Long-term Brain Magnetic Resonance Imaging Changes After Optic Neuritis in Patients Without Clinically Definite Multiple Sclerosis FREE

At the time of monosymptomatic optic neuritis, magnetic resonance imaging (MRI) of the brain often demonstrates white matter T2 signal abnormalities.^{1 - 6} The presence of such lesions increases the probability that the patient will develop additional neurological manifestations sufficient for a diagnosis of clinically definite multiple sclerosis (CDMS).⁷ Although longitudinal changes on MRI in patients who have CDMS have been well studied,⁸ there are limited data on long-term MRI changes in patients who do not develop CDMS.

Continued follow-up of the cohort of participants enrolled in the Optic Neuritis Treatment Trial (ONTT) has provided the opportunity to evaluate brain MRIs 10 to 14 years after the onset of optic neuritis in patients who have not developed CDMS. The objective of this study was to determine the proportion of such patients who manifest new brain MRI lesions on follow-up MRIs.

Between July 1, 1988, and June 30, 1991, 457 patients with acute optic neuritis were enrolled in the Optic Neuritis Treatment Trial. The study protocol, including the informed consent process, has been reported in detail in previous publications.^{9 - 14}

At study enrollment (1988-1991), brain MRIs were performed using a standardized protocol and graded at a central reading center.¹⁵ Between March 23, 2001, and November 7, 2002, patients still in follow-up were asked to consent to have a follow-up MRI if they had a gradable baseline MRI and had not developed CDMS. A follow-up MRI was performed for 108 (71%) of the 153 patients who met these criteria. The frequency of baseline MRIs showing 1 or more lesions did not differ significantly between the 45 eligible patients who did not have a follow-up MRI and those who did have a follow-up MRI (31% vs 32%).

At baseline, most imagings were performed on a 1.5-T MRI machine with 5-mm-thick T2-weighted axial segments with a 2.5-mm gap. At follow-up, the MRI protocol included 3-mm-thick segments with no gap. A fluid-attenuated inversion recovery sequence was used in the follow-up but not in the baseline MRIs.

The follow-up MRIs were independently read by the same 2 neuroradiologists (J.A. and F.R.M.), who graded the baseline MRIs using similar methods.¹⁵ The classification system, consisting of 5 grades, was identical to that used for baseline MRIs. Grade 0 indicated a normal MRI; grade 1, changes not specific for demyelinative disease; and grades 2 through 4, changes suggestive of demyelination, with the severity increasing with higher grades. Subclasses were established within grades 1 through 4. To be considered a lesion, the signal abnormality had to be at least 3 mm. Smaller lesions were called “punctate.” A direct comparison was made with the baseline MRIs to determine if new lesions were present on the follow-up MRI that were absent on the baseline MRI.

Analyses were conducted using SAS version 8.2 statistical software (SAS Institute Inc, Cary, NC). For comparisons of patients with and without new or additional lesions, categorical variables were compared with the Fisher exact test, and continuous variables with an independent samples t test. The association between the number of baseline lesions and the presence of new lesions in patients with baseline lesions (≥3 mm) was assessed by logistic regression. All reported P values are 2-tailed.

The 108 patients (22% male, 78% female) had a mean age of 32.7 years at the time of the baseline MRI (age range, 18.2-46.0 years) and 44.8 years at the time of the follow-up MRI (age range, 30.9-59.4 years). The mean time between the baseline and follow-up MRIs was 12.2 years (range, 10.2-14.5 years). At baseline, 61 (56%) of the 108 patients had a normal MRI, 35 patients (32%) had at least 1 lesion (≥3 mm), and 12 (11%) patients had 1 or more punctate lesions but no lesions 3 mm or larger.

On the follow-up MRI, at least 1 new lesion (≥3 mm) was seen in 27 (44%) of the 61 patients who had no baseline lesions (Table 1). Among patients who had no baseline lesions, the 34 patients who remained lesion free (no lesions ≥3 mm) were similar to the 27 patients who developed lesions in sex (female, 76% vs 74%; P = .99) and race/ethnicity (white, 88% vs 78%; P = .31) but tended to be younger (mean age at randomization, 30.8 vs 34.2 years; P = .07). The presence of at least 1 lesion (≥3 mm) on the follow-up MRI was not associated with whether the optic disc was swollen or normal during the optic neuritis episode at baseline (40% vs 48%, P = .61).

On the follow-up MRI, at least 1 new lesion (≥3 mm) was seen in 26 (74%) of the 35 patients with 1 or more lesions at baseline (Table 2). The number of baseline lesions did not predict whether new lesions were present on the follow-up MRI (P = .69). The frequency of a new lesion was 74% when 1 or 2 lesions were present on the baseline MRI and 75% when more than 2 lesions were present on the baseline MRI. There were no demographic or clinical characteristics of the 9 patients who did not develop new lesions that distinguished them from the majority who developed new lesions.

Twelve patients had punctate lesions (<3 mm) as the only abnormalities on the baseline MRI; 7 patients had 1 punctate lesion, and 5 patients had 2 or more. Nine of these 12 patients were found to have at least 1 lesion that was 3 mm or larger on the follow-up MRI: 4 had 3 or more nonovoid-nonperiventricular lesions and 5 had 3 or more lesions at least 1 of which was periventricular.

Magnetic resonance imaging has taken on an increasingly important role in the diagnosis and monitoring of patients with multiple sclerosis (MS).¹⁶ Current MS diagnostic criteria following a monosymptomatic presentation incorporate changes in serial MRI findings as documentation of dissemination in time.¹⁷ Accordingly, the long-term MRI characteristics of patients with monosymptomatic optic neuritis who do not develop MS on clinical grounds are of great interest.

Among 61 patients with a normal baseline MRI who had not developed clinical evidence of MS after 10 years, 27 (44%) exhibited at least 1 new 3-mm or greater lesion on follow-up brain MRIs. Subclinical demyelination is the most logical explanation for the new MRI findings within this relatively young population although, for some of the patients, it is possible that the lesions were present at the time of the initial MRI but were undetected owing to the MRI technique used. However, the fact that 34 patients (56%) did not develop clinical signs or MRI evidence of demyelination after 10 years suggests that there are many cases of optic neuritis that may be unrelated to MS.

Among the 35 patients whose baseline MRI showed at least 1 T2-weighted lesion 3 mm or larger, 26 patients (74%) developed at least 1 new lesion 3 mm or larger on follow-up MRI in the absence of a clinical diagnosis of MS. This phenomenon merely underscores the well-known dissociation between MRI findings and clinical expression of MS. The fact that an abnormal baseline MRI was more likely to show additional lesions than a normal baseline MRI emphasizes the predictive value of the initial MRI. The presence of even a single lesion predicted the development of further lesions. However, the development of new lesions does not necessarily indicate that the patient will develop clinical signs of MS even after 10 years.

Among the 12 patients with only punctate T2-weighted hyperintensities on baseline MRI, 9 patients (75%) exhibited changes on long-term MRI, a frequency similar to that in the patients with at least 1 lesion 3 mm or larger on baseline MRI. This finding suggests that a focal signal abnormality of any size may predict that additional signal abnormalities will occur in this population.⁷

These data have several limitations. Magnetic resonance imaging technology continues to advance, and multicenter, serial MRI studies are often forced to compare MRIs obtained with different magnets, field strengths, and protocols. Repositioning errors on serial imaging is also a source of potential difference between baseline and follow-up MRIs. The use of higher-field-strength magnets and smaller-slice-thickness MRIs at follow-up may have overestimated MRI changes over time. However, changes in these parameters appear to affect the assessment of lesion volume more than lesion numbers.^{18 - 20} In 1 series of patients with monosymptomatic demyelinating syndromes, 27% exhibited asymptomatic spinal cord lesions on MRI.²¹ Imaging confined to the brain would likely produce an underestimation of the total burden of T2-weighted changes over time. It is possible that not all T2-weighted MRI changes over time were the result of demyelination. Vasculopathic risk factors such as advanced age, diabetes mellitus, and hypertension may contribute to T2-weighted MRI lesions, but this seems unlikely to explain a significant portion of the change observed in this relatively young cohort. Despite the limitations on interpretation of the results imposed by the differences in MRI technique from the baseline to the follow-up MRIs, these differences influence only the incidence of new lesions and not our observed proportion of patients who have remained lesion free after 10 years.

These data are unique in reporting the long-term MRI changes following monosymptomatic optic neuritis in the absence of the development of clinical signs of MS. The results support the notion that not all cases of monosymptomatic optic neuritis are necessarily related to MS since a subset of patients manifest neither clincal signs nor MRI evidence of demyelination after more than 10 years of follow-up. In addition, the results indicate thatMRI signal abnormalities may accumulate without causing any clinical manifestations of MS even when the patient is followed up for more than a decade. This information is useful in counseling patients who develop first-episode optic neuritis.

Correspondence: Robin L. Gal, MSPH, Jaeb Center for Health Research, 15310 Amberly Dr, Suite 350, Tampa, FL 33647 (rgal@jaeb.org).

Accepted for Publication: March 24, 2004.

Author Contributions:Study concept and design: Beck, Trobe, Kaufman, Savino, and Smith. Acquisition of data: Beck, Gal, Trobe, Arrington, Bhatti, Brodsky, Buckley, Chrousos, Corbett, Goodwin, Kaufman, Keltner, Kupersmith, Miller, Moke, Orengo-Nania, Nazarian, Shults, and Smith. Analysis and interpretation of data: Eggenberger, Beck, Gal, Xing, Trobe, Arrington, Murtagh, Brodsky, Katz, Kupersmith, Miller, Smith, and Wall. Drafting of the manuscript: Eggenberger, Beck, Gal, Xing, Trobe, Moke, and Smith. Critical revision of the manuscript for important intellectual content: Beck, Gal, Trobe, Arrington, Murtagh, Bhatti, Brodsky, Buckley, Chrousos, Corbett, Goodwin, Katz, Kaufman, Keltner, Kupersmith, Miller, Orengo-Nania, Nazarian, Savino, Shults, Smith, and Wall. Statistical expertise: Beck and Xing. Obtained funding: Beck and Brodsky. Administrative, technical, and material support: Eggenberger, Beck, Gal, Arrington, Murtagh, Bhatti, Kaufman, Keltner, Kupersmith, Miller, Moke, and Smith. Study supervision: Beck, Gal, Trobe, Brodsky, Chrousos, Katz, Kaufman, Miller, Orengo-Nania, and Smith. Ongoing collection of data from participants: Savino.

Funding/Support: This study was supported by cooperative agreement U10 EY09435 from the National Eye Institute, National Institutes of Health, Bethesda, Md.

Writing Committee:Lead authors: Eric R. Eggenberger, DO; Roy W. Beck, MD, PhD; Robin L. Gal, MSPH; Dongyuan Xing, MPH; Jonathan D. Trobe, MD; JohnArrington, MD; F. Reed Murtagh, MD. Contributing authors: M. Tariq Bhatti, MD; Michael C. Brodsky, MD;Edward G. Buckley, MD; Georgia A. Chrousos, MD; James J. Corbett, MD; James A. Goodwin, MD; Barrett Katz, MD; David I. Kaufman, DO; John L. Keltner, MD; Mark J.Kupersmith, MD; Neil R. Miller, MD; Pamela S. Moke, MSPH; Silvia Orengo-Nania, MD; Sarkis Nazarian, MD; Peter J. Savino, MD; William T. Shults, MD; Craig H. Smith, MD; Michael Wall, MD.