The Effect of Corticosteroids for Acute Optic Neuritis on the Subsequent Development of Multiple Sclerosis

Fifteen clinical centers in the United States enrolled 457 patients from July 1, 1988, through June 30, 1991, according to a protocol reported elsewhere and approved by the investigational review boards^14,16,17. Because the primary objective of this study was to evaluate risk factors for the development of definite multiple sclerosis in patients with isolated optic neuritis, we excluded 65 patients (14.2 percent) from the original cohort because they had been given diagnoses of either definite (35 patients) or probable (30 patients) multiple sclerosis at entry into the trial. Also excluded were two patients found after entry not to have optic neuritis and one patient who dropped out before the base-line neurologic examination. Hence, the study cohort comprised 389 patients.

The criteria for entry into the original study included a diagnosis of acute unilateral optic neuritis with visual symptoms of eight days or less, age between 18 and 46 years, no history of optic neuritis or ophthalmoscopic signs of optic atrophy in the affected eye, no evidence of a systemic disease associated with the optic neuritis, and no previous treatment with corticosteroids for optic neuritis in the other eye.

Treatment Assignments

The randomization scheme used a permuted-block design with a separate sequence for each clinical center. Patients were randomly assigned to one of three treatment regimens: 250 mg of intravenous methylprednisolone every six hours for 3 days followed by 1 mg of oral prednisone per kilogram of body weight per day for 11 days (the intravenous-methylprednisolone group), 1 mg of oral prednisone per kilogram per day for 14 days (the oral-prednisone group), or oral placebo for 14 days (the placebo group). The first two regimens were each followed by a tapering regimen of prednisone consisting of 20 mg on day 15 and 10 mg on days 16 and 18. Treatment was begun within 8 days of the onset of visual symptoms, after a mean (±SD) interval of 5.0 ±1.6 days. The patients in the methylprednisolone group were hospitalized for the three days of intravenous treatment. Oral doses, rounded to the nearest 10 mg, were prescribed as single morning doses. Whereas the patients in the oral-prednisone and placebo groups were not informed of their treatment assignments, those in the intravenous-methylprednisolone group were aware of their assignments.

Base-Line Evaluations and Determination of Outcome

At study entry, the patients underwent ocular and neurologic examinations, visual-function testing, and brain MRI according to standardized protocols. Follow-up neurologic examinations were performed after 6 and 12 months and then yearly. Although they were not formally kept unaware of the patients' treatment-group assignments, the study neurologists were generally not aware of these assignments at the time of the follow-up visits.

The development of probable or definite multiple sclerosis was determined solely on the basis of clinical criteria. A demyelinating attack was defined as an episode of symptoms, documented on examination, that was indicative of a neurologic abnormality attributable to acute demyelination in one or more regions of the central nervous system; the symptoms had to last more than 24 hours and had to be separated from a previous attack by at least four weeks¹⁸. The optic neuritis present at entry into the study was considered to constitute one attack and clinical evidence of one lesion. Definite multiple sclerosis was diagnosed when there was a second attack with a new neurologic abnormality that was confirmed by examination. Probable multiple sclerosis was diagnosed when symptoms consistent with a new demyelinating event occurred for which there was no confirmatory examination. Recurrent episodes of optic neuritis in either eye were not considered in the diagnostic criteria for multiple sclerosis. All records of neurologic examinations were evaluated by a reviewer unaware of the patients' treatment assignments, in order to verify the character and dating of neurologic events. A new attack of optic neuritis was diagnosed if a patient reported new visual loss in either eye that was documented on visual-function testing and verified by a masked review of the records.

Unenhanced MRI scans obtained at study entry for 352 of the 389 patients were classified at a single center with a masked grading system². As previously described,² each signal abnormality was characterized by its size (≥ 3 or <3 mm), location (periventricular or nonperiventricular), and shape (ovoid or nonovoid). Scans showing no signal abnormalities were classified as grade 0; those showing one or more focal signal abnormalities, all of which were either smaller than 3 mm or nonperiventricular and nonovoid, as grade 1; those showing one periventricular or ovoid signal abnormality at least 3 mm in size, as grade 2; those showing two such abnormalities as grade 3; and those showing three or more such abnormalities as grade 4.

Statistical Analysis

The assumptions used in the calculation of sample size have been described elsewhere¹⁴. All reported P values are two-tailed.

The cumulative incidence of the development of definite multiple sclerosis was calculated for each treatment group with use of Kaplan-Meier estimates. Incidence rates were compared by a generalized Wilcoxon test for the first two years of follow-up and, in a separate analysis, for the entire follow-up period (up to four years in the case of some patients)^19,20. Unadjusted rate ratios and test-based confidence intervals for the development of definite multiple sclerosis within two years were based on person-time of follow-up^21,22. Adjusted rate ratios were determined from a proportional-hazards model²³. The assumption of proportional hazards was tested for the treatment groups with use of a time-dependent covariate and found to be appropriate.

Table 1. Table 1. Demographic and Clinical Characteristics of the Patients at Study Entry.

Of the 389 patients included in the study, 134 were randomly assigned to the intravenous-methylprednisolone group, 129 to the oral-prednisone group, and 126 to the placebo group. The number of patients enrolled at each clinical center ranged from 12 to 40 (median, 27). The patients' mean (±SD) age was 31.7 ±6.7 years; 77 percent were women, and 85 percent were white (Table 1).

As reported elsewhere, compliance with medication was excellent and side effects of treatment generally mild^14,24. Two patients in the intravenous-methylprednisolone group had serious adverse effects, an acute psychosis and acute pancreatitis. Both conditions resolved without sequelae.

Missed Visits and Loss to Follow-up

The aggregate rate of completed six-month, one-year, and two-year visits was 91 percent (90 percent in the intravenous-methylprednisolone group, 93 percent in the oral-prednisone group, and 90 percent in the placebo group).

Outcome Assessments

The data on 50 patients were censored without the patients' having completed the two-year visit. This group included 21 patients in the intravenous-methylprednisolone group, 14 in the oral-prednisone group, and 15 in the placebo group. The base-line characteristics of these patients were similar. Subsequent examination of 8 of these patients and telephone contact with 26 others revealed that only 1 patient in each group had symptoms of a neurologic event consistent with multiple sclerosis within the first two years.

Comparisons of Treatment Groups

Table 2. Table 2. Percentage of Patients in Whom Multiple Sclerosis Developed or Who Had New Attacks of Optic Neuritis within Two Years of Study Entry.

Definite multiple sclerosis developed within two years in 50 patients (12.9 percent): 16.7 percent of the placebo group, 7.5 percent of the intravenous-methylprednisolone group, and 14.7 percent of the oral-prednisone group (Table 2). The unadjusted two-year rate ratio for the development of definite multiple sclerosis in the intravenous-methylprednisolone group as compared with the placebo group was 0.43 (95 percent confidence interval, 0.21 to 0.89), and the rate ratio in the oral-prednisone group as compared with the placebo group was 0.86 (95 percent confidence interval, 0.37 to 2.00).

Table 3. Table 3. Estimates of the Effect of Treatment on the Development of Definite Multiple Sclerosis within the First Two Years of Follow-up, According to MRI Grade.

Most of the treatment effect was observed in the patients with abnormal MRI scans at study entry. Among patients with grade 3 or 4 scans, definite multiple sclerosis developed within two years in 35.9 percent of 39 patients in the placebo group, 32.4 percent of 37 patients in the oral-prednisone group, and only 16.2 percent of 37 patients in the intravenous-methylprednisolone group (Table 3). Regardless of treatment assignment, the rate of development of definite multiple sclerosis in patients with grade 0 or 1 MRI scans was so low that therapeutic efficacy could not be determined.

Figure 1. Figure 1. Kaplan-Meier Curves Showing the Cumulative Incidence of Definite Multiple Sclerosis, According to Treatment Group.

The numbers of patients still at risk for definite multiple sclerosis at the beginning of each six-month period are shown below the figure for each treatment group. P = 0.09 and P = 0.12 by the Wilcoxon test for the comparison of the curves for the three groups, evaluated together, for the first two years and the entire follow-up period, respectively. The two-year cumulative incidence of definite multiple sclerosis was lower in the intravenous-methylprednisolone group than in the placebo group (P = 0.03), but the value in the prednisone group did not differ significantly from that in the placebo group (P = 0.54).

Life-table analysis of the interval before the development of definite multiple sclerosis revealed that the two-year cumulative incidence of definite multiple sclerosis was lower in the intravenous-methylprednisolone group than in the placebo group (P = 0.03) but was not significantly different between the oral-prednisone group and the placebo group (P = 0.54) (Figure 1). The difference between the curves for the intravenous-methylprednisolone group and the placebo group appeared to lessen after two years. This trend was also found when the life-table analysis was limited to patients (71 in the intravenous-methylprednisolone group and 79 in the placebo group) who completed three years of follow-up, indicating that this finding was not due to a cohort effect.

Table 4. Table 4. Comparisons of Treatment Groups with Regard to Combined Outcomes in the First Two Years of Follow-up.

In the proportional-hazards model, the two-year adjusted rate ratio for the development of definite multiple sclerosis in the intravenous-methylprednisolone group as compared with the placebo group was 0.34 (95 percent confidence interval, 0.16 to 0.74), and in the oral-prednisone group as compared with the placebo group it was 0.90 (95 percent confidence interval, 0.48 to 1.71). For the combined outcomes of probable or definite multiple sclerosis, definite multiple sclerosis or a new attack of optic neuritis in the other eye, and definite multiple sclerosis or a new attack of optic neuritis in either eye, the rate in the intravenous-methylprednisolone group was consistently lower than that in either of the other groups (Table 4).

New attacks of optic neuritis within two years occurred more frequently in the oral-prednisone group than in the other two groups (Table 2 and Table 4).

Base-Line Covariates

Table 5. Table 5. Unadjusted and Adjusted Rate Ratios for the Development of Definite Multiple Sclerosis within the First Two Years of Follow-up for Base-Line Covariates.

Signal abnormalities of the brain, expressed as the grade of the MRI scans, were the strongest indicator of risk for the development of definite multiple sclerosis within two years (Table 5). Among the 202 patients with a normal or grade 1 scan at entry into the study, definite multiple sclerosis developed in only 5.0 percent, as compared with 24.7 percent of the 150 patients with grade 2, 3, or 4 scans.

Other variables that had high rate ratios for the development of definite multiple sclerosis included a history of optic neuritis in the other eye, ill-defined neurologic symptoms, a family history of multiple sclerosis, and white race, although the 95 percent confidence intervals for the last two of these variables included 1.0. Age and sex were not important indicators of risk.

In this controlled study, patients treated with intravenous methylprednisolone followed by oral prednisone had a reduction in the risk of new clinical manifestations of multiple sclerosis within the next two years, as compared with patients receiving either placebo or oral prednisone. This protective effect was most apparent in the patients at highest risk for multiple sclerosis -- namely, those with multiple focal brain MRI abnormalities. Patients treated with oral prednisone alone, in dosages typical of those prescribed in clinical practice, had a higher rate of new attacks of optic neuritis than patients in the other two groups.

Inasmuch as the clinical and demographic characteristics of our patients conformed to the profile accepted for acute, isolated optic neuritis^25,26 and the two-year risk of definite multiple sclerosis in our placebo group was approximately the same as in previous studies,^8-12 we believe our findings can be generally applied to patients with this diagnosis. Interpretation of the results must, however, be tempered by the fact that evaluating the randomized treatments with regard to the development of multiple sclerosis was not the primary study objective, and 14.2 percent of the originally randomized patients were not part of the current analysis because they were diagnosed as having multiple sclerosis at entry.

Nevertheless, the probability is high that the strong protective effect of intravenous methylprednisolone followed by oral corticosteroids found in this study is real and unlikely to have resulted from chance, confounding, or bias. The magnitude of the treatment effect was consistent among most of the base-line covariates, and the direction of the effect was consistent among clinical centers (data not shown). Both unadjusted and adjusted analyses yielded similar estimates of effect and similar confidence intervals. Missing data for patients who did not complete the two years of follow-up did not appear to be an appreciable source of bias.

Although the patients in the intravenous-methylprednisolone group were aware of their treatment, it is doubtful, for three reasons, that this awareness caused bias. Visual recovery was excellent in all three treatment groups, and there was no previous information to suggest that corticosteroid treatment would reduce the rate of new attacks of multiple sclerosis. Thus, neither the patients nor the neurologists evaluating them had reason to believe that intravenous methylprednisolone treatment was superior. Furthermore, a survey of the study neurologists at the conclusion of the two years of follow-up, while they remained unaware of the results, found no sentiment to indicate that that treatment would be superior. Second, the neurologic assessment was recorded in a structured format that resulted in a standardized examination. Third, patients were classified as having definite multiple sclerosis only after a masked independent review.

It is difficult to explain why such a strong and long-lasting protective effect against the development of new demyelinating events would result from a 3-day course of intravenous methylprednisolone followed by an 11-day course of oral prednisone, particularly since a 14-day course of oral prednisone alone had no such effect. Since oral prednisone is well absorbed and produces physiologic effects similar to those of methylprednisolone, the efficacy of intravenous therapy may be ascribed to the higher dosage. The explanation for the prolonged protective effect of the high-dose intravenous therapy is uncertain, considering that the antiinflammatory effects of corticosteroids on the immune response¹⁵ and on reduction of the permeability of the blood-brain barrier²⁷ are believed to be short-lived. Treatment administered at or near the onset of clinical disease may have interfered with the elaboration of the immune response, perhaps by causing depletion and delayed reconstitution of activated lymphocyte subgroups.

Four smaller studies^5,28-30 support our finding that signal abnormalities present on brain MRI at the time of the development of optic neuritis markedly increased the likelihood that new clinical signs of multiple sclerosis would develop. In the aggregate, in patients with optic neuritis, clinical evidence of multiple sclerosis developed after a mean follow-up of 0.9 to 4 years in only 4 percent of 80 patients with normal MRI scans, as compared with 30 percent of 108 patients with abnormal scans.

Two other factors that were found in our study to confer increased risk for the development of definite multiple sclerosis -- namely, a history of optic neuritis in the contralateral eye and ill-defined neurologic symptoms -- have not previously been well documented^8-11. The risk associated with previous attacks of optic neuritis in the same eye cannot be determined from our study, because patients with this condition were excluded from entry. Female sex, considered a risk factor for multiple sclerosis in a long-term study of patients with optic neuritis,¹⁰ was not confirmed as such in our study.

On the basis of this trial, the beneficial effect of the intravenous methylprednisolone regimen in reducing the rate of new demyelinating events over a two-year period justifies consideration of this treatment, even though it has only a marginal effect on visual recovery. Although brain MRI may not be needed to diagnose optic neuritis, scanning is valuable in establishing the two-year risk of additional manifestations of multiple sclerosis. Patients with multiple signal abnormalities on brain MRI were those who most clearly benefited from treatment. Because the rate of development of definite multiple sclerosis was so low in patients with normal MRI scans, the value of treating this group could not be assessed.

To our knowledge, apart from the present study, only in the recent controlled trials of interferon beta-1b,^31,32 which reduced the rate of exacerbation over a two-year period in patients with well-established multiple sclerosis, has the natural history of this disease been convincingly altered. Although a reduction in the development of new demyelinating attacks in patients with optic neuritis may not be the same as a reduction in the number of exacerbating events in patients with clearly established multiple sclerosis, the similarities between the two are likely to be greater than the differences. This increases the need to investigate pulsed-dose corticosteroids further, not only in patients with isolated optic neuritis but also in those with established multiple sclerosis.

A complete listing of the members of the Optic Neuritis Treatment Trial Study Group has been published previously¹⁷. In addition to the study authors, the following are the major participants in the study group.

University of South Florida, Tampa: B.J. Sellers, L. Zajac, P. Moke, C. Newman, R. Murtagh, and J. Arrington; George Washington University, Rockville, Md.: J.C. Backlund, D. Kenny, N. Loring, S. Campbell, P. Gilbert, W. Watson, and J. Zablotny; University of California, Davis: J. Keltner, C. Johnson, J. Spurr, and H. Hakim; National Eye Institute, Bethesda, Md.: C. Atwell and R. Mowery; University of Arkansas, Little Rock: M. Brodsky, S. Nazarian, B. Lam, B. Lyon, and S. Cain; Baylor College of Medicine, Houston: L. Rolak, J. McCrary, S. Orengo-Nadia, R. Gross, B. Slight, and M. McMaster; California Pacific Medical Center, Smith-Kettlewell Eye Research Institute, San Francisco: B. Katz and T. Ambrosio; Duke University, Durham, N.C.: E. Buckley, W. Massey, S. Pollock, M. Anderson, and G. Valentine; University of Florida, Gainesville: J. Guy, S. Zam, D. Shamis, and R. Watson; Center for Sight, Georgetown University, Washington, D.C.: G. Chrousos, J. Kattah, E. Burt, and S. Lauber; University of Illinois, Chicago: J. Goodwin, J. Nichols, and E. Sullivan; University of Iowa, Iowa City: S. Thompson, J. Corbett, M. Wall, R. Kardon, and C. Musser; Wills Eye Hospital, Thomas Jefferson University, Philadelphia: P. Savino, R. Sergott, T. Bosley, S. Ward, and M. Devlin; Wilmer Eye Institute, Johns Hopkins Hospital, Baltimore: N. Miller, M. Repka, D. Buchholz, S. Reich, and C. Krich-Putzulo; Kellogg Eye Center, University of Michigan, Ann Arbor: W. Cornblath and C. Caudill; Michigan State University, East Lansing: J. Kokinakis, J. Froehlich, and T. Moore; New York University, New York: F. Warren, A. Addessi, and J. Weinman; Devers Eye Institute, Good Samaritan Hospital, Portland, Oreg.: W. Shults, L. Diehl, R. Wilson, and R. Herndon; University of Washington, Seattle: C. Smith, D. Kuder, C. Wredberg, and P. Ernst.