the inflammatory insult. A potential explanation for this
may be changes in sodium channel distribution and
function in demyelinated axons. Sodium channels may
December 2011
be distributed along the internodal axonal membrane
with a beneficial effect of restoring nerve conduction and
hence function. However, there may also be greater
influx of sodium into the axon with increased intracellular sodium that in turn is followed by reversal of the
sodium–calcium exchanger and increasing intracellular
calcium, leading to axonal death.34
Our observations of the putative relationship between
the VEP latency and axonal loss encourage the notion of
seeking therapeutic interventions that are effective in protecting demyelinated axons against excitotoxic inflammatory
mediators during acute inflammation and/or that enhance
early remyelination and thereby avoid a vulnerable phase
for the postinflammatory demyelinated axon.
In the final analysis, none of the MRI markers used
in this study had a significant independent association
with axonal loss. This may be because the currently available MRI markers are relatively nonspecific and/or insensitive in capturing the microstructural pathology that
occurs in the symptomatic optic nerve lesion. There was
a univariate association between the length of the gadolinium-enhanced lesion and RNFL loss, suggesting that
disruption of the blood–brain barrier (BBB) and associated inflammation are a weak contributory factor in the
development of RNFL loss, but the stronger and independent association with VEP latency suggests that
demyelination was more important than inflammation
per se. A caveat, however, is that certain types of inflammation other than BBB disruption may not be reflected
on gadolinium-enhanced MRI.
Notably, we did not find a significant independent
association between baseline or 3-month optic nerve MTR
and final RNFL thickness (although there was a trend to
significance between baseline lesion MTR and RNFL
loss). MTR reflects myelin content, and can decline due
to both demyelination and axonal loss.35 In a serial study
of MTR following optic neuritis, Hickman and colleagues14 found a correlation between time-averaged MTR
and VEP, but no association between MTR and VEP in
cross-sectional analyses. It is possible that optic nerve
MTR during the first few months following an episode of
optic neuritis reflects both functionally intact myelin and
nonfunctional myelin breakdown products that have not
yet been cleared from the optic nerve; in the serial study
of Hickman and colleagues, MTR gradually declined to a
nadir over approximately 8 months, which may represent
the gradual clearing of myelin breakdown products rather
than ongoing loss of functional myelin.
A limitation of the study is the use of time-domain
OCT to study RNFL thickness; newer frequency domain
OCT systems provide higher resolution and should
of Neurology
Sharpe JA, Sanders MD. Atrophy of myelinated nerve fibres in the
retina in optic neuritis. Br J Ophthalmol 1975;59:229–232.
Hickman SJ, Brex PA, Brierley CM, et al. Detection of optic nerve
atrophy following a single episode of unilateral optic neuritis by
MRI using a fat-saturated short-echo fast FLAIR sequence. Neuroradiology 2001;43:123–128.
Hickman SJ, Brierley CMH, Brex PA, et al. Continuing optic nerve
atrophy following optic neuritis: a serial MRI study. Mult Scler
Costello F, Coupland S, Hodge W, et al. Quantifying axonal loss
after optic neuritis with optical coherence tomography. Ann Neurol 2006;59:963–969.
Klistorner A, Arvind H, Nguyen T, et al. Axonal loss and myelin in
early ON loss in postacute optic neuritis. Ann Neurol 2008;64:
Henderson APD, Altmann DR, Trip AS, et al. A serial study of retinal changes following optic neuritis with sample size estimates for
acute neuroprotection trials. Brain 2010;133:2592–2602.
Ferris FL, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts
for clinical research. Am J Ophthalmol 1982;94:91–96.
Farnsworth D. The Farnsworth-Munsell 100-hue and dichotomous
tests for color vision. J Opt Soc Am 1943;33:568–578.
Kinnear PR, Sahraie A. New Farnsworth-Munsell 100 hue test
norms of normal observers for each year of age 5-22 and for age
decades 30-70. Br J Ophthalmol 2002;86:1408–1411.
Kinnear PR. Proposals for scoring and assessing the 100-Hue test.
Vision Res 1970;10:423–433.
Hickman SJ. A serial MRI study following optic nerve mean area in
acute optic neuritis. Brain 2004;127:2498–2505.
Hickman SJ, Toosy AT, Jones SJ, et al. Serial magnetization transfer imaging in acute optic neuritis. Brain 2004;127:692–700.
Hickman SJ, Toosy AT, Miszkiel KA, et al. Visual recovery following acute optic neuritis—a clinical, electrophysiological and magnetic resonance imaging study. J Neurol 2004;251:996–1005.
Brusa A, Jones SJ, Plant GT. Long-term remyelination after optic
neuritis: a 2-year visual evoked potential and psychophysical serial
study. Brain 2001;124:468–479.
Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for
multiple sclerosis: 2005 revisions to the ‘‘McDonald Criteria.’’ Ann
Neurol 2005;58:840–846.
Beck RW, Cleary PA, Anderson MM, et al. A randomized, controlled trial of corticosteroids in the treatment of acute optic neuritis. N Engl J Med 1992;326:581–588.
Evangelou N, Konz D, Esiri MM, et al. Size-selective neuronal
changes in the anterior optic pathways suggest a differential susceptibility to injury in multiple sclerosis. Brain 2001;124:1813–1820.
Silverman SE, Hart WM, Gordon MO, Kilo C. The dyschromatopsia of optic neuritis is determined in part by the foveal/perifoveal
distribution of visual field damage. Invest Ophthalmol Vis Sci
Beck RW, Cleary PA. Recovery from severe visual loss in optic
neuritis. Arch Ophthalmol 1993;111:300.
Beck RW, Cleary PA, Backlund JC. The course of visual recovery
after optic neuritis. Experience of the Optic Neuritis Treatment
Trial. Ophthalmology 1994;101:1771–1778.
Trip SA, Schlottmann PG, Jones SJ, et al. Retinal nerve fiber layer
axonal loss and visual dysfunction in optic neuritis. Ann Neurol
Beck RW, Gal RL, Bhatti MT, et al. Visual function more than 10
years after optic neuritis: experience of the optic neuritis treatment trial. Am J Ophthalmol 2004;137:77–83.
Frisén L, Hoyt WF. Insidious atrophy of retinal nerve fibers in multiple sclerosis. Funduscopic identification in patients with and
without visual complaints. Arch Ophthalmol 1974;92:91–97.
Jenkins TM, Toosy AT, Ciccarelli O, et al. Neuroplasticity predicts
outcome of optic neuritis independent of tissue damage. Ann
Neurol 2010;67:99–113.
Halliday A. Evoked potentials in clinical testing. 2nd ed. Edinburgh, UK: Churchill Livingstone, 1993.
improve the accuracy of point estimates of effect in
future similar studies. However, we have achieved a coefficient of variation for repeated measures of 2.5 to 3.5%
with time domain OCT,36 which we feel is sufficient to
enable meaningful interpretation of our study findings.
The identification of reliable early predictors of
axonal loss will be helpful in selecting patients who are
most suitable for inclusion in trials of potential neuroprotective or neuroreparative agents in acute optic neuritis. Further studies are warranted to investigate whether
impaired color vision and prolonged VEP latency can be
confirmed as markers of a higher risk for axonal loss and
whether such measures may enhance the design of future
clinical trials that aim at preventing axonal loss following
optic neuritis.
The NMR Research Unit is supported by the MS Society
of Great Britain and Northern Ireland and the Department
of Health’s Comprehensive Biomedical Research Centre at
University College London Hospitals. P.G.S. was supported by the Guide Dogs for the Blind Association.
Potential Conflicts of Interest
SJJ has nothing to declare. DFG-H’s was paid a consulting
and lecture fee by Carl Zeiss Meditec, as well as a lecture fee by
OptoVue. A.P.D.H.: speaking fees, Novartis, Biogen Idec,
Bayer Schering Pharma. S.A.T.: travel expenses, Teva
Pharmaceutical Industries Limited. K.A.M.: consultancy,
Biogen Idec for image analysis for phase 3b multicentre drug
trial in subjects with RRMS, Novartis for image analysis in
multicentre drug trial in patients with PPMS. G.T.P.:
employment, NHS. D.H.M.: board membership, MS Trials
Advisory Board—Biogen Idec, MS Trials Advisory Board—
Bayer Schering, MS Trials Advisory Board—GlaxoSmithKline, MS Trials Advisory Board—Novartis, consultancy,
Biogen Idec, GlaxoSmithKline, Novartis; grants/grants
pending, MS Society (UK), Biogen Idec, Novartis, GlaxoSmithKline, NIHR, Genzyme, National MS Society, UK
Stem Cell Foundation; speaking fees, GlaxoSmithKline,
Novartis, Biogen Idec, National Institute of Health; royalties,
McAlpines Multiple Sclerosis, 4th Edition; travel expenses,
Biogen Idec, GlaxoSmithKline, Consortium of MS Societies.
Volume 70, No. 6
Henderson et al: Axonal Loss after Optic Neuritis
Halliday AM, McDonald WI, Mushin J. Delayed visual evoked
response in optic neuritis. Lancet 1972;1:982–985.
Halliday AM, McDonald WI, Mushin J. Visual evoked response in
diagnosis of multiple sclerosis. Br Med J 1973;4:661–664.
McDonald WI, Sears TA. Effect of demyelination on conduction in
the central nervous system. Nature 1969;221:182–183.
Hall JI. Studies on demyelinated peripheral nerves in guinea-pigs with
experimental allergic neuritis. A histological and electrophysiological
study. II. Electrophysiological observations. Brain 1967;90:313–332.
Redford EJ, Kapoor R, Smith KJ. Nitric oxide donors reversibly
block axonal conduction: demyelinated axons are especially susceptible. Brain 1997;120:2149–2157.
Trapp BD, Peterson J, Ransohoff RM, et al. Axonal transection in
the lesions of multiple sclerosis. N Engl J Med 1998;338:278–285.
December 2011
Smith KJ, Kapoor R, Hall SM, Davies M. Electrically active axons
degenerate when exposed to nitric oxide. Ann Neurol 2001;49:
Dutta R, McDonough J, Yin X, et al. Mitochondrial dysfunction as
a cause of axonal degeneration in multiple sclerosis patients. Ann
Neurol 2006;59:478–489.
Trapp BD, Stys PK. Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis. Lancet Neurol 2009;8:280–291.
Schmierer K, Scaravilli F, Altmann DR, et al. Magnetization transfer
ratio and myelin in postmortem multiple sclerosis brain. Ann Neurol 2004;56:407–415.
Henderson AP, Trip SA, Schlottmann PG, et al. An investigation of
the retinal nerve fibre layer in progressive multiple sclerosis using
optical coherence tomography. Brain 2008;131:277–287.
Без категории
Размер файла
127 Кб
loss, factors, optics, associates, axonal, neuritic, early
Пожаловаться на содержимое документа