Participation in this open-label, single-cross-over study was offered to patients with clinically definite 19, relapsing-remitting MS, followed in a natural history study with monthly Gd-enhanced MRI. Patients could enter that study no earlier than three months after the last relapse or course of IV-MP. In each individual patient, scans were performed at the same time of the day, and blood was drawn from an antecubital vein before each scan. We first selected those patients who had at least one Gd-enhancing lesion on the first three scans, and continued monthly visits for another three months. Final inclusion in the treatment trial required that the first six monthly scans showed at least three Gd-enhancing lesions (i.e. a mean of 0.5 Gd+ lesions per scan). Further inclusion criteria were: one or more relapses during the two years before the first MRI scan, contraindication to immunosuppressive or immunomodulatory therapy or decline of these therapies by the patient, and no contraindications to glucocorticoid treatment. Upon inclusion, monthly MRI and blood sampling was continued until month 12. Treatment started after the MRI at month 6 (see below).

The study was performed and funded by the participating institutions, with no external funding. It was approved by the ethics committee of the Bavarian State Board of Physicians (Bayerische Landesärztekammer). Patients gave written informed consent.

Twenty-five patients were screened by three monthly MRI scans. Ten patients fulfilled the inclusion criteria, but we excluded one patient with a very high baseline activity to reduce the risk of regression to the mean. See table 1 for clinical and demographic characteristics of the finally included nine patients, and figure 1 for a study flow chart. None of the patients had ever been treated with immunosuppressive or immunomodulatory drugs; mean annualized relapse rate during the entire disease duration was 0.68, and during the two years prior to study inclusion, 0.8.

Once every month, brain MRI was performed on the same 1.5T Signa Echospeed (GE Medical Systems), including axial T1-weighted sequences (TR = 640 ms, TE = 14 ms, 4 mm thick, 1 mm gap) as well as a proton density and T2 weighted dual fast spin echo sequence (TR = 3600 ms, TE = 13 and 91 ms, resp.) and a FLAIR sequence (TR = 10 s, TE = 133 ms, TI = 2200 ms) inherently coregistered with the T1 axial scans. Axial and coronal (3 mm thick, 1.5 mm gap) T1-weighted sequences were acquired after i.v. injection of 0.1 mmol/kg gadolinium chelate (Omniscan(R)), with the interval between injection and first axial scan standardized at 6 minutes.

The number of Gd-enhancing lesions on brain MRI was determined by a blinded radiologist (E.S.) on the scans presented in random order, with the date of acquisition masked. For each patient, the mean number of active lesions per scan was calculated for baseline and treatment period. To determine lesion volume and BFV, a neural network based image analysis system was used as described 20. Briefly, data preprocessing included manual delineation of the intracranial cavity contents, automatic grey value correction (coil inhomogeneity and interslice differences) and extraction of a small training data subset by interactive contour tracing of regions representing the structure classes: normal appearing grey matter (GM), normal appearing white matter (WM), cerebro-spinal fluid (CSF), and MS lesion. A hierarchical vector quantization approach was then deployed to segment the multi-dimensional image data into five classes: GM, WM, CSF, lesions and other (comprising meninges, vessels). The BFV was determined as the ratio of pixels belonging to either brain, i.e. GM, WM or MS lesions, to those belonging to brain or CSF. CSF segmentation was based on the image information obtained by all acquired MRI sequences; MS-lesions were identified based on T2, PD, and FLAIR data. The segmentation performance compared favorably (higher retest stability) with the more standard region growing segmentation algorithm for T2 lesion load (T2-LL) and angle image 21 for BFV 20. T2-LL and BFV were calculated for four time points (start and end of baseline: months 1 and 6; and four weeks after the first and the last MP infusion: months 7 and 12).

From month six to eleven, patients received 500 mg MP i.v. after each MRI scan, followed by an oral tapering dose of 40, 20 and 10 mg MP for one day each to prevent any potential temporary glucocorticoid deficiency.

Blood was collected, between 15:00 and 18:00 in the afternoon and always before the MRI scan, into tubes containing EDTA and trasylol, and plasma was stored at -80 degC until processing. sTNF-RI, sTNF-RII, IL-1ra and sVCAM1 were quantified by commercial enzyme-linked immunosorbent assay (ELISA) kits (s-TNF-RI and RII: Biosource Europe, Belgium; IL-1ra and VCAM-1: R&D Systems, Minneapolis, USA). Hormone concentrations were measured using commercial radioimmuno (RIA) or chemiluminescence assays (cortisol: RIA, DRG Instruments, Germany; ACTH: RIA, and prolactin: Chemiluminescence Immunoassay, both Nichols Institute Diagnostics, San Juan Capistrano, CA, USA).

The dexamethasone-CRH test (Dex-CRH test) was performed, and values were derived, as described 13, once before the start of treatment, and once between months 11 and 12.

In addition to a full physical and neurological examination including EDSS 22, blood chemistry (including calcium, phosphate and vitamine D3), blood count, thyroid function tests, urinanalysis and pyrrolidine cross-links in urine were determined monthly. At baseline and after six months of treatment, patients underwent the short ACTH test (i.v. injection of 200 microgram tetracosactid [Synacthen(R), Novartis] at 08:00, with determination of plasma cortisol concentration before and 60 minutes after injection) and CT densitometry of the lumbar spine. All procedures were performed on an outpatient schedule.

The mean number of Gd enhancing lesions per monthly MRI scan was defined as the primary outcome measure. In accordance with published guidelines and sample size estimates 1718, we planned to enroll ten patients. Secondary outcome variables were T2 lesion volume and monthly plasma hormone concentrations. Serum immune variables and results of neuroendocrine tests were analysed in an exploratory fashion. Number of relapses, EDSS, routine laboratory results and bone mineral density measures were safety variables, with study termination criteria predefined for the laboratory results. Brain fractional volume (BFV) was determined to control for global changes in brain volume, potentially resulting in false results of T2 volume analyses; BFV was not an outcome measure.

According to the data structure, the effect of treatment on the number of active lesions, on hormone and cytokine receptor concentrations was analyzed using the non-parametric Wilcoxon matched-pairs signed-ranks test and a one-factorial univariate analyses of variance (ANOVA) with repeated measures design. The continuous dependent variables were transformed with the ln-transformation [Y = ln (x+1)] before ANOVAs. Associations over time between the number of active lesions to plasma hormone concentrations and immunological parameters were analysed by cross-correlation analysis. Changes of T2-LL and BFV, and all other laboratory results were analyzed by MANOVA, and followed by univariate F-tests where appropriate. As nominal level of significance, p = 0.05 was accepted; it was corrected (according to Bonferroni correction procedure) for all post-hoc tests, in order to keep the type I error less or equal to 0.05. Data are given as mean +/- standard error of the mean (SEM).

The comparison group was extracted from the database maintained at the Sylvia Lawry Center for Multiple Sclerosis Research. This database includes clinical and imaging information on patients examined according to clinical trial protocols at separate centers, including untreated control groups from several large clinical trials and natural history studies. Patients were selected based on the following criteria: A diagnosis of relapsing-remitting multiple sclerosis, no current immunosuppressive or immunomodulatory treatment, longitudinal Gd-MRI follow-up over twelve months, presence of at least 1 new Gd+ lesion during the first three months, and at least three new Gd+ lesions during the first six months. This yielded a group of 83 patients (55 females, mean age 34.4 +/- 0.8 years, mean duration of disease of 7.5 +/- 0.6 years). The MRI protocols applied 0.1 mmol/kg of gadolinium chelate. The mean interval between injection and post-contrast T1w imaging was 5.7 minutes; the delay was 5 minutes in 69 of the selected 83 patients, one minute in one patient, and between 7.5 and 15 minutes in the remaining patients. Monthly scans over the entire twelve-month period were available for seven patients; among the remaining patients, four to six scans over the first six-month-period were available for 75 patients, and for both six-month-periods for 57 patients. The mean number of new Gd+ lesions was calculated for each individual patient for the first and the second six-month-period and expressed as mean number of Gd+ lesions per scan.

This study is the first to show a favorable effect of methylprednisolone on T2 lesion volume. A potentially confounding effect of the dehydrating properties of high-dose glucocorticoids, leading to pseudoatrophy 2930, can be largely excluded because (i) no accompanying brain volume decrease was seen and (ii) serial MP infusions led to a cumulative decrease of T2 lesion volume. Previous studies over short 233031 and longer periods 7 failed to demonstrate changes in T2 lesion load. In short-term studies of MP in acute exacerbations 233031, an anti-edematous effect can be expected to be particularly prominent, while our patients received a single infusion in the absence of acute relapse. Edema resolution thus appears to contribute less to this change in T2 signal than other histological changes (extent of gliosis, macrophage infiltration or remyelination).

The parallel time-course of prolactin and Gd-enhancing lesions during both baseline and treatment with repeated steroid infusions is intriguing with respect to a possible role of prolactin in the autoimmune process. Elevated plasma levels of prolactin have been described in MS relapse, with return to normal concentrations upon recovery 16. In other studies, plasma levels were higher in MS patients than in controls, albeit in the normal range 32, or normal 3334. While the role of prolactin in human immune function is a matter of debate (e.g., 3536), experimental data appear to support an immunostimulatory effect of prolactin 373839. Elevated levels of endogenous or therapeutic glucocorticoids are associated with clinically relevant immunosuppression 40 through several mechanisms, and at the same time reduce expression and secretion of prolactin 4142. A mere side-effect of steroid administration is unlikely given the significant cross-correlation with enhancing lesions during the baseline period. It remains to be determined if lower prolactin during treatment contributes to the reduction of inflammatory disease activity, or if blunted inflammation leads to a reduced level of general stressors which non-specifically stimulate prolactin secretion.

The panel of immunological analyses performed in this study was chosen based on specific hypotheses rather than being designed as a comprehensive immunological assessment of pulsed corticosteroid treatment. Focusing on these selected parameters, we did not observe significant changes associated with monthly IV-MP. However, this does not argue against a lasting immunological effect of pulsed steroids, which potentially could be measured by additional markers.

We are aware that the small number of patients studied and lack of long-term data prohibit direct clinical consequences from this study and we do not suggest pulsed corticosteroids as routine treatment. However, our study demonstrates an anti-inflammatory effect with almost no relevant short-term side effects, and no indication for adrenal insufficiency or osteopenia, in line with recently published data 43. Our results thus support further research 17.