Fibrinogen is a soluble 340 kDa dimeric glycoprotein that is synthesized in the liver, secreted into the plasma and able to signal via a number of receptors expressed on cells of the hematopoietic, immune and nervous systems [¹,²]. Apart from its pivotal role in thrombogenesis, inflammation, immune responses and atherogenesis, it is also a prominent acute-phase reactant. Transiently elevated plasma fibrinogen levels have been described in acute infectious diseases, in acute stroke and myocardial infarction; chronically raised plasma fibrinogen levels have been associated with an increased risk for cardiovascular diseases [³]. Whilst the significance of plasma fibrinogen is well established, the determination of fibrinogen in cerebrospinal fluid (CSF) has so far been restricted to descriptions of elevated fibrinogen degradation products in the CSF of patients with subarachnoid haemorrhage, traumatic brain injury and Guillain-Barré syndrome (GBS) [^4-6].

Accumulations of fibrinogen and its degradation products have been demonstrated in the central nervous system (CNS) tissue of stroke, bacterial meningitis (BM), HIV-encephalitis, Alzheimer's disease and multiple sclerosis (MS) patients [⁷]. In MS, the leakage of fibrinogen into the CNS was found to result in microglial activation and plaque formation [⁸]. In experimental allergic encephalomyelitis (EAE), an animal model of MS, a close correlation was found between perivascular fibrin deposition and relapses and, moreover, pathology was ameliorated after administration of anti-coagulants or fibrinogen depletion [⁹,¹⁰].

The aim of this study was to determine the value of fibrinogen as biomarker in plasma and CSF of patients with MS compared to other neurological diseases, and to analyze potential clinical and neuroradiological correlations.

Groups differed significantly in age and sex with an older age and female predominance in the GBS group (Table 1). As expected, patients with acute inflammatory disease of the CNS (BM and VM) had significantly elevated CSF WCC and elevated protein levels. Additionally, the group with BM had a significantly reduced GluR. The MS cohort had pathologically raised intrathecal immunoglobulin production of the IgG subtype and 97.5% had CSF-specific oligoclonal IgG bands (Table 1). Cerebral MRI showed significantly more contrast enhancing lesions (CEL) in RR-MS patients as compared to patients with CP-MS (2.6 ± 6.2 versus 0.5 ± 0.8, number of CEL, mean ± STD, p < 0.001), but significantly less T2 hyperintense lesions (26.9 ± 25.9 versus 46.5 ± 27.3, number of T2 hyperintense lesions, mean ± STD, p < 0.001).

MS patients had significantly lower fibrinogen levels in CSF and plasma compared to patients with BM (both p < 0.001), VM (p < 0.001, p < 0.01) and GBS (both p < 0.01). Additionally, OND patients had lower fibrinogen levels in CSF and plasma compared to patients with BM (p < 0.01; p < 0.05, respectively; Figure 1A and 1B).

The fibrinogen quotient in MS patients was significantly lower than in patients with BM and VM (both p < 0.001; Figure 1C). Qalb was significantly lower in MS than in BM or VM (both p < 0.001) and also lower than in GBS (p < 0.05), whereas there was no difference to OND (Figure 1D). Using data for all patients, Qalb was significantly correlated to the fibrinogen quotient (rho = 0.769, p < 0.001, Figure 1E). The fibrinogen index, which was calculated in order to discriminate between a passive transfer of fibrinogen through the BCSFB and intrathecal production of fibrinogen, showed a non-significant trend towards a lower index in patients with BM when compared to patients with VM and patients with OND. However, the overall statistical significance between groups was p < 0.012 (Figure 1F).

Significant correlations were found between serum CRP and CSF fibrinogen (rho = 0.483, p < 0.001) and plasma fibrinogen (rho = 0.738, p < 0.001; data not shown). Fibrinogen was independent of age for all investigated groups. Fibrinogen levels did not differ for different clinical disease courses or during acute relapses in MS patients when compared to times of remission or chronic progression. Furthermore, CSF, plasma fibrinogen and fibrinogen quotient were independent of the number of T2-hyperintense lesions and the number of CEL on cerebral MRI scans. Patients with BM and VM did not differ in clinical outcome using GOS with respect to their fibrinogen levels. In the GBS group, the fibrinogen status was independent of the electrophysiological form of the disease and the severity of the disease at nadir (data not shown).

This study was designed to investigate a possible involvement of CSF fibrinogen in MS pathogenesis during times of acute inflammation, since vascular abnormalities and accumulation of perivascular fibrin are key findings in MS [⁸]. Thus, the finding of normal fibrinogen values in CSF and plasma of MS patients in our study was unexpected. A possible explanation for this observation could be a restriction of blood-brain barrier (BBB) breakdown to local areas of acute inflammatory demyelination as opposed to a diffuse and widespread BBB breakdown in infectious meningoencephalitis. Although BBB breakdown in active MS plaques enables the deposition of fibrinogen, it is not necessarily associated with a more permeable BCSFB, thereby preventing a detectable increase of fibrinogen in the CSF compartment [¹⁶]. Restricted and focal damage of the BBB in MS may also account for the absence of elevated CSF fibrinogen levels during acute relapses and the lack of correlation to CEL. The high molecular weight of fibrinogen as compared to the lower molecular weight marker gadolinium used for imaging studies may be an additional negative confounder.

However, our study found highly increased fibrinogen levels in the acute inflammatory conditions that are associated with a diffuse alteration in the state of the BCSFB, where it is a highly sensitive marker for measuring acute neuroinflammation. The subsequent altering of the extracellular matrix of the brain parenchyma by fibrinogen has been shown to inhibit neuroregeneration and potentiate inflammation in experimental animal models [⁹,¹⁷]. The finding of elevated CSF fibrinogen holds not only true for the CNS, but also for an acute inflammatory condition of the PNS. The finding of increased CSF fibrinogen levels in GBS is in line with previous studies [⁶] and of particular interest since the impairment of the blood-nerve barrier in this context is irrespective of an infectious agent.

Due to the relatively small number of patients investigated and the retrospective design, our study was not powered to detect other influencing factors on the plasma fibrinogen status [³] and the conclusions may therefore be limited. However, our study has demonstrated elevated levels of CSF fibrinogen in patients with acute inflammatory diseases of the CNS and the PNS, where to our best knowledge influencing factors have not been thoroughly investigated. In these acute inflammatory disorders fibrinogen may enter the CSF directly through the BCSFB or indirectly through the BBB and make an important contribution to inflammatory and regenerative processes. In contrast, CSF fibrinogen is not increased in MS patients, where an apparently intact BCSFB prevents fibrinogen entering the CSF. This feature may be useful to distinguish MS from acute CNS inflammatory diseases.

BBB: blood-brain barrier; BCSFB: blood-CSF barrier; BM: bacterial meningoencephalitis; CEL: contrast enhancing lesions; CSF: cerebrospinal fluid; CNS: central nervous system; CP: chronic progressive; EAE: experimental allergic encephalomyelitis; GBS: Guillain-Barré syndrome; GluR: CSF/serum glucose ratio; GOS: Glascow Outcome Scale; HGS: Hughes Functional Grading Scale; MS: multiple sclerosis; OND: other neurological controls; PNS: peripheral nervous system; Qalb: CSF/serum albumin quotient; RR: relapsing-remitting; VM: viral meningitis; WCC: white cell count.

The authors declare that they have no competing interests.

Conceived and designed the study: RE BK MR. Patient recruitment: RE BK FdP PL AL CG HH TB. Analyzed the data: RE PL WS MS FD MR. Wrote the paper: RE MR. All authors read and approved the final manuscript.

The study was supported by research grants of the Gemeinnuetzige Hertiestiftung (Frankfurt, Germany) and the Interdisciplinary Center for Research and Treatment (IFTZ) of Innsbruck Medical University. The authors thank Wolfgang Loescher for critical review of the manuscript and Kathrin Schanda for excellent technical assistance.